J. J. K. Daemen Auth., K. Fuenkajorn, J. J. K. Daemen Eds. Sealing of Boreholes and Underground Excavations in Rock

Sealing of Boreholes and Underground
Excavations in Rock
JOIN US ON THE INTERNET VIA WWW, GOPHER, FTP OR EMAIL:

WWW: http://www.thomson.com
GOPHER: gopher.thomson.com
A service of lOOP
FTP: ftp.thomson.com
EMAIL: findit@kiosk.thomson.com
Sealing of Boreholes
and Underground Excavations
in Rock
Edited by
K. Fuenkajorn
Rock Engineering International,
Tucson, USA
and
J. J. K. Daemen
Mining Engineering Department,
University of Nevada, USA
CHAPMAN & HALL

London Weinheim . New York Tokyo
Melbourne Madras
Published by Chapman & Hall, 2-6 Boundary Row, London SEt 8HN, UK
Chapman & Hall, 2-6 Boundary Row, London SEI 8HN, UK
Chapman & Hall GmbH, Pappelallee 3, 69469 Weinheim, Germany
Chapman & Hall USA, 115 Fifth Avenue, New York, NY 10003, USA
Chapman & Hall Japan, ITP-Japan, Kyowa Building, 3F, 2-2-1 Hirakawacho,
Chiyoda-ku, Tokyo 102, Japan
Chapman & Hall Australia, 102 Dodds Street, South Melbourne,
Victoria 3205, Australia
Chapman & Hall India, R. Seshadri, 32 Second Main Road, CIT East,
Madras 600 035, India
First edition 1996
1996 Chapman & Hall
Softcover reprint of the hardcover 1st edition 1996
Typeset in lOj12pt Times by Thomson Press (I) Ltd., Madras
ISBN-l3: 978-94-010-7173-4 e-ISBN-13: 978-94-009-1505-3

001: 10.1007/978-94-009-1505-3
Apart from any fair dealing for the purposes of research or private study, or
criticism or review, as permitted under the UK Copyright Designs and
Patents Act, 1988, this publication may not be reproduced, stored, or
transmitted, in any form or by any means, without the prior permission in
writing of the publishers, or in the case of reprographic reproduction only in
accordance with the terms of the licences issued by the Copyright Licensing
Agency in the UK, or in accordance with the terms of licences issued by the
appropriate Reproduction Rights Organization outside the UK. Enquiries
concerning reproduction outside the terms stated here should be sent to the
publishers at the London address printed on this page.
The publisher makes no representation, express or implied, with regard to
the accuracy of the information contained in this book and cannot accept
any legal responsibility or liability for any errors or omissions that may be
made.
A catalogue record for this book is available from the British Library
Library of Congress Catalog Card Number: 95-71862
~ Printed on acid-free text paper, manufactured in accordance with

ANSljNISO Z39.48-1992 (Permanence of Paper).
Contents
List of Contributors ix
Preface xi
1. Introduction . . . . . . . . . . . . . . . . 1
J. J. K. Daemen
1.1 Background.............. 1
1.2 Sealing Requirements-Rules and Regulations . 1
1.3 Current Practice . . . . . . . . . . . . . . . . . 2
1.4 Recent Borehole Sealing Research-An Introductory
Overview . . . . . . . . . . . . . . . . . 2
1.5 Sealing of Shafts, Tunnels, Mine Adits,
Portals and Drifts . 4
1.6 Book Summary 4
1. 7 The Future . . 7
Acknowledgements . 8
2. Laboratory Performance of Cement Borehole Seals . 9

D. L. South and K. Fuenkajorn
2.1 Introduction . . . . . . . 9
2.2 Conceptual Approach . 10
2.3 Experimental Apparatus . 11
2.4 Experimental Procedures 13
2.5 Analysis . . . . . . . . . 15
2.6 Summary of Test Results . 19
2.7 Design Implications 26
Acknowledgements . . . . . . . 27
3. Strength Parameters of Cement Borehole Seals in Rock . 28

H. Akgiln
3.1 Introduction . . . . . . . . . . . . . . . 28
3.2 Push-Out Test Experimental Procedure. 29
3.3 Push-Out Test Mechanical Interactions. 29
3.4 Finite Element Analysis and Discussion. 31
VI Contents
3.5 Push-Out Test Results and Discussion . 36

3.6 Conclusions and Recommendations 39
Acknowledgements . . . . . . . . . . . . . . 39
4. Dynamic Loading Impact on Cement Borehole Seals . 40

G. S. Adisoma
4.1 Introduction...... 40
4.2 Experimental Methods 42
4.3 Flow Test Results . . . 44
4.4 Dye Injection Test Results 50
4.5 Dynamic Loading Test Results . 52
4.6 Design Considerations. 56
4.7 Discussion .. 60
4.8 Conclusions . 62
Acknowledgements 64
5. Performance of Bentonite
and Bentonite/Crushed Rock Borehole Seals . 65
S. Ouyang and J. J. K. Daemen
5.1 Introduction . . . . . . . . . . . 65
5.2 Materials............. 66
5.3 Permeability Tests and Results. 70
5.4 Piping and Flow of Bentonite. . 82
5.5 Prediction of Bentonite Permeability . 86
5.6 Conclusions and Recommendations 93
Acknowledgements . . . . . . . . . . . . . 95
6. In situ Performance of a Clay-Based Barrier. 96

B. H. Kjartanson, N. A. Chandler and A. W L. Wan
6.1 Introduction............ 96
6.2 The Canadian Nuclear Fuel Waste
Management Program . . . . . . 97
6.3 Evaluation of Buffer Physical Performance. 99
6.4 In situ Experiments . . . . . 100
6.5 Buffer/Container Experiment 102
6.6 Isothermal Test 117
Acknowledgements . 125
7. In situ Hydraulic Performance Tests

of Borehole Seals: Procedure and Analyses . 126
W. B. Greer
7.1 Introduction....... 126
7.2 Proposed Test Schemes. 127
7.3 Seal Tests . . . . . . . . 131
Contents vii
7.4 Conclusions.. 152

Acknowledgements. 156
8. In situ Hydraulic Performance of Cement Borehole Seals . 157

W. B. Greer and D. R. Crouthamel
8.1 Introduction................ 157
8.2 Preliminary Survey for Seal Installation. 157
8.3 Seal Tests at Oracle Ridge Mine. . . . . 158
8.4 Discussion of Testing at Oracle Ridge Mine. 175
8.5 Seal Testing at Superior, Arizona. . . . . . . 178
8.6 Recommendations for Testing In situ Borehole Seals. 181
Acknowledgements . . . . . . . 183
9. Sealing Boreholes in Rock Salt . 184

J. C. Stormont and R. E. Finley
9.1 Introduction........ 184
9.2 General Practice in Sealing Boreholes in Rock Salt. 186
9.3 Rock Salt Properties Relevant to Borehole Sealing. 190
9.4 Materials for Sealing Boreholes in Rock Salt. 201
9.5 Seal System Performance . 211
9.6 Design Considerations 218
Acknowledgement . . . . . 224
10. Design of Underground Plugs. 225

F. A. Auld
10.1 Introduction . . . . . . . 225
10.2 Types of Plugs . . . . . . 226
10.3 Factors to be Considered in Design of Plugs. 227
10.4 Design Calculations . . . . . . . . . . . . 233
10.5 Construction Aspects . . . . . . . . . . . 246
10.6 Plug Sealing and Resistance to Leakage. 249
10.7 Case Studies . . . . . . . . . . . . . 254
10.8 Conclusions and Recommendations . 264
11. Design of Borehole Seals: Process, Criteria

and Considerations . . . . . . . . . 267
J. J. K. Daemen and K. Fuenkajorn
11.1 Introduction . . . . . . . . . 267
11.2 Applications, Objectives and Requirements. 268
11.3 Considerations of Site Characteristics. . . . 270
11.4 General Design Criteria. . . . . . . . . . . 274
11.5 Material Selection and Placement Methods. 276
11.6 Summary.. 279
Acknowledgements . . . . . . . . . . . . . . . . . . 279
Vlll Contents
12. Sodium Bentonite as a Borehole Sealant. 280

1. E. Papp
12.1 Introduction . . . .. 280
12.2 Geographical Origin. 282
12.3 Geological Origin . . 282
12.4 Structure of Sodium Bentonite. 282
12.5 Properties of Sodium Bentonite. 284
12.6 Bentonite as a Borehole Sealant. 285
12.7 Bentonite as an Annular Sealant. 286
12.8 Types of Bentonite Sealants. . . . 287
12.9 Placement of Bentonite . . . . . . 292
12.10 Case History Featuring High-Solids
Bentonite Grout 294
12.11 Conclusions 297
Bibliography . 298
Index .... 325

Contributors
G. S. Adisoma
Independent Mining Consultants, Inc., 2700 East Executive Drive, Suite
140, Tucson, Arizona 85706, USA
H. Akgful
Department of Geological Engineering, Orta Dogu Teknik Universitesi,
Middle East Technical University, Ankara, Turkey
F. A. Auld
I. W. Farmer & Partners, Ltd, Jilland House, 329 Bawtry Road, Doncaster,
DN4 7PB, England
N. A. Chandler
AECL Research, Whites hell Laboratories, Pinawa, Manitoba, Canada ROE
1LO
D. R. Crouthamel
Stone & Webster Engineering Corporation, 245 Summer Street, Boston,
Massachusetts 02210, USA
J. J. K. Daemen
Mining Engineering Department, Mackay School of Mines, University of
Nevada, Reno, Nevada 89557-0139, USA
R. E. Finley
Sandia National Laboratories, Geotechnical Investigations Department,
MS-1325, P.O. Box 5800, Albuquerque, New Mexico 87185, USA
K. Fuenkajorn
Rock Engineering International, 7226 West Rivulet Drive, Tucson, Arizona
85743, USA
W. B. Greer
U. S. Bureau of Reclamation, Water and Power Resources Management
Division, 2800 Cottage Way, Sacramento, California 95825, USA
B. H. Kjartanson
Department of Civil and Construction Engineering, Iowa State University,
Ames, Iowa 50011-3232, USA
x Contributors
S.Ouyang
Industrial Technology Research Institute, Energy and Resources Labora-
tories, Building #24, 195-6 Chung Hsing Road, Section 4, Chutung.
Hsinchu, Taiwan 310, R.O.C.
J. E. Papp
Colloid Environmental Technologies Company, 1350 West Shure Drive,
Arlington Heights, Illinois 60004-7803, USA
D. L. South
Washington State Department of Ecology, Northwest Regional Office,
State of Washington, 3190-160th Avenue, S. E., Bellevue, Washington
98008-5452, USA
J. C. Stormont
Department of Civil Engineering, University of New Mexico, Albuquerque,
New Mexico 87131, USA
A. W.L. Wan
AECL Research, Whiteshell Laboratories, Pinawa, Manitoba, Canada
ROE 1LO
Preface
Sealing of boreholes and underground excavations has not received much

engineering attention until fairly recently. The growing awareness of and
sensitivity to environmental concerns of the technical community as well as
of the public at large has resulted in an increasing recognition of the fact
that these geological penetrations may have an environmental impact. The
issue of possible contamination resulting from migration along boreholes,
adits, shafts or tunnels unquestionably has been raised most forcefully with-
in the context of nuclear waste disposal. Several nuclear waste disposal
programs, notably the Civilian and the Defence programs of the US De-
partment of Energy, the US Nuclear Regulatory Commission and the
Canadian and Swedish radioactive waste disposal programs have conducted
major research efforts aimed at developing adequate seal designs for penet-
rations in host rock formations for high-level nuclear waste repositories.
While a considerable data base has been gathered over the last two decades
or so with regard to the performance of seals, most of the information is
presented in research reports and widely scattered papers in journals and
proceedings of conferences. Hence, the materials are not readily accessible
to potential users such as designers, contractors or regulators who are not
familiar with nuclear waste disposal programs. Although many government
agencies have implemented regulations requiring that unused boreholes and
underground excavations in rock formations be sealed, these regulations
tend to be generic and broad, and rarely allow for taking into account
site-specific conditions. As a result, it is probable that, for example, they are
excessively conservative for some locations and inadequate for others.
We organize and structure the available information on sealing boreholes
and underground excavations in a format that makes it much more readily
accessible. This book presents a comprehensive integrated summary of re-
cent laboratory and in situ experiments conducted to assess the mechanical
and hydraulic performance of the emplaced seals and host rocks. The con-
tents are structured so as to highlight design considerations and recommen-
dations under various aspects of seal and host rock characteristics, and
installation environments. Uncertainties in terms of design, performance
testing and performance predictions are recognized, and lead naturally to a
discussion of remaining research needs.
This book is immediately applicable as a compiled literature summary to
senior college students and research graduates. It is intended to be a
xii Preface
reference source for design professionals and researchers in many disciplines

of engineering and sciences, including mining, geological, petroleum, nu-
clear, civil, sanitary and environmental engineers, as well as geologists,
hydrologists and material scientists. The present testing techniques, ap-
proaches and recommendations will assist graduates and experimentalists
in designing their test schemes to develop new sealants and installation
methods for a specific application or to answer remaining questions on the
sealing of underground space. The data and design considerations presented
will allow engineers to plan and design seals with improved confidence.
K. Fuenkajorn
J. J. K. Daemen
CHAPTER ONE
Introduction
J. J. K. Daemen
1.1 BACKGROUND
Sealing of underground openings has assumed a growing importance with

the recognition that penetrations of geological formations can have a detri-
mental impact on the environment. Boreholes that penetrate aquitards may
allow migration and mixing of groundwaters of different qualities, and may
contaminate aquifers. Poorly sealed or open boreholes may allow prema-
ture and unnecessary depressurization of formations, and may result in
wasting of natural resources. Inadequately sealed mine adits may allow
unacceptable discharges of acid mine drainage.
This book presents an overview of sealing methods that can be used to
prevent or minimize the deterimental effects that may result from leaving
geological pentrations open. The emphasis of this book is to present results
of research investigations that address the difficult question of what level of
sealing performance can be expected from various sealing methods. The
work presented is oriented predominantly towards the sealing of boreholes
in rock. Implications for seal design are given.
1.2 SEALING REQUIREMENTS-RULES AND REGULATIONS

Because of the recognition that leaving boreholes unsealed can have a
variety of detrimental impacts, rules and regulations that govern the pro-
cedures to be followed to close boreholes have been implemented for many
years in many jurisdictions. Among the earliest regulatory requirements
were those imposed by the Texas Railroad Commission on sealing of oil
and gas wells. A prime objective of these regulations was to prevent prema-
ture depressurization of oil and gas formations.
Sealing of Boreholes and Underground Excavations in Rock.

Edited by K. Fuenkajorn and J. J. K. Daemen.
Published in 1996 by Chapman & Hall. ISBN 0 412 57300 8.
2 Introduction
In recent years many countries and states have introduced or strengthened

regulatory requirements for borehole sealing. An exhaustive
set of US regulations and related documents has recently been complied
and published by Baroid Drilling Fluids, Inc. and Environmental Plan-
ning Group, Inc. (1995), who also will provide annual updates of the
regulations. The availability of this documentation should facilitate great-
ly the difficult task of trying to remain updated about the regulatory
requirements.
Regulatory requirements historically have been dominated by very prac-
tical prescriptions in terms of how borehole sealing is to be accomplished.
For reasons that will become very clear later in this book, the obviously
more attractive approach of making the regulations performance oriented
or based remains exceedingly difficult to implement, because testing the
performance of borehole seals remains difficult.
1.3 CURRENT PRACTICE
Current borehole sealing practice depends largely upon the industrial con-
text within which boreholes are drilled. Within the petroleum industry the
use of cementitious seals predominates. In the construction, geotechnical
and water well industries clay, primarily bentonitic, seal materials share a
substantial fraction of the market. While the casual 'sealing' of boreholes
with telephone poles, beer cans and various types of debris may not yet
have been abandoned completely, it certainly no longer is acceptable
practice.
Cementitious sealing of deep oil and gas wells requires sophisticated
emplacement technology. The necessary equipment may be on site as part
of the well completion equipment. Such technology rarely if ever is available
for the sealing of geotechnical or ore exploration holes, or for water well
sealing. Grout pumps may be available for the sealing of such holes. Careful
downhole pumping of cement or bentonite grouts has been widely accepted
as a borehole sealing practice. Bailer emplacement is an acceptable alterna-
tive for cement grouts. Bentonite chips can be dropped in holes containing
standing water, and provide a low permeability seal.
1.4 RECENT BOREHOLE SEALING RESEARCH-AN

INTRODUCTORY OVERVIEW
For nearly two decades borehole sealing research has received a major
impetus, in several countries, within the context of sealing nuclear waste
repositories. While the objectives of these programs may be rather unusual,
these investigations have contributed substantially to the development of
sealing technologies and to the determination of sealing performance that
Recent borehole sealing research 3
are of value in general sealing practice, i.e. in the far more numerous and
mundane applications that dominate sealing requirements.
Sandia National Laboratories (e.g. Hansen et al., 1995) has conducted
and continues to conduct a major investigation of the sealing of penetra-
tions, primarily boreholes and shafts, through rock salt and, more generally,
evaporite formations. This work, some aspects of which are presented in
Chapter 9, has included extensive laboratory and field experiments on the
sealing performance of cementitious, salt and clay seals. The work can be
considered as an extension and continuation of work initiated by Oak
Ridge National Laboratory (e.g. McDaniel, 1980) and continued by the
Office of Nuclear Waste Isolation (e.g. Roy, Grutzeck and Wakeley, 1985).
Sealing of penetrations through salt and evaporites is particularly import-
ant because of the solubility of halite, and because of the numerous penetra-
tions of such formations, whether for mining, waste disposal or oil and gas
production purposes.
Within the context of the Swedish nuclear waste disposal program
Pusch and associates, e.g. Pusch (1983), have conducted extensive studies
on the sealing performance of bentonite, especially of highly compacted
bentonite. These studies have resulted in an exhaustive data basis on the
sealing performance of highly compacted bentonite, and in a dramati-
cally improved understanding of the performance characteristics of ben-
tonite.
The Office of Nuclear Waste Isolation has supported various borehole
sealing studies, conducted, among the principal research groups, at the
Materials Research Laboratory at Pennsylvania State University (e.g.
Roy, Grutzeck and Wakeley, 1985) and at Waterways Experiment Station
of the US Army Corps of Engineers (e.g. Buck, Burkes and Rhoderick,
1981). The main emphasis of this research has been on sealing of evaporite
formations with cementitious materials. The research has continued for
the Waste Isolation Pilot Plant (WIPP) (e.g. Wakeley, Harrington and
Weiss, 1993). IT Corporation, formerly D' Appolonia Consulting engineers
(e.g. D'Appolonia Consulting Engineers, 1981), also has participated ex-
tensively in the salt repository sealing studies (e.g. International Technol-
ogy Corp., 1987). The Canadian nuclear waste disposal program has
conducted investigations on borehole sealing, aspects of which are
summarized in Chapter 6.
The Department of Mining and Geological Engineering at the University
of Arizona has been sponsored by the US Nuclear Regulatory Commission
(NRC) from the late 1970s to the late 1980s to conduct investigations on
borehole and shaft sealing. Several chapters in this book (2-5,7 and 8) are
based on work sponsored by NRC.
The Federal Highway Adiministration has initiated a research study on
the sealing of exploratory boreholes, primarily geotechnical holes in soils,
initial results of which are summarized by Lutenegger and DeGroot
(1993).
4 Introduction
1.5 SEALING OF SHAFTS, TUNNELS, MINE ADITS,

PORTALS AND DRIFTS
Sealing of excavations larger than boreholes, such as mine shafts or adits, or

tunnels, may have a variety of purposes. The level of technology applied for
particular cases tends to be highly variable, depending upon the objectives
of the sealing operations. For many abandoned or to be abandoned mine
entries the prime purpose of closure is to prevent human intrusion, with the
ultimate objective of forestalling accidents that frequently result when
people enter old mines. For many such mine closures the technology tends
to be rather rudimentary, aimed at a minimal physical barrier that will
satisfy the requirements to minimize the risk of penetrations.
At the other extreme are situations where the objective of permanent
closures is to provide a true seal, e.g. tight to both air and water. Historical
examples of this type include primarily major case studies of mine flooding
and of the design of plugs to allow controlling the flooding (e.g. Garrett and
Campbell Pitt, 1961). Current examples tend to be oriented primarily
towards reducing the environmental impact from mine discharges (e.g.
Einarson and Abel, 1990). Design approaches for sealing large excavations,
as well as recent examples, are given in Chapter 10.
1.6 BOOK SUMMARY
Chapter 2 presents the results of flow testing of cement grout seals emplaced
in boreholes in rock. The study was intended primarily to look at the
performance of seals in deep boreholes in strong intact rock. Flow tests
were conducted in triaxial cells, with particular attention given to the influ-
ence of changing lateral stresses on the sealing performance. Detailed atten-
tion also was given in this study to the risk of interface flow, i.e. flow along
the contact between rock and seal cement grout. It was found that for the
materials tested, and under the conditions for which the tests were conduc-
ted, the interface did not usually form a preferential flow path. While drying
of the cementitious seals tested here does induce shrinkage and does in-
crease the hydraulic conductivity and hence allows increased flow, resatura-
tion with water flow reverses the sealing performance noticeably.
Chapter 2 includes a summary of the results of an extensive numerical
study of the flow paths that can be expected around a borehole seal as a
function of the hydraulic conductivity of the seal relative to that of the host
rock. No significant reduction in flow through the seal system is obtained
by reducing the hydraulic conductivity of the seal below about an order of
magnitude less than that of the host rock. Because these results are sum-
marized in dimensionless form, they are valid for seals of any diameter.
They provide guidance for the selection of seal permeability goals, assuming
that the permeability of the host rock is known or can be estimated.
Book summary 5
Chapter 3 briefly summarizes a study of bond strength of cement grout

seals in rock. Bond strength usually is not an issue for cementitious
borehole seals, whose length generally far exceeds the length necessary to
provide adequate bond strength. Bond strength often is a critical design
parameter for shaft and tunnel seals, at least for plugs at great depths that
may be subject to high axial loading. Chapter 3 provides bond strength
values determined experimentally on plugs emplaced in small-diameter
boreholes, and guidance on how these results might be scaled up to larger
diameter sealing plugs.
Chapter 4 summarizes an investigation of dynamic effects on the per-
formance of cement grout borehole seals. It is found that vibrations, even at
considerable acceleration, do not significantly deteriorate sealing perform-
ance. Conversely, allowing cementitious borehole seals to dry can rapidly
enhance their permeability and reduces their bond strength. Of considerable
interest is the observation that these seals tend to recover significantly upon
resaturation. The complex drying/resaturation sequence and effects deserve
further study, because they almost certainly are significant with respect to
the long-term performance of many seals.
Chapter 5 presents the results of experimental and modeling investiga-
tions on the borehole sealing performance of bentonite and mixtures of
bentonite and crushed welded tuff. The incentive for the investigation is the
proposition that benefits may be derived under certain circumstances from
loading bentonite with crushed rock. Such mixture sealants may maintain
the attractive features of bentonite seals, notably their low permeability,
self-healing, durability, swelling deformability and sorptive characteristics.
The addition of crushed rock may have some potential benefits, e.g. decreas-
ing the shrinkage potential during drying, increasing the strength or bearing
capacity and increasing the achievable density. The influence of adding
crushed rock on swelling potential remains rather uncertain.
The investigation reported in Chapter 5 addresses primarily the questions
of the influence of crushed tuff on hydraulic conductivity and of the condi-
tions under which bentonite might be lost from a seal, e.g. as a result of
bentonite flow into fractures in the host rock within which the seal is emplaced.
The experimental investigations are conducted by means of permeability
tests on reference samples of bentonite only, and mainly by permeability
tests on mixtures of crushed tuff and bentonite. Bentonite weight percen-
tages of 15, 25 and 35% have been tested. Several different gradations of
crushed tuff size distributions have been tested. Piping tests are reported,
aimed at determining the conditions under which bentonite may be lost
from the seals.
Modeling efforts consist of determining empirical correlations between
various characteristics of the seals and the permeability or resistance to
bentonite loss, as characterized by the yield stress.
The authors conclude that crushed rock/bentonite mixtures preferably
should contain at least 25% bentonite, in order to achieve permeabilities
6 Introduction
comparable to those of bentonite only. Compaction is a significant variable

affecting sealing performance. The mixtures, as compacted here, show an-
isotropy and heterogeneity.
Chapter 6 presents the results of extensive experimental investigations on
the sealing performance of bentonite-sand-aggregate earthen barriers. The
investigations were intended to identify an optimum sand/bentonite mix
that would give the best overall combination of physical performance char-
acteristics. The authors identify some serious difficulties with conventional
quality control procedures of compaction, and for that reason implemented
stringent controls on compaction procedures to be followed. The authors
recognize and stress the need for large-scale field testing, even though many
aspects of barrier performance can be studied adequately in laboratory
testing. The experimental investigations have focused strongly on the conse-
quences of moisture migration from, into and through the sandjbentonite
barriers. The observations during the experiments confirm experience
gained elsewhere (e.g. Gaudette and Daemen, 1988) that moisture migration
through bentonitic barriers is an exceedingly complex phenomenon. Never-
theless, current observations on these continuing tests indicate that the
barriers will provide a very low hydraulic conductivity.
Chapter 7 presents methods to determine the in situ hydraulic perform-
ance of borehole seals. Experimental determination of the in situ hydraulic
performance of borehole seals is difficult, at least when the seal performance
is in the low hydraulic conductivity range desirable for sealing openings in
intact low-permeability rock, because the conductivities are very low in-
deed. Depending on the hydraulic characteristics of the host rock, the flow
paths may be influenced significantly by flow through the host rock. Chap-
ter 7 outlines systematic experimental and analyses strategies for testing the
performance of such low-conductivity seals.
Chapter 8 describes field tests on cementitious seals emplaced in
boreholes in rock. The test methods and interpretation procedures intro-
duced in Chapter 7 are illustrated and applied here for several case studies.
One of the case studies, with a cement grout seal in a nearly horizontal
hole, demonstrated the satisfactory closure of an initial interface gap as a
result of cement grout expansion. The other experiments, on seals in vertical
holes, confirmed the excellent sealing performance of cementitious seals
emplaced by relatively simple and conventional procedures.
Chapter 9 deals with the sealing of boreholes in rock salt. Although a
somewhat specialized geological environment, it is rather widespread. Evap-
orite aquitards are frequently penetrated by boreholes and shafts, and seal-
ing failures in soluble evaporites can have and have had notoriously
disastrous consequences. Some aspects of rock salt behavior, notably its
potential for creep and for healing, are favorable with respect to sealing,
while others, especially its solubility, are potentially detrimental. The
authors stress the need to consider the entire seal system, i.e. including seal,
seal-rock interface and host rock, when seal design and performance are
The future 7
considered. For that reason they include a discussion of rock salt behavior
and characteristics.
Materials discussed in Chapter 9 for the purpose of sealing boreholes in
rock salt include granular rock salt, cement grouts, concrete and clays.
Results of laboratory and field tests on the performance of various seals are
presented. In their discussion of design implications the authors place
considerable emphasis on the preference for truly permanent seals, a pref-
erence which may have major implications for the selection of sealing
materials.
Chapter 10 presents an approach to the design of plugs for shafts, ramps,
drifts and tunnels. The chapter introduces a functional classification of the
main types of plugs used in underground excavations, and identifies the
main factors that need to be considered in the design of plugs. Detailed
design calculations are included, as well as allowable stresses for rock,
concrete and steel. Construction procedures for concrete plugs are given,
with emphasis on the need for careful attention to many details. The author
discusses procedures for reducing leakage parallel to plugs and gives several
design and construction case studies.
Chapter 11 outlines some design considerations for borehole seals. It is
recognized that most regulatory borehole sealing requirements prescribe
materials and emplacement procedures primarily. Chapter 11 gives a broad
overview of the many factors that need to be accounted for when planning
and designing borehole seals.
Chapter 12 presents a discussion of the use of sodium bentonite for the
sealing of the boreholes. The increasing popularity and acceptance of vari-
ous forms of bentonite for the sealing of boreholes is based on the wide
recognition of its desirable properties for sealing purposes: low hydraulic
conductivity, geochemical compatibility in many natural environments,
great capacity for self-sealing and ease of installation in a variety of forms.
Bentonite, by itself or as an admixture, has been used extensively for sealing
purposes in a variety of applications.
In Chapter 12 some of the origin and geological aspects of bentonite are
introduced. The use of bentonite as a borehole sealant is presented. Dis-
cussed in this context are various forms of bentonitic sealants, i.e. chips and
tablets, grouts, drilling fluids and granular bentonite. Placement methods of
bentonite are described, and illustrated with a case history.
1.7 THE FUTURE
Borehole sealing technology is well developed in a number of areas, particu-

larly petroleum engineering, water well sealing, and exploratory and
monitoring hole sealing. The necessity to seal holes upon decommissioning
and abandonment is widely recognized. However, some uncertainties re-
main. Among the high priorities that deserve attention are the improvement
8 Introduction
of methods to detect abandoned holes, the development of non-invasive

methods to verify that seals have been emplaced adequately and will per-
form satisfactorily, and the development and validation of methods to pre-
dict the performance of seals over an extended period.
ACKNOWLEDGEMENTS
The editors wish to acknowledge the support of the US Nuclear Regulatory

Commission for the research support provided over many years, which has
allowed us to perform in-depth studies of borehole sealing. We particularly
wish to acknowledge the valuable insights and guidance provided by con-
tract monitors Dr F. Larry Doyle and Mr Jacob Philip.
CHAPTER TWO
Laboratory Performance of
Cement Borehole Seals
D. L. South and K. Fuenkajorn
2.1 INTRODUCTION
This chapter presents an experimental method to assess the performance

of cement borehole seals (plugs) under laboratory conditions. One of the
prime goals is to obtain experimental data regarding the effectiveness of
sealing.
Laboratory conditions represent a highly idealized approximation of
field conditions. This permits the determination of the best possible seal-
ing performance that may be expected because emplacement of a plug
may be done under carefully controlled conditions. In the field, plugs
are emplaced in deep holes, in highly inaccessible locations, frequently
in a wellbore filled with mud. Many variables are introduced in such a
process, few of which can be controlled; some may not even be recog-
nized.
Laboratory testing permits a systematic, controlled variation of the par-
ameters that influence plug behavior. For the work presented here, the
parameters considered are rock type (granite, basalt and tuft), cement plug,
and mechanical plug-rock interaction under varying stress fields. Borehole
plug performance under varying stress field is investigated; temperature is
held constant and saturated conditions obtained. A significant part of the
effort is the development of laboratory equipment for performing the ex-
periments. Investigation of the cement borehole seals is concentrated on
their mechanical and hydrological performance. Where possible, results are
presented as ratios to facilitate application to other rock types and plug
materials. Applications of test results to some engineering aspects of
borehole seal design are discussed.

Edited by K. Fuenkajorn and J. 1. K. Daemen.
10 Laboratory performance of cement borehole seals
2.2 CONCEPTUAL APPROACH
The conceptual approach used to evaluate borehole seals in this work is to

compare flow through a sealed borehole in rock with flow through intact
rock. Rock cores 15 cm in diameter and 30 cm long (nominal dimensions)
have 2.54 cm diameter holes drilled from each end, leaving a rock bridge in
the center of the specimen (Figure 2.1). Water pumped into the top hole
flows through the specimen to the bottom hole. Steady-state flow rates are
measured. The rock bridge is then drilled from the sample and replaced
with a seal and the experiment repeated. This allows direct comparison of
flow rate through intact rock with flow rate through the same rock after a
small portion has been removed and replaced by a seal.
One of the main areas of interest is the performance of the seal under
varying stress conditions. The intact rock specimen is placed under axial
and confining stresses approximating a near-litho static stress field at a
depth of about 1000 m. The intact rock is tested, the rock bridge cored from
the specimen, and a plug placed and tested while the specimen remains
under this stress field. Axial and confining stresses are then lowered to
simulate depths of about 600 and 300 m, and flow through the plug-rock
I Nul
2 Washer
3 Top Plate
4 Loadmq Plalen
5 Piston Plug
6 Locd Cell
1 Piston
8 Cell Cap
9 Specimen
10 Pressure Cell
II Boll
12 Cent.finC} Pin
13 Aluminium P'oten
14 Neoprene Gasket
15 Bottom Plote
16 Bottom Plug:
CENTIMETERS
Fig. 2.1 Specimen and permeameter configuration.

Experimental apparatus 11
system measured. Lowering the stress field is a more severe condition with
respect to the plug-rock interface than raising it because radial stresses
across the interface are lessened.
The method of investigation is chosen to meet two criteria. First, and
most generally, experimental data are desired, as opposed to computer
simulation. This has the advantage of having a rock-plug interface, as
opposed to measurement on individual rock cores and individual cement
cores. Second, the method permits simple, straightforward data analysis
with few assumptions necessary. The main disadvantage of the method
chosen is that the experiments take a long time to perform.
Intact rock is chosen, as opposed to fractured rock, because a borehole
plug seals against the intact zone of a rock mass; that is, the area between
joints and fissures is the area in which sealing occurs. Plug material adjac-
ent to a fracture will result in the diversion of fluid reaching that point into
the fracture and out of the borehole.
In this work intact rock is considered as rock in which no fractures are
visible upon careful examination by the unaided eye.
2.3 EXPERIMENTAL APPARATUS
2.3.1 Permeameter design

An assembly drawing of the permeameter design is shown in Figure 2.1.
Small black rectangles indicate O-ring seals. The permeameter is designed
to accept a 15 cm diameter, 30 cm long rock cylinder with 2.5 cm diameter
holes drilled along its axis. Aluminum platens are usually closer to the
stiffness properties of rock; stainless steel platens are used for most of this
work because they are more chemically inert.
A normal axial stress of up to 21 MPa may be applied to the rock
cylinder by tightening the bolts. The load is measured using a load cell.
Fluid may be pumped into the top hole, the bottom hole, and about the
annulus between the rock cylinder and the pressure cell. Neoprene gaskets
are used to seal the ends of the rock cylinder, isolating the annulus from the
top hole and bottom hole. These gaskets are shown as heavy lines on the
assembly drawing.
Nominal maximum fluid pressure is 21 MPa. Access to the interior of the
specimen is provided by removing the piston plug and the bottom plug.
This may be done while the specimen is under an axial stress and, if press-
ure is maintained about the annulus, under a confining stress.
A centering pin in the bottom plug is used to align the specimen when it
is placed in the permeameter; the pin is removed during testing.
The specimen is coated with epoxy on the outside to prevent fluid seep-
age from the annulus through the rock to the center holes. Confining stress
is applied by pressurizing water in the annulus between the specimen and

the pressure cell. At this point pressure in the top and bottom holes of the
specimen may be zero. The neoprene gaskets at the ends of the specimen
maintain the confining pressure. Sealing by the gaskets depends upon an
axial stress higher than the confining stress. It is not possible to maintain a
higher confining stress than axial stress. Using an axial stress of 22.6 MPa it
is possible to maintain a confining stress of 19.6 MPa within 2% over a
period of 24 hours. The confining stress is applied with a manually operated
pump. Pressure is either maintained or drops as water leaks through the
seals. This equipment, with slight modifications, can be used to investigate
both temperature and moisture variations.
2.3.2 Pump design

Water must be supplied to a specimen in the permeameter at a constant
pressure and at a very slow rate. To compare experiments it is desirable to
fix the pressure and let flow rates be controlled by the permeability of the
rock-plug system. To achieve this a constant pressure pump is designed
and four are constructed. Compressed nitrogen is supplied to a large diam-
eter cylinder of the pump, forcing a piston downward. A smaller piston is
thus forced into a cylinder containing water, forcing water out of the bot-
tom of the small cylinder at constant pressure. The fitting connected to the
bottom of the water cylinder is connected with tubing to a fitting on the
permeameter.
Pressure intensification is approximately 11.5. A nitrogen pressure of 1.38
MPa yields a water pressure of 15.86 MPa. The pumps are made of stain-
less steel and have performed well. The main limitation is that pressures
below about 1 MPa cannot be well maintained at constant pressure because
of O-ring friction, which is also responsible for pressure fluctuations of
0.1-0.2 MPa at higher pressures.
2.3.3 Data acquisition system

Data are collected using an automatic data-logging system. The volume of
water pumped into the sample is measured by the pump piston displace-
ment with linear encoder. The amount of water flowing out of the sample is
collected in a flask which sits on a force transducer. Fluid pressure is
measured using National semiconductor pressure transducers Model
LX-1450 AF and Model LX-1460AF. The Model LX-1450 AF has a range
of 0-14 MPa and the Model LX-1460 AF a range of 0-21 MPa. The lower
pressure range transducer is used to measure pressure in the top hole. The
higher range pressure transducers are analog devices and are connected to
the analog-to-digital voltage converter card. Axial stress is monitored using
50 ton (45455 kg) capacity load cells.
Experimental procedures 13
2.4 EXPERIMENTAL PROCEDURES
2.4.1 Sample preparation
Rock specimen
Five different rock types are tested: two granites (Oracle granite and Char-
coal granite), basalt from the Sentinel Gap area on the Columbia Plateau
and two tuffs (Topopah Spring tuff and Apache Leap tuft). South and
Daemen (1986) describe in detail the petrographical and mechanical proper-
ties of the rocks. Fuenkajorn and Daemen (1991a, 1992a) describe the pro-
perties of Apache Leap tuff.
Cylindrical rock specimens are obtained either by laboratory coring of
boulders collected from the field or by field drilling. The cylinders are cut to
length, 30 cm, with a diamond saw and the ends ground flat and parallel.
Grinding is one of the most important steps as flat, parallel ends are
necessary to provide good end seals and to provide a uniform stress dis-
tribution. Specimen ends are prepared to specifications set forth by the
International Society for Rock Mechanics (1981) for preparing samples for
uniaxial compressive strength testing. This specification states that the ends
shall be flat to 0.02 mm and shall be parallel to within 0.10 mm in 50 mm.
Specimen flatness and parallelism are checked with a dial gauge.
Next, 2.54 cm diameter holes are drilled along the specimen axis from
each end to a depth of one-third the total length. A blind bit is used to
flatten the bottom of the hole.
In order to prevent seepage of water into the sides and ends of the
specimens, several coats of epoxy are applied. Next, the sample is placed in
the permeameter and a small axial load is applied to keep the top plate
secure. The permeameter is turned over, the bottom plug removed and the
bottom hole filled with distilled water. Enough water is poured into the
bottom hole so that when the bottom plug is replaced water is forced from
the valve, ensuring that no air is entrapped. The permeameter is righted and
connected to a pump, ready for saturation and testing.
Cement borehole plug

The cement used is a proprietary formulation provided by Dowell. It is
composed of Ideal Class A cement (Tijeras Canyon) with 50% water content
and proprietary Dowell additives D53 (10%), an expansive agent, and D65
(1 %), a dispersant. All percentages are weight percent with respect to ce-
ment. The cement is mixed according to American Petroleum Institute
specifications (American Petroleum Institute, 1986). Dowell indicates that
the slurry has density of 1.88 glcc, a strength of 26.2 MPa after 14 days of
curing at 43C, and 0.18% expansion after 14 days. Radial expansive stresses
of the cement tested here have been measured by various investigators (South
and Daemen, 1986; Fuenkajorn and Daemen, 1987; Akgun and Daemen,
1991c). They conclude that after 25 day of curing under water the radial
expansive stress of cement can increase by up to 4 MPa. Measurements of
the cement permeability by South and Daemen (1986) yield an estimate of
95 nanodarcy.
2.4.2 Testing
When the specimen is first tested it has a rock bridge in place. The testing
procedure is to first apply an axial and confining stress and to then apply a
vacuum to the top hole. At this time the bottom hole is filled with water.
The vacuum draws the air from the top hole and the pore space of the
specimen. Distilled water is then injected into the top hole through a mani-
fold which allows the water to be injected without admitting air to the top
hole. Hence, the air has been withdrawn from the specimen and water is
now being injected to saturate it.
Once the specimen has been saturated, as evidenced by water flowing
from the bottom hole at the same rate as it is injected into the top hole,
testing begins.
Table 2.1 summarizes the normal test schedule. With the specimen under
an axial stress of 23 MPa and a confining stress of 20 MPa, a fluid pressure
of 10 MPa is applied to the top hole by the constant pressure pump. Flow
occurs through the specimen to the bottom hole, which is at zero (atmos-
pheric) pressure.
Following the test at 10 MPa top hole pressure, tests are performed at 7.0
MPa and 3.5 MPa to provide data on the variation in flow rate with
injection pressure. Following the tests at the three different top hole fluid
pressures, the rock bridge is cored from the specimen. Axial and confining
stress are maintained during this operation.
A rubber stopper is placed at the location of the bottom of the rock
bridge. A plug of cement is then placed. The cement is then covered with
Table 2.1 Nominal schedule for rock bridge and plug flow testing
Step No. Description
1 Load permeameter, check seal

2 Saturate, establish flow
3 Test at axial stress = 23 MPa and confining stress = 20 MPa
(Injection pressures = 10.0, 7.0 and 3.5 MPa)
4 Core out specimen, place plug
5 Cure plug
6 Re-establish flow
7 Test at axial stress = 23 MPa and confining stress = 20 MPa
8 Reduce stress, axial stress = 15 MPa and confining stress = 13.5 MPa
9 Reduce stress, axial stress = 8.5 MPa and confining stress = 7.0 MPa
(Reduce injection pressures to 3.5 and 1.7 MPa)
Analysis 15
water and allowed to cure. Following curing the rubber stopper is removed
and the same series of tests performed on the rock bridge is performed on
the cement plug. Flow through the plug will thus be directly comparable
with flow through the intact rock.
Next, the axial and confining stresses to which the sample is subjected are
reduced and the test series repeated. Axial stress is reduced to about 15
MPa and confining stress to about 13.5 MPa. Following this test series
axial and confining stress are again reduced, to 8.5 and 7.0 MPa. respective-
ly. To avoid inducing tensile stresses, and to maintain the end seals, injec-
tion pressures of 3.5 and 1.7 MPa are used.
2.5 ANALYSIS
2.5.1 Flow calculation

Finite element analyses are performed using the program FREE SURF
(Neuman and Witherspoon, 1970) to determine the permeabilities of the
rock bridge and plug from the measured flow rates and to estimate the
amount and path of water flow through the specimen. For this analysis, the
model is designed for a hollow cylinder having length-to-diameter ratio
(LID) = 2, outer-to-inner diameter ratio = 6, with the rock bridge (or plug)
having length-to-diameter ratio = 1 at the mid-section of the hole.
This analysis is made in axisymmetry, assuming that the specimen and
plug are homogeneous, isotropic and fully saturated, and that Darcy's law
is applicable. The side and ends of the specimen are no-flow (impermeable)
boundaries. The boundaries of the top and bottom holes are constant-head
boundaries, the top hole at the head matching the injection pressure, the
bottom hole at zero (gage). Figure 2.2 shows the finite element mesh and
boundary conditions used in the flow analysis. Due the two symmetry
planes, only one-fourth of the specimen geometry is analyzed.
Due to the direct proportionality between the flow rate (Q) and the head
difference (Ilh- between the top and bottom holes) and between the flow
rate and specimen diameter (D), the calculated flow rates are presente<J in
the following forms: QIllhD, referred to here as 'apparent flow velocity', and
QIllhDkR, called 'normalized flow rate', where kR = hydraulic conductivity
of rock specimen. The first form has a velocity unit and the second is
dimensionless. The analysis is, therefore, independent of the head difference
and specimen size.
Figure. 2.3 shows the result from the finite element analysis, which pro-
vides a solution to obtain rock bridge conductivity from the measured flow
rate. Examples of the results are shown in the form of flow net in Figures
2Aa and 2Ab. Figure 2Aa shows flow path in a specimen having plug
permeability 10 times higher than rock permeability (k p = 10 kR). Figure
2Ab gives the flow path in a specimen having plug permeability 10 times
lower than rock permeability (k p = 0.1 kR). The normalized flow rate is
...........-'m.....m. _ Boundary
V
t
R
D
1 - - - - - - - ~ - - - - -....
Fig. 2.2 Finite element mesh and boundary conditions used in flow analysis of
permeameter specimens. (L = specimen length, D = specimen diameter).
-7
10
-8
10
~
E -9
~ 10
QI~
-10
10
-11
10
I I I
-9 -8 -7
10 10 10
Fig. 2.3 Apparent flow velocity as a function of rock specimen permeability for
rock bridge testing (kR = k p), calculated by finite element method.
z
eo"
8.
10
8.
(a)
8.%
8.
7.
eo
L-__~LL~~~~-L-L-L-L~__ R
(b) (b) 1020 30 40 50 10 70 80 oocr,
Fig.2.4 (a) Flow net for test specimens with plug permeability 10 times higher than
rock permeability (k p = 10kJ. (b) Flow net for test specimens with plug permeability
10 times lower than rock permeability (k p = 0.1 kJ.
calculated as a function of plug-to-rock permeability ratio (kp/k R , as shown

in Figure 2.5. This relation is used to calculate k p after kR has been deter-
mined from rock bridge testing. The normalized flow rate tends to be
independent of the kp/k R ratio when the ratio is less than 1. The flow rate
linearly increases with increasing ratio if kp> 100 k R For this case the
borehole plug permeability may be calculated by assuming that all water
flows through the plug (i.e. one-dimensional flow).
This finite element analysis assumes also that the plug and interface
permeabilities are identical. If the actual interface has higher permeability
than the plug (Fuenkajorn and Daemen, 1986b), this analysis will overesti-
mate the plug permeability.
2.5.2 Stress analysis

The program PLANE2D-FE (Desai and Abel, 1972) is used to determine
tangential and radial stress distributions within the specimen subjected to
various sets of confining stress and borehole pressure. The analysis is made
in axisymmetry and assumes that the materials are linearly elastic, homo-
geneous and isotropic, and that the rock-plug interface is a welded surface.
For this analysis the rock cylinder is welded tuff and the borehole plug is
cement. Elastic modulus is 22.6 GPa for the tuff and 5.65 GPa for the
cement. Fuenkajorn and Daemen (1991b) describe in detail the boundary
and loading conditions of the model. One of the objectives is to identify
tension zones at the plug-rock interface. The tension zone is of concern
because it may induce a separation at the interface, particularly near the
upstream end of the plug.
2
10
10
-1
10
-2 -1
10 10 10
Fig.2.5 Normalized flow rate as a function of plug-to-rock permeability ratio

(k p / kR)
Summary of test results 19
Results from the analysis indicate that a tension zone will not be induced
at the interface if the borehole pressure is less than 75% of the confining
stress. Akgun and Daemen (1991c) give detailed stress analysis for cement
borehole plugs in welded tuff subjected to a variety of applied pressures.
Figures 2.6 a nd 2.7 show the normalized tangential stress (O"O/O"la,) and
normalized radial stress (O"r/O"la,) at the mid-sections of the specimens with
the rock bridge and with the cement plug. The lateral stress, O"lal' represents
the fluid pressure in the annular zone. The top hole pressure or injection
pressure is equal to 0.50"1a" Comparison of the stress distributions between
the rock bridge specimen (Figure 2.6) and the cement plug test specimen
(Figure 2.7) indicates that the low Young's modulus of the cement (about
one-fourth of the tuft) reduces the stress concentration within the plug
region (i.e. it provides a more uniform stress distribution within the plug).
This implies that if the effective stress had a significant effect on the flow
behavior, the flow path in a cement plug specimen would be less compli-
cated than that in a rock bridge specimen.
2.6 SUMMARY OF TEST RESULTS
Table 2.2 summarizes the results from rock bridge testing and borehole
plug testing in rock specimens. Discussions below are concentrated on
testing of Charcoal granite specimens, which typifies the permeameater
test results.
-rrT-------lo.sfl--------------- A
1.2 1.5 0.5 0.8 0.9

1.3
1.05 1.1 1.151.21.3 1.6 0 0.60.8 0.9

_1 ______ 1 __ L_LJ.1:~~:9.___ ____ .:P~_l ___ L _________ _
I
0.2
A'
Fig.2.6 Normalized tangential and radial stresses ((fe/ (fla' and (fr/ (fla,) at central
section of rock bridge test specimen under confining stress (fla,' with top hole pres-
sure = 0.5 (fla"
r------ - -------------- A
1.5
A'
Fig.2.7 Normalized tangential and radial stresses (aela la , and arla la ,) at central
section of cement plug test specimen under confining stress ala" with top hole
pressure = 0.5 ala'.
Table 2.2 Summary of rock and borehole plug permeability

Rock type Rock permeability Cement plug
(nanodarcy) permeability
( nanodarc y)
Charcoal granite 52-85 <52

Sentinel Gap basalt 0.16-0.17 10-73
Oracle granite
With fracture 50000
Without fracture 9 8000a
Topopah Spring tuff 1700-4800 2000-2500b
Apache Leap tuff 1-10 10-100
a This permeability is calculated assuming water flows exclusively through the plug (the value
is highly uncertain).
b High permeability may result from an alkali-aggregate reaction between the tuff and the
cement.
2.6.1 Effects of stress variation

The results shown in Figure 2.8 are representative of tests performed using
granite with cement plugs. The cement used is Ideal Class A cement with an
expansive agent and a dispersant added. Data from the test on the rock
bridge are shown as a solid line, curve a. Dashed lines show data from tests
6 Axial and Confining

Stresses. IMPal
"'ax ..!L
a 22.9 19.7
b 022.9 19.6
5 c 22.8 19.7
d 0 15.6 13.8
. . . 8.6 7.0
'i:
'E
.... 4
"'eu
",
'2
~ 3
Q
Ji
:!
"
II:: 2
~
Ii:
,
./
2 4 6 8 10
Tap Pressure, PT , (MPa)
Fig. 2.8 Typical test sequence and results from permeameter test. Inflow rate as
a function of top hole (injection) pressure, for a Charcoal granite specimen with
cement borehole plug.
on the cement plug. The latters a, b, c, d and e are ordered in time (i.e. a was
run first, then b, etc.).
Curve a, flow rate through the rock bridge, is the baseline. Curve b is
derived next. It shows a lower flow through the cement plug than the rock
bridge under similar axial and confining stresses. Curve c repeats the condi-
tions of curve b. It is derived to check repeatability; obviously, curve c
shows a lower flow rate than curve b. This is believed to be due to decreas-
ing cement plug permeability resulting from forcing high-pressure water
through the plug. Reducing the axial and confining stresses increases the
flow rate (curve d), but not until axial and confining stresses are reduced to
about one-third their initial values does flow rate through the plug-rock
system exceed the initial flow rate through intact rock.
Figure 2.9 shows the flow rate as a function of permeability calculated by
the program FREE SURF for the granite specimen with top-hole injection
pressures (PT ) of 3.5, 7.0 and 10.1 MPa. Bottom hole head is zero; the sides
and ends of the specimen are modeled as no-flow boundaries. The flow rate
is a linear function of the permeability at a given injection pressure. This is
expected as the calculation is based on Darcy's Law.
( 'O.OMPa
,7.0 MPa
c10~2 ~3.5MPa
1
:1
"
<r

~IO-3
CG-I02
3.3
4.2 ROCK BRIDGE
4.6
10- 8 10-7 em/min

1.6 x 10-7 1.6 x 10-6 darcy
Permeability
Fig. 2.9 Flow rate as a function of permeability for specimen with rock bridge
tested at top hole pressures of 10.1, 7.0 and 3.5 MPa, calculated by program
FREESURF.
Based on the measured flow rates from a granite specimen, the rock has
permeabilities of 54.4, 69.2 and 75.8 nanodarcy for top-hole injection press-
ures of 3.5, 7.0 and 10.1 MPa, respectively. The permeability increases with
increasing injection pressure because the higher pore water pressure tends
to increase the size of the connected pore space in the specimen, increasing
its permeability. Microscopic examination of thin sections of the granite
indicates that pore space in the granite exists along mineral grain bound-
aries and as microfractures through grains.
Figure 2.10 shows the variation in sample permeability with the first
stress invariant (I1) at each top-hole injection pressure. As the sample is
subjected to higher stress conditions, that is, as 11 increases the sample
permeability decreases, probably due to decreasing pore sizes and fracture
widths within the sample. Permeability increases as the injection pressure
increases at a given stress level, 1 1 , because increasing the injection pressure
decreases the effective stress; increasing injection pressure tends to open
pores and fractures.
Figure 2.11 shows the variation in flow rate with variation in plug per-
meability, The abscissa is the ratio of plug permeability to intact rock
permeability, kp/k R. kR' the intact rock permeability, is that permeability
determined for the specimen with rock bridge in place. At 10, ( = 1) the
plug and rock permeabilities are equal, the case for intact rock. The 10.1
MPa injection pressure curve was calculated using kR = 75.8 nanodarcy,
and the 7.0 and 3.5 MPa curves using their corresponding intact rock
permeabilities. The measured flow rates through the plugged specimens at
60
50
45 Measured
Calculated
35
20
15
10
0L---~10----~20----~~----4~0~~5~0--~6~0--~70
IIIMPa)
Fig.2.10 Permeability as a function of first stress invariant for a Charcoal granite

specimen with rock bridge tested at different top hole pressures, PT'
3.5MPa
'C10- 2
~u
oS
!
"
a:
~ I _ _ _ _ _ _ _ _ _--r=--..::::::::7.0MPa
f!.IO- 3 t-
1 - - - - - - 3.5 MPa
CG-I02
CEMENT 1 PLUG
Fig.2.11 Flow rate plotted as a function of plug-to-rock permeability ratio for a

Charcoal granite specimen tested under top hole pressures of 10.1, 7.0 and 3.5 MPa,
calculated by program FREESURF.
10.1,7,0 and 3.5 MPa, with axial and confining stresses of 22.9 and 19.6
MPa, respectively, are shown as arrows pointing to the k p = kR line. All
measured flow rates fell below the theoretical curves, possibly due to stress
redistributions resulting from the coring out of the rock bridge and the
expansiveness of the cement used as the plug material.
These test results supported by the finite element analysis illustrate three
points significant to choice of materials for borehole sealing:
Borehole plug materials with permeabilities less than the rock being
sealed do not significantly reduce flow in the vicinity of the borehole.
Flow through the borehole plug only begins to increase significantly
when the plug material becomes an order of magnitude of greater in
permeability than the rock being sealed.
Significant reduction of rock stress did not greatly increase flow through
a cement plug in granite. However, the granite had a higher Young's
modulus than the cement plug. Decreasing the triaxial stress state on a
rock/plug combination in which the rock has a lower Young's modulus
than the plug could yield different results.
2.6.2 Effect of drilling method

Granite specimens are tested to evaluate the effects of differing drilling
methods. Two specimens from the same location are obtained by diamond
coring in the field. One specimen first had a percussion hole drilled into it in
the field; it was then overcored. The other specimen was first cored in the
field and returned to the laboratory. Top and bottom holes were diamond
drilled to leave a rock bridge in place. The specimen with the rock bridge
was then tested, the rock bridge cored from it and a cement plug placed. A
cement plug was also placed in the percussion drilled specimens. Flow rates
were measured in both specimens under similar axial and confining stresses
with similar injection pressures.
Using the flow rate obtained on the specimen with a rock bridge as a
reference value of 100%, the specimen with the percussion-drilled center
hole had a flow rate of 55% and the specimen with the diamond-drilled hole
had a flow rate of 71 %. The cement plug in the percussion-drilled specimen
was air dried for 27 days at a temperature of 21C and still had a flow rate
75% of the reference flow rate of the specimen with the rock bridge in place.
This very direct comparison of diamond-drilled with the percussion-drilled
specimens indicates drilling method does not create a borehole wall damage
zone which significantly affects seal performance. This finding agrees with
the experimental results obtained by Mathis and Daemen (1982) and
Fuenkajorn and Daemen (1984, 1986a). They conclude that even though
different drilling methods (diamond coring, percussion and rotary drilling)
induce different characteristics of damage (roughness, intensity and extent
of fracture zone) around the borehole wall, their effect on the hydraulic
performance of the cement borehole seals is not significant. It should be

noted that the effect of drilling fluid (mud) is not considered here. In deep
in situ boreholes the remains of drilling fluid on the hole wall (mud cake)
could degrade the cement plug performance.
2.6.3 Effect of pre-existing fractures

One granite specimen with a visible fracture in its top third was tested with
a rock bridge in place and with the rock bridge cored from the specimen
and a cement plug placed. The fracture was thin, partially healed and less
than 1 mm in aperture. It was barely visible upon visual inspection prior to
testing and was not detected during routine core logging.
The test configuration was different from the tests described in the previ-
ous sections; this was tested in divergent and convergent flow. The most
significant results regarding borehole plugging for general engineering ap-
plications came from the convergent flow tests. In these tests water flow was
from the annulus to the top and bottom holes. Equivalent permeabilities of
the top third of the specimen, which contained the fracture, and the bottom
third of the specimen were calculated.
Using the equivalent permeability of the intact rock as a reference of 1,
the rock with the cement plug had an equivalent permeability about 1000
times greater. (This was the only cement/granite combination in which the
cement-plugged specimen showed a significantly higher permeability than
the intact specimen.) The equivalent permeability of the rock with the frac-
ture was about 7000 times greater than the permeability of the intact rock,
and 7 times greater than the permeability of the plugged rock. This demon-
strates the dominant effect of fractures on flow in the vicinity of a borehole.
Results from a similar test configuration obtained by Fuenkajorn and Dae-
men (1984, 1986a) also indicate that pre-existing fractures intersecting the
borehole near the plug location usually become a preferential flow path for
fluid to bypass the plug.
2.6.4 Effect of plug-rock interface

Based on the dye tests performed, the plug-rock interface zone is not a
preferential migration path in the granite tested. Variation of the triaxial
stress state on the two basalt specimens does not significantly affect flow
through the cement plugs, nor is the basalt permeability measurably af-
fected. Increased flow rate in the granite specimens with decreasing triaxial
stress state is due to variation of rock permeability, rather than plug per-
meability.
Interaction pressure at the cement plug-rock interface resulting from the
cement expansion is important for maintaining a good hydraulic bond at
the interface. This mechanical interaction is governed by the expansive
capability of the cement after solidifying, and by the stiffness of the
surrounding rock. For the type of cement used here, the interaction pressure
at the interface could be as high as 4 MPa when it is installed in a stiff rock,
such as granite.
2.6.5 Drying effect

Two specimens with cement plug are subjected to drying to ascertain the
effects of temperature/moisture variations. One specimen is dried at room
temperature for 27 days. This is found to impair slightly the performance of
its cement plug. The other specimen is dried for 42 dyas at 55C. This
impairs plug performance, increasing plug permeability from about 10-60
nanodarcy to an initial value of 190 000 nanodarcy. As flow continues the
plug becomes less permeable, reaching 2500 nanodarcy after 50 days. In-
creasing the triaxial stress state on the sample does not greatly affect plug
permea bili ty.
The results obtained here indicate that temperature/moisture variations,
specifically heating in air and allowing the cement to dry, degraded cement
plug performance significantly over even the relatively short term of 6
weeks. Microscopic studies on the dried-out cement plugs by Fuenkajorn
and Daemen (1986b) indicate that widths of the plug-rock interface are
about 0.1 mm for specimens dried at 45C for 7 days, and about 0.2 mm for
specimens dried at 260C for 1 h. The dried cement plugs themselves also
show shrinkage cracks with an average width of 0.01 mm and intensity of 1
per cm. More discussions on the effect of drying on the hydraulic perform-
ance of cement plug are given in Chapter 4.
2.7 DESIGN IMPLICATIONS
Taken together, the laboratory tests and numerical analyses indicate several
points to be considered when plugging boreholes:
Currently available expansive cements are adequate to provide good per-
formance for borehole seal under changing stress conditions. However, if
the plug material is stiffer (i.e. it has a higher Young's modulus) than the
rock being sealed, stress relief could result in an increase in borehole
diameter greater than the resulting expansion of the plug material.
Interaction pressure at cement plug-rock interface is governed by the
swelling pressure of the cement and stiffness of the rock. Good hydraulic
bond at the interface is expected when cement is installed in stiff rock.
Making the plug material less permeable than the surrounding rock,
including fractures, will not significantly reduce fluid flow. In fractured
rock, flow through the natural fracture system will dominate provided the
seal material is no more than one order of magnitude more permeable
than the intact rock.
Acknowledgements 27
Borehole sealing is not sensitive to the drilling method used to produce
the borehole.
Drying can significantly increase cement plug permeability by several
orders of magnitude. Performance partially recovered upon resaturation.
Temperature/moisture variations, specifically heating in air and allowing
the cement to dry, degraded cement plug performance significantly over
even the relatively short term of 6 weeks.
A combination of cement and bentonite placed in appropriate sections of
a borehole is expected to give the best results. Cement should be placed
at each end of a section with a bentonite plug between. The cement will
provide strength and the bentonite will be able to accommodate strains
resulting from stress changes.
ACKNOWLEDGEMENTS
The work was part of research effort sponsored by the US Nuclear Regula-
tory Commission, under contracts NRC-04-78-271 and NRC-04-86-113.
Support and permission to publish this chapter are gratefully acknow-
ledged.
CHAPTER THREE
Strength Parameters of
Cement Borehole
Seals in Rock
H. Akgiin
3.1 INTRODUCTION
Sealing of penetrations (e.g. boreholes, shafts, mine drifts and tunnels) may
be required for a variety of reasons. Penetrations of and near a high-level
nuclear waste repository need to be sealed reliably to retard any radionu-
clide migration to the accessible environment (US Nuclear Regulatory
Commission, 1983, 1985). Design of seals may also be required in (1) water
dams, barriers, water wells, mine drifts or shafts, to prevent flooding of
underground operations (e.g. Garrett and Campbell Pitt, 1961; Loof-
bourow, 1973), (2) diversion tunnels for the construction of hydroelectric
power plants (e.g. Mitchell, 1982; Kinstler, 1983; Moller et al., 1983; Pett-
man, 1984), (3) mine openings in order to control mine effluents (e.g. Mining
Waste Study Team 1988; Einarson and Abel, 1990), (4) oil, gas and chemical
disposal wells (e.g. Smith, 1976; Calvert, 1980; Halliburton Services, un-
dated), and (5) blasthole stemming studies (e.g., Konya, Otounye and
Skidmore, 1982; Otuonye, Konya and Skidmore, 1983).
Axial loads on seals or plugs may be due to water, gas or backfill press-
ures, or due to temperature changes induced subsequent to waste and plug
emplacement. These axial loads induce shear stresses along the contact
between plug and host rock. These shear stresses may cause cracking and
increased permeability along the plug-rock interface. Under extreme
conditions they could cause dislodging or slipping of plugs. Therefore, the
interface between the plug and rock is a critical element for the design and
performance of plugs in boreholes, shafts or tunnels.

Push-out test mechanical interactions 29
The objective of this study is to determine the strength of cement grout
borehole plugs in welded tuff cylinders as measured through push-out tests.
The push-out test involves the use of a steel rod to dislodge experimentally
a cement grout plug emplaced within the coaxial borehole of a hollow rock
cylinder (Figure 3.1). The rock specimens were taken from the densely
welded brown unit of the Apache Leap tuff near Superior, Arizona. The
expandable cement grout formulation, provided by Dowell-Schlumberger, was
Self-Stress II cement grout. Cement grout mixing was performed according to
American Petroleum Institute (API) specifications, API Standard No. RP-lOB
(1986). The push-out tested tuff cylinders had inside radii ranging from 6.4 mm
to 51 mm, outside radii ranging from 38 mm to 94 mm and lengths ranging
from 102 mm to 178 mm. The plugs were cured under water for 8 days prior to
testing. Loads were incremented stepwise at intervals of 5 min.
3.2 PUSH-OUT TEST EXPERIMENTAL PROCEDURE
Figure 3.1 shows the push-out test arrangement. A cylindrical steel rod
applies an axial load to a neat cement grout plug installed in a rock cylin-
der. The L VDT (Linear Variable Displacement Transducer) and dial gage
that measure the .vertical displacement of the top of the plug are mounted
on horizontal arms connected to the loading rod. The top L VDT and dial
gage displacement monitoring points rest on horizontal brackets clamped
to fixed vertical reference bars. The steel platen underneath the sample has a
slit on one side to allow the downward movement of the horizontal arm of
the bottom vertical displacement monitoring assembly. A vertical rod,
screwed into the bottom of the cement plug, is connected to the horizontal
arm which supports the bottom L VDT and dial gage monitoring points for
bottom plug displacement measurements. The bottom LVDT and dial gage
are clamped to fixed vertical reference bars. A steel pipe is placed around
the rock specimen to provide confinement.
The tuff cylinders tested had inside radii of 6.4, 13, 25 or 51 mm, outside
radii ranging from 38 to 94 mm and lengths ranging from 102 to 178 mm.
The tuff cores were plugged with nearly centered Self-Stress II cement grout
plugs having length-to-radius ratios ranging from 2.0 to 8.0. The cement
grouts of the push-out specimens were initially loaded to 4450 N. The load
was kept approximately constant and increased by 4450 N every 5 min until
the plug failed. The load and displacements were recorded every 30 s upon
failure. Experimental details are given by Akgun and Daemen (1991c).
3.3 PUSH-OUT TEST MECHANICAL INTERACTIONS
Many structures are mechanistically similar to an axially loaded plug em-

placed within the coaxial borehole of a hollow rock cylinder, e.g. composite
30 Strength parameters of cement borehole seals in rock
Dial gaga LVTD
Rock
sample
Vertical steel rod Vertical steel rod
Dial gage
c:::::====t'=t===S:=~='====~3---Circular steel plate
Fig. 3.1 Schematic drawing of push-out test set-up.
materials, reinforced concrete, piles and rock bolts. Akgun and

Daemen (1991c) present a review of mechanical analyses of interaction
configurations similar to push-out testing.
The shear stress distribution induced by push-out loading along the ce-
ment plug-rock interface can be calculated in several ways (Akgun and
Daemen, 1991c).
Assuming a uniform shear stress distribution gives an average shear stress
(3.1)
where ray = average shear stress along the plug-rock interface, Ci zo = axial
stress applied to the plug, a = plug radius or rock cylinder inside radius
and L = plug length .
Assuming an exponential shear stress distribution (elastic solution) results
in
Ci zo af3 cosh[f3(L-z)]
(3.2)
r=-2- sinh(f3L) ,
Finite element analysis and discussion 31
where r = exponential shear stress along the plug-rock interface;
[32 = {[I - 2vp(VSF)]j[a 2(EpjER) In (rja)(l + vR)]}; r = critical radius beyond
which shear stresses in rock are negligible = {a + [2.0 exp( - 0.33
(EpjE~)]L}; EpjER = ratio of the Young's moduli of plug and rock; (the
vertical stress function) VSF = {vp (l - (ajR)2 j[(l - (ajRf (1 - vp) +
(1 + v~(EpjER) [(1 - 2vR)(ajR)2 + I]]}, Vp , VR = Poisson's ratio of plug and
rock, respectively; R = rock cylinder outside radius; a, a zo and L are as
defined by Equation (3.1). Equation (3.2) is valid for the following condi-
tions: 2.0::::; Lja::::; 8.0, vp::::: VR and 0.10::::; EpjER ::::; 10.
It follows from Equation (3.2) that the peak interface shear stresses occur
at the loaded end of the plug - rock interface (i.e. at z = 0). The peak shear
stress (rp) can thus be expressed as
a zoa[3
r - ---=---=--- (3.3)
p - 2tanh([3L)
The push-out tests reported herein have been performed on unconfined
tuff cores with finite outside-to-inside radius ratios. Different radius ratios
lead to different stiffnesses, and hence to different radial (contact) stresses at
the plug - rock interfaces. As the axial stress applied to a borehole plug is
directly proportional to the radial stress through Poisson's effect, the axial
stress at failure (or the axial strength) for a push-out specimen with a finite
radius ratio can be normalized for that of an infinite rock mass by the
following equation (Akgun and Daemen, 1991c):
a(vpjVSF)
a (3.4)
(1 - vp ) + (EpjER)(l + vR)
=~--------~~--~
n
where an = normalized axial strength for a rock mass; a = axial strength for
a push-out specimen with a finite outside-to-inside radius ratio; VSF, VR, Vp
and EpjER are as defined by Equation (3.2).
The average Young's modulus and Poisson's ratio of the cement grout
plug are 5.3 GPa and 0.22, that for the welded tuff are 23 GPa and 0.20.
The modulus ratio (i.e. the ratio of the Young's modulus of the cement
grout plug to that of the rock) is 0.23 (Akgun and Daemen, 1991c).
3.4 FINITE ELEMENT ANALYSIS AND DISCUSSION
3.4.1 Introduction
The main objective of the finite element analysis is to study the interface
shear stress and tensile stress distribution within and in the vicinity of an
axially stressed borehole plug. The secondary objective is to assess the
validity of the analytical interface shear stress distribution solution pres-
ented by Equation (3.2).
An axisymmetric finite element program, PLANE-2DFE, was used to

determine the stress distribution within an axially stressed push-out speci-
men. The finite element analysis simulated a two material push-out test on
an elastic material with an applied axial plug stress of 1.0 MPa. Desai and
Abel (1972) give the finite element formulation for the four-node
isoparametric axisymmetric elements used in this analysis.
Figure 3.2 gives the finite element mesh and boundary conditions for a
plugged cylinder with an outside-to-inside radius ratio of 6.0, and plug
length-to-radius ratio of 2.0. The finite element model is designed for a tuff
cylinder of radius 76 mm, length 127 mm and with a 13 mm radius coaxial
hole. The cement grout plug is centered halfway down the hole and has a
length-to-radius ratio of 2.0. The mesh consists of 256 elements and 295
nodal points. The hatched area in Figure 3.2 shows a particular rock sec-
tion along which the normal and shear stress distributions were studied in
order to determine the critical radius, r, given by Equation (3.2), beyond
which the shear stresses in rock are considered to be negligible. Akgun and
Daemen (1991c) give the finite element mesh and boundary conditions for
plugged tuff cylinders with outside-to-inside radius ratios of 60 and plug
length-to-radius ratios of 2.0, 4.0 and 8.0, respectively. The objective of
using a radius ratio of 60 is to simulate an in situ rock mass. Randolph and
Wroth (1978) propose the use of a radius ratio of 50 for in situ rock mass
simulation.
,----,
p
Om
-"
!
i
i
,-
i
i
i
i
~M ~
'oR
I 25mm I
Fig. 3.2 Finite element mesh and boundary conditions for a push-out specimen.
3.4.2 Shear stress distribution

Figure 3.3 gives the interface shear stress near the loaded end of a borehole
plug (i.e. at zlL = 0.1, where z is the distance from the initial location of the
loaded end of the plug, and L is the plug length) per unit applied axial
stress. The plot is given as a function of the modulus ratio (EplER) and plug
length-to-radius ratio (Lla), and allows a comparison of the results of the
finite element analysis with that of the closed-form solution given by Equa-
tion (3.2). Figure 3.3 is for a 13 mm radius borehole plug with a cylinder
outside-to-inside radius ratio (Ria) of 60 and for a Poisson's ratio of the
plug and rock of 0.22 and 0.20, respectively. The interfacial shear stress
decreases with increased modulus ratio and increased plug length-to-radius
ratio.
Figure 3.4 gives the shear stress distribution along the plug-rock inter-
face per unit axial stress applied to the borehole plug. The plot presents
results from the finite element analysis along with those obtained through
the closed-form solution and is given as a function of plug length-to-radius
ratio. Figure 3.4 is for a push-out sample with a modulus ratio of 0.23, a
plug radius of 13 mm, and for Poisson's ratios of the plug and rock of 0.22
and 0.20, respectively. The peak shear stresses increase with decreasing plug
length-to-radius ratios. The shear stresses do not distribute over the entire
lengths of longer plugs. The analytical solution overestimates the peak
interfacial shear stresses by up to 32% (when compared to the finite element
analysis) and is conservative for plug design.
0.8
0.7 - - - Finite element analysis

- - - Analytical solution
0.6
0.5
2
~ 0.4
~
0.3 -7)a-:4--
0.2 --- -----
0.1
-----------
0.0 '":---~---'-----'----'-----
0.1 2.0 4.0 6.0 8.0 10.0
EplER
Fig.3.3 Interface shear stress per unit applied axial stress ('!'p/O'zo) near the loaded
end of plug as a function of modulus ratio (Ep/ E R) and plug length-to-radius ratio
(L/a).
'CIGzo
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0.0 r----r--,--,---.---r----.....----r-.....---.-....,
Ua=4
0.1
0.2
0.3
0.6
0.7
0.8
- _ _ Finite element analysis
0.9 Analytical solution
1.0L.-----------------I
Fig.3.4 Interfacial shear stress distribution per unit applied axial stress (Tj(JzO) for
a modulus ratio (Epj EJ of 0.23.
3.4.3 Tensile stresses

An axial stress applied to a borehole plug creates tension both in the plug
and in the rock. The determination of these tensile stresses is important
since any tensile fracturing may cause preferential pathways around seals.
The finite element program, PLANE-2DFE, was used to determine the
tensile stress distribution within an axially stressed push-out specimen. The
finite element analysis simulates a push-out test with an axial stress applied
to the plug of 1.0 MPa. The finite element meshes given by Akgun and
Daemen (1991c) were used to analyze the principal tensile stress distribu-
tions. The meshes represent 13 mm radius borehole plugs with length-to-
radius ratios of 2.0, 4.0 and 8.0, respectively. All meshes have cylinder
outside-to-inside radius ratios of 60 and represent borehole plugs in an
infinite medium.
Figure 3.5 gives the zones of tensile radial and tensile tangential stress in
a specimen with a length-to-radius ratio of 2.0 and a modulus ratio of 0.23,
representative of an axially stressed Self-Stress II cement grout plug em-
placed in an Apache Leap tuff cylinder. The tensile radial stress zones (on
the right-hand side of Figure 3.5) and the tensile tangential stress zones (on
Fig.3.5 Percentage normalized radial stress (ar/a zo ; right) and tensile tangential
stress (a 8 /a ZO; left) contours for an axially stressed borehole plug in rock. Modulus
ratio of 0.23 and plug length-to-radius ratio of 2.0.
the left-hand side of Figure 3.5) are plotted as a percentage of the axial
stress applied to the plug. Figure 3.6 gives the tensile stress distribution for
a plug length-to-radius ratio of 8.0. The magnitudes of the tensile stresses
and the volumes under tension decrease with increased plug length.
Table 3.1 summarizes the principal tensile stresses in three critical loca-
tions where at least one of the principal stress components gives maximum
tension. The tensile stresses are presented as a function of the applied axial
stress and plug length-to-radius ratio.
The modulus ratio of the push-out specimen in Table 3.1 is taken to be
0.23, which is representative of a Self-Stress II cement grout plug emplaced
within an Apache Leap tuff cylinder. The Poisson's ratios of the plug and
rock are 0.22 and 0.20, respectively. The tensile stresses are reckoned
negative.
Table 3.1 shows that the maximum tension in the plug occurs at the top
corner. At this location only one stress component is tensile. The bottom
center of the plug is in biaxial tension. The maximum tension in rock occurs
at the upper contact between plug and rock. At all three critical locations
the magnitude of tension decreases with increasing plug length-to-radius
ratio. No axial tensile stresses are observed within the plugs.
The most likely axial stress on a borehole plug in rock is due to water
pressure. If a borehole in which the plug is emplaced at 1000 m below the
surface fills up with water, this creates a water pressure of 9.8 MPa on the
plug. This water pressure creates a maximum tensile stress of 6.7, 4.9 and
Fig.3.6 Percentage normalized radial stress (aria zo; right) and tensile tangential
stress (aolazo; left) contours for an axially stressed borehole plug in rock. Modulus
ratio of 0.23 and plug length-to-radius ratio of 8.0.
Table 3.1 Principal tensile stresses in plug and in rock at three criticalloca-
tions a
Plug length Top corner b Bottom corner C In rock d adjacent
to radius ratio of plug of plug to top corner of
plug
2.0 -69 -11 -76

4.0 -49 -1.2 -60
8.0 -26 -0.2 -41
"The results are given as a percentage of the axial stress applied to the borehole plug
(0"3/0" z); buniaxial tension; cbiaxial tension; dtriaxial tension.
2.6 MPa at the top comer of plugs with length-to-radius ratios of 2.0, 4.0
and 8.0, respectively (Table 3.1). The mean tensile strength of medium-
strength concrete is 2.9 MPa (Neville, 1981). Therefore, tensile failure of plugs
with length-to-radius ratios of 2.0 and 4.0 are very likely. The tensile strength
of plugs with length-to-radius ratios of 8.0 exceeds the maximum tensile
stress developed at the top comer of the plug. Hence, a plug with a length-to-
radius ratio of at least 8.0 should be used to avoid tensile fracturing.
3.5 PUSH-OUT TEST RESULTS AND DISCUSSION
Figure 3.7 shows the applied axial stress vs. the top and bottom plug
displacements of a typical push-out test. In all push-out tests the behavior
was elastic. The bottom plug displacements were small as compared to
Push-out test results and discussion 37
the top axial displacements prior to bond failure. Upon failure the differ-
ence between the top and bottom axial plug displacements decreased, prob-
ably due to stress relief. The push-out specimens failed after periods ranging
from 3 to 30 min. The tests were continued for up to 2 h. Residual axial
strengths ranged from 10 to 33% of the peak axial strengths. On
average, the tests gave a residual axial strength of about 20%. One of the
51 mm radius push-out samples showed tensile splitting during push-out
testing.
Table 3.2 gives the mean strength measures of the push-out specimens
tested. The axial strength, bond strength and the peak shear strength are the
applied axial stress at failure, the average shear stress at failure and the peak
shear stress at failure, respectively. The bond strength and the peak shear
strength are calculated from Equations (3.1) and (3.3). Values in square
brackets represent strength measures that are normalized to an infinite rock
mass. Equation (3.4) is used to obtain the normalized axial strength. The
normalized axial strength is used in Equations (3.1) and (3.3) to obtain the
normalized bond strength and normalized peak shear strength, respectively.
All three strength measures decrease with increasing plug radius and with
decreasing plug length.
Equations (3.5) and (3.6) give the best fit of the normalized axial strength
(or the applied axial stress at failure normalized for an infinite rock mass) as
45.0
I - - - Top plug displacement

40.0 Bottom plug displacement
I
I
35.0
~ 30.0
rf
~
gj 25.0
I
w I
~
...J
I
20.0 II
(
\
~ \
\
0
w
::::;
c..
15.0 \
\ ,,
c..
(
10.0 ""...,
5.0
0.0 ':-_-'::_ _J........_--1.._ _..1.-_--I._----I

0.0 0.5 1.0 1.5 2.0 2.5 3.0
AXIAL PLUG DISPLACEMENTS (mm)
Fig.3.7 Applied axial stress vs. axial plug displacements of a push-out sample.
Table 3.2 Mean strength measures of push-out specimens one standard

deviation'
Plug Plug Axial Bond Peak
radius length to strength strength shear
(mm) radius ratio (MPa) (MPa) strength (MPa)
6.4(14) 2.0 51 6.2 11 1.2 354.1

[526.3J [11 1.2J [35 4.1J
13(6) 2.0 354.4 7.3 1.0 243.0
[354.4J [7.4 1.0J [243.0J
13 (5) 4.0 669.0 8.1 1.0 446.1
[679.1J [S.2 1.0J [446.2J
13 (3) S.O 1702.5 11 0.4 115 1.7
[1722.5J [11 O.4J [116 1.7J
25 (3) 2.0 23 5.5 6.3 1.9 204.8
[24 5.8J [6.62.0J [21 5.0J
51 (3) 2.0 151.7 3.90.7 17 1.9
[18 2.0J [4.60.8J [202.2J
"The number of push-out samples tested is given in parantheses; values in square
brackets represent strength measures normalized to an infinite rock mass.
a function of plug radius and plug length-to-radius ratio, respectively. Both

equations obey a power law. Equation (3.5) is valid for a plug length-to-
radius ratio of 2.0; Equation (3.6) is for a plug radius of 13 mm:
O"n = 133 a- O . 52 r = 0.90 (3.5)
O"n = 16 (Lja)1.1 r = 0.98 (3.6)
where O"n = normalized axial strength (MPa), a = plug radius (mm),
Lja = plug length-to-radius ratio and r = correlation coefficient.
Equations (3.5) and (3.6) show that the normalized axial strength de-
creases with increasing plug radius and with decreasing plug length. Size
effects on the normalized axial strength might follow from the size-strength
studies performed on pillars (e.g. Farmer, 1985). When a fracture occurs in a
cubic pillar with a side length, L, a certain amount of strain energy, U, is
required to satisfy the energy balance for each unit area of the fracture
surface created. If the fracture results from brittle breakdown and the strain
energy per unit fracture area (U j L 2) is assumed to be constant, the strength
of the pillar is inversely proportional to the square root of one of the pillar
linear dimensions (or the plug diameter in the case of a push-out specimen).
This might explain the inverse proportionality between the axial strength
and plug radius.
Acknowledgements 39
3.6 CONCLUSIONS AND RECOMMENDATIONS
The analysis performed herein shows that a borehole plug with a modulus
ratio of 0.23 and plug length-to-radius ratio of 8.0 has a higher axial
strength and shows little probability of tensile failure when compared to
shorter plugs. Plugs with smaller radii and greater lengths give higher axial
strengths and lower peak shear stresses. The axial strengths represent lower
bounds due to the absence of confining pressure.
As Equation (3.3) presented for calculating the peak shear stresses is
linearly elastic, it overestimates the peak shear stresses and is conservative
for plug design. The calculation is further conservative due to the utilization
of zero confinement and due to ignoring progressive failure which leads to
reduced peak shear stresses. It is recommended that plug design be based
on limiting the elastic peak shear stress to well below the peak shear
strength. Some guidance about the peak shear strength is given in the last
column of Table 3.2. A conservative design is believed to result if the
strengths of the 13 mm radius plugs in Table 3.2 are divided by a factor of
22 for full size (e.g. 5.0 m radius) plugs. It is also recognized that the factor
22 is uncertain as it is based on an extrapolation far beyond the range
over which measurements have been made. This uncertainty confirms the
desirability of performing some experiments on larger radius borehole
plugs.
The results of the detailed numerical and experimental analyses of the
mechanical performance of seals performed by Akgun and Daemen (1991c)
and some reported herein indicate that permanent abandonment plugs
should be designed with a length-to-radius ratio of 8.0. This conservative
length criterion will prevent the development of excessively detrimental
tensile stresses within and near an axially loaded borehole plug. It is of
considerable interest that a very similar geometrical design recommenda-
tion results from detailed hydrological analyses of water flow through plugs
and host rock (Greer and Daemen, 1991), assuming reasonably similar
hydraulic conductivities for the plug and the rock. This may not be surpris-
ing in the light of the parallelism between the governing equations for fluid
flow and for elastic stress analysis.
ACKNOWLEDGEMENTS
This work was part of a research effort sponsored by the US Nuclear

Regulatory Commission, under contract NRC-04-86-113. Support and per-
mission to publish this chapter are gratefully acknowledged.
CHAPTER FOUR
Dynamic Loading Impact on

Cement Borehole Seals
G. S. Adisoma
4.1 INTRODUCTION
One of the concerns with regard to the performance of borehole seals is the
impact of earthquakes, or other types of dynamic loading such as large-
scale blasting, on seal integrity. Of all possible scenarios (erosion, glaciation,
tectonic and other natural processes), seismic motions are the most likely
effects to be experienced by borehole and excavation seals. In the United
States many locations are sufficiently close to seismic regions to be affected
thereby (as can be seen from the seismic hazard map shown in Figure 4.1),
while sufficiently far not to experience major effects such as rock fall, slipp-
age and/or breakage. The map in Figure 4.1 indicates the effective peak
acceleration that might be expected to be exceeded during a 50 year period
with a 10% probability.
A review of available literature indicates that deep underground struc-
tures in competent rocks experience less damage than surface structures,
openings at shallow depth and openings in fractured rocks, when subjected
to earthquakes and subsurface blasts. This is based on surveys on the effect
of earthquakes on wells, tunnels, mines and other underground structures
(Nazarian, 1973; Stevens, 1977; Dowding, 1977; Dowding and Rozen, 1978;
Pratt et ai., 1979; Owen and Scholl, 1981; Marine, Pratt and Wahi, 1982), as
well as the impact of commercial blasting, underground explosion tests
(using conventional TNT explosives) for the US Army Corps of Engineers,
and subsurface nuclear blasts in connection with Project Hard Hat (Holmes
et aI., 1963; Bauer and Calder, 1971; Langefors and Kihlstrom, 1978, Owen
and Scholl, 1981; Oriard, 1982; Labreche, 1983; Holmberg, Larsson and
Sjoberg, 1984).

Introduction 41
Fig. 4.1 Seismic hazard map of the United States, showing contours of effective
peak acceleration in units acceleration of gravity. (From Applied Technology
Council, 1978.)
While the information may provide considerable guidance about the like-
ly response of sealed openings to dynamic loading, the foregoing review also
shows that specific data on the effect of dynamic loading on rock and seal
permeability are virtually non-existent. Yet, there is a considerable need for
this kind of information. Borehole sealing has long been used in the petro-
leum industry to isolate and stabilize the lower portion of a producing well.
More recently, a similar technique has been adopted by the solution mining
industry. Many abandoned oil and gas wells have been sealed with cement
grout. However, no data are currently available regarding their sealing
performance with respect to seismic loading.
This information is also crucial in sealing abandoned underground min-
es to prevent groundwater migration into the mine and the leakage of
harmful acid to the outside environment. Underground space utilization
for the storage of oil, natural or liquefied gas, chemicals, wastes and others
requires proper sealing of all access boreholes and wells to maintain the
integrity of the storage caverns. Burial of high-level nuclear waste in geo-
logic media poses perhaps the most important challenge for borehole
sealing application. This is due to the stringent requirement that all access
pathways to the deep underground repositories must eventually be sealed,
and the sealing performance must not be compromised by events such as
earthquakes.
42 Dynamic loading impact on cement borehole seals
Despite the diversity in application, borehole and underground excava-

tion seals generally hold a similar function: as a protective measure in
containing often volatile commodities or their byproducts (acid, crude oil,
radionuclides, etc.) from seeping out into the environment. In many of these
engineering applications, groundwater migration seems to be the most like-
ly scenario for the breach in sealing.
In this chapter an experimental sealing performance assessment of cement
borehole plugs that have been subjected to dynamic loading is provided.
This includes a study of plugs that have dried, as well as of plugs that have
remained wet throughout the testing period. The basic approach in this
study was to establish steady-state flow through a cement borehole plug in
a rock cylinder, subject the plugged cylinder to dynamic loading using a
shaking table, and assess the influence of shaking on plug permeability. The
parameter measured during the flow test prior to and after dynamic loading
was the flow rate at various injection pressures. This was used to determine
hydraulic conductivity of the cement plug. In the dynamic loading phase the
tests were carried out with increasing duration (up to 5 min) and peak
acceleration (up to 2 g).
The experimental work performed in this chapter is based on the assump-
tion that external shaking of a plugged rock cylinder will adequately simu-
late some aspects of dynamic waves impacting a sealed opening. Of
importance in sealing performance testing, analysis and design consider-
ations is the scaling effect of size, which will influence the acceleration and
duration of the dynamic loads as well as the dynamic displacement and
diffusion processes that control fluid flow.
4.2 EXPERIMENTAL METHODS
A total of eight granite cylinders of Precambrian age (from Charcoal Black

Quarry in Minnesota) were used in this study. Their typical dimensions
were 15 cm diameter by 30 cm long. Each cylinder had a 2.5 cm diameter
hold drilled coaxially from both ends. In the middle portion of the hole,
either a rock bridge was left (for baseline rock permeability measurements)
or a cement plug was installed to seal the borehole.
The cement plug consisted of type A Portland cement, mixed with 50%
distilled water, an expansive agent and a dispersant. The cement mixture
was poured on top of a stopper in the hole and cured for at least 8 days
under water, at room temperature and atmospheric pressure. These plugs,
as well as the rock bridge, were of varying lengths, from 3 cm to 11 cm.
Four of the cement plugs remained cured under water (wet) until testing
commenced. Three others were allowed to dry (after curing) at ambient
room temperature for a period of 3-7 months, before they were rewetted
and tested. Upon completion of the test one of the wet plugs was oven dried
to simulate rapid plug drying from the heat generated by some external
Experimental methods 43
source (e.g. geothermal heat flow or heat generated from nuclear waste); this
specimen was then retested.
During the flow test, which lasted up to 9 months, distilled water was
injected under pressure on top of the plug and the outflow was collected
underneath it. Hydraulic conductivities were calculated from the measured
flow rates through the plug (or rock bridge). The schematic diagram of flow
test lay-out is described in detail in Figure 4.2.
Once a long-term steady-state flow trend had been established, the cylin-
ders were subjected to dynamic loading on a shaking table. Shaking was
performed at various accelerations (up to 2 g), and for different durations
(up to 5 min). Flow testing was continuing before, during and after the
shaking. Figure 4.3 shows the dynamic loading test set-up.
During the later stage of post-shaking flow test, a dye solution was
injected into four rock cylinders ~ two with wet cement plugs and two
having dried cement plugs. Upon completion of the test these specimens
were sawed in half lengthwise and visual observations of the flow patterns
were conducted. Test procedures, equipment and materials used in this
study are described in detail elsewhere (Adisoma, 1987; Adisoma and
Daemen, 1988).
L R
10
4
I. Nitrogen oos tank

2. Pressure reoulator
3. Low-pressure(oos) cylinder of pressure intensifier
4. High-pressure(water) cylinder of pressure intensifier
5. Woter injection pressure goge
6. Rotameter (flowmeter)
7. Rock sample
8. Borehole plug
9. Measuring pipets for outflow collection
10. 0101 oaOe for piston displacement measurement
II. Stainless steel connector
12. Rubber stopper
Fig.4.2 Schematic diagram of flow test layout. Using two Permatex-sealed rubber
stoppers, the longitudinal flow through the plug (used to calculate hydraulic conduc-
tivity) and the peripheral flow through the rock around the plug are collected
separately in the Rand L pipets, respectively.
Fig.4.3 A plugged rock cylinder sitting on top of a shaking table is shown in the
dynamic loading test configuration (foreground). Four plugged rock cylinders un-
dergoing simultaneous flow testing are visible in the background.
4.3 FLOW TEST RESULTS
During each individual flow test (which may last from a few minutes for a
dried plug to a couple of days for a wet plug), inflow and outflow were read
at a given time interval. The flow rate was determined by linear regression,
i.e. the slope of outflow vs. time plot (Figure 4.4). Injection pressure is
maintained constant during each flow test.
For a given specimen the flow test was repeated many times, often using
different injection pressures. The resulting flow rate from each test was
plotted as a function of the elapsed time (in days) since testing on that
specimen commenced. Figure 4.5 shows an example of such plot (a rock
cylinder with a dried plug, the same specimen used in Figure 4.4). Each data
point in Figure 4.5 is a result of (and represents the flow rate from) a single
flow test described in Figure 4.4.
4.3.1 Hydraulic conductivity determination

Flow rate through a medium is one method of quantifying its sealing per-
formance. Since flow rate depends on the length of the medium and on the
injection pressure, which are not uniform during the flow test, hydraulic
conductivity (K) is used to quantify and compare the effectiveness of various
1.6.----------------------------------r~
1.2
1:
u
3: 0.8
0
-1
lJ..
0.4
~
~
10 20 30
TIME (MIN)
Fig.4.4 Result from a flow test, showing outflow plotted as a function of time.
Flow rate is determined from the slope of the best-fit line. This is an example from a
dried cement plug; 1.5 MPa injection pressure.
'"~
'"
'"~
'"
'"~
z
;:
.......
0
w
.... '"~
'"'" '"
a:
~
'"c '"~"j
..J
'"~
N ,
...'"'0
~
50 100 150 200 250

ELRPSED TIME (DRYS J
Fig.4.5 Flow rate through a dried cement plug as a function of total test time for
the specimen (the same specimen as in Figure 4.4). Injection pressure was 1.5 MPa.
sealing materials. Also known as coefficient of permeability or simply per-

meability, it is calculated from the flow rate using Darcy's law for one-
dimensional flow through a porous medium, and is independent of distance
and pressure gradient. Its derivation can be found in most hydrology tests
(e.g. Harr, 1962; Freeze and Cherry, 1979). Hydraulic conductivity has a
unit of velocity; often times darcy is also used (1 darcy is approximately
equal to 10- 3 cmls for water at 20DC).
For flow through a preferential path such as that of dried cement
plug-rock interface (section 4.4), a single fissure flow model, rather than
Darcy's law for uniform flow through a porous medium, is more appropri-
ate (Snow, 1968; Zeigler, 1976). The model can be conveniently analyzed by
the equivalent parallel plate concept. Darcy's law can then be applied to
determine a coefficient of fissure permeability (K j ) as a function of an
equivalent parallel plate aperture.
4.3.2 Permeabilities of Charcoal granite and cement seals

To provide a benchmark for measuring cement plugging effectiveness,
the permeability of Charcoal granite must be determined. Figure 4.6a
shows the permeability of a granite rock bridge specimen. It indicates a
very low permeability rock, from 4 to 30 nanodracy (4 x 10- 9 to
3 x 10 - 8 darcy).
Figures 4.6b, c and d show the permeabilities of wet cement seals in three
specimens. During 9 months of flow testing the values remain fairly con-
stant with time. One exception is the specimen in Figure 4.6d, which shows
a decreasing permeability with time. Compared to the permeability of gran-
ite in Figure 4.6a, wet cement seals are equal to or less permeable, with
hydraulic conductivity of between 0.4 and 30 nanodarcy.
Figure 4.7 shows permeability vs. time plots for two specimens whose
cement plugs were allowed to dry after curing. These plugs dried out in
ambient room temperature (7 months for the specimen in Figure 4.7a, 3
months for that in Figure 4.7b) prior to flow testing. This caused the
cement plug to shrink and the plug-rock interface to open, creating a
preferential flow path. The dye injection test (section 4.4) verified the
existence of this preferential flow path (see also Adisoma and Daemen,
1984).
The plots in Figure 4.7 indicate that fissure permeabilities decrease rapid-
ly during the first 2 months of testing. They continue to decrease, albeit at a
slower rate, until the flow test is concluded 4-6 months later. This suggests
renewed cement expansion upon resaturation, which results in partial clos-
ing of the gap in the plug-rock interface. However, the final permeabilities
are still several orders of magnitude higher than those of cement plugs
which are always wet. During this period fissure permeability decreases
from 20 to 0.35 darcy (Figure 4.7a), and from 0.95 to 0.012 darcy (Fig-
ure 4.7b).
Flow test results 47
8
of
i
~ 'I'
Ii!

8
or
Ii!
or
8
.;
'0 60 120 180
ELAPSED TU'IE 'OAYS J
240
(a)
8
ri
Ii!
ri
E!
'I'
i Ii!
of
"
8
or
Ii!
or
8
.;
'0 60 120 180 240 300
(b) ElAPSEO TIllE 'QAYS'
Fig.4.6 Results of permeability measurement in (a) Charcoal granite and (b)-(d) in

three specimens having wet cement seals. The results indicate a very low-permeabil-
ity granite, and an even lower-permeability cement seal when maintained wet
throughout.
4.3.3 Summary of permeabilities

The experimental results are summarized in a simplified plot shown in
Figure 4.8. This composite plots shows the permeabilities of Charcoal gran-
ite and of wet and dried cement plugs as a function of time. The values
obtained here compare favorably with granite and cement permeabilities
g
of
8
of
Iii
of
~
'0 120 lID 240
(c) ElNSEO tillE I DAYS,
8
~
Iil
~
~
m
Iil
;
8
of
Iil
of
~
'0 60 120 180 240 300
(d) ELA'SEO TIME lOAYS'
Fig. 4.6 (C), (d).
obtained by other investigators (e.g. Brace, Walsh and Frangos, 1968; Cobb,
1981; South and Daemen, 1986).
Drying the cement plugs increases their previously very low permeabili-
ties by seven to nine orders of magnitude, depending on the drying period
and temperature. The dried cement plugs exhibit a similar response when
they are resaturated. The flow rates decrease rapidly for the first 2 months,
and continue to decrease at a slower rate thereafter. An oven-dried cement
Flow test results 49
L.l
;;"') ci
S
8
.;
Ii!
?
l'l
'0 50 100 150 200 250
ELAPSEO TItlE IDAYS I
(a)
l'l
Ii!
.;
j~
,,<' ,

!l
,
Ii!
,
~'0
50 100 150 200 250
(b) EUl'SED Til (~YSI
Fig.4.7 Fissure (interface) permeabilities of dried cement plugs as a function of

time. The permeability of specimen in (a), which was dried for 7 months, is 25 times
that in (b) which was dried for 3 months. These values are several orders of magni-
tude higher than those of wet plugs.
plug in basalt tested by South and Daemen (1986) exhibits very similar
behavior. Drying is potentially very detrimental to cementitious plugs.
Sealing performance is only partially recovered when the plugs are re-
saturated.
/Cement plug room-clried

for 7 months
ement plug
~oven-dried at
90C for 5 days
"
10- 3 md
/Granite rock
10'"
/Soturated cement plug
f-----'---
IO-IOO!;-t;""~IOO;;-;;'''''!;;-;2;;;;OO~2';;-0-;:300~300
Elapsed Time (days)
Fig.4.8 Permeabilities of Charcoal granite, wet cement plugs and dried cement
plugs as a function of time.
4.4 DYE INJECTION TEST RESULTS
A liquid concentrate dye marker was injected into the permeant (water)
during the later stage of flow testing in two rock cylinders with wet cement
plugs. Dye was also injected into a rock cylinder with a room-dried cement
plug and another with an oven-dried cement plug. This test was performed
upon completion of the dynamic loading test and the subsequent post-
shaking flow test. After the dye injection test had been concluded the rock
specimens were sawed in half along their lengths. This allowed visual obser-
vation of the flow pattern of both wet and dried plugs, as well as verifica-
tion of results obtained from the flow tests.
Figure 4.9a shows the cross-section of a rock cylinder with a wet cement
plug. Dye was injected during the last 41 days of flow testing on this
specimen. The dye-colored permeant (water) penetrated the cement grout
seal uniformly. The plug body remains intact and no visible cracks can be
observed. Preferential flow path along the plug-rock interface is non-exist-
ent. This observation confirms the very low flow rate observed during flow
testing on this specimen. Judging from the similar flow rate in other speci-
mens with wet cement plugs, this flow pattern seems to be typical for all wet
cement borehole seals as well.
Figure 4.9b is a photograph of a sawed half of a rock cylinder containing
a dried cement plug. This cement plug was left to dry at room temperature
for 7 months and was subsequently flow tested for 8 months. In this
Dye injection test results 51
(a)
(b)
Fig.4.9 Cross-section of (a) a wet cement plug and (b) a dried cement plug in
Charcoal granite. Dye-colored permeant penetrated the wet cement plug body uni-
formly, without any preferential flow path (a). In the dried plug, dye traces in the
interfacial fissures and in a crack across the plug body indicate preferential flow
paths (b).
particular specimen dye was injected during the last 39 days of the flow test.
The photograph clearly shows traces of dye along the plug-rock interface.
The dye also highlights a crack which is visible across the plug body,
extending from the left to the right interface. This crack and the interfacial
fissures clearly acted as preferential flow path, since none of the dye marker
penetrated the main body of the plug.
This experiment has shown that when a cement seal is dried it shrinks
and cracks. This is especially true in cement grout with an expansive agent
added to it. The rate of shrinking and cracking is apparently related to the
temperature during the drying process and the duration of drying. Cement
shrinkage causes the plug-rock interface to open, thus creating a preferen-
tial flow path. Cracking in the plug body creates additional flow paths. The
presence of preferential flow paths seems to be typical in dried cement
plugs. They explain the high flow rates observed during the flow tests on
these specimens.
4.5 DYNAMIC LOADING TEST RESULTS
Rock cylinders with wet and dried cement seals were subjected to dynamic
loads on a shaking table during the later stage of the flow test. This is
especially critical for the dried plugs, where the rapidly decreasing flow rate
with time (Figure 4.7) as a result of the cement's renewed expansion during
the early part of the flow test may hinder the interpretation of the dynamic
loading test results. Flow tests were conducted prior to and after the appli-
cation of dynamic loads. Therefore, the resulting flow rates can be com-
pared directly.
The following summarizes the dynamic loading test condition:
acceleration amplitude from 1 to 2 g;
velocity amplitude from 6.4 to 10.4 mls (21 to 34 ft/s);
displacement amplitude from 2.8 to 3.8 cm (1.1 to 1.5 in);
motion frequency from 2.6 to 3.6 Hz;
dynamic load duration from 20 to 326 s;
injection pressure from 1 to 4 MPa.
Figure 4.10 shows the longitudinal flow rates through the plug, at differ-
ent injection pressures, prior to and after the application of dynamic loads
of various durations (indicated by the dashed vertical lines). Figure 4.10a is
an example for a wet plug and Figure 4.10b is for a dried plug. The rock
cylinder in Figure 4.10a was subjected to dynamic loads at an acceleration
of 2 g and the injection pressures ranged from 1 to 4 MPa. For the dried
plug in Figure 4.10b the acceleration was 1 g and injection pressures varied
between 1 and 4 MPa. The flow rates in Figure 4. lOb is two orders of
magnitude higher than those in Figure 4. lOa. Yet, in both cases they re-
mained practically unaffected by the application of dynamic loads.
0
~
dynamic loadings at a-2g

0 {('20s t 2""40s t 3-805 t 4-1605 t 5'" 300s
II?
~'...\
I I I
/' /
I I
I I I 1 I
p= 4 MPa 1 1 I 1 I
0 1 1 I I
.4 ~ 1 I I 1
I
~ I : :
x I I
:!':
,.r,'- 0
r
u -
II?
i'a::!
'"
'"~
0 0
~
0
II?
0
0
0
140 162 184

ELRPSED TlMEI DRYS)
206 228 250
(a)
Oynamic loadings at occeleration=I.Og

tc=20s tc=45s tc=86s tc=166s tc=326s
I
I
j :I
I
:I
1 I I I
1 I 1 1
I : I

-I
c I
E I
;;;. 1
E I I
~ I I
III I 1 1
-
l-
<t 10-3
I I 1
0
Ir 0 1 a a 10 a :
1
!it P=2MPo a I 1 I au
0 I a a
-'
LL
I a a
0
o
o
10-~'-:IO-----'1~20--~13~0-15-t7-1~40-14-+7~15-0-15f-4-16~0-1631-~17-0-1-77-1-1~80-
(b) ELAPSED TIME (Days)
Fig.4.10 Longitudinal flow rates through the plug before and after dynamic loads
are applied, for a specimen with (a) a wet cement plug and (b) with a dried cement
plug. Flow rates at various injection pressures are not affected by dynamic loadings
(indicated by the dashed lines) administered for durations of up to 5 min.
In Figures 4.11 and 4.12 the detailed impact of shaking is presented in

terms of the change (or lack thereof) in plug permeability. Figure 4.11 shows
a permeability vs. time plot for the same specimen shown in Figures 4.6b
and 4. lOa. The dashed vertical lines represent the application of dynamic
loads for the durations indicated (20- 300 s). A peak acceleration of 2 g was
used throughout. The peak particle velocity was 10.4 mls (34 ft/s), at a dis-
placement amplitude of 3.8 cm (1.5 in) and a motion frequency of 3.6 Hz.
The plot indicates that subjecting a wet cement plug in a granite cylinder to
-7.00..,-----------------,
SP[CI"'EN 005309-06
Dynamic: Iooding' at a _ 2 II
20, 40, SO, 1!iOt 300,
-7.50
cC,
~ c :1:1
C?::C II
~ :uoo ~n i
Ci 10
-8.SO
TIME (DAYS)
Fig.4.11 The effect of dynamic loading at a = 2 g on the permeability of a wet

cement grout seal. The change observed in seal permeability as a result of dynamic
loading is less than the permeability variation without dynamic loading. Dashed
vertical lines denote the time of dynamic load application.
-7.00 . , - - - - - - - - - - - - - - - - - ,
, -,,
SPEOWEN ~309-00
Oync:lmie loading_ Qt
-40. 80. 160, '20
-'.eo
1-.. 00
,' . ,
iT
~
a-
~
IU C
C c
-8.50
-IiI,OO
(a)
"0 '50 "0
nME (DAYS)
no ,'"
Fig.4.12a
Dynamic loading test results 55
-.O O - y - - - - - - - - - - - - - - - - - ,
SPECIt.IEN CG5J09-31\1
~om!c: loodinQ' at a .. I %
""~l(
: b:
-9.50
-10.00
,'" 200
TIME (DAYS) '''' '00
(b)
-0.50
,,
SPECIMEN CG5JD9-21!
Dynomie Ioodi~. Ilt 0'"
20. 4~ 8 1!6 8 166 8 326.
t;
r
L.",
,,- 0,,- C
0' c 0, i
"9 iW U cO' ~ice
-2.00
"i~ T

DtiJ
(c)
-2'"
'" 19'
"" '" "0 '50
TIME (DAYS)
"0
1.00
SPECIIilEN CGSJ09-Dl
4081)y80~io~dI,1j. CIt a .. 2 ~o.
'.00
t;
I
>c.-
S
'.00
"-, ' . yO
-1.DD
12' ,... '" 10'
(d) TIME (DAYS)
Fig.4.12 The impact of dynamic loading on (a) and (b) wet cement grout seals and
(c) and (d) dried cement grout seals in granite specimens. All dynamic loads were at
a = 1 g, except for specimen in (d) where a = 2 g. No significant change can be
observed in both the permeability of the wet cement seals and the fissure permeabil-
ity of the dried cement seals after dynamic loads were applied.
dynamic loads of this magnitude does not change the plug permeability by
more than its variation prior to the application of dynamic loads.
Figure 4.12 shows the results for the remaining specimens. Two granite
cylinders with wet cement plugs in Figures 4.12a and 4.12b (the same speci-
mens shown in Figures 4.6c and 4.6d, respectively) were tested in tandem
atop the shaking table. These specimens were subjected to dynamic loads
for up to 320 s. The peak acceleration was 1g and the peak particle velocity
was 7.3 mls (24 ft/s) at 3.8 cm (1.5 in) stroke length and 2.6 Hz frequency. The
permeabilities of these plugs were unaffected by shaking. Figure 4.12b
shows that some permeability data before and after dynamic loading appli-
cation are missing. This is because no outflow could be detected during the
dynamic loading cycle of this specimen, due to problems in measuring very
low flow rates (at the lowest limit of resolution of flow test instrumentation).
The specimen in Figure 4.12c (the same specimen shown in Figures 4.7b
and 4.10b) was dried for 3 months at room temperature. It was subjected to
five shaking cycles during the last 7 weeks of flow testing, for durations
ranging from 20 to 326 s. A peak acceleration of 1g was used, at a stroke
length of 2.9 cm (1.1 in) and a nominal frequency of 3 Hz; the peak particle
velocity was 6.4 mls (20.9 ft/s). During this period the fissure permeability
remained practically unchanged.
The specimen in Figure 4.12d (the same as in Figure 4.7a) had a cement
plug that was dried for 7 months at room temperature. Dynamic loadings
were applied in five cycles, ranging in duration from 40 to 300 s. A 2 g
acceleration was used throughout. At a peak displacement amplitude of
3.8 cm (1.5 in) the frequency was 3.6 Hz and the peak particle velocity
was 10.4 mls (34 ft/s). The result shows that even at the worst drying
condition, the fissure permeability of dried cement plugs in granite is
unaffected by shaking under the most severe condition experienced in
this test.
Results presented in this section apply only for longitudinal flow through
the plug. The results of dynamic loading on the peripheral flow through the
surrounding rock is given in Adisoma (1987) and in Adisoma and Daemen
(1988), as well as details of the test for each specimen. The general con-
clusion is that repeated application of shaking does not affect the peripheral
flow rate through the rock surrounding the plug.
4.6 DESIGN CONSIDERATIONS
This section deals with some general recommendations on the design of

borehole and underground excavation seals under a dynamic loading envi-
ronment. The content is by no means exhaustive and is limited to several
factors worthy of consideration in the design process.
Design considerations 57
4.6.1 Dynamic response of underground openings

Underground structures in rock such as tunnels and mine shafts are differ-
ent from surface structures, for example buildings, bridges and others. With
the latter, the geological medium forms only the foundation for the struc-
ture, while for tunnels and other openings in rock, the geological medium is
a major component of the structure. In cases where the rock does not
require support or reinforcement, the geological medium is the structure.
The response of underground structures such as tunnels to seismic motion
may be understood in terms of three principal types of deformation: axial,
curvature and hoop (Figure 4.13).
Axial and curvature deformations develop when seismic waves propagate
parallel or obliquely past a tunnel. Axial deformation induces alternating
regions of compressive and tensile strain that travel as a wave train along
the tunnel axis (Figure 4.13a). Curvature deformation creates alternate re-
gions of negative and positive curvature propagating along the tunnel (Fig-
ure 4.13b). Hoop deformation results when waves propagate normal or
Tension Compression
(a)
Negative Curvature
(b)
Tunnel Cross Section

before Wave Motion
(c)
Tunnel during
Wave Motion
Fig.4.13 The principal types of deformation in a tunnel due to seismic motion. (a)
Axial deformation along tunnel, (b) curvature deformation along tunnel and (c)
hoop deformation of cross-section. (After Owen and Scholl, 1981.)
nearly normal to the tunnel axis. One effect of the hoop deformation is the
distortion of the cross-sectional shape (Figure 4.13c).
4.6.2 Data gathering for design

The following factors have some impact on the design of borehole and
excavation seals in the dynamic loading environment. In order to under-
stand better their influence, quantitative measurements and data gathering
should be conducted to gather baseline data for design purpose.
Seismic site investigations

Information such as the seismic hazard map shown in Figure 4.1 is essential
for designing underground facilities that require sealing. For a given site the
expected maximum ground motion is determined based on historic earth-
quake records. This, together with the design response spectra specified for
the facility, is used to prescribe the motion for which the structure must be
designed. It is important to note, however, that such a probabilistic express-
ion of seismic hazard is based solely upon seismic history. To make this
information more useful, the distribution of active faults must be consider-
ed. It is important that seismic design response criteria be developed and
incorporated into the design of facilities. Unfortunately, the state of the art
in seismic design technology in rock is still poorly developed (Owen and
Scholl, 1981).
Dynamic deformations
Axial deformation parallel to the openings (Figure 4. 13 a) is likely to induce
differential longitudinal (axial) strain between the opening wall and the seal.
The most obvious test configuration to simulate this condition is push-
out/pull-out testing. This could include cyclic loading at a range of frequen-
cies and amplitudes corresponding to the likely upper limits for in situ
differential strains. Push-out tests could be performed by having a plug,
contained in a hole within a fixed rock cylinder, loaded by a rigid cylinder.
It can also be performed with a pulsating liquid or gas pressure applied on
one side of the plug. The latter would facilitate concurrent application of
the dynamic load with fluid flow (permeability) testing, but could be per-
formed only in compression.
Curvature effects (Figure 4.l3b) are unlikely to be significant for shaft or
drift plugs, given the short length of the plugs with respect to opening
diameter and especially to wave lengths. For borehole studies they could be
investigated most readily by subjecting sealed boreholes to cyclic bending.
In addition, it would be preferable to run flow tests on the plugs prior to,
during and after the dynamic loading.
The transverse relative deformation (Figure 4.13c) may well be the most
severe loading condition for rigid plugs. It will result in direct tension across
some parts of the plug-rock interface, while compressive stresses are in-
duced in a perpendicular direction. This combination would seem to be an
open invitation to inducing tensile fractures across rigid plugs. Experimen-
tal simulation of this configuration may be most feasible by cyclic line
loading of plugged cylinders.
Impact loading of nearby blast

At the very extreme case, dynamic loading can be considered as impact or
transient loading, such as when blasting is carried out adjacent to a
plug-rock system. The induced stresses in the plug-rock interface could be
simulated by a hammer blow to the sides of a rock specimen containing a
borehole plug. Frequencies are likely to be extremely high. It would be
desirable to perform these experiments with controlled hammer (projectile)
impacts, such that short wavelength pulses of known magnitude are gener-
ated. These are designed to maximize differential deformation between
borehole and the plug. The impacts should be generated both longitudinally
and transversely. The result should give some indication of the upper bound
values of stresses in the plug-rock system due to dynamic loads, and how
they might affect the permeability.
Scaling factor
Dynamic loading tests performed in the experiment described in this chap-
ter were much more severe in terms of acceleration, velocity amplitude and
duration than what might be realistically experienced. However, plug and
opening size will affect performance during an earthquake. In centrifuge
tests using models with linear scaling of lin, an acceleration scaling factor
of n times the field acceleration is commonly used (Schofield, 1981; Craig,
1982). Assuming that this scaling hold, an acceleration of 2 g in a model seal
2.5cm (1 in) in diameter (as used in these tests) results in stresses identical
with those produced by an acceleration of 0.02 g in a prototype seal 100
times larger, i.e. about 2.5 m (8.3 ft) in diameter. As a comparison, a peak
ground acceleration of 0.64 g was recorded near the center of the aftershock
zone (within 1 km of the San Andreas fault) during the October 17, 1989,
Loma Prieta earthquake in the San Francisco Bay area (Dames and Moore,
1989). Considering that typical shaft diameters range from 4 to 12 m (10 to
over 30ft), it would be clearly desirable to perform shaking tests on larger
diameter cylinders and seals, and at higher accelerations.
Another factor to be considered in dynamic loading simulations is the
duration of the applied load. For models having a linear scaling of lin, the
time scaling for dynamic displacement which eventually results in increased
flow is also lin. On the other hand, for diffusion processes or fluid flow the
scale factor is 1/n 2 Hence, the maximum duration for which shaking has
been applied here exceeds the likely, even scaled, duration of an actual
earthquake. For example, the duration of strong ground shaking recorded
during the 1989 Loma Prieta earthquake was only lOs, admittedly a much
shorter time period that one would predict for a magnitude 7.1 event
(Dames and Moore, 1989). Difficulties may arise from the conflicts in select-
ing the various scaling factors and a more detailed study along these lines is
required. Coates (1981) mentions the problems in trying to fulfill all the
similitude requirements (i.e. geometric, kinematic and dynamic similitudes)
between the model and the prototype in some cases. Smith (1977) discusses
the problems associated with centrifugal modeling in geotechnical engineer-
ing, i.e. time scaling and viscous effects in dynamic problems and stress path
considerations.
Drying of cement seals

Drying of cement seals has proven to be a critical factor controlling the
sealing performance, if cement is to be used as a sealing material. The
cement and interface behavior under different moisture conditions are still
not fully understood. A test could be devised to measure the expansive
stress and stress relief of cement seals at different stages, from pouring,
during curing and hardening, at saturated condition, during drying and
finally at resaturation. It would be desirable to perform such tests at tem-
perature ranges likely to be encountered in the host environment, as well as
over ranges of degrees of saturation of the host rock. The latter should
include testing of seals emplaced in rock with permeabilities and porosities
representative for emplacement in unsaturated formations. By controlling
the environment in which the blocks are emplaced during curing a reason-
able simulation of curing and aging in unsaturated zone should be possible.
Experiments simulating severe yet realistic environmental situations that
induce performance degradation of the seals due to drying deserve high
priority. They should include simulation of cement curing in unsaturated
host rock, as well as the effects of drying, airflow and possibly rewetting on
older, hardened cementitious seals. A worst-case scenario of truly dried
cement plugs subjected to shaking can be simulated, using airflow instead of
waterflow. These plugs should be subjected to longitudinal as well as trans-
verse shaking.
4.7 DISCUSSION
Figure 4.13 illustrates the influences most likely to be experienced by

borehole seals. These can be used as a guideline for the design of an experi-
mental program to identify uncertainties and potential problem areas asso-
ciated with seal performance under dynamic loading. Analyses of these
Discussion 61
deformation patterns, when considering plug inclusions, could benefit great-

ly from procedures developed for buried pipelines (O'Rourke and Wang,
1978; Ariman and Muleski, 1979) and from wave impact analyses (e.g. Pao
and Mow, 1973; Miklowitz, 1978).
A number of questions remain on the performance of borehole and un-
derground excavation seals with respect to earthquake shaking. Dynamic
loading test on laboratory-scale specimens produced minimal impacts on
sealing performance of cement borehole plugs. Care must be taken, how-
ever, in extrapolating the laboratory results to actual in situ conditions.
Scale effect of size may influence the dynamic (acceleration, duration of
shaking) and flow parameters differently. An evaluation of the potential
impacts also needs to consider probable material response. For this purpose
it is appropriate to consider the two most likely sealing options to be
pursued, i.e. relatively rigid cementitious seals (concrete shafts and drift
plugs, borehole and fracture cement grout) and relatively soft earthen
(bentonite, bentonite/crushed rock) seals.
Most of these seismically induced deformations are likely to remain with-
in the elastic strain range of the rock unless the initiating event is exceeding-
ly large and sufficiently close to the openings (Monsees and Merritt, 1984).
However, in jointed rock at considerable depth Wahi et al. (1980), Ross-
Brown, Trent and Wahi (1981), Wahi and Trent (1982) and Marine, Pratt
and Wahi (1982) illustrate numerically some cases where the combined
effects of high in situ stress, nuclear waste-induced thermal stress and earth-
quake-induced dynamic stress can cause substantial discontinuous failure.
Rigid plugs should significantly reduce differential stresses around sealed
openings and hence should reduce the risk of failures. Conversely, under
extreme loading conditions they might become subject to excessive stress
themselves. Relatively soft backfill (e.g. crushed rock/betonite mixtures) may
be less effective in reducing small rock displacements, but may accomodate
larger deformations without excessive detrimental effects on its sealing capacity.
Considerable drying may take place for plugs located in hot, dry or at
least unsaturated host rock. This will be worse for a seal location in an
initially unsaturated environment, where subsequent water infiltration is
very slow. Further deterioriation in seal performance will result if consider-
able airflow through the plugs takes place. These conditions are likely to be
encountered in the application of borehole and shaft sealing for high-level
nuclear waste respositories, for example at the first US repository site in
Yucca Mountain, Nevada. A dry condition will pose challenging require-
ments, especially for cementitious seals but also for earthen (e.g. bentonitic)
seals. Particular attention should be given to shrinkage control, i.e. by
mixing the cement with sand, aggregates or other materials, or by using
techniques such as carbonization. The influence of seal emplacement in an
unsaturated environment, thermally driven seal de saturation and drying of
the seal emplacement areas need to be integrated in a realistic manner into
testing, analyses and design of borehole and excavation seals.
4.8 CONCLUSIONS
The scope of this chapter is to provide an experimental performance as-

sessment of cement borehole plugs subjected to laboratory dynamic load-
ing. This includes the study of dried plugs as well as of plugs that have
remained wet throughout the testing period. Review of the literature indi-
cates that deep underground structures in competent rocks are safer than
surface structures, openings at shallow depths and openings in fractured
rocks, when subjected to earthquakes and large-scale subsurface blasts. It
can be expected, therefore, that shaft or borehole seals installed at depth
and in intimate contact with the surrounding rock walls should show
minor (if any) effects from shock waves impacting on the sealed openings.
This is especially true if the seals can be matched mechanically to the
surrounding rock mass in order to minimize any impedance differences,
and is particularly likely for boreholes which, in all probability, will be
filled over their entire length. The situation is less obvious for shaft plugs,
which may be relatively short, i.e~ with lengths in the same order of magni-
tude as the shaft diameter, and certainly much shorter than the likely
wavelengths.
Steady-state water flow tests have been conducted on cement borehole
plugs installed in Charcoal granite cylinders, as well as on the granite rock
bridge left in place when the axial holes are drilled from both ends of the
cylinder. The tests are performed by injecting distilled water into the seal
and measuring the inflows as well as the outflows to obtain the flow rate
through the seal. Dye markers have been injected during the last stage of
the flow tests to identify the flow patterns in wet and in dried cement
plugs. All experiments have been conducted on rock cylinders that are not
loaded, except for the injection pressure and (unknown) cement swelling
pressure, both of which induce a tensile tangential stress in the rock cylin-
der. This should impose a rather severe condition with regard to interface
flow. Only at plug locations with exceedingly anisotropic in situ stress-
fields should a more disadvantageous sealing condition, with respect to
the stress state, be encountered. All flow analyses assume that the plug,
rock and interface gap, if any, are saturated, and that steady-state condi-
tion prevails. The latter is verified by a finite element analysis for one-
dimensional flow.
The flow test results indicate that wet cement grout seals, i.e. cement
borehole plugs that are never allowed to dry, are equal or less permeable
than intact Charcoal granite. Sealing performance can degrade severely
when cement grout seals are allowed to dry. Plugs suffer from decoupling
along the cement-rock interface, and the permeability increases by several
orders of magnitude. Upon resaturation the fissure permeability decreases
rapidly, indicating partial closure of the interfacial gap due to cement swell-
ing. This usually continues during the first 2 months of resaturation and is
Conclusions 63
followed by a much slower decrease in fissure permeability. However, this

permeability still remains several orders of magnitude higher than the
permeability of wet cement plugs that are never allowed to dry. The seal
performance is not fully recovered.
The extent of seal performance degradation seems related to the length of
drying and to the drying temperature. Longer drying time and/or higher
drying temperature seem to correlate with more severe performance degra-
dation. This is due in part to cracking in the cement body itself, which acts
as an additional preferential flow path. Visual inspection after sawing the
specimens in half, upon completion of dye injection tests, shows that the
flow penetrates the wet plug uniformly, but occurs nearly exclusively along
the plug-rock interface and along the cracks in the dried plugs, which act
as preferential flow paths.
Dynamic loading was applied using a shaking table during the later stage
of ongoing flow tests. The (un scaled) applied dynamic loads are consider-
ably more severe than what would be encountered during any likely earth-
quake or subsurface blast loading. Nevertheless, the effect on seal
performance is minimal. The peripheral flow rate through the rock immedi-
ately surrounding the seal remains within the original range after dynamic
loading tests are concluded. This indicates the absence of new fracture
development in the rock as a result of the applied dynamic loads. Tests
conducted on wet cement plugs show that flow rate through the seal (or seal
permeability) does not change significantly after dynamic loads are applied.
The change in permeability as a result of dynamic loading is less than
permeability variation without dynamic loading. Even the fissure permea-
bility of dried cement plugs is not affected by the dynamic loads. It needs to
be pointed out that the dynamic shaking is applied only after the dried
plugs have been flow tested with water for several months, during which
time considerable resaturation and re-swelling of the cement plugs take
place. Shaking is not performed on truly dry plugs, in which a pronounced
shrinkage gap exists between cement plug and borehole rock wall. More-
over, the shaking cycle is applied externally to the plugged rock cylinder,
which are accelerated as a unit. These experiments do not induce the poten-
tially most damaging deformations, such as relative longitudinal strain be-
tween plug and rock wall, borehole bending or hole deformations.
The effect of size, which influences the acceleration and duration of the
dynamic loads, necessitate a special consideration in extrapolating the result
of laboratory-scale samples to the actual design conditions. The scale effect
also influences dynamic displacement and diffusion process that control
fluid flow. Sealing in an unsaturated environment may have a negative
impact on the drying (curing and aging) conditions of cementitious seals, as
well as on the structure of earthen seals. An unsaturated environment will
need to be integrated realistically into sealing performance test, analysis and
design.
ACKNOWLEDGEMENTS
This work was part of research effort sponsored by the US Nuclear Regula-
tory Commission, under contract NRC-04-78-271. Support and permission
to publish this chapter are gratefully acknowledged.
CHAPTER FIVE
Performance of Bentonite
and Bentonite/Crushed
Rock Borehole Seals
S. Ouyang and J. J. K. Daemen
5.1 INTRODUCTION
Bentonite is an excellent sealant material because of its low permeability,

desirable swelling and self-healing characteristics, sorptive qualities and lon-
gevity in nature (e.g. Meyer and Howard, 1983; Pusch, 1983). Bentonite is
used extensively for sealing boreholes. Under certain circumstances advan-
tages can be gained by sanding bentonite or by mixing bentonite with
crushed rock.
According to Dixon, Gray and Thomas (1985) the addition of up to 50%
of sand increases the achievable compacted density, does not change (or
decrease) the swelling pressure developed by the clay (Taylor et al., 1980, p.
151, show a nearly steady decrease of swelling pressure over a sand content
range from 25 to 90%), decreases the shrinkage potential, increases thermal
conductivity and increases the bearing capacity of the backfill, minimizing
creep or settlement. The addition of crushed rock to bentonite may deliver
similar advantages.
Borehole seal components may be required to retain adequate sealing
performance over a long period of time (Fernandez et aI., 1987). The effect
of piping and erosion, as learned from the failures of earth dams, embank-
ments and natural slopes (e.g. Sherard, Decker and Ryker, 1972; Rosewell,
1977; Goodman and Sundaram, 1980) is thus important, especially if the
seals are installed in locations intercepted by joints and/or fractures.
This chapter includes a laboratory investigation of the sealing per-
formance of bentonite and of bentonite/crushed tuff plugs under diverse

Edited by K. Fuenkajorn and 1.1. K. Daemen.
66 Bentonite and bentonite/crushed rock borehole seals
conditions. The sealing performance assessments include high llljection

pressure permeability tests, polyaxial permeability tests, high-temperature
permeability tests and piping tests. Theoretical analyses are performed to
develop permeability and piping models for bentonite.
5.2 MATERIALS
5.2.1 Bentonite
The bentonite used in this study is American Colloid cis granular Volclay
from Upton, Wyoming. The bentonite is a highly colloidal, expansive clay
that is an alteration product of volcanic ash (Mitchell, 1976, p. 39). The
volcanic ash was first deposited in a saltern sea near the Black Hills. Due to
later uplift the ash particles were subjected to weathering and were altered
from their fragile, glassy state to the swelling, paste-forming Na-rich ben-
tonite (Jepsen, 1984).
The cis granular bentonite consists primarily of the clay mineral mont-
morillonite, with traces of quartz, feldspar and biotite. This highly plastic
clay has the ability to adsorb nearly five times its weight in water and can
swell to 10 to 13 times its dry size upon complete hydration (Jepsen and
Place, 1985). The bentonite has an average specific gravity of 2.92, liquid
limit of 433%,plastic limit of 50% and plasticity index of 383%. The average
initial (air-dried) moisture content is 9.56%, over a range of 9.41 to 9.69%.
The moisture-density relation gives the optimal moisture content as 23.5%
and a maximum dry density of 12.21 kN/m 3 , determined following the
compaction method A described in ASTM Standard D698-78.
Approximately 90% of the cis granular bentonite is montmorillonite
(American Colloid Company, Data No. 202). Structurally, montmorillonite
is classified in the expansive 2: 1 clay mineral group. The 2: 1 designation
indicates a clay composed of an octahedral sheet sandwiched between two
silicate tetrahedral sheets. The theoretical composition of 2: 1 clays is
(OH)4SisAI402on(interlayer) H 20 (Mitchell, 1976, p. 37), which is almost
never found in natural occurrence due to isomorphous substitution in the
crystal lattice. An approximate chemical formula for the Wyoming benton-
ite (American Colloid Company, Data No. 202) is (AI, Fe1.67 Mg o.33 )
Si4 0lo(OH)2Na, Ca O. 33 ' indicating a significant isomorphous substitution
of AI3+ by Fe 2+ and a predominance of Na+ions adsorbed.
The negatively charged montmorillonite clay particle creates an electric
field. Exchangeable cations are adsorbed on the clay surface to balance this
field. In the dry state the cations are held tight on the clay surface. Any
cations in excess of those necessary for charge balancing form a soluble
salt precipitate (with its associated anions) in and around the clay particles
(Mitchell, 1993, p. 11). In the presence of free pore water the salts go
into solution. This process leads to the formation of a higher molar
Materials 67
concentration near the clay surfaces, and the cations tend to diffuse into the
surrounding lower-concentration pore water to obtain a homogeneous ion
concentration throughout the clay-water system. The cation diffusion is
opposed by the negative electric field of the clay particle. The negatively
charged clay sheet and the positively charged cation distribution in the pore
water are called the 'diffuse double layer' or simply the 'double layer'
(Mitchell, 1976, p. 113; Wu, 1976, p. 398).
The adsorbed ions are exchangeable with other ions in an aqueous envi-
ronment. The capacity of such ion exchanges is termed the cation exchange
capacity, which is measured in milliequivalents per 100 grams of dry clay.
The equivalent weight of an element is its atomic weight (in grams) divided
by its valence. The smectite clay group, which includes montmorillonite,
typically has a cation exchange capacity of 80-150 milliequivalents/100 g
(Grim, 1953, p. 129).
The development of the diffuse double layer and cation exchange can
account for many observed changes in the engineering behavior of clays.
Grim (1953, p. 127) points out the great sensitivity of the physical properties
of clays to the type of exchangeable ions carried. Singh (1982) indicates that
the liquid, plastic and shrinkage limits, as well as the permeabilities of
various montmorillonites, are a function of the adsorbed cation, either Na +
or Ca 2 +. Results from Endell et al. (1938), reported by Grim and Guven
(1978, pp. 242-3), indicate that mixtures of Ca montmorillonites with sand
have permeabilities two orders of magnitude higher than similar mixtures of
Na montmorillonites with sand. Similar permeability ratios for the two
bentonite types are also reported by Mesri and Olson (1971). The difference
may be explained by the double layer thickness and cation exchange.
5.2.2 Apache Leap tuff

The rock used to produce crushed tuff is from the densely welded brown
unit of Apache Leap tuff. This tuff shows low porosity and permeability.
Physical and mechanical properties of Apache Leap tuff are given by
Fuenkajorn and Daemen (1991a). (The rock samples were collected along
old highway 60, approximately 1 mile (L61 km) east of Superior, Arizona.)
Tuff is a volcanic ash which has been compressed under its own weight and
sometimes is welded due to its high temperature during deposition. Con-
stant head radial flow tests performed on two hollow tuff cylinders give
permeabilities of 6.3 and 2.3 x 10 - 10 cm/s.
5.2.3 Bentonite/crushed tuff

A mixture of bentonite and ballast material (e.g. quartz sand and crushed
rock) is being considered as backfill and sealant for nuclear waste reposito-
ries (Pusch, Jacobsson and Bergstrom, 1980; Nilsson, 1985; Holopainen,
1985; Yong, Boonsinuk and Wong, 1986; Williams and Daemen, 1987).
Immediate gains from adding crushed rock to bentonite are to alter the
natural geochemical properties as little as possible (Holopainen, 1985) and
to reduce the amount of waste rock to be disposed of (Smith et ai., 1980).
Bentonite content and the gradation of the ballast material are two deci-
sive factors in the design of mixture plugs. If the grading is not proper or
the mixture not thoroughly homogenized, or if the amount of bentonite is
not sufficient to fill the ballast pores, bentonite gel can be displaced and
torn-off fragments can be transported through channels that form at a
relatively low water overpressure (Pusch, 1987, Borgensson and Ramquist;
Pusch, Erlstrom and Borgensson, 1987).
Bentonite content
Pusch, Jacobsson and Bergstrom (1980) report that the permeability of
water-saturated bentonite/quartz mixtures with a weight ratio varying from
1: 10 to 1: 5 (i.e. 9 and 16.7 wt% sodium bentonite, respectively) ranges
between 10- 7 and lO- 11 cm/s. Holopainen (1985) indicates permeabilities of
5 x 10- 7 to 10- 8 cm/s for bentonite/crushed rock mixtures with 15% so-
dium bentonite. Bentonite/sand percentages of 10/90 and 20/80% have been
used by Nilsson (1985). Bentonite contents of 5, 15,25 and 35 wt% have been
tested by Williams and Daemen (1987) in mixes with crushed basalt. Their
flow test results indicate massive failure for the mixtures with 5% bentonite.
After an extensive study of clay/crushed granite mixtures, Yong, Boonsinuk
and Wong (1986) conclude that a candidate backfill should have a clay
content between 20 and 30%.
Three bentonite weight percentages (15,25 and 35%) have been chosen
for the bentonite/crushed tuff plugs tested in this study.
Crushed tuff gradation

Five different crushed tuff gradations have been used. Three gradations
(Types A, Band C) are derived from the 'coarse' type by first eliminating all
particles finer than 0.74 mm (US mesh #200) and then successively remov-
ing crushed tuff particles larger than 9.42, 6.68 and 4.75mm. The 'coarse'
gradation is obtained from crushing tuff chunks of approximately
15 x 15 x 20cm (6 x 6 x 8 in), using a jaw crusher and an adjustable roller
crusher in sequence. The other two gradation types (FA and FC) are ob-
tained using n = 0.5 and Dmax = 9.42 and 19.05 mm, respectively, in the Ful-
ler-Thompson grading equation (Equation (5.1)):
P 100(_d_)n
w =
Dmax
(5.1)
where Pw = weight percent passing sieve aperture d, Dmax = maximum par-

ticle size, and n = exponent. The Fuller-Thompson grading curve is con-
sidered to be an ideal grading which may result in the densest possible state
Materials 69
of packing (Winterkorn, 1975; Head, 1980, p. 150). The US Bureau of Public
Roads (US Bureau of Public Roads, 1962; as cited by Winterkorn, 1975)
recommends 0.45 for n as the best overall value; while Head (1980, p. 150)
suggests an n value of 0.5. The Fuller-Thompson equation has been used in
formulating backfill material for a nuclear waste disposal vault (Pusch and
Alstermark, 1985, in which the exponent of n is not reported; Yong,
Boonsinuk and Wong, 1986, n = 0.25).
Figure 5.1 shows the grain-size distributions of the five gradation types
together with the 'coarse' one. The figure also shows the coefficients of
uniformity (d 60 /d 10 )' The uniformity coefficients for Types FA and FC are
theoretical values computed from Equation (5.1). In actuality, the grain-size
analysis of crushed tuff aggregates smaller than 0.074mm is not performed.
To obtain crushed tuff of the FA and FC gradations requires additional
efforts in material handling (e.g. sieving and blending). Crushed tuff of A, B
and C gradations is obtained easily from the 'coarse' aggregates, as dis-
cussed earlier.
5.2.4 Permeant
Permeability testing on bentonite and bentonite/crushed tuff plugs with the
local (site-specific) groundwater is desirable to incorporate possible physi-
cal-chemical interactions (Gaudette and Daemen, 1988; Neuzil, 1986). When
U.S. MESH: No.4 No.10 No.20 No.40 No.60 No.100 No.200

100 0
Gradation type Cu (derld10)
0 A 16.5 0
(!)
80 B 15 20 zw
z ~
C 13
iii
en Coarse
FA ~n = 0.5l
11 FC n=0.5
36a
36a ~
~
~ 60
14
w
a:
40 ~
w a: Theoretical value
w
0 0
a:
w a:
a... w
a...
40 60
!i:(!) !i:
(!)
jjj jjj
== 20 80 ==
0 100
10 1 0.1
GRAIN SIZE (mm)
Fig. 5.1 Grain-size distribution of crushed tuff used in sealing experiments.

the site-representative groundwater is not available, testing can use deaired

distilled water as the primary permeant, thus providing a standard reference
condition. Several waters with different chemistries have been tested.
5.3 PERMEABILITY TESTS AND RESULTS
5.3.1 Sample preparation and installation

Bentonite plugs have been constructed using the sedimentation method by
gradually dropping 16.45 g of air-dried bentonite powder into a water col-
umn. The samples have been allowed to hydrate for between 3 weeks and a
month prior to testing. Saturation of the samples has been aided by inter-
mittent vacuuming from the top. The applied vacuum was less than 34.5
kPa to avoid upward movement of the samples.
Bentonite of the optimal water content (23.5%) has been used to prepare
the compacted samples. To achieve a uniform water content distribution, a
sprayer is used to distribute evenly the (distilled) water across the clay while
mixing. The bentonite is cured for 72 h before installation. For sample
diameters greater than 50.8 mm (2 in), a manual rammer as specified by
ASTM 0698-78 (Standard Proctor Method), 03.2.1, is used to compact the
samples. A hammer compactor with hammer weight of 0.53 kg (1.l6Ib) and
circular specimen contact of 25.4 mm (lin) in diameter is used for compact-
ing samples of smaller diameters. The drop height and the number of drops
per layer are adjusted to provide the same energy input as for the Standard
Proctor Test (593 kJ/m3; 12275 ft-Ib/ft 3).
The bentonite samples compacted in PVC permeameters have no im-
mediate confinement on top, and swell vertically during hydration. They are
saturated under a water pressure of 24.5 kPa (2.5 m water column) from the
bottom port for about 75 days. A vacuum pressure of 34.5 kPa (5 psi) is
applied continuously to the top port to help the saturation process. For the
compacted samples totally confined in stainless steel permeameters, satura-
ting has been attmepted by an injected water pressure of 345 kPa (50 psi)
and an intermittent vacuum of 103.5 kPa (15 psi) for approximately 35 days.
Such samples are later subjected to a low water pressure of 69 kPa (10 psi)
for about 45 days and then to a pressure of 24.5 kPa for another 2 weeks,
prior to flow testing.
The mixture plugs are prepared by thoroughly mixing crushed tuff (con-
stituting a selected gradation) with a predetermined amount of bentonite
containing 23.5% water. A mixed sample is stored and cured in an air-tight
plastic jar for 72 h. The sample is then transferred into a permeameter using
a small plastic shovel. This emplacement method is selected to minimize
particle segregation. All the mixture samples receive a compaction energy
equivalent to that of the Proctor compaction method, except for the three
samples mixed with crushed tuff of the FC gradation. The latter three
Permeability tests and results 71
samples received only 25% of the Proctor compaction energy. The pro-
cedure for sample saturation follows the method recommended by ASTM
(D2434-68, 6.6.4). To minimize changes in sample structure, the samples are
saturated under a 2.5 m water column (24.5 kPa) from the bottom for
approximately 2 months, frequently aided with a vacuum pressure of
103.5 kPa (15 psi) from the top.
5.3.2 Permeability test methods

Constant head, standard falling head (Lambe, 1951), and modified falling
head (double-pipet falling head (Williams and Daemen, 1987) methods are
used to determine permeabilities of the plugs. They are steady and quasi-
steady (i.e. the falling head) flow test methods.
Precise measurement of permeability in highly deformable media (e.g. the
bentonite studied here) is difficult with any technique (Smiles and
Rosenthal, 1968; Kharaka and Smalley, 1976; Olson and Daniel, 1981). In
this investigation, low hydraulic gradients are used for the determination of
permeability to minimize pressure-induced deformation of the samples dur-
ing flow testing. The flow tests with large hydraulic gradients are aimed at
studying the sealing performance under high injection pressures. Consider-
able efforts are devoted to balance inflow and outflow.
5.3.3 Permeability test results of bentonite seals

Fourteen bentonite samples have been constructed: (1) four sedimented
2.5 cm (1 in) diameter plugs for flow testing with different water chemistries,
(2) six compacted samples of 2.5,5 and 10 cm (1,2 and 4 in) diameter (two
for each size) for studying the size effect on permeability, (3) four compacted
plugs having diameters of 2.654, 3.475, 6.01 and 10.246 cm, respectively, for
the high injection pressure flow testing.
I nfiuence of water chemistry

Samples B-S-1-A, B-S-1-B, B-S-1-C, and B-S-1-D were constructed for flow
testing with different permeants. [Sample designation is as follows. First
letter: material type (e.g. B for bentonite); second letter: construction
method (S for sedimentation, C for compaction); digit in third place: nom-
inal diameter in inches, last letter: individual sample identification.] Sample
descriptions and permeants are shown in Table 5.1; test results are given in
Figure 5.2.
Sample B-S-1-A, deposited in and tested with distilled water, showed a
permeability as low as 6.9 x 1O- 8 cm/s at the end of a testing period of
21 days. To replace distilled water with 2% sodium pyrophosphate solution
(a dispersing agent), the sample was flushed with the solution for about 37
days before the flow testing resumed. Two months later the permeability
Table 5.1 Sample descriptions of sedimented bentonite plugs

Initial sample Type of Length Permeant 1 Permeant 2
number fluidfor (em)
sedimentation"
B-S-I-A Distilled water 15 Distilled water 2% Sodium

(pH = 6) pyrophosphate
B-S-1-B 2% Sodium 9 2% Sodium Distilled water
pyrophosphate pyrophosphate
(pH=9)
B-S-1-C 2% Calcium 14.7 Distilled water 4% Sodium
hydroxide pyrophosphate
(pH = 12)
B-S-1-D Synthetic waterb 19.7 Synthetic water 2% Sodium
(pH = 6) pyrophosphate
'Percentage: chemical added as weight percentage of distilled water in sample; bproduced by
boiling distilled water in the presence of A-Mountain crushed tuff.
--~B-S-1-A
. .....
B-S-1C(SP)
........
B-S-1-~"'_>=-_'
-
_ _o::It::iir-....
4
10~ O~~~~~~~1~O~~~~~~2~O~~~~~~~
TIME (days)
Fig.5.2 Permeability of 2.5 cm (1 in) diameter bentonite plugs with first permeants.
Sample and permeant details are listed in Table 5.1. Graph B-S-1-C(SP) shows the
permeability of sample B-S-1-C when permeated with 4% sodium pyrophosphate
solution.
dropped to 2 to 3 x 1O-8 cm/s. Sample B-S-1-B, deposited in and permeated

with a 2% sodium pyrophosphate solution, showed a permeability of
1.4 x 1O-8 cm/s after about 25 days. The sample was flushed with distilled
water for 30 days and subsequently tested with the same water for 83 days.
A permeability of 1.6 x 1O-8 cm/s was measured at the end of the test.
Relatively high permeability values, about 1O-5 cm/s were measured for
sample B-S-1-C during a test period of 8 days. This sample was deposited in
a 2% calcium hydroxide suspension (a flocculating agent) and later tested

with distilled water. Subsequently, the sample was flushed with 4% sodium
pyrophosphate solution for 48 days. The permeability testing was then
resumed and continued for 72 days. A reduced permeability of
3.4 x 10 - 7 cmls was measured at the conclusion of the test (Figure 5.2)
Sample B-S-1-D, sedimented in synthetic water (distilled water boiled in the
presence of crushed tuft), had a permeability of 7.7 x 10- Bemis when tested
with the synthetic water and 2 x 10- Bemis with the 2% sodium pyrophos-
phate solution.
Influence of sample size

Two compacted bentonite plugs of 2.5, 5 and 10 em (1, 2 and 4 in) diameter
each have been constructed in PVC permeameters. During saturation the
samples are allowed to adsorb water and to swell vertically. This hydration
swelling has continued throughout the subsequent flow tesing period. Chan-
ges in sample length as a result of the hydration-swelling process are sum-
marized in Table 5.2. The sample length at the end of permeability testing is
the maximum possible expansion, given the restraint imposed by the end
caps of the permeameter. Considering that the bentonite can swell to 10-13
times its dry size upon complete hydration (Jepsen and Place, 1985), the
final sample lengths in Table 5.2 suggest that the samples are still capable of
swelling further. Due to the cap restraint, a certain amount of swelling
pressure must have been generated. This is confirmed by the fact that the
bottom cap plate of the permeameter housing sample B-S-1-B has been
detached from the sample chamber (originally glued on with epoxy).
Results of first permeability test sequence

Permeabilities of the six compacted bentonite plugs have been measured
continuously for approximately 85 days, using the double-pipet falling head
method. For sample B-S-4-B the testing method was changed to the
Table 5.2 Changes in sample length of compacted bentonite plugs

Sample number Sample length (cm)
After compaction Before flow testing After flow testing
B-C-I-A 3.40 11.5 12.8

B-C-I-B 3.33 11.35 12.8
B-C-2-A 3.18 9 9.1
B-C-2-B 3.76 9.2 9.4
B-C-4-A 3.35 10.1 10.2
B-C-4-B 3.35 9.7 10.0
standard falling head method after the detachment of the bottom cap was
noticed. The inflow and outflow are out of balance throughout the flow
testing. This imbalance most likely is due to the continuing hydration and
expansion of bentonite.
Permeabilities of all six compacted bentonite plugs are of the order of
10- 9 cm/s. Recognizing that the samples are not yet completely hydrated,
the permeabilities appear reasonable when compared with the
6.9 x 10- 8 cm/s (for sample B-S-l-A, Figure 5.2) for which complete hy-
dration can be assumed.
Permeability varies only slightly, and not systematically, as a function of
diameter. Differences typically amount to two to three times the permeabil-
ity obtained. The small variations suggest that similar test results can be
obtained if the same procedures for sample preparation, installation and
flow testing are repeated. Permeabilities of the bentonite plugs appear in-
variant with sample diameters used in this study.
Results of second permeability test sequence

In an attempt to improve the mass balance and to examine changes of
permeability with continued hydration, five of the six compacted bentonite
plugs have been subjected to flow testing for an additional 43 days. The
permeability results show a relatively small variation over the additional
test period. The inflow and outflow remain out of balance, which may,
again, be due to continuing hydration and expansion of the bentonite.
During the second flow test sequence, samples B-C-l-A and B-C-l-B
yield an average permeability of 2 x 1O- 8 cm/s, about two times higher
than the value obtained from the later measurements of the first test se-
quence. Samples B-C-2-A, B-C-2-B and B-C-4-A show essentially the same
permeabilities as measured before, from 2 to 6 x 1O- 9 cm/s. No evidence
can be discerned of any size effect on permeability.
Results of high injection pressure flow tests

This experimental series includes flow tests of samples B-C-l-A-S, B-C-l
3/8-A-S, B-C-2 3/8-A-S and B-C-4-A-S. Samples dimensions, initial water
content, bulk density and porosity are given in Table 5.3.
Before flow testing, the samples are subjected to an injection water press-
ure of 345 kPa (50 psi) for about 2 months, and intermittently to vacuuming
at the top. The double-pipet falling head method is then used to determine
permeabilities. With approximately 1.2 m of water head difference across the
samples, flow testing has continued for more than a month. No positive
outflows have been observed. The samples have again been SUbjected to
vacuum and the test set-up has been replaced with the constant-head
method driven by compressed helium. The outflow is monitored by the
movement of an air bubble in a horizontal pipet. The inflow is calculated
Table 5.3 Sample characteristics of compacted plugs installed in stainless steel per-
meameters
Sample Sample Sample Initial Initial Saturated Saturated Porosity
number length diameter water bulk water bulk
(em) (em) content density content density
(%) (g/cm 3 ) (%) (g/cm 3 )
B-C-I-A-S 8.89 2.654 23.55 1.431 57.13 1.720 0.625

B-C-l 3/8-A-S 9.623 3.457 32 1.482 88.32 1.536 0.721
B-C-2 3/8-A-S 14.95 6.01 23.55 1.384 60.26 1.603 0.638
B-C-4-S 13.125 10.246 23.55 1.358 62.05 1.683 0.644
from the drop of the water column in a PVC water reservoir. No outflow
has been detected for sample B-C-l 3/8-A-S.
For this test series the hydraulic gradients range from 47 to 720 (injection
pressures from 68.9 to 620.5 kPa, i.e. 10 to 90 psi). The permeability values
primarily range between 2 x 10 - 9 and 5 x 10 -11 cm/ s (e.g. Figure 5.3) and
appear to decrease with increasing hydraulic gradient.
5.3.4 Flow tests on bentonite/crushed tuff seals

Mixtures of 15, 25 and 35 wt% bentonite and crushed tuff gradation types
A, Band C form the majority of the samples. In addition to the crushed
rock gradation and bentonite weight percent, other variables studied
include sample size, flow direction, hydraulic gradient and temperature.
1()-8..--_ _ _ _ _ _ _ _---.
I

I I
~1()-8
E
.!:!- I
~
II
::::;
ffi
;J
II

w
~ I I
II:
~ 1010
I
I I

10-11 ':-_--'--_ _' - _ - ' - _ - - - '

0.0 150.0 300.0 450.0 600.0
HYDRAULIC GRADIENT
Fig.5.3 Permeability vs. hydraulic gradient for sample B-C-2 3/8-A-S.

Testing conducted includes longitudinal flow tests, polyaxial flow tests, high
temperature flow tests and piping tests.
In this test series PVC permeameters have nominal inside diameters of
10.16 and 30.5 cm and stainless steel permeameters of 10.16 and 20.3 cm. A
rectangular plexiglass permeameter constructed for polyaxial flow testing is
12.5 x 11.4 x 11.4 cm in size. Thin porous plates can be installed at inflow
and outflow ends. Circular openings in the walls of some PVC per-
meameters simulate openings in the walls of holes where seals are emplaced.
The openings are plugged during sample installation and saturation and are
unplugged for flow testing. No porous stone or filter materials (e.g. sand)
are emplaced at these side openings because they might impede particle
movement. For the majority of the plugs tested the inflow and outflow
differs only by 5 to 1O%.
Longitudinal flow tests

Average permeabilities obtained from the downward flow testing under low
hydraulic gradients ( < 35) are summarized in Figure 5.4. The permeability
decreases with increasing bentonite content. The permeabilities of the plugs
containing 25% or more bentonite are close to those of plugs constructed of
bentonite only. Figure 5.4 also demonstrates the effect of crushed tuff grada-
tion on sealing performance: the greater the uniformity coefficient (d 6o / d1o )
of crushed tuff, the lower the permeability. The effect of sample size on the
sealing performance is not clear. The inconsistency in the permeabilities
measured for different plug sizes may be due to variations in the stiffness of
the permeameters, compaction and the ratio of grain size to permeameter
diameter. Upward flow testing was performed to study the action of an
upward seepage force. The upward permeability is about three times higher
than the downward permeability.
Permeability of the mixtures is relatively constant for the low range of
imposed hydraulic gradients. As the hydraulic gradient increases, permea-
bility decreases, causing a breakdown of the linear relationship between
flow rate and hydraulic gradient. This behavior has been observed for all
high injection pressure flow tests; an example is shown in Figure 5.5. The
sudden jumps in permeability may result from the radial expansion of pores
due to rapid large increases in injection pressure, 390 kPa for the first jump,
762 kPa for the second one. No permanent increase in permeability can be
detected, indicative of the excellent healing capability of bentonite. The
departure from the early linear flow rate~hydraulic gradient relationship is
abrupt. The sudden change may indicate the onset of bentonite flow in the
pores between crushed tuff particles. This assumption was confirmed when
samples were examined upon completion of testing. For the sample contain-
ing 25% bentonite, the clay has flowed downward, leaving the top quarter
of the sample consisting primarily of crushed tuff. Bentonite flow was less
developed in the sample containing 35% bentonite, and can only be discer-
... pvc, 10.16 cm, Type A

t. PVC, 10.16 cm, Type B
o PVC,10.16cm,TypeC
~ PVC, 10.16cm, Type FA
8.8.,10.16 em, Type A
8.8.,20.3 cm, Type A
x PVC, 30.5 cm, Type A
~
i
~
:::; 10.7
OJ

w
:::l;
cr
W
Q.
I
10.8
I
[J 100% bentonite,
void ratios: 6.1-8.8
100% bentonite,
void ratios: 1.67-1 .81
o 10 20 30 40 50 60
BENTONITE WEIGHT PERCENT
Fig.5.4 Permeability of bentonite/crushed tuff mixtures as a function of bentonite

content Hydraulic gradients are lower than 35.
ned by examing the subtle difference in texture between the top and bottom
of the sample. The hydraulic gradient at which bentonite flow takes place
appears to vary with bentonite content. The gradient ranges for mixture
samples of type A crushed tuff containing 25% and 35% bentonite were
120-150 and 280-300, respectively. The pressure required to initiate ben-
tonite flow is probably a function of the yield stress of the bentonite.
Polyaxial permeability tests

Adequate sealing ability in the transverse direction may be necessary to
minimize any flow of contaminated groundwater or gases laterally into the
connected fracture system. Moreover, compromising the sealing ability in
the transverse direction may jeopardize the entire sealing performance if
10.7 r - - - - - - - - - - - - - - - - - ,

~ 10.8
-!!!
l
I' I I i
~
::::;
iii
.--..
<
w
:;
a:
w
10.9
I I
Ilf
0.
10.10 '::-_ _'--_----''--_----'_ _--L_ _- . I

0.0 400.0 800.0 1200.0 1600.0 2000.0
HYDRAULIC GRADIENT
Fig.5.5 Flow rate vs. hydraulic gradient for sample B/AL-C-8-35/A-S.
piping occurs radially. This consideration can be significant if seals are

installed at locations intercepted by joints and/or fractures.
Polyaxial permeability testing has been performed on two rectangular
samples. One sample consists of 75% crushed tuff and 25% bentonite, the
other of 65% crushed tuff and 35% bentonite. The crushed tuff is of the type
A gradation. The test results are summarized in Table 5A
High permeability is observed in one horizontal direction, for both sam-
ples. The two high horizontal permeabilities are very similar. The high
horizontal permeability exceeds the vertical permeability by almost one
order of magnitude for the sample containing 25% bentonite and by more
than two orders of magnitude for the sample containing 35% bentonite.
High-temperature permeability tests

Two samples installed in stainless steel permeameters were immersed in a
constant-temperature water bath for flow testing at up to 60C. The sam-
ples were constructed by mixing crushed tuff of type A gradation with 25%
and 35% bentonite. Table 5.5 summarizes the results. The 'measured' per-
meability is computed directly using the experimental measurements. As-
suming the major temperature effects are the changes in viscosity and
density of the permeant (i.e. water), permeabilities of the plugs at other
temperatures can be calculated based on the average permeability measured
Table 5.4 Polyaxial permeability test results

Wt% Permeability( cm/s)
Bentonite/Tuff
Vertical Horizontal (l ) Horizontal (2)
25/75 5.2-1.7 X 10- 8 1.8-1.4 X 10- 7 1.8-1.5 X 10- 8

35/65 9.2-2.5 X 10- 9 11-1.9 X 10- 7 12-6.3 X 10- 9
Table 5.5 Results of permeability tests at elevated temperature

Temp. Total Test Total Permeability
CC) duration inflow outflow
(days) (cm 3 ) (cm 3 ) Measured Corrected a Specific
(cm/s) (cm/s) (m 2 )
Sample 1: 25% bentonite, 75% type A crushed tuff

(x 10- 8) (xlO- 8 ) (x 10- 17 )
21 28 29.25 28.6 1.280.40 1.28 1.28
35 23 23.25 22.6 2.320.15 1.72 1.72
45 21 21.5 21.8 2.350.12 1.62 1.45
60 6 8.05 8.6 2.390.31 2.05 1.16
35 17 17.07 15.65 2.260.10 1.72 1.68
Sample 2: 35% bentonite, 65%type A crushed tuff
(x 10- 9 ) (x 10- 9 ) (x 10- 18 )
21 28 8.1 7.25 3.06 1.16 3.06 3.05
35 23 7.25 6.15 5.02 1.38 4.13 3.72
45 21 6.45 6.1 5.26 0.90 3.88 3.24
60 4 0.8 1.2 4.26 1.34 4.92 1.58
35 17 3.65 3.67 4.27 1.22 4.13 3.17
Hydraulic gradients applied are less than 10; 'corrected: permeant (water) viscosity and
density calculated for true temperature.
at the room temperature (21q, provided that the permeant properties are
adjusted. The results are shown in the sixth column of Table 5.5 and the
specific permeabilities calculated are given in the last column of Table 5.5.
The specific permeability reaches a maximum at 35C, at 60C it is
reduced by 10% for the sample containing 25% bentonite, and by 50% for
the sample containing 35% bentonite, when compared to the specific per-
meability at 21C.
Piping tests
Piping tests have been performed on two samples installed in perforated
10.16 cm diameter PVC permeameters. The samples contain 25% bentonite
and 75% crushed tuff of type A gradation. A 2.25 mm diameter hole drilled
through the walls of the permeameters simulates an opening in the walls of
boreholes or shafts where seals have been emplaced. The bottom of the
samples is approximately 1 cm below the center of the holes.
The vertical permeabilities are in the range 1.2-1.8 x lO- s cm/s, under
hydraulic gradients less than 32. One sample was then subjected to an
injection pressure of 32.75 kPa with the bottom outlet closed and the side
hole opened. Approximately 2 days later bentonite and fine tuff particles
along with a small amount of water moved through the hole into the
connecting tubing. The bottom outlet was reopened a week later to allow
determination of the vertical permeability. The permeability had increased
by two orders of magnitude, into the lower 10- 6 cm/s range. In subsequent
flow testing, the bottom outlet was closed again. About 12 h after the
injection pressure had been raised to 113 kPa, a spill of water was observed
and the inflow reservoir was completely drained. The gross hydraulic gradi-
ent induced by the injection pressure was 116.
During flow testing of the second sample, the bottom outlet was left open
while the injection pressure was increased. Bentonite and fine tuff particles
appeared in the tubing connected to the side hole at an injection pressure of
27 kPa. No signs of piping were detected for injection pressures up to
367 kPa. The amount of water flowing out of the side hole is less than 2% of
the amount flowing out of the bottom outlet. The latter outflow was used in
calculating the vertical permeability. This sample maintains a relatively
constant permeability, in the low lO- s cm/range, under injection pressures
up to 145 kPa. The permeability decreases as the injection pressure in-
creases above 145 kPa. This behavior parallels what has been observed in
the high injection pressure flow testing described earlier.
5.3.5 Discussion
Bentonite permeability as a function of void ratio

Permeability of eleven bentonite (cis granular) plugs under low hydraulic
gradients are summarized in Table 5.6. Table 5.7 shows five bentonite (MX-
80) permeability measurements reported by Borgesson, Hokmark and Kar-
nland (1988). Regressed equations for permeability (in cm/s) as a function of
void ratio are given by Equation (5.2) for 11 samples and by Equation (5.3)
for 16 samples, respectively.
K = 3.72* 10- 11 (e)2.S3996 (5.2)
K = 6.68*10-11 (e)3.15926 (5.3)
Equation (5.3) gives a better fit and therefore is more suitable for estimating
the permeability of bentonite.
Table 5.6 Measured and predicted permeabilities of bentonites (c/s granular)

Sample Porosity Water Hydraulic Measured Predicted Ratio
number (n) content gradient K 1 (cm/s) K2 (cm/s) (K 1 /K 2 )
(%)
B-S-I-A 0.933 475.8 < 10 7.0 X 10- 8 6.15 X 10- 8 1.14
B-C-I-A 0.879 249.7 <12 1.2 x 10- 8 8.39 X 10- 9 1.43
B-C-1-B 0.880 251.5 < 12 1.3 x 10- 8 8.57 X 10- 9 1.52
B-C-2-A 0.831 168.1 <12 2.5 x 10- 9 2.44 X 10- 9 1.02
B-C-2-B 0.836 175.0 <12 3.7xlO- 9 2.74 x 10- 9 1.35
B-C-4-A 0.846 188.5 <12 3.3 x 10- 9 9.21 X 10- 10 3.58
B-C-4-B 0.848 191.7 <12 3.5 x 10- 9 9.57 X 10- 10 3.66
B-C-1-A-S 0.625 57.1 <57 9.0xlO- 11 6.23 x 10- 11 1.44
B-C-13/8-A-S 0.721 88.3 <77 2.2 x 10- 9 1.73 X 10- 9 1.27
B-C-23/8-A-S 0.638 60.3 <50 1.1 x 10- 10 8.58 X 10- 11 1.25
B-C-4-A-S 0.644 62.1 <57 3.0 x 10- 10 9.39xlO- 11 3.19
Predicted permeabilities were obtained from Equation (5.16), using a specific gravity of 2.92 for
the solids; wp = 50% and So = 800 m 2 /g.
Table 5.7 Measured and predicted permeabilities of bentonite (MX-80)

Saturated Porosity Water Measured a Predicted b Ratio
density (n) content K1(cm/s) K z (cm/s) (K 1 /K 2 )
(g/cm 3 ) (%)
2.1 0.421 25.08 3.0xlO- 12 2.90 x 10- 12 1.03

1.9 0.526 38.31 1.9 x 10- 11 2.36 x 10- 11 0.80
1.7 0.631 59.10 2.0 x 10- 10 6.75 X 10- 11 2.96
1.57 0.7 80.45 6.0 x 10- 10 1.60 x 10- 10 3.75
1.295 0.845 187.62 3.8 x 10- 9 2.90 X 10- 9 1.56
'From SKB Report 88-30, Figure 4.2; specific gravity = 2.9, wp = 70 (Borgenson et aI., 1988).
bCalculated from Equations (5.16) and (5.17) with So = 800m 2 /g.
Sealing performance of bentonite/crushed tuff plugs

Permeability of bentonite/crushed tuff plugs decreases with increasing ben-
tonite content. The permeabilities of the plugs containing 25 to 35% are
close to those of plugs constructed of bentonite only. The greater the uni-
formity coefficient (d 6o / dio) of crushed tuff (e.g. types FA and A), the lower
the permeability (Figure 5.5). The results suggest that an appropriate com-
position for the mixture plugs would contain at least 25 wt% bentonite
mixed with well-graded crushed rock.
The effect of sample size on sealing performance is not clear. The incon-
sistency in the permeabilities measured for different plug sizes may be due
more to variations in the stiffness of the permeameters, compaction and the
ratio of grain size to permeameter diameter (Ouyang and Daemen, 1989,

1990). For the normally consolidated mixture samples, the upward permea-
bility is about three times higher than the downward permeability.
According to the results for samples B/AL-C-8-25/A-S and BjAL-C-8-
35/A-S, the potential for piping damage to the sealing performance is small
if the maximum hydraulic gradient does not exceed approximately 120 and
280, respectively. The piping test results of sample B/AL-C-4-25/A-P-B sup-
port this deduction.
Polyaxial permeability tests indicate a difference of up to one or two
orders of magnitude between vertical and horizontal permeabilities. The
high horizontal permeability results from the uneven bentonite distribution
in the pores between rock particles due to particle segregation resulting
from the actions of sample installation and compaction. The permeability
difference may be reduced by introducing a thin layer of bentonite on top of
each compacted layer, an approach which deserves further investigation.
Temperature seems to have no negative effects on the sealing performance
within the test range from room temperature to 60C. The specific perme-
ability reaches a maximum at 35C.
5.4 PIPING AND FLOW OF BENTONITE
A breakdown of the linear relation between flow rate and hydraulic gradi-
ent has been observed in all high injection pressure tests on crushed
tuff/bentonite plugs. The abrupt departure from the initial linear relation-
ship is believed to indicate plastic flow of bentonite.
The pressure required to initiate bentonite flow likely relates to the yield
stress of the bentonite, which depends primarily upon its water content if
the type of adsorbed cations and pH of pore water remain the same. If the
yield stress of bentonite can be established as a function of water content,
the critical hydraulic gradient at which flow of bentonite takes place can be
calculated, provided the water content of bentonite and the mean pore size
of the crushed tuff skeleton are known. The long-term sealing performance
may not be impaired by bentonite loss if the maximum possible hydraulic
gradient expected in the field will not exceed the critical gradient.
The flow properties of a clay slurry or paste lie between those applicable
to liquids and solids (Scott Blair and Crowther, 1929). Bingham (1916)
introduced the concept of material which does not flow until a shear stress
'r, is reached and thereafter flows at a rate proportional to the excess shear
stress, , - 'r. Clay slurries or pastes do not follow this idealized law exactly
(Marsland and Loudon, 1963). When the shear stress, 'z, reaches 'r, shear
failure occurs near the wall of the capillary and the slurry moves as a plug.
As the pressure gradient increases, the diameter of the plug becomes smaller
until the material flows in a streamline manner like a viscous liquid. The
rate of flow then increases linearly with the pressure gradient.
Piping and flow of bentonite 83
The flow of clay within a matrix of sand or crushed rock particles is

analogous to the flow of clay paste in capillaries. For bentonite dispersed in
distilled water the yield stress depends primarily upon its water content
(Marsland and Loudon, 1963), and is expected to assume a minimum value.
Yield stresses of bentonite pastes mixed with distilled water are used to
evaluate the critical hydraulic gradient required to initiate the flow of be-
ntonite.
5.4.1 Determination of yield stress of bentonite

Bentonite pastes having nominal (distilled) water contents of 75,100,200
and 500% have been prepared and allowed to cure for 72 h in air-tight
containers prior to testing. The pastes are driven through glass capillaries
by compressed gas (helium).
Experience from pre-trials indicates that a constant shearing surface con-
dition is difficult to achieve due to the slow flow rates of the thick bentonite
pastes. To prolong the test duration solves the problem but brings about a
significant change in water content of the clay paste, resulting from the
migration of water. Shortening the capillaries was found undesirable be-
cause the clay slurry near the outlet dried out due to evaporation and
therefore impeded the advance of the paste. Since the yield stress of benton-
ite as a function of water content is the ultimate interest in determining the
critical hydraulic gradient for bentonite flow, the experiments of bentonite
flow in glass capillaries are aimed at obtaining such a relationship. The
yield stress of a bentonite paste is determined by narrowing the driving
pressures down to a small range within which a slight change of driving
pressure results in either the flow or no-flow condition. The yield stress is
computed based on the no-flow condition.
Figure 5.6 shows the yield stress of bentonite pastes vs. bentonite wt%.
The squares represent the early shear stress measurement after bentonite
flows into the capillaries. The circles indicate the shear stress at which no
advance of the clay can be detected with a measuring tape of 0.5 mm
resolution. Time elapsed before the no-flow conclusion is reached varies
from 30min to more than 24h, depending upon water content of the sam-
ples. The rate of shear strain is lower than 2.1 x 1O- 4 /s. The shear stress
computed for the 'no-flow' condition is then assumed to be the
yield stress of bentonite. Also included in Figure 5.6 are yield stresses of
Wyoming bentonite grouts (triangles) determined by Marsland and Loudon
(1963).
It is not always easy to find the no-flow condition quickly. When it takes
a long time to identify such a condition the water content of bentonite in
the capillary can be quite different from its initial value due to the migration
of water. A careful examination of experimental records is necessary to
calculate the water content corresponding to the yield stress. For the circles
in Figure 5.6 the averaged initial water content is related to the yield
~ 103
~
W
II:
!ii
c
...J
W
;;: 102
o 20.0 40.0 60.0

BENTONITE WEIGHT PERCENT
Fig.5.6 Yield stress of bentonite pastes vs. bentonite wt%. Triangle points are
from Marsland and Loudon (1963).
stress based on the early no-flow condition. Water content of bentonite

paste inside the capillary, determined immediately after testing, is used for
the no-flow condition established in the later part of a test.
5.4.2 Relation between water content and yield stress of bentonite

The following equations were obtained from curve fitting.
1. Thick bentonite pastes (70% < w < 510%):
log('f) = 13.728w-O.2818, (5.4)
where 'f is yield stress of bentonite (Pa), and w is water content (%).
log('f) = 1.6964 + 0.0417x, (5.5)
where x = bentonite wt%. Correlation coefficients (R2) for Equations (5.4)
and (5.5) are 0.999.
2. Thin bentonite pastes (720% < w < 2300%, based on the data from Mar-
sland and Loudon, 1963).
log(,d = 8.552 exp( - 0.001715w) (5.6)
Piping and flow of bentonite 85
or
log('I) = 3.841 - 0.00214w (5.7)
10g('I) = - 1.209 + 0.3017x, (5.8)

where wand x are defined earlier. Correlation coefficients are 0.998, 0.971
and 0.998, respectively. Regression on the pooled results (thick and thin
pastes) gives
log('f) = 4.47 exp( - 0.00118w) (5.9)
and
10g('I) = -1.57 + 1.4 In (x), (5.10)
with R2 = 0.934 and 0.938. Excluding the anomalous point for x = 12.5%
(w = 723%)in Figure 5.7, the regression gives
log('f) = 8.81 - 1.07 In (w) (5.11 )
and
log('f) = 1.76 + 1.44 In (x). (5.12)
R2 values are 0.938 and 0.976, respectively. These equations can be used to
predict the yield stress of bentonite when mixed with distilled water.
5.4.3 Prediction of critical gradient using the relation between

yield stress and water content
The relation between yield stress and water content has been incorpora-
ted in a model for predicting the critical pressure gradient for bento-
nite/crushed rock plugs. The model requires the water content of bentonite
and the representative pore size of the crushed rock skeleton. The critical
pressure gradient (ic,p) can be computed as follows:
.
I
2'f
=- (5.13)
c,p Rm
where Rm is representative pore radius of crushed rock skeleton. In this

study, d so (sieve aperture at 50% passing) instead of d s or d 10 commonly
used for predicting permeability of granular materials (Kenney, Lau and
Ofoegbu, 1984; Hazen, 1892), is assumed to represent Rm' This selection is
based on the consideration that the bentonite will first start flowing in the
large pores between crushed rock particles for a given yield stress. For a
crushed rock mixture of gradation type A, the value of d so is 3.9 mm. This
implies a pore diameter of 8 mm, which is close to the maximum particle
size of 9.42mm for crushed rock constituents of type A. Such an estimation
for the representative size of large pores appears to be reasonable,
considering the separation of rock particles due to the bentonite filler. The
predicted critical gradients are compared with the experimental ones for
nine bentonite/crushed tuff samples in Table 5.8.
The proposed model overestimates the critical hydraulic gradient by a
factor of between 1.11 and 2.12. The discrepancy may be due to several
factors.
The neglect of slip at the wall of the capillary resulting in the overestima-
tion of Cf.
In driving the clay paste through a capillary, part of the energy must have
been consumed by the accompanying migration of moisture; the actual
force effective for the advance of clay paste is less than the product of the
driving pressure multiplied by the cross-sectional area of the capillary.
The computed water content of bentonite at saturation is an average one.
The water content of bentonite changes as the pore pressure varies during
flow testing. The variation of pore pressure leads to consolidation near
the outflow end and swelling near the inflow end. This time-dependent
process cannot be eliminated and creates a non-uniform distribution of
water content in the sample. Because of the swelling, the water content of
bentonite near the inflow end is expected to be higher than the average
water content. The flow of bentonite therefore should first occur in the
upper part of the sample. The critical gradient extracted from a flow
rate-hydraulic gradient curve most likely corresponds to the critical
gradient for bentonite flow at a higher water content.
The yield stress of bentonite for a given water content is computed based
on the no-flow condition. Such a condition is established on the basis of
observations and is consequently limited by the resolution of the measur-
ing tape. If the condition identified actually resided in a flow region, the
yield stress thus computed would be overestimated.
5.5 PREDICTION OF BENTONITE PERMEABILITY
The frequent necessity of obtaining the permeability of a material, and the

difficulty of measuring it directly, has lead to the development of theoretical
and empirical models for the prediction of permeability. Empirical perme-
ability formulae for sand and for clay are reviewed by Loudon (1952) and
by Tavenas et al. (1983a,b), respectively. They indicate that the formulae are
limited to the type of material and the range of void ratios studied. Theor-
etical permeability models can be grouped, according to Lagerwerff,
Nakayama and Frere (1969), into two types: grain models and pore models.
'Grain' does not necessarily mean the 'soil grain' only but may include the
immovable water layers attached to the solid surface. The models developed
by Carman (1939), Schmid (1957) and Lagerwerff, Nakayama and Frere
(1969) are of the first type. Pore models include those of Childs and
Table 5.8 Predicted and experimental critical pressure gradients for bentonite flow
Sample number Saturated Sample Critical Gradient Pore radius
water content length hydraulic gradient ratio Rm
of bentonite (cm) predictedj (mm)
(o/~ Predicted Experimental experimental
BjAL-C-4-25jA-P-B 108.07 10.27 244 115 2.12 3.9

BjAL-C-8-25jA-S 115.93 10.9 207 120 1.72 3.9
BjAL-C-8-35jA-S 83.42 10.65 492 280 1.76 3.9
BjAL-C-4-25jA 120.61 9.5 189 170 1.11 3.9
BjAL-C-4-25jB 122.38 9.8 290 250 1.16 2.45
BjAL-C-4-25jC 133.33 9.8 339 273 1.24 1.725
BjAL-C-4-35jA 79.1 9.0 530 424 1.25 3.9
BjAL-C-4-35jB 92.94 9.5 560 352 1.59 2.45
BjAL-C-4-35jC 106.43 10.0 572 360 1.59 1.725
Collis-George (1950), Marshall (1958), Millington and Quirk (1959) and

Paterson (1983). Both types of theoretical models are based on the
Hagen- Poiseuille equation for laminar flow in a circular pipe.
The pore models cited above are handicapped for application to benton-
ite by the fact that they do not allow for swelling. The pore size distribution
required for the models is usually derived from the soil-moisture character-
istic, which may be subject to the effects of swelling (Lagerwerff, Nakayama
and Frere, 1969). This jeopardizes the application of the pore models to
swelling materials (e.g. bentonite). The grain models suggested by Schmid
(1957) and Lagerwerff, Nakayama and Frere (1969) require several flow
tests to predetermine some key parameters before the prediction of permea-
bility is possible. The applicability of the two grain models is thereby great-
ly reduced.
The Kozeny-Carman equation (Carman, 1937) gives good estimates for
the permeability of clean sand (Taylor, 1948; Loudon, 1952) and of quartz
powder and spherical glass particles (Carman, 1939), but fails to predict the
permeability of clays (Michaels and Lin, 1945; Lambe, 1955). The modified
Kozeny-Carman equation (Carman, 1939), incorporating the concept of
stationary water films held at the surface of clay particles, yields better
results, but only to a certain extent.
5.5.1 Kozeny-Carman equation

The Kozeny-Carman equation can be expressed in the following form:
n3
(5.14)
k= mt 2SJ(1-n)2'
where k is intrinsic permeability (cm 2 ), n is porosity, m is the shape factor of

conducting pores, t is tortuosity and So is the specific surface of the soil
particles (cm 2 /cm 3 ).
According to Carman (1937), m = 2.5 and t 2 = 2 suit most materials. The
modified Kozeny-Carman equation is given by
n3
(5.15)
k = mt2 SJ(l- n)2'
where ne is effective porosity and other parameters are as defined above.

The derivations for Equations (5.14) and (5.15) are based upon the
Hagen-Poiseuille equation and Darcy's law, and can be found in Carman
(1939) and Yong and Warkentin (1975, pp.144-6), and Carman (1939),
respectively. The assumptions involved in the derivation are uniform and
equidimensional pores and laminar fluid flow (Olsen, 1962).
According to Michaels and Lin (1954), discrepancies between measured
and computed (from Equation (5.14)) permeabilities in clays are likely to be
due to (1) interfacial phenomena (the influence of electrical forces
Prediction of bentonite permeability 89
concentrated at the liquid-solid interfaces that act on the permeating fluid)

and (2) particle packing characteristics (degree of particle dispersion and
particle orientation). Their studies on kaolinite indicate that the effects of
particle packing characteristics are primarily responsible for the discrepan-
cies. The effects of interfacial phenomena are minor. They further conclude
that the latter effects could be attributed primarily to counter electro-osmo-
sis and consequently that the thickness of immobilized liquid films on the
surface of solids must be extremely small, less than 4% of the diameter of
the pores. The conclusion of the limited thickness of immobilized liquid
films may be supported by the results reported by Rosenquist (1955) and
Aylmore and Quirk (1960). An important deduction from Equation (5.15) is
that clays may have zero permeability at considerable porosities (where the
effective porosities become zero), e.g. at n = 0.207 for a clay soil and
n = 0.355 for a plastic clay (Carman, 1939).
Olsen (1962) investigated the effects of several factors on the failure of
Equation (5.14) in saturated clays: (1) possible violation of Darcy's law, (2)
electrokinetic coupling, (3) high viscosity, (4) tortuous flow paths and (5)
unequal pore sizes. His results show that (1) the possible violations of
Darcy's law and electrokinetic coupling are insignificant, (2) high viscosity
and/or tortuous flow paths fail completely to account for the discrepancies
and (3) unequal pore sizes can explain all the discrepancies.
Discrepancies between measured and predicted flow rates for kaolinite,
illite and Boston blue clay, obtained from consolidation permeation tests by
Olsen (1962), are explained by a model he proposed. The model consists of
clusters that are equidimensional, uniform in size and porous. Three par-
ameters define the model pore geometry: (1) N, the number of particles per
cluster; (2) ec, the intra-cluster void ratio; and (3) ep ' the inter-cluster void
ratio, which equals the total void ratio minus the intra-cluster void ratio,
eT - ec. Since flow rates are proportional to the fourth power of pore radii,
the contribution of the flow component through the cluster pores is as-
sumed to be negligible.
Based on an assumed relationship between the total, intra- and inter-
cluster void ratios, Olsen was able to produce discrepancies for systems of
clusters which are similar to those measured. At high total void ratios the
compressibility of the individual clusters is considered negligible compared
to that of the cluster skeleton. When the clusters approach a density corre-
sponding to the densest possible packing of spheres, the clusters themselves
begin to compress as the total void ratio is decreased. At this stage the
inter-cluster pores and consequently the flow rates are likely to change only
a little, while the predicted flow rates are smaller due to the reducing total
void ratios. This behavior may explain why, at porosities less than about
0.4, measured flow rates decrease less rapidly with decreasing porosity than
predicted.
Although the cluster model provides possible and reasonable explana-
tions for the discrepancies between measured and predicted permeabilities
in clays, a refinement of the Kozeny-Carman equation is hindered by

the difficulties in determining the parameters needed to describe the pore
geometry.
5.5.2 Refined Kozeny-Carman equation for clays

The Kozeny-Carman equation needs to be modified for the prediction of
permeabilities in clays. Such an improvement requires a description of the
complex pore geometry of clays. The modified Kozeny-Carman equation
(Equation (5.15)) only takes care of the effects of the immobilized liquid
films at solid surfaces on the pore size. Considering the influences of compli-
cated water-solid-electrolyte interactions on the soil structure, a complete
mathematical description is not attempted. Instead, a possible collective
factor (a water content ratio) to account for the changes in pore geometry of
clays is examined.
Atterberg limits are water contents where the soil behavior changes
(Holtz and Kovacs, 1981, p.36). Depending on its water content, a fine-
grained soil can exist in any of four states: solid, semisolid, plastic and
liquid state. The water content corresponding to the transition between
adjacent states is termed shrinkage, plastic and liquid limit, respectively
(Lambe and Whitman, 1979, p. 33). At the plastic limit the particles or units
of particles slide past one another upon application of force but there is still
sufficient cohesion to allow them to retain a shape (Y ong and Warkentin,
1975, pp. 66-7). At the liquid limit the cohesion becomes too small to retain
a definite shape and the material acts as a liquid.
Water contents such as plastic and liquid limits and other related indices
correlate with engineering properties such as the undrained shear strength,
compression index and compression ratio (e.g. Skempton, 1944; Worth and
Wood, 1978; Nagaraj and Srinivasa Murthy, 1983, 1986; Pandian and
Nagaraj, 1990). The correlations lead to postulation that water content of a
fine-grained soil describes collectively a possible equilibrium state of the
soil's structure, and therefore a corresponding state of pore structure.
The vertex of the parabola-like curves describing clay permeabilities as a
function of porosity (Olsen, 1962) should indicate a unique state of pore
structure. According to Olsen's cluster model, this unique state may corre-
spond to the densest possible packing of clay clusters. Further reduction in
total void ratio will be due primarily to the compression of the clusters
themselves. The water content related to the unique pore structure of a soil
should be very close to its plastic limit, at which particles or units of
particles slide past one another upon application of force. The plastic limit
shall be used to represent (indirectly) the unique state of pore structure.
The Kozeny-Carman equation may then be refined as
(5.16)
Prediction of bentonite permeability 91
for w> wp, where w is water content and wp is the plastic limit; and
k _ (wp) n3 (5.17)
- w mt 2S~(1 - n)2
for w < wp (i.e. highly compacted clays).
Conceptually, saturated fine-grained soils having a water content less
than the shrinkage limit should be considered as solids which have zero
permeability. The lower bound of water content, w, for Equation (5.17) may
therefore be set at the shrinkage limit. This lower bound can also be a water
content at which all the water is held firmly by the solids, based on the
concept of immobilized liquid films. The water content of a soil sample at
saturation can be expressed as
1 n
W=--- (5.18)
Gs (1- n)'
where Gs is the specific gravity of the solid. Using this relation, Equations
(5.16) and (5.17) can be reduced to
k= (~J (~J mt2S6~: - n)3 (5.19)
and
(5.20)
5.5.3 Validation of the refined Kozeny-Carman equation

Approximately 90% of American Colloid cis granular and MX-80 benton-
ite are montmorillonite (American Colloid Company, Data No. 202).
Montmorillonite has an estimated specific surface of 760-800m2 Ig (Quirk,
1968; Shainberg, Bresler and Klausner, 1971; Mitchell, 1976, p. 45; Yong
and Warkentin, 1975, p. 46). Using the specific surface value of 800m 2 /g
and assuming the other 10% of the materials have negligible effects on the
permeability, the refined Kozeny-Carman equation has been used to check
measured permeabilities (cis granular bentonite), as well as those of MX-80
bentonite in the literature (Borgesson, Hokmark and Karnland, 1988). The
plastic limit and specific gravity are 50% and 2.92 for the cis granular, and
70% and 2.9 for the MX-80 bentonite. The predicted k (cm 2 ) values are
converted to k (cm/s) values for water at 21C by multiplying by
9.799 X 10 4 . The results are shown in Tables 5.6 (next to last column) and
5.7, respectively.
For 11 cis granular bentonite samples the ratios between measured and
predicted permeabilities vary from 1.02 to 3.66. For five MX-80 samples the
ratios range from 0.8 to 3.75. Such narrow deviations substantiate the
usefulness of the refined Kozeny-Carman equation proposed for the predic-

tion of permeabilities in clays. The credibility of the equation is further
enhanced by the wide range of porosities (0.42-0.93) covered in the model
validations. Moreover, according to Equation (5.19), permeability should be
linearly related to the porosity function of n4/(1 - n)3 for clay samples
having a water content greater than wp (plastic limit). Such a relationship is
demonstrated in Figure 5.7.
5.5.4 Discussion
The refined Kozeny-Carman equation can predict permeabilities of
Wyoming sodium bentonite mixed and permeated with distilled water.
The equation is believed to be able to handle other fine-grained soils and
situations of different pore-water chemistry. The deduction is based on
three reasons: (1) the refined equation can account for the common parab-
ola-like discrepancies observed between measured and predicted (from the
Kozeny-Carman equation) permeabilities in different clays (Olsen, 1962);
(2) the specific surface and the plastic limit change with materials, there-
fore implicitly accounting for the material type, (3) changes in pore-water
chemistry should result in different values of the plastic limit as they do
/
/
/
~/
/
/
6f!';..
/",
/
/
/
'"
;.;' /
'"/ /
~
10-11 L....l-l-L..U.Ju.L---''--'-....L.J....l..1..1.l.L----1---,-....L..1....LJ...J..l.L----1--.J
10 100 1000
POROSITY FUNCTION
Fig.5.7 Relationship between the permeability of bentonite and the porosity func-
tion n4 /{1 - n)3.
Conclusions and recommendations 93
for the liquid limit (Borgesson, Hokmark and Karnland, 1988); the sugges-
ted water content ratios may still account for the effects of changes in the
pore geometry.
The same liquid as used for the permeation determination should be
employed in determining the liquid limit. Further investigations are recom-
mended.
5.6 CONCLUSIONS AND RECOMMENDA nONS
5.6.1 Conclusions
Permeability tests on sedimented bentonite plugs indicate the dependence of
permeability on the composition of molding water and permeant. The ben-
tonite sample deposited in and permeated with the synthetic groundwater
gives a permeability (7 x 10- 8 cm/s) very similar to that of the sample
prepared and tested with deaired distilled water. Permeability of bentonite
can be reduced by molding or by percolating with a dispersing solution (e.g.
2% sodium pyrophosphate solution).
Mixtures of bentonite and crushed densely welded Apache Leap tuff can
be engineered to yield a permeability close to that of pure bentonite. One
appropriate composition to reach that goal would contain at least 25 wt%
bentonite mixed with well-graded crushed rock. A mixture containing 25%
bentonite and 75% crushed tuff of type A (maximum particle size of
9.42 mm) gradation appears to be a promising seal material. Limited test
results suggest that crushed tuff of FA or FC gradations (Fuller-Thompson
gradations, n = 0.5 and Dmax = 9.42 mm and 19.05 mm, respectively) may
also be good candidates for mixing with bentonite. The sealing performance
of mixture plugs is enhanced by increasing the amount of bentonite to 35%.
The increase in bentonite content improves the bentonite occupancy per-
centage and reduces the water content of bentonite at saturation, giving
better resistance to piping, erosion and flow. Similar effects have been ob-
served if crushed rock constituting a Fuller-Thompson grading curve (e.g.
type FA with n = 0.5) is used.
Compaction and the amount of bentonite are decisive factors in produc-
ing good mixture seals. The effectiveness of compaction in reducing porosity
is hindered by the soft bentonite buffer. To reduce the bulk porosities of the
mixture plugs containing 25 wt% or more bentonite, a compaction energy
higher than that of the standard Proctor compaction is necessary.
Bentonite/crushed tuff mixtures tested in this study exhibit heterogeneity
and anisotropy. A difference of up to one or two orders of magnitude can be
expected between the vertical and horizontal permeabilities. The higher
horizontal permeability results from the uneven bentonite distribution
in the pores between crushed rock particles due to particle segregation
resulting from sample installation and compaction. Moreover, the contact

between individually compacted layers may serve as a preferential flow
path. Increasing bentonite content from 25 to 35% in the mixtures reduces
the vertical permeability by nearly one order of magnitude but results in
little change in the horizontal permeability. Adequate sealing ability of the
mixture plugs in the transverse direction may be necessary to minimize the
possibility of the flow of contaminated groundwater or gases laterally into a
connected fracture system in a host rock formation. Moreover, compromis-
ing the sealing ability in the transverse direction ultimately may jeopardize
the entire sealing performance if piping occurs radially. This consideration
can be significant if borehole seals are installed at locations which are
intercepted by joints and/or fractures.
Temperature has no negative effect on the sealing performance of benton-
ite/crushed tuff plugs within the test range from room temperature to 60C.
The specific permeability reaches a maximum at 35C and decreases with
increasing temperature, indicating the effect of temperature on the structure
of the samples. The decreases in the specific permeability are likely to be
due to the thermal expansion of crushed tuff particles and the expansion of
the diffuse double layer of bentonite. The structural change is reversible
within the temperature range tested.
The possibility of piping in passageways created by the radial expansion
of pores due to an increasing injection pressure is small except for plugs
containing a low bentonite content (e.g. 15 wt%). The effect of pore expan-
sion is believed to be counteracted by pore clogging resulting from the
migration of fine particles. The fine particle migration is evidenced by
bentonite flow between crushed tuff aggregates. The migration may be
further supported by the breakdown of the linear relation between flow
rate and hydraulic gradient demonstrated in all high injection pressure
flow tests. For mixtures of type A crushed tuff and 25% or 35% bentonite,
piping damge is small if the maximum hydraulic gradient does not exceed
approximately 120 and 280, respectively. Piping can occur if bentonite is
lost externally. The piping model developed in this study combines yield
stress characteristics of bentonite and the flow of bentonite through capil-
laries. The model provides an analytical means to determine the critical
pressure gradient at which bentonite of a given water content may start
to flow.
The modification made to the Kozeny-Carman equation includes a cor-
rection factor to account for the microstructural changes in clays, respon-
ding to the changes in water content. Permeability measurements of 11
samples in this study, together with five measurements reported in the
literature, are used to examine the validity of the model. Predicted bentonite
permeabilities agree well with the experimental ones over a wide range of
void ratios. The ratio between the predicted and measured permeabilities
varies from 0.8 to 3.75. The prediction error is within 60% of the measured
permeability for 11 of 16 samples.
Acknowledgements 95
5.6.2 Recommendations for further studies
High injection pressure flow tests and piping tests in the transverse direc-
tion are recommended to evaluate the consequences of the permeability
anisotropy.
Methods are needed to minimize the particle segregation and to assure a
uniform distribution of bentonite. Permeability anisotropy may be reduc-
ed by emplacing a layer of bentonite on top of each compacted layer. The
crushed rock, during subsequent compaction, should carve into the be-
ntonite layers above and below to tie together adjacent lifts.
The proposed permeability model can predict permeabilities of Wyoming
sodium bentonite mixed and permeated with distilled water. Different
material type and pore-water chemistry are likely to result in changes
only in parameters of the specific surface and plastic limit. Further studies
are recommended to verify these postulates.
For bentonite molded with distilled water, the yield stress is expected to
assume a minimum value due to the development of a dispersive micro-
structure. When the pore-water chemistry changes, bentonite can have a
flocculated structure and thus a higher yield stress. Studies of the influ-
ence of pore-water chemistry on the yield stress of bentonite are recom-
mended.
The effect of bentonite loss into fractures on the sealing performance
deserves further investigation. This effect may be evaluated by conducting
flow tests on seals installed in permeameters with rectangular slits of
carefully controlled dimensions. Such a test configuration is more repre-
sentative of in situ conditions than a circular opening in the wall of a
permeameter.
Flow of bentonite in capillaries deserves further investigation. The macro-
scopic analysis presented here may oversimplify the flow patterns and
mechanics. A detailed observation and description of bentonite flow in
capillaries should assist in identifying any major shortcomings in present-
ly available flow models.
ACKNOWLEDGEMENTS
tory Commission, under contract NRC-04-86-113. Support and permission
to publish this chapter are gratefully acknowledged.
CHAPTER SIX
In situ Performance of a
Clay-Based Barrier
B. H. Kjartanson, N. A. Chandler and A. W L. Wan
6.1 INTRODUCTION
Bentonite, a processed clay product composed primarily of the clay mineral

montmorillonite, has been used or specified for use as a barrier material in a
variety of waste containment facilities. When compacted and applied in an
earthen waste containment barrier the surface-active clay mineral mont-
morillonite can impart the desirable properties of low hydraulic conductiv-
ity, low contaminant diffusivity and chemical retardation. Moreover,
depending on the rigidity of confinement, the dry density of bentonite with-
in the barrier and pore fluid/contaminant chemistry, bentonite barriers may
develop up to tens of MPa of swelling pressure against rigid confinement.
Bentonite/soil mixtures with bentonite content ranging from 100% benton-
ite, in a highly compacted form, to 5% bentonite, by dry mass proportion,
have been specified for use as barrier materials. The 'soil' component may
be a specifically designed sand-aggregate. The design of the particular ben-
tonite/soil mixture will depend to a large degree on the physical and
geologic setting of the proposed barrier, the nature of the wastes coming
into contact with the barrier material and barrier construction procedures.
For example, barriers for underground waste containment may use 100%
highly compacted bentonite in situations in which the bentonite will be
rigidly confined by the surrounding rock and rigid restraining structures
such as concrete plugs and bulkheads. As an example, the Swedish KBS-III
concept for radioactive waste disposal in an underground repository is
considering the use of highly compacted bentonite to seal waste emplace-
ment boreholes, tunnels, shafts and exploratory boreholes (Gray, 1993). In
contrast, lean bentonite/soil mixtures typically ranging from 5% to 20%

Edited by K. Fuenkajorn and 1. J. K. Daemen.
Canadian nuclear fuel waste management program 97
bentonite by dry mass have been specified and used for landfill liners and
covers where locally available soils would by themselves not meet the per-
formance requirements (USEPA, 1988; USEPA, 1991). Moreover, lean ben-
tonite mixtures have been designed and applied for the containment of
brine in surface impoundments (e.g. Haug, Barbour and Longval 1988) and
assessed for use as waste containment liners (e.g. Evans and Quigley, 1992;
Kenney et al., 1992).
In all cases the design, performance assessment and installation of ear-
then barriers required a combination of laboratory and field tests. For
example, the design of the appropriate bentonite/sand mixture, assessment
of compaction characteristics and assessment of contaminant/barrier ma-
terial compatibility and contaminant transport properties (e.g. Evans and
Quigley, 1992) would most effectively be carried out through laboratory
testing. However, in order to include the effects of scale in performance,
such as secondary permeability features including macropores and fissures,
in situ hydraulic conductivity tests should be carried out (Daniel, 1989).
This chapter describes the processes of physical performance assessment
of a bentonite-sand earthen barrier material specially designed for nuclear
waste containment in an underground repository environment. Laboratory
tests have been carried out to arrive at and define an appropriate benton-
ite/sand mixture which would give the best overall combination of physical
performance characteristics. Moreover, these tests have been used to define
important fundamental properties, such as hydraulic conductivity and ionic
diffusivity. Laboratory-scale model tests have been carried out to examine
the barrier's physical performance under potential varying boundary condi-
tions. This chapter focuses on the assessment of the in situ performance of
the barrier material within a representative geological setting. While aspects
and results specific to the performance of the nuclear waste containment
barrier are described, the methodologies developed and applied and lessons
learned could readily be applied to the performance assessment of engineer-
ed barriers for other applications.
6.2 THE CANADIAN NUCLEAR FUEL WASTE MANAGEMENT

PROGRAM
In order to investigate aspects of disposal vault design and operation, a

conceptual design study has been completed (Simmons and Baumgartner,
1994). In this study, corrosion-resistant fuel-waste containers will be placed
in boreholes drilled into the floors of disposal rooms and separated from
the host rock by a bentonite-sand buffer material (Figure 6.1). As the vault
will be below the natural groundwater table, exploration boreholes, access
shafts and ramps, disposal rooms and disturbed rock surrounding
the underground excavations will need to be sealed effectively to control
98 In situ performance of a clay-based barrier
1.2 m diameter
Emplacement
Boreholes
Fig 6.1 The emplacement borehole design for nuclear fuel waste disposal.
groundwater flow and to inhibit waste dissolution and the potential release
of radionuclides.
In the Canadian disposal concept a near-field seal, termed the buffer, is
required to inhibit the transport of water and radionuclides, to protect and
support the containers, and to conduct effectively radiogenic decay heat
from the container to the surrounding rock. Keeping these requirements in
mind, a research program was carried out to define the properties of candi-
date materials for constructing seals in waste repositories (Dixon and Gray,
1985). Through this research it was found that a 1: 1 dry mass ratio of
sodium bentonite clay and silica sand compacted to its maximum dry den-
sity at its optimum moisture content would give the best overall combina-
tion of self-sealing ability, low hydraulic conductivity, low ionic diffusivity,
strength and thermal conductivity. For vault operational reasons, the buffer
would be compacted in situ. According to the studies by Dixon and Gray
(1985), the buffer should be compacted to a dry density of 1.67 Mg/m 3 ,
corresponding to 95% modified Proctor maximum dry density at a moist-
ure content near optimum (17-19% by mass). In this state the initial degree
of saturation of the buffer is between 75 and 85%.
Shortly after emplacement, container temperatures will rise in response to
the heat generated by the wastes. The buffer near the container could
undergo drying, shrinkage and cracking due to heat-induced moisture
Evaluation of buffer physical performance 99
movement down the thermal gradient away from the container
(Radhakrishna et al., 1992). These processes tend to reduce the thermal
conductivity of the buffer, resulting in higher container temperatures and
perhaps further moisture movement. The thermal, hydraulic and mechan-
ical properties of the buffer near the rock will depend on the degree of
thermally induced moisture redistribution and the moisture boundary con-
ditions. Water availability and hydraulic pressures at the buffer-rock
boundary will be controlled largely by the hydraulic characteristics of the
disturbed rock immediately adjacent to the excavations.
In the longer term, after final closure of the vault, groundwater pressures
in the rock will increase in response to the regional groundwater levels, and
water uptake by the buffer, driven by total water potential gradients, should
increase the degree of saturation of the buffer mass. The buffer will swell
and develop pressure against the container, the surrounding rock and the
backfill as it saturates. Under these conditions it is envisaged that any
cracks and gaps formed during the early stages from thermally induced
drying should self-seal (Dixon et ai., 1993).
To a large extent, the thermal-hydraulic-mechanical performance of the
buffer depends on local moisture conditions and the moisture flux bound-
ary conditions (i.e. hydraulic interaction between the buffer and the rock).
Conversely, distributions of water content and suction in the buffer are
influenced by temperature, temperature gradients and hydraulic fluxes.
Thus an understanding of fundamental processes prior to buffer saturation
and their potential impact on long-term buffer performance is required.
6.3 EVALUATION OF BUFFER PHYSICAL PERFORMANCE
Laboratory-scaled physical model tests have been carried out to examine

heater- buffer interactions in a simulated emplacement borehole environ-
ment (Radhakrishna et ai., 1989; Selvadurai, 1991). These tests allow a lar-
ger volume of buffer to be tested (up to about 35000 cm 3 ) and allow an
evaluation of the adequacy of the fundamental material properties to de-
scribe buffer performance under a variety of well-controlled boundary con-
ditions. While the laboratory tests offer the advantages of simplicity of test
set-up, and easily varied and well-controlled boundary conditions, a major
issue still remains. The adequacy of the laboratory-derived fundamental
material properties to describe field performance at or near full scale in a
realistic environment needs to be confirmed. Of particular importance is an
assessment of the effects of in situ hydraulic and mechanical boundary
conditions on buffer performance. As noted above, issues such as the degree
of thermally induced drying, the rate of water uptake and swelling pressure
development/self-sealing all depend on the hydraulic boundary conditions.
These issues are being addressed through large-scale in situ testing. In the
Canadian program, an Underground Research Laboratory (URL) has been
constructed in the plutonic rock in the Canadian Shield, near Lac du

Bonnet in Manitoba. The URL provides a representative environment for
conducting large-scale multidisciplinary experiments relevant to the dis-
posal concept.
6.4 IN SITU EXPERIMENTS
The Buffer/Container Experiment was installed in the URL to assess the

in situ performance of the buffer material within a representative geological
setting. The experiment is a full-scale simulation of one element of the
in-floor borehole emplacement configuration for waste disposal. The experi-
ment arrangement (Figure 6.2) permits the assessment of buffer performance
under an applied thermal load from a central heater and under conditions
of available moisture from the intact, unfractured rock. The main objectives
Fig 6.2 Configuration of the Buffer/Container Experiment.

In situ experiments 101
of the experiment are to examine the interactions between the buffer, rock
and container under elevated temperature, and to develop the technologies
for future in situ experiments.
The second experiment was the Isothermal Buffer/Rock/Concrete Plug
Interaction Test (Figure 6.3). Temperature gradients in the Buffer/Container
Experiment act as a driving mechanism for moisture movement in the
buffer, and temperature changes in the rock influence the near-field pore
pressures. Both these effects influence the interpretation of moisture transfer
within the buffer and the rock. Hence, an isothermal experiment was carried
out to examine water uptake by the buffer under isothermal conditions.
In addition to the assessment of buffer performance, the Buffer/Container
Experiment and the Isothermal Test provide data for the qualification and
further development of numerical and conceptual models for the processes
of heat and moisture transfer. Experience is also gained with respect to
geotechnical instrumentation, underground materials handling, in situ buf-
fer compaction, large-diameter borehole drilling in hard crystalline rock
and concrete plug design and placement. The information gleaned from
these experiments is directly applicable to the design of future sealing ex-
periments at the URL.
+
legGl'!d:
~
" P1Iy!:htoo\el$r
~ PLAN = Eal'lh Pr_un! Cell
Pookllfstrlngs m PooklIf String ,
Plezomateffl , Pnlll!l'm\llc Ple%00I$I :
~ t ~ Hydraulic ~r ;
Aa4ial strain OOI!$ ! Radial Strain C'lOlI!"mmmJ
Mil'
P~1ar!I
~
DimeMlona in rom Eal ~NI oo/Is
JeT eIoo!rode$
SECTION
Fig 6.3 Configuration of the Isothermal Test.

6.5 BUFFER/CONTAINER EXPERIMENT
6.5.1 Experiment design

The Buffer/Container Experiment was installed in a full-scale emplacement
borehole drilled into granitic rock. The borehole was 1.24 m in diameter
and 5.0 m deep, in which the buffer was compacted in situ. By conducting
the experiment at full scale, any possible influence of scale was removed.
The geometry of the experiment heater is the same as the design geometry
of a waste container. The heater is operating at a constant setting of
1200 W, producing a near steady-state skin temperature of 85C. The heater
was designed with triple redundancy in its heater elements and double
redundancy in its controls in order to ensure continued operation of the
heater over the lifetime of the experiment. The cavity in which the heater
sits was created by compacting the buffer in the annulus region created by a
735mm diameter metal form and the rock wall of the 1240mm diameter
borehole. The form was removed prior to heater installation, and the 50 mm
void between the heater surface and the buffer was filled with dry sand. The
sand was placed in the void by means of the controlled pouring technique.
One meter of backfill material, which consists of 75% granite aggregate
and 25% clay, was placed at the top of the experiment as shown in Figure
6.2. The entire system of buffer, heater, sand and backfill was sealed using a
steel cap and neoprene gasket. The void between the cap and backfill was
grouted with a high-strength sand-cement mortar mixture. A restraint
system was designed to minimize the displacements at the top of the experi-
ment. Prevention of the upward movement of buffer and backfill inhibits
reduction in measured earth pressures caused by vertical displacement. The
six columns of the restraint system were designed to transfer the load gener-
ated below the steel cap up to the roof of the excavation. The columns were
prestressed to reduce the compliance between components of the system,
and were designed so that a design swelling pressure of 2 MPa from the
buffer below the cap would result in about 1 mm of vertical displacement.
Conventional geotechnical instrumentation was installed in the experi-
ment to measure changes in temperature and pressure, and deformation in
the buffer-rock-container system during the course of the experiment. To
measure temperatures, thermistors and thermocouples were installed within
the buffer, on the heater and grouted in boreholes drilled into the rock. The
moisture contents of the buffer were monitored by thermal needles and
psychrometers, while earth pressures were monitored using both vibrating
wire and pneumatic sensors. In the rock, pore pressures were monitored in
packer isolated zones and by pneumatic and hydraulic piezometers, while
vertical displacements were measured using extensometers, and horizontal
stress changes were recorded by borehole radial strain cells. In all, over 600
sensors were installed in the Buffer/Container Experiment, most of which
are data logged automatically.
Buffer/container experiment 103
6.5.2 Location
The Buffer/Container Experiment was installed at the 240 Level of the
URL. The rock in the experiment room is classified as medium-grained
gneissic to schlieric grey granite, and hosts various steeply dipping pegma-
tite and granadiorite dykes and shallow dipping leucocratic sills (Woodcock
et al., 1991). The axis of the experiment room has been aligned with the
direction of the major principal in situ stress in the rock. The magnitudes of
the principal stresses at a depth of 240 m have been determined to be about
30 and 15 MPa subhorizontally and 12 MPa subvertically. At the 240 Level,
rock displacements near excavated openings are elastic, with little or no
stress-induced damage to the rock.
6.5.3 Excavation and excavation damage

The experiment room was excavated using a pilot-and-slash, controlled
blasting technique. The room is 31 m in length with the nominal width and
height being 4.8 m and 5.8 m respectively. The blast sequences were designed
to minimize excavation damage in the rock due to blasting. However, exam-
ination of the cores from holes drilled into the floor of the room indicate an
excavation damage zone that varies from a few millimeters in the wall to as
much as a meter in the floor. The excavation-induced fractures extend to a
depth of 40 cm in the floor at the location of the Buffer/Container Test.
Hydraulic characterization of the uppermost meter of rock was carried
out using pressure pulse tests in 38 mm diameter boreholes. The transmis-
sivities measured in the uppermost 20 to 40cm of the rock increased by two
to five orders of magnitude. This coincides with the mapped depth of exca-
vation damage of 40cm. Transmissivities measured using pressure pulse
tests at depths of 0.4 -15.0 m are relatively constant with depth and are
consistent with those of the intact rock, with hydraulic conductivity of
about 10- 13 m/s.
6.5.4 Hydrogeology
The hydrogeology of the URL is largely controlled by the presence of
low-dipping fracture zones. The 240 Level is located in a wedge of essential-
ly unfractured grey granite bounded by Fracture Zone 2 (over 50m below
the Buffer/Container Experiment) and Fracture Zone 2.5 (approximately
45 m above). Although hydraulic pressure in the rock at large distances
from the 240 Level excavations approaches the hydrostatic value of about
2000 kPa, the drawdown effect of the URL reduces the pore-fluid pressure
in the vicinity of both the Buffer/Container Experiment and Isothermal Test
to about 1600kPa (Figure 6.4). Due to the lack of fractures or other hy-
draulically conductive zones at the chosen sites for the Buffer/Container
Buffer/Container
Experiment
P~
Isothermal
T
",# A
~~4
-=:J
o 20m
130 m Level Borehole Intersection

Packer Location
3 Zone Number
-1S00-Pressure Contour 1991 May
(kPB)
Fig 6.4 Drawdown of hydraulic pressure at the 240 Level of the URL.
and Isothermal experiments, the water available for moisture uptake by the
buffer material will only be that capable of flowing through the intact rock.
To assess the availability of water at the boundary of the emplacement
borehole, the inflow into the borehole was collected and measured immedi-
ately after drilling the hole (Chandler et al., 1992). Five conical-shaped, vinyl
collection rings were installed in the 5 m deep emplacement borehole, with
consecutive rings being 1 m apart (Figure 6.5). The water flowing into each
ring was collected for a period of 5 weeks, at which time the rate of inflow
into the emplacement borehole had stabilized. Inflows into the lower four
collection rings ranged from 3 to 8 mLjday. This rate of inflow is consistent
with a far field hydraulic boundary of 1600 kPa and a hydraulic conductiv-
ity of 10- 13 to lO- 12 mjs.
Inflow into the uppermost ring was three orders of magnitude higher.
This high inflow was attributed to water flowing through the fractures in
the 40cm thick excavation damage zone in the floor of the room. The
source of the water was a nearby reservoir on the 240 Level. A 50 cm deep
inflow barrier was therefore constructed by drilling overlapping boreholes
around the collar of the emplacement borehole and filling them with grout.
Subsequent inflow measurements confirmed the effectiveness of this grout
curtain.
Styrofoam
Cover Plate
O.Om
Water Collection
Pipe
1.0
Steel Ring/Gland
Vinyl Liner
2.0
~Vibrating Wire
Pressure
Transducer
3.0 Transducer
Cables
Packer
Inflaction
4.0 Lines
Emplacement
Borehole
5.0
Fig 6.S System of measuring inflow into the emplacement hole of the Buffer/
Container Experiment.
6.5.5 Experiment implementation

A number of technology development tasks had to be completed before the
Buffer/Container Experiment and the Isothermal Test could be installed.
These included the following:
selection and assessment of instrumentation, and assessment and devel-
opment of installation procedures for the instrumentation to monitor the
performance of the buffer;
development of specialized equipment for drilling the emplacement
borehole;
development of equipment and procedures for in situ buffer and backfill
compaction; and
design, construction and qualification of an electric heater to simulate the
waste container.
6.5.6 Instrumentation assessment

Instrumentation to monitor the performance of the buffer had to meet a
number of operational requirements. The instruments had to operate
reliably of temperatures up to 90C and total pressures (pore pressure plus
swelling pressure) of up to about 4 MPa for at least 2 years. In addition, the

instruments needed to be robust enough to allow installation in conjunction
with in situ buffer compaction and have sufficient corrosion resistance to
operate in the elevated temperature, saline pore fluid environment. Stan-
dard or slightly modified geotechnical instruments were given the highest
priority: simple, robust and compact instruments which meet the geo-
metric limitations of the experiment were selected and assessed for use in
the tests.
Before installation in the tests a laboratory assessment and calibration
program was carried out to evaluate the suitability and durability of the
candidate instruments, to gain a better understanding of the response of the
instrument in contact with buffer and to develop and optimize interpreta-
tion methods for the sensors. Temperature sensors, earth pressure cells and
transducers, piezometers and moisture and water potential sensors were
evaluated. Details of this assessment program are presented by Kjartanson
et al. (1989b, 1990).
Recognizing the importance of in situ water content and suction measure-
ment in understanding the in situ performance of the buffer in terms of
moisture mass balance and water uptake, significant work has continued on
the enhancement of monitoring and interpretation technologies for the
moisture content sensors. Conventional thermocouple psychrometers are
used as moisture content sensors to track spatial and temporal distributions
of suction and water content in the buffer. The psychrometers used in both
the Buffer/Container Experiment and the Isothermal Test are manufactured
by Wescor Inc. in Utah, USA. The main part of the instrument, which
measures 6 mm in diameter and 25 mm in length, is basically a thermo-
couple circuit protected by a ceramic shield. The instrument measures the
relative humidities in the soil from which the total suction can be inferred
using the Kelvin equation (Hillel, 1980). The procedures for assessing water
contents in the buffer involve using psychrometers to determine suctions in
the buffer, and then converting suctions to water contents using the suc-
tion - water content relationship determined from laboratory tests. Recent
work has shown that the use of thermocouple psychrometers in compacted
clays such as the buffer are influenced by total stresses, and by thermal and
vapour pressure transients in the compacted materials. Results from labora-
tory studies show that under 'undrained' constant mass and moisture condi-
tions, compressive total stresses in the buffer could lead to an increase in
humidity in the unsaturated material. Under these conditions the psych-
rometer will measure a suction which does not properly reflect the actual
water content in the unsaturated buffer. Furthermore, sensitivity calcula-
tions show that, under the influence of thermal gradients, water vapour
movement within the buffer could lead to 100% relative humidities in the
unsaturated buffer without significant increases in the moisture content of
the soil water. Under this circumstance the effectiveness of psychrometers as
moisture content sensors is seriously limited.
6.5.7 Emplacement borehole drilling

The 1.24 m diameter emplacement borehole for the experiment was drilled
using a specially developed high-pressure rotary water-jet drill (Kjartanson
et aI., 1989a; Puchala et al., 1989). The drill was capable of advancing the
borehole at a rate of 50 mm/h. Unfortunately, the drill and high-pressure
components, such as the pump and cutting nozzles, were prone to wear and
failure, which reduced overall production. The hole was advanced using 1 m
long drill runs, with the rock stub broken out using expansive cement. The
final depth of the hole below the top of the concrete floor slab in the test
room was 5.8 m.
6.5.8 In situ buffer compaction

An important aspect of the Buffer/Container Experiment was the develop-
ment of an in situ buffer and backfill compaction technique compatible with
the underground geometric constraints and test requirements. The tech-
nique must yield a low statistical variation in the compaction density, allow
installation of instruments without damage and allow compaction of the
buffer both full face and around the collapsible liner used to form the heater
cavity (Kjartanson et aI., 1992). A manually operated hydraulic-powered
impact hammer was developed and qualified in laboratory compaction
trials.
Field trials with the impact hammer were carried out at the URL in a test
emplacement borehole to refine the compaction procedures for an under-
ground environment and to develop quality control and assurance pro-
cedures for the Buffer/Container and Isothermal Tests. In these field trials a
number of different conventional quality-control procedures, including the
sand replacement and rubber balloon methods, were examined. The con-
clusion reached after the trials was that the quality-control methods were
generally not accurate nor reliable, or could not be properly conducted
within the confines of the emplacement borehole. The most reliable method
was to remove an entire lift and directly calculate the density using the wax
density method, but this would be too disruptive. It was decided, therefore,
to develop, verify and use a method specification.
Detailed compaction procedure checklists were developed to provide
quality assurance for the method specification. Ten lifts were compacted in
the trial emplacement borehole following these procedures and checklists to
ensure that the method specification provides an acceptable average lift
density with an acceptable variation about that average. The results of the
trials with both buffer and backfill (Figure 6.6) show that following the
method specification would produce uniform, compacted lifts to the re-
quired specification, i.e. densities greater than 95% of the modified Proctor
maximum dry density with lift density standard deviations well within the
1.80
AROUND LINER : FULL LIFT
1.78
I
I
I
1.76 I
MODIFIED PROCTOR MAXIMUM DRY DENSITY
1.74
'"E 1.72
"0
:::;
~ 1.70
enZ
w 1.68
I OVERALL AVERAGE DRY DENSITY
C
>-
It
I
C
1.66 --- --- ---950/'; -r;,,-obH=lED -P-Roct?FfMAXiMUriin5i=i'rEfENSIW- -- ---- ------ ----
I
1.64
: + AVERAGE LIFT DENSITY
1.62
I I STANDARD DEVIATION LIMITS
I
I COMPACTION TIME - 40 minim"
1.60
2 3 4 Average 5 6 7 8 9 10 Averege
LIFT NUMBER
Fig 6.6 Lift by lift variability during in situ compaction trials on buffer.
required 0.03 Mg/m3. The procedure was shown to be repeatable and

operator independent.
6.5.9 Data collection

There are upwards of 800 instruments, in 20 different instrument types,
being logged for the Buffer/Container Experiment and the Isothermal Test.
A database has been established to facilitate the handling and presentation
of information from both experiments. All the instrumentation is logged
automatically, with the exception of pneumatic earth pressure cells and
pneumatic piezometers. The raw data are stored on a VAX computer at the
URL and are backed up daily on a separate system at Whiteshell Labora-
tories, some 15 km away. Routines have been created to facilitate visual
daily inspection of the data, as a check on instrumentation failures. After
600 days of heating, with some instruments installed for over 3 years, only
18 instruments have failed in the Buffer/Container Experiment, and an
additional four failures in the Isothermal Test.
Data collection rates have changed over the course of the Buffer/Con-
tainer Experiment to reflect the rate of change of observed transients. After
turning on the heater, temperatures were logged every 30 min for the first
month to capture the transient thermal phase. The rate of change in earth
pressure cells and moisture content sensors was anticipated to be slower.
Therefore, these instruments were logged hourly during this phase. As the
thermal transients diminished the rate of logging of all instruments was
decreased appropriately with most instruments now being logged every 6 h.
6.5.10 Thermal regime

A heater power of 1000 W was originally selected to produce a target heater
surface temperature of 85C. After 16 days, with the measured heater tem-
perature tracking about 7SC below that predicted by theoretical calcula-
tions, a back analysis was carried out and the power output was increased
to 1200 W. The major change to the input parameters used in the back
analysis was an increase in sand thermal conductivity by about 15% over
its initial value. The difference is possibly due to compression of the sand
brought about by thermal expansion of the heater upon activation. After
600 days of heating the heater skin temperature was about 85C. Figure 6.6
shows that, at 600 days, the temperatures of the buffer immediately adjacent
to the sand annulus and at the borehole wall were about 62C and 43C
respectively. Because of the spatial and temporal consistency in thermal
gradients observed (Figure 6.7) and the observed symmetry of the thermal
field, it may be assumed that no significant change in the thermal properties
of the sand, buffer and rock has occurred.
80
BUFFER END OF ROOM
t
eo
I-
:I:
(!)
[ij r ROCK
:I: 70
I
0
~
0:
W eo
l-
e(
W
:I:
I- 50
e(
D
E
w 40
0:
::>
l-
e(
0: 30
W
a.
::;
w
I- 20
10
0
DISTANCE FROM CENTER OF EMPLACEMENT HOLE

(M)
Fig 6.7 Temperatures measured in the buffer and the rock in three different direc-
tions, illustrating axial symmetry.
6.5.11 Rock hydrogeology and buffer moisture regime

After the start of the heating phase, pore pressures in the rock responded
dramatically to changes in temperature (Figure 6.8). The thermal expansiv-
ity of water is an order of magnitude greater than that of the rock, hence
expansion of the water will be sufficient to cause the measured increase. The
migration of the water into cooler rock away from the experiment has
resulted in a subsequent gradual pressure decrease. Figure 6.9 shows the
pore pressure contours surrounding the Buffer/Container Experiment and
demonstrates that the pore pressures and the direction of flow are affected
significantly by increasing temperatures.
The pore-water pressure gradients in the rock are directly related to the
rate of moisture supply to the buffer. Piezometers installed in the rock at
the mid-height of the heater were used to determine the hydraulic gradients
near the experiment (Figure 6.10). The flow towards the emplacement
borehole, as calculated using these gradients, is approximately 0.06 mL/min
into the entire emplacement borehole. It can be inferred from the data that
flow patterns are complicated (Figure 6.9). The flow of water indicated by
the contours of pore pressure implies a more variable degree of water
supply from the rock throughout the length of the emplacement hole.
It may be generally stated that water uptake was initially evident in the
buffer near the buffer-rock boundary (Figure 6.11). Furthermore, signifi-
cant increases in measured earth pressures slightly above the top of the
heater could be taken as an indication of a notable increase in moisture in
this area. Although not evident in the temperature measurements, the
ELAPSED DAYS FROM THE START OF HEATING (Nov 20,1991)
o 200 400 600

-
........ _-----
_____ ~~~_~_'--l
~~HGl1
... - - - - - - ________ 1~~_,, ___ ...,
1 RW22
-
0
1RW23
1RW23_ _ ,w _______
------ 1RW24
-
200 1RW25
Buffer/Container Experiment Borehole HG11
1RW22
o ~ J
NOV DEC JAN FEB MAR APR MAY JUN JUl AUG SEP DEC JAN FEB MAR APR MAY JUN JUl AUG SEP
1991 1992 1993
Fig 6.8 Measured pore pressure increases following heater activation.

Scale:
r""'1"iii""
Pore pressure contours in kPa and

possible flow paths ( - ) in the
Buffer/Container Experiment.
Fig 6.9 An interpretation of pore pressure contours (kPa) surrounding the

Buffer/Container Experiment and potential flow paths.
250 Pressure = 62.6 + 223.3 Ln{r} kPa
Flow =143.1 k (m /s)

3 RP5
Of200 Hydraulic conductivity (k) =1.4 x Hi'2 mls
!2
i150
too
50
or-~~-----------------------------------1
0.8 1.2 1.4 1.6 1.8 2 2.2

Radial Distance from Borehole Center (m)
Fig 6.10 Pore pressure distribution in the rock at the heater mid-height.
Fig 6.11 Moisture content changes in the buffer after 350 days (from psychrometer
and thermal needle data).
moisture sensors and earth pressure cell responses adjacent to the heater in
the buffer annulus region indicate thermal drying in the buffer annulus.
6.5.12 Buffer mechanical response

In situ compaction of the buffer resulted in 'locked-in' pressures of several
hundred kPa being recorded on both the contact and free-field earth press-
ure cells. Most of the cells show a gradual increase in pressure prior to
heater activation (Figure 6.12). Earth pressure cells adjacent to and below
the heater responded strongly to both heater activation and the power
output change to 1200 W. The cells adjacent to the heater subsequently
showed pressure decrease after the initial sharp increase. The pressure in-
creases result from thermal expansion of the heater, buffer and other com-
ponents of the experiment, whereas the decrease in total pressure is
interpreted as being associated with consolidation of a zone of saturated
buffer adjacent to the borehole wall under fixed displacement boundary
conditions. The measured total pressures below and immediately above the
heater continue to increase with time. The cells immediately adjacent to the
.~
GEOHOR CELL OATA

EARTH PRESSURE VS DATE
NOT TEJ.ftRATlRE COARECTEO
Fig 6.12 Total earth pressure response at the buffer-rock interface vs. time.
top of the heater (BG4 in Figure 6.12) have shown a particularly strong
increase with time, probably a result of thermally induced vapour transport
and subsequent swelling pressure development, as described above. The
total stresses in the annulus region continue to decrease with time, with
BG6 reading near zero pressure. These readings tend to support buffer
drying and shrinkage in the annulus region, as described above.
6.5.13 Geomechanical response

Instrumentation was installed in the rock to monitor the strains induced by
heating and by mechanical interaction with the buffer. Vertical displace-
ments in the rock surrounding the emplacement borehole are measured in
two extensometer strings. The strings are anchored 15 m into the rock and
each has eight displacement transducers measuring vertical movement of
the rock. Radial strain cells measuring horizontal stress changes were instal-
led in boreholes adjacent to the emplacement hole. These cells measure
changes in borehole diameter which can be converted into stresses using
elastic theory.
The extensometers and radial strain cells have been responding largely as
expected. The measured vertical and horizontal strains are more or less
proportional to temperature changes in the rock. The influence of the buffer
on the rock displacements appears to be negligible. Of interest, however, is

the fact that the in situ thermal expansivity of the rock appears to be greater
than the thermal expansivities measured on rock samples in the laboratory.
6.5.14 Analysis and interpretation

History-matching analyses have been carried out during the course of the
experiment to
track the progress of the responses and ensure that the experiment is
operating more or less as expected;
assist with experiment operation (e.g. provide quantitative data to change
the power output, as was required);
develop a better understanding of the system response during the experi-
ment; and
identify areas of continuing uncertainty to assist with the design of ex-
periment decommissioning.
The analyses, which examined the coupled heat-moisture and mechanical
response of the buffer, used updated and revised boundary conditions and
material properties, as appropriate. The temperature changes, moisture
changes and mechanical effects were systematically evaluated with the re-
sults of one analysis providing constraints and input on performance for the
other. For example, as the heat transfer and moisture conditions of the
buffer are intrinsically linked, the thermal response of the buffer would
provide constraints on how significantly the moisture conditions could
change to correspond to that observed thermal response. Similarly, the
mechanical response of the buffer depends intrinsically on the moisture
regime in the buffer. The Ontario Hydro Research Division coupled heat
and moisture flow code TRUCHAM (Radhakrishna and Lau, 1992) and
deformation analysis code TISDA (Lau and Radhakrishna, 1992) were used
in the analyses.
Based on history matching, the in situ thermal conductivity of the sand
infill material was increased from the initial value of 0.38 W /m;oC to
0.45W/m;oC. In addition, to more closely reflect reality, TRUCHAM was
run isothermally to account for potential water uptake during the installa-
tion phase; the resulting moisture content distribution was used as the
initial condition for the heating phase. The rock-buffer boundary was
modeled as a permeable boundary; the buffer was given free access to water,
and water could move across the boundary, depending on the driving forces
and transport coefficients. Other details of the analyses are reported in
Radhakrishna and Lau (1992). Calculated and measured temperature pro-
files through a cross-section of the buffer are shown in Figure 6.13. Tem-
peratures in the system generally match to within 1C. Calculated moisture
content variations across the buffer annulus at the mid-height of the heater
are shown in Figure 6.14. Moisture is seen to be driven radially outwards
90
Heater Buffer Rock

80
Calculated
....................... Actual
70
G 60
~
e
::l
~ 50
~
E
Q) 201 Day
I- 40
54 Day
30
-._-......
20 26 Day
B Day
10
o 500 1000 1500 2000 2500 3000
Radial Distance from Borehole Center (m)
Fig 6.13 Comparison of measured and calculated thermal gradients.
0.38 .1
0.36 ~
Rock
~.--.--.---.--.--.--.--.
C 0.34
(I)
C
o
() 0.32 ~
(I)
:;
1ii
'0 0.30
:;
.g
Q) 0.28
E
=>
g
0.26
0.24 Heater
0.22
o 50 100 150 200 250 300
Elapsed Time Since Start 01 Heating (days)
Fig 6.14 Calculated moisture content variations across the buffer annulus.
toward the rock. These results are consistent with the trends of the moisture
content sensor readings and the observed thermal response.
Using the calculated moisture distribution (Figure 6.14) and laboratory-
derived relationships between moisture content, elastic parameters, dry den-
sity and volume change (Lau and Radhakrishna, 1992), the response of the
contact earth pressure cells on the borehole wall at the heater mid-height
was modeled. The results are plotted on Figure 6.15. Two analysis cases
were carried out: one with the sand annulus assumed to be compressible,
and the other with the sand annulus assumed to be rigid. The graph indi-
cates that the model forecasts the pressure responses reasonably and, in
addition, matches the observed pressure drop near the end of the observed
time period.
The results of history tracking of the experiment demonstrated the useful-
ness of the observational approach used in this study when dealing with a
system with complex interactions, and uncertainties are involved in the
required material properties. Although the above analyses are encouraging
and experiment performance has been excellent, several uncertainties still
exist. Whereas the trends of the moisture content changes and the evolution
of the moisture regime is sensible, quantitative interpretation of moisture
content changes is extremely difficult. Interpretation of the moisture content
sensors is complicated by total stress effects from heater thermal expansion
and swelling pressure development, and by thermal expansion-induced
excess pore pressures in the buffer. For example, the psychrometer may
600
;\ ..............0 ....................... [1
500
:;\~,'f,....... .................-........... --,. . . ...-r 1BG8
400
... \1'-". . . .
-----.
/' \
'iii ..,/ ,\ "'''' .......

a.. 300 .../ / '\ I\.. '" '" _._---IeG7--r
~ .. ...-.r"''''-__'-
!!!
~
II>
II>
!!!
a..
200 1: ",'" ... Y;I"
Calculated Pressure Response
.... 0 .... incompressible sand layer
100 - - - - compressble sand layer
. - - Measured Pressure Response

o _. oj. _ _ 04 _ ~ ,", -
-100OL--L.--1-00L.----J--200...L...-...L.....--:3~00---L.--:-40~0:--......L..--'---'
Time (number of days starting from May 1.1991)
Fig 6.15 Calculated and measured total earth pressure response.

Isothermal test 117
register a change in total potential (vapour pressure) which reflects thermal-

ly induced vapor pressure changes rather than a change in local moisture
content. The uncertainty of the moisture content regime leads to difficulties
in interpretation of the total stress responses. Are the changes in total
stresses largely due to local water content increases or is there a mechanical
rebound component (release of 'locked-in' stresses) as well? Moreover, un-
certainty still exists regarding the thermal and mechanical performance of
the sand annulus material since its in situ properties were inferred from
history-matching back analyses. These uncertainties can only be clarified
through a well-designed and executed test decommissioning program.
6.6 ISOTHERMAL TEST
The Isothermal Buffer/Rock/Concrete Plug Interaction Test, as mentioned

earlier, is being conducted at the URL to provide supporting information
for the Buffer/Container Experiment. The primary objective of the Isother-
mal Test is to assess the rate at which the buffer takes up moisture from the
rock under constant temperature conditions. Installation of the Isothermal
Test was completed in November, 1992 and, at the time of writing, the test
is still in progress.
6.6.1 Experiment design

Both the Isothermal Test and the Buffer/Container Experiment were instal-
led in boreholes 5 m deep by 1.24 m diameter in granite. However, the
Isothermal Test design differs from that of the Buffer/Container Experiment
in a few respects. The lack of a heater in the Isothermal Test means there is
no internal cavity or sand installed in the test. Also, only the bottom 2 m of
the Isothermal Test borehole were filled with buffer material. This allowed
room within the emplacement borehole for a 1.25 m thick concrete plug to
resist vertical expansion of the buffer. The restraint relies primarily on the
bond strength between the rock and the concrete. However, should this
bond fail with time, restraining bars embedded in the concrete and extend-
ing outward into boreholes in the rock will provide vertical resistance,
preventing further movement of the plug.
Essentially the same types of instrumentation were installed in the Iso-
thermal Test as were used in the Buffer/Container Experiment. Data from
220 sensors are being recorded regularly for the Isothermal Test. In addi-
tion, a method for remotely assessing the buffer moisture content was im-
plemented, and 32 electrodes are installed around the circumference of the
emplacement borehole. By varying the electrodes used as the anode and
cathode and by measuring the potential at the remaining electrodes, a
tomographic image of electrical impedance within the buffer can be created
(Strobel, 1993). The relationship between the impedance and the moisture
content of the buffer has been established through a series of laboratory and
field tests.
6.6.2 Geology and hydrogeology

The granite in the vicinity of the Isothermal Test has essentially identical
thermal, mechanical and hydraulic properties to those determined for the
rock surrounding the Buffer/Container Experiment. However, the granite at
the Isothermal Test location exhibits pink alteration (hematization of feld-
spars) caused by proximity to Fracture Zone 2 (approximately 8 m below
the base of the emplacement borehole). Within the borehole the rock is
predominately medium-grained pink granite with some lenses of coarse-
grained pegmatitic granite occurring near the base of the test hole. There
are no natural fractures intersecting the borehole or within the experiment
room. Although the thickness of the zone of visible excavation damage
below the concrete floor is 0.5 m, the top of the concrete plug, and hence the
top of the experiment, is 1.5 m below the floor.
The background pore-water pressures in the vicinity of the Isothermal
Test is approximately 1600kPa, as shown in Figure 6.3. Inflows into the test
hole were monitored immediately before placement of the buffer and the
measured inflow rate was consistent with a rock hydraulic conductivity of
between 10 - 13 and 10 - 12 m/s.
6.6.3 Concrete plug

The primary purpose of the concrete plug is to act as a restraint to upward
movement of buffer resulting from swelling. The principal component of
resistance is the bond strength between the concrete and the rock. Further-
more, if required, secondary restraint is provided by 32 stainless steel bars
embedded in the concrete and extending outward into horizontal boreholes
in the rock. It is anticipated that, when saturated, at a dry density of
1.73 Mg/m 3 , the buffer will exert a maximum swelling pressure of 1.5 MPa
on the plug. Therefore, using a s~fety factor of 3, the bars were designed to
resist shear forces under this maximum swelling pressure.
The final dimensions of the concrete plug are 1.25m long and 1.24m in
diameter. The concrete was a mixture of sulphate-resisting cement, crushed
granite aggregate, silica sand, silica fume, superplasticizer and potable
water. Approximately 3800 kg of fresh material was used for constructing
the plug. Preparation of the fresh concrete was carried out underground
near the test hole. A portable gravity mixer driven by an electric motor was
used for the mixing. Results from laboratory tests show that the hardened
concrete has an unconfined compressive strength of about 100 MPa at 28
days and the maximum long-term shrinkage of the hardened concrete
should be about 0.2%.
I sothermal test 119
6.6.4 Test results
Temperature
The temperatures of the buffer, rock and concrete plug vary between 11 De
and 13e (Figure 6.16). However, short-term thermal transients were noted
in the buffer and the surrounding rock during the construction of the plug.
The increase in temperature in the buffer and the rock was attributed to the
heat of hydration effect resulting from the hardening process of the fresh
concrete used in formation of the plug. In response to the thermal effects
from the concrete, the temperature of the buffer immediately adjacent to the
plug rapidly rose from its initial temperature of 13e to a maximum tem-
perature of 30C. The excess temperature in the system dissipated within 20
days. At steady state, a temperature gradient of about OAoe/m is noted
along the length of the buffer mass, increasing towards the opening of the
test hole. Furthermore, the temperatures in the buffer and the rock appear
to be tracking the ambient underground temperature, which varies from
l20e in the winter to 16e in the summer. Generally, differences in tem-
perature between the buffer and the rock are within 1C.
Tomp$f:/ut-$ <;:cnfn:;:t0
lJ.,i/C
Fig 6.16 Thermal and pore pressure conditions surrounding the Isothermal Test.
Earth pressure
Total pressure cells were installed at the buffer-rock boundary, at the
buffer-plug boundary and within the buffer mass to track the temporal and
spatial changes in the total pressure of the buffer. 'Locked-in' pressures
attributed to the in situ compaction process, which ranged from 100 to
400 kPa, were noted in the buffer mass prior to the installation of the
concrete plug. In general, the majority of the pressure cells show systematic
increases in the total earth pressure with time. The rate of increase of
pressure varies from 50 to 100 kPa/month. Higher pressures are measured
near the base of the test hole and, at the time of writing, the total pressures
in the buffer are as high as 1000 kPa. This can be compared to the maxi-
mum swelling pressure of 1500-2000 kPa measured in water-saturated com-
pacted sand-bentonite samples using laboratory-scale rigid test apparatus
(Dixon et aI., 1986). One year after the installation of the buffer, the upward
pressure at the base of the concrete plug was between 450 and 550 kPa,
with the higher pressure measured closer to the rock-buffer interface. It can
be inferred from these observations that water uptake in the buffer zone
close to the rock has influenced the vertical stress.
Concrete performance
Four vibrating wire strain cells are used to measure the internal deforma-
tion of the plug. The cells were embedded at four different elevations in the
fresh concrete during the construction of the plug. Measurements from the
cells suggest that immediately after the placement of the fresh concrete the
plug first underwent rapid shrinkage, followed by a gradual expansion.
Most of the shrinkage strains occurred during the hardening phase of the
fresh concrete material. The amount of shrinkage strain appears to be
greatest at both ends of the plug, measuring about 1500 {leo In contrast,
smaller shrinkage strains totalling approximately 800 {le were measured for
the centre of the plug. Shrinkage of the hardened concrete appeared to be
complete 100 days after the placement of the concrete.
The magnitude of the expansion strains in the hardened concrete are
significantly less than that of the shrinkage strains, by about two orders of
magnitude. About 350 days after the installation of the plug the maximum
expansion strain measures approximately 50 {leo Axial expansion of the plug
is occurring at a steady rate of about 4 {le/month.
Eight linear displacement transducers are used to monitor the vertical
and horizontal movements of the concrete plug. The instruments were in-
stalled immediately after the concrete had hardened. Two of the transducers
are mounted horizontally to track the radial displacements while the re-
maining six transducers are mounted vertically to detect any movement of
the plug along the axis of the test hole. Two of the six vertical transducers
are mounted directly on top of the plug. The other four vertical transducers
Isothermal test 121
are used as 'telltales' and measure the differential displacements in the plug
0.25 m and 0.50 m below the surface.
Figure 6.17 shows the axial displacements of the plug with time. Positive
and negative values denote respectively upward and downward displace-
ments. The data in the figure show that shrinkage of the concrete was
measured by all the transducers during the initial phase of the test. The two
surface transducers show that significant upward displacements of the plug
began to occur approximately 80 days after the concrete had hardened. The
timing coincides with the completion of the shrinkage phase and the initi-
ation of the expansion phase of the concrete as noted by the strain cells.
Therefore these initial displacements, as measured by these two surface
transducers, are interpreted as expansion displacements of the concrete.
At about 150 days after the installation of the plug, upward displace-
ments at depths of 0.25 and 0.5 m below the plug surface initiated (Fig-
ure 6.17). The timing coincides with the development of the total pressures
on the underside of the plug, as noted by the two pressure cells located
immediately below the plug. The displacements, as determined by the tell-
tales, are interpreted as the upward movement of the plug. It can be inferred
from the similarity in the trend and magnitude of the response of the
tell-tales that little differential movement has taken place within the plug;
that is, the plug moves as a whole. About 350 days after the installation of
the plug, a maximum average upward movement of about 0.2 mm by the
plug has been noted.
0.40
Displacements at the top of the concrete
plug
0.30
E
.s
1: 0.20
CD
E
CD
0
<11
Ci
U>
'6 0.10
(ij
'x
<t
0.00
Displacements
within the plug
-0.10
0 50 100 150 200 250 300 350 400
Elapsed time in days since 4 November 1992
Fig 6.17 Measured displacement at the surface of the concrete plug and at 0.25 m
and 0.5 m below the surface.
Hydrogeology
The pore pressure contours surrounding the Isothermal Test are shown in
Figure 6.16. In general, the pore pressures as measured in the packer sealed
boreholes are constant or slightly increasing with time. Together with the
pore pressure data from the piezometers, there is a definite pressure gradi-
ent towards the emplacement hole. However, the three pneumatic piezo-
meters nearest the borehole are showing zero pressure, implying that
unsaturated or suction conditions exist within 0.5 m of the buffer-rock
interface.
The pore pressure contours surrounding the Isothermal Test imply ap-
proximately radial flow towards the test section. The pore pressure in the
rock has not been affected by changes in temperature; therefore, the inflows
calculated from measured hydraulic gradients are the same as collected
inflows prior to experiment installation. The rate of flow towards the test
section of the borehole is approximately 0.04 mL/min.
Total suction and moisture content

The total suctions in the buffer are measured by 24 psychrometers installed
at four different levels within the buffer mass. In general, the initial suctions
within the buffer varied from 3800 kPa to 4300 kPa. The psychrometers
show that moisture uptake by the buffer is immediate and that the change
in total suction with time is gradual and systematic. In general, the psych-
rometers closest to the rock show a faster rate of suction decrease than the
mner ones.
Figure 6.18 shows the total suction and the implied moisture content
profiles in the buffer 350 days after buffer installation. The profiles represent
the moisture conditions of a half-section perpendicular to the axis of the test
tunnel. The moisture contents in the buffer are determined by converting
the total suction values using the specific moisture capacity relationship
developed from laboratory tests.
The data in Figure 6.18 clearly show the systematic variations in suction
and moisture content within the buffer. As expected, the buffer closest to the
rock shows lower suctions or higher moisture contents than the buffer in
the centre. From the figure it is also evident that the largest changes in
suction and moisture content are located at the bottom corner of the test
hole, as a result of increased moisture availability from the rock in both the
vertical and horizontal directions.
6.6.5 Analysis and interpretation

The Isothermal Test is progressing systematically. In general, good correla-
tion of the data is achieved between the different types of instruments
installed in the test.
Conclusions 123
Rock
Concrete Plug
High Suction
As-Compacted
Moisture Content
Low Suction Near Saturation
Total Suction Moisture Content
Fig 6.18 Total suction and moisture content profiles in the buffer of the Isothermal
Test.
The increase in the total earth pressure and the decrease in the total
suction in the buffer implies that the buffer is gradually taking up moisture.
Zero or negative pore-water pressures still persist in the rock immediately
next to the borehole. This therefore implies that moisture movement within
the buffer is likely to be dominated by vapour flow.
Calculations using finite element (Thomas et al., 1993) and finite difference
techniques were carried out to provide an understanding of the time required
for saturation of the buffer under the in situ hydraulic boundary conditions.
These calculations indicate that the time for saturation varies from 3 years to
several hundred years. The accuracy of these calculations is greatly influenced
by the assumptions and material parameters adopted in the analyses.
6.7 CONCLUSIONS
The physical performance of a clay-based barrier may be assessed through

both laboratory and in situ studies. The laboratory studies allow the
performance to be assessed under various boundary conditions and at rela-

tively modest cost, but the complex interactions and effects of the in situ
boundary conditions for the prototype can neither be replicated nor ob-
served in the simplified laboratory experiments.
This chapter has described two in situ experiments designed to examine
the performance of a bentonite-based barrier material for the purpose of
nuclear fuel waste containment. In the Buffer/Container Experiment, after
600 days of heating the thermal regime in the buffer has essentially reached
steady state and shows symmetry. Measurements by the total pressure cells
and the moisture sensors imply that moisture is being driven away from the
annulus zone of the buffer and is being redistributed in the buffer above and
below the heater. The consistency in the measured temperature gradients
with time implies that the redistribution of moisture within the buffer does
not significantly influence the thermal properties of the material.
The Isothermal Test is progressing systematically. Measurements by the
total pressure cells and the psychrometers correlate well and imply that the
buffer is gradually taking up moisture from the host rock.
The in situ experiments carried out on the clay-based buffer material have
demonstrated the ability to characterize and monitor the hydraulic re-
sponse of near-field intact, very low permeability rock;
highlighted the importance of the sand annulus between the heater (con-
tainer) and the buffer in the thermal, hydraulic and mechanical responses
of the system;
highlighted the potential importance of thermal expansion effects, which
have had a dominating influence on the response of the Buffer/Container
Experiment; and
demonstrated the feasibility of technologies such as in situ buffer compac-
tion and monitoring instrumentation and techniques.
In terms of future repository monitoring, these in situ experiments have
demonstrated that system temperatures may be measured reliably over ex-
tended periods of time, but that the hydraulic and mechanical response of
the buffer is significantly more difficult to measure and interpret. The ex-
periments have demonstrated, however, that the response of one instrument
type can be used to corroborate and support the response of another, as
was shown with the moisture content sensors and earth pressure cells. The
importance of accurate assessment of the time-dependent moisture response
of the buffer in the repository needs to be weighed against the measurement
and interpretation difficulties. We should only place reliance on perform-
ance parameters which can be measured reliably. In this study we have
demonstrated the important role that the observational approach and back
analyses play in understanding the complex interactions between the con-
tainer, sand, buffer and rock during transient water uptake by the buffer.
The modeling, however, requires further development to permit the study
of thermo-hydraulic-mechanical interaction of the buffer with other
Acknowledgements 125
components in the repository in terms of suction potentials rather than

moisture gradients.
ACKNOWLEDGEMENTS
The work described in this chapater is part of the Canadian Nuclear Fuel
Waste Management Program which is funded jointly by AECL Research
and Ontario Hydro under the auspices of the CANDU Owners Group.
CHAPTER SEVEN
In situ Hydraulic
Performance Tests of
Borehole Seals: Procedures
and Analyses
W B. Greer
7.1 INTRODUCTION
The performance of seals in boreholes, shafts and tunnels is an important

concern in mine safety, in the control of water in operating and abandoned
mines and in the underground containment of hazardous wastes (Einarson
and Abel, 1990; Mining Waste Study Team, 1988). Despite the clear need
for effective seals, few documented test results are available concerning the
performance of in situ seals. Not only are test results lacking, but methods
for conducting such tests are not well developed. This chapter presents four
in situ tests of the hydraulic performance of borehole seals: the steady
constant-head, the transient constant-head, the head buildup and the tracer
travel-time tests. Each test is described and the mathematical models used
for analysis are presented and discussed. The tests and analysis methods are
applicable to seals and the surrounding rock, which may be treated as
homogeneous, isotropic, saturated porous media and permit fluid flow in
accord with the equation of groundwater flow. The methods are applicable
to tests using water, or other slightly compressible fluid, as the permeant.
With some modifications (not presented here), the test and analysis tech-
niques may be applied using gas. The study considers only the bulk motion
of water through the seal and adjacent rock mass. The transport of solutes
via diffusion or hydrodynamic dispersion and chemical interaction of fluid,

Proposed test schemes 127
seal and rock are not considered. The mechanical interactions of fluid, seal
and rock are not considered, except in the very limited sense implied by the
use of the equation of groundwater flow. The study focuses on the testing of
seals in open, or uncased, boreholes or borehole sections. The seals con-
sidered are 'parallel' seals; that is, they are in the shape of a right cylinder
with a diameter equal to that of the borehole. Field experience with the tests
is presented in Chapter 8.
7.2 PROPOSED TEST SCHEMES
7.2.1 Approach to testing
The test methods are intended for seals with a length-to-diameter ratio less
than about five. While borehole seals often greatly exceed this ratio, short
seals are necessary for testing in order to obtain measurable results in
practical test times and to approximate the conditions assumed in models
used for analysis. Further, in at least some hydraulic tests involving axial
flow, the flow properties of portions of the seal beyond about five diameters
from the point of flow initiation are not reflected in tests (Greer and
Daemen, 1991).
Two basic test configurations are used: one-side and two-side. The one-
side configuration (Figure 7.1) is used when only one side of the seal is
accessible. In this configuration the hydraulic head on the inaccessible side
of the seal is assumed to be constant and equal to He. An isolated, water-
filled interval, the injection zone, is created on the accessible end of the seal,
usually by installation of a pneumatic or mechanical packer. Water may
be delivered at constant pressure to the injection zone to maintain a head
H (H > HJ in the interval. The flow rate to the injection zone, or injection
rate, Qi' may be monitored as well as the pressure in the injection zone.
The injection zone may be 'shut in' by closing a valve on the water-delivery
line.
The two-side configuration, used when there is access to both ends of the
seal, is shown in Figure 7.2. Isolated, water-filled intervals are created on
both sides of the seal, again, usually by packers. One of the isolated inter-
vals is an injection zone with head equal to H. The other interval, the
collection zone, may be maintained at constant head He (He < H) by allow-
ing the interval to drain freely. The outflow rate from the collection zone, or
collection rate, Qe' may be monitored, along with the pressure in the collec-
tion zone. A valve on the outflow line may be closed to shut in the collec-
tion zone. The two test configurations allow the seal to be subjected to a
variety of hydraulic loadings, the responses to which may be modeled in
terms of the hydraulic characteristics of the seal or of the seal and rock
mass.
128 In situ hydraulic performance test of borehole seals
fV'.l.._ _ Shut-in valve
Packer
Head = H--==---,
Grout seal
Head - H, _-==-""1
(H,<H)
Fig_ 7_1 One-side configuration for testing seal.
7.2.2 Approach to analysis

In analyzing a flow test of a seal, the simplest approach is to assume one-
dimensional axial flow through the seal. With such an approach, the analy-
sis is closed-form and only the hydraulic characteristics of the seal are
involved. In this study one-dimensional analytical flow models are pres-
ented for the analysis of steady and transient seal tests. For some test
conditions the approach provides reasonable estimates or upper-limit
values of the hydraulic properties of the seal. However, the seal is not
isolated hydraulically from the rock mass. Such isolation is probably not
possible if flow through the interface between rock and seal is to be in-
cluded. For tests in which flow in the rock mass is significant, a more
realistic analysis accounts for flow components in both the seal (i.e. seal
core and seal-rock interface) and the rock mass (i.e. damaged zone and
undamaged rock mass). In this study axisymmetric three-dimensional flow
models are used in the analysis of some tests. For these models the flow
domain includes both the seal and the surrounding rock mass. The rock
mass is taken as the annular volume between two concentric cylinders. The
inner cylinder has a diameter equal to that of the borehole. The diameter of
the outer cylinder is finite but large with respect to seal dimensions. The
axial dimensions of the concentric cylinders are also finite but large. The
Proposed test schemes 129
0 0 - - Shut-in valve
Packer
Head = H --t:::==--'
Grout seal
Head = He ---==--.
(H,<H)
Packer
[x-l--Shut-in valve
~
Q, = collection rate
Fig_ 7.2 Two-side configuration for testing seal.
axial and radial boundaries are surfaces of constant and uniform head. The
axisymmetric analyses, which are accomplished using a suitable computer
program for groundwater flow, yield hydraulic properties of both the seal
and rock mass.
7.2.3 Hydraulic characteristics of seals

The hydraulic characteristics of the seal are expressed by the hydraulic
conductivity and specific storage of the seal, Ks and Sss, respectively. The
conductivity and specific storage of the rock mass, Kr and Ssr' express the
characteristics of the rock mass. Hydraulic conductivity, K, for an isotropic

medium may be defined by the following expression of Darcy's law:
ij
K=-- (7.1)
Vh'
where h is the hydraulic head [L], Vh is the hydraulic gradient, and q is the
specific discharge vector [LIT] or volumetric flow rate through a unit
cross-sectional area normal to the direction of the hydraulic gradient. Hy-
draulic conductivity has units of [LIT]. It depends upon properties of both
the fluid (i.e. water) and the medium and may be expressed as
K = kpgl/1, (7.2)
where p is the density [MIL 3] and /1 is the dynamic viscosity [M/(LT)] of

water, g is acceleration due to gravity [LjT2], and k is the intrinsic permea-
bility (or permeability) of the porous medium [L 2] (Bear, 1972). Permeabil-
ity is generally considered to depend only on the medium and may be used
as an alternative to hydraulic conductivity to express the transmissive char-
acter of the porous medium.
The specific storage [L -1] of a porous medium is the volume of water
released from storage per unit volume of porous medium per unit decline in
hydraulic head. The specific storage accounts for the release of water from
storage, due to the expansion of water as the fluid is depressurized and due
to the compression of the medium. Constitutive expressions for specific
storage have been derived by Jacob (1950), De Wiest (1966), Brace, Walsh
and Frangos (1968), and van der Kamp and Gale (1983). The expressions
vary somewhat in form, depending on the assumptions used in the
derivation.
7.2.4 Determination of hydraulic properties

Obtaining hydraulic properties from seal tests using the one-dimensional
and axisymmetric models is an inverse procedure. Hydraulic properties
obtained from analyses with the one-dimensional models and the steady
two-side axisymmetric model are clearly unique. Properties obtained from
analyses with the other axisymmetric models are likely not to be unique.
Hydraulic properties obtained from all of these analyses are 'model' proper-
ties; that is, they are values which cause the model, with its inherent limita-
tions, to reproduce, as nearly as possible, the response to the test observed
in the field. Analysis of the same seal test with different models may yield
somewhat different model properties, all of which may differ to some extent
from actual in situ properties.
In analyses involving unsteady flow, the technique of curve-matching is
utilized to obtain hydraulic properties. The technique was developed by
Theis (1935). Variations of the method have been applied by others, inclu-
Seal tests 131
ding Walton (1962) and Neuzil et al. (1981). A detailed explanation of the
method is provided by Wenzel (1942).
7.3 SEAL TESTS
7.3.1 Steady constant-head test

This test may be performed using either the one-side or two-side configur-
ation (Figures 7.1 and 7.2). For the two-side configuration, water is de-
livered to the injection zone at constant head, H. Water outflow from the
seal and rock mass is collected in the collection zone, which is maintained at
a head equal to Hc(Hc < H). Flow conditions are assumed to be steady. The
steady injection rate, Qi' and the rate of flow out of the collection zone, Qc'
are measured. In the one-side configuration there is no collection zone and
only the injection rate, Qi' is measured. For both configurations the ambient
head in the rock mass (Hb) is also measured, if possible. Analysis in terms of
the one-dimensional model yields the hydraulic conductivity of the seal.
Analysis with the axisymmetric model will usually yield the hydraulic con-
ductivities of both the seal and rock mass.
Related Tests
The test is very similar to the constant-head permeameter test for determin-
ing the hydraulic conductivity of a soil or rock specimen in the laboratory
(Lambe and Whitman, 1969; Bear, 1972). Numerous workers have perfor-
med constant-head permeameter tests on intact, low-permeability rock sam-
ples with water as the permeant (e.g. White et al., 1979; Trimmer et al., 1980;
McDaniel, 1980; Jones and Owens, 1980; Bernabe, Brace and Evans, 1982;
South and Daemen, 1986).
The steady constant-head injection test is a field procedure which bears
similarity to the steady constant-head seal test. In the injection test an
isolated interval in a borehole is established between two packers or be-
tween a packer and the bottom of the hole. The rock mass is assumed to be
saturated. Water is injected into the interval under constant pressure until
approximately steady flow is achieved. A number of analytically derived
expressions and expressions based on flow-net, analog or numerical studies
have been developed to determine the hydraulic conductivity of the rock
mass using the injection head (i.e. head in excess of ambient head), steady
flow rate and geometry of the isolated interval. The most widely used
expressions are those presented by Hvorslev (1951) and one based on the
Thiem equation for steady, radial flow to a well (Bear, 1979). Workers who
have performed steady constant-head injection tests in low-permeability
rock include Davison (1980), Haimson and Doe (1983), Andersson and
Persson (1985) and Spane and Thorne (1985).
One-dimensional analysis
In the one-dimensional model, flow is assumed to be steady and axial
through a porous-medium seal. For this model, application of Darcy's law
gives the following upper limit expression for the hydraulic conductivity
(K.) of the seal:
K = Qld L s (7.3)
s A(H-He) ,
where Qld is the steady flow rate through the seal [L 3 IT], H is the constant
injection head [L], He is the head in the collection zone [L], and Ls and A
are the length [L] and cross-sectional area [L 2], respectively, of the seal.
Estimates of the hydraulic conductivity of the seal may be obtained for
the two-side test (i.e. a test using the two-side configuration) by substitution
of Qi or Qe for Qld in Equation (7.3). For the one-side test Qi may be
substituted for Qld to obtain an estimated Ks. If flow is very nearly one-
dimensional through the seal, Qi and Qe are approximately equal and yield
about the same value of Ks. In general, the injection flow Qi includes
significant components into the rock mass as well as axial components into
the seal. Similarly, Qe usually includes components from both rock mass
and seal. Assuming that He = 0 and that at some axial and radial distances
from the seal the hydraulic head in the rock mass is constant and equal to
zero (i.e. Hb = 0), the following inequality must hold:
(7.4)
If Equation (7.4) holds, the values of Ks obtained using Qi or Qe in Equation
(7.3) are upper limit values. The value obtained with Qe is generally a
smaller upper limit than that obtained with Qi. However, if the head in the
rock mass (i.e. H b) is sufficiently greater than zero, the injection rate, Qi'
may be less than Qe. In this case the conductivity based on the injection rate
is a smaller upper limit than the value based on Qe. If the rock mass head is
yet greater, Qi may be less than Qld' making the hydraulic conductivity
based on the injection rate a lower limit on the actual conductivity of the
seal. For all values of head in the rock mass greater than or equal to zero,
Qe is greater than Qld' provided He = O. Thus conductivity based on the
collection rate is always an upper limit for K s' so long as H e ~ 0 and H b ~ O.
These relationships are examined in detail in Greer and Daemen (1991).
Axisymmetric analysis
When analyzed in terms of the axisymmetric model, the two-side test yields
the hydraulic conductivity of both the seal and the rock mass. On the other
hand, the axisymmetric analysis of a one-side test yields an infinite set of
pairs of seal and rock conductivities, anyone of which, when used in the
model, reproduces an observed injection flow rate. The unique solution pair
Seal tests 133
of conductivities for a given one-side test can be selected from the infinite
set if the conductivity of the rock mass is known by an independent means.
The flow domain for the axisymmetric model of the two-side test is
shown in Figure 7.3 and Table 7.1. Using dimensional analysis, the follow-
ing relationship may be shown to hold for this model:
(7.5)
where D = diameter of the borehole or seal [L], Kr = hydraulic conductivity

of the rock mass [LIT] and other quantities are as previously defined.
To analyze the two-side steady constant-head test, the following steps are
taken.
1. Perform a series of steady-state computer simulations of the test. In the

simulation runs use a range of Krl Ks values which brackets the actual
ratio of conductivities. For the runs the flow field dimensions and the
boundary heads, H, Hb and He are known. An arbitrary value of Ks is
r1 r2
K. Kr r5
ra
r7 rs
Fig.7.3 Flow field for axisymmetric analysis of steady constant-head seal test.
Table 7.1 Boundary conditions for flow field in Figure 7.3
Boundary Physical description Hydraulic condition
r1 Injection face of seal h = injection head = H

r2 Borehole wall in h = injection head = H
injection zone
r3 Gland of packer 8h/8r= 0
r4 Axial boundary of h=Hb
rock mass
rs Radial boundary of h=H b
rock mass
r6 Borehole wall in h=Hc
collection zone
r7 Collection face of seal h=Hc
rs Center line of borehole 8h/8r = 0
assumed and used throughout all the simulations. The only varying
quantity in the runs is K r The output of each run includes Qi and Qc.
2. Using the results of each run, calculate KsHD/Qc' KrHD/Qc and QjQc.
These are the first three dimensionless products in Equation (7.5). The
fourth and fifth products are known and are constant for all simulations.
3. Make semilog plots of KsHD/Qc vs QjQc and KrHD/Qc vs QjQc using
the values calculated in step 2. Plot Qj Qc on the arithmetic scale.
4. Calculate Qj Qc using the approximately steady injection and collection
rates achieved in the test. Using this QjQc, find the corresponding
KsHD/Qc and KrHD/Qc from the plots in step 3.
5. Using the values computed in step 4 and the known H, D and Qc values
for the test, calculate Ks and K r These are the 'solution' conductivity
values for the axisymmetric model. With these values of Ks and K r , the
model yields the observed injection and collection rates for the test.
If Hb is unknown, analysis may be made for K" Kr and H b, provided results

are obtained for tests at two or more injection heads, H. A solution for this
case is presented in Chapter 8.
Axisymmetric analysis of one-side test

For a one-side constant-head test, the relationship between the variables
may be expressed as
K,HD KrHD =f (Kr Hb) (7.6)

Q1. 'Q.1 1 K'
,
H .
A procedure for analysis consists of these steps:

Seal tests 135
1. Do step 1 of the procedure for the two-side test. Ignore the references to
He and Qe'
2. Using the results of each run, calculate KsHD/Qi and KrHD/Qi'
3. Using the values computed in step 2 and the known values of H, D and
Qi from the test, calculate a pair of Kg and Kr values for each run.
4. Use the pairs of Ks and Kr values found in step 3 to construct a log-log
plot of Ks vs K r. This plot is the locus of 'solution' conductivity values for
the test. Any pair of Ks and Kr values from the plot will cause the
axisymmetric model to yield the Qi value observed in the test.
5. If Kr is known by a means independent of the test, it may be used to
obtain Ks from the plot in step 4.
7.3.2 Transient constant-head test

This test may be performed with either the one-side or two-side configur-
ation. For both configurations, head prior to the test is assumed to vary
linearly from H 1 (H 1 > 0) at the injection side of the seal to a head defined
to be zero at the collection side. For the two-side configuration, head in the
collection zone is kept equal to zero throughout the test. At the start of the
test the injection head is changed to positive H 2(H 2 not equal to H 1) and
maintained at that level throughout the test. For the two-side configuration
the test consists of recording the transient injection rate, Qi and transient
collection rate, Qe' until steady state is approached. For the one-side con-
figuration the test is similar, except that there is no isolated collection zone
and only the transient injection rate is monitored. One-dimensional analysis
of the transient flow rate(s) gives the hydraulic diffusivity (Ks/ Sss) of the seal.
If Ks is obtained using the steady flow data from either before or after the
test, Sss' the specific storage of the seal, may be obtained from the diffusivity.
Axisymmetric analysis may yield Kr/Sss (or K,/Sss) and Ssr/Sss' If Ks and Kr
are determined by axisymmetric analysis of the steady flows either preced-
ing or following the test, K s' Sss' Kr and Ssr may be determined.
Related methods
The test has been applied by AI-Dhahir and Tan (1968) to obtain hydraulic
properties of laboratory samples. In their laboratory test, H 1 is assumed to
be zero and generally only the transient injection flow is monitored. The
analysis yields the hydraulic conductivity of the sample from the final
steady-state flow rate achieved in the test. The hydraulic diffusivity, and
thence the specific storage, may be obtained from the slope of a plot of
inflow rate vs (time)-o.s. AI-Dhahir and Tan applied their technique to a
clay sample in an oedometer consolidation test at constant total stress. The
test yielded the conductivity and coefficient of consolidation (a form of the
diffusivity) for the specimen. The values compared closely to results ob-
tained by other methods on similar samples.
The transient constant-head injection test, which is sometimes used by
hydrogeologists to determine formation hydraulic properties, has much
similarity to the one-side transient constant-head seal test. The set-up and
conduct of the transient injection test are the same as for the steady injec-
tion test except that the transient rate of injection is monitored. Van
Everdingen and Hurst (1949) and Jacob and Lohman (1952) independently
derived an analytical solution for the transient injection test for purely
radial flow. Hydraulic conductivity and specific storage of the formation
may be obtained using the transient flow data, either by matching against a
dimensionless type curve or by a straight-line graphical method. Hantush
(1959) extended this work to include constant-head radial flow in a cylindri-
cal formation with vertical leakage and constant head equal to zero at the
outer boundary; radial flow in a closed, cylindrical formation with vertical
leakage; and radial flow in an infinite formation with vertical leakage.
Gibson (1963) obtained the analytical solution for the case of spherical flow
into an infinite formation with head initially uniformly equal to zero.
Wilkinson (1968) discussed the use of Gibson's solution to obtain hydraulic
properties. Ferris et al. (1962) presented the analytical solution for the tran-
sient constant-head test for the semi-infinite case of one-dimensional flow.
For this test, initial head is assumed uniformly equal to zero. When based
strictly on the transient flow data, the test yields only the product of hy-
draulic conductivity and specific storage.
Spane and Thorne (1985) used the solution of Jacob and Lohman (1952)
in their analysis of multiple transient constant-head injection tests in a
single interval of low-permeability basalt. The conductivity values they ob-
tained agreed closely with values determined by other transient and steady-
state test methods. They did not report values of specific storage determined
from the tests. Doe et al. (1982) utilized the transient constant-head test to
determine the hydraulic properties of individual fractures. They assumed all
observed flow to be flow through the tested fracture. They used the solution
of Jacob and Lohman and the extensions of Hantush to analyze infinite and
finite fractures and a fracture intercepting a source of leakage.
The one-dimensional transient injection and collection rates are given by
the following dimensionless expressions:
(7.7)
(Qo - QJH 2 2~
L. [ cos (nn
) exp ( - n2 n 2 tds)] ,
( _ HI ) =
QoH2 -
n=l
(7.8)
Seal tests 137
where
(7.9)
and
(7.10)
The quantity tds is a dimensionless expression for time in terms of the

hydraulic properties and length of the seal. Qo is the steady-state flow rate
achieved at the end of the test. Equations (7.7)-(7.10) are derived in Greer
and Daemen (1991), based on a general analytical solution presented in
Carslaw and Jaeger (1959).
Semilog 'type curve' plots of Equations (7.7) and (7.8) are presented in
Figures 7.4a and 7.4b. Values used for the plots are shown in Appendix A of
Greer and Daemen (1991). The hydraulic diffusivity may be obtained using
either the injection type curve (Figure 7.4a) and a plot based on the transi-
ent injection data or the collection type curve (Figure 7.4b) and a plot based
on the transient collection rate data.
Solution procedure
The following curve-matching procedure may be used for the analysis.
1. Calculate Ks using Equation (7.10). Alternatively, Ks may be calculated
using a similar expression in terms of H 1 and the steady flow rate prior
to the test.
2. Prepare semilog plots of (Qi - Qo)H 2/(Qo(H 2 - H 1)) and/or (Qo - Qc)
H 2/(Qo(H 2 -H 1 )) vs. tds using Equations (7.7) and (7.8). Dimensionless
time tds is plotted on the logarithmic axis. These plots are, respectively,
the injection and collection type curves.
3. Using the data from the test, make semilog plots of ((Qi - Qo)H 2/
(Qo(H 2 -H 1 )) and/or (Qo-Qc)H 2 /(Qo(H 2 -H 1 )) vs. t. These plots are
the injection and collection data curves. Plot the data curves on paper
having scales of the same size as those used for the type curves with time,
t, on the logarithmic axis.
4. Superimpose the data curve on the corresponding type curve. Keeping
the logarithmic axes (i.e. axes for tds and t) collinear, move the data curve
over the type curve until the data curve coincides as well as possible with
the type curve.
5. Select an arbitrary 'match' point from the overlapping portions of the
curves. Obtain tds and t for the match point.
6. Calculate the hydraulic diffusivity (Ks/ Sss) using Equation (7.9) and the
match-point values of tds and t. The specific storage may be obtained
from the diffusivity since Ks is known from step 1.
10.0
(s)
~ 5.0
~
I
g
1.0
(b)
0.0 L:---L..--1....L...L..J....LJ..I..I......:,---1.---L....J......L~J.L..,,.-I--I.-L.:lI:.J...J..L.1J
10 4 10~ 1~
fm,= K.t/( SssL.2)
Fig.7.4 (a) Injection type curve for one-dimensional model of transient constant-
head test (Equation (7.7)). (b) Collection type curve for one-dimensional model of
transient constant-head test (Equation (7.8)).
The graphical method of AI-Dhahir and Tan (1968), which uses the steady-
state flow rate and the transient injection flow data, may be used to obtain
the hydraulic properties for the case in which the head is initially uniformly
zero.
The flow field for the axisymmetric model is the same as for the axisymmet-
ric analysis of the steady test (Figure 7.3 and Table 7.1), except that the head
in the injection zone is H 2 and the specific storages of the seal (Sssl and rock
mass (Ssr) must be included. For the axisymmetric model of the transient
test, the collection zone head is set equal to He. The following relationship
between the variables of the axisymmetric model is obtained by dimensional
Seal tests 139
analysis and is the basis for the solution procedure which is presented:
Qi Qc _ (Hb H cHI Kr t5 t ) (7.11)

H D' K r2
K r2 H D - f2 H 2
'H2'H
2 'K s ' , dr ,
where
(7.12)
and
(7.13)
The solution procedure consists of the following steps.
1. Solve for Ks and Kr using the steady-state injection and collection rates
at the beginning and/or end of the test and the axisymmetric procedure
for the steady constant-head test.
2. Perform a series of computer simulations of the transient test using a
range of t5 values which brackets the actual ratio of specific storages for
the field test. In the simulations use the field values for heads H b' H c and
H 2 and the hydraulic conductivities, Ks and K r, obtained in step 1.
Assume an arbitrary value of Sss and use it in all simulation runs. To
obtain the initial head distribution for the transient simulations, run the
model to steady state using HI as the injection head. The steady-state
heads obtained from the run are the initial heads for the flow field. In the
transient simulations the only varying input quantity is Ssr' Output from
each run includes Qi and Qc as functions of time.
3. Using the results of step 2, plot a family of log-log type curves of
Qj(KrH 2D) and Qc/(KrH 2D) vs. t dr for the range of t5 values. Plot all of
these curves on one graph.
4. Make a log-log plot of Qj(K rH 2 D) and Qc/(KrH2D) vs. time, using the
injection and collection rates and Hand D from the test and using Kr
obtained from step 1. Make the plot on transparent graph paper using
the same size scales as for the type curves.
5. Superimpose the data curves (from step 4) on the type curves. Keeping
the time and t dr axes collinear, slide the data curves over the type curves
to obtain a 'match'. A match is achieved when the points of the data
curves are lined up, as closely as possible, with the best-fitting injection
and collection type curves.
6. Once the match is obtained, select an arbitrary match point from the
overlapping portions of the graphs. For the match point, note the values
of time and t dr and the value of t5 for the matched type curve. Using these
values, calculate Ssr with Equation (7.12) and Sss with Equation (7.13).
If Ssr is known by a means independent of the test, the need to compare the
data curve to each of the type curves in the matching process is eliminated.
For this case, the procedure is as follows.
1. Do steps 1-3 of the preceding procedure.
2. Transform the observed Qi' Qc and time data from the test to values of
QJ(KrH 2D), Qc/(KrH 2D) and t dr , respectively. Plot the transformed data
directly on the type curves. Interpolate to determine the value of b for the
type curve which matches the transformed data curve.
3. Use the value of b found in step 2 and equation (7.13) to calculate Sss.
For the general case (i.e. K s' Sss' Kr and Ssr are unknown), the axisymmetric
procedure may be performed using tds instead of t dr . If performed with t ds '
analysis gives Sss in step 6 using Equation (7.9) and Ssr using Equation
(7.13).
7.3.3 Head-buildup test

Two variations of the head-buildup test are presented: recovery and
impulse. Both variations may be performed with either the one-side or two-
side configuration. In the recovery variation, at the start of the test, hydrau-
lic head is assumed to vary linearly from zero (or the head is defined to be
zero) in the collection zone and at the collection face of the seal to positive
H at the injection face. The injection zone is maintained at constant head,
H, throughout the test. The test is begun by 'shutting in' (i.e. closing off the
outflow from) the water-filled collection zone. The test consists of recording
the buildup in head in the collection zone as a function of time. The impulse
variation is similar, except that head is assumed to equal H throughout the
length of the seal initially. In general, a one-dimensional analysis of the test
yields seal properties Ks and Sss. Axisymmetric analysis may yield Ks and
Sss and also rock properties Kr and Ssr. The first head-buildup test on an
in situ seal known to the author was conducted by Sandia National Lab-
oratories in 1979-1980 as part of what became known as the Bell Canyon
Test. The test was performed on a borehole seal subjected to high artesian
pressures. The analysis of that test was one-dimensional. Specific storage of
the seal was assumed equal to zero (Christensen and Peterson, 1981).
Related laboratory methods

A procedure which is very similar to the head-buildup test is the transient
pulse test for the determination of hydraulic properties of laboratory sam-
ples. A number of variations of the transient pulse test are used. In the most
general form of the test, one end of a saturated sample is in hydraulic
contact with a fluid-filled, pressurized reservoir called the upstream reser-
voir. The other end of the sample is in contact with a similar reservoir, the
downstream reservoir. Before the test, the hydraulic head in both reservoirs
Seal tests 141
and in the sample is uniform. To begin the test a small volume of fluid is
suddenly injected into the upstream reservoir, causing a small rise in head.
The test consists of monitoring the decline of head in the upstream reservoir
and the buildup in head in the downstream reservoir as fluid flows through
the sample. In general, analysis of the transient head data yields the hydrau-
lic conductivity and specific storage of the specimen. Most variations of the
test arise from making one of the reservoirs relatively very large so that it
remains at essentially constant head throughout the test. Both water and
high-pressure gas have been used as the permeant in published studies.
Hsieh et al. (1981) derived a general analytical solution for the transient
pulse test. Neuzil et al. (1981) described practical aspects of the solution
technique. In the latter paper the general solution is applied to the analysis
of tests on two specimens of shale. A summary of the historical development
of the transient pulse test and its application by numerous workers to
low-permeability samples is presented in Greer and Daemen (1991).
Related field methods

The well-testing procedure of hydrogeology closest to the recovery vari-
ation of the head-buildup test is the Theis recovery test (Theis, 1935). The
test is conducted following a period of pumping at a constant rate in a well
drawing from the full depth of a confined aquifer. After pumping is termin-
ated the recovery of the water level in the well is recorded as a function of
time. Analysis of the recovery data yields the hydraulic conductivity of the
formation near the well. In petroleum engineering the pressure-buildup test
following constant-rate pumping, as proposed by Horner (1951), is essen-
tially the same as the recovery test. Petroleum engineers have also develop-
ed analytical solutions for pressure buildup following pumping at constant
head (Ehlig-Economides and Ramey, 1979).
The pulse test of hydrogeology is a procedure with much similarity to the
impulse variation of the head-buildup test. The pulse test is performed in an
isolated, water-filled section of a borehole after it has equilibrated hydrauli-
cally with the surrounding formation. At the start of the test a volume of
water is suddenly injected into (or removed from) the interval, which pro-
duces a sudden increase (or decrease) in head in the interval. The test
consists of recording the decay (or buildup) in head within the interval as a
function of time. The test yields the hydraulic conductivity and specific
storage of the interval. The analytical solution to the initial boundary value
problem describing the test for radial flow to/from a well of finite radius
was presented by Cooper, Bredehoeft and Papadopulos (1967) and
Papadopulos, Bredehoeft and Cooper (1973). The original application of the
method was to borehole intervals which are open to the atmosphere. In
such intervals recovery of head occurs slowly as water flows into the inter-
val and rises in the open borehole or standpipe above the interval. Such
recovery may require excessive time in low-permeability formations.
Bredehoeft and Papadopulos (1980) were the first to apply the method to
fully isolated, water-filled intervals. In such intervals recovery is accom-
plished by the compression of the water in the interval and thus occurs
much more rapidly, making the method suitable for low-permeability for-
mations. A summary of field applications of the pulse test is presented in
Greer and Daemen (1991).
Recovery variation
The recovery variation of the head-buildup test is performed after a long
period of constant-head testing, at which point conditions of steady flow
and linear variation in head across the seal are assumed established. The
groundwater flow equation and the following boundary and initial condi-
tions describe the one-dimensional model of the test (Figure 7.5).
Boundary conditions:
h(Ls' t) = H for t~0 (7.14)
h(O, t) = hc(t) for t~0 (7.15)
~ dh c _ 8hl -0 for t>O (7.16)

KsA dt 8z z = 0
Fig.7.S Flow field of one-dimensional model of head-buildup test. Initial condi-

tions depend on variation of test used.
Seal tests 143
Initial conditions:
zH
h(z,O) = - for 0 < z < Ls (7.17)
Ls
(7.18)
where he is the head in the collection zone [L] and Se is the compressive
storage of the collection zone [L 2]. The compressive storage is defined as
the volume of water added to the collection zone per unit increase in
hydraulic head in the zone. Se is determined experimentally and apart from
the test.
The general solution to the initial boundary value problem of the one-
dimensional model was obtained by Daemen et al. (1986, Appendix A). The
solution is as follows:
00
h(z, t) = H + 2H L
n=l
where the quantities <Pn are the roots of the equation

S AL
tan <P = _s_s _s . (7.20)
Se<P
Evaluating Equation (7.19) at z = 0, dividing the numerator and denomina-
tor of the series term by sin<Pn' and rearranging gives this expression for
head-buildup in the collection zone:
he = 1- 2 I: [exp( - <P~tds)
] (7.21)
H n=l ",2 + <p~+ <p! '
'f'n e e 2
where
e= SssALso (7.22)
Se
e
The dimensionless parameter is equal to the ratio of the compressive
storage of the seal (Sss x seal volume) to that of the collection zone. Equa-
tion (7.20), whose roots are the values <Pn' may be expressed in terms of eas
follows:
<ptan<p = e (7.23)
Impulse Variation
In the impulse variation of the head-buildup test, the head in the seal is
assumed to be uniform and initially equal to H. The initial boundary value
problem describing the impulse variation is the same as for the recovery
variation except that the initial condition is changed to
h(z, 0) = H for 0 < z < Ls (7.24)
Greer and Daemen (1991) obtained the solution for the impulse variation by
applying superposition to a special case of the general solution of Hsieh et al.
(1981) for the transient pulse test. The solution is given by Equation (7.25):
(7.25)
where n and ~ are as defined for Equation (7.21). Further discussion of the
impulse, recovery and other variations is presented in Greer and Daemen
(1991). That presentation includes, for each variation, solutions for the limit-
ing cases where ~ approaches zero or infinity.
Type curves of the analytical solutions

To analyze a head-buildup test (either recovery or impulse variation), a plot
of the observed buildup data is matched against a family of dimensionless
type curves of the analytical solution. Type curves for the recovery variation
are presented in Figures 7.6 and 7.7. In Figure 7.6 dimensionless head
(hel H) is plotted against tds for a range of ~ values. Figure 7.7 is similar
except that the abscissa is tds~' Either set of type curves may be used. The
purpose in presenting the two sets is to show the behavior of the solution
for two limiting cases: one in which ~ approaches infinity and the other in
which ~ approaches zero. In the figures, scales of the same size are used.
Curves for the same ~ value are the same size and shape in both figures
because corresponding points in the figures have the same ordinate values
and abscissa values which differ by a constant multiplier. In Figure 7.6 the
curves for values of ~ greater than about 100 converge to a single limiting
curve for ~ approaching infinity. For ~ values greater than 100, dimension-
less head is a function of tds only. In Figure 7.7 the curves for ~ values less
than 0.1 converge to a limiting curve for ~ approaching zero. For ~ values
less than 0.1, dimensionless head buildup depends on tds~ only. For values
of ~ between 0.1 and 100, dimensionless buildup depends on tds (or tds~) and
on the parameter ~. Type curves of the analytical solution of the impulse
variation are presented in Figures 7.8 and 7.9. In Figure 7.8 he! H is plotted
against tds~2 and in Figure 7.9 against tds~' In Figure 7.8 the curves coincide
for ~ greater than about 10. Thus for ~ in this range, hel H is a function of
tds~2. In Figure 7.9 the curves coincide for values of ~ less than about 0.01.
Thus, for ~ in this range, he! H is a function of tds~ only. Between ~ equals
Seal tests 145
1.00
.:.;; 0.50 ~= 00, 1000,100

.r:.0
10
Fig.7.6 Type curves (hjH vs. tdJ) for recovery variation of one-dimensional head-
buildup test.
1.0
~ 0.5
.r:.
10
Fig.7.7 Type curves for recovery variation of one-dimensional head-buildup test

(Equation (7.21)).
0.01 and 10, hc/H depends upon both tds~2(or tds~) and ~. The values used
for plots in Figures 7.6-7.9 are in Greer and Daemen (1991, Appendix A).
General solution procedure

A curve-matching procedure may be used to obtain the hydraulic properties
of a seal (Ks and Ss.) from the results of a head-buildup test. The procedure
is as follows:
1.0
~ 0.5
0.0 L.J...u..wul.::!:::UuIlll:::i:fclIiiL...L.J.JJ.UiiI......I..U.wuL...L..1.JwwL....L...LI.lWIi.....u.J..Wi.lL...J.JL.WWI
10.6 10.5 10-4 10-3 10- 2 10- 1 10
tds/;2 = K,. Sas A'4! SC2
Fig.7.8 Type curves for impulse variation of one-dimensional head-buildup test.
1.0
~ 0.5
.c:
10
Fig.7.9 Type curves for impulse variation of one-dimensional head-buildup test

(Equation (7.25)).
1. Prepare the type curves appropriate to the variation of the test.

2. Make a semilog plot on transparent graph paper of he! H vs. t using the
data from the test. Plot the data using the same scales as for the type
curves, with t on the logarithmic horizontal axis.
3. Overlay the plot of the test data on the set of curves for hel H vs. t ds ('
Keeping the horizontal axes of both plots superposed, move the data
curve horizontally, comparing the experimental curve to the type curves.
Seal tests 147
Align the data curve with the type curve with which it most nearly
matches in shape. If it appears that the best match is to be attained with
the type curve for a large value of ~, perform the matching procedure
using the other set of type curves (i.e. hel H vs. tds or tds~2). If the best
match is attained with one of the limiting curves, go to the procedure for
limiting curves (next section). Otherwise, go to step 4.
4. Select an arbitrary 'match' point from the overlapping portions of the
data and type curves. Note the values of tds~ and t corresponding to this
point. Also, note the value of ~ of the type curve matching the data
curve.
5. Use Equations (7.9) and (7.22) and the values of tds and t noted in step 4
to obtain the conductivity Ks' With the value of ~ from step 4, calculate
Sss using Equation (7.22).
Solution procedure for limiting cases

If the data curve matches the limiting curve for ~ approaching zero, only Ks
may be determined from the test. To obtain K s' select a match point and
calculate Ks using Equations (7.9) and (7.22) and the values of tds~ and t
corresponding to the selected point. An alternate graphical approach based
on the exponential form of the limiting solution is presented in Greer and
Daemen (1991).
For the recovery variation, if the data curve matches the type curve for ~
approaching infinity, only the hydraulic diffusivity (Kj Sss) may be deter-
mined from the test. To obtain Kj Sss' select a match point and calculate
Ksl Sss using Equation (7.9) and the values of tds and t for the match point.
For the impulse variation, if the data curve matches the type curve for ~
approaching infinity, only the product KsSss may be obtained from the test.
To calculate KsS ss' choose a match point and use Equations (7.9) and (7.22)
and the values of tds~2 and t for the selected point.
The axisymmetric analysis of both recovery and impulse variations with
either the one-side or two-side configuration is performed in the same man-
ner. However, as will be discussed subsequently, the analysis is considerably
simpler and more reliable for two-side recovery or impulse tests. In the
axisymmetric analysis the seal and rock mass are each assumed to be
homogeneous and isotropic with unknown hydraulic properties Ks and Sss
and Kr and Ssr' respectively. The test is analyzed using a transient ground-
water flow computer program. In this analysis, the collection zone is treated
as part of the flow domain with an hydraulic conductivity much greater
than Ks or Kr and a specific storage, Sse' equal to ScI~, where ~ is the
volume of the collection zone. A solution consists of the values of K s' Sss' Kr
and Ssr which cause the model to produce the observed head-buildup
curve, he vs. t.
The flow domain for the axisymmetric analysis is shown in Figure 7.10
and Table 7.2. Using dimensional analysis, the following relationship may
be shown to hold among the variables of the axisymmetric model:
(7.26)
where
(7.27)
To achieve a solution, an iterative procedure is followed using multiple

series of computer simulations based on the relationship of Equation (7.26).
The quantities H, H b , Se and ~ and the flow field dimensions are assumed
to be known. An arbitrary value of Ks is assumed and used throughout all
simulations. The inputs for a particular simulation are Sss' Kr and Ssr The
outputs include he as a function of t. The steps of the procedure are as
follows.
1. Make a semilog plot on transparent graph paper of he! H vs. t (t on the

log axis) using the observed data from the test. This is the data curve.
Kc (KcK. and KcKrl

Sse = ScI Vc
Fig.7.10 Flow field for axisymmetric analysis of head-buildup test.

Seal tests 149
Table 7.2 Boundary conditions for flow field in Figure 7.10a

Packer head ah/az = 0

Center line of ah/ar= 0
borehole
aBoundaries r 1 to r 5: same as in Table 7.1.
2. Compute estimates of Sss and Ssr using constitutive expressions and rep-
resentative material properties (Greer and Daemen, 1991, Appendix L).
Use these to compute 'reasonable' values of ~ and ,. Holding these
quantities constant, generate a series of simulations for a range of Krl Ks
values. Plot the results as a family of semilog type curves of he! H versus
tds~ (tds~ on the log axis). Use the same size scales for the type curves as
for the data curve.
3. Match the data curve against the type curves of step 2 in the same way
as in the one-dimensional analysis. The estimated solution value of
Krl Ks is the conductivity ratio of the type curve which most nearly
matches the data curve. For two-side tests in which H #- H b , accurate
matching, particularly of the middle to late time portions of the curve, is
usually relatively easy to accomplish. In such tests the buildup curve at
later dimensionless times is strongly dependent upon the ratio Krl Ks. In
fact, the peak value of hel H achieved at steady state for these tests is
solely dependent upon Krl Ks. (Note that in most one-side recovery or
impulse forms of the test, H, which is on the inaccessible side of the seal,
equals H b For this condition the peak or steady-state value of he! H is
one, regardless of Krl Ks. Thus the type curves for different values of
Krl Ks are very similar in shape and matching is very difficult.)
4. Holding the value of Krl Ks constant and equal to that found in step 3,
perform simulations for a range of ~ and , values which bracket the
values found in step 2. For each, plot a family of semilog type curves
(helH vs. tds~)' with a separate curve for each value of ~.
5. Match the data curve against the curves obtained in step 4. Note the
values of ~ and, of the type curve which most closely matches the data
curve.
6. Using the values of ~ and, found in step 5, repeat steps 2 and 3. If
the value of Krl Ks obtained previously in step 3 is still the best match,
go to step 7. Otherwise, repeat steps 4-6 using the new match value of
KrlKs
7. Matching the data curve with the best-fitting type curve, select an arbit-
rary match point from the overlapping portions of the data and type
curves. Calculate K s' Sss, Kr and Ssr using the match-point values of t and
tds~' the values of Krl K s' ~ and, of the best-fitting type curve and the
equations defining ~, C and tds~.
The families of type curves found in steps 2 and 4 are bounded by limiting
curves in much the same way as in the one-dimensional analysis. If the
best-fitting curve is one of the limiting curves, less hydraulic information
may be obtained from the test. It must be emphasized that the properties
obtained from this analysis are not necessarily unique; there may be other
combinations of property values which yield an equally good or, perhaps,
better match.
If Kr and Ssr are known apart from the test, the analysis is simplified
because parameter, is known. For this case the known, value is used in
step 2. In step 4 simulations are made only for the known' and a range of ~
values. The need to compare the data curve to each of the type curves in
steps 3 and 5 can be eliminated by means of the following. In step 1,
transform the observed he and t data to values of heH and
t dr , = Krt/(SseD2). For the transformation use the known values of K r, Sse' D
and H. In steps 3 and 5 use t dr , instead of tds~ in the type curves and plot
the transformed data directly on the type curves. Interpolate to determine
the value of Kr/ Ks for the type curve which matches the transformed
data in step 3. Interpolate to find the ~ value of the matching type curve in
step 5.
Determination of compressive storage of collection zone

The compressive storage of the collection zone (SJ is due to the compres-
sion of the water in the collection zone and the deformation of the collec-
tion zone walls (including packer, rock and tubing) (Hsieh et al., 1981;
Neuzil, 1982). The compressive storage may be determined in the field by
injecting precisely known volumes of water into the water-filled collection
zone and measuring the immediate increase in pressure. The compressive
storage is the ratio of the volume injected into the collection zone to the
change in pressure head in the collection zone. In the mathematical model
presented, the compressive storage is a constant.
7.3.4 One-dimensional tracer travel-time test

The one-dimensional tracer travel-time test, as presented herein, is a simple
method for estimating an upper limit for the minimum travel time for water
marked with tracer to flow through or around a seal under conditions of
steady flow. Comparison of this travel time to travel times calculated using
known or estimated properties of the seal and rock mass may indicate the
existence of a high-velocity flow path through/around a seal, contrary to the
assumption of a homogeneous seal and rock mass.
The test is performed after a long period of constant-head injection in
which steady flow is approximately established. The test is begun by sud-
denly introducing a slug of tracer to the injection zone without (or with
minimal) interruption of constant-head injection. After introduction of the
Seal tests 151
tracer the collection zone is sampled on a regular basis until the tracer is
detected. The elapsed time between injection of the tracer and its first
detection is an upper limit value for the minimum travel time. The travel-
time test is not intended to be used by itself to determine seal properties.
Rather, it is intended to help interpret or clarify results of the previously
discussed tests. To the author's knowledge, the first tracer travel-time test
on an in situ seal was performed by Sandia National Laboratories in the
1979-1980 Bell Canyon Test (Christensen and Peterson, 1981).
Related methods
The test is presented in several texts (e.g. Cedergren, 1977; Todd, 1980) as an
expedient field technique for estimating the hydraulic conductivity of a zone
of groundwater flow between two wells. In this application a tracer slug is
introduced into a well which is screened in the zone of interest. Samples are
taken from a second well, similarly screened and positioned along the direc-
tion of flow from the first. Sampling is continued until the peak concentra-
tion of the tracer is detected. The time required for detection of the peak
concentration divided into the length of the flow path between the wells is
an estimate of the average linear velocity of the groundwater intercepted by
the wells. That velocity is then used with the average hydraulic gradient
between the wells, an estimate of the average porosity, and Darcy's law to
obtain an estimate of hydraulic conductivity.
The average axial velocity through a seal may be expressed as
(7.28)
and as
(7.29)
where Ns is the effective porosity of the seal [dimensionless] and 1'. is the
travel time for axial flow through the seal from the injection to the collec-
tion side [T]. Substituting Equation (7.29) into (7.28) and the result into
Equation (7.3) gives
(7.30)
For a steady constant-head test or tracer test on a seal in a homogeneous

and isotropic rock mass, the minimum possible travel time, T,.m' for flow
through the rock mass from the injection to the collection zone is the travel
time for the flow path just within the rock mass along the seal. The travel
time along this flow path is the minimum possible because the flow path is
the minimum possible length (i.e. Ls) and the gradient driving flow along
this path is the maximum possible gradient (i.e. (H - HJ/L.). Following the
development of Equation (7.30), the expression for the minimum possible
travel time in the rock mass is
T = __N----'--rL---':'-----_ (7.31)
rm Kr(H -HJ
where N r is the effective porosity of the rock. For a tracer travel-time test, T.
and T.m may be computed using H, He' and Ls from the test and using
known or estimated values of K s' N s' Kr and N r. Let t* be the travel time
obtained from the test. Time t* is an upper limit for the minimum travel
time for flow from the injection to the collection zone. Time t* may be
compared to T. and T.m If t* < < minimum [T., T.m]' there is strong indica-
tion that a high-velocity flow path through or around the seal exists, con-
trary to the assumption of a homogeneous seal and rock mass. If t* is
comparable to minimum [T., T.mJ, the result is consistent with the assumed
flow field. The effects of hydrodynamic dispersion, which probably are mini-
mal for a short seal, are neglected in this analysis.
7.4 CONCLUSIONS
Four tests may be used to characterize the hydraulic performance of

borehole seals: the steady and transient constant-head tests, the head-build-
up test and the tracer travel-time test. The constant-head tests and the
head-buildup test may be analyzed in terms of one-dimensional models in
which all flow is assumed to be axial through a homogeneous and isotropic
porous seal. These tests may also be analyzed using axisymmetric three-
dimensional models in which axisymmetric flow occurs in both the seal and
surrounding rock mass, which are each taken as a homogeneous and iso-
tropic porous material. The tracer test is analyzed in terms of a one-dimen-
sional model only. The hydraulic properties obtained from analysis of the
tests are summarized in Table 7.3. The tests may be performed using either
the one-side or two-side configuration (Figures 7.1 and 7.2).
7.4.1 Steady constant-head test

Of the tests presented, the steady constant-head test is the simplest to
perform and most straightforward to analyze. As flow is assumed to be
steady, the test yields only hydraulic conductivity. In general, an upper limit
for the hydraulic conductivity of the seal is obtained by one-dimensional
analysis of the test. The one-dimensional analysis utilizes a simple express-
ion based on Darcy's law. Axisymmetric analysis of a two-side test gives the
conductivity of the seal and of the rock mass. Axisymmetric analysis of the
Conclusions 153
Table 7.3 Hydraulic properties obtained from test analysis
Test Hydraulic properties from analysis
One-dimensional Axisymmetric
Steady constant-head K, K"K r

Transient constant-head K,/Sss KriSsr (or
K,/Sss)' S,rlSss
Head-buildup K" Sss K" sss, K r, Ssr
Trace travel-time Existence of high-
velocity flow path
one-side test gives an infinite set of pairs of seal and rock conductivities, any
one of which causes the model to reproduce the field test results. The actual
solution pair of seal and rock conductivities may only be determined from
the infinite set if the rock conductivity is known by an independent means.
Of the axisymmetric analyses presented, that for the steady test is the
simplest to perform, requiring only one series of simulations.

The transient constant-head test is probably the most difficult of the tests to
perform because it requires the monitoring of what may be very rapid
changes in very small flow rates. One-dimensional analysis yields the hy-
draulic diffusivity (Ks/ Sss) of the seal. By determining Ks from the steady
flow rate either preceding or following the test, Sss may be obtained from
the diffusivity. Axisymmetric analysis yields the diffusivity of either the rock
or the seal and the ratio of rock and seal specific storages (15 = Ssr/ Ss.). If Ks
and Kr are determined by axisymmetric analysis of the steady flows either
before or after the test, Sss and Ssr may be obtained from the diffusivity
and b.
If the hydraulic properties of the rock mass (K r and Ssr) are known by a
means independent of the test, the analysis for Ks and Sss is simpler and
more reliable.
The one-dimensional analysis may be performed using either the transi-
ent injection flow rates or the transient collection rates or both. In principle,
the axisymmetric analysis could also be performed with just the transient
injection rate or just the transient collection rate (assuming that both the
steady injection and collection rates at the end of the test are known).
However, in the axisymmetric analysis the most reliable calculation of hy-
draulic properties is accomplished when matching both injection and collec-
tion data curves.
As discussed in Chapter 8, obtaining accurate plots of the observed tran-
sient injection and collection flows is quite difficult. In the absence of
accurate data plots which can be matched against the type curves, a
one-dimensional analysis may be advisable, as a reliable axisymmetric
analysis may not be possible. Greer and Daemen (1991) showed that one-
dimensional analysis of either the transient injection or collection rate yields
seal diffusivity (Ks/ Ss.) values within a factor of three to four of those
obtained by axisymmetric analysis for Kr/ Ks :::;; 0.1 and J:::;; 1. The one-
dimensional analysis yields seal diffusivities nearly identical to those of the
axisymmetric analysis for Kr/ Ks :::;; 0.01 and J :::;; 0.01.

Two variations of this test are presented: recovery and impulse. Both vari-
ations may be performed with either the one-side or two-side configuration.
One-dimensional analysis of tests made using either variation yields both
Ks and Sss'
Axisymmetric analysis of the test is difficult because matching must be
performed against multiple series of type curves. However, for the axisym-
metric analysis of a two-side test in which H =1= H b , the step involving
matching against type curves for a range of Kr/ Ks values (step 3) is relative-
ly easy to perform accurately because the shape of the buildup curve at
later dimensionless times is strongly dependent upon the ratio Kr/ Ks. For
most tests following the one-side configuration, this strong dependence
does not exist and matching is much more difficult. Successful axisym-
metric analysis yields K s' Sss, Kr and Ssr. In an axisymmetric analysis the
properties obtained are not necessarily unique; there may be other combi-
nations of property values which yield an equally good or, perhaps, better
match.
The families of type curves used in both the one-dimensional and axisym-
metric analyses are bounded by limiting curves. If matching is achieved with
a limiting type curve, less hydraulic information may be obtained from
the test.
If Kr and Ssr are known apart from the test, the analysis for Ks and Sss is
simpler and more reliable.
7.4.4 Tracer travel-time test

The tracer travel-time test may be analyzed with a one-dimensional, steady-
state, porous-medium model in terms of the hydraulic conductivity and
porosity of the seal and of the surrounding rock. The travel-time test is
intended to be used as a complement to the other tests. It is not presented
as a means to determine hydraulic properties, but rather to indicate the
likely existence of a high velocity flow path through or around the seal.
The tracer test and its analysis are relatively simple to perform. The
greatest potential difficulty in the analysis is obtaining reliable estimates of
the effective porosities of the seal and rock mass.
Conclusions 155
7.4.5 Comparison of one-side and two-side configurations

The two-side configuration has a number of advantages compared to the
one-side configuration.
In the steady constant-head test, use of the two-side configuration allows
determination of both Ks and Kr from an axisymmetric analysis. Axisym-
metric analysis of a one-side test permits determination of an infinite set
of pairs of Ks and Kr values, anyone of which will provide the observed
test flows. The particular solution pair may only be obtained with addi-
tional information or testing.
While using the one-dimensional model for steady tests, the collection
flow from the two-side test generally gives a lower upper limit on seal
hydraulic conductivity than the injection flow from the one-side test.
In an axisymmetric analysis of a transient constant-head test, the transi-
ent injection curves vary in shape much less than the collection curves.
For the one-side test, with only injection curves, obtaining an unambigu-
ous match is very difficult for the axisymmetric model. The task is more
easily accomplished for the two-side test with curves for both injection
and collection.
In axisymmetric analyses of two-side head-buildup tests in which H # H b ,
Krl Ks may usually be determined simply and reliably with only estimates
of the other hydraulic properties, thus greatly simplifying the solution
process. With the one-side configuration, H usually equals H b; as a result
the type curves for different Krl Ks vary much less in shape, making
unambiguous matching much more difficult.
The performance of the tracer travel-time test on a seal accessible from
only one side is much more difficult. It generally requires placement of a
tracer injector, capable of delayed activation, on the inaccessible side of
the seal prior to seal installation.
In general, control of experimental equipment in two-side tests, with
access at both ends, is superior to that in one-side tests.
The primary advantage of the one-side configuration is that it may be
employed to test any seal which is accessible from one or both ends of the
borehole. On the other hand, the two-side configuration requires access to
both ends of the seal, which usually necessitates access from both ends of
the hole. Another advantage of the one-side configuration is that natural
fluid pressure can often be utilized to pressurize the injection zone.
7.4.6 Comparison of one-dimensional and axisymmetric analyses

The advantages of one-dimensional analyses compared to axisymmetric
include the following.
One-dimensional procedures are simpler and more straightforward. 'Ana-
lytical expressions are used for one-dimensional analyses whereas
multiple runs with a computer flow model are required for axisymmetric
analyses. All the one-dimensional analyses yield unique solutions;
axisymmetric solutions of the head-buildup test and possibly of the tran-
sient constant-head test may not be unique. Obtaining unambiguous
matches on multiple series of type curves for the axisymmetric analysis of
the head-buildup test is extremely difficult compared to the matching
required in the one-dimensional analysis.
The accuracy of a one-dimensional analysis may be adequate for many
purposes. For example, the upper limit on seal conductivity provided by
one-dimensional analysis of the steady constant-head test may be suffi-
ciently low that the added effort required for an axisymmetric analysis is
not justified.
Axisymmetric analyses have advantages compared to one-dimensional
analyses.
Axisymmetric analyses provide a more realistic simulation of hydraulic
behavior. In any field test there will be flow components in the rock mass
which may influence test results significantly, particularly for Krl Ks > 0.1.
These components usually can be accommodated in the axisymmetric
model.
Axisymmetric analysis often permits determination of rock hydraulic
properties, as well as those of the seal.
ACKNOWLEDGEMENTS
ledged.
CHAPTER EIGHT
In situ Hydraulic
Performance of Cement
Borehole Seals
W B. Greer and D. R. Crouthamel
8.1 INTRODUCTION
This chapter presents techniques and instrumentation for the in situ hydrau-
lic performance tests of cement borehole seals. These tests were conducted
at the Oracle Ridge mine site, where the seal was installed in a nearly
horizontal borehole, and at Superior site, where the seals were installed in
vertical boreholes. The methods used in the data interpretation and analysis
given in Chapter 7 were applied at both sites. Test results and problems
associated with the testing are discussed. Recommendations on the in situ
hydraulic performance test of borehole seals are given in section 8.6.
8.2 PRELIMINARY SURVEY FOR SEAL INSTALLATION
Preliminary surveys to assist in the selection of seal location within the

boreholes consisted of logging of core, video logging of the borehole and
constant-head injection tests using a straddle-packer unit. Injection testing
was performed in selected intervals of the boreholes. The criteria for select-
ing the seal location were
absence of hydraulically significant fractures;
low effective hydraulic conductivity of the interval; and
convenience for seal installation and testing.

158 In situ hydraulic performance of cement borehole seals
8.3 SEAL TESTS AT ORACLE RIDGE MINE
Oracle Ridge Mine is located in the Santa Catalina Mountains near

Tucson, Arizona. At the underground mine an expansive cement grout seal,
10 cm in length, was installed in a diamond cored borehole penetrating the
recrystallized limestone of the Escabrosa Formation. The hole was 10 cm in
diameter, 33 m long, dipped at 10.4 and connected two drifts. With access
to both ends of the borehole, tests using the two-side configuration (Fig-
ure 7.2) were performed. The location selected for the seal was about 3 m
from the lower end of the borehole, which permitted relatively convenient
access to the seal from that end. Results in the interval selected for the seal
indicated an effective hydraulic conductivity of 3 x 10- 10 to 3 x W- 9 cm/s.
A temporary mechanical plug was installed in the hole at the lower end of
the selected interval. A mass of cement grout slurry, 1.5 m in length, was
placed above the plug using a dump bailer. Groundwater in the hole, if any,
was minimal during placement. After 13 days of curing the temporary plug
was removed and the grout mass drilled out partially from both ends,
leaving a seal Wcm long.
8.3.1 Hydraulic conditions in the rock mass

The following observations are made about the distribution of water and
hydraulic head in the rock mass near the seal.
The walls of the drift near the lower end of the borehole where the seal
was located were slightly moist or dry throughout the year. The walls of
the borehole below the seal were moist or slightly moist.
Injection above the seal was nearly continuous for over 3 years at heads
of 1900-3400cm (Head is measured with respect to the horizontal plane
passing through the centroid of the lower face of the seal.)
Two piezometers were installed in the rock mass at the elevation of the
seal. The piezometer tips were 3.6 m apart and the seal lay approximately
midway between them. One of the piezometers consistently indicated
heads of less than 500 cm as injection continued and heads less than
400cm for 7 months after the termination of injection. The other
piezometer registered heads between 1600 and 1900 cm during injection
and between 1200 and 1600cm after its termination.
During the head-buildup test, when the collection zone was shut in and
the collection rate was near zero, the injection flow (at a head of about
1900 cm) was in reverse; that is, flow was from the formation into the
injection zone.
For 3 months of constant-head testing before and for nearly 2 months of
testing after the head-buildup test, the injection rate was significantly less
than the collection rate. The injection heads during these periods were
between 1810 and 1950cm.
Seal test at Oracle Ridge Mine 159
Based on these observations, the rock mass around the seal was probably
saturated. The relatively high heads in the formation indicated by the one
piezometer, the flow reversal, and by injection rates less than collection
rates suggest that at least part of the rock mass near the seal was connected
hydraulically to a source of natural (i.e. not related to the injection) head in
excess of 1900 cm. Assuming that the rock mass was generally saturated, the
absence of significant moisture on drift and borehole walls near the seal
indicates that fractures were tight and effective hydraulic conductivity of the
rock mass was low.
Steady constant-head test

The steady constant-head test using the two-side configuration was per-
formed on the seal nearly continuously for most of the 3.5 year testing
period. Water was injected at constant pressure into a water-filled injection
zone formed on the upper side of the seal and maintained at a constant
head H. The outflow from the seal was collected in a water-filled collection
zone on the seal's lower side. The collection zone was kept at a nearly
constant head, He (with He < H) by allowing it to vent to atmosphere
through a pipet. The flow rate into the injection zone (i.e. injection rate, Qi)
and flow rate out of the collection zone through the pipet (i.e. collection
rate, Qe) were measured. From these rates the hydraulic conductivity of the
seal was estimated from a one-dimensional analysis and the conductivity of
the seal and of the surrounding rock mass from an axisymmetric analysis.
Test set-up
A gas-over-water pump (Figure 8.la) delivered water at constant pressure to
the injection zone. The injection zone (Figure 8.1b) was formed by inflating
a pneumatic packer just above the seal. Before inflating the packer the
borehole above the seal was filled with water, and the injection pump
tubing and injection line were purged of any trapped air. Head in the
injection zone is the pressure head supplied by the compressed gas driving
the injection pump plus the height of the water level in the pump flow tube
above the datum. During testing, the head in the injection zone generally
varied slightly due to the drop in water level in the flow tube and variations
in the pressure supplied by the compressed gas. The average head in the
injection zone for a test period was taken as the average of the heads at the
beginning and end of the period. The water injected was from a sump of
mine drainage and was filtered prior to use in the pump. During testing, the
borehole above the injection packer was filled with water. With respect to
the datum, the hydraulic head supplied by this column of water was about
540cm.
Vent
Bourdon-tube
From the pressure gage
sump
O.85cm
1.0. flow
tube
N2 cylinder
Injection line
(a)
Injection packer
(b)
Fig. 8.1 (a) Gas-aver-water pump; the flow tubes are made of translucent Pvc. (b)
Injection packer and injection zone.
The collection system measures the water flow from the seal and forma-
tion to the collection zone during a test period. The collection system
evolved as testing continued (Figure 8.2). A mechanical plug was positioned
just below the cement seal and oriented so that one of its two ports (Fig-
ure 8.2) was at the crown of the borehole, the other at the invert. Before
testing, the collection zone between the seal and the mechanical plug was
filled (except for a small air pocket) with water from the reservoir. Air
bubbles were purged from all tubing. During a test period, valves A, Band
D were closed; value C was open. The flow of water into the collection zone
from the seal and surrounding rock caused an upward displacement of the
water level in the collection pipet. The pressure in the collection zone rose
slightly as the water level in the collection pipet rose. As a result, the
volume of the small air pocket decreased. During a test period the inflow to
Reservoir
25 ml collection
pipet
c 25 ml drain
D
Grout seal pipet
Flexible
tubing
A Valve
Fig.8.2 Collection system used for steady-state constant-head testing.
the collection zone equaled the volume displaced in the collection pipet
plus the reduction in the volume of the air pocket. The volume of water
displaced in the pipet is called the 'uncorrected' collection volume, as it
does not account for the reduction in the air pocket. The 'corrected'
collection volume was obtained at the end of the test period by opening
value D and releasing water from the collection pipet to the drain pipet
(Figure 8.2). By adjusting the water level in the collection pipet back to its
level at the start of the test period, the air pocket was returned to its initial
volume. The volume of water displaced in the drain pipet during the
draining process equals the inflow to the collection zone during the test,
corrected for the compression of the air pocket. The average corrected
collection rate for a test period is obtained by dividing the corrected
volume by the length of the period. The average uncorrected rate is ob-
tained in a similar manner using the uncorrected volume. The average
hydraulic head in the collection zone for a test period is taken as the
average of the water-level elevations in the collection pipet at the begin-
ning and end of the period.
Steady constant-head testing began 18 days after the seal was installed
and continued for 3.5 years, with brief interruptions for other tests. Com-
plete results of the testing are given in Greer and Daemen (1991). In the first
injection test, flow was visually observed issuing from the top portion of the
seal on the collection side. In the subsequent tests the flow which could be
seen coming from the seal was greatly reduced. Evaporation was monitored
for 18 months in a pipet identical to, and placed next to, the collection
pipet. The evaporation rate was found to be negligible. For the 17 months
during which both corrected and uncorrected volumes were measured, the
corrected volumes were an average of 5% greater than the uncorrected
volumes.
The axisymmetric analysis of the steady constant-head test is based on the
average heads and flow rates for three testing seasons spanning 7 of the 10
months. The average values for the seasons are presented in Table 8.1.
These particular time intervals were selected because the injection and col-
lection heads are reasonably constant within each interval and different
injection heads are used for the intervals. For two of the three seasons,
corrected collection rates are not available. For these seasons corrected
rates are estimated by increasing the uncorrected rates by 5%.
The flow domain for the axisymmetric analysis is shown in Figure 8.3 and
Table 8.2 The axial and radial boundaries of the rock mass are sufficiently
large that injection and collection rates are relatively insensitive to changes
in the positions of these boundaries. The head (Hb) at the axial and radial
boundaries of the rock mass is assumed to be unknown but uniform and
constant and is to be determined as part of the analysis.
To perform the analysis the following steps are taken.
1. For each of the three testing seasons, the procedure described for the
two-side configuration (Section 7.3.1) is executed for a range of Hb
values. The average injection and collection rates and heads are used. As
a result of this step, solution values of Ks and Kr (designated K: and K:)
are obtained for the range of Hb values, for each of the three seasons.
2. A semilog plot is made of K:
(on the logarithmic axis) vs. Hb for
each season (Figure 8.4). A similar plot is made with K:
and Hb
(Figure 8.5).
3. The intersection of any two of the three curves in Figure 8.4 and of the
two corresponding curves in Figure 8.5 constitutes a solution. If the test
conformed perfectly to the axisymmetric model, the three curves of Fig-
ure 8.4 would intersect at a common point whose coordinates are the
hydraulic conductivity of the seal and the boundary head, H b Also, the
curves of Figure 8.5 would intersect at the point whose coordinates are
the conductivity of the rock mass and H b .
Table 8.1 Average test results for three seasons of steady constant-head testing
Testing Average Average Average Average Average QJQc based
data injection collection injection uncorrected corrected on average
seasons head zone head rate collection collection injection
(em) (em) ( cm 3 /h) rate rate and corrected
(cm 3 /h) (cm 3 /h) rates
2/14-4/22 1850 33 0.0641 0.0898 0.0943" 0.680

5/2-7/3 2520 33 0.146 0.0894 0.0939" 1.56
9/24-12/20 3340 34 0.253 0.0956 0.100 2.53
"Estimated by increasing the uncorrected rate by 5%.
z=520
z= 88.0
z= 37.0
r1
z= 5.0
K,
rS
r 12
z=-5.0 Ks
r 11
z--7.0
z= -10.0
I
(=5
I
(=485
Fig.8.3 Flow field for axisymmetric model of steady-state constant-head test.

Units are in centimeters.
In Figure 8.4 the curves intersect at points 1,2 and 3. The corresponding
points in Figure 8.5 are 1',2', and 3'. The conductivity and head values
corresponding to the intersection points are presented in Table 8.3. If only
the results of seasons 2/14-4/22 and 5/2-7/3 are considered, the solution is
given by the conductivity and boundary-head values corresponding to
points 1 and 1'. The solution values for boundary head for points 1 and l'
should be equal. The small difference in the values is probably due to errors
in reading the plots in step 1. Similarly, points 2 and 2' and 3 and 3'
correspond to the solutions for the cases in which seasons 5/2-7/3 and
9/24-12/20 and seasons 2/14-4/22 and 9/24-12/20, respectively, are con-
sidered.
Table 8.2 Boundary conditions for flow field in Figure 8.3

r1 Injection face of seal h = injection head (H)

r2 Borehole wall in h = injection head (H)
injection zone
r3 Gland of packer ohjor= 0
r4 Wall of water-filled
borehole above packer h= 540cm
rs Axial boundary of h=Hb
rock mass
r6 Radial boundary of h=H b
rock mass
r7 Axial boundary of
rock mass h=Hb
rs Wall of open borehole
below seal h = 0 (appro x.)
r9 Gland of mechanical
plug ohjor = 0
r 10 Borehole wall in h = collection zone
collection zone head (Hcl
r l1 Collection face of seal h = collection zone
head (He)
r 12 Center line of borehole ohjor= 0
10~r---------------------------------'
2114 - 4122
~
1 10.9
r-~_~
ic" 9/24 - 12120
10. 100.'-'--'-'-'-L...J....L...J..~27:00:::0'-'--J...J..-'-'....I..J'-40~OO~....L..J--'-J...J..~67:000
BOUNDARY HEAD (em)
Fig.8.4 Solution values of seal conductivities.
The average solution for Ks is about 4 x 10- 10 cm/s. For Kr the average
is about 3 x 10- 10 cm/s. The solution value for head along the axial and
radial boundaries is about 3200 cm. The close grouping of points 1,2 and 3
and 1',2' and 3' tends to confirm the appropriateness of the axisymmetric
model.
10~c---------------------------------,
i 10.1 L-~---
';.:
10. 11 0~-'-'--'--'-'-"---'---':-2000~--'-'-'--'-'--'--'-4'-=OO':c0:'-'-.J-1--'-'-'--"--':--:'6000
BOUNDARY HEAD (em)
Fig. 8.5 Solution values of rock conductivities.
Table 8.3 Solution values of seal conductivity (Ks), rock con-

ductivity (Kr) and boundary head (Hb) for axisymmetric analy-
sis of steady constant-head test
I nterseetion Ks Kr Hb
point number" (xlO-lOem/s) (xlO-loem/s) (em)
6.0 3130
l' 2.7 3210
2 3.1 3380
2' 3.0 3280
3 4.3 3090
3' 2.9 3190
aFrom Figures 8.4 and 8.5. Points 1 and l' are the intersections of
solution curves (Figures 8.4 and 8.5) for seasons 2/14-4/22 and
5/2- 7/3. Points 2 and 2' are the intersections of solution curves for
seasons 5/2-7/3 and 9/24-12/20. Points 3 and 3' are the intersections
of solution curves for seasons 2/14-4/22 and 9/24-12/20.
Assuming steady axial flow through a homogeneous and isotropic seal, the
hydraulic conductivity of the seal is obtained from Equation (7.3) using
either injection rate, Qi' or collection rate, Qe' for Qld' Simulations (Greer
and Daemen, 1991) show that if the test conforms to the axisymmetric
model and He:::::; 0, the value of Ks computed using Qi in Equation (7.3) is an
upper limit value provided Hb < 1.9 H. The simulations show that Ks cal-
culated using Qe is an upper limit for any value of H or H b , provided both
are greater than zero and He:::::; 0.
The hydraulic conductivity of the seal is calculated using Qi in Equa-
tion (7.3). The minimum values of H occur in the first flow pt;riod, with the
smallest value being 1810 cm. Assuming that the axisymmetric model holds
and that Hb equals 3200 cm (as obtained in the axisymmetric analysis),
Hb/H = 3200/1810 = 1.8 < 1.9, which implies that the conductivity values
based on the injection rate are all upper limit values. From the one-dimen-
sional analysis an upper limit for Ks is about 1 x 10- 9 cm/s. The values of
Ks based on injection and collection rates slowly decline over the course of
testing at a given injection head.

A transient constant-head test in which only the collection rate was
monitored was performed after a period of steady constant-head testing. At
the start of the test a step decrease in injection head was applied. The test
consisted of monitoring the transient collection rate until approximately
steady conditions were again obtained. The test is analyzed in terms of the
one-dimensional model to obtain an estimate of the hydraulic diffusivity of
the seal (KJSss).
Test set-up
The equipment used for the transient constant-head test was the same as for
the steady test, except that a buret with a capacity of 5 ml and subdivisions
of 0.01 ml was used as the collection tube. The buret was inclined in order
to minimize the effects due to the compression of the air pocket.
The inclined buret was installed without interrupting ongoing constant-
head injection. The buret volume readings taken during the transient con-
stant-head test were uncorrected; that is, the readings did not take into
account the compression of the small air pocket at the apex of the collection
zone. An experimental technique was presented for making this correction.
The technique is only practical for the steady constant-head test, in which
the time interval between readings is of the order of several days. It is not
suited to the correction of buret readings taken in the transient test in which
the interval is a few tenths of an hour. To correct the buret readings, an
algorithm was prepared based on the assumptions that the air pocket com-
presses according to Boyle's law and that the collection zone is a cylindrical
volume with a horizontal axis. The algorithm utilizes uncorrected and cor-
rected collection tube volumes from previous steady testing to determine
the volume of the air pocket (idealized) at the start of the test. It then
computes a 'corrected' volume reading for each 'uncorrected' reading ob-
tained in the test.
The test is interpreted in terms of the one-dimensional model for the transi-
ent constant-head test by plotting the test data as shown in Figure 8.6. In
developing the data for the plot, the initial and final driving heads are used
for H 1 and H 2' respectively. Qo is the steady collection rate achieved at the
end of the test, 0.041 cm 3 /h. Superimposing the plot in Figure 8.6 on the
type curve for collection flow (Figure 7.4b) yields the following match-point
values: t d, = 1.0 and t = 3.8 h.
Substituting t = 3.8 h = 1.37 X 104 sand Ls = 10 cm into the expression
for tds yields K,/Sss = 7.3 x 10- 3 cm 2 /s. The hydraulic conductivity may be
computed using the steady collection rate which either preceded or followed
the test. Using the steady rate established after the test (i.e. Qo = 0.041 cm 3 /
h), a driving head of 1960 cm, a cross-sectional area of 78.5 cm 2 and a seal
length of 10 cm, the conductivity is calculated to be 7.4 x 10- 10 cm/s by
Equation (7.10). With this conductivity the specific storage for the seal is
about 1 x 1O- 7 cm- 1 .
The data curve (Figure 8.6) only approximately matches the type curve.
Three of 12 data point are not shown on the curve as the measured flow
rates were less than the final flow rate, Qo' Of the points which were plotted,
three diverge considerably from the theoretical curve. These variations
probably result from the inability of the measurement system to measure
accurately the very low and rapidly changing collection rate.
The hydraulic conductivity of the seal calculated using the steady collec-
tion rate and driving head which preceded the test, 0.090 cm 3 jh and
3350 cm, is 9.5 x 10- 10 cm/s. This value is in reasonable agreement with
that obtained using the flow rate and head following the test.
1.0,..--------------------,
Q o = 0.041em3/h
=
H1 3350cm
H2 = 1960 em

:
I
'"
2=0
o
~ 0.5
o
I
1-

0.0 ' -_ ___'__~..L......J___'_........

.&.J..._ _'____'___'_..........'_'_........
10.1 1 10
ELAPSED TIME (h)
Fig.8.6 Results of transient constant-head test at Oracle Ridge mine.


A head-buildup test, recovery variation, was performed after a long period
of constant-head testing in which near-steady conditions were well estab-
lished. To start the test a valve controlling the outflow of water from the
collection zone was abruptly closed. The test consisted of recording the
buildup in head in the collection zone following the closure. The results are
analyzed in terms of the one-dimensional model of the test to obtain esti-
mates of the hydraulic conductivity and specific storage of the seal. An
axisymmetric analysis is attempted, yielding estimates of both seal and rock
hydraulic properties.
Test set-up
The injection zone used in the head-buildup test was the same as that used
in the steady constant-head test. The average head maintained in the injec-
tion zone was 1920 cm. The equipment set-up for the collection zone is
shown in Figure 8.7. Due to the higher pressures expected to develop in the
collection zone during the head-buildup test, the mechanical plug used in
constant-head tests was replaced by a pneumatic packer with an inflated
gland length of 46 cm. The packer was placed so that the air-removal tube
(Figure 8.7) touched the lower face of the seal at the crown of the borehole.
The length of the collection zone for the head-buildup test was 25 cm. A
bourdon-tube gage (with range of 0-200 psi or 0-1380 kPa, and subdivi-
sions of 0.5 psi or 3.4 kPa) was attached at point E to monitor collection
zone pressures. The tubing from the packer to valves A, Band C and to the
250 ml graduated
cylinder Reservoir
Collection tubes 25ml

pipet
I-'E=-----if7\ Bourdon-tube
vpressuragage
A
B Valve
Fig.8.7 Collection zone and system for monitoring pressure for the head-buildup
test.
pressure gage was of stainless steel, 6.3 mm in outside diameter. The collec-
tion zone was filled, except perhaps for a small air pocket, with water from
the reservoir. Air escape from the collection zone during filling was accom-
plished by opening valve B. Air was carefully bled from all tubing. To
expedite filling of the collection zone and purging of air from the system, a
peristaltic pump was installed in the line from the reservoir. With valves C
and D open and B closed, the large collection tube was filled to overflowing.
At overflowing the pressure gage read 2.4psi(169cm). Valve A was then
closed. The system was left in this configuration for 5 days to establish the
initial conditions required for the test.
The test was begun by closing valve C (Figure 8.7). Readings of the
pressure gage were taken periodically. The time interval between readings
was short initially but was lengthened over the course of the test. The test
was terminated 45 days later at a pressure reading of 23.8 psi (164 kPa).
Prior to the start of the test, with the 250 ml cylinder filled to overflowing,
the pressure gage read 2.4 psi or 169 cm. By survey, the elevation of the top
of the cylinder was 47 cm above the datum. Therefore, the hydraulic head
corresponding to each pressure gage reading is obtained by subtracting
122 cm (i.e. 169 - 47 cm) from the reading. These head values are used in the
axisymmetric analysis of the test. In the one-dimensional analytical model
the initial head in the collection zone is assumed to be zero. To adjust the
head values to those compatible with the one-dimensional model, 47 cm,
which is the initial head in the collection zone, is subtracted from each head
value. Similarly, the injection head is reduced by 47 cm to 1870 cm to
conform to the model.
During the head-buildup test the injection flow was reversed for nearly
the entire period of the test; that is, the direction of flow in the injection line
was toward the injection pump rather than toward the injection zone.
Determination of compressive storage of collection zone

The compressive storage of the collection zone (SJ is defined as the volume
of water added to the collection zone per unit increase in hydraulic head in
the zone. The compressive storage is due to the compression of the water in
the collection zone and the deformation of the walls of the zone (including
the packer, rock and tubing) (Hsieh et al., 1981; Neuzil, 1982). The compres-
sive storage was determined in the laboratory. The pneumatic packer and
steel tubing from the field test were installed in 4 in diameter (10 cm) sched-
ule 40 steel pipe filled with water. The collection zone was the space be-
tween the packer and a steel end plate bolted to one end of the pipe. The
pressure in the collection zone was systematically raised from 0 to 2110cm
in steps of 140cm by injecting water into the zone using a precision piston-
displacement pump. The volume of water required for each increase in
pressure was recorded. Sc, the ratio of the incremental volume of water
added to the incremental pressure change, which should be approximately
constant, was found to vary by a factor of 5 over the tested pressure range.
The average value for Sc was 1.7 x 1O- 2 cm 2 .
In the one-dimensional analysis of the test, a plot of the observed buildup
data is matched against a family of dimensionless type curves of the analyti-
cal solution for the one-dimensional recovery variation. The method is
discussed in Chapter 7; the data curve is shown in Figure 8.8. The ordinate
values of the data points are obtained by dividing the adjusted head values
by 1870 cm, the adjusted injection head.
Matching is performed by superimposing the data curve on the type
curves of hjHvs. tds~ (Figure 7.7). Keeping the horizontal axes of the two
plots collinear, the data curve is moved horizontally and compared to each
of the curves. The data curve is found to match most closely the type curve
with ~ = 3. Convenient match-point values of t = 1000 hand tds ~ = 8 are
obtained. As shown in Table 8.4, these values, along with the ~ value of the
matched type curve, yield a hydraulic conductivity and a specific storage of
the seal of 5 x 10- 9 cm/s and 7 x 10- 5 cm -1, respectively.
The data curve matches the type curve for ~ = 3 closely until about
t = 240 h. After that time the data curve flattens more rapidly than the type
curve. The flattening is due to flow components into the rock mass which
cause the head in the collection zone to approach ultimately a lower head
than that of the injection zone.
1.0r------------------------.

:r:
l' 0.5

I
/
.'

10 10 2 10 3 10 4
TIME (h)
Fig.8.8 Results of head-buildup test.

Table 8.4 Calculation of hydraulic conductivity and specific storage based
on one-dimensional analysis of head-buildup test
Quantity Value
Dimensionless time (tds~) at 8
match point
Time (t) at match point 1000h
Compressive storage (Sc) of O.Ol72cm2
collection zone
Length of seal (LJ IOcm
Cross-sectional area of seal (A) 78.5 cm 2 (seal diameter = 10 cm)
Hydraulic conductivity 2 x 10- 5 cmjh = 5 x 10- 9 cm/s
(Ks) = tds~ ScLJ(At)
(Equations (7.9) and (7.22))
Dimensionless parameter (~) of 3
type curve matching data curve
Specific storage (S,,) = ~ Sc/(ALs)
(Equation (7.22))
An analysis of the head-buildup test was performed in terms of the axisym-
metric numerical model (described in Chapter 7) in which the assumed flow
field consisted of the seal and surrounding rock mass with the head at the
axial and radial boundaries of the rock mass equal to zero. The analysis, an
iterative procedure in which the test buildup data is matched against
multiple series of type curves, is presented in detail in Greer and Daemen
(1991). The solution values of seal and rock hydraulic properties were not
unique; that is, the data curve appeared to match reasonably well more
than one type curve. However, the seal properties for all solutions were
quite consistent: a seal conductivity of 6 x lO-9 cmls and specific storage of
about 1 x lO-4 cm -l. The analysis gave a rock conductivity of
1 x lO - 10 cmls and specific storage of 2 x 10- 7 cm - 1 or smaller. An im-
proved analysis could be made by assuming that the head at the axial and
radial boundaries of the rock mass is equal to 3200cm, as determined in the
analysis of the steady test.
8.3.4 Tracer travel-time test

A tracer test was performed on the seal, to determine an upper limit to the
minimum travel time for tracer-marked water to move, under a constant
injection head, from the injection to the collection side of the seal. The test
was performed after a long period of injection at constant head. The test
began with the interruption of injection for 3.5 h to replace the water in the
injection zone with a water solution of potassium bromide. Injection was

then resumed and the collection zone sampled periodically. Samples were
examined using liquid chromatography for the presence of the bromide
tracer. The test results are analyzed in terms of a one-dimensional porous-
medium flow model.
Test set-up
The injection system used before and during the tracer test was the same as
used in steady constant-head testing. The collection and sampling system
for the test is shown in Figure 8.9. The mechanical plug used in steady
constant-head testing was installed below the seal, creating a collection zone
about 2 cm long. The mechanical plug was oriented so that one of its two
ports was at the crown of the borehole, the other at the invert. Prior to the
test the collection zone, sampling bulb and tubing (Figure 8.9) were filled
with water from the reservoir.
The circulation of water with a peristaltic pump helped to flush air from
the collection zone and tubing to the sampling bulb, where it was expelled
from the system. Then, with the pump turned off, the connecting line to the
tank was disconnected at point B and that end of the connecting line held at
about the same elevation as the reservoir water level. At this point, the
collection system was ready to collect outflow from the seal and surround-
ing formation. The mechanical plug was then installed.
To take a water sample, the peristaltic pump was run for 5 min at a low
rate to achieve uniform mixing within the collection zone, sampling bulb
and tubing. Then, with the pump still operating, the water tank was connec-
ted into the system and a 60 ml sample was removed by inserting a syringe
needle through the septum of the sampling bulb. The tank was then discon-
nected from the system and the pump shut off.
Reservoir
Grout seal
~ Indicates direction of flow
Fig. 8.9 Tracer collection and sampling system.

A volume of 11 L of a 10000ppm bromide solution was prepared by

dissolving 164 g of potassium bromide crystals in 11 L of filtered mine
water. The mine water (without the bromide salt) was found to contain a
negligible concentration of the bromide ions. The densities of the tracer
solution and the mine water at room temperature were 1.007 and
0.996 giml, respectively.
Test procedure
Constant-head injection was interrupted, the injection-side packer removed
and a suction pump used to pump out the water in the hole above the seal.
Next the potassium bromide solution was poured into a PVC pipe which
extended from the upper end of the borehole down to the top of the seal.
The quantity of solution was enough to fill the borehole for a length of
about l.4m above the seal. After removing the PVC pipe, the injection-side
packer was reinstalled in the borehole. Injection was resumed at a head of
about 1900 cm. Samples of about 60 ml were taken periodically. An attempt
was made initially to analyze samples for the bromide tracer using an
ion-selective electrode. However, in trial runs before the tracer test the
method did not give repeatable results and therefore was abandoned. All
analysis was performed by high-performance liquid chromatography
(Stetzenbach and Thompson, 1983) by the Analytical Chemistry Laboratory
of the University of Arizona. The detection limit of the chromatography
procedure was about 0.1 ppm.
First detection of the tracer occurred in the ninth sample, which was
taken 565 min or 9.42 h after the injection zone was pressurized. During the
test and for 2 months prior to the test the head in the injection zone
averaged 1920 cm. The average head in the collection zone during the test
was about 30 cm.
A minimum travel time of not more than 565 min indicates relatively rapid
flow through a fracture or an opening along the seal-rock interface and is
not consistent with either flow through the intact cement of the seal or
through the intact limestone. This conclusion can be supported by calculat-
ing the travel time for axial flow across the seal and the minimum possible
travel time for flow in the rock mass under the assumption that the seal and
rock are homogeneous and isotropic porous media. These calculations are
made in Table 8.5 using Equations (7.30) and (7.31) together with known or
conservatively estimated values of hydraulic conductivity and effective po-
rosity. Based on the calculations in this table, the travel time for the intact
cement is over 3000 times the value of 565 min. The minimum travel time
for intact limestone is more than 10 times the observed 565 min.
Table 8.5 Calculation of travel time for axial flow across seal and minimum possible travel time for flow in rock mass
around seal
Hydraulic Effective Seal length Difference in Travel time
conductivity porosity (Lscm) hydraulic head (s)
(cm/s) across seal
(H -He) (cm)
Seal (cement grout) 1 x 10- 10 (KJa 0.2 (Ns)b 10 1890 1.06 x 10 8

(N sL; / [Ks (H ~ H clJ)
(Equation (7.30)
Rock mass 3 x 10- 9 (Kr)e 0.02 (Nr)d 10 1890 3.53 x 10 5
(Escabrosa Limestone) (NrL; /[Kr(H-HclJ
(Equation (7.31))
'South and Daemen, 1986, p. 72; bassumed value; 'from packer tests before installation; dFreeze and Cherry, 1979, p. 37.
Discussion of testing at Oracle Ridge Mine 175
8.3.5 Dye injection and seal removal

As a final test, a water-soluble red rhodamine dye was introduced to the
injection zone and constant-head injection was continued for 48 days. At
the end of that period, the seal was removed by overcoring from the lower
end. Visual inspection revealed a concentration of dye along the seal-rock
interface at the top of the borehole. The staining suggests that significant
flow occurred along this portion of the interface.
8.4 DISCUSSION OF TESTING AT ORACLE RIDGE MINE
In the first steady constant-head injection test, which was performed 18

days after seal installation, water was visually observed issuing from the top
portion of the seal at a relatively high rate. In similar tests 2 weeks later, the
flow was greatly reduced. The reduction in flow (both injection and collec-
tion) continued throughout the testing campaign at a gradual rate. The high
initial flow rate from the top of the seal is consistent with a gap or relatively
permeable zone along the roof of the seal, resulting from gravity settling of
cement materials during initial curing. This phenomenon has been noted by
others (e.g. Kelsall et at., 1982).
The large initial reduction in flow rate was probably due to expansion
and hydration of the cement grout. The cause of the gradual decline in
injection and collection rates over time is not known with certainty. How-
ever, the decline does not appear to be due to changes in hydraulic head in
the rock mass. Most head changes in the rock mass would affect the injec-
tion and collection flows in an opposite manner. For example, if head in the
rock mass were to increase, the injection rate would tend to decrease but
the collection flow would increase. Thus it seems most likely that the effec-
tive conductivity of the seal or the rock, or both, slowly decreased. A slow
decrease in seal hydraulic conductivity would be expected as a result of
continuing cement curing (e.g. Mindess and Young, 1981; Neville, 1981).
The flow from the top of the seal in the first constant-head test indicates
that, at least initially, this interfacial region was a dominant flow path. In
the dye-injection test the concentration of staining in this same interfacial
area suggests that this zone was a significant flow path throughout the
testing period. Together, the results of the tracer, dye-injection and initial
constant-head tests indicate that the interface at the top of the seal was a
significant, high-velocity flow path.
Based on the axisymmetric analysis of the steady test, the seal has an
effective hydraulic conductivity of 4 x 10- 10 cm/s. This conductivity, suffi-
ciently low for most sealing purposes, is somewhat higher than the conduc-
tivity of the intact cement grout, which is about 1 x 10 -1 0 cm/s (South
and Daemen, 1986). The higher conductivity may be due to leakage along
the top of the seal. The hydraulic conductivity of the seal, based on the
axisymmetric analysis of the steady test, probably is the most accurate

estimate of the property to come from the tests. This contention is based on
the steady test being the test with fewest potential sources of error and
producing consistent results over an extended testing period. Also, the
axisymmetric model is a more realistic representation of the nature of the
test than the other model considered (i.e. the one-dimensional model), as it
accounts for flow in both the seal and rock mass. Further, the test appears
to conform closely to the axisymmetric model (as shown in Figures 8.4
and 8.5).
During the three steady constant-head testing seasons, the seal conductiv-
ity based on one-dimensional analysis using injection rates averaged
2 x 10- 9 cm/s. Seal conductivity based on one-dimensional analysis of col-
lection rates averaged 1 x 10- 9 cm/s. Assuming that the axisymmetric
model reasonably describes the test, these may be taken as upper limit
values.
A very similar value for the hydraulic conductivity of the seal was ob-
tained by the one-dimensional analysis at different injection heads. This is
consistent with Darcy's law and suggests that injection pressures were be-
low levels which might open fractures or alter the conductivity of the seal or
rock mass.
The values of seal conductivity obtained from the head-buildup test
(5 x 10- 9 cm/s for the one-dimensional model and 6 x 10- 9 cm/s for the
axisymmetric model) are in reasonable agreement with those obtained from
the steady test. However, these values are considered less reliable because
only one test was performed and because of uncertainty over the compres-
sive storage of the collection zone.
The specific storage of the seal is determined from the one-dimensional
analysis of the head-buildup test to be 7 x 10 - 5 cm -1 and 1 x 10 - 4 cm - 1
from the axisymmetric analysis. Based on the diffusivity obtained from the
transient constant-head test and a value for Kg obtained by one-dimen-
sional analysis of the final collection flow rate in the transient test, the
specific storage of the seal is 1 x 10- 7 cm -1. These values for SSg seem high
for the grout seal, as suggested by calculations made by Greer and Daemen
(1991, Appendix L). In those calculations the specific storage of the grout
was calculated using the constitutive expression of van der Kamp and Gale
(1983) and representative material properties. The calculated specific stor-
age was about 3 x 10 - 8 cm - 1. Discrepancies of several orders of magnitude
are common in values of specific storage calculated from field tests. Addi-
tional testing would have been necessary to determine this property more
precisely.
The conductivity of the rock mass is determined from the axisymmetric
analysis of the steady constant-head test to be 3 x 10- 10 cm/s. This agrees
reasonably closely with the conductivity determined in injection tests before
seal installation. A slightly lower value (1 x 10- 10 cm/s) is obtained by the
axisymmetric analysis of the head-buildup test. The specific storage of the
Discussion of testing at Oracle Ridge Mine 177
rock mass is 2 x 10 - 7 cm or smaller, as determined in the axisymmetric

analysis of the head-buildup test. Using the expression of van der Kamp
and Gale and representative material properties, Greer and Daemen (1991,
Appendix L) calculated the specific storage of the rock mass to be about
1 x 1O- 9 cm- 1 .
According to the axisymmetric analysis of the steady test, the hydraulic
head in the rock mass at the axial and radial boundaries is about 3200 cm.
This compares reasonably well with the head values ranging from 1600 to
1900cm recorded by one of the two piezometers at the elevation of and
about 1.6 m from the seal. The second piezometer, on the opposite side of
the seal, consistently recorded heads of less than 500 cm. Comparable head
variations for closely spaced piezometers in rock have been noted by others
(e.g. Sowers, 1976).
The injection system maintained constant head in the injection zone
satisfactorily throughout testing. Measurement of injection rates was ad-
equate for long (i.e. a week or longer) injection intervals between readings.
The system could have been improved by providing additional ports in the
bulkhead of the injection packer to allow for circulation of a tracer. This
provision would have eliminated the need to temporarily halt injection and
remove the packer to introduce the tracer at the beginning of the travel-
time test. The injection system was not suitable for measuring rapidly
changing flow rates in a transient constant-head test.
The collection system for steady constant-head testing evolved over the
course of the testing program. The systems used during later testing were
quite satisfactory for the accurate measurement of collection volume for test
intervals from a few days to about 2 weeks. The system was not adequate to
measure accurately flow rates in a transient constant-head test.
The collection system for the head-buildup test was generally adequate.
However, exclusion of all air from the collection zone and tubing was
difficult if not impossible to achieve. Additionally, because of the magnitude
of the expected pressure buildup during the course of the test, a pneumatic
packer was used to form the collection zone instead of the mechanical plug.
The design of the packer made necessary a collection zone about 25 cm in
length. The collection zone required 46 days to approach equilibration. A
smaller collection zone, made possible by a more compact packer design,
would have equilibrated sooner. With a pneumatic packer there is always
the possibility of leakage of inflation gas into the collection zone, which
could result in error in the measured buildup curve. No leakage is known to
have occurred in the test.
In analysis of the head-buildup test the calculated values of Ks and Sss
depend on the compressive storage, Sc. The compressive storage was ob-
tained in the laboratory in a steel pipe and varied by a factor of five over
the pressure range of the collection zone in the field test. The variation may
have been due to the nonlinear compliance of the packer gland or possibly
to the presence of air in the system, though every effort was made to
eliminate air. The compressive storage could have been determined in the
field by injecting a known volume of water into the water-filled collection
zone and measuring the resulting rise in pressure. The volume of water
injected divided by the rise in pressure (expressed as pressure head) is Sc.
The laboratory determination of Sc might have been less accurate than a
field determination because the compliance of the laboratory collection
system (i.e. pipe, packer and tubing) did not exactly duplicate that of the
field test. Use of a smaller injection head to reduce the magnitude of build-
up in a test may help minimize effects due to variation in Sc. With the
collection system on the lower side of the seal, it was difficult to remove all
air from the collection zone. Since the presence of air can significantly affect
the shape of the buildup curve (Gale and Raven, 1979), the test might have
been improved by establishing the collection zone for this test on the upper
side. The collection system for the tracer test seemed satisfactory. The circu-
lation system provided adequate mixing in the collection zone prior to
sample removal.
Pneumatic packers were used in the pre-installation injection tests, to
form the injection zone in all seal tests, and to form the collection zone in
the head-buildup test. Packers were also used to isolate the tip sections of
the two piezometers. Leakage of packer inflation gas into the collection
zone in the head-buildup test, into the piezometer tip section, or into the
pressurized zone between injection test straddle packers, can lead to erron-
eous pressure or flow rate measurements. Leakage of packer inflation gas is
not known to have been a problem in the Oracle Ridge Mine tests. How-
ever, such leakage has been a significant problem in other tests (e.g. Greer
and Daemen, 1991). Inflation of packers with water rather than gas may
reduce leakage. Another alternative which deserves consideration is the use
of mechanical packers. A simple mechanical plug was used successfully to
form the collection zone for all but the head-buildup test at Oracle Ridge
Mine. The plug provided a waterproof seal for pressures up to about
50kPa.
8.5 SEAL TESTING AT SUPERIOR, ARIZONA
This section describes results of in situ hydraulic performance test of cement

seal installed in vertical boreholes. Three vertical test holes of diameter
15 cm were cored to a depth of 7 m in the densely welded Apache Leap tuff,
near Superior, Arizona. Inclined holes (5 cm diameter) were drilled at 55 0
from the ground surface to intersect near the bottom of the vertical holes
and to provide access to beneath the seal (Figure 8.10). The test method and
instrumentation were similar to those used at the Oracle Ridge Mine.
Fuenkajorn and Daemen (1991a, 1992a) describe the mechanical and pet-
rographical properties of Apache Leap tuff. Hydraulic conductivities of the
tuff around the vertical and inclined boreholes were determined using a
Seal testing at Superior, Arizona 179
Injection line
Flush line _ _ _./
Vertical borehole
T E
...~
Injection stand--lP'!'Io"III
o 0.
Sanded cement
Fig.8.10 Test boring at Superior test site.
straddle packer to locate an intact zone for emplacement of the seal in the
vertical borehole. Results from the constant-head tests indicated that the
hydraulic conductivities of intact zone were about 10- 11 cm/s. The hydrau-
lic conductivities of the fractured zone ranged from 10 - 4 to 10 - 6 cm/s.
8.5.1 Borehole seal testing

Prior to installing a seal in the field, a full-scale laboratory model was
constructed to aid determination of the optimum cement grout plug instal-
lation techniques. Cement plugs were emplaced with a custom-built dump-
ing bailer both under water and without water present. These trials
indicated that the installation of cement seals with standing water in the
borehole can cause significant piping channels along the plug and borehole
wall interface, creating pathways for preferential flow. Cross-section of the
plugs after curing revealed dark streaks as open channels, confirming the
effect at the interface. The piping effect was recognized for seals installed
with any level of standing water present. In addition to piping, significant
mixing of the cement with the water was evident. This mixing could notice-
ably increase the porosity and permeability of the seal and could weaken
the seal.
Based on the laboratory results, the in situ seal was installed in dry hole.
The cement was lowered into the hole with the bailer to minimize disturb-
ance of the mixture and released above the previously installed injection
stand cemented into the bottom of the test hole (Figure 8.11).
Collection zone _ _ t=,-:::::-T-...J

--t--'--.
Collection tube port ---+---'
Cement borehole plug
0.15 - 0.25mm diameter sand
0.25 - 0.43 mm diametersand
Porous stone
Flush Itne _ _..f
Fig.8.11 Installed cement borehole seal, injection test system underneath the seal
and collection zone above the seal.
The inclined hole allowed for access to both ends of the seal for hydraulic
testing without compromising the seal performance by having to penetrate
the seal with hydraulic test lines. This allowed injection and collection of
fluids on either side of the seal and hence the measured flow rates could be
verified through a balance check between the inflow and outflow. A series of
one-dimensional constant-head tests and transient tests was performed over
a 2 year period. The instrumentation and methods of testing and analyzing
the data were similar to those used at the Oracle Ridge Mine (section 8.3).
The hydraulic conductivity of the cement seal was confirmed through four
steady-state and three transient hydraulic tests to be of the order of
10 - 10 cm/s. Detailed description of the results is given by Crouthamel
(1991).
8.5.2 Discussions of test results

The hydraulic conductivity values obtained from the seal in welded tuff at
Superior site are approximately one order of magnitude lower than those
Recommendations for resting in situ borehole seals 181
obtained from the Oracle Ridge Mine. This may be due to the combinations
of the following factors.
Control of the seal installation at Superior site was made easier through
shallow boreholes (4.5 m vs. 32-40 m). This allowed more precise and
rapid placement of seal materials and more precise control of instrumen-
tation.
Hydraulic access to both ends of the seal was available in the tests perfor-
med at Superior. This allowed verification of the seal permeability through
the collection outflow and by monitoring and duplication of equipment at
both ends. This minimized the interference of pneumatic packers and the
problems associated with packer leakage and compressibility.
All cement seals at Superior site were placed in dry holes. Standing water
in test holes at Oracle Ridge Mine site probably had a negative impact
on the quality of seals.
The hydraulic conductivities of cement borehole seals at Superior site were
about an order of magnitude higher than those tested in the laboratory
(Chaper 2). This was probably due to the following factors.
Installation procedure. This related primarily to the mixing duration and
temperature and humidity during installation, which affected the swelling
capacity and porosity of the cement.
Conditions of the borehole wall. The borehole wall of the laboratory
specimen was clean, smooth and free of fractures, and hence resulted in a
high interaction pressure between the seal and rock, and a good hydraulic
bond at the interface. For the in situ testing the seal installation in a rough
borehole and near fractures might result in a lower interaction pressure
and a poorer hydraulic bond at the interface (Fuenkajorn and Daemen,
1986b).
Injection pressure. Laboratory test results (Chapter 2) indicated that in-
creasing the injection pressure resulted in a decrease of borehole seal
permeability. The injection pressures for in situ testing, usually less than
0.5 MPa, were significantly lower than those in the laboratory, which
usually exceeded 5 MPa.
8.6 RECOMMENDATIONS FOR TESTING IN SITU

BOREHOLE SEALS
This section summarizes recommended procedures for in situ testing of seals

in boreholes, shafts, ramps, tunnels, drifts or other underground excava-
tions. These recommendations are based on borehole seals with a length-to-
diameter ratio of about one. Full-scale testing of seals is likely to be affected
by several deviations from this reference case. Real borehole seals are likely
to be considerably longer and may be amenable only to one-side testing.
While conceptually the test procedures should be applicable to much larger

seals, such as seals in shafts, the practical implementation of such testing
will be difficult. Whenever reference is made to packer testing in the sub-
sequent paragraphs, it is implied that dams or bulkheads are likely to be
needed for testing larger seals. As the seallength-to-diameter ratio increases,
hydraulic connection between the two ends of the seal weakens (which, after
all, is the intent of seal design), thus making testing more difficult. Negative
results, i.e. inconclusive results, from testing such seals in accordance with
the recommendations that follow will still be valuable, in that they will
indicate the absence of pervasive or throughgoing highly preferential flow-
paths.
Before a seal is installed, the borehole should be cleaned thoroughly to
remove debris and drilling mud or sediment from the walls. The general
condition of the rock and the width and spacing of fractures should be
determined from drilling core and/or downhole videotape or photographs.
The hydraulic conductivity and, if practical, the specific storage of the rock
mass should be determined by means of packer tests. In such tests the tested
intervals should be the planned length of the injection zone or less and
should overlap. Highly conductive intervals found in this process should be
grouted and then retested.
Hydraulic head conditions in a saturated rock mass should be assessed.
As a minimum, the interval to be sealed should be shut in or isolated by
means of packers and the equilibration pressure determined. If practical,
piezometers should be installed at seal depth in nearby boreholes and
monitored before seal installation and throughout testing. Cross-hole tests
(Hsieh and Neuman, 1985a, 1985b) may be used to assess the large-scale
conductivity of the rock mass and the degree of interconnectiveness of the
fractures. It must be noted that the placement of the seal may possibly alter
the hydraulic properties of the rock mass near the seal. Such altering could
result from stresses imposed by a swelling seal, from a skin created by the
seal material, or from penetration of the seal material into fractures in
the rock.
For testing with the two-side configuration, as a minimum the steady
constant-head and tracer travel-time tests should be conducted. The steady
test is simple to perform. Its results are readily analyzed using either the
one-dimensional or axisymmetric models. The steady tests should be con-
ducted at two or three injection heads to permit comparison of results. The
tracer test is also simple to perform and may indicate unambiguously
whether or not a high-velocity flow path exists through or around the seal.
If practical, the head-buildup and transient constant-head tests also should
be performed. Specific storage may be determined only from these (or other
transient) tests. Accurate measurement of rapidly changing velocities is the
greatest difficulty in the transient constant-head test. Transient injection
and collection rates both should be monitored, if possible.
For the one-side configuration, the steady constant-head test should be

conducted, as a minimum. The test should be conducted at two or three
injection heads to permit comparison of results. The tracer test cannot be
performed from this configuration unless the required instrumentation can
be positioned on the inaccessible side and appropriate communication es-
tablished across the seal. If practical, the transient constant-head and head-
buildup tests should be performed for the one-side configuration. However,
with the one-side configuration, usually only the injection rate may be
monitored in the transient constant-head test. Greer and Daemen (1991)
found that the variation in shape of the transient injection type curves for
the axisymmetric analysis is generally minimal; thus, matching in an
axisymmetric analysis based on the transient injection data alone is difficult.
In most forms of the head-buildup test performed with the one-side con-
figuration, the injection head, H, and the formation head, H b , are equal,
which makes axisymmetric analysis extremely difficult (Chapter 7).
The minimum time required for the steady test is of the order of a few
days to several weeks or months per injection pressure, depending on the
seal and rock conductivities. Longer times are needed when these conduc-
tivities are low because flow rates are small and long periods (a few days or
weeks) are required to inject or collect an adequate volume of fluid for an
accurate determination of flow rate. Also, time is required for transient
effects to subside and for flow to approach a steady rate. For the head-
buildup test, the time required depends upon seal and rock conductivity
and storage values and also on the compressive storage of the collection
zone. By minimizing the volume and flexibility of the collection zone, this
latter quantity may be minimized and testing time reduced.
Removal of the seal by overcoring is preferable to coring. Overcoring
preserves the interface, which may be a significant flow path. Injection of
dye prior to seal removal successfully marked a major flow path along the
interface of the seal at Oracle Ridge Mine. In cases in which coring is
performed rather than overcoring, video-logging after the coring may reveal
significant information about the interface area. At Oracle Ridge Mine
inspection of the removed seal and overcore was limited to visual examin-
ation. Density, strength, porosity and permeability tests and chemical
analyses could also be performed to assess the quality of the cement grout.
ACKNOWLEDGEMENTS
ledged.
CHAPTER NINE
Sealing Boreholes
in Rock Salt
J. C. Stormont and R. E. Finley
9.1 INTRODUCTION
Effective sealing of penetrations (boreholes) in geological materials is be-

coming increasingly important as it is recognized that groundwater resour-
ces need to be protected from potential contamination. In the undisturbed
geological system, potential contaminants are separated from freshwater
aquifers by low-permeability aquitards including evaporite deposits such as
bedded and domal halites. Improper plugging, sealing or abandonment of
boreholes which penetrate aquitards can allow the borehole to act as a
conduit between the contaminated and freshwater regions. Borehole seal
performance has been recognized as a potential environmental concern by
numerous researchers (e.g. USEPA, 1977, 1987; Freeze and Cherry, 1979;
National Research Council, 1985; Smith and Browning; 1993; Smith et at.,
1993).
There are thousands of boreholes which penetrate rock salt (halite) and
other evaporite deposits. Because 25 % of all continental areas are underlain
by evaporites (Blatt, Middleton and Murray 1980), many deep boreholes
often penetrate rock salt formations. Exploration and production wells fre-
quently penetrate rock salt formations because evaporite deposits can act as
natural traps for oil and natural gas. Rock salt deposits are being used for
or are under consideration as host rocks for the terminal storage of radioac-
tive and hazardous wastes. In addition, these deposits can be used to store
petroleum, liquified natural gas and compressed gas. These facilities have
numerous penetrations (boreholes and shafts) associated with them. All
boreholes which penetrate rock salt deposits require sealing to satisfy regu-
latory or operational requirements. The real or perceived failure of these

Introduction 185
seal systems to isolate contaminants from groundwater supplies can result
in costly litigation and remediation, loss of public confidence and support
and destruction of groundwater resources.
Rock salt formations have naturally low permeability, which makes them
ideal locations for seals in boreholes which penetrate them. Furthermore,
rock salt has unique properties and behaviors (creep, healing and reconsoli-
dation) which allow the construction of seals which may be effective for very
long periods of time. On the other hand, rock salt can become damaged by
the creation of the borehole. This damaged zone can act as a preferential
pathway around the seal. It is therefore necessary to develop an under-
standing of rock salt properties and behaviors to design and construct
effective seals.
It is important to view borehole sealing in evaporites as a system consist-
ing of the host rock, the rock - seal interface and the seal material. An
effective seal design should include a careful evaluation of the behavior of
each component of the system. The seal system can be designed so that the
predicted response satisfies the required performance. It is only through
such a rigorous engineering design process that high-quality, predictable
seal systems can be constructed.
In this chapter we discuss some of the factors which affect the perform-
ance of borehole seals in salt. These factors cover a wide range of disciplines
and are closely intertwined. Important factors relevant to all sealing appli-
cations in salt and evaporites include:
Host rock properties;

strength and failure;
creep rates;
dilatancy and permeability increases;
healing and permeability decreases;
Seal material characteristics;
permeability and porosity;
strength;
chemical compatibility with the host rock and hydrologic enriroment;
emplacement techniques;
Seal system performance;
interaction between seal and creeping host rock;
seal system strength;
seal system permeability.
The remainder of this chapter discusses the design and performance of

each of the system components. Although we focus exclusively on sealing
boreholes, many of the discussions can be extrapolated to sealing other
penetrations such as shafts and tunnels.
186 Sealing boreholes in rock salt
9.2 GENERAL PRACTICE IN SEALING BOREHOLES

IN ROCK SALT
9.2.1 Applications
Some of the important applications which require sealing of boreholes
which penetrate halite deposits are briefly described below.
Nuclear waste disposal

The Waste Isolation Pilot Plant (WIPP) near Carlsbad, New Mexico, USA,
is planned as the first mined geological repository for transuranic (TRU)
wastes generated by the defense programs of the United States Department
of Energy (DOE). The WIPP Facility is constructed at a depth of approxi-
mately 650 m and is located within a thick sequence of bedded halite with
thin interbeds of anhydrite and clay. Sealing systems will be designed to
limit the release of liquid and gaseous waste to the accessible environment
by preventing boreholes and shafts from becoming preferred pathways
(Stormont, 1987; Nowak, Tillerson and Torres, 1990). An ongoing experi-
mental program to develop seal systems and designs emphasizes in situ
testing.
Performance requirements for WIPP penetrations are described by
Nowak, Tillerson and Torres (1990). Currently, the WIPP large-scale shaft
seals are required to be constructed with a short-term (less than 100 years)
permeability equivalent to 1 x 10- 18 m 2 ( ~ 1 microdarcy). This permeability
is achieved largely by high-quality concrete and bentonite seals. The
long-term seal performance (beyond 100 years) is expected to achieve per-
meabilities similar to the surrounding halite 1 x 1O- 22 m 2 ) by using re-
consolidated granular salt as the primary seal material. In addition to these
permeability performance requirements, the seals are expected to be struc-
turally sound and readily constructable with reasonably available technol-
ogies and equipment.
Germany is also investigating the disposal of nuclear waste in under-
ground evaporite deposits. Plans for sealing penetrations associated with
disposal facilities are based to a large extent on sealing experience for gas
storage projects in potash. These projects have developed sealing designs
similar to the radioactive waste disposal programs in the United States and
are reported elsewhere (e.g. Sitz and Forster, 1989). These designs typically
call for composite seals consisting of concrete, asphalt or bitumen, and
clays.
Oil and gas

The United States Environmental Protection Agency (USEPA) (1987) esti-
mates that there are approximately 1.2 million oil and gas production wells
General practice in sealing boreholes in rock salt 187
in the United States alone. In addition, there are an unknown number of

brine injection wells used to inject oil field brine produced during produc-
tion activities. Many of these wells penetrate halite and evaporite deposits,
most notably in the Permian Basin region of West Texas and Eastern New
Mexico and along the Gulf Coast region. It has been emphasized recently
that oil and gas wells that are improperly plugged or abandoned can pose
significant environmental problems (Smith et aI., 1993). Recent newspaper
articles have raised concern over some notable borehole seal failures (Suro,
1992). These wells typically penetrate through a number of formations
which usually include a low-permeability cap rock which serves as a trap
for the petroleum and gas. In many areas potable water is found in aquifers
overlying these low-permeability traps. In its reports to Congress (1977,
1987), the USEPA has suggested that these wells, if improperly plugged, can
serve as vertical conduits through which contaminants can be transported
to the groundwater resources or to the ground surface. Currently, each state
has the authority to regulate abandonment requirements for these wells
within broad guidelines developed by the USEP A. Typical requirements
include sealing of any water producing zones by the use of cements. These
wells typically are cased to depths below potable water and requirements
for annular grouts between the casing and the strata vary from state to state
and also depend on the date the well was completed.
Performance requirements for abandoned oil and gas wells are currently
regulated on a state by state basis within broad guidelines defined by the
USEPA. Typically, abandonment requirements for oil and gas wells are
defined not on a performance requirement as is typical for the nuclear waste
disposal industry, but rather in terms of the length of cement plug within
the boreholes at locations specified relative to stratigraphic interfaces, water
bearing zones and the ground surface. Specific requirements for perform-
ance in terms of permeability and chemical compatibility appear to be
lacking. However, plugs can be 'tested' by applying axial loads to cement
plugs using the dead weight of a drill rig and drill string.
Strategic petroleum reserve

The US Strategic Petroleum Reserve (SPR) was originally established to
provide a 500million barrel reserve of petroleum (since expanded to
750 million barrels) stored in salt caverns on the US Gulf Coast in the states
of Louisiana and Texas. These caverns include solution-mined openings
and a non-working salt mine. Six storage facilities are on-line, with approxi-
mately 670 million barrels of petroleum stored (Ehgartner, 1991). Access to
these facilities include boreholes and raise bores up to 2 m in diameter.
The SPR storage caverns are planned to be taken 'off-line' as they age.
Each cavern will then be refilled with brine and sealed following an as yet
undefined plan. These seals must restrict the upward flow of brine, which is
potentially oil-tainted, toward overlying environments. Currently, the SPR
is investigating cementitious, epoxy-based and granular salt sealing

materials.
Performance requirements for the SPR access boreholes are not clearly
defined; however, it is probable that they will have to at least meet EPA
standards typical of similar oil and gas production wells which are regu-
lated on a state by state basis. Standard cementing techniques with per-
formance measures based solely on cemented lengths may be required.
Others
Numerous other facilities include penetrations through halite and evaporite
formations which may require sealing. These include salt and potash mines,
European radioactive waste disposal facilities, solution mining penetrations,
energy storage (compressed air and gas, and liquified natural gas) and
hazardous waste disposal. The unique nature of each of these facilities
defines the sealing needs and strategies required, although the concepts
discussed here have applicability to all such sealing needs.
9.2.2 Existing regulations and conventional practice

There is minimal design guidance for borehole seals from existing regula-
tions. States and governmental agencies normally define plugging and aban-
donment requirements in terms of length and location of materials in the
boreholes. Typically, plugs are required for abandoned wells at production
zones and/or freshwater zones. Plug lengths of up to 100 m may be re-
quired. Cementitious materials are usually used for these plugs. The base-
line practice for sealing boreholes in salt or other rocks has been open-hole
cementing of abandoned wells performed by the petroleum industry. Open-
hole cementing is typically carried out to plug abandoned wells, to assist in
directional or follow-on drilling, to minimize or restrict lost circulation, or
for isolation of specific zones in the borehole to restrict water inflow or for
emplacement of instrumentation.
The recent investigations into the use of halite as a storage and disposal
medium has focused much more attention on the factors that affect the
quality of a borehole seal. The long-term or permanent performance-based
requirements for these projects have necessitated careful study of factors
and mechanisms affecting seal performance. Depending upon design goals
such as seal system permeability and longevity, alternatives to conventional
open-hole cementing may result in a more effective seal.
The USEPA (1987) has identified a need for more stringent evaluation of
borehole seal emplacement. The typical tests performed to assure seal qual-
ity include a careful and complete review of construction data and pressure
tests of the seal system as a check for cracks and leaks in the casing and
seal-rock interface. These tests are conducted at the top of the plug and are
intended to assure immediate performance. The nature of the plug construc-
General practice in sealing boreholes in rock salt 189
tion (placement across production and freshwater zones) suggests that the
standard practice for these tests may not identify seal problems until per-
formance is compromised.
It is inevitable that there will be increased future regulatory requirements
for borehole seal performance to protect natural resources. While these
regulations may be motivated by the few exotic applications such as nuclear
waste disposal, improving the quality of seals for the thousands of aban-
doned oil and gas wells may have the greatest effect on environmental
quality. It is likely that these requirements will focus on the longevity of the
borehole seal.
9.2.3 Sealing options

Different options exist for sealing abandoned wells and penetrations
through halite and other evaporites, depending on the sealing objective.
Options include short-term, long-term and permanent seals. Short-term
seals include temporary seals used to effect a change in the downhole
condition, such as grouting a water producing zone until a longer-term
option is required. Long-term seals are considered to be those seals em-
placed using conventional equipment and materials, other than granulated
salt or evaporite. Permanent seals are those emplaced using granular ma-
terials whose chemical compatibility and permanence can be assured. Each
borehole should be evaluated independently, based upon the combination
of these factors.
Short-term seals
Short-term seals are emplaced for the temporary control of deleterious
conditions encountered in the borehole or underground penetration. The
techniques and equipment used for these emplacements are discussed subse-
quently in this chapter. These seal techniques include grouting and typically
incorporate cementitious or clay materials as slurries. The removal of casing
or borehole wall treatment is usually not called for except under extreme
situations. These types of seals have only limited longevity as true seals, due
to the nature and control associated with the emplacement techniques.
Interbeds of anhydrite and clay are common occurrences in most evap-
orite deposits. Interbeds can play either a major or minor role in sealing,
depending on their extent, their fracturing, whether they are significant
water producers, etc. A careful sealing strategy will consider the location(s)
of such zones and will attempt to place permanent seals in regions away
from such interbeds. These interbeds may require short-term seals them-
selves during the course of emplacement of permanent seals in the halite to
minimize water production, which could have a negative effect on the em-
placement techniques used for the permanent seals. Standard practice for
temporary seals typically consists of sulfate-resistant grouts and concrete

plugs.
Long-term seals
Long-term seals are those seals emplaced using the equipment and tech-
niques described subsequently in this chapter. These seals also include those
composite seals designed for nuclear and hazardous waste isolation, which
include cementitious, polymer, clay and asphalt-based materials. The reason
for not including these seals in the permanent category is the uncertainty
associated with the long-term geochemical stability of the cementitious and
polymer-based materials, together with the mechanical stability of the as-
phalt-based and clay seal materials. Asphalts and clays are found naturally
in halite and other evaporite deposits; however, their mechanical stability in
the seal designs is based on the mechanical stability or containment pro-
vided by the cement and polymer-based seal materials. Therefore, none of
these materials should be considered permanent from a sealing standpoint.
Permanent seals
It is suspected that the common practices currently employed for sealing
boreholes in halite and evaporites do not or cannot satisfy permanence
requirements for sealing, especially for hazardous materials. These require-
ments do recognize the need for chemically compatible high-quality seal
materials, but use materials such as cements or others whose permanent
compatibility cannot be demonstrated. An alternative technique would be
to demonstrate the high-quality emplacement of clearly compatible
materials such as granulated salt for boreholes in halite to form a high-
quality (low-permeability, high-strength, chemically stable) permanent seal.
Permanent sealing techniques cannot be used for all boreholes in halite or
other evaporites. However, these techniques can provide an additional tool
when sealing of penetrations in these materials is required to be permanent.
9.3 ROCK SALT PROPERTIES RELEVANT TO

BOREHOLE SEALING
Borehole sealing in rock salt requires an understanding of the behavior of

rock salt. Open boreholes will close with time due to creep of the rock salt.
Creep closure of a borehole sealed with a relatively stiff material such as a
cement-based seal will result in an appreciable interface stress. The closure
will also compact granular seal materials if they are emplaced at low den-
sity. A zone of rock surrounding the borehole dilates in response to stress
changes during excavation. This damaged zone can be significantly more
permeable than the intact host rock or the seal material, and thus can be
Rock salt properties relevant to borehole sealing 191
the principal flow path through the seal system. Stress buildup due to
closure on a borehole seal can heal the damaged zone. Quantification of
these phenomena for specific borehole sealing applications may not be
possible or even necessary. However, an appreciation of these concepts is
required for defensible performance evaluations for seals in halite or other
evaporites.
9.3.1 Mechanical behavior

The mechanical behavior of rock salt is complex. Its response to load is
inelastic, nonlinear, time-dependent and loading rate-dependent. At low
stresses and fast loading rates, rock salt behaves as a relatively brittle
material and fails along a well-defined failure plane. At increasing confining
stresses and/or slower loading rates the behavior of rock salt becomes
increasingly ductile. At sufficiently great confining stress and/or slow load-
ing rates the response of rock salt is perfectly plastic. Rock salt also experi-
ences significant creep; that is, it deforms plastically under sustained
deviatoric loads. In the following discussion we will distinguish between the
relatively short-term response of rock salt, such as that immediately upon
excavation, and the long-term response due to creep.
Mechanical properties of rock salt

Even under test conditions which are relatively fast, the pronounced time
dependency of rock salt makes it virtually impossible to determine its truly
time-independent response and some time dependency is embedded in all
measured response. The most common short-term laboratory tests on rock
salt samples have been the conventional triaxial compression (CTC) test, in
which cylindrical samples are subjected to a constant confining pressure
while the axial stress is increased. All quasi-static CTC tests reveal a near-
zero yield stress, followed by a strain-hardening response. At low confining
pressures (0-3.5 MPa), rock salt tends to fail in a brittle manner; that is, at a
maximum load the sample develops failure surfaces and loses strength. The
unconfined compressive strength of rock salt falls between approximately
15 and 30 MPa for a wide range of bedded and domal salts (Hansen,
Mellegard and Senseny, 1984; Fuenkajorn and Daemen, 1988a, b; Desai and
Varadarajan, 1987; Wawersik and Hannum, 1980). Hansen, Mellegard and
Senseny (1984) found the average compressive strength of bedded salt to be
about 20% greater than domal salt. From microscopic observations on
samples which had been deformed during CTC test, Carter and Hansen
(1983) conclude that for confining pressures of 3.5 MPa and below, the
predominant mode of deformation is cataclastic flow with fracturing along
grain boundaries and shearing along cleavage planes.
As the confining pressure is increased, the failure becomes increasingly
ductile so that above about 3.5 MPa the material exhibits a perfectly plastic
to strain-hardening response for axial strains as great as 10%. By 14 MPa

confining pressure, deformation is dominated by crystal plasticity (Carter
and Hansen, 1983). The rate of loading or straining will affect the observed
stress~strain response. In general, decreasing the stress or strain rate de-
creases the confining pressure at which the overall sample behavior be-
comes ductile (Fuenkajorn and Daemen, 1988a).
With confining pressures up to about 20 MPa, the failure of rock salt can
be represented by a linear relationship between deviatoric and mean stresses
(Hambley, Fordham and Senseny, 1989; Stormont, Daemen and Desai,
1992; Desai and Varadarajan, 1987). The strength is dependent on the stress
path. In terms of the conventional Mohr~Coulomb strength criteria, the
effective friction angle for rock salt varies from about 50 for compressive
stress paths and 30 for extensional stress paths (Stormont, 1990c). At
greater confining stresses the ultimate shear strength appears to be a nearly
constant at 30 MPa during a conventional CTC test (Hansen, Mellegard
and Senseny, 1984; Farmer and Gilbert, 1981).
All CTC tests reveal that rock salt dilates as it is deformed at confining
pressures up to 20 MPa. For comparable amounts of axial strain, dilatancy
increases as the confining pressure is decreased. Limited tests reveal that
rock salt dilates under different stress paths and the response depends on
mean stress (Wawersik and Hannum, 1980).
Brazilian disk tests reveal that the tensile strength of rock salt is very
small, of the order of 1~2 MPa (Fuenkajorn and Daemen, 1988a, b; Hansen,
Mellegard and Senseny, 1984; Pfeifle, Mellegard and Senseny,1983). Hansen,
Mellegard and Senseny (1984) found bedded salt to have about a 10%
greater tensile strength than domal salt.
Although rock salt exhibits nonlinear inelastic response during loading,
the unloading response tends to be linearly elastic, and elastic properties
have been determined from unloading under different stress paths (Singhal,
1974; Wawersik and Hannum, 1980; Pfeifle, Mellegard and Senseny, 1983;
Hansen, Mellegard and Senseny, 1984; Desai and Varadarajan, 1987;
Fuenkajorn and Daemen, 1988a, b; Stormont and Daemen, 1992). Typical
elastic properties are Young's modulus of 30 GPa and Poisson's ratio of
0.35, essentially independent of the type of salt.
Rock salt creep

Rock salt creeps or deforms plastically under sustained deviatoric loads.
Creep is generally expected to enhance borehole sealing in rock salt be-
cause, as the borehole closes by creep, the interface between the seal and
salt is reduced and compressive stresses develop at the interface. The closure
will continue until the seal provides a backstress on the order of the
lithostatic state of stress. This will compact granular seal materials, presum-
ably decreasing their permeability and providing a tight interface for cern en-
titious seals.
There have been many attempts to develop useful models of salt creep.
These attempts have resulted in rheological, empirical and physical models.
Extensive summaries of the measurement and modeling of rock salt creep
can be found in Carter and Hansen (1983), Fuenkajorn and Daemen
(1988a, b) and Munson and Wawersik (1992). Developing widely applicable
constitutive models for salt creep is hindered by differences in test condi-
tions and variations among the tested material, as shown by the differences
in reported strengths and creep rates. Factors which may influence the
value of creep parameters include sample origin, history and size. Impuri-
ties, including clay and anhydrite, clearly affect the derived creep
parameters. The study of Fuenkajorn and Daemen (1988a, b) is of particular
value because they used tests on a single source of bedded rock salt to
assess the ability of 20 different constitutive models to represent borehole
closure. The models ranged from simple rheologic models to complex mech-
anistic-based models. Model parameters were derived from one set of ex-
periments and used to predict the behavior under different conditions. They
concluded that none of the models was sufficiently robust to be applicable
under the entire range of test conditions (stress and strain rates). They
attributed this conclusion to the inability of the constitutive models to
describe the different mechanisms which dominate at different stress and
strain rates and levels, including fracture propagation, plastic flow, healing
and dislodging of crystals.
Measured closure rates reported for boreholes and shafts which penetrate
rock salt vary widely. Unger and Howard (1986) measured borehole closure
with caliper logs in boreholes drilled through a deep overthrust belt salt in
Utah. They found that 'dirty' salt (containing appreciable amounts of silt-
stone) at 3300m depth closed to a diameter of 12.7 cm from 31.1 cm diameter
in only 14h. Clean salt at a similar depth did not close at all. Flak and
Brown (1990) ran caliper logs in boreholes more than 4000m deep in salt
formations in Colorado. The borehole was actually oversized through the salt
sections, apparently due to washouts. Subsequent caliper logging revealed no
measurable closures. Munson et al. (1992) measured closure of a 6.2 m diam-
eter shaft at WIPP. At a depth of 612 m about 75 mm of closure was meas-
ured after 3 years. Cook (1983) measured about 50 mm of closure of a shaft
through a halite formation after about 6 months at a depth of about 1000 m.
Results such as those cited above give rise to the question as to whether
any single constitutive model can account for the behavior of rock salt. As
complicated as many of these models are, many aspects of rock salt behav-
ior which are important for certain applications may not be included. For
example, most of the these models assume isovolumetric deformation
whereas it is well known that salt near excavations dilates or increases its
porosity (e.g. Borns and Stormont, 1989). Another shortcoming of most
creep models is that they do not account for the presence of pore fluids in
terms of the well-known effect of moisture on creep (e.g. Spiers, Urai and
Lister, 1988).
Although there is no universally accepted model of salt creep today, there

is considerable empirical evidence of the strong dependence of creep on
temperature and on some measure of deviatoric stress. These observations
are captured in a simplified model for the steady-state creep rate (B):
(9.1)
where A and n are empirical constants, ad is deviatoric stress, Q is activation
energy, R is the gas constant and T is temperature. From this relatively
simple relation we can draw the following inferences regarding creep of
boreholes.
Creep will increase with depth, due to both the increased stress and
temperature with depth. Infante and Chenevert (1989) predict that these
effects can increase closure in the first hour of drilling from 5% to 27% at
depths of 3810m and 4270m, respectively.
As a borehole creeps in on a seal material, the backstress which develops
will reduce the deviatoric stress, which in turn reduces the creep rate.
Increasing mud weight, essentially to provide backstress to reduce the
deviatoric stress, is used to limit closure of boreholes and maintain stabil-
ity in rock salt (Ungar and Howard, 1986; Infante and Chenevert, 1989).
The creep rate is very sensitive to the stress term due to the stress expo-
nent. Experimentally derived exponents vary by more than a factor of
two (Pfeifle, Mellegard and Senseny, 1983), resulting in closure rates
which vary by orders of magnitude (Kelsall and Nelson, 1983).
Given enough creep deformation, the salt may eventually fail or rupture.
Surface spalling of the workings of salt and potash mines are believed to be
caused by excessive deformation (Nair and Singh, 1974). However, there are
no reported observations of long-term borehole failures in salt to determine
the magnitude of this effect.
9.3.2 Rock salt permeability

The very low permeability of intact rock salt is its principal advantage as a
medium for hydrocarbon storage and waste disposal. Rock salt crystals are
essentially impermeable; tests on a single crystal of salt reported by Suther-
land and Cave (1980) and by Gloyna and Reynolds (1961) indicate a less
than 10- 21 m 2 and zero permeability, respectively. The very small porosity
of intact rock salt (typically 0.1 %) consists of brine-and/or gas-filled pores
with pressures near the lithostatic stress. The effective permeability of rock
salt under these conditions is very small and difficult to measure. However,
rock salt dilates and may fracture around excavations including boreholes,
and the permeability of this zone dramatically increases. As with any rock
type and sealing application, the nature of the rock adjacent to the seal may
be the most significant factor in the degree of sealing that can be achieved.
It has long been recognized that rock salt experiences some type of
'damage' during the process of sample collection and preparation (e.g. Baar,
1977; Guessos, Ladanyi and Gill, 1988). This disturbance has been identified
as the principal cause of the relatively large permeabilities of as-received
core (Gloyna and Reynolds, 1961; Sutherland and Cave, 1980; Peach et al.,
1987; Stormont and Daemen, 1992). Hydrostatic stress reduces the perme-
ability with time, apparently to some very small, limiting value. Subsequent
hydrostatic unloading indicates that the majority of the permeability reduc-
tion is irreversible (Sutherland and Cave, 1980). A typical test showing the
healing effect of hydrostatic stress on a rock salt core is given in Figure 9.1.
Sample healing has been attributed to plastic flow along grain boundaries
by Sutherland and Cave (1980). Typically, the permeabilities of healed sam-
ples under hydrostatic confining pressure are below the resolution of the
test method (usually < to- 20 m 2 ). Measurements by Gloyna and Reynolds
(1961) indicate a much lower permeability for bedded salt, possibly due to
the different content of impurities which restrict flow or the lower porosity
of the bedded salts.
To confirm the low permeability of intact rock salt inferred from labora-
tory tests, high-resolution permeability measurements have been made from
the underground workings of the WIPP Facility. When the test region is
: At=10h
o 3 6 12 15
Fig. 9.1 Permeability vs. hydrostatic stress data from tests on core recovered from
WIPP Facility. (From Stormont and Daemen, 1992.)
well away from the existing excavation, brine permeability measurements

indicate that the formation has a brine permeability of between 10 - 20 and
10 - 21 m 2 , a porosity of 0.1 % and a pore pressure of up to 14 MPa
(Peterson, Lagus and Lee, 1987; Nowak and McTigue, 1987; Saulnier and
Avis, 1988). Because the porosity of undisturbed rock salt < 1%) is
saturated with brine under relatively high pressures, the gas permeability is
effectively zero (Stormont, 1990b; Stormont, Howard and Daemen, 1991).
9.3.3 The disturbed rock zone
Measurements of the disturbed rock zone

Rock salt dilates and may fracture around excavations including boreholes,
and the permeability of this zone dramatically increases. This region is
referred to as the disturbed rock zone (DRZ). A DRZ surrounding excava-
tions in rock salt was initially detected from numerous gas permeability
measurements made from WIPP excavations (Stormont, Peterson and
Lagus, 1987). In the test intervals which comprised solely of rock salt and
were within the first meter from an excavation, the interpreted permeabili-
ties varied from about 10- 13 to 1O- 19 m 2 . The interpreted gas permeabili-
ties rapidly decrease with distance from an excavation so that by one-half of
an effective radius away from an excavation, the gas permeability is im-
measurably small. These gas flow measurements imply that the DRZ is a
dilated, partially saturated zone which extends a limited distance from an
excavation. Dilation and desaturation of the rock salt surrounding WIPP
excavations have also been inferred from electrical resistivity measurements
and seismic tomography (Borns and Stormont, 1989) and from ultrasonic
velocity and attenuation (Holcomb, 1988).
To simulate excavation-induced damage, healed rock salt samples have
been subjected to deviatoric stresses in laboratory tests (Stormont and Dae-
men, 1992). For the complete range of confining pressures used (0-15 Mpa),
the rock salt samples strain harden, dilate and become increasingly per-
meable. The permeability increase is a result of a largely interconnected
pore structure which develops along dilated grain boundaries (Stormont,
1990c). These results imply that the dilation of rock salt, which has been
measured under a wide range of conditions (e.g. Hunsche, 1993), is accom-
panied by a dramatic increase in permeability.
Lingle et al. (1982) investigated drilling-induced damage in rock salt by
testing large cores under simulated downhole conditions. Holes of diameter
20 cm were drilled in cores 34 cm in diameter by 91 cm long. A damaged
region adjacent to the borehole was detected, measured by an increase in
permeability after drilling of two orders of magnitude. The extent of damage
adjacent to the 20 cm diameter holes was estimated at 5 cm into the rock, or
to a distance of 1.5 radii (R) from the borehole center.
An in situ test was conducted to quantify the DRZ surrounding boreholes

in rock salt by monitoring the hydrological response of a rock salt layer to
nearby drilling (Stormont, Howard and Daemen, 1991). An array of twelve
small-volume pressurized brine- and gas-filled test intervals located about
8 m from an underground room was first established. Their pressure re-
sponse was monitored with time prior to, during and after the drilling of a
nearby large-diameter hole (1 m diameter). Sometime later, gas and brine
injection tests were conducted in the boreholes. The emphasis of the
measurement and analyses was to quantify the changes in permeability as a
result of drilling. The response of both the gas- and brine-filled boreholes
prior to the drilling of the mine-by borehole is consistent with the assump-
tion that the formation is a very low-permeability, low-porosity, porous
medium with a significant pore (brine) pressure. Pressurized gas does not
flow from the monitoring boreholes out into the formation, indicating that
the effective gas permeability of the formation prior to the mine-by is very
small. The interpreted brine permeabilities and formation pore pressures are
in the range expected for undisturbed rock salt. During the drilling of the
large-diameter borehole, the pressure changes in the gas and brine monitor-
ing wells reveal the instantaneous creation of a partial saturated zone ex-
tending 1.25 R to 1.5 R into the formation. Beyond 2 R the gas well pressure
is unaffected by the drilling. Subsequent injection tests indicate the forma-
tion has a measurable gas permeability and porosity at 1.25 R, between
5 x 10 - 18 and 10 - 15 m 2 . At 1.5 R the formation is apparently still brine
saturated but gas can be injected into the formation and displaces brine.
Beyond 1.5 R there is no indication that the formation has any permeability
to gas. Experimental results are summarized in Figure 9.2 and confirm that
a damaged region forms coincident with drilling within about 1.5 R from a
borehole in rock salt.
Predictions of the disturbed rock zone

There are a number of approaches to predicting the extent and magnitude
of a DRZ in rock salt. One approach is to identify the regions surrounding
an excavation in which the stress state is sufficient to dilate or microcrack
the rock salt. A criterion based on a very simple model of frictional micro-
crack initiation and growth has been used successfully to delineate zones of
enhanced permeability about both drifts (Stormont, 1990a) and boreholes
(Stormont, 1990c).
The extent of the zone of increased permeability can also be predicted
using a constitutive model for the rock salt which realistically accounts for
dilation. An elastoplastic constitutive model accounting for strain harden-
ing, a non-associative volumetric response and unequal behavior in com-
pression and extension was used to simulate rock salt behavior surrounding
boreholes and drifts (Stormont, 1990c). The extent of the predicted dilated
zone compared well with the region of the formation, with an enhanced gas
10- 14
o Post-Excavation
o Gas Permeability
10- 16 Brine PermeabHity
1 Brine Permeability Interpreted
Approximate Assuming Test Interval Gas
~
:s 10- 18
~ '!I :' Saturation Displacing Formation Brine
l Boundary
~
,
0..
I 10-20 '-- -- --
r- --: - - !'~'~.p--- -- Brlne Permeability
!---..........
~ 0 ~.. , (Pre-Excavation
10-22
l t :
l
:
i
i-"'''"Gas PermeabHhy
immeasurably Smail)
4.0 _
Borehole
Cenlerline ----r ----r ----- ----- ----- ----- ----- Range of
E Pre-Excavation
:::J
~~ 3.0 i i Pressures
._ co R ~ - l - l --..."....---
=0.. . ::
5-::E
w~
. .6
:
2.0
iii!!!
~:::J
Borehole
Wall
.
.!!1
r:::!!!
m
-0..
1.0 Post-Excavation
Borehole o o Gas - FRied Test Interval
~ 1 Brine - Filled Test Interval

0.0
0.0 1.0 2.0 3.0 4.0 5.0
Distance from Center of Mine by Borehole (Distance = rlR)
Fig.9.2 Permeability and test interval pressure as a function of distance from borehole. (From Stormont,
Howard and Daemen, 1991.)
permeability measured in situ. These efforts provide no information on the

magnitude of the permeability within the DRZ.
Another approach seeks to quantify the permeability increase within the
DRZ, most conveniently by developing a stress-permeability relationship.
A phenomenological model of permeability changes in rock salt resulting
from damage has been developed by Stormont, Daemen and Desai (1992).
The model begins with the equivalent channel model, which gives the per-
meability as a function of the porosity, hydraulic radius (flow path aperture)
and empirical constants. The permeability model is recast in terms of easily
calculated or measured quantities, i.e., stress and strain.
k = cu</>s, (9.2)
where k is permeability, c and s are empirical constants, u is a stress
measure and </> is porosity. The model parameters were developed from
permeability measurements during quasi-static compression tests. Predic-
tions of permeability changes were made by incorporating this model into a
time-independent plasticity finite element code. This model was used to
predict the permeability changes surrounding a 0.95 cm diameter borehole
in rock salt at a depth of about 650 m below ground surface. The calculated
zone of dilation surrounding the borehole extends 0.15m into the forma-
tion, comparing well with that inferred from field measurements. The pre-
dicted porosity is quite small, an order of magnitude less than that inferred
from the field measurements. As a consequence, the predicted permeability
is well below the measured values. This model has been extended to include
the effect of the pronounced time dependency (creep) of rock salt (Stormont
and Fuenkajorn, 1994). The time-dependent behavior of rock salt has very
little impact on the predicted permeability changes because of the very
strong dependence of permeability on dilation. Because creep tends to re-
duce the deviatoric stresses, the greatest potential for dilation occurs im-
mediately upon excavation. This result has been confirmed by
measurements which indicate that, after the initial increase in permeability
following excavation, there is very little change in the extent or magnitude
of the DRZ (Stormont, Howard and Daemen, 1991).
9.3.4 Healing of rock salt

A key design issue for sealing in rock salt is that under certain conditions
the DRZ will eventually 'heal', or return to a condition comparable to its
pre-disturbed state. When a relatively stiff inclusion (such as grout or con-
crete immediately after placement and granular salt after it consolidates
appreciably) is located in an opening in rock salt, the tendency of the rock
to creep will reduce the deviatoric stresses in the vicinity of the inclusion.
These stresses are expected to reverse the disturbance (including a decrease
of permeability) in the adjacent rock by literally forcing the rock back
together.
The mechanisms of healing are presently not well understood. Healing

has been attributed to plastic flow along grain boundaries (Sutherland and
Cave, 1980) and may be related to fracture healing in rock salt (Costin and
Wawersik, 1981) or granular salt consolidation (Zeuch and Holcomb, 1991).
Healing is probably a function of mean stress, deviatoric stress, stress rate,
time, moisture content and clay content.
Healing tests under hydrostatic stress conditions have been measured by
Stormont and Daemen (1992). As-received cores served as the 'damaged'
samples. Initial permeability measurements were made on the samples and
were repeated as a function of time. The test results are summarized in
Figure 9.3. For clarity, only three permeability values are given for the
hydrostatic loading portion of each test: the initial permeability at 2.4 MPa
hydrostatic stress, the permeability immediately after the hydrostatic stress
is increased to 14.5 MPa and the final permeability after the stress is held at
14.5 Mpa for 10 h or more. At the initial hydrostatic stress state of 2.4 MPa,
the permeabilities of all samples fall in the range between 10 -1 7 and
10- 18 m 2 . The increase of the hydrostatic stress from 2.4 to 14.5 MPa causes
an immediate decrease of the permeability by about 50%. The permeability
typically decreases by more than four orders of magnitude in less than 24 h
under a sustained 14.5 MPa hydrostatic stress. In fact, in all but three cases
no pressure changes were detected for some time and the final permeability
given represents a conservative upper bound or maximum value.
10',6
10.17
10.,8
1
~
i
10.,9
~
.r 10.20
10. 21
10.22 L----'_----'_........L_--..l 10 22
5 10 15 20 0 10 15 20 0 10 15 20
Hydrostatic Stress (MPa) HydrostatIc Stress (MPa) HydrostatIc Stress (MPa)

(I) (b) (c)
Fig.9.3 Permeability vs. hydrostatic stress data on rock salt cores. (a) Intial per-
meability at 2.4 MPa hydrostatic stress, (b) the permeability immediately after the
hydrostatic stress was increased to about 14 MPa and (c) the final permeability after
the stress was held for some period of time (typically overnight). (From Stormont,
1990c.)
Materials for sealing boreholes in rock salt 201
One of the largest uncertainties associated with healing is the effect of

deviatoric stresses on the healing process. Healing and damage probably
occur simultaneously, with the net effect of healing or damage determined
by the stress state. At low deviatoric and high mean stresses, healing domi-
nates. As deviatoric stresses are increased, less healing and more damage
will occur. At some point the combination of deviatoric and mean stresses
will be sufficient such that the net effect is to damage the rock.
There are limited data of healing under deviatoric stress conditions.
Some apparent healing was observed on rock salt samples subjected to a
large deviatoric stress overnight (confining stress = 7.6 MPa, deviatoric
stress = 40 MPa) (Stormont, 1990c). These results are not conclusive as
other tests under comparable deviatoric stresses did not indicate healing.
9.4 MATERIALS FOR SEALING BOREHOLES IN ROCK SALT
The following is a discussion of various candidate seal materials, including

granular rock salt, cement and concrete, bentonite and other materials. The
best possible seal material would return a borehole to a condition compar-
able to its undisturbed state within a predictable period of time and whose
long-term performance could be demonstrated. Further, the economics of
the sealing task require compatibility of seal materials and emplacement
techniques with the required performance. The emplacement techniques
used for seals in salt and other evaporites are perhaps as important as any
other sealing issue. The various sealing materials (cements and concretes,
bentonites, crushed salt, etc.) require proper equipment to ensure that these
materials are transported to the seal location in the proper condition.
9.4.1 Granular rock salt
Granular rock salt consolidation and permeability

Rock salt has the potential to be an effective, simple seal material. Experi-
mental evidence suggests that granular or crushed rock salt consolidates
under certain conditions, resulting in a porosity and permeability that de-
crease toward values comparable to intact salt. For granular salt emplaced
in an opening in a rock salt formation, the consolidation is driven by the
creep closure of the adjacent host rock. Other advantages of granular salt
are its availability, low cost and obvious compatibility with the host rock.
The time-dependent properties of granular salt have been measured by
numerous laboratory researchers. At a given stress the single most import-
ant parameter in the consolidation of granular salt is the presence of a small
amount of water. Small amounts of water accelerate consolidation and the
accompanying permeability decreases in comparison with dry granular salt
(Zeuch and Holcomb, 1991; Holcomb and Shields, 1987; International
Technology Corp., 1987; Shor, Baes and Canonico, 1981; Pfeifle, 1991}. The
effects of other variables, such as particle size, are secondary and not as
obvious.
The dependence of salt consolidation on added water can be illustrated
by considering the experimental result of Holcomb and co-workers
(Holcomb and Hannum, 1982; Holcomb and Shields, 1987). The 1982 tests
were conducted on dry (no additional water) granular salt, whereas the 1987
tests involved small amounts ( < 3 wt%) of additional water. The volume
strain data, d VIVo, from both sets of data can be reasonably described by
(Holcomb and Hannum, 1982; Holcomb and Shields, 1987)
dV
- = a log(t} + b, (9.3)
Vo
where a and b are constants and t is time in seconds. The constant, b, is a
measure of the initial condition of the sample. To compare times to achieve
the same volumetric strain for tests under similar initial and loading condi-
tions, Equation (9.3) can be rewritten as
(9.4)
where the subscripts d and w refer to dry and wet conditions, respectively.
The constant a w for a wet test is five to ten times greater than ad from a
comparable dry test. Therefore, for dry granular salt to experience the same
strain under similar test conditions requires a time five to ten orders of
magnitude greater than that for the wet sample.
Sjaardema and Krieg (1987) developed and implemented a constitutive
relationship for the consolidation of granular salt based on the laboratory
data of Holcomb and co-workers. Numerical calculations of wet granular
salt consolidation in WIPP shafts and drifts were then conducted to deter-
mine the influence of the presence of the granular salt on the closure of the
shafts and drifts. Up to a fractional density of 0.95 (the extent of the labora-
tory data the model was based on), the results indicate that no substantial
backstress (resistance) develops in the granular salt; that is, the closure is
largely unaffected by the presence of granular salt until the porosity in the
granular salt reached 5%. This result is not conclusive as a limited number
of laboratory tests indicate some bulk modulus increase for compressed
granular salt with fractional densities as low as 0.85 (Gerstle and Jones, 1986).
As expected, as consolidation proceeds the permeability of the granular
salt decreases. Figure 9.4 shows permeability vs. fractional density for two
sets of tests that proceeded to high fractional densities (Butcher, 1991). In
general, permeability values for samples with a fractional density of 0.85 or
less are 10- 15 m 2 or greater. Between fractional densities of 0.85 and 0.95,
however, the permeability drops dramatically. By 0.95 fractional density,
the permeability of the granular salt is of the order of that of intact salt. A
similar trend of a dramatic decrease in permeability at a fractional density
10.10
/ .,
10.11 "
".
10.12
Assumed Pure Crushed Salt Curve ' . '.
(IT Data, Test 3,1987) ",
"
"
~_ Old Backfill
10.13 ~ (German)
\
Assumed Pure Crushed Salt Curve
\ III
10.14 \
., \
(Holcomb Data, 1987) Compacted
\ I
N'
. 10.15
I"
c
'. \
: - Data
(German)
...
~ " ,
:aas \
E
CD 10.16 /"'"
Assumed Salt/Bentonite Curve '''k t, +

CD
a.. ,
10.17 " +
'''!-,
,
'-t
10.18 r--------------------------------" ,
EZZZJ
+
Pure Crushed Salt (Kappel. 1986) ,,
10.19
30% Bentonite (Pleille. 1991;
Pleille and Brodsky, 1990) ,,
)(
C
15% Bentonite (Stroup and Senseny, 1987)
5% Bentonite (Stroup and Sanseny, 1987)
,,
If Pure Crushed Sail (Liedtke, ,,
10.20
as Reported by Kappel, 1986)
Pure Crushed Salt (IT Data, Test 3, 1987) ,,
Pure Crushed Salt (Holcomb and Shields, 1987) ,,
10.21
0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00
Fractional Density (fs)
Fig.9.4 Granular salt permeability as a function of fractional density. (From

Butcher, 1991.)
of 0.95 has been observed in experiments on calcite to simulate the alteration

of permeability and porosity of rocks by plastic flow processes (Evans, 1983).
This figure also shows data reported from Germany (Kappei, 1986). The
higher permeabilities of the German data could be a result of experimental
techniques, using liquid permeability tests. Achieving complete saturation of
the salt with the test fluid is very difficult and enhancement of channeling by
dissolution is possible. Sample disturbance effects similar to those for intact
salt are also possible. On the other hand, the permeabilities derived from
gas measurements could be artificially low due to capillary threshold press
ure and relative permeability effects (Butcher, 1991).
The exact mechanism(s) of consolidation of granular salt are not under-

stood. Clearly, water plays some important role. Yost and Aronson (1987)
dismiss dislocation mechanisms of creep as a primary mechanism for con-
solidation of wet salt, and suggest pressure solution and/or the Joffe effect
as the dominant mechanism(s). Holcomb and Shields (1987) discuss the
possibility of a pressure solution mechanism for consolidation in view of
their experimental data, and conclude that further investigation is required.
Post-test analyses were conducted on consolidated samples (International
Technology Corp., 1987) and it was concluded that water played an import-
ant role in salt consolidation (and the accompanying permeability decrease)
by facilitating pressure-solution or recrystallization. Zeuch (1990) adapted a
model for isostatic hot pressing to the consolidation of nominally dry
granular salt and found good agreement between the model and Holcomb
and Hannum's laboratory data. Interestingly, this model predicts consolida-
tion approaching intact salt densities over periods of less than 50 years
under conditions expected at a depth of about 650 m, in contrast to simple
extrapolations of laboratory data.
While small amounts of water have been demonstrated to benefit consoli-
dation, larger amounts may be detrimental for sealing. It is conceivable that
if the salt becomes saturated while substantial porosity remains, further
consolidation could be impeded by the low compressibility of the entrapped
water (Nowak and Stormont, 1987). Tests by Baes et al. (1983) indicate that
water can be readily squeezed out of salt so as to not impede consolidation,
even to low permeabilities. The results of Zeuch and Holcomb (1991) indi-
cate that saturated granular salt consolidates in a manner similar to that of
granular salt with much less water. However, these laboratory tests have
been on drained samples; it is not obvious to what degree water in large
emplacements could be expelled during consolidation.
Nowak and Stormont (1987) developed a model of granular salt consoli-
dation which couples simplified and idealized representations of shaft or
borehole closure, salt consolidation, water influx from the host rock and
inflow from the overlying water-bearing strata. The model predicts the
porosity decrease of the granular salt due to closure concurrent with the
filling of the porosity from water influx and inflow down the penetration.
As a worst case, consolidation was assumed to cease when the salt became
saturated. (This assumption was made to allow simple, conservative calcula-
tions to be conducted. We expect that the greater rate of closure at depth will
force fluid upward, so water saturation will not necessarily preclude consoli-
dation.) Effective consolidation was assumed to be achieved when the poros-
ity of the granular salt decreased to 5% or less. The model provides
conservative estimates of the final condition of granular salt (saturated poros-
ity) and the corresponding time to achieve its final condition as a function of
depth. The representations of closure, water influx, inflow from the overlying
water-bearing strata, initial porosity of the granular salt and time of emplace-
ment after excavation were varied in order to assess the sensitivity of the
model to these parameters. The following conclusions were reached.
Granular salt will effectively consolidate in less than 200 years in the
bottom 100~200 m of the 650 m deep shaft.
During this period of consolidation, the granular salt must be protected
from substantial inflow of water from the overlying water-bearing strata
by other seal components.
The emplaced density of the granular salt has a profound effect on the
extent and time it takes to become an effective seal.
Emplacement of granular salt seals

The emplacement technique will dictate the initial density of the granular
salt. The preceeding analysis indicates that the careful control over the
emplacement process for granular salt is essential to create an effective seal.
The current technology for the placement of crushed salt seal components is
similar to those used for bentonite pellets. These materials can be trans-
ported to the seal location using a dump bailer or can be simply poured
into shallow boreholes. However, these techniques do not provide adequate
control over the emplacement process. Granular salt components emplaced
in this way will have low density and will be of poor quality from a
permanent sealing standpoint. The granular materials may have a tendency
to 'bridge' in the borehole, resulting in low strength and allowing preferen-
tial flow paths. Also, studies indicate (e.g. Nowak and Stormont, 1987) that
high-quality granular salt should be emplaced with only limited water avail-
able to minimize the buildup of pore pressure in the voids. Such a pore-
pressure buildup could either not allow or restrict the consolidation of this
material to the low porosities required for high-quality, low-permeability
permanent seals.
Higher emplaced densities will result in a lower permeability and will
increase confidence in the long-term performance. Holcomb and Hannum
(1982) found that merely pouring granular salt into a mold with no added
compaction will result in a fractional density of only about 0.60. Compacted
blocks have achieved fractional densities of up to 0.85 (Stormont and
Howard, 1987; Torres, Howard and Finley, 1992). Quarried salt blocks, cut
from intact salt, could have fractional densities in excess of 0.95. However,
placement of either compressed or quarried blocks in remote boreholes is
improbable under even the best circumstances. Also, large-diameter
boreholes may require systematic placement of numerous blocks, creating
numerous joints which may act as preferential flow paths for long time
periods after emplacement.
Recently, the replacement and compaction in situ of granulated salt,
whose long-term performance can be demonstrated, into boreholes as a
permanent sealing material has been investigated. Studies of the dynamic
compaction of granular salt suggest that fractional densities approaching
0.90 can be obtained in remote boreholes. These tests, and other dynamic
consolidation concepts under development, are believed to be compatible
with existing downhole equipment. Field emplacements using these tech-
niques have not yet been attempted. There are several advantages of in situ
compaction of granular salt as a permanent seal material for borehole seals
and other mined openings in evaporite deposits.
The permeability of granular salt decreases significantly when the relative
density approaches 90-92% of the intact density. These are density values
approachable using in situ compaction techniques under development.
The compaction process in situ can be shown to 'lock in' horizontal stress
in the compacted material. This horizontal stress enhances the formation
of a tight interface and creates a stress condition in the surrounding host
rock that accelerates the healing of the DRZ that develops during
borehole drilling.
In situ compaction techniques using native materials can provide excel-
lent quality control over the emplacement process, thus providing confi-
dence that the seal has consistent properties throughout.
The compatibility of the seal material assures that the long-term and
permanent performance of the seal can be conservatively estimated from
the initial emplaced densities. The long-term performance will be enhanced
as creep of the surrounding rock reduces the porosity, and hence the
permeability, of the emplaced material to that approaching the host rock.
Porosities as low as about 10% have been achieved using techniques
compatible with emplacement in boreholes. Compacted granular materials
should be used for those seals in which a high-quality permanent seal is
required. These components can be used in conjunction with other, more
standard, seal materials. Actual seal designs should be developed on a
borehole by borehole basis.
9.4.2 Cementitious materials

Cementitious grouts have been used for many years to seal surface-drilled
well bores for disposal of chemical and toxic wastes, and to seal abandoned
oil and gas boreholes. Cementitious materials have historically been used as
a borehole seal material because of their availability, relatively low cost and
ease of use. Furthermore, properties of standard cements, such as strength
and permeability, are generally understood and considered adequate for
typical sealing applications.
The single consistent conclusion from historical experience of borehole
sealing is that the cement itself is relatively impermeable and that observed
leakage is predominately attributable to the cement/formation or cement
casing interface and the adjacent rock. Probable causes for flow at the
interface are cement shrinkage, poor rock quality and interaction between
the cement and the host rock. Typical cement formulations for sealing in
salt include an expansive (or at least shrinkage compensating) agent to

create a tight interface, and salt-saturated mix water to reduce dissolution
at the interface (e.g. Eilers, 1974).
The WIPP Experimental Program has supported the development of
cementitious grouts for sealing excavations in rock salt since 1975. In 1979
an expansive grout was developed for the plugs placed and tested in
borehole AEC-7 near the WIPP Facility. This test, the Bell Canyon Test
(BCT), involved placing a grout plug in an anhydrite layer at a depth of
1370 m below ground surface. Both freshwater and saltwater expansive
grouts were studied (using Class H cement, fly ash and an expansive
additive (calcium sulfate)}. Class H cement produced in the southwest
USA is a coarse-grind cement with a low tricalcium aluminate (C 3 A)
content. The coarseness of the cement is beneficial for reducing the water
content and for achieving high strength and density and low permeabil-
ity. Maximum practical use of a high-range water reducer also assists in
achieving pumpability and the hardened physical properties. Low C 3 A
content in cement has been shown to be a major factor in the durability
of concretes, particularly with respect to sulfate attack. Fly ash has been
included in all mixtures (30% by solid volume replacement for cement) to
reduce an early temperature rise and to contribute to later-age strength
and durability. The expansive additive for this cement system was gyp-
sum hemihydrate or plaster. To reduce the potential for dissolution of
the adjacent rock salt, salt-saturated water was used in the mixture
designated BCT-1F (Table 9.1) Because anhydrite is more soluble in
brine than in fresh water (Gulick, Boa and Buck, 1980), a mixture desig-
nated BCT-1FF was developed with fresh mix water for sealing in an-
hydrite layers. Results from laboratory durability studies have been
reported with no appreciable deterioration (Gulick, Boa and Buck, 1980,
1982; Rhoderick and Buck, 1981; Buck, Burkes and Rhoderick, 1981a,b;
Burkes and Rhoderick, 1983).
A principal concern regarding the use of cementitious materials as a
long-term seal material are their durability or longevity. Cementitious ma-
terials will not be in chemical equilibrium with their environment (Lambert,
1980). Potential mineralogical phase changes could manifest themselves as
(1) the formation of a soluble, friable, or permeable phase in the seal or
nearby rock or (2) shrinkage or degradation of adhesion, opening the inter-
face between the seal and the rock (Lambert, 1980).
Cements are especially susceptible to degradation in environments with
high sulfate contents (Lea, 1971). Hart (1983) reports that concrete liners
which pass through the formation above salt mines in the northeastern US
degrade or corrode due to formation water leaking through the liner, result-
ing in a reduction of the concrete thickness of about 3mm per year, An
examination of a 20 year old shaft liner in the Carlsbad potash district
suggests that sulfate attack has caused appreciable deterioration of the
concrete liner (D'Appolonia, 1981). Heinmann, Stanchell and Hooton (1986)
demonstrated that the presence of clay accelerates the dissolution of some

cementitious materials.
Nowak et al. (1992) analyzed samples of concrete from a WIPP shaft
liner. Degraded concrete was found behind the liner and at joints between
concrete lifts. The magnesium-rich brine collected from the annulus between
the liner and the host rock was implicated in causing the deleterious reac-
tions. Mechanisms of degradation include a chemical alteration of the con-
crete paste by the magnesium causing a loss of strength; precipitation of
magnesium in the concrete pores, resulting in cracking; and reactions be-
tween chloride ions and alumina in the concrete, resulting in a more per-
meable paste.
On the other hand, there is evidence for the longevity of cementitious
materials in certain environments. Evaluation of some ancient cementitious
materials reveals they have survived in apparently good condition for cen-
turies (Malinowski, 1982; Monastersky, 1987). Buck and Burkes (1979)
examined a 17 year old grout specimen from a shaft liner in contact with
rock salt. The Class C-based grout was found to be in tight contact with the
host rock and have no obvious deterioration.
A series of cementitious materials including the BCT-1F mixture were
emplaced in boreholes in the WIPP Facility in 1987. These materials are
intended to provide a test bank of candidate seal materials for periodic
sampling to examine them for evidence of degradation. In 1993 some sam-
ples were removed and examined. The condition of the samples was asses-
sed by compressive strength and pulse velocity measurements. There has
been no measurable deterioration of the materials after 6 years (Wakeley,
Harrington and Weiss, 1993).
From the above observations, we conclude that the empirical evidence for
cementitious materials is mixed. Clearly, the presence of water at the host
rock -seal interface is a key factor in the stability of these materials. If they
remain relatively dry, they will apparently remain stable indefinitely.
Cementitious materials in contact with water show observable degradation.
Another approach to assessing the long-term stability of cementitious ma-
terials is through analytical models. However, there is presently no compre-
hensive model of the complicated system of cementitious materials, the host
rock, the formation water and their interaction sufficient to make reliable
predictions of long-term (hundreds of years), time-dependent performance.
Indeed, the problem is so complex (involving kinetics, thermodynamics and
chemistry) that resolution of all issues in the near future seems remote. The
emplacement of the cement will have a large impact on its effectiveness.
Techniques for emplacement of cements and other plugging materials are
well known from the petroleum and other industries and have been re-
ported extensively elsewhere (e.g. Montgomery and Smith, 1961; Suman and
Ellis, 1977a; Herndon and Smith, 1978; George and Faul, 1985; Smith,
1990). Cements are probably the most commonly used materials for plug-
ging and sealing of boreholes in salt and have a long history of use. The
common emplacement techniques take advantage of the favorable ma-

terials-handling characteristics of these products. Cementitious materials
are typically placed using one of several methods, including balanced plug,
dump bailer, two-plug method, or low or high pressure squeeze. Tremmie
pipes can be used for shallow boreholes.
9.4.3 Clays
Clays have found many applications as fluid barriers in underground exca-
vations (e.g. National Coal Board, 1982; Sitz, Koeckritz and Oellers, 1989),
as components of earth dams (e.g. Simi and Harsulescu, 1979) and in con-
tainment of hazardous wastes (e.g. Johnson et al., 1984; Leppert, 1986). In
particular, sodium bentonite is under consideration as a seal material for
geological nuclear waste repositories (e.g. Stormont, 1987). Bentonites have
low permeability, swell on contact with water to fill voids and establish a
tight interface with the rock, are relatively stable and can sorb or retain
certain contaminants. Bentonites are composed principally of montmoril-
lonite, a smectite mineral responsible for their characteristic swelling. Be-
ntonite mixed with filler or ballast material is being considered as a seal
material as a matter of economy, as well as to minimize the loss of the
bentonite through small fractures. Sitz, Koeckritz and Oellers (1989) found
that the sand in a bentonite/sand mixture stopped bentonite losses through
fractures with maximum widths of 2-4 mm.
The permeability of mixtures of bentonite and various filler materials has
been measured by numerous investigators in the laboratory (Wheelwright
et al., 1981; Peterson and Kelkar, 1983; Radhakrishna and Chan, 1985;
Stroup and Senseny, 1987). There is considerable variability in the data
owing to differences in test methods, sample density, working fluids, etc. In
general, the permeability of the mixtures to water and brine was found to
fall off to 1O-18 m 2 or lower somewhere between 25 and 50wt% bentonite.
This is probably coincident with the bentonite becoming the continuous
phase of the mixture. Pusch (1987) determined that the permeability of
bentonite to brine is about an order of magnitude greater than that to fresh
water.
Another important property of mixtures containing bentonite is the
swelling pressure developed when the mixture is confined and saturated
with water. Swelling is expected to fill voids and heal fractures within the
bentonite seal and perhaps to a limited degree in the adjacent host rock.
Gray, Cheung and Dixon (1984) and Pusch (1980) demonstrated that swell-
ing pressures of bentonite mixtures are dependent on the effective clay
density; that is, the mass of the bentonite divided by the volume of the
bentonite and any voids. Thus, the sand or other filler material is merely an
inert filler. Clay structures can fail in the presence of seepage by erosion
along pre-existing cracks or piping (internal retrogressive erosion). The pre-
dominant factors involved in failure by both mechanisms are loosening of
interparticle cohesive forces upon saturation (dispersivity), permeability and

swelling potential (Resendiz, 1976). The risk of failure is increased as the
first two factors increase and the third decreases. Clays rich in montmoril-
linite (e.g. bentonite) are generally too expansive to permit cracks to remain
open and too impervious to allow seepage velocities large enough to induce
piping (Resendiz, 1976). Further, bentonite is relatively plastic and can
withstand considerable deformation prior to failure. The tendency for ero-
sion or piping failures is increased at the interface between the clay and
dissimilar materials (Penman and Charles, 1979), i.e. the seal-rock interface.
Pusch et al. (1987a, b) demonstrated the effectiveness of bentonite in cre-
ating a tight interface by swelling. Pusch (1983) investigated the possibility
of the migration of bentonite into rock fractures and the subsequent erosion
of the bentonite by flowing groundwater. He concluded that bentonite will
migrate a few tenths of a meter into fractures wider than 0.1 mm over the
course of thousands of years and should not be significantly eroded by
groundwater.
Clays exist naturally in geologic formations, including bedded salt, and
are therefore appealing as long-term seal compone~ts. Clay sealants have
been used by man for long periods of time; Lee (1985) documented the
effectiveness of a clay sealant for periods as long as 2100 years. While
bentonite alteration to other clays does occur under some conditions, at
non-elevated temperatures bentonite transformations are expected to be
very slow, of the order of millions of years (Meyer and Howard, 1983; Roy,
Grutzeck and Wakeley, 1983). Krumhansl (1984) found from experiments in
WIPP-specific aqueous solutions that bentonite is expected to maintain its
desirable mineralogic characteristics indefinitely.
Bentonite is commonly placed as compressed pellets or slurries using
techniques similar to those used for cements. An alternative technique has
been demonstrated by Pusch (1983) for the emplacement of bentonite as a
sealing material in boreholes in granite. This technique takes advantage of
the bentonites natural swelling potential by placing a sacrificial perforated
copper pipe in a borehole, forcing highly compressed cylinders of bentonite
into the pipe and lowering the pipe into the borehole to the seal location.
The bentonite swelled to press through the perforations in the pipe, effecting
a low-permeability seal within the borehole. Alternative techniques for ben-
tonites are simply to 'pour' pellets into the borehole, thereby creating a seal
at the water level in the borehole. This technique does not allow adequate
estimates of the quality of seals emplaced in this way. Bridging of the pellets
can occur, causing a weakness in the seal together with possible preferential
flowpaths.
9.4.4 Other sealing materials

Asphalt is a bituminous material produced by the distillation of crude oil.
In the construction industry, asphalts have a wide variety of applications
Seal system performance 211
because they are durable, plastic, highly waterproof, strong and highly resis-
tant to the action of most acids, alkalies and salts (Herubin and Marotta,
1977). Bacterial degradation requires microorganisms and moisture. Even if
these conditions are present, the degradation is expected to be very slow
(ZoBell and Molecke, 1978). Many properties of asphalt, including density
and viscosity, can be tailored by the distillation process and by the addition
of weighting materials and blending and dissolving agents. Liquid asphalt
has been utilized as a key component in the construction of waterproof
liners in strata overlying salt and potash deposits (Hart, 1983). Sitz,
Koeckritz and Oellers (1989) describes the use of asphalt as a component of
an elaborate seal for an underground gas storage facility in domal salt.
Various chemical grouts have been used in shaft grouting programs and
chemical seal rings. Polymer materials have been considered as seal ma-
terials for boreholes in salt. These materials typically are costly and some
require special handling procedures to mitigate health concerns. The long-
term durability of polymers is unknown (Coons, Meyer and Kelsall, 1982).
A recent study (Ehgartner, 1991) investigated the use of low-density epoxy
grout as a temporary seal and mechanical base for other undefined perma-
nent seals. These laboratory and bench-scale tests demonstrated strengths in
excess of 10 MPa which were considered acceptable for their intended use.
9.5 SEAL SYSTEM PERFORMANCE
Borehole seal performance cannot be deduced solely from an understanding

of its components, that is, the host rock and the seal material. Rather, it is
more useful and accurate to think of the borehole seal as a system com-
prised of the intact host rock, the rock adjacent to the borehole which has
altered properties, the seal~rock interface and the seal itself. Because of the
favorable properties (low permeability, adequate strength) of the intact salt
and other evaporites and most seal materials, it is most often the seal~rock
interface and the host rock adjacent to the borehole which dominate the
performance of the seal system.
9.5.1 Laboratory tests of seal systems in salt

Mechanical push-out (bond strength) and hydraulic tests are performed on
seal~rock and seal~casing systems to provide a measure of the seal quality
(e.g. Carpenter, Brady and Blount, 1992). A few laboratory tests have been
performed on borehole seal systems in rock salt. Grout plugs have been cast
in central holes drilled in salt cores and subjected to various tests as de-
scribed below.
Push-out or bond strength tests were conducted on cementitious plugs
cast in rock salt cores by Akgun and Daemen (1986). They emplaced an
expansive salt water-based grout in 2.5 cm diameter holes in large-diameter
salt cores. The interpreted shear strengths along the interface depend upon
the assumed shear stress distribution. Assuming a uniform shear stress dis-
tribution yields local shear strengths from 2 to 12 MPa. Dissolution occur-
red along the interface in a number of samples and tended to reduce the
strength of the bond. They reconditioned a number of failed samples by
drying and rewetting them, and then retested these samples. They found
that the reconditioned samples achieved an average of 60% of the original
strength.
Laboratory tests under simulated downhole conditions were conducted
on 38 cm diameter by 122 cm long cylinders of domal salt from Avery
Island, Lousiana (Bush and Piele, 1986). While the samples were subjected
to a confining pressure of 15.9 MPa and an axial stress of 17.2 MPa, a
borehole was drilled through the center of the samples. A 20 cm diameter
diamond core bit with a bit weight of 89 kN was operated at 60 rpm to
create a central borehole in the sample. Saturated brine was used as the
drilling fluid and was maintained at a pressure of 13.8 MPa in the borehole.
A grout borehole seal was installed under downhole conditions. The grout
was a salt water-saturated, expansive grout similar to the BeT-IF grout.
An arrangement to simulate a downhole bailer was used to place the grout.
After curing for 28 days, flow tests were conducted on the borehole seal
system with saturated brine containing a fluorescent dye. A permeability of
approximately 10- 18 m 2 was inferred from the test (Bush and Piele, 1986).
Numerous cores were taken of the borehole seal-rock system and exam-
ined. Some damage of the rock adjacent to the borehole was evident. There
was also some evidence of local dissolution, apparently due to undersatura-
tion of the brine. Dye was observed along the seal-rock interface and in the
rock adjacent to the seal. Permeability tests on the rock -seal interface
indicated that it possessed an equivalent permeability of the order of
10- 12 m 2 (Bush and Piele, 1986). None of the post-test measurements or
observations suggested that the grout was an inadequate seal material
(Scheetz, Licastro and Roy, 1986).
9.5.2 Field measurements of seal system performance

A series of seal system tests have been performed in the underground work-
ings of the WIPP Facility. These tests utilize materials and geometries
similar to those being considered for the eventual sealing of the WIPP
Facility. These tests were conducted in boreholes from 15.2 cm to 96.5 cm in
diameter. The emplacements holes are oriented both vertically into the floor
and horizontally into the ribs. The seal system comprises the seal, seal-rock
interface and the rock adjacent to the seal. Figure 9.5 show a generalized
configuration for the tests conducted to date. Typically, an emplacement
hole and an access hole (if used) are drilled either in the rib or floor of the
opening. The seal material, in some cases containing instrumentation, is
emplaced over a predetermined interval in the emplacement hole. A packer
Pressure and
Flow Meters
00
Receiving Packer (Optional)
Tracer Detection Interval
Access Hole In Some Cases,

Instrumentation
Is Emplaced in Seal
and Adjacent Rock
Injection Straddle
Test Interval
Packer System
Not to Scale
Fig.9.5 General test configuration for in situ seal tests. (From Stormont, 1986.)
system capable of using brine or gas can be placed in the access hole (where
used), and the interval beneath the seal can be pressurized for fluid flow
measurements.
Concrete seals
Concrete seals were emplaced in nine holes located in the floor and rib of an
experimental room in the underground workings of the WIPP Facility (Stor-
mont, 1986; Stormont and Howard, 1986). The seal material was an expan-
sive salt-based concrete based on the BCT-IF grout shown in Table 9.1
(Wakeley and Walley, 1986). The concrete was mixed on the surface and
transported underground by means of concrete buckets suspended beneath a
skip on the hoist. Acceptance criteria for the concrete included slump, limited
bleed, segregation, limited air entrapment, self-leveling behavior and worka-
bility. For the vertical seals the concrete was poured into place via a tremmie
pipe to reduce the potential for entrapped air during free-fall. The seals were
underlain by a granular salt layer. One instrumented and one uninstrumented
seal of each of the following dimensions were placed: 91 cm diameter, 91 cm
long; 41 cm diameter, 60cm long; and 15 cm diameter, 30cm long. Instrumen-
tation included stress and strain measurements in the concrete seals and the
adjacent rock. For the horizontal seals, the concrete was pumped into
pre-positioned forms. All the horizontal seals were of the same dimensions,
namely 91 cm diameter and 91 cm long.
Many of the seals have been subjected to brine and/or gas flow tests
(Peterson, Lagus and Lie, 1987). A packer system was placed into the access
hole and the test region beneath the seal was pressurized. Tracer gases were
included in some of the gas flow tests. Because of the test configuration, it is
necessary to make assumptions about the flow geometry in order to inter-
pret permeability from most of these tests. The reported seal permeabilities
were calculated as if all of the flow was through the seal system and thus
they represent maximum values. The data can also be interpreted by assum-
ing all of the flow is into the formation and none through the seal system.
The test results are summarized in Table 9.2. Tracer gas breakthrough
was detected in only one of the vertical seals. Two of the three brine flow
tests on vertical seal systems were dominated by flow into horizontal frac-
tures which intercepted the test region beneath the seal. In the other seal
tested with brine, no flow of brine through the seal system was observed
during the 140 day test. The interpreted seal system permeability from the
gas and brine tests on this seal are essentially identical, about 10- 18 m 2 .
(Peterson, Lagus and Lie, 1987). It is important to bear in mind that the
reported permeabilities are not definitive measures of intrinsic permeability,
principally because the flow paths are unknown. In addition, interpreted
brine permeabilities may be affected by dissolution of the host rock salt and
creep closure of the holes, and interpreted gas permeabilities are affected by
the Klinkenberg phenomenon and threshold pressure effects (Peterson,
Lagus and Lie, 1987).
The horizontal seals were tested only with gas. The test results are sum-
marized in Table 9.3. The interpreted permeability of each of the seals is
quite low, but tracer gases were detected on the low-pressure side of all
three seals in less than 1 h. Estimated fracture apertures are given in
Table 9.3. In the two instrumented seals the instrumentation cables which
penetrate the seal were determined to be the predominant flow paths. There
was no detectable leakage along the seal-rock interface. For the unin-
strumented seal the leak was identified to be along the seal-rock interface.
Table 9.1 Components and proportions of salt-saturated grout (BCT-1F)

Component %of total %of total Batch weight
by mass solids by mass for 1 m 3 (kg)
Class H cement 48.3 61.2 991.1

Class C fiy ash 16.2 20.6 333.0
Plaster 5.7 7.2 116.9
NaCI 7.9 10.0 161.7
Dispersant 0.78 1.0 16.0
Defoamer 0.02 0.Q2 0.3
Water 21.1 433.9
Total lOO.O 100.0 2052.9
Table 9.2a Summary of vertical concrete gas flow tests
Seal" Seal Formation Tracer Estimated
permeabilityb permeabilityC arrival maximum fracture
(m 2 ) (m 2 ) (h) apertured ( cm )
MAE 11 3 x 10- 19 5X 10- 21 none < 10- 6

MAE12 6x1O-19 4x 10- 21 none < 10- 6
MAE21 4x 10- 19 1X 10- 20 none < 10- 6
MAE22 3x 10- 19 1X 10- 20 none < 10- 6
MAE31 e 2x 10- 17 2X 10- 18 1.5 1.3 X 10- 5
MAE32 5x 10- 19 3X 10- 20 none <2 x 10- 6
"Length/diameter of seals (cm): MAEll, MAE12 (30.4/15.2); MAE21, MAE22 (60.0/40.6);
MAE31, MAE32 (91.4/91.4). bCalculated permeability assumes all flow through seal with
porosity of 0.01. cCaicuiated permeability assumes all flow through formation with porosity of
0.001. d< symbol indicates tracer arival should have occurred by the end of the test if the
fracture aperture was greater than the value shown. Obvious leak through seal.
All of these seals were retested approximately 1 year later. No tracer gas
was detected through any of the seals. These results suggest that stress
build-up on the seal due to the closure of the host rock salt can improve the
seal system performance.
The mechanical (stress, strain) response of some of the expansive salt-
water concrete seals was measured (Stormont, 1987). Within the first 100
days a number of interacting processes yield a complicated stress-strain
response within the seal. These processes include chemical expansion/con-
traction, thermal expansion/contraction and thermally induced creep of the
neighboring rock salt. Beyond 100 days, creep deformation of the rock salt
is the dominant mechanism which controls the stress-strain response of the
seal system.
Compressive stresses are developed and maintained at the seal-rock in-
terface and no stress or strain measurements indicate concerns about stabil-
ity or impending failure. Thermomechanical modeling of the concrete seals
can reproduce some of the observed behavior, but is far from an accurate
predictive tool (Van Sambeek and Stormont, 1987; Labreche and Van
Sambeek, 1988). Uncertainties in the modeling of the host rock salt adjacent
to excavations is the biggest single difficulty in modeling the concrete seal
systems (Labreche and Van Sambeek, 1988).
Granular salt and bentonite seals

Seals consisting entirely of granular salt were emplaced in 1 m diameter
horizontal and vertical holes. The seals were constructed from pressed salt
blocks made with an automated block-pressing machine (Stormont and
Howard, 1987). Small amounts of water were necessary to produce durable
blocks. Relative densities of the seals were typically greater than 82%.
Table 9.2b Summary of vertical concrete brine flow tests
Seal a Seal permeability Formation permeability Estimated maximum Comments

(m 2 )/porosity (m 2 )/porosity fracture aperture e (cm)
MAE12 - b <7 X 10- 7 No visible flow through seal

during 140 day test with
1.7 MPa brine pressure
MAE22 2 x 1O- 19 /O.OO3 c 1.2 x 10- 2 %.01 to O.OOld < 1 x 10-6 No visible flow through seal
b
MAE32 < 2 x 10-6 No visible flow through seal
"Length/diameter of seals (cm): MAEI2, 30.4/15.2; MAE22, 60.0/40.6; MAE32, 91.4/91.4. bSeal and formation permeabilities could not be
determined, as brine flowed to adjacent boreholes, possibly along horizontal fractures. The test pressures were reduced from the 3.5 MPa
design level (at which the flow rates continued to increase as a function time) to 1.8 MPa (where the rates remained constant). cAssumes seal
is initially unsaturated. dAssumes formation is saturated. e< symbol indicates tracer arrival should have occurred by the end of the test if the
fracture aperture was greater than the value shown.
Measurements made on the seals included both mechanical (deformation

and stress) and gas permeability measurements. These seals were emplaced
at an equivalent relative fractional density of approximately 0.82-0.83 and
after 5 years have achieved fractional densities up to 0.86, due to creep
closure of the surrounding rock. (Torres, Howard and Finley, 1992).
Measurements of stress within the seals do not indicate significant loading
of the seals. The measured gas permeabilities exceeded the measurement
capacity of the available equipment (approximately 30000 cm 3 / min). These
materials are expected to provide high-quality seals after long periods of
time. The emplacement of block seals proved to be a labor-intensive process
and it is not known what time period is required for creep closure of the
surrounding rock to create a high-density seal.
Seals composed of mixtures of bentonite and salt and pure bentonite
have also been emplaced. Seals of 50% sodium bentonite and 50% granular
salt have been evaluated and reported by Stormont and Howard (1987) and
Finley and Jones (1994). The tests on the 50% bentonite and 50% granular
salt block seals show bulk permeabilities of the order of 10 - 15 to 10 - 16 m 2
(Stormont and Howard, 1987; Finley and Jones, 1994). These studies also
showed that incipient piping along the interface between the seal and host
Table 9.3 Comparison of April 1986 and May 1987 horizontal concrete tracer gas
flow tests
Seal" Seal Formation Tracer Estimated Comments
( date) permeability permeability arrival maximum
(m 2 ) (m 2 ) (h) fracture
aperture ( cm)
MBE31 1 x 10- 18 8 X 10- 20 <0.3 >2xlO- 5 Leak through

(4/86) fracture in
formation
below seal
MBE31 2 x 10- 19 1 X 10- 20 None <8 x 10- 7
(5/87) (264 h)
MBE32 4 x 10- 19 3 X 10- 20 <0.2 >3 x 10- 5 Leak along
(4/86) instrumentation
bundle
MBE32 2 x 10- 19 1 X 10- 20 None < 3 x 10- 6
(5/87) (26 h)
MBE33 2 x 10- 19 1 X 10- 20 <0.1 >4xlO- 5 Leak along
(4/86) instrumentation
bundle
MBE33 7 x 10- 19 5xlO- 2O None < 3 x 10- 6
(5/87) (24 h)
aAll seals: length 91.4 cm, diameter 91.4 cm; bcalculated permeability assumes all flow through seal
with porosity of 0.01; ccalculated permeability assumes all flow through formation with porosity of
0.001; d < symbol indicates tracer arrival should have occurred by the end of the test if the fracture
aperture was greater than the value shown.
rock interface may have been arrested by migration and subsequent swell-
ing of the bentonite. This was determined by post-test evaluation of clay
contents from various locations within the seal. Also, it was shown that the
rate of filling of the test interval may have contributed to the failure of one
seal, again due to piping along the interface.
Seals composed of 100% bentonite blocks compressed to densities of
1.8 g/cm 3 and 2.0 g/cm 3 have also been emplaced (Finley and Tillerson,
1992). All the seals are comprised of compressed blocks emplaced manually
in the boreholes. These seals are intended to evaluate the effectiveness of
pure bentonite as a sealing material in large-diameter openings (1 m) and to
evaluate differences in emplaced densities of the bentonite. Preliminary re-
sults of these tests show permeability to brine to be about 10- 18 m 2 (Finley
and Tillerson, 1992). These seals include internal pressure measurements
which show only limited bentonite swelling pressure under approximately
0.5 MPa brine pressure.
9.6 DESIGN CONSIDERATIONS
Design and performance information for sealing boreholes in rock salt is

contained throughout this chapter. In this section we highlight and reiterate
some of the most important design considerations.
It is of paramount importance to consider the entire seal system when
developing and defending designs for sealing boreholes in rock salt. The
complex interactions between the seal and host rock must be recognized so
that reasonable predictions of the seal performance can be made. The com-
plicated nature of the in situ system suggests that seals should be designed
on a case by case basis.
The design begins with the definition of the performance requirements for
the borehole system. We defined short-term, long-term and permanent seals
to represent the range of performance requirements possible for a borehole
seal system. We suggest that when the hole is to be abandoned, permanent
sealing represents the best option. Restoring the borehole to a condition
comparable to the formation by means of the consolidation of granular salt
represents a permanent seal. On the other hand, relying on the performance
of a cementitious seal whose properties will eventually degrade is not an
optimal approach, even if the degradation is very slow.
The difference in performance of granular salt and cementitious seal sys-
tems are given in Figure 9.6. Many design considerations are embedded in
these figures and they serve as a useful illustration of the design process.
The figures provide the permeability of the seal systems as a function of
time. In the figures the values for time are qualitative, and for permeability
are approximate. A similar figure could be developed for other sealing
materials, notably bentonite. However, bentonite requires structural
~n-'l
Cementitious seal Cementitious seal
(Immediate "penneability) ... 1-1-1-1-1- ------------
1'; 1 maximum K possible
/ (Breakdown into
=10 -18 11 GranularConstituents)
1
DRZAround
Borehole Healing I Geochemical
,... .... All -Degradation of
M _ Cementitious seal _ 1"; 1 CementitiousMaterials
Possible
~
minim~m Kpossible i
I
- I
SW
J
... I I I
. --------,- ---------------- ------------------------l------ ------!~~~
I
I
I
I
I
I
I
Borehole Borehole Time Geochemical
Drilled Sealed (Dimensionless) ~radation (years)
Fig. 9.6 a Seal permeability as a function of time for cementitious seals.

Open hoi
-14 2
----------- 80% density granular salt (-10 m)
-15 2
85% density granular salt (-10 m
-17 2
......
----------------- 90% density granular salt (-10 m)
N
g
>. -20 2
:= ___________ 92% density granular salt (-10 m)
is
~
E
~
Intact Halite
I
I _10-22 - --
I
I
I
I
I
Time to I
high density !
LowK- I
years to 100s I
of years I
~!
time
Borehol Borehole (Dimensionless) Time to Complete Reconsolidation to
Drilled Sealed and Intact K - years to 100s of years
Emplaced
Fig. 9.6 b Seal permeability as a function of time for granular salt seals.
confinement for it to remain effective and the performance of seals is

dependent on cementitious seal performance, as presented in Figure 9.6a.
Prior to the drilling of the borehole, the initial permeability of the system
is that of the formation. For rock salt this is very low, of the order of
10 - 22 m 2 or less. The permeability increases to that of an open borehole
upon drilling. The emplacement of the seal material causes an immediate
reduction in the seal system permeability.
The permeability of the cementitious seal system indicates an almost
immediate decrease upon seal placement to a value of about 10- 17 to
1O- 18 m 2 . This is considered to be the minimum permeability of a carefully
constructed cementitious seal system immediately after placement, based on
experimental evidence of similar systems (Peterson, Lagus and Lie, 1987).
The permeability of the system can be attributed to the seal-rock interface
and the adjacent disturbed rock zone. Creep of the host rock will tend to
reduce the permeability of the disturbed rock zone and to create a tight
seal-rock interface. After some time the seal system attains its lowest per-
meability, which is equivalent to the permeability of the seal material itself.
For example, high-quality cements have permeabilities as low as 1O- 2 m 2 .
After some period of time, cementitious seal materials may begin to degrade
because of the chemical incompatibility between the seal, the host rock and
formation water. Assuming that degradation proceeds until the seal is re-
duced to its constituents, the ultimate condition of the seal may be compar-
able to sand and have a permeability of the order of 10- 15 m 2 . Because the
permeability of the cementitious seal system will never achieve a permeabil-
ity as low as the host rock and the seal is likely to degrade eventually,
cementitious seals are considered to be long-term but not permanent seals
by themselves.
The highly time-dependent behavior of a granular salt seal system is
illustrated in Figure 9.6b. For low initial emplacement densities (say 0.8
fractional density), the initial permeability of the seal system is largely a
function of the seal material itself. These materials will continue to consoli-
date as a consequence of the creep closure of the borehole until they achieve
a permeability equivalent to that of the host rock. The time required to
achieve this degree of consolidation is of the order of hundreds of years for
typical creep rates. Emplacement of granular salt of higher initial density
offers an attractive alternative to cementitious seals. An initial emplaced
density of 90% can provide immediate performance comparable to the
cementitious seals (about 10- 17 m 2 ). As shown in Figure 9.6b, higher initial
densities would be result in even lower permeabilities. The time required for
the permeability to decrease to levels comparable to the intact formation for
90% or greater initial fractional density may be less than 100 years. It is
important to note that the granular salt seals are not expected to degrade
with time, due to the natural compatibility of the seal materials with the
host rock. Granular salt seals are therefore considered to be permanent
seals.
A design approach which takes advantage of the attributes of both

cementitious and granular salt seals is the use of composite seals comprised
of both types of seals. Granular salt seals are placed in the substantial rock
salt sections. Cementitious seals are located between water-bearing zones
and rock salt formations, preferably sealing across the contact of the salt
and adjacent units. This approach utilizes cementitious seals to limit fluid
movement into the granular salt as it is consolidating. Once the granular
salt has consolidated, the performance requirements for the cementitious
seals diminish.
Successful design of a permanent borehole seal requires an understanding
of the properties and behavior of the host rock salt. Borehole closure due to
creep of the surrounding rock will drive the consolidation of the granular
salt and eventually reduce the permeability of the disturbed rock zone.
Creep rates vary from site to site, and even within a borehole, due to stress,
temperature, moisture and impurity variations. Bounds on the expected
creep rates must be developed, preferably from field measurements.
Another important consideration in the permeability of the granular salt
seal system is the magnitude and extent of the disturbed rock zone. This
region, which extends about 0.5-1 borehole radius into the rock, can be
sufficiently permeable to serve as a bypass of the seal system. For example,
for a seal system with a granular salt seal emplaced with an initial permea-
bility of 10- 17 m 2 , a disturbed rock zone with a permeability of 10- 16 m 2
will be the principal flow path through the system. Fortunately, the permea-
bility of the disturbed rock zone will decrease to intact values as the forma-
tion creeps in and consolidates the granular salt seal.
The inflow of water into the borehole may compromise a granular salt seal
system. It is possible that if sufficient water enters the pore spaces of the
granular salt it could impede further consolidation. One source of water is
brine from the rock salt formation itself. Because the brine inflow rates are
low they are difficult to measure. The few measurements of brine inflow
which have been made suggest that brine inflow can be estimated, based on
measurements of the formation permeability and pore pressure. Water inflow
from adjacent water-bearing zones will probably be the principal source of
water. The seal system (probably cementitious) between the granular salt seal
and the water-bearing zone will be relied upon to limit this quantity.
The emplaced density of the granular salt seal is a principal factor in the
performance of the seal system. Simply dumping granular salt into the
borehole will result in a fairly low initial fractional density (60-80%). The
time to achieve consolidation will increase, the performance requirements
on other sealing components such as cementitious seals will increase, and
the certainty of the overall seal system performance will be reduced. The
greater the initial density of the granular salt, the better. As with most
granular materials, tailoring the moisture content and the particle size dis-
tribution can increase the initial density. The greatest improvement in initial
densities, however, is achieved with some type of downhole compaction.
Conclusions 223
The above considerations have been combined into a model to predict

the performance of a granular salt seal system (Nowak and Stormont, 1987).
This model could be adapted or modified for other sealing applications, or
a separate model could readily be developed. A model of this type requires
estimates and simplified representations to be made of closure, brine inflow,
water inflow from adjacent water-bearing units, the disturbed rock zone,
and granular salt consolidation.
Because it is important to limit the amount of water which enters the
consolidating granular salt seals, an understanding of the performance of
the portion of the seal system which protects the consolidating granular salt
is necessary. These seals will consist principally of cementitious materials
and perhaps other materials such as clays. The flow of water through these
seal systems will depend in part upon the hydraulic pressure of the water-
bearing zones and the permeability of the rocks between the water-bearing
zones and the rock salt section. The permeability of the seal material and
the seal-formation interface are obviously also important. It is reasonable
to assume that the seals will degrade with time, although degradation rates
and ultimate conditions are difficult to estimate. A model which predicts
flow through a seal system above a consolidating granular salt seal was
developed by Stormont and Arguello (1988). The model accounts for the
formation pressure, formation permeability, seal and seal-rock interface
permeability and seal degradation. This approach is one method of devel-
oping an estimate of the performance of the portion of the overall seal
system designed to protect the consolidating granular salt.
9.7 CONCLUSIONS
A borehole seal design should strive for a seal system which is a permanent
solution. This approach is necessary to satisfy existing or future regulatory
requirements and to obviate the need ever to perform remedial measures on
a borehole seal system.
Sealing boreholes in rock salt requires an understanding of the host
rock's response to the drilling of the borehole. Rock salt dilates and be-
comes permeable in the region immediately adjacent to the borehole. This
region can be the most permeable component of the seal system, and thus
can dominate its performance. However, as the host rock salt creeps inward
and contacts the seal materials, the resultant stress build up tends to 'heal'
the rock to a condition comparable to that before excavation.
Granular rock salt has the potential to be the ultimate seal material for
sealing penetrations in rock salt. Granular rock salt consolidates in a
borehole which is closing from creep, reducing its porosity and permeability
to that comparable to intact rock salt. Granular rock salt is chemically
compatible with the host rock salt and therefore should be effective indefi-
nitely. It is readily available and inexpensive.
In order to design an effective granular salt seal, the following informa-

tion is required: (1) borehole closure rate, (2) extent and magnitude of the
disturbed rock zone, (3) water inflow rate and (4) emplaced density of the
granular salt. Experience has shown that borehole closure estimates need to
be derived on a case by case basis. It should be well established that the
borehole is closing at a measurable rate before emplacing a granular salt
seal. The disturbed rock zone is expected to extend 0.5-1 borehole radius
into the formation and to have a permeability which can be as great as
1O-16 m 2. Water inflow into the borehole can come from overlying or
underlying water bearing zones, or from the rock salt itself. Conventional
techniques such as pumping tests or drill stem tests can be used to estimate
water inflow rates of producing formations. Because water inflow from the
host rock salt is typically not measured, it may be desirable to use the
limited inflow rates available from the literature. The emplaced density of
the granular salt is a function of the emplacement technique. Techniques
being developed which employ downhole compaction offer the best chance
of constructing seals with minimum porosity and consequently will achieve
a low permeability with minimal borehole closure and time.
High-quality cementitious seals are an important part of the design of an
effective borehole seal system in rock salt. These seals will isolate the con-
solidating granular salt from water-bearing zones above and below the salt
formations. These seals also provide redundancy and can be used to seal
across interbed layers of anhydrite, clay or other materials, if desired or
required during seal construction. Clays, asphalt and epoxy are other ma-
terials which can serve as effective components in a seal system. The longev-
ity of these materials is difficult to estimate, but is probably of the order of
hundreds of years.
9.8 ACKNOWLEDGEMENT
The research and experience of the authors was developed while working
for Sandia National Laboratories under contract to the US Department of
Energy.
CHAPTER TEN
Design of Underground Plugs

F. A. Auld
10.1 INTRODUCTION
To sink mine shafts and drive inclined drifts, underground roadways or

tunnels successfully, experience and skill is needed to maintain excavation
stability and to deal with and control groundwater. The presence of the
latter is possibly the most serious threat to working in the underground
environment and the miner must always operate with care when approach-
ing known zones of water-bearing strata.
During development work in shafts and tunnels, techniques are available
whereby strata water can be controlled temporarily prior to installing a
water-tight lining. Such methods are pumping, where the amount is not
excessive; pre-grouting of the strata for reducing water to within the avail-
able pumping capacity; and freezing, if excessive amounts are expected.
Before commencing development work hydrogeological boreholes are
normally drilled from the surface to locate the water-bearing zones approxi-
mately. Pressure recovery tests are carried out within the boreholes are to
provide data for estimating water inflow quantities which could be expected
during excavation. Subsequently, forward probe drilling is carried out prior
to each section of excavation to locate the water exactly.
Such procedures allow development to take place safely irrespective of
the presence of water. However, it is not always possible, or economic, to
provide fully watertight linings for shafts and tunnels and, throughout the
life of the underground system, ground relaxation and stress readjustment
may allow further ingress of groundwater.
Accidental inrushes of large quantities of water are also a potential haz-
ard if mining takes place too close to undetected sources and ground insta-
bility occurs, or if drilling interconnects with unexpected water-bearing
zones. Therefore, it can be seen that, in many cases, water will be prevalent
in underground workings, whether it is expected or unexpected, and the
means must be provided for sealing off areas of the workings either for

226 Design of underground plugs
temporary water control while pumping to disposal or on a permanent

basis. Plugs of concrete with a designed specific length and which fill the
shaft or tunnel cross-section are used for this type of sealing.
The design of underground plugs is well documented for the gold mines
of South Africa where reasonably hard rock and relatively high water press-
ures are experienced at deep levels (Garrett and Campbell Pitt, 1958, 1961;
Lancaster, 1964). However, very little new information has been forthcom-
ing since 1964 and published design data concerning other situations in
softer rocks and with lower imposed hydrostatic pressures are virtually
non-existent.
This chapter therefore sets out to review underground plug design, with
the object of bringing the subject to prominence and more up to date. An
attempt has also been made to rationalize the design process in relation to
current practice. The following five sections of this chapter consist of a
description of various types of plug; a discussion of the factors to be con-
sidered in plug design; detailed design calculations; construction aspects;
and plug sealing and resistance to leakage.
To elucidate the contents of these five sections more fully, the sixth
section comprises three case studies of actual plugs. Based on the overall
concepts contained in this chapter, conclusions and recommendations for
plug design are formulated at the end.
10.2 TYPES OF PLUGS
Four different catagories of underground plugs can be defined: (1) pre-

cautionary plugs, (2) control plugs, (3) emergency plugs and (4) temporary
or consolidation plugs. Basic descriptions follow outlining the functions
of each type.
10.2.1 Precautionary plugs

The plugs are normally constructed in underground roadways to limit the
area of flooding should water inrushes occur. Watertight doors are built
into them which can be shut when any danger of flooding arises. Pre-
cautionary plugs are installed as a safety measure prior to development in
areas known to be potential water-bearing zones and such plugs are design-
ed to withstand full hydrostatic pressure from surface level.
10.2.2 Control plugs

Sealing off or controlling the inflow of water from abandoned mining areas
involves the introduction of control plugs. Plugs constructed in boundary
pillars between adjacent mines also fall into this category. They are referred
Factors to be considered in design of plugs 227
to as boundary plugs and serve to prevent water flowing from abandoned

areas of one mine into the workings of an adjacent mine.
No means of access to the sealed off areas is provided through control
plugs but normally drain pipes, with valves, are cast into them. These plugs
are designed to resist full hydrostatic pressure from surface level or the
pressure imposed by the head of water to the highest overflow point.
10.2.3 Emergency plugs

Plugs of this type are constructed to seal off unexpected inrushes of water
either temporarily or permanently. No means of access to the sealed off
areas is provided in such plugs and they are usually designed to withstand
full hydrostatic pressure from surface level.
10.2.4 Temporary or consolidation plugs

Plugs which allow inflow water to be controlled or stopped while simulta-
neously providing the resistance for high-pressure grouting and consolida-
tion operations are known as temporary or consolidation plugs. They are
normally removed after the water pressure zones are sealed. Full hydro-
static pressure from surface level may again be the dominant design par-
ameter for these plugs.
10.3 F ACTORS TO BE CONSIDERED IN DESIGN OF PLUGS
When designing underground plugs the following factors need to be con-

sidered: (1) the purpose for which the plug is to be constructed; (2) the type
of excavation in which the plug is to be installed (shaft or tunnel); (3) where
the plug is to be sited in relation to the prevailing rock and working
conditions; (4) plug shape; (5) head of water to be withstood by the plug; (6)
the condition of, and the stress in, the rock surrounding the plug; (7) the
strength of, and stresses in, the material of the plug; and (8) the method of
plug construction.
10.3.1 Purpose
Each of the four categories of plug described above has a different specific
function and the form of a particular plug will be dependent upon the
prevailing situation.
10.3.2 Type of excavation

Undisturbed ground stress conditions alter locally in the areas surrounding
an excavation. The adjusted stresses differ, depending upon whether the
excavation is for a vertical shaft or a horizontal tunnel. A more uniformly

distributed stress occurs around a shaft excavation whereas, for a tunnel,
the vertical ground pressure may be different from the horizontal, causing
stress variations around the perimeter. In highly stressed ground a fracture
zone may surround the excavation and its extent will also depend upon
whether it encompasses a shaft or a tunnel. Therefore, the installation of a
plug in a shaft will require different design considerations than for construc-
tion in a tunnel.
10.3.3 Location of plug

One of the most important factors in deciding where to place a plug is the
condition of the surrounding rock. Preferably, the ground should be free
from geological disturbances which could provide leakage paths for water.
However, there could be limitations for the choice of site and the presence
of faults or dykes in the immediate vicinity may have to be accommodated.
It is not advisable to site plugs in or near the fracture zones of highly
stressed ground resulting from mining excavations although it is probably
impossible always to avoid such situations. Control plugs may have to be
located near mined-out areas to restrict outflow of water and the distance of
boundary plugs from the workings depends upon the width of the boundary
pillars in which they are installed. Boundary plugs need careful inspection
at all times as boundary pillars which are too thin will be pervious to water
and the danger of plug failure could be present under high hydrostatic
pressures.
Plugs should be sited in ground which is not likely to be affected by
subsequent gound movements resulting from mining operations. Damage to
both the plug and the surrounding strata would annul the grout sealing
integrity and introduce fresh leakage paths. The prevailing working condi-
tions could also influence the choice of plug site. When there is an inrush of
water, depending on the amount of water flowing into the workings, prefer-
ence will be shown for a site which can be temporarily dammed upstream,
providing relatively dry construction conditions for the plug.
Ventilation would be another criteria to be considered, particularly for an
underground environment where high temperatures prevail. However, in
an emergency an adequately ventilated site may not necessarily be forth-
coming.
10.3.4 Plug shape

Three basic forms of solid plug can be considered (Figure 10.1). The first
consists of a thin reinforced concrete wall (Figure 1O.la) or unreinforced
arch (Figure 1O.lb) keyed into the excavation all around the perimeter in
contact with the ground. Design of the slab involves calculation of bending
moments, shear forces and axial forces, sufficient strength being incorpor-
Factors to be considered in design of plugs 229
---
Possible water leakage paths Possible water leakage paths
through strata through strata
, ; - - .........
*- ....
* ~~..~...~~.
JllIYii~II!l. ".' ~~
. !.
..
Water
pressure
..
Water
pressure
", ........ _-_/

Possible water leakage paths Possible water leakage paths
through strata through strata
(a) (b)
...Water
--
pressure
\' 1l\t".~''1I\",,,. In\\\'~
(e) (d)
Steel bulkhead door Steel bulkhead door
(I (I. .. o' 0
.0'
O. D ,'o~' ~ ...0' o.
lIS
(e) (f)
Fig.lO.l Plug shapes. (a) Reinforced concrete slab in rectangular opening

(adequate strength but insufficient leakage resistance). (b) Unreinforced concrete
arch in rectangular opening (adequate strength but insufficient leakage resistance).
(c) Unreinforced concrete tapered plug in rectangular opening (adequate strength
and leakage resistance but uneconomical). (d) Unreinforced concrete parallel plug in
rectangular opening (economical with adequate strength and leakage resistance). (e)
Unreinforced concrete cylindrical parallel plug, with human access, in circular open-
ing. (f) Unreinforced concrete cylindrical parallel plug, with roadway access, in
circular opening.
ated in the structure to resist the applied pressure. The amount of keyed-in
area is related to the bearing resistance of the surrounding ground. A solid
plug of the second type possesses a longer length, no reinforcement and
incorporates a taper to provide the ground-bearing area (Figure 10.1c). Par-
allel plugs are the third type (Figure lO.1d) and resistance to the applied end
hydrostatic pressure is achieved through mechanical interlock with the
rough excavation face of the surrounding rock.
Garrett and Campbell Pitt (1958, 1961) consider plug length to be gov-
erned more by leakage resistance around the sides and through the sur-
rounding rock than by structural strength. The longer length required for
leakage sealing also ensures low shearing or bearing stresses at the con-
crete-rock interface. Thin barriers, although economic on materials, have
very short, unsealable leakage paths at their extremities and so are not
suitable for underground plugs. Tapered plugs, when compared with paral-
lel plugs, require more rock excavation, which introduces further rock de-
stressing, extra construction time and added cost. Increased quantities of
concrete are involved and tapered plugs are subjected to larger pressures,
resulting from the greater projected end area at the maximum cross-section
dimensions. Such factors are a disadvantage when plugs are required to be
installed under emergency conditions. Although the leakage resistance
paths are adequate with tapered plugs, the other factors are prohibitive.
Garrett and Campbell Pit (1958, 1961) have reported the results of tests
in South Africa on an experimental plug at West Driefontein, and on a
Virginia/Merriespruit boundary plug, which show conclusively that parallel
but rough-sided excavations will retain a plug without any sign of failure
under very heavy load conditions. On this basis, all further discussion on
plugs in this chapter is focused predominantly on parallel plugs. The section
on plug length based on bearing strength of concrete or rock at the interface
(discussed later) does however contain tapered plug design theory. Solid
plugs installed in shafts or tunnels will have a cross-section of the excava-
tion shape in which they are constructed. Shaft plugs will generally be
circular in cross-section whereas for drifts, roadways or tunnels the shape
may be square, rectangular, D-shaped, circular or otherwise. For
precautionary plugs with access ways through them, either purely for
human entry (Figure lO.le) or roadway access for materials transportation
(Figure 1O.1f), a different concept is required.
To resist high strata grouting pressures, which are applied in the trans-
verse direction for leakage sealing purposes, only the circular shape pro-
vides adequate strength. Precautionary plugs with access through them
therefore need to be in the form of concrete cylinders with sufficient length
for leakage resistance, adequate mechanical interlock automatically being
provided. In plugs incorporating access roadways the dimensions required
for clearance govern the inner diameter while strength to resist radial grout
pressure determines the wall thickness. These two criteria apply for part of
the length in a plug which is only provided with human access, as a structural
F actors to be considered in design of plugs 231
concrete wall can be incorporated integrally with the concrete cylinder at

the upstream face (Figure lO.le). In this case the strength of the wall is
adequate, the concrete cylinder acting as a sufficiently long sleeve to provide
leakage resistance and mechanical interlock with the surrounding ground.
In addition to the concrete cylinder, two other steel components are
necessary for the successful operation of a precautionary plug with an
access way: the bulkhead door for sealing off the plug in an emergency and
a load transfer cylinder (Figures lO.1e, lO.lf). The steel load transfer cylinder
allows the bulkhead door pressure to be carried by the concrete cylindrical
plug through bearing on the ring flanges. Enough flanges are provided to
reduce the bearing stresses to permissible limits. Load-transfer cylinders
must also be designed with sufficient wall thickness to permit their interface
with the concrete to be grouted up to the required pressure for sealing off
water penetration resulting from the hydrostatic pressure on the plug.
10.3.5 Head of water to be withstood

For the majority of plugs the design head of water will be that from ground
surface to the level of plug installation. This should be taken as normal for
all designs unless very clearly defined lower overflow levels are shown to
exist below ground surface which produce heads of water that cannot under
any circumstances be exceeded.
10.3.6 Condition of, and the stress in, the surrounding rock
The successful sealing of water flow by the introduction of a plug depends
on the capacity of the surrounding rock to prevent leakage. Any discontinu-
ities in the strata will make the task of sealing off more difficult. Fissures of
geological origin or fracture planes resulting from high ground stress could
endanger plug performance and installation of plugs in such areas should be
avoided wherever possible, as indicated in section 10.3.3.
The type of rock in which a plug is constructed is also a very important
factor in governing how well leakage paths can be sealed or how efficient
the shearing resistance or bearing capacity will be along the concrete - rock
interface. The presence of weak beds of shale, clay, sandstones or conglom-
erates will increase leakage potential and reduce interface shearing resis-
tance or bearing capacity.
As already indicated, undisturbed ground stresses alter once excavation
takes place and the magnitude and variation of such stresses around an
opening differ for shafts and tunnels. High ground stresses, which cause
rock fracture, depend upon the following factors: (1) the depth below
the surface, (2) the size of the opening, (3) the proximity of other mining
excavations and (4) the proximity of geological disturbances which may

introduce tectonic stresses.
The subject of stress evaluation around underground openings is a com-
plex one and is too large a topic to be introduced into this chapter. Never-
theless, it is a subject which must be fully understood if a true evaluation of
concrete plug-rock interaction is to be formulated and more research into
this area is required.
10.3.7 Strength of, and stresses in, plug material

Five points warrant consideration when evaluating the stresses in, and
strength of, underground plugs: (1) concrete compressive strength, (2)
the early-age development of strength, (3) the shear or bearing stress at the
plug-rock interface, (4) the pore-water pressure in the concrete and (5) the
possible end spalling of the plug due to high stresses set up by ground
pressure.
Provided the recommendations of current Codes of Practice for struc-
tural concrete (in the UK, British Standards Institution, 1969b, 1985, 1987)
are followed, with Grade 25 concrete (characteristic strength 25 N mm - 2) as
the minimum specified requirement, then dense, impermeable and durable
concrete ought to be achieved easily. On this basis, in conjunction with the
length required for sealing which ensures low stresses, no problems of
strength should be encountered.
Early-age strength development is important from two aspects. First, it is
essential that plugs develop their specified strength without any detrimental
effects occurring from shrinkage, thermal changes or ground pressures. Pro-
vided care is taken to overcome these factors, the integrity of the concrete
mass will be protected and leakage paths through plugs minimized. The
second aspect is how quickly a plug needs to be sealed. It is possible to use
higher-strength concrete mixes than are required purely for design strength.
This allows higher strengths to be achieved at earlier ages and hence the
problem of sealing can be tackled more quickly.
The factor of safety against shearing or bearing failure in the rock or
concrete of the plug at their interface depends upon the magnitude of the
induced stresses, which in turn is related to plug length. Since the length of
a plug should be determined with leakage sealing in mind, which means
providing a longer length than is necessary for structural strength purposes,
relatively low interface stresses are inherent in good plug design. Knowledge
of pore-water pressure behavior within a concrete plug is limited. A press-
ure gradient exists from the hydrostatic pressure at the face in contact with
the impounded water to zero at the opposite end. How the pressure and
induced stresses are dissipated throughout the system and into the sur-
rounding ground is a matter for conjecture at the present time and this
area, in conjunction with rock stress evaluation, needs further research.
Theoretical stress distributions in shaft plugs have been determined by Sitz,
Design calculations 233
Koeckritz and Oellers (1989) using finite element analysis. However, it is

unlikely that spalling of the free face of a plug will occur due to pore-
water pressure unless non-homogeneous irregularities occur in the concrete
mass.
It is conceivable that high localized ground stresses at the ends of plugs
could cause spalling at these points, reducing the effective resistance to
applied pressure and leakage. Careful choice of site related to a study of the
induced rock stresses and rock strength can avoid or minimize this risk.
Additional plug length would also contribute to solving this problem.
10.3.8 Method of plug construction

For precautionary, control and temporary or consolidation plugs, which
can generally be constructed in phase with and under normal mine operat-
ing conditions in a relatively dry environment, the method of construction
has little influence on design. However, in conditions of emergency, ma-
terials access problems and water inflow quantities may require consider-
ation of different methods of construction. Normal concrete transportation,
placing and compaction can be replaced by grouted concrete in which a
mixture of cement, sand and water is introduced into pre-placed aggregate
(Chamber of Mines of South Africa Code of Practice, 1983). This technique
is particularly suitable for the construction of plugs in areas where access is
difficult or for plugs installed under water in flooded shafts. Concrete can
also be placed by tremie under water if necessary. Resulting from the chosen
method of construction, different concrete-rock interface allowable shear or
bearing stresses may have to be used, depending upon how dense and
impermeable the plug mass is expected to be and how integral a contact can
be achieved with the surrounding work.
10.4 DESIGN CALCULATIONS
10.4.1 Formulae for calculating plug length and strength

Plug length based on shear strength of concrete or rock at the interface
Garrett and Campbell Pitt (1961) quote the following formula which can be
applied to parallel-sided plugs with rectangular cross-sections if interface
shearing is accepted as the governing failure mechanism:
pbh = 2(b + h) IPpe' (1O.1a)
where p is the intensity of applied pressure, b is the width of the plug, h is

the height of the plug, I is the length of the plug and Ppe is the permissible
punching shear stress of the rock or concrete at the interface.
By transposing Equation (1O.1a), the length of plug can be obtained:
1= pbh (1O.1b)
2(b + h) Ppe
For a square cross-section, Equation (1O.1b) becomes
1= pb/4p pe (lO.1c)
The length of circular plugs, of radius r, can be obtained from

pnr2 = 2nrlp pe ' (lO.2a)
giving
(1O.2b)
Plug length based on bearing strength of concrete

or rock at the interface
Although the shear strength concept of the previous section can be em-
ployed, Garrett and Campbell Pitt (1961) also considered that, alternatively,
the mechanism of interaction between concrete plugs and the surrounding
rock could be more in the form of direct bearing rather than shearing at the
interface. Mechanical interlocking action is achieved at an excavation face
through the various inclined planes of its surface. Orientation of these
planes can be in any direction lying between the extremes parallel with or
normal to the general direction of the excavation face. An assumption can
be made that half the inclined planes resist movement by direct bearing
(Figure 1O.2a) while the others are subjected to tensile stresses and therefore
can be neglected.
For a parallel plug, consider an element of the excavation face ABC
(Figure 1O.2b) with a horizontal length AC = 1', which contributes to the
plug bearing resistance over the element of length 1'/2. The permissible
bearing stress in the concrete or the rock is Pbe and F~ represents the total
bearing resistance over the element of plug bearing length BC, inclined at
an angle of ()( to AC such that
BC cos ()( = 1'/2. (10.3)
In the triangle of forces (Figure 1O.2b) P' is the element of applied horizon-
tal force which is resisted by the horizontally resolved component of F~.
Therefore
P' = F~ sin ()(; (10.4)
however
(10.5)
L - _ - - ' - - _ - - L_ _.L-_.....L_Total effective resistance

(a) length = i
-Compression component
ex t of force on plug
AII+,______ ~_D ________ ~

(b)
Fig. 10.2 Evaluation of parallel plug length based on bearing strength of concrete
or rock at the interface. (a) Plug bearing resistance. (b) Element of plug bearing
resistance. (Garrett and Campbell Pitt, 1961.)
From Equation (10.3)

Be = [1/(2 cos IX) (10.6)
and combining Equations (lOA), (10.5) and (10.6) gives

pi = (PbJ/2) tan IX (10.7)
Summing all the forces on the plug results in

[' [
L pi = P = pbh = L Pbe 2: tan IX = 2(b + h) 2: Pbe tan IX (1O.8a)
from which
[= pbh (10.8b)
(b + h) Pbe tan IX
Since the surface planes will be inclined at angles of between 0 and 90 to 0
the direction of thrust, Garrett and Campbell Pitt (1961) considered the
assumption that the average inclination IX = 45 for a parallel-sided plug
was justified. Equation (1O.8b) becomes
1= pbh (1O.8c)
(b + h)Pbe
For a square cross-section, Equation (10.8c) reduces to
1= pb/2Pbe' (1O.8d)
The length of circular parallel plugs can be obtained from
(1O.9a)
glvmg
[= pr/Pbe (1O.9b)
for IX = 45. Tapered plugs can also be considered if appropriate amend-
ments are made to Equations (10.3)~(1O.9) (Figure 10.3). The element of
bearing length BC (Figure 10.3b) is now inclined at an angle IX + {3 to the
horizontal and
BC cos IX = ['/2 cos {3, (10.10)
where (3 is the angle of plug taper. From the triangle of forces
P' = F~ sin(IX + (3); (10.11)
however
['
F' = P BC = P (10.12)
b be be 2 cos IX cos (3 .
Combining Equations (10.11) and (10.12) gives

, [' sin(IX + (3) ['
P = Pbe ~ {3 = Pbe ~ (tan IX + tan (3). (10.13)
2 cos IX cos 2
l'
I P' = P = pb max hmax = I Pbe"2 (tan IX + tan (3)
1
= 2(bay + hay) "2 Pbe(tan IX + tan (3), (10. 14a)
where bmax is the maximum plug width at the water face, hmax is the maxi-
mum plug height at the water face; bay is the average plug width along its
length and hay is the average plug height along its length.
.. Water pressure x end

area = total force (p)
(8)
B
Compression component
of force on plug
f (1 - tana tan~)
(b)
Fig. 10.3 Evaluation of tapered plug length based on bearing strength of concrete
or rock at the interface (developed from Garrett and Campbell Pitt, 1961). (a) Plug
bearing resistance. (b) Element of plug bearing resistance.
From Equation (1O.14a)
(1O.14b)
For rJ. = 45 Equation lO.14b becomes
(1O.14c)
and for a square section
(1O.14d)
The length of circular tapered plugs can be obtained from
pnr!ax = n(r max + rmin) [( "2

1)2 + (r max - rmin)2
J1/2
X Pbe(tan rx + tan /3), (1O.15a)

where rmax is the maximum plug radius at the water face and rmin is the
minimum plug radius at the face remote from water. Equation (1O.15a) gives
2
1= 2 [ P (
r!ax
2 2 /3)2 - (r max - rmin)
2J1 /2 (1O.15b)
rmax + rmin) Pbe (1 + tan
for rx = 45.
An alternative form of bearing calculation for tapered plugs is that for a
smooth-faced wedge driven into an opening. On this basis the whole surface
area acts in bearing and the element of bearing length becomes AC (Fig-
ure 10.3b), inclined at an angle /3 to the horizontal where
AC cos/3 = l' (10.16)
and
P' = F~ sin /3; (10.17)

however
F~ = Pbe AC = Pbe1' /cos /3. (10.18)

Combining Equations (10.17) and (10.18) gives
P' = Pbe/' tan /3. (10.19)

Ip' = P = pbmaxhmax = I Pbe 1' tan /3 = 2(b av + hay) IPbe tan /3 (1O.20a)
from which
1= pbmaxhmax (1O.20b)
2(b av + haJ Pbe tan /3
is derived. For a square section
1= pb!ax (1O.20c)
4b av Pbe tan /3
Comparing Equations (1O.20b) and (1O.20c) with Equations (1O.14c) and
(1O.14d), respectively, if rx = 45 is replaced by rx = 0 in the latter two equa-
tions then compatibility is achieved, except for the anomaly of reducing the
length by half in the case of the wedge theory due to use of the full bearing
area.
The equivalent length of circular tapered plugs based on the smooth-

wedge principle can be obtained from
(1O.21a)
where
(10.21b)
Cylindrical plug strength

The strength of cylindrical parallel plugs (Figures 1O.le and 10.10 can be
determined using the standard Lame elastic design theory for thick cylin-
ders (Auld, 1979, 1982a):
2Pr(t + ri (10.22)
(J = t(t + 2r i) ~ Pc'
where (J is the maximum tangential stress in the concrete cylinder wall,

occurring at the inside face, Pc is the permissible concrete compression
stress, Pr is the externally applied radial pressure, r i is the inside radius of
the cylinder and t is the concrete cylinder wall thickness.
Bearing strength of cylinder walls

Plugs of the type shown in Figure 10.le, which carry load from a circular
face wall back through a cylindrical rear section and thence into the sur-
rounding rock, must have sufficient strength at the interconnection between
the wall and cylinder. The cylinder end area must be sufficiently large to
reduce the bearing stress imposed by the end wall to a value within the
permissible limit. Hence, the calculated concrete bearing stress,
(1O.23a)
where Pb is the permissible concrete bearing stress, ro is the outside radius of
the cylinder and P is the horizontal applied force on the cylindrical rear
section; i.e.
P= pnr; (1O.23b)
- [the concrete or rock permissible surface resistance over length 1* of the
front wall].
Combined stress at the interconnection between the

face wall and the cylindrical rear section
The cylinder stress and the bearing stress determined from the above sec-
tions act together at the interconnection to produce a combined compres-
sive stress situation. Care should be taken to ensure that the calculated
combined compression and bearing stress,
(1O.23c)
Punching shear of front wall

The punching shear resistance of the front wall against the cylindrical rear
section (Figure 1O.le) must also be adequate. Therefore, the calculated con-
crete punching shear stress,
fp = pnr;;2 nrr ~ Pp' (10.24)
where Pp is the permissible concrete punching shear stress.
10.4.2 Formulae for designing steel load transfer cylinders
Load transfer to concrete by flanges

The concrete bearing stress,
(10.25)
where rof is the outside radius of the load transfer cylinder flange, ro is the
outside radius of the steel cylinder wall and n is the number of flanges.
Flange bending
The bending moment is given by
f; (r of - r0) (r of - r0 + t). (10.26)
The steel bending stress,

3fb
f ms = -2 (rOf - rO)(rof - ro + t) ~ Pms' (10.27)
tf
where Pms is the permissible steel bending stress and tf is the flange thick-
ness.
Compression resistance to radial grouting pressure

This aspect is important for the sealing in of the steel load transfer cylinder
into the concrete plug to prevent leakage along the interface.
If d is the distance between the centers of the flanges, the flange cross-
sectional area is (r of - r0) t f and the cross-sectional area of the steel cylinder
wall between the flange centers becomes td.
Evaluating the effective membrane thickness for both the cylinder wall
and the flange gives the effective membrane thickness per unit length,
to = [n(rof - ro) tf + (n - 1) tdJ/I. (10.28)
The hoop stress for the steel load transfer cylinder, fes' can be evaluated
using Equation (10.28):
(10.29)
where Pes is the permissible steel compression stress. To enable the cylindri-
cal wall and flanges to act as a composite structure, any welding of flanges
to the cylinder must be capable of carrying the interaction stresses.
Direct bearing of end load on cylinder wall

The bearing area of the cylinder wall must be capable of transferring the
end load from the bulkhead door to the bearing flanges without overstress-
ing, i.e. steel bearing stress,
P1tr~f
f bs = 1t(r~ _ rf) ~ Pbs
(10.30)
where Pbs is the permissible steel bearing stress.
Resistance to axial compression

In addition to end bearing, the cylindrical wall must be capable of acting as
a column between the flanges to allow flange bearing to be effective.
Timoshenko and Gere (1961) give the critical load, Per' for a column
with built-in ends, which is a conservative approach for a cylinder wall as it
neglects additional strength due to curvature, as
(10.31)
where E is the modulus of elasticity for steel (2.1 x 1011 Pa) and I is the
moment of inertia of the cylinder wall equal to t 3 /12 per unit length.
For the cylinder, multiplying Per by the circumference and dividing by the
total applied end pressure will produce the factor of safety of
(10.32)
Provided the flanges are not spaced too far apart, satisfying the criterion of
the section on 'compression resistance to radial grouting pressure' will auto-
matically produce a large factor of safety in Equation (10.32).
10.4.3 Bulkhead door design

Generally, bulkhead door pressures will be relatively large and therefore the
best shape to resist the load is a spherical segment. The shell thickness for
such doors can be determined on the basis that the meridional and hoop
forces per unit length of the shell are equal to pal2 (Fliigge, 1967), where a is
the radius of curvature of the shell. Dividing this value by the shell thick-
ness, t, gives the compressive stress in the steel, f es' as
fes = pal2t ~ Pes (10.33)
It should be noted that thin shell domes are prone to buckling, and
stiffening for the door should be provided to avoid any possibility of insta-
bility under load. The subject of steel bulkhead door design lies in the
specialist field of pressure vessels and is outside the scope of this chapter.
Operation and sealing of such doors are prime parameters to be considered
in design and recourse should be made to specialist design and fabrication
manufacturers for the supply of such elements.
10.4.4 Rock, concrete and steel permissible stresses
Permissible shear and bearing stresses for rock and concrete

at the plug interface
The proposed formulae for determining the length of plugs, either on the
basis of shear strength or on one of the two bearing strength philosophies
are a very simplified form of a much more complex stress system. Both the
rock and the concrete are in a confined state along their interface. The
compression strength of concrete in the UK is quoted on the basis of
150mm cubes tested at 28 days in an unconfined compression testing ma-
chine (British Standards Institution, 1983). It is known that concrete, when
tested in a confined state, shows an increase in strength over the unconfined
condition (Jaeger and Cook, 1979). The confining action of the surrounding
rock against the concrete plug could modify the bearing force calculated by
the formulae in the section entitled 'plug length based on bearing strength
of concrete or rock at the interface'. However, the true resistance probably
lies somewhere between that given by the bearing capacity and the resis-
tance provided by shear. Shearing in this context would be of a punching
nature, as opposed to the traditional structural engineering form of beam
shear, and even this could be modified, depending upon the magnitude of
the interface confining stresses. Hence, the ultimate validity of the permis-
sible stress values for shearing and bearing will depend upon the effective-
ness by which the concrete of the plug is confined by the surrounding rock.
The plug concrete can be considered as a homogeneous material on the
assumption that good construction practice has been observed. However,
the surrounding rock will be anything but homogeneous, being cracked and
fissured before excavation takes place. Destressing also occurs during and
subsequent to excavation and therefore, when grouting and hydrostatic
pressures are applied to the rock, movement inevitably will occur. The
direct strains will be accompanied by movement in the direction of cracks
and bedding planes and the effectiveness of the confining action will be
dependent on this movement.
Irrespective of which theory is applied to define the stress conditions, the
governing factor remains the stress in the rock. As indicated above, more
research is needed to understand how the stresses in the surrounding rock
are modified by confined plugs subjected to end pressure. Until this aspect
is investigated in detail the validity of any formulae utilized in defining plug
and rock stress conditions will be in question. At the present time, with the
formulae available, it will be necessary to check the shear and bearing
stresses for both the concrete and the rock and to base the design on the
weaker material.
Concrete permissible stresses are contained in Table 10.1, based on the
current UK Codes of Practice (British Standards Institution, 1969b, 1985).
The values are all related to the concrete characteristic strength, this being
the lower limit below which not more than 5% of the cube test results
would fall based on a statistical analysis of samples tested. Both of the
Codes of Practice are specifically for reinforced concrete and neither treats
the unreinforced concrete situation realistically, particularly with regard to
punching shear philosophy. However, Manning (1961) quotes the safe punc-
hing shear stress to be about one-fifth of the safe compressive stress and this
has been included in Table 10.1. The maximum allowable values for
Pc, Pb' Pbe' Pp and Ppe are heavily outlined in Table 10.1 as these are the
suggested values to be adopted in design. The reason for using a factor of
safety equal to 4 for Pbe and Ppe is explained below.
It is much more difficult to propose realistic permissible stresses for rock.
The strengths of rocks are normally determined by testing cylindrical sam-
ples and, as a result of the non-homogeneity of the material, it is normally
only the best pieces from which the samples are obtained. It must always be
remembered that strengths of rocks which are determined from such testing
will not be typical of the actual strength in situ and appropriate adjustments
should be made to allow for this.
Assuming that the grouting process for strata water sealing is carried out
methodically and conscientiously, most of the rock bedding planes and
fissures local to the plug interface should be filled and consolidated. This,
allied with the confining action of the surrounding rock, could allow the
lower-strength concrete permissible stresses to be taken as being representa-
tive of the rock also. For the purposes of design, this would be an alterna-
tive if no actual data were forthcoming. The practice in South Africa is to
use a permissible shear stress value of 0.59 N mm - 2(85lb in - 2) for concrete
placed in the normal manner and 0.83 N mm - 2(120 lb in - 2) for grouted
Table 10.1 Concrete permissible stresses
Type of stress
Bending Direct Concrete-rock Beam shear Beam shear Punching Concrete-rock

compression compression interface (Nmm-2) (Manning, shear interface
Pc (Nmm- 2) or bearing, bearing a , 1961) 0.1 Pc (Manning, punching
Pb (Nmm- 2) Pbe( = 3.75 Ppe) (Nmm- 2) 1961) sheara Ppe
(Nmm- 2) Pp = 0.2 Pc (Nmm-2)
(Nmm- 2)
CP114:1969 fcul2.73 0.75fcul2.73 0.75fcul4 (fcul50) + 0.27 0.2fc)4

value b (Maximum
= 0.90)
Grade 25 9.16 6.87 4.69 0.77 0.92 1.84 1.25

(characteristic 30 10.99 8.24 5.63 0.87 1.10 2.20 1.50
strength,fc) 35 12.82 9.62 6.56 0.90 1.28 2.56 1.75
40 14.65 10.99 7.50 0.90 1.47 2.94 2.00
45 16.48 12.36 8.44 0.90 1.65 3.30 2.25
50 18.32 13.74 9.48 0.90 1.83 3.66 2.50
55 20.15 15.11 10.31 0.90 2.02 4.04 ~
BS811O:Part 1: 0.67fcul 0.4fcul1.4 0.67fc)4 (0.2 x 0.67fc)/4
1985 value c (1.4 x 1.5
Grade 25 7.98 7.14 4.19 Not represen- 0.80 1.60 0.84
(characteristic 30 9.57 8.57 5.03 tative of 0.96 1.92 1.01
strength,fc) 35 11.17 10.00 5.86 unreinforced 1.12 2.24 1.17
40 12.76 11.43 6.70 concrete in 1.28 2.56 1.34
45 14.36 12.86 7.54 plugs 1.44 2.88 1.51
50 15.95 14.29 8.38 1.60 3.20 1.68
55 17.55 15.71 9.21 1.76 3.52 1.84
a Based on a factor of safety = 4; h British Standards Institution (1969b) (withdrawn); C British Standards Institution (1985).
concrete where positive contact between the concrete and the surrounding
rock is assured by subsequent grouting (Lancaster, 1964; Chamber of Mines
of South Africa Code of Practice, 1983). This is a general rule and is not
related specifically to concrete or rock strength; neither does it take into
account the rock condition. The values are therefore unrealistic, particularly
with regard to the increased concrete strengths currently being achieved in
underground construction, due to the improved workability and quality
control procedures adopted in conjunction with better batching, transporta-
tion and placing techniques. Therefore, it is considered that the values in
Table 10.1 are more appropriate.
Permissible concrete stresses other than at the plug interface

with the rock
These are also covered by Table 10.1.
Permissible steel stresses for load transfer cylinders

and bulkhead doors
Typical permissible steel stress values for steel (Grade 43) are contained in
Table 10.2. These are taken from the current UK Code of Practice for the
use of structural steel in building (British Standards Institution, 1969a).
10.4.5 Factors of safety

The factors of safety for structural concrete quoted by the UK Codes of
Practice (British Standards Institution, 1969b, 1985) are given in Table 10.1.
CP 114:1969 (now withdrawn) introduced a value of 2.73 to relate the
characteristic strength to the permissible compression stress in bending, pc'
BS 8110: Part 1:1985 is more specific in its breakdown of safety factor. The
actual compression strength of concrete in a structure is equivalent to
0.85 x characteristic cylinder strength (Comite Europeen du Beton, 1970),
where for plug construction the 0.85 factor takes account of the difference
between instantaneous loads on cylinders at an age of 28 days and loads
Table 10.2 Permissible stresses in steel (Grade 43)a

Thickness Type of stress
(mm)
Bending, p Axial compression, Bearing, Pbs
(Nmm- 2 ) illS pcs(Nmm- 2 ) (Nmm- 2 )
~40 165 155 190

>40 150 140 190
a(BS449: Part 2: 1969, British Standards Institution, 1969a).
applied for a longer duration on specimens of the same age. Since British
Standard practice (BS 1881: Part 116:1983) uses the 28 day cube test as a
means of strength control, a correction factor of 0.8 is needed to convert the
cube strength to the equivalent cylinder strength. Hence, in relation to the
characteristic cube strength, feu' the actual in situ strength of concrete is
represented by 0.68 x characteristic cube strength, the value of 0.68 being
equal to 0.85 x 0.8. A figure of 0.67 is employed in BS 8110 : Part 1:1985.
Partial safety factors for load, Yf' and strength, Ym' are used in the ultimate
limit state approach to the design of concrete structures (BS
8110: Part1:1985). On the basis of it normally being of a long-term perma-
nent nature, the value of Yf for hydrostatic loading can be taken as 1.4. For
Ym , which is introduced to account for possible strength differences between
test specimens and the actual structure caused by such aspects as insuffi-
cient compaction and differences in curing, the specified value is 1.5. The
effective factor of safety in accordance with BS 8110: Part 1:1985 is therefore
1.4 x 1.5 = 2.1 when related to the actual strength of concrete in situ, or
(1.4 x 1.5)/0.67 = 3.13 when compared against the characteristic strength as
given by 28 day cube test results.
For the steel stresses in Table 10.2, the factor of safety to yield will be
approximately 1.5 with probably the same again to failure. This gives a
probable minimum factor of safety to failure equal to 2.25. The factors of
safety for concrete and steel which have been built into the permissible
stresses, in the range 2-3, are acceptable because the performance of the
material under load is well established and quality control ensures consist-
ency. For the mechanism of resistance at the plug-rock interface, the true
behaviour is not understood fully and the rock shear and bearing permis-
sible stresses cannot be established realistically. On this basis, and because
plugs are normally installed as a safety measure, it would be prudent to
adopt a higher safety factor when determining plug length using the shear
or bearing resistance criteria. A minimum factor of safety of 4 is recommen-
ded in line with South African practice (Lancaster, 1964) and this has been
introduced into Table 10.1 for the Pbe and P pe values.
10.5 CONSTRUCTION ASPECTS
10.5.1 Batching, transporting and placing concrete

In any concrete construction work it is necessary to have the right batching
plant, geared to the demand. This is particularly important for underground
construction, where pours must be completed with minimum interference
from external factors. Rates of pouring are governed by the physical restric-
tions in particular areas underground and by the problem of access for
materials to those areas. A surface batching plant is preferred because ag-
gregate and cement weighing, together with the metering and addition of
Construction aspects 247
water and admixtures, can be controlled in the most effective manner. How-
ever, such a system depends on being able to transport the premixed con-
crete underground and, in some circumstances, an underground bat ching
plant may be necessary. This does not relieve the problem of having to
transport the concrete mix constituents underground as separate items. For
grouted concrete, again a surface grout mixing set-up would be preferred.
Normally, with surface batching plants large quantities of concrete or grout
can be mixed and transported underground rapidly to give a constant
uninterrupted supply. This is particularly advantageous for the construction
of emergency plugs.
Current trends in UK mining development have favored the employment
of established ready-mixed concrete suppliers. By adopting such suppliers,
high quality is achieved through the utilization of their specialist expertise
in the production of concrete. Quality control measures for the use of
concrete underground are the same as those for surface works and are in
accordance with the relevant Code of Practice (BS 5328:Part 4: 1990). Inde-
pendent approved organizations are normally employed for cube testing, or
any other testing of the hardened concrete which is required.
The preferred method of transporting and placing concrete underground
is pumping. With shaft access, concrete can be dropped down a vertical pipe
for further transportation underground by pumps situated at the bottom of
the shaft. Drift access allows pumping from the surface, down the incline
and then further underground directly to the point of plug installation.
Once pipelines are installed, minimum interference with mining and plug
construction operations is achieved and large volumes of concrete can be
delivered and placed rapidly. Structural concrete mixes (Grade 25 and
above) can now be transported from the surface, down shafts or drifts,
directly to the point of placement underground through pipelines with small
diameters, typically in the order of 32-50 mm (Martin, 1989).
10.5.2 Concrete mix design

Similar to any concrete construction, care must be taken to provide the
right balance of ingredients, first of all to suit the particular mode of trans-
portation and placing being used, and secondly to make sure that the
optimum design is achieved. In addition to strength, the most important
factor of the mix design is to obtain the correct workability. For under-
ground work, with its restricted placing environment, it is essential that
high-workability mixes are used. In the author's opinion (Auld, 1982b,
1982c), successful construction with concrete underground is entirely de-
pendent upon the inclusion of plasticizing admixtures for transporting and
placing.
Two mix designs which have been used successfully by Cementation Min-
ing Limited are contained in Table 10.3, on page 258. Cement replacement
materials were incorporated to minimize the thermal effects which are dis-
cussed in more detail below.
10.5.3 Thermal and shrinkage effects

If it can be achieved, it is preferable to pour a concrete plug in one oper-
ation to avoid construction joints which are potential leakage paths
through the plug itself. Although normal structural concrete mixes (BS
5328: Part 1: 1990) could be used for plug construction, the large volumes
required for mass filling can be subjected to detrimental thermal effects
during setting. This is dependent upon the amount of cement included in
the mix. Internal buildup of heat within the mass due to the cement hy-
dration process could induce high thermal stresses. The strength integrity of
the structure would be impaired and, on cooling, thermal cracking may
result. By cooling the aggregates and mixing water, the ultimate tempera-
ture attained by normal mixes can be reduced but it is preferable, where
possible, to use cement replacement materials to minimize heat of hydration
gain. Table 10.3 contains two such mixes.
An additional factor which assists concrete thermal control is the embed-
ment of service, water control and grouting pipes in the plug mass. Heat
will be dissipated through these pipes, particularly in the case of temporary
or consolidation and emergency plugs if water is flowing through them.
Shrinkage will not normally be a problem with the designed concrete mixes
currently being employed in underground construction. The use of plas-
ticizers enables the water/cement ratio of the basic mix to be kept to a
minimum, therefore ensuring very low water loss during the curing stage,
which prevents excessive shrinkage. Three other factors also contribute to
shrinkage reduction. These are the underground environment, in which no
rapid drying out conditions normally prevail, the limited facial exposure to
drying elements in the environment and the relatively thick concrete sec-
tions used.
10.5.4 Construction points of detail
Excavation
Care should be taken during excavation to minimize damage to the sur-
rounding strata. Machine cutting and hand trimming is preferred to drilling
and blasting.
Plug installation
Two factors assist in reducing leakage paths at the concrete-rock interface.
Before beginning to pour concrete for a plug, the floor should be thorough-
ly cleaned to remove any debris or construction dust. At the roof of the
Plug sealing and resistance to leakage 249
plug, to ensure a tight seal, concrete must be discharged as high up as

possible and a crown feed pipe, which can be withdrawn as topping up
takes place from one end of the plug to the other, should preferably be
installed. Air bleed pipes, which subsequently can be used for contact zone
grouting, are also beneficial at roof level.
Grout seals
Where mass concrete is cast directly against rock it is necessary to grout up
the contact zone to prevent leakage through any shrinkage gaps. It is very
difficult to obtain full tight contact with the surrounding rock over a large
surface area with grouting. Therefore, it is preferable also to provide one or
more narrow chases, surrounding the plug cross-section completely, in
which grout can be injected and pressurized to provide a tight ring seal.
Temporary water control

For temporary or consolidation plugs and emergency plugs, control of shaft
water is essential to allow good construction. In roadways, floodwater will
need to be temporarily dammed upstream and the water led off through
valved pipes cast into the plug. Debris grills will need to be installed at the
upstream ends of pipes. Temporary consolidation plugs in shafts also need
to include vertical steel rising mains through which shaft inflow water can
be withdrawn during construction, to prevent pressurizing of the underside
of the plug whilst hardening.
Services
Pipes need to be installed in precautionary plugs to carry services. These
pipes should be fitted with sealing glands at each end for plug water tight-
ness when the bulkhead doors need to be closed.
10.6 PLUG SEALING AND RESISTANCE TO LEAKAGE
10.6.1 Grouting procedure

Design calculations can be carried out as shown previously to determine
plug dimensions. However, an integral part of the successful installation of
an underground plug is the means by which leakage past the plug is mini-
mized or eliminated. Grouting is the process by which this is achieved.
The science or 'art' of grouting depends very much upon the knowledge
and experience of mining development contractors, and cannot be discussed
in detail here. As a process, grouting consists of the pressurized injection of
cement or chemical grouts into the strata to fill voids, fissures, bedding
planes and any other anomalies in the rock surrounding a plug. Its purpose
is to seal off all water paths and grouting of the plug itself may be needed,
depending upon whether construction joints are incorporated and also
upon the standard of workmanship.
Injection of grout at the contact surfaces between the plug and the rock is
also necessary to fill shrinkage gaps, porous zones due to placing difficulties
and cracks in the rock adjacent to the plug due to destressing. Pressures of
up to twice (Garrett and Campbell Pitt, 1958) and 2.5 times (Garrett and
Campbell Pitt, 1961; Lancaster, 1964) the pressure which the plug has to
resist have been recommended for this grouting. These pressures are used in
the deep gold mines of South Africa where generally strong rocks and
relatively high water pressures are encountered. Even with localized fracture
zones around such excavations, opening up of the cracks under high press-
ures to accommodate the entry of grout is not detrimental. However, in
softer rocks at shallower depths, as in the UK coal measures, such pressures
would be damaging and are not to be recommended. Precautionary plugs
of the cylindrical type should only be stressed to a maximum of 1.25 times
the hydrostatic pressure, related to surface level, this being the value by
which the normal structural concrete permissible stresses can be exceeded
for short-term loading (CP 114: 1969). Hence, the post-stressing of the plug
and rock, which is advocated for the South African conditions (Garrett and
Campbell Pitt, 1958) will generally not be as effective in UK practice for
enhancing the confining action. The radial Poisson's ratio effect, resulting
from the end pressure, will also be less effective with regard to increasing
the interlocking resistance.
Leakage associated with plugs can occur at the following places: (1)
through the plug concrete, (2) along the concrete-rock interface, (3) through
the rock surrounding the plug and (4) along the interface between the plug
concrete and the steel load transfer cylinder if access through the plug is
provided.
10.6.2 Leakage through the plug concrete

Three possible reasons exist for leakage through the plug concrete: (1) po-
rosity of the concrete, (2) construction joints and (3) the presence of cracks.
The presence of highly porous concrete is very unlikely, due to the dense,
impermeable and durable mixes currently used in underground construc-
tion. High workability, achieved with the use of plasticizers, ensures full
compaction and with good quality control and careful placing techniques
there should not be any excess porosity problems. Grouting will help to seal
off the more porous zones if they do occur.
Wherever possible, construction joints should be avoided but, with the
current mix designs, if they are necessary, very little joint preparation is
required. Good bonding should be easily achieved between consecutive
pours.
Cracks in the concrete can be caused by excessive water pressure behind

the plug, thermal effects during setting and maturing or excessive stresses
and strains transmitted to the concrete by the surrounding rock. Provided
the design is processed correctly in relation to the applied water pressure,
adequate measures are introduced in the mix design to minimize the ther-
mal effects and the chosen plug site is competent, a homogeneous structure
can be constructed free from defects.
10.6.3 Leakage along the concrete to rock interface

Interface leakage could result from (1) shrinkage gaps at the interface, (2)
shear cracks at the interface due to plug movement under high water press-
ure, (3) cracks caused by excessive ground stress and (4) poor contact with
the surrounding rock, caused by debris and construction dust not removed
from the floor prior to casting and also as a result of air and water pockets
trapped at the underside of the roof.
As discussed in section 10.5.3, shrinkage in underground concrete should
be minimal. The grouting process also enables gaps caused by shrinkage to
be sealed up.
The possibility of plug movement will only occur if the plug length is too
short and, as mentioned previously, the capacity to seal leaks is the prime
factor in determining plug length. A longer length is needed to provide
leakage resistance than is required for structural purposes. This will ensure
that the interface shear stresses are sufficiently small to avoid any plug
movement under high hydrostatic pressure.
The radial adhesion between the plug concrete and the rock at the inter-
face will be small and highly stressed rock conditions could damage the
intimate contact. Judicious choice of the plug site could avoid such failure.
Good workmanship will prevent problems such as construction debris
not being removed prior to casting the plug concrete. Provision of the
correct concreting facilities should assist in attaining close contact with the
roof.
10.6.4 Leakage through the rock surrounding the plug

Leakage through the strata can occur as a result of (1) geological fissures or
other discontinuities in the rock and (2) cracks in the rock formed by
ground stress or by strain from mining operations.
If possible, plugs should be sited away from faults in the rock. However, if
they are unavoidable, the grouting process will help to seal and stabilize
conditions.
The problem of rock failure under high stress is experienced when other
mining excavations encroach too closely or at great depth where overbur-
den pressures become excessive. Every effort should be made to avoid over-
stressed areas.
10.6.5 Determination of plug length required for sealing

The problem of leakage associated with underground plugs has been dis-
cussed previously on the basis of where it occurs and how it can be mini-
mized or stopped by grouting. It is relatively straightforward to determine
the plug length which conforms to the permissible punching shear and
bearing stress values at the concrete-rock interface. Quantifying exactly the
length which is required for a leakage-free plug is much more difficult.
Published data concerning the subject are scarce and what information is
available is related to specific ground conditions which cannot be applied
on a general basis.
Garrett and Campbell Pitt (1958, 1961) published the results of tests on
an experimental plug, 1.220 m (4 ft) square by 3.350 m (7 ft 8.5 in) long and
situated in sound quartzite, at West Driefontein. The static water pressure
was approximately 20.7N mm - 2 (3000 lb in - 2). An extensive system of tap-
ping points and holes in the rock were incorporated for studying leakage at
the steel load transfer cylinder interface with the concrete, at the con-
crete-rock interface and through the strata. Leakage quantities were ob-
served before grouting and after various stages of pressure grouting were
completed. From the test results, Garrett and Campbell Pitt (1958) pro-
posed certain concepts which are given below.
The resistance of a plug to the passage of water either along its contact
with rock or through the adjacent fractured rock depends on two factors:
the length of the plug and the resistance of the rock to the passage of
water.
The latter, being a condition of the rock which varies greatly with differ-
ent types and mining conditons, can be regarded as the practical con-
sideration for determining plug length.
The two factors can be interrelated using the pressure gradient through
the rock as the linking medium.
Results from the West Driefontein test plug form the basis for the graphs
contained in Figure 10.4 which are reproduced from Garrett and Campbell
Pitt (1958). These results refer only to the rock and pressure conditions
described. The graphs are (A) the minimum length of plug that would be
required if the contact between plug and rock was ungrouted
[P/l=0.23Nmm- 2 m- 1 (20.8Ibin- 2 ft- 1 )]; (B) the minimum length when
the contact is grouted but before the rock is grouted [P/l = 3.64
Nmm- 2 m- 1 (161lbin- 2 ft- 1 )]; (C) the minimum length when normal
grouting of the rock was 41.4Nmm- 2 (6000Ibin- 2 ) [p/l=9.14
N mm - 2 m - 1 (4041b in - 2 ft -1)]. This is normal to South African practice,
being twice the hydrostatic pressure, but is not normal to the UK. The
graph (D) is similar to C but with the addition of chemicals to seal rock
fissures. Graph C is applicable in South Africa to a normally grouted plug
but has no safety margin.
OOr-------~-------r--_r------,_----c__,
70
I
I-
C)
Z 30
~
\\ ....... .
4 \\'f.~ ..... ;,;.
...........
~
~--a c
- ~
~~~=::::J.D
2 3 4 6 6 10 12 14 16
HEAD(XlO' It) (WATER TABLE TO PLUG)
Fig. 10.4 Required plug lengths to resist hydrostatic pressure based on leakage
resistance and bearing (Garrett and Campbell Pitt. 1958). Results in imperial units
are used as presented by these authors. Dotted lines are based on bearing (Equation
(1O.8d)), all other lines relate to leakage resistance of the test plug.
Garrett and Campbell Pitt (1958) suggested from this that plug length
should be such that a leakage factor of safety should not be less than 4 and
may be as much as 10. The choice depends on an assessment of many
factors which include fracture of rock during excavation and subsequent
destressing, porosity of the rock and its acceptance of grout. In Figure lOA
the graphs show plug lengths when factors of safety of 4, 6, 8 and 10 are
applied. These depend on the plug-rock contact and the rock being
grouted to at least the same pressure as that which the plug is designed to
resist. Plug lengths for various square section sizes based on a Phe value of
4.14Nm- 2 (600Ibin- 2 ) are also included on the basis of Equation (1O.8d)
(shown by dotted lines). The value of 600 lb in - 2 was used by Garrett and
Campbell Pitt (1958).
As far as the author is aware, this is the only published information which
attempts to quantify directly leakage resistance in relation to plug length,
apart from records of past plugs which have been successful. It has been
emphasized throughout this section that the data put forward relates only
to the particular test conditions. This leaves the plug designer very much to
his own initiative and experience in determining the plug length which will
provide adequate sealing. Further research is therefore necessary into this
area. However, if a plug is constructed with sufficient length to resist move-

ment but cannot prevent leakage through the surrounding rock, the length
can always be increased. In an emergency this may be important, because
the plug could be constructed to prevent flooding and subsequently
lengthened to reduce leakage.
10.7 CASE STUDIES
10.7.1 British Gypsum Ltd, Sherburn Mine, England, 1980

(Emergency plug)
In 1980 a pressure pad was constructed by British Gypsum Ltd,
(section A-A of Figure 10.6) in an attempt to seal off water inflow into the
area of the pump sump (1 East 5 South in Figure 10.5). The general dip of
the strata was from left to right in Figure 10.5 (west to east) and it was
considered that water was following the interface between the gypsum, in
which the roads were driven, and the marl bed below and making its way
down the strike. Excavation of the gypsum appeared to have encroached on
the marl bed and allowed water to come up through the floor. The main
access to the mine was via the 1 in 4 adit which was close to the inflow
position. Upon failure of the pressure pad during grouting operations,
Cementation Mining Ltd were asked on 21 October 1980 to design a new
scheme for sealing off the water. The water inflow at that time was es-
timated to be 182 L s -1 (2400 gal min -1) (Figure 10.7). Various structural
schemes for pressure pads and combinations of pads and plugs were con-
sidered and discarded in favor of the complete plug solution shown in
Figure 10.6 for simplicity, speed of construction and permanency.
Urgency was the main criteria as, within 6 days of Cementation Mining
Ltd being called in (27 October), the water inflow had risen to 379 L S-l
(5000galmin-1) and it was rapidly becoming obvious that there was a
danger of losing the mine.
The plug scheme adopted is detailed in Figure 10.6. It was deemed pru-
dent not to disturb the remaining sections of the original pressure pad.
Another gravel bed was laid over the top in which six more water control
pipes were placed in addition to the two pipes (one 200mm diameter and
one 100 mm diameter) previously installed below the original pressure pad.
The additional pipes were four of 200 mm diameter and two of 300 mm
diameter and carried the water to a new sump position adjacent to the
proposed plug site. Additional rising mains were installed in the shaft to
cope with the increasing inflow.
The first concrete was poured on 28 October and Figure 10.6 shows the
concreting stages. Because of the large mass of concrete involved, construc-
tion joints were necessary and a low heat of hydration mix, incorporating a
cement replacement material, was used (Table 10.3). Concrete was pumped
Case studies 255
.. '
. . . Rising main
6>~~'
q~ borehole
.o~~r to surface
--------.~~~~----~~~~
Emergency plug _JI;~L==IT~~
D[
~
~
'il-
0I 5I 10 15 20 25
,---,
I I I I
meters
\i I
Fig.tO.S British Gypsum Ltd, Sherburn Mine, England. Underground layout
showing position of emergency plug.
from the surface down the 1 in 4 adit, through a 100 mm pipe, directly into
position in the plug. The 4 week time period for placing the concrete
resulted from various equipment, labour and general construction problems
but once concreting had commenced the water inflow was controlled at a
peak level of 606 L s - 1 (8000 gal min - 1) (Figure 10.7).
Minimal true design was required for the Sherburn Mine plug. The depth
below ground level was 48 m, which resulted in a hydrostatic pressure of
0.47Nmm- 2 (68lbin- 2 ). This is not excessive and the length of the plug
was extremely long. However, length in this case was governed by practical
considerations to suit the particular situation. The pressure gradient from
>-------------------8- 35.300 ,~,-,~:LI\\j"
~ Illy )\'1111 . _
\\\\, __ .. 1)\)\\"
\\'11 --\\',\,'))')\ )"1. ..
'"1111\' '''1111 111)\' B~ ''0'111\)1\11 if""IIII" , \"'!II) \\\\'" '>\'-)1\\\\\\,,\--"'"
HI
r
A " A
l 1\ c======= ==---------------- __ Pipes to be led out j
~ ~======= ~ ------------------ ..
through opening
~'/;,-~/~-::,:':::./i;.::.-~/~" '/i"'::::7/":",:,,~//~?/-_,\\,,,'.::. I--//~::::.:: ,11\\. //:.;..-, , \ / / Y....
before concreting sump
<~.~
j"[, -!: ;: :;: : ~; ~ i.,Jtl~:! ! !;-!i! !1!1I!~! ! ! ! l l i! !~ !I;
I. 5600 .1
(a) Plan
'1l11fjjjf B./ Mass concrete plug inflow

; water control pipes
,\1 C
o, 2, !3 ,4 5,
'jt~i.iil~$&7' J meters
(c) Section B-B (d) Section c-c
Concrete infill behind block
walls, dowelled into rock to
hold down Original pressure
pad edges
10
Failed zone of original

pressure pad
(b) Section A-A ... _'H/'"
Fig 10.6 Emergency plug at British Gypsum Ltd, Sherburn Mine, England, 1980. (a) General arrangement plan and sections (b)
A-A, (c) B-B and (d) C-C showing proposed concreting sequence.
Case studies 257
~ l8
8.CD -~i
_.gE
-I! eCDE
-S8
g 8oe>
-8
.!rl Various
~",s
I~:;!
e.
;: ill
concreting stages >
"I
LL
~~I~~~~~ E __
peakValu8
5000
4000
3000
2000
1000
300
200
100L-~~~~~J-J-~~~-L~~~-7~~~
22 27 2 7 12 17 22 27 1 6 11 16 21 26 1 6 11 16
SEPT OCT NOV DEC 1980
Fig 10.7 British Gypsum Ltd, Sherburn Mine, England. Water inflow quantities
related to concrete poured in emergency plug, showing effectiveness of water stop-
ping (water inflow quantities in imperial units as recorded).
one end to the other was only 0.47/35.3 = 0.013 N mm - 2 m- 1

(0.59Ibin- 1 ft- 1)and the 35.3m (116 ft) length (Figure 10.6a) was eventually
extended out to the adjacent access roadway.
Figure 10.7 indicates how effective the plug was in stopping water. On
completion of the various concrete stages, the control pipe valves were
closed and the inflow almost completely stopped. Final sealing by grouting
commenced after the valves were shut off and involved a combination of
grout pipe positions. Some were previously cast into the plug to reach
places which would have been inaccessible by drilling from the two plug
faces. These, in addition to injection at the inflow point through the water
control pipes and to the contact zones through other holes drilled from
both faces, enabled the water to be sealed off completely on a permanent
basis. Only cement grout injection was necessary.
10.7.2 Proposed precautionary plug, 1981, for overseas contract

by Cementation Mining Ltd
Figure 10.8 contains a proposal for three precautionary plugs to be installed
near the bottom of a shaft for access protection in the case of an inrush.
Construction of the plugs was to be in a thin limestone bed, 15 m thick,
Table 10.3 Cement replacement mixes previously used by Cementation Mining Ltd
for underground plugs
Emergency plug in roadway: Temporary consolidation plug
Grade 30 (0 PC replacement in shaft: Grade 55 (OPC
with PFA); 30Nmm- 2 replacement with Cemsave
ground granulated blast
furnace slag); 55Nmm- 2
Site British Gypsum Ltd, Sherburn National Coal Board, North

Mine Selby No.1 shaft
Supplier Topmix Ltd Topmix Ltd
Total cementitious 400 kg m -3(250 kg m -3 OPC, 500 kg m - 3 (30% OPC, 70%
content 150kgm- 3PFA) Cemsave)
Sand 770 kg m - 3(Elvaston Zone 2) 595 kg m - 3 (Blaxton Zone 3)
Sand % of total 42 34
aggregate
Coarse aggregate 1050 kg m - 3 (Elvaston Gravel) 1150 kg m - 3 (Blaxton Gravel)
Water 180Lm- 3 180Lm- 3
Water: cement ratio 0.45 0.36
Slump without 50mm 60mm
plasticizer
Plasticizer Flocrete N (Cementation Flocrete N (Cementation
Chemicals Ltd) Chemicals Ltd)
Slump with 160mm 160mm
plasticizer
situated above and below weak, water-bearing strata zones at a depth of

542.5 m (1780ft) [5.43Nmm- 2 (787Ibin- 2 ) hydrostatic pressure].
Design of the plug, load transfer cylinder and bulkhead door was carried
out in accordance with the design calculation section, Grade 35 concrete
being specified. The concrete-rock interface calculated punching shear
stress was 0.63Nmm- 2 (91lbin- 2 ) and the pressure gradient 5.43/12=
0.45Nmm- 2 m- i (19.9 Ibin- 2 ft-i).
The proposed grouting scheme depended on the actual ground condi-
tions at the level of the plug when the shaft was eventually sunk. However,
care needed to be taken above and below the plug in order not to encroach
too close to the water-bearing zones.
10.7.3 National Coal Board North Selby Mine, England, 1982

(temporary consolidation plug)
During December 1982 Cementation Mining Ltd were sinking two shafts at
North Selby for the National Coal Board Selby project. Both shafts had
reached the stage of sinking through the Ackworth Rock, which is a Coal
Measures sandstone and an aquifer, with No.1 shaft sump (Figure 10.9)
standing at - 540.2 m (1772 ft) below surface level [hydrostatic pressure
2500 4000
1- -I -,
I 12000 I I
E
4.500 i I /4.500 ~
~ E
I I I
I I I
I r7 ~~ I I I~
I I I
I
~ ..
-~.v..,-= . _~#_,J,v. -~A':. ~A-&, ~V' -_~A"-"~"
".'-~": ~~~'(/#:~ ;~~-==
01 Load transfer cylinder. 6 No. rings

Ie of 7 No. segments each fabricated
..; _ fro~ 45 p.!!!..l!'ick,_Gra~ ~ ste'!!~--II .1-._._ Water pressure-
1 No. grout port per segment. Rings
fitted with rubber recesSed seals. Access ladder
Removable decking
~
Grout seal 500 mm wide by 100 mm deep-+--~

annulus with 8 No. 50 mm grout tubes
through steel sheet bolted to strata ~ ~
o 234 5
1 1 1 I
(a) meters
Fig 10.8 Proposed precautionary plug for overseas contract by Cementati9n Mining Ltd, 1981. (a) General arrangement elevation,
(b) section A-A, (c) view on B-B and (d) plan at inbye end.
Top of limestone
oI 2
I
3
I
4
I
5
I
meters
I
I
I
I /
/
Service pipes to be
fitted with glands at
each end
~
- -
Dewatering pipe. Inbye ~~~~~~~~~

end to upstream sump /1
-- --
with debris ~rill. Outbye
end fitted With valve in
valve pit. / Development roadway
/ I width 5000 \
/ I \
(b) I \
I \
I \
Bottom of limes:=to::..:n.=.eL-_
Bulkhead door with human access tube.
Fabricated in 4 No. segments from
45 mm thick, Grade 43 steel. Segments
to be bolted together underground and
all joints fully profile welded.
, ~
(e) (d)
Fig. 10.8 (b), (c) and (d).

Case studies 261
<a>
\---+-...-Strata grouting holes

and standpipes
o, 234
""
5
meters
Fig 10.9 National Coal Board, North Selby Mine, England. Section through shaft,
showing (a) temporary consolidation plug and (b) plan at pump lodge level.
5.4 N mm - 2 (7831b in - 2)]. The previous sump level in No.1 shaft stood
13.8 m (45 ft) above the new sump level and strata cover grouting was
carried out from the previous level.
During the period of strata cover grouting, problems of grout standpipe

installation were experienced due to the poor rock conditions and deterio-
ration and heave of the sump took place. The length of cover grouting was
also long (over 40m) whereas the preferred maximum length was approxi-
mately 30m. To enable the wall of the cover grouting cone to be less prone
to leakage at the lower level of treatment and to guarantee satisfactory
grout standpipe installation, it was decided to sink to the - 540.2 level and
install a concrete plug. This would be closer to the zone of strata requiring
the major grout treatment and, by casting the grout standpipes into the
plug, a pressure pad for the next cover of strata grouting could be provided.
Due to the potential water inflows for sinking below the plug, it was
necessary to install a pump lodge (Figure 10.9) for stage pumping to surface.
No choice of position was available for the pump lodge other than immedi-
ately below the last cast section of the shaft wall.
At the time of placing the plug, shaft water inflow to the sump was
approximately 11 L s -1 (150 gal min -1). Figure 10.10 shows the framework
for supporting the grout pipes and the water control rising mains during
casting of the plug. The concrete mix design for the plug is given in
Table 10.3. Minimal heat of hydration existed in the concrete mass, due to
use of the cement replacement material (Cemsave) and additional heat re-
moval occurred through the rising mains and grout pipes. The Grade 55
concrete was the same as the shaft lining concrete. However, designing the
plug on the basis of 7 day cube test results, (2/3) x 55 = 36.7 N mm - Z
(5317lbin- Z), allowed pressurizing of the plug for water stopping at the
earliest opportunity. The benefit of the 28 day strength was taken for the
wall-bearing resistance. The recommendations given in the design calcula-
tion section for cylindrical plugs were followed for the plug design.
Neglecting the bearing resistance of the tapered plug, the calculated punc-
hing shear stress for the concrete-rock interface was 0.89 N mm - Z
(129 lbin -Z) and the pressure gradient was 5.4/17.3 = 0.31 N mm -z m- 1
(13.7Ibin- 2 ft-1).
Grouting up of the plug started from the bottom through 50 mm grout
pipes installed in the rising mains. These pipes were grouted in, leaving the
bottom free for injection into the gravel bed, and also secured by high-
pressure flanges bolted together at the top of the rising mains. The bottom
injection was phased to follow backwall injection of the shaft wall above the
plug, and controlled using the standpipes as 'tell-tales' before closing off for
final pressurizing.
The shaft water make was reduced to approximately 0.45 L S-l
(6 gal min -1) before final tightening up, this amount being predominantly
from behind the shaft lining above the pump lodge. The pump lodge was
restricted to a position close to the plug.
To enable the plug to be subsequently broken out without damaging the
shaft wall, the bottom surface of the wall was painted with a bond-breaking
agent, Setcrete 11 (Don Construction Chemicals Ltd); the hanging rod ends
Contact zone grout
pipes (8 No.)
Etll~
_ ~~ 1{
L ~J~nIMm;og
:< ' grout pipes
fA ~~.. . . 4. \ ' .. ~ Al
./ '\ \,~ Lightweight
Test hole chain ties
standpipes (8 No.)
fa Bl
Visqueen Strata grouting Steel angle
membrane standpipes (48 No.) support frames
rc Cl
Gravel be<
(a)
(c)
Steel channel frame

for pipe support ? ~ ~ t r
meters "'" ~. Steel channel
frame pipe
supports
Rising mains set in

perforated oil drums
(b) (d)
Fig 10.10 National Coal Board, North Selby Mine, England. (a) General arrangement elevation of temporary consolidation plug,
and sections (b) A-A, (c) B-B and (d) C-C showing supporting framework for cast in grout stand pipes and water control rising mains.
were sleeved and two water bars were incorporated, the inner one protected
and the outer one sacrificial for plug sealing.
10.8 CONCLUSIONS AND RECOMMENDATIONS
The first objective of this chapter was to review underground plug design
for the purpose of bringing the subject to prominence and more up-to-date.
As an additional objective, design rationalization was attempted on the
basis of current practice.
The author considers the first objective to have been achieved. However,
much more work needs to be carried out to quantify, in greater detail, strata
leakage resistance in relation to plug length before the design procedure can
be regarded as being completely rationalized.
The philosophies of design included here are based predominantly on the
excellent work of Garrett and Campbell Pitt which was reported in 1958
and 1961. In addition to the normally accepted punching shear stress con-
cept of design for plug interaction with the surrounding rock, they proposed
a bearing stress concept which was related to the surface roughness and
also carried out tests on both experimental and service plugs to quantify
plug length in relation to leakage resistance. This is the only published work
known to the author which relates to the latter factor. However, the work
carried out by Garrett and Campbell Pitt is specifically applicable to the
gold mines of South Africa, where hard rocks of the quartzite type are
encountered at deep levels and high water pressures are experienced
(Figure 10.4). The quoted data are not directly applicable to any other rock
conditions, particularly those of the softer sandstone, limestone, marls and
coal measures experienced in the UK (Figure 10.4), where the aquifers are
closer to the surface and the hydrostatic pressures are much less. A study of
the Garrett and Campbell Pitt work was essential here to form the basis for
applying their principles to other rock conditions, in line with modern
construction codes of practice, as it appears that very little forward progress
has been made in the subject of plug design during the last 3 decades.
Considering the two parallel plug length design theories, one based on
punching shear stress and the other on bearing stress, which have been
proposed for resistance to horizontal thrust at the concrete-rock interface,
it would appear that they are incompatible. Comparing Equations (1O.1c)
and (lO.8d), giving 1 = pb/4p pe and 1= pb/2Pbe respectively, using the value
for Pbe = 3.75p pe from Table 10.1 indicates that the length based on permis-
sible punching shear stress, as given by Equation (1O.1c), will always be the
longer by a factor of 1.875. Based on the concept of length being a priority
for resistance to leakage, the bearing stress concept can be neglected in the
design of parallel plugs. It should, however, be pointed out that although
the permissible shear stress concept is recommended for determining length,
Conclusions and recommendations 265
in order to assist sealing by increasing the leakage resistance, the actual

strength will be greater because of the bearing action.
As already mentioned in section lOA, the two tapered plug design the-
ories based on different bearing stress concepts are also not compatible. In
this case the longer length is given by the Garrett and Campbell Pitt rough
surface-bearing resistance philosophy, as opposed to the smooth-faced
wedge principle, and the former is therefore the recommended approach,
based on the longer length required for leakage resistance.
With regard to the permissible stresses quoted in Table 10.1, the values of
Pc,Pb and Pp are realistic for the current types of concrete now being used
underground. The factor of safety equal to 4 used in connection with the Pbe
and Ppe values at the concrete-rock interface is also probably realistic.
However, care should always be taken to study rock strength and condition
to confirm the values. It is interesting to note that the permissible shear
stress values for the interface, which are quoted by South African practice,
are less than the values recommended in Table 10.1. Although the South
African values are not related directly to concrete or rock strengths, nor to
the rock condition, they result in longer plug lengths which err on the safe
side for leakage resistance. On this basis it can be seen that the stronger,
and better quality, concretes now being employed in underground construc-
tion will give shorter plug design lengths for strength but could have in-
herent leakage problems if sufficient length is not provided.
At the present time it is not possible to define the exact length which is
needed for sealing in relation to any particular ground conditions. The
pressure gradient concept of Garrett and Campbell Pitt (1958, 1961) would
appear to be a practical means of quantifying the resistance of rock to the
passage of water through specific lengths but insufficient data is available as
yet for general application of the principle. The allowable pressure gradient
of 9.14 N mm- 2 m- 1 (4041b in- 2 ft- 1), which the South Africans would
accept for normally grouted rock, should not be adopted in the UK as it is
based on plug-rock interface grouting pressures of 2-2.5 times hydrostatic
pressure. Such high pressures would not be adopted in the UK, values of
1.25-1.5 being more representative.
Comparing the pressure gradients from the case studies with the Garrett
and Campbell Pitt data in Figure lOA, the Sherburn Mine emergency plug
value of 0.013 N mm - 2 m -1 (0.59 lb in - 2 ft -1) is much less than that given
by graph A [0.23 N mm - 2 m -1 (20.8 lb in - 2 ft -1)] showing that it pos-
sessed a satisfactory leakage resistance without grouting.
For the proposed precautionary plug the pressure gradient of
0045 N mm- 2 m- 1 (19.9Ibin- 2 ft-1) was much less than that given by
graph B [3.64 Nmm- 2 m- 1 (161Ibin- 2 ft-1)]. This indicates that although
leakage would occur before grouting of the contact zone it would not leak
after grouting the interface. The North Selby temporary consolidation plug
was also in this category, possessing a pressure gradient greatly reduced
[0.31 N mm- 2 m- 1 (13.7Ib in- 2 ft- 1)] from that given by graph B.
It would appear that the Garrett and Campbell Pitt pressure gradient of
3.64 N mm- 2 m- 1 (1611b in- 2 ft-1) could be used as an upper limit in the
UK for plugs with the contact zones and strata grouted. However, much
lower pressure gradients will result in the ability to seal off leakage more
easily.
Each plug scheme will generally be an individual design tailored to the
particular situation. The above recommendations for pressure gradients
should be used with caution and the rock leakage resistance in situ, which is
associated with each design, must be investigated as thoroughly as possible
prior to preparing any scheme. Successful plug design therefore will rely
heavily on the mining contractor's experience and knowledge.
Current concrete mix designs, using plasticizers for high workability, are
much more easy to place and provide much tighter contact with the sur-
rounding rock. Improved sealing will be achieved and leakage resistance
will be much greater. Increased pressure gradients should be capable of
being withstood by shorter lengths of plug and therefore in future the
quantifying of such data by experiment and in situ monitoring is essential to
progress and improve underground plug design.
Understanding of plug mechanisms of resistance to horizontal thrust,
when confined by the surrounding rock, can be enhanced by further studies
into rock stresses resulting from excavations. Modification of these stresses
during interface pressure grouting and the accompanying plug stressing
needs to be investigated. Finally, study of the effects of end pressures on
such a combined stress system would lead to knowledge of how stresses are
dissipated throughout the whole and possibly a clearer picture of the ulti-
mate behavior of the interface under load would emerge. Future research
and experiment are therefore imperative to advance the state of the art of
plug design.
CHAPTER ELEVEN
Design of Borehole Seals:

Process, Criteria and
Considerations
J. J. K. Daemen and K. Fuenkajorn
11.1 A INTRODUCTION
This chapter presents design process, criteria and considerations for sealing
boreholes in rock, with a main emphasis on the hydraulic and mechanical
performance of the emplaced seals. The scope of the present information is
intended to cover various types of borehole and sealing materials designed
for different objectives and requirements. While the main emphasis of this
chapter is on borehole sealing, some recommendations are applicable to
larger underground excavations (e.g. shafts and tunnels). Discussions are
focused on the hydraulic bond and mechanical bond of the seals under a
variety of site characteristics and environments. Some design criteria recom-
mended here are based on the results of laboratory and in situ experiments
conducted specifically to determine the hydraulic and mechanical perform-
ance of neat cement, commercial grade bentonite and mixtures of bentonite
and crushed rock. Other relevant industry guidelines will be referred to,
where applicable.
The proposed designs are not intended to replace the industry guidelines
and regulations enforced by the local states and government agencies.
Rather, they should be incorporated into the existing guidelines for specific
applications, such as those given by Gray and Gray (1992) and US Bureau
of Mines (1994) for the sealing of abandoned boreholes in the mining indus-
try; by Smith (1986, 1990) for sealing oil and gas wells; and by ASTM
D5299 and USEPA (1975) for the sealing of groundwater and monitoring
wells. Most regulations on borehole sealing are generic, broad and rarely

Edited by K. Fuenkajorn and 1. 1. K. Daemen.
268 Design of borehole seals
allow prOVlSlon for taking into account site-specific conditions. They

usually do not provide sufficient information about the expected perform-
ance of the seals after installation, and hence do not allow for adequate
assessment or prediction of their sealing effectiveness. A proper integration
of the present recommendations into the relevant industry guidelines will
improve the understanding of the hydraulic and mechanical performance of
borehole seals under specific site characteristics and environments, and will
ensure that the designed seals perform their intended functions and fulfil the
sealing requirements.
The recommendations are limited to the technical and engineering as-
pects of the design process. Applicable codes and legal and regulatory pro-
cedures should be followed even though they are not addressed here. The
main aspects of the design process for borehole sealing discussed here in-
clude (1) applications, objectives and requirements, (2) considerations of site
characteristics, (3) general design criteria and (4) material selection and
placement methods.
Three main types of sealing materials discussed here are cementitious
materials bentonite and mixtures of bentonite and sand or crushed rock.
The term 'seal system' represents all types of seals emplaced in a borehole
and the interface between the seals and rock surrounding the borehole.
11.2 APPLICATIONS, OBJECTIVES AND REQUIREMENTS
Boreholes are drilled for many purposes. They can range from shallow,
small-diameter exploratory holes to investigate a light foundation to the
very deep, large-diameter holes for the production of oil and gas. The
best-known applications of borehole sealing probably are in the areas of oil
and gas exploration and production, and groundwater monitoring and pro-
duction. Other applications of borehole sealing include boreholes used in
underground mining, underground storage caverns in rock salt and solution
mining; deep holes used for monitoring tectonic stresses for earthquake
prediction; wells used for geothermal recovery and steam production; holes
used for coal gasification; and exploratory holes for construction of dams,
bridges, road tunnels, buildings, subways, etc. One somewhat unusual but
potentially important application is the sealing of boreholes (and other
underground excavations) that penetrate the geological barrier around a
high-level nuclear waste repository. The most common reason for sealing
for these applications is to prevent groundwater contamination. Other rea-
sons for sealing are to support the casing, protect casing from corrosion,
seal off zones of lost circulation and to prevent blowouts (in oil and gas
wells). Of growing importance, especially for water wells and for various
exploratory holes in areas where groundwater is present, is the recognition
that leaving holes open or inadequately sealed may reduce the artesian
pressure and may result in a wasteful outflow of groundwater.
Applications, objectives and requirements 269
Regardless of the sealing objectives and applications, study of the per-
formance of emplaced seals has been focused on two aspects: hydraulic
bond and mechanical bond. From a point of view of potential leakage, the
hydraulic bond is important in order to assure that flow through the seal
and along the host rock-seal interface remains within an acceptable magni-
tude, depending upon types of application. Providing adequate mechanical
bonding (shearing resistance at the interface) assures that a seal will not be
dislodged, e.g. under the action of gas or liquid pressure.
Types of borehole sealing have been classified by various industries,
based on their sealing objectives. For example, Gray and Gray (1992) clas-
sify the sealing of mine openings, including boreholes, into three categories:
permanent, temporary and semi-permanent. Smith (1994) classifies sealing
of groundwater wells into three categories: temporary sealing, sealing ac-
tively used boreholes and sealing for permanent decommissioning.
To cover various types of boreholes, the sealing is classified here into two
main categories: (1) sealing actively used boreholes and (2) sealing unused
boreholes. The first category involves sealing of the annular zones between
casing (or pipe) and the surrounding rock, and sealing of open boreholes
that will be used in the future. The seals for this category are designed not
only to isolate the zones of gas or liquid, but also to support the operation
of the boreholes (e.g. to stabilize the casing, borehole instruments and sur-
rounding rock, and to prevent corrosion). The second category represents
permanent sealing, which mainly involves sealing of any to-be-abandoned
boreholes or wells. The primary function of seals for this category is to
isolate zones of gas or liquid, with the main emphasis on environmental
protection. The design of these seals is aimed at long-term performance
because they will eventually become a permanent part of the rock forma-
tions.
Regulatory sealing requirements for borehole abandonment frequently
are specified in terms of the types of material to be used, emplacement
procedures to be used and the length of seal to be emplaced. Seal designs
are rarely based on explicit performance requirements. It is usually implied
that seal performance will be adequate if certain material selection criteria
and emplacement procedures are followed. Within the context of present
detailed prescriptive regulatory requirements, design concerns are largely
pre-empted.
In several common fields of practice, in particular petroleum engineering,
but also, although probably to a lesser extent, in water well abandonment,
sealing and geotechnical or exploration drilling practices are fairly well
established and regulatory specifications are explicit and detailed, so that
explicit sealing design requirements frequently may not be necessary.
A rational approach to seal design requires the identification of necessary
seal performance criteria, and the selection of materials and emplacement
procedures to ensure that the performance requirements will be satisfied.
For seals intended to prevent groundwater pollution, a basis for required
performance can be established by determining allowable levels of con-

taminant flow through the seal. Such an analysis will result in an acceptable
flow rate, which in turn can be converted into a combination of seal proper-
ties, in particular hydraulic conductivity, and seal geometry, i.e. necessary
seal length. When determining the allowable flow rates it is important to
take into consideration the entire seal zone system. i.e. including interface
flow and flow through the host rock.
11.3 CONSIDERATIONS OF SITE CHARACTERISTICS
The site characteristics should be considered in the seal design because they
normally have direct impacts on the seal performance. It is important that
borehole seals are designed specifically for a particular site. If possible,
technical specifications, sealing materials and placement methods should be
derived and selected for a specific 'seal location', i.e. considering the depth of
the seals, characteristics of the surrounding rock mass and in situ fluid
pressures and flows, rock stresses and temperatures. The site characteristics
which are tentatively important to the success of borehole sealing are dis-
cussed below. The significance of these individual factors depends on type,
location and layout (depth, orientation and diameter) of the boreholes, and
on the performance requirements of the seals.
Consideration of the effect of each individual factor sometimes may not
be sufficient because the potential combined effect of applicable site charac-
teristics may be the controlling factor in determining whether or not a seal
system will continue to perform its intended function. Consequently, the
combined effects should be considered in the design as well so that the
design will be conservative.
11.3.1 Borehole condition

One of the most important factors to be considered when planning the
sealing of boreholes is the condition of the holes. At the very least, it is
necessary to know, to the greatest extent practicable, the geometry of the
hole, i.e. its true diameter, and of its casing, its condition and location and
any installations in the hole (e.g. pumps, pipes and cables). For long-aban-
doned holes it must be considered likely that some debris may have been
dumped in the hole. Depending on the rock types penetrated, the probabil-
ity of hole failures, collapses or sloughing must be evaluated.
In preparation of borehole sealing it is desirable to caliper log the hole.
Determining the actual size of the hole, as a function of depth, will allow a
correct determination of the amount of sealing materials to be emplaced.
Caliper logging may also indicate where there are stability problems sec-
tions, which are probably not preferred locations for seal emplacement.
Video logging is highly desirable, as it will allow a visual determination of
Considerations of site characteristics 271
downhole conditions. Geophysical logging can provide specific properties

about the downhole environment, particularly about lithology, casing con-
dition and cementing condition.
Any damage or weakness zone around the opening, which has been
induced by drilling, weathering, deterioration, high stress concentration or
by casing removal, could soften the surrounding rock and hence reduces the
hydraulic and mechanical bonds at the seal- rock interface. Fractures
around a seal location could become a leakage path for gas or water to
bypass the seal.
11.3.2 Geological considerations

Available logging data and geological cross-sections should be used to
determine the stratigraphy of the rock unit(s) covering the entire length of
the borehole. The mechanical, hydrological and chemical properties and
fracture characteristics of the surrounding rock units, particularly those
around the seal location, should be investigated and considered in the seal
design and material selection. Strength and stiffness of the surrounding rock
mass usually affect the mechanical interaction at the seal- rock interface
which governs the axial strength of the emplaced seal. Permeability of the
surrounding rocks should be used to determine the desirable range of per-
meabilities of the designed seals.
11.3.3 Hydrological considerations

The depth of the groundwater table is one of the main factors governing the
location and length of seals. Infiltration, surface flooding and potential for
fluctuation of the groundwater table should be taken into consideration.
These hydrological conditions control the magnitude of the hydraulic press-
ures (gradients) imposed on the seals. The moisture content of the surround-
ing rock may affect the long-term properties of the seals. Of particular
importance is determining whether or not liquid (e.g. groundwater) flow is
taking place at the seal locations. Significant flow rates, especially at high
pressure, are certain to have a detrimental effect on seals unless they are
placed using special procedures. The flow of liquids or gases into cementi-
tious grouts will dilute the seal-materials, and will increase the permeability
and strength of the seal. Where possible, every effort should be made to
reduce the flow rate as much as possible immediately before emplacing a
seal. The preferable option is to direct flow from the hole to be sealed by
pumping nearby boreholes.
Emplacing cementitious seals in standing water generally will reduce the
quality of the seal. It is therefore desirable to reduce the water levels in holes
to be sealed to the greatest extent possible.
If a seal is installed in the unsaturated zone above the groundwater table,
any perched water zones in the unsaturated rock units should be considered.
Potential flow of the perched water into the sealed borehole may alter the
hydrological conditions of the seals, and may increase the amount of dis-
charge on the seals.
11.3.4 Gas and liquid pressures

To prevent the dislodging of cementitious seals or piping and erosion of
bentonitic seals, the maximum pressure gradients induced by water or gas
at the seal ends should be determined. For sealing of groundwater wells, the
pressure gradient is related to the depth of seal and the hydraulic head of
groundwater. The pressure gradient is an important design requirement for
sealing a borehole connected to liquified gas and compressed air storage
cavern, and for sealing oil- and gas-producing wells. The maximum pressure
gradients usually determine the minimum length of borehole seal.
11.3.5 Geochemistry
Chemical compatibility between the seals, surrounding rock and ground-
water may be insignificant for temporary or short-term sealing. For long-
term or permanent sealing the chemical compatibility is a significant factor
and should be considered in the material selection process. One of the
reasons for mixing bentonite with crushed rock (obtained from the seal
location) to form a seal is to minimize the potential for chemical incompati-
bility between the seal and the existing environment. The type of cement
selected for the preparation of cementitious seals should be compatible with
groundwater chemistry.
11.3.6 Impact from nearby activities

This factor involves geomechanical and hydrological interferences from
nearby excavation, underground operation, mining-induced displacements
or subsidence induced by excavation or by groundwater withdrawal. This
situation includes boreholes in underground mines (e.g. exploratory
boreholes in the pillar, roof or floor of an active mine opening) and
boreholes connected to storage caverns with fluctuation of storage pressures
due to pumping and retrieval process.
Potential changes of confining pressure, hydrological conditions and
characteristics of the surrounding rock should be considered in the seal
design. For example, if large displacement is anticipated it might be prefer-
able to install a flexible seal (e.g. bentonite) rather than a stiff seal (e.g.
cement or concrete).
A major influence on seal performance from nearby activities is the devel-
opment of excessive deformations. Large strains may be imposed as a result
of ground subsidence, whether induced by oil and gas production, ground-
water withdrawal, or mining. Boreholes in or near subsidence areas may be
Considerations of site characteristics 273
subjected to tension, compression, bending and torsion. Moreover, as sub-

sidence develops and progresses, different sections of boreholes may be
subjected to different deformation mechanisms, while each particular sec-
tion of a borehole will be subjected to different differential strain histories.
Rigid borehole seals with strong mechanical bonding, i.e. cement grout
and concrete seals, will try to follow the ground deformations. Deformable
seals, especially seals with favorable healing characteristics, e.g. highly plas-
tic bentonitic seals, can much more readily accommodate relatively large
straining without significant detrimental effects on sealing performance.
In areas where relatively large ground movements can be expected it is
recommended that at least an estimate be made of the strains to ensure that
a few localized transverse fractures in otherwise sound long borehole seals
will not necessarily have a significant detrimental impact on overall
borehole seal performance. Conversely, pervasive longitudinal fractures or
extended separation along the seal- borehole wall contact may create high-
ly undesirable preferential flow paths.
11.3.7 In situ stresses

From the view point of the stability of the borehole to be sealed, the effect
of in situ stresses is considered to be insignificant for shallow boreholes. For
sealing of deep boreholes, the magnitude and ratio of horizontal principal
stresses are important, particularly where the surrounding rock has relative-
ly low strength. A high concentration of tangential compressive stresses
could induce failure (i.e. borehole breakout) at the borehole boundary and
hence degrade the seal performance. The in situ stresses relate directly to the
confining pressure around the seal, which govern the mechanical interaction
at the seal- rock interface. Confining pressure at the seal location is an
important consideration when seals with high expansive or swelling stresses
are used (e.g. cement grout with an expansive agent and highly compacted
bentonite).
11.3.8 Thermalloads
The effect of elevated temperatures is important for deep oil and gas wells,
geothermal wells, some ore exploratory wells and in deep holes or in areas
with high geothermal gradients. The temperature at the seal location will
affect the curing period and hydration of cementitious seals. Elevated tem-
peratures may have detrimental effects on the long-term sealing perform-
ance of bentonitic seals.
11.3.9 Seismic activities

The characteristics and frequency of occurrence of seismic activities caused
by earthquakes or blasting should be considered. A major consideration for
designing seals that support backfill is whether the backfill might be subjec-
ted to liquefaction (National Coal Board, 1982).
Tectonic movement along planes intersecting sealed boreholes is likely to
shear off plugs. If the movement is sudden, a stiff brittle seal (e.g. cured
cement grout or concrete) is likely to be sheared off completely. A relatively
soft and plastic clay seal may withstand considerable deformation without
major detrimental impact on the longitudinal hydraulic conductivity of the
seal system.
11.4 GENERAL DESIGN CRITERIA
The design requirements for seals should be translated from the design
objectives while considering the site characteristics. Detailed technical spec-
ifications for seals should be developed to satisfy the requirements. General
design criteria for any applications of borehole sealing are discussed as
follows .
The borehole seal and each component of the seal system are designed for
specific seal locations. At each seal location physical, petrological, mech-
anical, chemical and hydrological properties of the rock mass and the
in situ stress state are taken into consideration. For example, Fuenkajorn
and Daemen (1987) and Akgun and Daemen (1991c) experimentally dem-
onstrated that the swelling stress of expansive cement increases with a
decrease of boundary deformation, thereby increasing the hydraulic and
mechanical bonds between the seal and the surrounding rock. However,
the swelling pressure should not be too high and cause tensile failure of
the surrounding rock or opening up of unfavorably oriented fractures or
joints in the host rock. The swelling pressure limitation can be readily
determined for intact rock, when the long-term tensile strength and the
in situ stress field are known .
The effective permeability of the seal system (seals, seal- rock interface)
should be reasonably close to that of the surrounding rock mass. For
unusually short seals, i.e. shorter than a few borehole diameters, no sig-
nificant sealing performance is gained, i.e. no significant reduction in flow
rate is obtained, by reducing the seal permeability to less than 10 times
that of the host rock (e.g. Figures 2.4 and 2.5). For seals whose length is
much greater than their diameter the ratio of 10: 1 for seal permeability
to host rock permeability probably can be increased substantially with-
out the borehole becoming a highly preferential flow path. Where the
10: 1 ratio can be maintained, without excessive cost, it provides an
acceptable sealing goal. A formal site-specific hydrological analysis can
provide more detailed insight as to whether a further relaxation of the
seal permeability is acceptable.
General design criteria 275
The strength and stiffness of the seal system are adequate to provide
mechanical support to the surrounding rock mass. This is to minimize the
inward movement of the surrounding rock, which could create a preferen-
tial flow path for water or gas to bypass the seals.
If cementitious seals are used they should be installed in intact portions
of the boreholes, to the extent that this is practicable. Experimental re-
sults obtained by South and Daemen (1986), Greer and Daemen (1991),
Crouthamel (1991), Akgun and Daemen (1991c) and Fuenkajorn and
Daemen (1991 b) have indicated that the mechanical and hydraulic per-
formance of cement borehole seals is enhanced when they are installed in
the intact portion of the boreholes. It is therefore highly desirable that the
rock around cementitious borehole seals is free of fractures.
Cementitious seals in boreholes are designed such that the shearing resis-
tance between the cement seal and surrounding rock mass is adequate to
sustain the axial load induced by the cement seal itself and by any overly-
ing backfill. For seals that will or may be subjected to significant axial
loads, a bond strength design may be required. Most seals to be emplaced
in boreholes are intended for permanent hole abandonment, suggesting
that a prudent conservative design approach would be most appropriate.
For the unusual situation where relatively short seals are emplaced, i.e.
seals with a length of less than five diameters, we recommend that an
elastic stress analysis be performed and that all bond stresses as well as
stresses in seal and host be kept well below the elastic limit. For the more
common case where relatively long borehole seals are emplaced the issue
probably is moot, because the length of the bond area can more than
adequately compensate for some local overstressing.
The length of cementitious seals should be sufficient such that they can
maintain their mechanical stability. Based on the results from laboratory
testing and numerical analyses on the shearing strength of borehole seal
in rock, Akgun and Daemen (1991c) conclude that permanent abandon-
ment borehole seals should be designed with a length-to-diameter ratio of
four or greater. This length criterion will prevent development of excess-
ively detrimental tensile stresses within and near an axially loaded seal.
The similar geometrical design is also recommended by Greer and
Daemen (1991) who assess experimentally the in situ hydraulic perform-
ance of borehole seals in rock.
Mechanical interactions between the cementitious seals and the host rock
are taken into consideration in the design. Different seal components or
types, however, may exhibit different mechanical impacts on the adjacent
and surrounding media. Such impacts should therefore be considered in
the design of the seal system.
The impact of nearby or intersecting excavations (open or sealed) on the
seal system behavior is taken into consideration, where applicable. Such
impacts include, for example, stress interference, alteration of rock mass
characteristics and change of hydraulic gradients.
The seal system is designed considering the existing conventional practi-

ces and state-of-the-art technologies, if and when applicable, such as
those given by the American Petroleum Institute (1986) and Smith (1986,
1990) for cement; Crouthamel (1991), Greer and Daemen (1991), Garrett
and Campbell Pitt (1958, 1961), Auld (1983), Chekan (1985), and Gray
and Gray (1992) for concrete; Dixon and Gray (1985), and Pusch et al.
(1987a,b) Pusch (1987, 1988, 1990), and Ouyang and Daemen (1992) for
bentonite; US Bureau of Mines (1994) for mine backfill; Lutenegger and
DeGroot (1993) and Smith (1994) for sealing of groundwater wells.
11.5 MATERIAL SELECTION AND PLACEMENT METHODS
11.5.1 Selection of seal materials

Discussed here are considerations and general criteria relevant to the selec-
tion of seal materials.
Seal materials should be selected for specific seal locations considering
the local hydrological characteristics, borehole conditions and in situ stres-
ses and their subsequent changes, as well as the geomechanical and
chemical properties of the surrounding rock mass. For permanent sealing,
the physical and chemical compatibility among seals, backfill and surround-
ing rock should be taken into account so that the originally intended seal
functions are maintained.
Granular materials (sand, gravel or crushed rock), if used for backfill,
should be sufficiently wellgraded. A coefficient of uniformity equal to or
greater than 16 is recommended by Ouyang and Daemen (1992). The maxi-
mum particle size of backfill should be sufficiently small, less than one-tenth
of the diameter of the borehole to be sealed.
Cementitious and bentonitic materials are the most widely used for
borehole sealing. Cementitious grout will provide relatively stiff and strong
seals. To the extent that contamination and emplacement desegregation are
avoided, cementitious seals can provide exceedingly low permeabilities. For
sealing applications where low permeability is necessary, strong preference
should be given to shrinkage-compensated or slightly expansive cement
grouts. For large diameter applications, sanded cement grouts, with addi-
tion of aggregate, will further reduce the risk of shrinkage.
If a mixture of bentonite and crushed rock (or sand and gravel) is used to
form a seal, the amount of bentonite should be sufficient to allow compac-
tion to be effective. Ouyang and Daemen (1992) suggest that the amount of
the bentonite in the mixture should be greater than 35 wt %. They indicate
that for a bentonite~crushed rock mixture, compaction and the amount of
bentonite are decisive factors in producing good mixture seals. The effec-
tiveness of compaction in reducing porosity is hindered by the soft benton-
ite buffer. For loosely or ineffectively compacted mixtures containing 25 %
Material selection and placement Methods 277
bentonite or less, the sealing performance can be degraded by dynamic

disturbance during compaction. The influence of such disturbances is great-
ly reduced when more bentonite is added. Seal mixtures with a bentonite
content of 35% exhibit a more homogeneous texture, more isotropic behav-
ior and lower permeability than those with lower bentonite contents.
11.5.2 Placement methods

Placement methods depend strongly on the industrial context within which
the sealing is to be performed, as well as the required level of sealing
performance. Sophisticated emplacement technology is particularly well de-
veloped in the petroleum industry, and allows the placement of high-quality
seals at great depth and under difficult conditions (e.g. high temperatures,
high gas or liquid pressures and/or flows).
Relatively simple emplacement procedures may be fully adequate for the
sealing of shallow holes and can provide sealing performance appropriate
for local conditions.
Where the primary purpose of backfill is to stabilize holes, i.e. to prevent
hole sloughing and collapse, a sufficiently strong granular backfill can be
emplaced by surface dumping, pouring or shoveling. In order to preclude
bridging, it is essential that average and maximum particle sizes are not too
large. A reasonable guideline is that the maximum particle size should not
exceed one-tenth of the minimum open borehole diameter and that the
average particle size should not exceed about one-twentieth of the borehole
diameter. Angular crushed rock will provide more strength than rounded
smooth gravel. Especially for relatively deep holes, for crooked holes and
for holes with weak (e.g highly corroded) casing or weak walls, it is highly
desirable to check frequently the depth to which backfill has risen, in order
to ensure that bridging has not occurred.
It is clearly undesirable to dump rock in holes with walls so weak as to be
highly susceptible to erosion and sloughing.
Whether or not the use of drill cutting for backfilling is appropriate
greatly depends upon site-specific conditions. Backfilling with drill cuttings
greatly reduces the volume of cuttings that need to be disposed of otherwise.
In holes with minor or no lithological changes along the hole length a
potentially significant benefit from the use of drill cuttings is the likelihood
of geochemical stability, i.e. reduced chances of chemical interactions be-
tween the host rock backfill and groundwater. For holes that penetrate
multiple different lithologies, it may be less obvious whether or not there is
compatibility between the host rock and cuttings. Similarly, before deciding
on the use of drill cuttings it is necessary to consider whether contamination
may have occurred during drilling, e.g. from drilling mud and/or additives.
It should be recognized that the strength of using backfill drill cuttings
usually will be significantly less than that of crushed rock. The strength (and
seal performance) can be improved by controlled compacted emplacement.
Cementitious seals can be emplaced by a number of different methods.

The once common practice of dumping the mix, or even unmixed materials,
down the hole must be considered unacceptable except for the shallowest of
holes, because segregation is virtually unavoidable with such an emplace-
ment procedure. Moreover, there is always a real possibility of bridging
leaving voids.
Bailer emplacement is possible and certainly is acceptable in dry holes. In
standing water or in holes contaminated with mud, some seal contamina-
tion may be expected.
The pumping of cement grouts or slurries is usually the preferred option.
In order to minimize permeability and segregation, and to maximize density
and strength, it is strongly recommended that injection takes place from the
bottom up, while gradually withdrawing the injection string.
Emplacement under pressure is essential when grout needs to be injec-
ted into the annulus behind perforated casing. Pressurized injection of
relatively fluid (low viscosity) grouts may assist in sealing fractures in the
host rock, and may reverse some of the flow path enlargement and per-
meability enhancement that may have resulted from borehole wall relax-
ation, especially in boreholes that have been open for a long time.
Conversely, the emplacement pressure should not be so large as to open
up new flow paths along the seal, e.g. as a result of inducing longitudinal
radial fractures.
Compaction is desirable for the installation of backfill or any seals for-
med by bentonite, clay, sand, gravel or crushed rock, or by combinations
of these materials. The sealing effectiveness of these particulate media can
be enhanced by compaction. Compaction reduces the permeability and
potential settlement, and increases the bulk density and strength of the
backfill. In situ compaction of backfill is usually recommended but is not
required for conventional mine closure practices.
Boreholes should be prepared prior to sealing. They should be cleaned; as
a minimum all loose debris should be removed. If possible, the casing
should be pulled, but if this is not possible, it should be perforated and/or
split in order to allow grout to flow into the annulus behind casing.
Particularly for holes that have been abandoned for a long time, that
have been in operation for a long time and that penetrate rock formations
susceptible to caving, consideration should be given to overdrilling and/or
reaming the hole (e.g. Lamb and Kenney, 1989).
Methods for removing casing should not induce damage to the borehole
wall. The borehole casing and the process of removing it may affect seal
performance. If casing is left in place, corrosion of the casing may eventually
give rise to preferential flow paths. The extent of the disturbed zone on the
host rock should be assessed and taken into account during seal design and
material selection.
11.6 SUMMARY
This chapter discusses the process, criteria, and considerations for design of
borehole seals in rock, with a main emphasis on the hydraulic and mechan-
ical performance of the emplaced seals. Discussions are concentrated on
cementitious and bentonitic seals. It is strongly recommended that design
and material selection for borehole seals are site-specific, i.e. taking into
consideration the hole conditions, hydrological and mechanical characteris-
tics of the surrounding rock, depth, and in situ stress at the seal locations.
Systematic planning and design for borehole sealing should include a deter-
mination of the borehole conditions. The objectives to be accomplished
need to be identified, as well as all regulatory requirements that may affect
sealing. Sealing materials should be selected to provide assurance that the
sealing performance objectives can be satisfied. The emplacement methods
should be selected so as to ensure that the as-emplaced seal will meet the
performance objectives. The general design considerations recommended
here can be integrated into applicable industry guidelines to enhance the
sealing effectiveness, particularly for long-term seal performance.
ACKNOWLEDGEMENTS
The experience of the authors was developed from conducting research

projects sponsored by the US Nuclear Regulatory Commission, under con-
tracts NRC-04-78-271 and NRC-04-86-113. Support and permission to
publish this chapter are gratefully acknowledged.
CHAPTER TWELVE
Sodium Bentonite as
a Borehole Sealant
1. E. Papp
12.1 INTRODUCTION
Improperly sealed boreholes pose a threat to the integrity of the natural

groundwater system and therefore require a hydraulic barrier. All effective
borehole sealants will minimize vertical migration of water and/or con-
taminants by hydraulically sealing off various strata to prevent infiltration
of surface water (Nielsen, 1993). To accomplish these objectives, borehole
sealants must have lower hydraulic conductivity than the surrounding for-
mations, compatibility with native soils to ensure a satisfactory seal at the
soil-seal interface, and longevity. Bentonite, a naturally occurring clay min-
eral, has become increasingly popular as a borehole sealant because of its
ability to perform these functions. The major constituent of this ore is
sodium montmorillonite. This chapter discusses the origin and properties of
bentonite and its uses and applications in today's industry. In addition to
this, a case history featuring bentonite for sealing deep monitoring wells will
be discussed.
Montmorillonite, the major constituent of bentonite, is a member of the
smectite clay mineral group. The term smectite denotes a group of clay
minerals, the most common of which are montmorillonite, beidellite, non-
tronite, saponite and hectorite. Smectite clays belong to the mica family of
phyllosilicates, but differ from true micas because of their low layer charge
and exchangeable interlayer cations. Another differentiating feature is that
their interlamellar surfaces and cations can be readily hydrated and dehy-
drated. Because of these characteristics, smectite clays are commonly called
'swelling clays'.

Edited by K. Fuenkajorn and 1. 1. K. Daemen.
Introduction 281
Montmorillonites are generally classified as sodium or calcium types,

depending upon the dominant exchangeable ion. When sodium predomi-
nates, a large amount of water can be absorbed in the interlayer, resulting in
significant swelling. As will be discussed below, the presence of sodium
facilitates the absorption of many layers of water molecules, a phenomenon
which does not occur when the predominant exchangeable ions are calcium
or magnesium.
Bentonite has a soft, plastic, slippery consistency. Its color ranges from
white to light green and light blue when removed from the ground. Upon
atmospheric exposure it becomes light cream in color, then gradually chan-
ges to yellow, red or brown.
Long called the 'clay of 1000 uses', sodium bentonite presently has more
individual uses than any other mineral known. Popular uses include civil
engineering, well drilling and foundry and pelletization applications.
Civil engineers have utilized sodium bentonite in numerous water and
fluid retention applications. Dry bentonite is mixed into a porous soil.
When hydrated, it swells and forms a low permeability barrier for ponds,
reservoirs, sanitary and industrial landfills, dams, lagoons and slurry
trench cut-off walls.
The well-drilling industry (for oil, water and groundwater monitoring)
has utilized bentonite as a drilling fluid additive since the 1920s. Benton-
ite provides lubricity, viscosity, density, gel strength and filtrate control.
Since the early 1980s drilling-related uses of sodium bentonite has in-
cluded the sealing of shallow exploration boreholes to prevent contami-
nation of groundwater aquifers by surface or interformational leakage.
This chapter will discuss specific types of seals formed with sodium ben-
tonite.
The foundry industry utilizes bentonite as bonding agent for sands used
in producing foundry cores and molds. The bentonite bonds with the
sand and water to form a plastic, cohesive material that can be com-
pressed around a pattern to prepare a mold. The bentonite ideally bonds
with the sand with sufficient strength to maintain the cavity before, dur-
ing and after the pouring of molten metal. It forms and maintains an
efficient high-strength bond at temperatures of as much as 3,000F
(1700C).
Sodium and/or calcium bentonite is used extensively as a binding agent
in the pelletization of animal feeds, iron ore and other fine-grained solids.
The use of sodium bentonite to pelletize iron ore derived from taconite
was introduced in the United States in the mid-1950s. Within 20 years it
became the largest single use for sodium bentonite.
Additional commercial applications for bentonite include absorbants for
pet litter and filtering agents for depolarizing vegetable and animals oils and
fats. Purified, water-washed forms of bentonite are used in water-based
paints, cosmetics, medicines and pharmaceuticals (Odom, 1986).
282 Sodium bentonite as a borehole sealant
12.2 GEOGRAPHICAL ORIGIN
Montmorillonites are by far the most abundant of the smectite clay min-
erals (Odom, 1986). They are the predominant mineral composing bentonite
and fuller's earth throughout the world. Fuller's earth is any natural ma-
terial which is generally composed of attapulgite or a smectite clay, with no
implication of origin or mineralogy. Both minerals are major components
of soils in arid and semi-arid regions. Bentonite deposits range in age from
Cretaceous to Recent and are commonly associated with beds of marine
origin (O'Driscoll, 1988). Associated beds may also be non-marine, such as
freshwater limestones, carbonaceous shale or beds of coal. Montmorillon-
ites in the form of bentonites or fuller's earth are found on every continent
except Antarctica. Large volumes of sodium bentonites occur in the western
United States in parts of South Dakota, Wyoming and Montana. Sizable
calcium bentonite deposits are found in Alabama, Arizona, Mississippi and
Texas. The vast majority of montmorillonite deposits outside the United
States are of the calcium type. Sodium bentonites, however, have been
found in Mexico, Argentina, South Africa, Turkey, Japan and Australia
(Odom, 1986). Individual beds of bentonite may be laminated or massive
and may be found at or below ground surface. Because of its ability to
expand when wet and contract while drying, bentonite outcrops frequently
have a 'popcorn' appearance (Colangelo and Upadhyay, 1990). Weathering
tends to increase an outcrop's colloidal properties. Bentonite layers vary in
thickness from less than 1 in (2.5 cm) to more than 20 ft (6 m). Thicker beds
typically consist of several different layers.
12.3 GEOLOGICAL ORIGIN
Most bentonite deposits in the United States were apparently formed by the
in situ process, during the Cretaceous Period. This commonly accepted
theory states that the bentonite was formed by volcanic ash. This ash was
generated from the massive volcanic plateau in the western United States.
The ash was spewed into the air and transported to shallow inland seas
where it underwent a devitrification process, which is believed to have taken
place after deposition and burial of the ash. Although not commercially
viable, a few bentonite deposits are known to have been formed by the
hydrothermal activity or groundwater alteration of glassy acidic extrusive
igneous rocks.
12.4 STRUCTURE OF SODIUM BENTONITE
The mineral montmorillonite is the constituent of bentonite, with minor

amounts of feldspar, gypsum, quartz, mica and gypsum present.
Structure of sodium bentonite 283
Montmorillonite consists of two basic types of sheet units that form a three-
component molecular lattice. The layers consist of two tetrahedral sheets
sandwiched around a central octahedral sheet (Figure 12.1); thus mon-
tmorillonite is sometimes called a 2: 1 mineral (Grim, 1968). The tetrahedral
sheets define the upper and lower units of the structure and contain tet-
ravalent (Si) and usually some trivalent (AP +) cations. The apexes of the
tetrahedra point toward one another and the oxygen anions at the apexes
form part of an octahedral sheet that contains primarily trivalent aluminum
(Figure 12.2). For montmorillonite the total negative charge contributed by
the sum of all the oxygen and hydroxyl anions exceeds the total positive
charge contributed by the sum of all the structural cations (Si4 +, AI3+,
Fe 2+, Fe 3 +, Mg2 +). This imparts a slight overall negative charge on the 2: 1
structure in the tetrahedral and octahedral sheets. This excess negative
charge on the sheets is counterbalanced by exchangeable cations which
exist between them. The thickness of each 2: 1 layer is typically 1 nm; the
diameter is typically 100 to 300 times the thickness.
I
I
I
I
I
I
I
I
I
I
~
~OI~mO )~ C
IAI I O.96nm
SI
Fig. 12.1 Schematic diagram of montmorillonite structures.

o Oxygens
Hydroxyls
Aluminum
Silicon
Fig. 12.2 Crystalline structure of sodium montmorillonite showing interlamellar

water layers.
12.5 PROPERTIES OF SODIUM BENTONITE
Sodium bentonite's unique properties include a large surface area, ion ex-
change capabilities and high water absorption and expansion. Bentonite
platelets are typically positioned in equidimensional and thin flake-shaped
units. The flat shape of the flakes results in an extremely large surface area,
often ranging from 700 to 800m 2 /g. With lateral dimensions 100 to 300
times greater than platelet thickness, only 12 g of sodium bentonite theoreti-
cally yields sufficient surface area to cover a football field.
Perhaps sodium bentonite's most widely known characteristics are its
high water absorption and swelling properties. Sodium bentonite is charac-
terized as capable of absorbing at least five times its weight in water and
expands when fully saturated with water to a volume 12-15 times its orig-
inal dry size. This tremendous water-absorbing property is the most obvi-
ous physical feature distinguishing it from other clays used for industrial
applications (Figure 12.3 illustrates the interlayer water system binding to
clay surfaces). Once sodium bentonite is added to water, the sodium ions
hydrate and partially dissociate from the bentonite platelet. High sodium
and high montmorillonite contents are indicative of high swelling or water-
holding properties.
Bentonite as a borehole sealant 285
Tetrahedralsheeto!
bentontte platelet
~i' : ~"
. ' \ 1 . . : . I'\! 0
..WJ-\Q\LoO}
. .' . 0
Oriented water
<>:::9<5:~o-Q.~""'''o-Q.~''

molecules
.. ~~::~~.. ~~....... ~
!.J~l:!J~l-~o
Tetrah~ral sheelo!
bentontte platelet
0 ~
V V 0 00
Fig. 12.3 Interlayer water system showing hydrogen binding to the adjacent clay
surfaces. (From Grim, 1968.)
12.6 BENTONITE AS A BOREHOLE SEALANT
Exploratory boreholes, if improperly sealed, provide an open conduit for

downward moment of surface water and/or contaminants, mixing of
aquifers, modification of groundwater flow conditions and may even cause
physical hazards.
One of the most significant features of bentonite as a borehole sealant is
its low permeability (hydraulic conductivity). The material sealing the
borehole should have a permeability equal to or lower than the least per-
meable formation layer penetrated during the drilling process. The defini-
tion of hydraulic conductivity is the rate of flow of water in gallons per day
through a cross-section of one square foot under a unit hydraulic gradient,
at the prevailing temperature. Hydraulic conductivity is usually expressed in
values of cm/s. Because sodium bentonite generally has a hydraulic conduc-
tivity value of 1 x 10 - 7 to 1 x 10 - 9 cm/s, it forms an excellent borehole
seal when hydrated (the lower the hydraulic conductivity, the less leakage
through a given material, and therefore the better seal). If the borehole is
backfilled with other material (i.e. recompacted drill cuttings, sand or bor-
row material), a low-hydraulic conductivity seal cannot be ensured, and the
borehole may then act as a conduit for vertical migration of potentially
contaminated groundwater (Nielsen, 1991). If drill cuttings are used to
provide the seal it is possible that they may be contaminated (especially
when drilling in contaminated zones) and thus adversely affect the ground-
water quality.
12.7 BENTONITE AS AN ANNULAR SEALANT
Because of its unique swelling characteristics and large surface area, sodium
bentonite provides a highly effective annular seal in environmental monitor-
ing wells. The annular seal is defined as the sealing material placed above
the filter pack in the annulus between the borehole and the well casing
(Nielsen, 1991). Its function for environmental monitoring wells is to provide
a low-permeability seal that will not adversely affect ambient groundwater
quality, will seal discrete sampling zones, and will prevent the mixing of
aquifers and the vertical migration of surface water.
If not properly sealed, the annular space may act as a conduit for water
and contaminants to enter into the aquifer through various pathways:
through a seal that has cracked, shrunk or deteriorated over time,
through the area created as a result of the sealing material not properly
bonding to the casing, or
through a void in the annular space created as a result of bridging of the
sealing material.
The design of the annular seal varies from well to well; however, the
American Society for Testing and Materials (ASTM) has established
standards on groundwater and vadose zone investigations (ASTM,
1991). ASTM D5092 (Standard Practice for Design and Installation of
Ground Water Monitoring Wells in Aquifers) promotes (1) durable and
reliable construction, (2) extraction of representative groundwater sam-
ples and (3) efficient and site hydrogeological characterizations (ASTM,
1992). Given these objectives, the annular seal should be installed to best
suit the monitoring well and limit any adverse affects from improper
sealing.
In a typical single-cased monitoring well, a 3~5ft (1~1.5m) bentonite
layer is placed above the filter pack. This layer (typically bentonite chips or
tablets) forms a low-permeability seal to prevent the infiltration of grout
into the filter pack. The infiltration of grouting materials may adversely
affect the quality of the groundwater samples collected from the screened
area. Following installation of the bentonite seal, a high-solids bentonite
grout is commonly installed. In certain instances bentonite chips and tablets
may not be able to be placed effectively above the filter pack (i.e. the
material may bridge, or is a deep well installation, etc.). In this case a
secondary filter pack of a finer-grained material is installed which allows for
a high-solids bentonite grout to be placed directly on top of the filter pack.
Bentonite platelets will form a filter cake at the interface with the finer-
grained, lower-porosity, secondary filter pack. This mechanism prevents the
migration of grouting fluids into the sampling zone. Cement based grouts
do not have the ability to form a filter cake and subsequently should not be
installed directly on top of the filter pack. To ensure that there will be no
migration of fluids into the screened area, the entire annular space should
Types of bentonite sealants 287
be sealed. Figure 12.4 illustrates the components of a typical groundwater
monitoring well.
12.8 TYPES OF BENTONITE SEALANTS
Since the early 1980s bentonites has become widely accepted for sealing
boreholes and the annular space in monitoring wells. This is due to its
ability to provide a low-permeability seal, its ease of use and the fact that it
will not affect the quality of existing groundwater. Bentonite processors
have made the material available in many forms and sizes to accomplish
this task successfully. Products available to the drilling industry include
high-solids bentonite grouts, bentonite chips and tables, bentonite drilling
fluids and granular bentonite.
12.8.1 High-solids bentonite grout

High solids bentonite grouts are products specifically engineered for use as
borehole sealants. The grouts consist of a blend of powdered bentonite and
Concrete
Borehole ------
Sand or
gravel pack
I=~r"'-'----Silotted well
screen
Fig. 12.4 A typical groundwater monitoring well.

fresh water mixed to a pumpable slurry containing a minimum of 20%

solids (ASTM, 1992). For example, grout with a minimum of 20% solids is
formed by mixing 24 gal (6.35 L) of water and 50lb (22.6 kg) of bentonite.
The higher the solids content, the lower the porosity and therefore the lower
the permeability. The solids content should only reflect the weight of be-
ntonite used in the slurry, not inactive filler materials (i.e. sand, drill cut-
tings, etc.) To determine the percentage of solids of a grout mixture, the
weight of the material is divided by the weight of material and water [(i.e.
(weight of dry material/weight of material + weight of water) x 100)].
Generally, a direct correlation exists between the percent solids and the
hydraulic conductivity of a high-solids grout. Figure 12.5 suggests a consist-
ent trend of decreasing hydraulic conductivity with increasing solids con-
ductivity (Edil and Muhanna, 1993). This graph also illustrates the increase
in mud weight (density) due to the increase in solids.
High-solids bentonite grouts should be evaluated for the following char-
acteristics.
Permeability. Grout material should have a permeability equal to or
lower than the least permeable formation layer penetrated.
Reactivity. Grout should not react with formation material or water nor
be capable of contaminating the aquifer.
Placeability. Grout must be in a form which can be positively and accu-
rately placed to fill all voids.
1.0xl0-7 1.3
- MudWeight
-e- Hydraulic Conductivity
1.25~
~If)
E 1.0x10-8 1.2
-
:9
.c.
OJ
~
.::.:: ~"0
::J
~
1.15
1.0xl0-9 1 - - - - - - - - r - - - - - - r - - - - - - - r - - - - - - - - 4 1.1
20 25 30 35 40
Solids Content(Ofo)
Fig. 12.5 Hydraulic conductivity and mud weight versus percent solids. (Edil and
Muhanna, 1993.)
Flowability. Grout should be self-leveling in the annulus, should not
bridge and should be uniform in consistency.
Settlement. Grout should have minimal penetration into a permeable
formation zone.
Stability. Grout should provide an element of structural stability when
set.
Bonding/shrinkage. Grout should be capable of bonding to a well casing
and borehole wall to provide a watertight seal; and
Workability. Grout should be easily handled and mixed, and should not
prematurely gel.
Alternatively, portland cements and cement/bentonite mixtures are also
often used to seal boreholes. These materials have both advantages and
disadvantages (Table 12.1). Shrinkage generally occurs with the cement,
causing it to pull away from the borehole wall and/or casing. Because of its
chemical nature, cement is a highly alkaline substance (pH from 10 to 12),
and thus introduces the potential for altering the pH of water with which it
comes in contact (Nielsen, 1991). Cement also produces heat when curing.
This heat, coupled with the weight of the material (typically 14-15Ib/gal
(1.68-1.80 kg/L)), can deform or collapse PVC casing. Bentonite grouts,
however, are less dense and do not generate heat while hydrating (Healy,
1991).
12.8.2 Bentonite chips and tablets

Bentonite chips and tablets are often used to seal shallow boreholes (less
than 50ft (15m)). Chips are irregularly shaped with sizes ranging from 0.25
to 0.75 in (0.63 to 1.91 em) diameter and moisture contents between 12%
and 16%. Because of their high moisture content and slow swelling tenden-
cies, chips can be dropped through a water column more readily than a
material with a low moisture content. Consequently, chips have been used
to seal wells/boreholes whose depths exceed 1000 ft (305 m).
Bentonite tablets consist of bentonite that is mined, field-dried, dried in a
rotary kiln, processed to a granular form and then compressed into a tablet.
Tablets are uniform in shape with typical sizes ranging between 0.25 and
0.5 in (0.63 to 1.26 cm) diameter and a moisture content of less than 10%.
Because of their low moisture content, tablets begin to hydrate almost
immediately upon contact with water. During placement through water, the
tablets may become sticky and clump together. Depending on the borehole
diameter and placement technique, bridging may occur during installation.
Generally, both tablets and chips should be placed in the saturated zone
to ensure complete hydration. However, if it is necessary to place them in
the vadose zone, they must be saturated with water applied manually
down the borehole. When completely hydrated, both bentonite chips and
Table 12.1 Advantages and disadvantages of various grouts
Grout material Description Advantages Disadvantages
High-solids Powdered bentonite Does not shrink during Has limited structural strength;
bentonite mixed with water to form curing; no heat of may desiccate in
a slurry with a minimum hydration; forms a low extremely dry fine grained
of 20% solids and a permeable flexible seal; soils; limited resistance to
density of 9.4 Ib/gal (1.13kg/L) will rehydrate if dried sulfates
Portland cement type I Most common type of Readily available; can Develops high heat of
cement with no special be delivered on site in hydration; shrinks and can
properties ready mix trucks; crack when curing; low
develops good structural resistance to sulfates; may
strength elevate pH levels
Portland cement type V High resistance to sulfates High resistance to sulfates Not readily available;
and brines; moderate heat lower set strength; shrinks
of hydration and can crack when curing;
may elevate pH levels
Cement and 3-5%added to most types Can reduce shrinkage of Decreases set strength;
bentonite mixtures of cement cement; decreases may lead to elevated pH
density making the levels; can shrink and
grout more pumpable; crack when curing
decreases heat of hydration
tablets have hydraulic conductivity values ranging from 1 x 10 - 7 to

1 X 10- 9 cm/s.
12.8.3 Bentonite drilling Ouids

Bentonite drilling fluids (made from powdered bentonite) are engineered for
use in mud-rotary drilling operations. The primary functions of bentonite
drilling fluids are (1) to stabilize the borehole, (2) to cool and lubricate the
drill bit and drill string, (3) to support the walls of the borehole, and (4) to
transport cuttings to the surface. Because of their ability to gain solids (i.e.
sand, drill cuttings, etc.) throughout the drilling process, bentonite drilling
fluids have been used as a slurry to seal boreholes. This mixture is usually
viscous and barely pumpable during placement in the borehole. A bentonite
slurry is typically used because it is convenient, easy to use, economical and
it appears to have a high solids content. A bentonite drilling fluid typically
has a solids content of 2% to 4%. If it is higher than this the fluid is so
viscous that the material is not pumpable. Bentonite drilling fluids are
designed to hydrate quickly and give high yield and viscosity at low solids.
This facilitates drill cutting removal and fast drilling production. It does
not, however, make an adequate borehole seal.
Because of its low solids content, a bentonite drilling fluid will separ-
ate, therefore dropping the bentonite to the bottom of the borehole and
thus compromising its ability to form an effective seal. A typical drilling
fluid slurry at 6.25% solids has a hydraulic conductivity of 8 x 10- 5 cm/s
(Bertane, 1985), making it unacceptable as a borehole seal. As a result of
bentonite drilling fluids containing organic polymers to extend the yield,
they are often prohibited as a seal in the construction of monitoring
wells.
12.8.4 Granular bentonite

Granular bentonites consist of dry bentonite clay in various mesh sizes to
seal shallow boreholes. Granular bentonites can be used dry or mixed with
water to form a borehole sealant. Because of its small particle size and low
moisture content (less than 10%), granular bentonite is often difficult to
place through standing water and often results in bridging. For this reason,
granular bentonite should only be placed in the vadose zone.
Granular bentonites are commonly mixed with water by gravity as a dry
product to provide a pumpable grout mixture. To attain a high-solids
grouting material, however, the granular bentonite should be mixed with an
organic polymer wetting agent. This organic polymer delays the bentonite's
hydration, enabling the user to increase the solids content of the grout.
However, the use of organic polymers may adversely affect the ambient
groundwater quality and is therefore not recommended.
12.9 PLACEMENT OF BENTONITE
12.9.1 Bentonite chips and tablets

Bentonite chips and tablets are commonly placed in shallow boreholes in
the dry form. The material is poured directly into the borehole and falls via
gravity into the area to be sealed. This method is therefore called the gravity
method. When the gravity method is used, special precautions should be
taken to ensure that the bentonite has reached its intended location. To
verify that the material is not bridging, a tamping rod should be used to
check the level of installed bentonite. When placing chips or tablets through
a lengthy water column, the gravity method leaves chips and tablets very
susceptible to bridging. However, as stated earlier, bentonite chips tend to
fall more readily through a water column due to their higher moisture
context compared to than tablets.
12.9.2 High-solids bentonite grouts

High-solids bentonite grouts should be pumped under positive pressure into
the borehole or annular area of the well. To accomplish this task, the high-
solids bentonite grout (powder) should be mixed with an appropriate mixer
to shear the individual platelets of the bentonite. The shearing process
allows bentonite to hydrate more fully by breaking apart the aggregate clay
particles and exposing the individual platelets. Mixers commonly used are
paddle mixers and recirculation systems.
A paddle mixer, which is the most commonly used machine, is a circular
tub with a vertical shaft and rotating horizontal blades which provide
moderate shear action. These mixers work well with both bentonite and
cement, and can handle cement additives such as sand. Paddle mixers
(Figure 12.6) are generally self-powered with their own pumps for grout
delivery.
Recirculation systems mix grout by pumping water from a mixing tank
through a constricted hose, which causes increased pressure and velocity. As
dry material is added to this stream of water, the turbulence and shearing
action of the pump adequately mixes the grout. Rotar-stator, piston and
centrifugal pumps are considered to be the most effective for recirculation
mixing (Healy, 1991).
Once the grout has been properly mixed, it should be pumped under
pressure through a side-discharge tremie pipe (Figure 12.7). A side-discharge
tremie pipe is used to prevent any erosion and infiltration of the grout into
the filter pack. The grout should be pumped into the borehole from the
bottom of the drilled hole to the top, displacing any water. Once the ben-
tonite mixture exits the borehole at the surface it should be weighed with a
mud balance to ensure that a correct mud weight has been achieved
throughout the entire length of the borehole.
Placement of bentonite 293
Fig. 12.6 Mixing of bentonite grout with a standard paddle mixer.
12.9.3 Granular bentonite

Granular bentonite, in its dry state, is commonly placed in the vadose zone
of boreholes to provide a low-permeability seal. This is typically performed
by pouring the granular material directly down the casing or annulus.
Granular bentonite is also often mixed with water to form a pump able
grout mixture that can be placed through water via a tremie pipe to its
desired depth. However, as a result of the bentonite's high swelling capabili-
ties, granular bentonite when mixed directly with water achieves only a
10 - 15% solids content. To achieve a higher solids content an organic
polymer is typically mixed with the water and bentonite to attain a higher
solids content. The polymer delays the hydration of the bentonite, enabling
a greater amount of bentonite to be added to the mixture, thus increasing
the solids content. However, as stated earlier, the use of polymers in grout-
ing materials may adversely affect the quality of the groundwater being
sampled.
In the late 1980s the Geological Survey of Canada patented a
sand/bentonite injector (SBI). The purpose of the SBI was to allow sand
and granular bentonite to be installed at predetermined levels in monitoring
wells and geotechnical boreholes (Figure 12.8). The injector uses air to in-
stall dry bentonite and sand, through a small-diameter tube, to the desired
Bentonite Slurry Int8te: from Grout Mixer
Annular Space -++-I~I
I
III
I 1II-l-l-==::'-BentonlU Annular Seal
III, '
=
;
1 ~: ::::iOf:rt+-t=~ Well Screen
m~:
=1 ~; _I;oj:++-~FilterPack
m:~:::=:~ III
III III
Fig. 12.7 Emplacement of grout with a tremie pipe.
level below water. The SBI comprises three pressure-rated cylinders which
require an air compressor for operation. Once pressurized, valves on each
tank are adjusted to allow the air to carry the appropriate material through
the tube to the bottom of the well or borehole. This allows immediate direct
placement and packing. As a result, the bentonite cannot swell prematurely
and the full swelling capabilities of the bentonite are seen at the site of the
emplacement. The applications where the SBI is most often used are
piezometer installations, sealing of cone penetrometer testing holes, multi-
level installations and situations where the size of the annulus to be grouted
is minimal.
12.10 CASE HISTORY FEATURING HIGH-SOLIDS

BENTONITE GROUT
12.10.1 Design and Construction of Deep Ground Water Monitoring

Wells (Lawrence D. Beard, 1991)
Deep groundwater monitoring wells were constructed at the Colbert
Landfill Superfund Project in Spokane County, Washington. The Colbert
Landfill is an inactive 40 acre (16.2 ha) municipal solid waste landfill. The
subsurface lithology at the facility consists of unconsolidated gravel, sand
and silt deposits overlying granitic bedrock. The purpose of installing deep
monitoring wells at the site was to investigate the presence and extent of
spent chlorinated solvents in two aquifers.
Case history featuring high-solids bentonite grout 295
Fig. 12.8 Diagram of a multilevel well installation sealed with the sand/bentonite
injector.
The groundwater monitoring wells installed as part of the Colbert Pro-

ject included design components applied by the consultant for the investiga-
tion and the use of a 30% solids bentonite grout. The 30% bentonite grout
became available shortly before the August 1989 start of field activities.
A typical monitoring well diagram for the Colbert Project is shown in
Figure 12.9.
Bentonite pellets were considered impractical because of potential ma-
terial placement difficulties. The annular sealant had to be paced through a
220 ft (67 m) water column. Bentonite pellets would swell significantly as
they settled through the water column, and frequent material bridging was
an anticipated problem.
Cement grout was not considered for the annular sealant because the
potential for elevated groundwater pH (due to the high alkalinity of ce-
ment). Elevated pH was a particular concern at the Colbert Project site
because well clusters were used, and the screened zone for shallow monitor-
ing wells are adjacent to the annular sealant for deeper monitoring wells.
The difficulty of maintaining a continous seal when grout placement spans
several days, the potential for permanent fluid migration pathways to form
during cement curing, and the potential impact of the heat of hydration on
PVC casing material strength characteristics also contributed to the deci-
sion not to use cement grout.
With the elimination of bentonite pellets and cement grout from con-
sideration as annular sealants, bentonite grout also posed some potential
performance problems, including filter pack intrusion, grout loss (in coarse
formations), dehydration (and cracking) in the unsaturated zone and settle-
ment. Many of these potential performance problems are directly or in-
directly related to the low percentage of solids typically obtained with
bentonite grout mixtures (typically 20%). A recently introduced product,
PureGold Grout, was considered because of the manufacturer's claim that it
achieves a 75% higher percentage of solids than other bentonite grouts.
Because Pure Gold was a new product, its physical and chemical proper-
ties were evaluated prior to use. Evaluation included a limited number of
laboratory tests to assess physical characteristics, and a review of chemical
analyses provided by the manufacturer. Laboratory tests verified that
PureGold Grout achieved 35% solids and maintained a workable consist-
ency when mixed to the manufacturer's specifications. For comparison, wet
Upper
Aquifer
Lower Aquifer
Fine Sand
PVC Screen
F;]:=P~+--- F'iHer Pack
~=Ii'r.a--~~~::~:e~eel
Fig. 12.9 Typical deep monitoring well installation for the Colbert Project.
Conclusions 297
densities were also determined for hydrated bentonite chips and bentonite
pellets. The wet density measured for the grout is about the same as that
measured for bentonite chips, and only about 5% lower than the measured
density of bentonite pellets. Although the high density of the grout reduced
the concern of grout intrusion, a 2 ft (0.6 m) layer of fine sand above the
filter pack was incorporated into the monitoring well design to reduce
further the potential for grout intrusion into the filter pack.
Samples of PureGold Grout were also subjected to repeated dehydra-
tion/rehydration cycles to determine its rehydration characteristics. Visual
observations indicated that the grout rehydrates even after oven-drying.
Thus, dehydration cracks that may form in the vadose zone would tend to
seal when rehydrated.
12.11 CONCLUSIONS
Since the early 20th century, sodium bentonite has been used more than
any other mineral known. The majority of sodium bentonite has been used
in the foundry industry, the pelletization industry and well drilling. With the
recent awareness in groundwater protection, sodium bentonite has been
used extensively to seal boreholes to prevent contamination of groundwater
by surface or interformationalleakage.
Because of its large surface area, high water absorption and swelling
properties, sodium bentonite is an excellent seal for boreholes and the
annulus in boreholes. Sodium bentonite generally has a hydraulic conduc-
tivity value of 1 x 10- 7 to 1 x 10- 9 cm/s when hydrated with fresh water.
The type of bentonite selected to seal a borehole and/or the annulus is
dependent on several factors. These factors include the diameter and depth
of the borehole, the annular space, the depth of the water column, etc.
Manufacturers of bentonite have developed products that are suitable for
most sealing scenarios.
Bibliography
Adams, L. M. and Lipscomb, 1. R. (1984) Evaluation of Mine Seals Constructed in

1967 at Elkins, Randolph County, Wv, US Bureau of Mines Rep. 8852.
Adisoma, G. S. (1987) The Influence of Dynamic Loading on the Sealing Perform-
ance of Cement Borehole Plugs, MS Thesis, Univ. of Arizona, Tucson.
Adisoma, G. S. and Daemen, 1.1. K. (1984) Laboratory assessment of the effect of
drying on the performance of cement borehole plugs. Proc. Waste Mgmt '84
Sym., Tucson, Vol. 1, pp. 579-83.
Adisoma, G. S. and Daemen, 1. 1. K. (1988) Experimental Assessment of the Influence
of Dynamic Loading on the Permeability of Wet and Dried Cement Borehole
Seals. US Nuclear Regulatory Commission Rep. NUREGjCR-5129,
Washington, DC.
Akgun, H. and Daemen, 1.1. K. (1986) Size Influence on the Sealing Performance of
Cementitious Borehole Plugs. US Nuclear Regulatory Commission Rep.
NUREGjCR-4738, Washington, DC.
Akgun, H. and Daemen, 1.1. K. (1989) Bond Strength of Cement Borehole Plugs
in Salt. US Nuclear Regulatory 'Commission Rep. NUREGjCR-5401,
Washington, DC.
Akgun, H. and Daemen, 1.1. K. (1991c) Bond Strength of Cementitious Borehole
Plugs in Welded Tuff. US Nuclear Regulatory Commission Rep. NUREGj
CR-4295, Washington, D.C.
Akgun, H. and Daemen, 1.1. K. (1994a) Performance assessment of cement grout
borehole plugs in basalt. Engineering Geology, 37, 137 -48.
Akgun, H. and Daemen, 1.1. K. (1994b) Strength of cement grout plugs in salt.
Transactions, Society of Mining, Metallurgy and Exploration, Inc., 296,1818-21.
Al-Dhahir, Z. A. and Tan, S. B. (1968) A note on one-dimensional constant-head
permeability test. Geotechnique, 18, 499-505.
American Petroleum Institute (1986) Specifications for Materials and Testing for
Well Cements, 3rd edn, API, Production Department, Dallas.
Andersson, I.-E. and Persson, O. (1985) Evaluation of single-hole hydraulic tests in
fractured crystalline rock by steady-state and transient methods: a comparison.
Materials Research Soc. Sym. Proc., Vol. 50, Stockholm, pp. 155-63.
Anon (1987) Shaft Abandonment Guidelines. Utah Dept. of Natural Resources, Div.
of Oil, Gas and Mining, Salt Lake City.
ANSIjAWWA A 100-90, AWWA Standards for Water Wells. American Water
Works Association, Denver, Colorado.
Applied Technology Council (1978) Tentative Provisionsfor the Development ofSeis-
mic Regulations for Buildings. US National Bureau of Standards, Spec. Publ.
510, Washington, DC.
Bibliography 299
Ariman, T. and Muleski, G. E. (1979) A review of the response of buried pipelines
under seismic excitations, Lifeline Earthquake Engineering-Buried Pipelines,
Seismic Risk, and Instrumentation. AS ME, PVP-34, New York, pp. 1-29.
Arnold, D. M. and Paap, H. J. (1979) Quantitative monitoring of water flow behind
and in wellbore casing. Soc. Petro. Eng. Trans., 287, 121-30.
ASTM D698-78, Test method for laboratory compaction characteristics of soil using
standard effort (12,400 ft-Ibf/ft 3 ) (600 kN-m/m 3 ). Annual Book of ASTM Stan-
dards, Vol. 04.08, American Society for Testing and Materials, Philadelphia.
ASTM D2434-68, Test method for permeability of granular soils (constant head).
Annual Book of ASTM Standards, Vol. 04.08, American Society for Testing and
Materials, Philadelphia.
ASTM D5092, Standard practice for design and installation of ground water
monitoring well in aquifers. Annual Book of ASTM Standards, Vol. 04.08, American
Society for Testing and Materials, Philadelphia.
ASTM D5299-92, Standard guide for decommissioning of groundwater wells,
vadose zone monitoring devices, boreholes, and other devices for environmental
activities, Annual Book of AS1M Standards, Vol. 04.08, American Society for
Testing and Materials, Philadelphia.
ASTM (1991) Standard of Ground Water and Vadose Zone Investigations. American
Society for Testing and Materials, Philadelphia.
Aul, G. N. and Cervik, J. (1979) Grouting Horizontal Drainage Holes in Coalbeds, US
Bureau of Mines Rep. 8375.
Auld, F. A. (1979) Design of concrete shaft linings. Proc. Instn Civil Engineers, Part
2,67, Sept., 817-32.
Auld, F. A. (1982a) Ultimate strength of concrete shaft linings and its influence on
design. Proc. Sym. on Strata Mechanics, Newcastle upon Tyne, Elsevier, pp.
134-40.
Auld, F. A. (1982b) Concrete in underground works, Concrete Society Technical
Report No. 105, The Concrete Society, London.
Auld, F. A. (1982c) Concrete in underground works. The Concrete Society North
West Region Sym., Concrete in the Energy Industry.
Auld, F. A. (1983) Design of underground plugs. Int. J. Mining Engineering, 1,
189-228.
Aylmore, L. A. G and Quirk, J. P. (1960) Domain or turbo static structure of clays.
Nature, 187, 1046.
Baar, C. A. (1977) Applied Salt-Rock Mechanics, Elsevier, Amsterdam.
Baes, C. F. Jr, Gilpatrick, L. 0., Kitts, F. G. et al. (1993) The Effect of Water in Salt
Repositories, Oak Ridge National Lab. Rep. ORNL-5950, Oak Ridge.
Barbreau, A., Heremans, R. and Skytte, J. B. (1980) Radionuclide migration in
geological formation. Proc. 1st European Community Conf Radioactive Waste
Management, Luxembourg, pp. 515-30.
Baroid Drilling Fluids, Inc. and Environmental Planning Groups, Inc. (1995) Water
Well, Monitoring Well and Borehole Abandonment Documents for the United
States, EPG Publishing, Barrington, IL.
Bartlett, J. W. and Koplik, C. M. (1979) Information Base for Waste Repository
Design, Volume 7-Executive Summary. US Nuclear Regulatory Commission
Rep. NUREG/CR-0495, Washington, DC.
Bauer, A. and Calder, P. N. (1971) The influence and evaluation of blasting on
stability. Proc. 1st Int. Conf on Stability in Open Pit Mining, Vancouver, pp.
83-95.
Bear, J. (1972) Dynamics of Fluids in Porous Media, American Elsevier, New York.
Bear, J. (1979) Hydraulics of Groundwater, McGraw-Hill, New York.
300 Bibliography
Beard, L. D. (1991) Design and construction of deep groundwater monitoring wells.

Current Practices in Ground Water and Vadose Zone Investigations, ASTM STP
1118, American Society for Testing and Materials, Philadelphia.
Benjamin, J. R. and Cornell, C. A. (1970) Probability, Statistics, and Decisionfor Civil
Engineers, McGraw-Hill, New York.
Bernabe, Y., Brace, W. F. and Evans, B. (1982) Permeability, porosity and pore
geometry of hot-pressed calcite. Mechanics of Materials, 1, 173-83.
Bernaix, J. (1969) New laboratory methods of studying the mechanical properties of
rocks. Int. J. Rock Mech. Min. Sci. 6, 43-90.
Bertane, M. J. (1985) The Use of Groutfor the Sealing af Monitoring Wells, American
Colloid Company Grout Report, Arlington Heights.
Bingham, E. C. (1916) An Investigation of the Laws of Plastic Flow, US Bureau of
Standards Scientific Paper No. 278.
Binnall, E. P., Benson, S. M., Tsao, L. et al. (1987) Critical Parameters for a High-
Level Waste Repository, Volume 2: Tuff. US Nuclear Regulatory Commission
Blatt H., Middleton, G. and Murray, R. (1980) Origin of Sedimentary Rocks,
Prentice-Hall, Englewood Cliffs.
Borgesson, L. (1990) Interim Report on the Laboratory and Theoretical Work in
Modeling the Drained and Undrained Behavior of Buffer Materials. SKB Rep.
90-45, Stockholm.
Borgesson, L., Hokmark, H. and Karnland, O. (1988) Rheological Properties
of Sodium Smectite Clay. Swedish Nuclear Fuel and Waste Management Co.
Technical Rep. SKB 88-30, Stockholm.
Borns, D. J. and Stormont, J. C. (1989) The delineation of the disturbed rock zone
surrounding excavations in salt. Proc. 30th US Rock Mech. Sym., W. Virginia
Univ., pp. 353-60.
Bourbie, T. and Walls, J. (1982) Pulse decay permeability: analytical solution and
experimental test. Sac. Petro. Eng. J., 22, 719-21.
Bowen, R. (1975) Grouting in Engineering Practice, Applied Sciences, London.
Brace, W. F. (1980) Permeability of crystalline and argillaceous rocks. Int. J. Rock
Mech. Min. Sci. & Geomech. Abstr., 17, 241-51.
Brace, W. F. Walsh, J. B. and Frangos, W. T. (1968) Permeability of granite under
high pressure. J. Geophys. Res., 73, 2225-36.
Braun, E. R. (1969) Geology and Ore Deposits of the Marble Peak Area, Santa
Catalina Mountains, Pima County, Arizona. MS Thesis, Univ. of Arizona,
Tucson.
Bredehoeft, J. D. and Papadopulos, S. S. (1980) A method for determining the hy-
draulic properties of tight formations. Water Resources Res. 16,233-8.
Bredehoeft, J. D. Neuzil, C. E. and Milly, P. C. D. (1983) Regional Flow in the
Dakota Aquifer: A Study of the Role of Confining Layers. US Geological Survey
Water-Supply Pap. 2237.
Bredehoeft, J. D., England, A. W., Stewart, D. B. et al. (1978) Geological Disposal af
High-Level Radioactive Wastes-Earth-Science Perspectives. US Geological
Survey Cir. 779.
British Standards Institution (1969a) BS 449:Part 2:1969, Specification for the use of
structural steel in building, Part 2. Metric units, British Standards Institution,
London.
British Standards Institution (1969b) CP 114:1969, The structural use of reiriforced
concrete in buildings, British Standards Institution, London (withdrawn).
British Standards Institution (1983) BS 1881:Part 116:1983, Testing concrete,
Part 116. Method for determination of compressive strength of concrete cubes,
British Standards Institution, London.
Bibliography 301
British Standards Institution (1985) BS 8110:Part 1:1985, Structural use of concrete,

Part 1. Code of practice for design and construction, British Standards Institu-
tion, London.
British Standards Institution (1987) BS 8007:1987, Code of practice for design of
concrete structures for retaining aqueous liquids, British Standards Institution,
London.
British Standards Institution (1990) BS 5328: Part 1:1990, Guide to specifying con-
crete mixes. BS 5328:Part 4:1990, Specification for the procedures to be used
in sampling, testing and assessing compliance of concrete, British Standards
Institution, London.
Brodsky, N. S., Zeuch, D. H. and Holcomb, D. 1. (1995) Consolidation and permea-
bility of crushed WIPP salt in hydrostatic and triaxial compression. Proc. 35th
Rock Mech. Sym., Lake Tahoe, pp. 497-502.
Buck, A. D. and Burkes, J. P. (1979) Examination of Grout and Rock from Duval
Mine, New Mexico. US Army Engineer Waterways Experiment Station, Misc.
Pap. SL-79-16, Vicksburg.
Buck, A. D. and Mather, K. (1982) Grout Formulationsfor Nuclear Waste Isolation
Office of Nuclear Waste Isolation Rep. ONWI-413, Columbus.
Buck, A. D., Burkes, J. P. and Rhoderick, J. E. (1981) Examination of ERDA-lO
Grout Specimens at Different Ages. US Army Engineers Waterways Experiment
Station, Misc. Pap. SL-81-20, Vicksburg.
Burkes, J. P. and Rhoderick, J. E. (1983) Petrographic Examination of Bell Canyon
Tests (BCT) I-FF Field Grouts over a Three-Year Period. Office of Nuclear
Waste Isolation Rep. SL-83-12.
Bush, D. D. and Piele, S. (1986) A Full-Scale Borehole Sealing Test in Salt Under
Simulated Downhole Conditions, Vol. 1 Office of Nuclear Waste Isolation Rep.
BMI/ONWI-573(1).
Butcher, B. M. (1991) The Advantages of a Salt/Bentonite Backfillfor Waste Isolation
Pilot Plant Disposal Rooms. Sandia National Lab. Rep. SAND90-3074,
Albuquerque.
Calvert, D. G. (1980) Shrinkage-compensating cements in oil well cementing. Sym.
on Expansive Cement, ACI, Detroit, pp. 193-7.
Carman, P. C. (1937) Fluid flow through granular beds. Trans. Inst. Chemical Engin-
eers, 15, 150-66.
Carman, P. C. (1939) Permeability of saturated sands, soils and clays. J. Agricultural
Sci., 29, Part 2, pp. 262-73.
Carpenter, R. B., Brady, J. L. and Blount, C. G. (1992) Effects of temperature and
cement admixtures on bond strength. J. Petro. Tech., 4, 880-5.
Carrera, J. and Neuman, S. P. (1982) Quasi three-dimensional finite element model
of the Madrid basin in Spain. J. Arizona-Nevada Academy of Science, 17, 44-5.
Carslaw, H. S. and Jaeger, J. C. (1959) Conduction of Heat in Solids, Clarendon
Press, Oxford.
Carter, N. L. and Hansen, F. D. (1983) Creep of rocksalt. Tectonophysics, Int. J.
Geotectonics and the Gelogy and Physics of the Interior of the Earth, 92, 275-333.
Cedergren, H. R. (1977) Seepage, Drainage, and Flow Nets, John Wiley and Sons,
New York.
Cetintas, A. (1994) Experimental, Assessment of Borehole Plug Stability. PhD Dis-
sertation, Univ. of Arizona, Tucson.
Chamber of Mines of South Africa Code of Practice (1983) Construction of under-
ground plugs and bulkhead doors using grout intrusion concrete.
Chandler, N. A., Kjartanson, B. H., Kozak, E. T. et al. (1992) Monitoring the
geomechanical and hydrological response in granite for AECL Research's Buffer/
Container Experiment. Proc. 33rd US Rock Mech. Sym., Santa Fe, pp. 161-70.
302 Bibliography
Chapman, S. L. (1972) Idaho Department of Water Administration closes flowing

artesian well. Water Well Journal, October.
Cheatham, J. B, and McEver, J. W. (1964) Behavior of casing subjected to salt load-
ing. JPT, AIME Trans., 16, 1069-75.
Chekan, G. J. (1985) Design of Bulkheads for Controlling Water in Underground
Mines, US Bureau of Mines Cir. 9020.
Childs, E. C. and Collis-George, N. (1950) The permeability of porous material.
Proc. Royal Soc., London, 20lA, 392-405.
Christensen, C. L. (1980) Sandia Borehole Plugging Program for the Waste Isolation
Pilot Plant, Sandia National Lab. Rep. SAND80-0390C, Albuquerque.
Christensen, C. L. and Peterson, E. W. (1981) The Bell Canyon Test Summary Re-
port, Sandia National Lab. Rep. SAND80-1375, Albuquerque.
Clauser, C. (1992) Permeability of crystalline rocks. Eos, Transactions, AGU, 73
233-8.
Coates, D. F. (1981) Rock Mechanics Principles, Monograph 874, Rev. 4th edn,
CANMET, Energy, Mines and Resources, Ottawa.
Cobb, S. L. (1981) A Laboratory Facility for Testing the Performance of Borehole
Plugs in Rock Subjected to Poly axial Loading. MS Thesis, Univ. of Arizona,
Tucson.
Colangelo, R. V. and Upadhyay, H. D. (1990) The origin and physical properties of
bentonite and its usage in the groundwater monitoring industry. 11th Annual
Natl. Conj., Superfund 1990, Washington, DC.
Co mite Europeen du Beton-Federation Internationale de la Precontrainte (1970)
International recommendations for the design and construction of concrete struc-
tures. Principles and Recommendations, FIP Sixth Congress, Prague.
Cook, R. F. (1983) Long-term closure of a shaft excavation through evaporite de-
posits. Proc. 1st Int. Potash Technology Conj., Saskatoon, pp. 275-81.
Coons, W. E., Meyer, D. and Kelsall, P. C. (1982) Evaluation of Polymer Concrete for
Application to Repository Sealing. Office of Nuclear Waste Isolation Rep.
ONWI-41O, Columbus.
Cooper, H. H., Bredehoeft, J. D. and Papadopulos, I. S. (1967) Response of a finite-
diameter well to an instantaneous charge of water. Water Resources Res. 3,
263-9.
Cornwell, F. E. (1953) Flow from a short section of test hole below groundwater
level. Theory and Problems of Water Percolation, US Bureau of Reclamation
Engineering Monograph, pp. 65-8.
Costin, L. S. and Wawersik, W. R. (1981) Creep Healing of Fractures in Rock Salt.
Sandia National Lab. Rep. SAND80-1375, Albuquerque.
Craig, W. H. (1982) Conflicts in dynamic modeling in geomechanics. Design for
Dynamic Loading, the Use of Model Analysis, Construction Press, London, pp.
1-6.
Crouthamel, D. R. (1991) In situ Flow Testing of Borehole Plugs. MS Thesis, Univ.
of Arizona, Tucson.
Crouthamel, D. R. and Daemen, 1. 1. K. (1991) Pressurized grout applications in
fractured tuff for containment of radioactive waste. Geotechnical and Geological
Engineering, 9, 53-62.
Crouthamel, D. R., Fuenkajorn, K. and Daemen, J. 1. K. (1993) In situ flow testing of
cement borehole plug in welded tuff. Int. J. Rock Mech. Sci., 30,1503-6.
Daemen, J. J. K., South, D. L., Greer, W. B. et al. (1983) Rock Mass Sealing-Annual
Report June 1983-May 1984. US Nuclear Regulatory Commission Rep.
Daemen, J. J. K., Greer, W. B., Adisoma, G. S. et al. (1985) Rock Mass Sealing-An-
nual Report June 1983-May 1984. US Nuclear Regulatory Commission Rep.
Bibliography 303

Daemen, J. J. K., Greer, W. B., Fuenkajorn, K. et al. (1986) Experimental Assessment
of Borehole Plug Performance. us Nuclear Regulatory Commission Rep.
Dagan, G. (1978) A note on packer, slug, and recovery tests in unconfined aquifers.
Water Resources Res. 14, 929-34.
Dames & Moore (1989) The October 17, 1989 Loma Prieta Earthquake, Dames &
Moore Special Report, Los Angeles.
Daniel, D. E. (1989) In situ hydraulic conductivity tests for compacted clay. J.
Geotech. Eng., 115, 1205-26.
D'Appolonia Consulting Engineers (1976) The Status of Borehole Plugging & Shaft
Sealing for Geologic Isolation of Radioactive Waste. Office of Nuclear Waste
Isolation Rep. ONWI-15, Columbus.
D'Appolonia Consulting Engineers (1980a) Repository Sealing Design
Approach-1979. Office of Nuclear Waste Isolation Rep. ONWI-55, Columbus.
D'Appolonia Consulting Engineers (1980b) Repository Sealing: Evaluation of
Materials Research Objectives & Requirements-1980. Office of Nuclear Waste
Isolation Rep. ONWI-108, Columbus.
D'Appolonia Consulting Engineers (1981) Sealing Consideration for Repository
Shafts in Bedded and Dome Salt. Office of Nuclear Waste Isolation Rep. ONWI-
225, Columbus.
Davison, C. C. (1980) Physical Hydrogeology Measurements Conducted in Boreholes
WN-1, WN-2, and WN-4 to Assess the Local Hydraulic Conductivity and Hy-
draulic Potential of a Granitic Rock Mass. National Hydrology Research Insti-
tute, Inland Waters Directorate, Environment Canada Technical Record TR-26,
Manitoba.
Desai, C.S. and Abel, J. F. (1972) Introduction to the Finite Element Method-A
Numerical Method for Engineering Analysis, Van Nostrand Reinhold, New
York.
Desai, C. S. and Varadarajan, A. (1987) A constitutive model for quasi-static behav-
ior of rock salt. J. Geophys. Res., 92, 11445-56.
De Wiest, R. J. M. (1966) On the storage coefficient and the equation of ground-
water flow. J. Geophys. Res., 71, 1117-22.
Dixon, D. A. and Gray, M. N. (1985) The Engineering Properties of Buffer
Material-Research at Whiteshell Nuclear Research Establishment. Atomic
Energy of Canada Limited Technical Record TR-350, Volume III, Pinawa.
Dixon, D. A., Gray, M. N. and Thomas, A.W. (1985) A study of the compaction
properties of potential clay-sand buffer mixtures for use in nuclear fuel waste
disposal. Engineering Geology, 21, 247-55.
Dixon, D. A., Gray, M. N., Baumgartner, P. and Rigby, G. L. (1986) Pressures acting
on waste containers in bentonite-based materials. Proc. 2nd Int. Conf on Radio-
active Waste Management, Winnipeg.
Dixon, D. A., Wan, A. W. L., Graham, J. and Campbell, S. L. (1993) Assessment of
self-sealing and self-healing abilities of dense, high-bentonite-content sealing
materials. Proc. 1993 Joint CSCE-ASCE National Conference on Environmental
Engineering, Montreal, Quebec.
Doe, T. W., Long, J. C. S., Endo, H. K. and Wilson, C. R. (1982) Approaches to
evaluating the permeability and porosity of fractured rock masses. Proc. 23rd
Sym. Rock Mech., Univ. of California, Berkeley, pp. 30-8.
Dowding, C. H. (1977) Seismic stability of underground openings. Proc. Rockstore
Conf, Stockholm, Vol. 2. Pergamon, Oxford, pp. 231-8.
Dowding, C. H. and Rozen, A. (1978) Damage to rock tunnels from earthquake
shaking. J. Geotech. Eng. Div., ASCE, 104, 175-91.
304 Bibliography
Dressel, W. M. and Volosin, J. S. (1985) Inverted Pyramid-Shaped Plugs for Closing

Abandoned Mine Shaft-Galena, KS, Demonstration Project. US Bureau of Mine
Cir.8998.
Driscoll, F. G. (1986) Groundwater and Wells, 2nd edn, Johnson Division, St Paul,
Minnesota.
Edil, T. B. and Muhanna, A. S. (1993) Discussion on characteristics of a bentonite
slurry as a sealant. Geotech. Testing J., 16, 135-7.
Edwards, A. L. (1972) TRUMP: A Computer Programfor Transient and Steady State
Temperature Distributions in Multidimensional Systems. Lawrence Livermore
National Lab. Rep. UCRL-14754, Livermore.
Ehgartner, B. (1991) Summary of Characterization Tests on a Low Density Epoxy
Grout for Use as a Bulkhead Seal Material in the US Strategic Petroleum
Reserve. Sandia National Lab. Rep. SAND91-1112, Albuquerque.
Ehlig-Economides, C. A. and Ramey, H. J. (1979) Pressure Buildup for Wells Pro-
duced at a Constant Pressure. Society of Petroleum Engineers of AIME, SPE
7985, Richardson.
Eilers, L. H. (1973) Borehole Sealing Final Report. Oak Ridge National Lab. Rep.
ORNL/Sub/15966-73/1, Oak Ridge.
Eilers, L. H. (1974) Sealing AEC No.1 Well at Lyons Kansas. Oak Ridge National
Lab. Rep. ORNL/Sub/33542-74/1, Oak Ridge
Einarson, D. S., and Abel, J. F. (1990) Tunnel bulkheads for acid mine drainage.
Proc. Int. Sym. Unique Underground Structures, Denver, Vol. II, Chapter 71.
Endell, K., Loos, W., Meischeider, H. and Berg, V. (1938) Ueber Zusammenhaenge
zwischen Wasserhaushalt der Tonminerale und Bodenphysikalischen Eigenschaften
Bindiger Boeden, (On the relations between water management by clay minerals
and the soil physical properties of cohesive soils), Veroeff. Dtsch. Forsch. Boden-
mech.5.
Engelmann, H.-J., Boochs, P. W., Hansel, W. and Peters, L. (1989) Dams as sealing
systems in rock salt formations-test dam construction and determination of
permeability. Sealing of Radioactive Waste Repositories, Braunschweig, Federal
Republic of Germany, May 22-25, a workshop organized by OECD Nuclear
Energy Agency and Commission of the European Communities.
Evans, B. (1983) Alteration of porosity and permeability by plastic flow processes in
rocks. Geol. Soc. Am. Abstr. WIth Programs, 15, 568.
Evans, E. A. and Quigley, R. M. (1992) Permeation of sand-bentonite mixtures with
municipal waste leachate. Proc. 4th Can. Geotech. Soc. Conf, Pap. 68.
Fairchild, D. M. and Canter, L. W. (1984) Abandoned wells and groundwater.
Ground Water Age, 18,33-36,39.
Fairhurst, c., Gera, F., Gnirk, P. et al. (1993) The international Stripa Project: an
overview. Tunnelling and Underground Space Technology, 8, 315-43.
Farmer, I. W. (1985) Coal Mine Structures, Chapman and Hall, London.
Farmer, I. W. and Gilbert, M. J. (1981) Time dependent strength reduction of rock
salt. Proc. 1st Conf. on Mechanical Behavior of Salt, Pennsylvania State Univ.,
pp. 3-18.
Fernandez, J. A. (1985) Repository Sealing Planfor the Nevada Nuclear Waste Stor-
age Investigations Project Fiscal Years 1984 Through 1990. Sandia National Lab.
Rep. SAND84-0910, Albuquerque.
Fernandez, J. A. and Freshley, M. D. (1984) Repository Sealing Concepts for the
Nevada Nuclear Waste Storage Investigations Project. Sandia National Lab.
Fernandez, J. A., Case, J. B., Givens, C. A. and Carney, B. C. (1993) A Strategy to
Seal Exploratory Boreholes in Unsaturated Tuff. Sandia National Lab. Rep.
SAND 93-1184, Albuquerque.
Bibliography 305
Fernandez, J. A., Kelsall, P. c., Case, J. B. and Meyer, D. (1987) Technical Basisfor
Performance Goals, Design Requirements, and Material Recommendationsfor the
NNWSI Repository Sealing Program. Sandia National Laboratories Rep.
SAND-84-1895, Albuquerque.
Fernandez, R., McGowen, c., Nolan, E. et al. (1976) Borehole Plugging by Compac-
tion Process. Charles Stark Draper Lab. Inc., Cambridge.
Ferris, J. G., Knowles, D. B., Brown, R. H. and Stallman, R. W. (1962) Theory of
Aquifer Tests. US Geological Survey Water-Supply Pap. 1536-E.
Finley, R. E. and Jones, R. L. (1994) Results of Brine Flow Testing and Disassembly of
a Crushed Salt/Bentonite Block Seal at the Waste Isolation Pilot Plant. Sandia
National Lab. Rep. SAND93-0808, Albuquerque.
Finley, R. E. and Tillerson, J. R. (1992) WIPP Small-Scale Seal Performance
Tests-Status and Impacts. Sandia National Lab. Rep. SAND91-2247, Albuquerque.
Flak, L. H. and Brown, J. J. (1990) Case history of an ultradeep disposal well in
western Colorado. SPE Drilling Engineering, 15, 39-44.
Fliigge, W. (1967) Stresses in Shells, Springer-Verlag, Berlin.
Fogg, G. E., Simpson, E. S. and Neuman, S. P. (1979) Aquifer Modeling by Numerical
Methods Applied to an Arizona Groundwater Basin. Natural Resource Systems Rep.
32, Department of Hydrology and Water Resources, Univ. of Arizona, Tucson.
Foreman, J. W. (1971) Deep mine sealing. Proc. Acid Mine Drainage Workshop,
Ohio Univ., pp. 19-45.
Forster, C. B. and Gale, 1. E. (1980) A Laboratory Assessment of the Use of Borehole
Pressure Transients to Measure the Permeability of Fractured Rock Masses.
Lawrence Berkeley Lab. Rep. LBL-8674, SAC-28, UC-70, Berkeley.
Forster, C. B. and Gale, J. E. (1981) A Field Assessment of the Use of Borehole
Pressure Transients to Measure the Permeability of Fractured Rock Masses
Swedish Nuclear Fuel Supply Co., Stockholm, and Lawrence Berkeley Lab.
Rep. 34, LBL-11829, SAC-34, UC-70, Berkeley.
Frazee, J. M. (1990) Abandoning test on geotechnical holes-criteria, regulations, and
resource concerns. National Drillers Buyers Guide, February, p. 6.
Freeze, R. A. and Cherry, J. A. (1979) Groundwater, Prentice-Hall, Englewood.
Fuenkajorn, K. and Daemen J. J. K. (1984) Experimental assessment of borehole
wall drilling damage in basaltic rocks. Proc. 25th US Rock Mech. Symp. North-
western Univ., Evanston, pp. 774-83.
Fuenkajorn, K. and Daemen, J. J. K. (1986a) Experimental Assessment of Borehole
Wall Drilling Damage in Basaltic Rocks. US Nuclear Regulatory Commission
Fuenkajorn, K. and Daemen, J. J. K. (1986b) Rock-cement interface effects on
borehole plug permeability. Proc. Waste Management '86 Sym., Tucson, pp.
251-6.
Fuenkajorn, K. and Daemen, J. J. K. (1987) Mechanical interaction between rock
and multi-component shaft or borehole plugs. Proc. 28th US Rock Mech. Sym.,
Univ. of Arizona, Tucson, pp. 165-72.
Fuenkajorn, K. and Daemen, J. J. K. (1988a) Borehole closure in salt. Proc. 29th US
Rock Mech. Sym. Univ. of Minnesota, Minneapolis, pp. 191-8.
Fuenkajorn, K. and Daemen, J. J. K. (1988b) Borehole Closure in Salt. US Nuclear
Regulatory Commission Rep. NUREG/CR-5243, Washington, DC.
Fuenkajorn, K. and Daemen, J. J. K. (1991a) Mechanical Characterization of the
Densely Welded Apache Leap Tuff. US Nuclear Regulatory Commission Rep.
NUREG/CR-5688, Washington, DC.
Fuenkajorn, K. and Daemen, J. J. K. (1991b) Cement borehole plug performance in
welded tuff. Proc. 32nd US Rock Mech. Sym., Univ. of Oklahoma, Norman, pp.
723-32.
306 Bibliography
Fuenkajorn, K. and Daemen, J. J. K. (1991c) Brine-mixed bentonite borehole plugs.

Proc. Waste Management '91 Sym. Tucson, pp. 775-83.
Fuenkajorn, K. and Daemen, J. J. K. (1992a) An empirical strength criterion for
heterogeneous tuff. Eng. Geo., 32, 209-23.
Fuenkajorn, K. and Daemen, J. J. K. (1992b) Drilling-induced fractures in borehole
walls. J. Petro. Tech., 44, 210-16.
Fuenkajorn, K. and Daemen, J. J. K. (1992c) Borehole Deformation and Stability in
Welded Tuff. US Nuclear Regulatory Commission Rep. NUREG/CR-5687,
Washington, DC.
Fuenkajorn, K. and Daemen, J. J. K. (1992d) Borehole sealing. Proc. 2nd Int.
Con[. Compressed-Air Energy Storage. Electric Power Research Institute, San
Francisco, pp. 5.1-5.21.
Fuenkajorn, K. and Serata, S. (1992) Geohydrological integrity of CAES in rock
salt. Proc. 2nd Int. Con[. Compressed-Air Energy Storage, Electric Power
Research Institute, San Francisco, pp. 4.1-4.21.
Fuenkajorn, K. and Serata, S. (1993) Numerical simulation of strain-softening and
dilation of rocks. Int. J. Rock Mech. Min. Sci., 30, 1303-6.
Fuenkajorn, K. and Serata, S. (1994) Dilation-induced permeability increase around
caverns in salt. Proc.1st North American Rock Mech. Sym. June 1-3, Austin, pp.
648-56.
Gaber, M. S. and Fisher, B. O. (1988) Michigan Water Well Grouting Manual.
Michigan Department of Public Health, Lansing.
Gale, J. E. and Raven, K. G. (1979) An Assessment and Evaluation of Compliance
Effects of Borehole Transient Test Equipment. Water Resources Branch, IWD,
Dept. of Fisheries and Environment, Ottawa.
Garcia, J. A. and Cassidy, S. M. (1938) Bulkhead for coal mines. Trans. AIME, 130,
284.
Garrett, W. S. and Campbell Pitt, L. T. (1958) Tests on an experimental under-
ground bulkhead for high pressures, J. S. Afr. Inst. Mining Metall., 59, 123-43.
Garrett, W. S. and Campbell Pitt, L. T. (1961) Design and construction of under-
ground bulkheads and water barriers. Trans. 7th Commonwealth Mining and
Metallurgical Congress, Johannesburg, pp. 1283-99.
Gass, T. (1981) The impact of abandoned wells on groundwater quality. Water
Well. J., March.
Gass, T. (1988) The impact of abandoned wells. Water Well J., January.
Gass, T. E., Lehr, J. H. and Heiss, H. W. Jr (1977) Impact of Abandoned Wells on
Ground Water. EPA-600/3-77-095, US Environmental Protection Agency, R.S.
Kerr Laboratories, Ada, Oklahoma.
Gaudette, M. V. and Daemen J. J. K. (1988) Bentonite Borehole Plug Flow Testing
With Five Water Types. US Nuclear Regulatory Commission Rep.
George, C. and Faul, R. (1985) Cementing techniques for solution mining wells and
salt storage domes: the state-of-the art. Proc. Sym. Solution Mining of Salts and
Brines, SME of AIME, New York, pp. 11-23.
Gerstle, W. H. and Jones, A. (1986) Mechanical Properties of Crushed Salt/Bentonite
Blocks. Sandia National Lab. Rep. SAND86-0707, Albuquerque.
Gibson, R. E. (1963) An analysis of system flexibility and its effect on time-lag in
pore-water pressure measurements. Geotechnique, 13, 1-11.
Gloyna, E. F. and Reynolds, T. D. (1961) Permeability measurements of rock salt. J.
Geophys. Res., 66, 3913-21.
Goodman, R. E. and Sundaram, P. N. (1980) Permeability and piping in fractured
rocks. ASCE J. Geotechnical Eng., 106, 485-98.
Bibliography 307
Gray, M. N. (1993) The OECD/NEA International Stripa Project, 1980-1992. Over-

view, Volume III-Engineered Barriers, SKB, Stockholm, Sweden.
Gray, M. N., Cheung, S. C. H. and Dixon, D. A. (1984) The Influence of Sand Con-
tent on Swelling Pressures and Structure Development in Statically Compacted
Na-Bentonite. Atomic Energy of Canada Rep. AECL-7825, Pinawa.
Gray, T. A. and Gray, R. E. (1992) Mine closure, sealing, and abandonment. SM E
Mining Engineering Handbook- Vol. 1, 2nd edn, Littleton, pp. 659-74.
Greer, W. B. and Daemen, J. J. K. (1991) Analyses and Field Tests of the Hydraulic
Performance of Cement Grout Borehole Seals. US Nuclear Regulatory
Commission Rep. NUREG/CR-5684, Washington, DC.
Greer, W. B. and Daemen, J. J. K. (1992) Hydraulic performance tests of a cement
grout borehole seal. SME Trans., 292, 1911-7.
Grim, R. E. (1953) Clay Mineralogy, McGraw-Hill, New York.
Grim, R. E. (1968) Clay Mineralogy, 2nd edn, McGraw-Hill, New York.
Grim, R. E. and Guven, N. (1978) Bentonites-Geology, Mineralogy, Properties and
Uses. Developments in Sedimentology, 24, Elsevier Scientific Publishing Com-
pany, Amsterdam.
Guessos, Z., Ladanyi, B. and Gill, D. E. (1988) Effect of sampling disturbance on
laboratory determined properties of rock salt. Proc. 2nd Conf on Mechanical
Behavior of Salt, Federal Institute for Geosciences and Natural Resources,
Federal Republic of Germany, pp. 137-58.
Gulick, C. W. Jr, Boa, J. A. Jr and Buck, A. D. (1980) Bell Canyon Test (BCT)
Cement Grout Development Report. Sandia National Lab. Rep. SAND80-1928,
Al buq uerq ue.
Gulick, C. W. Jr, Boa, J. A. Jr and Buck, A. D. (1982) Borehole Plugging Materials
Development Program-Report 3. Sandia National Lab. Rep. SAND81-0065,
Albuquerque.
Gupta, D. c., Nataraja, M. and Daemen, J. J. K. (1989) Regulatory questions and
concerns about sealing a HL W repository in an unsaturated environment. Proc.
Sealing of Radioactive Waste Repositories, Braunschweig, FRG, Workshop by
OECD NEA and CEC, Paris, pp. 201-12.
Hackney, R. M. (1985) A new approach to casing design for salt formations.
SPE/IADC Drilling Conf, New Orleans, pp. 79-89.
Haimson, B. C. and Doe, T. W. (1983) State of stress, permeability, and fractures in
the Precambrian granite of northern Illinois. J. Geophys. Res., 88, 7355-71.
Halliburton Services (undated) Cement Plug Studies, Houston, Texas.
Halliburton Services Company (1980) Halliburton Cementing Tables, Duncan,
Oklahoma.
Hambley, D. F., Fordham, C. J. and Senseny, P. E. (1989) General failure criteria for
saltrock. Proc. 30th US Rock Mech. Sym., Morgantown, pp. 91-8.
Hansen, F. D., Mellegard, K. D. and Senseny, P. E. (1984) Elasticity and strength
of ten natural rock salts. Proc. 1st Conf. on Mechanical Behavior of Salt,
Pennsylvania State Univ., pp. 71-83.
Hansen, F. D., Ahrens, E. A., Tidwell, V. C. and Tillerson, J. R. (1995) Dynamic
compaction of salt: initial demonstration and performance testing. Proc. 35th
Rock Mech. Sym., Lake Tahoe.
Hantush, M. S. (1959) Nonsteady flow to flowing wells in leaky aquifers. J. Geophys.
Res., 64, 1043- 52.
Harr, M. E. (1962) Groundwater and Seepage, McGraw-Hill, New York.
Hart, D. E. (1983) Current design and maintenance practice in salt mine shafts. Proc.
6th Int. Sym. on Salt, Vol. 1, Salt Institute, pp. 609-12.
Haug, M. D., Barbour, S. L. and Longval, P. (1988) Design and construction of a
308 Bibliography
prehydrated sand-bentonite liner to contain brine. Can. J. Civil Eng., 15,

955-63.
Hazen, A. (1982) Some physical properties of sand and gravel, with special reference
to their use in filtration. 24th Annual Report, Massachussets State Board of
Health, Boston, pp. 539-66.
Head, K. H. (1980) Manual of Soil Laboratory Testing, Volume 1: Soil Classification
and Compaction Tests, Engineering Laboratory Equipment Limited.
Healy, B. E. (1991) Fine tuning grouting jobs. Ground Water Age, 25, 13-6.
Heinmann, R. B., Stanchell, M. A. T. and Hooton, R. D. (1986) Short-Term Dissol-
ution Experiments on Various Cement Formulations in Standard Canadian Shield
Saline Solution in the Presence of Clay. Atomic Energy of Canada Ltd. Rep.
AECL-9059, Pinnawa.
Herndon, 1. and Smith, D. K. (1978) Setting downhole plugs: A state-of-the-art.
Petroleum Engineer, April, pp. 1-5.
Herubin, C. A. and Marotta, T. W. (1977) Basic Construction Materials, Reston
Publishing, Reston.
Hillel, D. (1980) Fundamentals of Soil Physics, Academic Press, New York.
Holcomb, D. J. (1988) Crosshole measurements of velocity and attenuation to detect
a disturbed rock zone in salt at the waste isolation pilot plant. Proc. 29th US
Rock Mech. Sym., Univ. of Minnesota, Minneapolis, pp. 633-40.
Holcomb, D. J. and Hannum, D. W. (1982) Consolidation of Crushed Salt Under
Conditions Appropriate to the WIPP Facility. Sandia National Lab. Rep.
SAND82-0630, Albuquerque.
Holcomb, D.1. and Shields, M. (1987) Hydrostatic Creep Consolidation of Crushed
Salt with Added Water. Sandia National Lab. Rep. SAND87-1990, Albuquerque.
Hollingshead, G. W. (1971) Stress distributions in rock anchors. Can. Geotech. J., 8,
588-92.
Holmberg, R., Larsson, B. and Sjoberg, C. (1984) Improved stability through opti-
mized rock blasting. Proc. 10th Con! Explosive and Blasting Tech., Orlando,
p.173.
Holmes, R. S., Kwan, L. T., Skinner, E. H. and Wong, E. Y. (1963) Loading, Re-
sponse, and Evaluation of Tunnels and Tunnel Liners in Granite, POR-1801,
Holmes & Narver, Inc., Los Angeles.
Holopainen, P. (1985) Crushed aggregate-bentonite mixtures as backfill material for
repositories of low- and intermediate-level radioactive wastes. Eng. Geol., 21,
239-45.
Holtz, R. D. and Kovacs, W. D. (1981) An Introduction to Geotechnical Engineering,
Prentice-Hall, Inc., Englewood Cliffs, New Jersey.
Horner, D. R. (1951) Pressure build-up in wells. Proc. Third World Petroleum Cong.,
The Hague, The Netherlands, pp. 503-20.
Hsieh, P. A. and Neuman, S. P. (1985a) Field determination of the three-dimensional
hydraulic conductivity tensor of anisotropic media 1. Theory. Water Resources
Res., 21, 1655-65.
Hsieh, P. A. and Neuman, S. P. (1985b) Field determination of the three-dimensional
hydraulic conductivity tensor of anisotropic media 2. Methodology and applica-
tion to fractured rocks. Water Resources Res., 21,1667-76.
Hsieh, P. A., Tracy, J. V., Neuzil, C. E. et al. (1981) A transient laboratory method
for determining the hydraulic properties of 'tight' rocks-I. Theory. Int. J. Rock
Mech. Min. Sci. & Geomech. Abstr., 18,245-52.
Hunsche, U. E. (1993) Failure behavior of rock salt around underground cavities.
7th Sym. on Salt, Vol. I, pp. 59-65.
Hvorslev, M. J. (1951) Time Lag and Soil Permeability in Ground-Water Observa-
tions. US Army Corps of Engineers Bull. No 36, Vicksburg.
Bibliography 309
Illinois Department of Mines and Minerals (1973) Sealing abandoned water well.
Water Well J., April.
Infante, E. F. and Chenevert, M. E. (1989) Stability of boreholes drilled through salt
formations displaying plastic behavior. SPE Drilling Engineering, 4, 57-65.
International Atomic Energy Agency (1984) Effects of Heat from High-Level Waste
on Performance of Deep Geological Repository Components. IAEA-TECDOC-
319, Vienna.
International Society for Rock Mechanics (1981) Committee on Laboratory Tests.
Suggested method for determining the uniaxial compressive strength and deforma-
bility of rock materials. pp. 111-6.
International Technology Corp. (1987) Laboratory Investigation of Crushed Salt
Consolidation and Fracture Healing. Office of Nuclear Waste Isolation Rep.
BMI/ONWI-631.
Jacob, C. E. (1950) Flow of groundwater, in Engineering Hydraulics, John Wiley,
New York, pp. 321-86.
Jacob, C. E. and Lohman, S. W. (1952) Nonsteady flow to a well of constant draw-
down in an extensive aquifer. Am. Geophys. Union Trans., 33, 559-69.
Jaeger, J. C. and Cook, N. G. W. (1979) Fundamentals of Rock Mechanics, 3rd edn,
Chapman & Hall, London.
Jakubick, A. T., Klein, R., Gray, M. N. and Keil, L. D. (1989) Characteristics of the
excavation response zone as applied to shaft sealing. Proc. NEA Workshop on
Excavation Response Zone in Geological Repositories for Radioactive Waste,
Winnipeg, pp. 157-74.
Jeffrey, R. G. (1981) Shaft or Borehole Plug-Rock Mechanical Interaction. MS
Thesis, Univ. of Arizona, Tucson.
Jeffrey, R. G. and Daemen, J. J. K. (1981) Analysis of the mechanical interaction
between a shaft or borehole plug and the surrounding rock. Proc. Waste Man-
agement '81 Sym., Tucson, Vol. 2, pp. 865-80.
Jepsen, C. P. (1984) Sodium bentonite: still a viable solution for hazardous waste
containment. Pollution Eng., April.
Jepsen, C. P. and Place, M. (1985) Evaluation of two methods for constructing
vertical cutoff walls at waste containment sites. Hydraulic Barriers in Soil and
Rock, ASTM STP 874, American Society for Testing and Materials, Philadel-
phia, pp. 45-63.
Johnson, A. I., Froebel, R. K., Cavalli, N. J. and Pettersson, C. B. (eds) (1984) Hy-
draulic Barriers in Soil and Rock. ASTM STP 874, American Society for Testing
and Materials, Philadelphia.
Jones, F. O. and Owens, W. W. (1980) A laboratory study of low-permeability gas
sands. J. Petroleum Technology, 32, 1631-40.
Kappei, G. (1986) Geotechnical Investigations on Backfill Materials in the Asse Salt
Mine. United States/Federal Republic of Germany Workshop on Sealing and
Backfilling of Salt Repository Rep. GSF T-250, Albuquerque.
Kelsall, P. C. and Nelson, J. W. (1983) Geologic and engineering characteristics of
Gulf region salt domes applied to underground storage and mining. Proc. 6th
Int. Sym. on Salt, Vol. 1, Salt Institute, pp. 519-43.
Kelsall, P. c., Case, J. B. and Chabannes, C. R. (1982) A Preliminary Evaluation of
the Rock Mass Disturbance Resulting from Shaft, Tunnel, or Borehole Excava-
tion. Office of Nuclear Waste Isolation Rep. ONWI-411, Columbus.
Kelsall, P. c., Case, J. B., Meyer, D. and Coons, W. E. (1982) Schematic Designs for
Repository Seals for a Reference Repository in Bedded Salt. Office of Nuclear
Waste Isolation Rep. ONWI-405, Columbus.
Kenney, T. c., Lau, D. and Ofoegbu, G. I. (1984) Permeability of compacted granu-
lar materials. Can. Geotech. J., 21, 726-9.
310 Bibliography
Kenney, T. c., van Ween, W. A., Swallow, M. A. and Sungalia, M. A. (1992) Hydrau-
lic conductivity of compacted bentonite-sand mixtures, Can. Geotech. J., 29,
364-74.
Kharaka, Y. K. and Smalley, W. C. (1976) Flow of water and solutes through com-
pacted clays. Am. Assoc. Petroleum Geol. Bull., 60, 973-80.
Kimbrell, A. F. and Daemen, J. J. K. (1986) Performance assessment of bentonite as
a borehole seal when tested in situ in a medium-grained granite. Proc. Waste
Management '86 Sym., Tucson, Vol. 2, pp. 257-63.
Kimbrell, A. F., Avery, T. S. and Daemen, J. J. K. (1987) Field Testing of Bentonite
and Cement Borehole Plugs in Granite. US Nuclear Regulatory Commission
Kinstler, F. L. (1983) Mackintosh and Murchison Dams-River Diversions. Civil
Eng. Trans., 25,51-6.
Kjartanson, B. H., Gray, M. N., Puchala, R. J. and Hawrylewicz, B. M. (1989a)
Excavating large-diameter boreholes and shafts in granite with high pressure
water jetting. Proc. IMM Shaft Engineering International Conference, Harrogate,
England.
Kjartanson, B. H., Hnatiw, D. S. J., Lucas, R. J. et al. (1989b) Buffer/Container
Experiment-Buffer Instrumentation Assessment Program. Whiteshell
Laboratories, Atomic Energy of Canada Limited Internal Rep.
Kjartanson, B. H., Chandler N. A., Wan, A. W. L. et al. (1992) Use of a method
specification for in situ compaction of clay-based barrier materials. Proc. 3rd
International High-Level Radioactive Waste Management Conference, Las Vegas.
Kjartanson, B. H., Radhakrishna, H. S., Lau, K. -c. et al. (1990) Geotechnical
instrumentation for in situ assessment of bentonite-sand buffer. Proc. Engineer-
ing in our Environment, the Annual Conference of the CSCE, Hamilton, Ontario,
II-2, pp. 670-87.
Klinkenberg, L. J. (1941) The permeability of porous media to liquids and gases.
Drilling and Production Practice-11th Mid- Year Meeting, Tulsa, pp. 200-13.
Konya, C. J., Otuonye, F. and Skidmore, D. (1982) Airblast reduction from effective
blast hole stemming. Proc. 8th Con[. on Explosives and Blasting Technique, New
Orleans, pp. 145-56.
Koplik, C. M., Pentz, D. L. and Talbot, R. (1979) Information Basefor Waste Reposi-
tory Design, Volume I-Borehole and Shaft Sealing. US Nuclear Regulatory
Commission Rep. NUREG/CR-0495, Washington, DC.
Kranz, R. L., Frankel, A. D., Engelder, T. and Scholz, C. H. (1979) The permeability
of whole and jointed Barre granite. Int. J. Rock Mech. Min. Sci., & Geomech.
Abstr., 16, 225-34.
Krumhansl, J. L. (1984) Observations Regarding the Stability of Bentonite Backfill in
a High-Level Waste (HLW) Repository in Rock Salt. Sandia National Lab. Rep.
Kruseman, G. P. and De Ridder, N. A. (1979) Analysis and Evaluation of Pumping
Test Data. International Institute for Land Reclamation and Improvement,
Bull. 11, The Netherlands.
Labreche, D. A. (1983) Damage mechanism in tunnels subjected to explosive loads.
Seismic Design of Embankments and Caverns, ASCE, New York, pp. 128-41.
Labreche, D. E. and Van Sambeek, L. L. (1988) Analysis of Data from Expansive
Salt- Water Concrete Seals in Series B Small-Scale Seal Performance Tests. Sandia
Lagerwerff, J. V., Nakayama, F. S. and Frere, M. H. (1969) Hydraulic conductivity
related to porosity and swelling of soil. Proc. Soil Sci. Soc. Am., 33, pp. 3-11.
Lamb, B. and Kenney, T. (1989) Decommissioning wells-techniques and pitfalls.
Proc.3rd National Outdoor Action Conf on Aquifer Restoration, Ground Water
Bibliography 311
Monitoring and Geophysical Methods. National Water Well Association, Dublin,

Ohio, pp. 217~28.
Lambe, T. W. (1951) Soil Testing for Engineers, John Wiley & Sons, New York.
Lambe, T. W. (1955) The permeability of fine-grained soils. ASTM STP 163,
American Society for Testing and Materials, Philadelphia, pp. 56~67.
Lambe, T. W. and Whitman, R. V. (1969) Soil Mechanics, John Wiley & Sons,
New York.
Lambert, S. J. (1980) A Strategy for Evaluating the Long-Term Stability of Hole-
Plugging Materials in Their Geologic Environments. Sandia National Lab. Rep.
SAND80-0359c, Albuquerque.
Lancaster, F. H. (1964) Report on Research into Underground Plugs. Transvaal and
Orange Free State Chamber of Mines Research Rep. No. 27/64.
Langefors, U. and Kihlstrom, B. (1978) The Modern Technique of Rock Blasting, 3rd
edn, John Wiley & Sons, New York.
Lau, K. -c. and Radhakrishna, H. S. (1992) Thermally Induced Stress Deformation
Analysis of Buffer for Borehole Emplacement Concept, Final Report to AECL
Research, Ontario Hydro Research Division, Toronto.
Lea, F. M. (1971) The Chemistry of Cement and Concrete, Chemical Publishing.
Lee, C. F. (1985) A Case History on Long-Term Effectiveness of Clay Sealant. Atomic
Energy of Canada Limited Rep. TR-338, Manitoba.
Leppert, D. (1986) Environmental uses of natural zeolites and bentonite. Proc. 22nd
Sym. on Engineering Geology and Soils Engineering, Boise, pp. 469~51O.
Lin, W. (1977) Compressible Fluid Flow Through Rocks of Variable Permeability.
Lawrence Livermore Lab. Rep. UCRL-52304, Livermore.
Lin, W. (1978) Measuring the Permeability of Eleana Argillite from Area 17, Nevada
Test Site, Using the Transient Method. Lawrence Livermore Lab. Rep. UCRL-
52604, Livermore.
Lingle, R., Stanford, K. L., Peterson, P. E. and Woodhead, S. F. (1982) Wellbore
Damage Zone Experimental Determination. Office of Nuclear Waste Isolation
Rep. ONWI-349, Columbus.
Long, J. C. S., Remer, J. S., Wilson, C. R. and Witherspoon, P. A. (1982) Porous
media equivalents for networks of discontinuous fractures, Water Resources
Res., 18, 645~58.
Loofbourow, R. L. (1973) Groundwater and groundwater control. SME Mining
Engineering Handbook, Soc. of Mining Engineers, New York, Vol. 2, pp. 26~55.
Loudon, A. G. (1952) The computation of permeability from simple soil tests.
Geotechnique, 3, 165~83.
Lutenegger, A. J. and DeGroot, D. J. (1993) Hydrologic properties of contaminant
transport barriers as borehole sealants, in Hydraulic Conductivity and Waste
Contaminant Transport in Soils. ASTM STP 1142, American Society for Testing
and Materials, Philadelphia, PA.
Malinowski, R. (1982) Ancient Mortars and Concretes~Durability Aspects, History of
Technology, 7th Annual Vol. (eds A. R. Hall and N. Smith) Mansell, London.
Manning, G. P. (1961) Reinforced Concrete Design, 2nd edn, Longmans, London.
Marine, I. W., Pratt, H. R. and Wahi, K. K. (1982) Seismic effects on underground
openings, in The Technology of High-Level Nuclear Waste Disposal, DOE/
TIC-4621, Vol. 2, pp.179~206.
Marinelli, F. (1984) Analysis of constant head injection tests in single, partially
penetrating boreholes. MS Thesis, Univ. of Arizona, Tucson.
Marshall, T. J. (1958) A relation between permeability and size distribution of pores.
J. Soil Sci., 9, 1~8.
312 Bibliography
Marsland, A. and Loudon, A. G. (1963) The flow properties and yield gradients of
bentonite grouts in sands and capillaries, in Grouts and Drilling Muds in Engin-
eering Practice, London, pp. 15-21.
Martin, J. (1989) Surface grouting of circular roadways. Mining Engineer, October,
161-215.
Martin, R. T. (1975) Feasibility of Sealing Boreholes with Compacted Natural Earthen
Materials. Oak Ridge National Lab. Rep. ORNLjSub-3950j2, Oak Ridge.
Mathis, S. P. (1982) Drilling induced damage to borehole wallrock: a theoretical,
laboratory, and field comparison of the effects of diamond and percussion
drilling. MS. Thesis, Univ. of Arizona, Tucson.
Mathis, S. P. and Daemen, J. J. K. (1982) Rock damage induced by drilling-An
experimental assessment of potential leakage around borehole seals, in Proc.
Waste Management '82 Sym., Tucson, Vol. 2, pp. 393-404.
McDaniel, E. W. (1980) Cement Technology for Borehole Plugging: An Interim Report
on Permeability of Cementatious Solids. Oak Ridge National Lab. Rep.
ORNLjTM-7092, Oak Ridge.
McGinty, J. E. and Calvert, D. G. (1975) Cementing off, plugging and redrilling.
Water Well J., July.
Mesri, G. and Olson, R. E. (1971) Mechanisms controlling the permeability of clays.
Clays and Clay Minerals, 19, 151-8.
Meyer, D. and Howard, J. J. (eds) (1983) Evaluation of Clays and Clay Minerals for
Application to Repository Sealing. Office of Nuclear Waste Isolation Rep.
ONWI-486, Columbus.
Michaels, A. S. and Lin, C. S. (1954) Permeability of kaolinite. Industrial Eng. Chem.,
46, 1239-46.
Miklowitz, J. (1978) The Theory of Elastic Waves and Waveguides, North Holland,
Amsterdam.
Millington, R. J. and Quirk, J. P. (1959) Permeability of porous media. Nature, 183,
387-8.
Mindess, S. and Young, J. F. (1981) Concrete. Prentice-Hall, Englewood Cliffs.
Mining Waste Study Team (1988) Mining Waste Study. Final Report Prepared by
the Mining Waste Study Team of the Univ. of California, Berkeley.
Mitchell, J. K. (1976) Fundamentals of Soil Behavior. John Wiley & Sons, New Yark.
Mitchell, J. K. (1993) Fundamentals of Soil Behavior, 2nd edn, John Wiley & Sons,
New York.
Mitchell, W. R. (1982) Examples of dewatering considerations and outlet arrange-
ments in H.E.C. dams, Tasmania, ANCOLD Bull., No. 61, May, pp. 26-82.
Moebs, N. N. and Krickovic, S. (1970) Air-Sealing Coal Mines to Reduce Water
Pollution. US Bureau of Mines Rep. 7354.
Moller, D. W., Minch, H., Welsh, J. P. and Rubsight, R. M. (1983) Pressure grouting
to control ground water in fractured granite rock, Helms pumped storage pro-
ject, in Innovative Cement Grouting. American Concrete Institute Pub. SP-83,
Detroit, pp. 129-51.
Monastersky, R. (1987) Heirs to ancient air. Science News, 132,172-3.
Monsees, J. E. and Merrit, J. L. (1984) Seismic design of underground struc-
tures-Southern California metro rail project, Lifeline Earthquake Engineering:
Performance, Design and Construction, ASCE, New York, pp. 52-66.
Montgomery, P. C. and Smith, D. K. (1961) Oil well cementing practices and
materials. Petroleum Engineer, May and June, 1-12.
Moore, J. G., Morgan, M. T., McDaniel, T. W. et al. (1979) Cement Technology for
Plugging Boreholes in Radioactive Waste Repository Sites. Oak Ridge National
Lab. Rep. ORNL-5524, Oak Ridge.
Bibliography 313
Morin, R. H. and Olsen, H. W. (1987) Theoretical analysis of the transient pressure

response from a constant flow rate hydraulic conductivity test. Water Resources
Res., 23, 1461-70.
Moye, D. G. (1967) Diamond drilling for foundation exploration. Civil Eng. Trans.,
The Institution of Engineers, Australia, pp. 95-100.
Munson, D. E. and Wawersik, W. R. (1992) Constitutive Modeling of Salt Behav-
ior-State of the Technology. Sandia National Lab. Rep. SAND91-1614C,
Albuquerque.
Munson, D. E., DeVries, K. L., Schiermeister, D. M. et al. (1991) Measured and
calculated closures of open and brine filled shafts and deep vertical boreholes in
salt, in Proc. 33rd US Rock Mech. Sym., Santa Fe, pp. 439-48.
Nagaraj, T. S. and Srinivasa Murthy, B. R. (1983) Rationalization of Skempton's
compressibility equation. Geotechnique, 33, 433-43.
Nagaraj, T. S. and Srinivasa Murthy, B. R. (1986) A critical reappraisal of compres-
sion index equations. Geotechnique, 36, 27-32.
Nair, K. and Singh, R. D. (1974) Creep rupture criteria for salt. 4th Sym. on Salt,
Vol. 2, Northern Ohio Geological Soc., Cleveland, pp. 41-9.
Narasimhan, T. N., Neuman, S. P. and Edwards, A. L. (1977) Mixed explicit-impli-
cit iterative finite element scheme for diffusion-type problems: II. Solution strat-
egy and examples. Int. J. Num. Meth. Eng., 11, 325-44.
Narasimhan, T. N., Neuman, S. P. and Witherspoon, P. A. (1978) Finite element
method for subsurface hydrology using a mixed explicit-implicit scheme. Water
Resources Res., 14, 863-77.
National Coal Board (1982) The Treatment of Disused Mine Shafts and Adits. Min-
ing Department, National Coal Board, London.
National Research Council (1985) Groundwater Contamination. National Academy
Press, Washington, DC.
Nazarian, H. N. (1973) Water well design for earthquake-induced motions. J. Power
Div., ASCE, 99,377-94.
Neuman, S. P. (1976) User's Guide for FRESURF 1. Dept. of Hydrology and Water
Resources, Univ. of Arizona, Tucson.
Neuman, S. P. and Narasimhan, T. N. (1977) Mixed explicit-implicit iterative finite
element scheme for diffusion-type problems: I. Theory. Int. J. Num. Meth. Eng.,
11,309-23.
Neuman, S. P. and Witherspoon, P. A. (1970) Finite element method of analyzing
steady seepage with a free surface. Water Resources Res., 6, 889-97.
Neuman, S. P., Preller, C. and Narasimhan, T. N. (1982) Adaptive explicit-implicit
quasi three-dimensional finite element model of flow and subsidence in multi-
aquifer systems. Water Resources Res., 18, 1551-61.
Neuzil, C. E. (1982) On conducting the modified 'slug' test in tight formations. Water
Neuzil, C. E. (1986) Groundwater flow in low-permeability environments. Water
Neuzil, C. E., Cooley, c., Silliman, S. E. et al. (1981) A transient laboratory method
for determining the hydraulic properties of 'tight' rocks-II. Application. Int. J.
Rock Mech. Min. Sci. & Geomech. Abstr., 18,253-8.
Neville, A. M. (1981) Properties of Concrete, 3rd edn, Pitman, London.
Nielsen, D. M. (1991) Practical Handbook of Groundwater Monitoring, Lewis Pub-
lishers, Michigan.
Nielsen, D. M. (1993) Correct well design improves monitoring. Environmental
Protection, 4, 38-49.
Nilsson, J. (1985) Field compaction of bentonite-based backfilling. Eng. Geol., 21,
367-76.
314 Bibliography
Nowak, E. J. and Stormont, J. C. (1987) Scoping Model Calculations of the Reconsoli-

dation of Crushed Salt in WIPP Shafts. Sandia National Lab. Rep. SAND87-
0879, Albuquerque.
Nowak, E. J. and McTigue, D. F. (1987) Interim Results of Brine Transport Studies in
the Waste Isolation Pilot Plant (WIPP). Sandia National Lab. Rep., SAND87-
0880, Albuquerque.
Nowak, E. J., Tillerson, J. R. and Torres, T. M. (1990) Initial Reference Seal System
Design. Sandia National Lab. Rep. SAND90-0355, Albuquerque.
Nowak, E. J., Wakeley, L. D., Poole, T. S. and Burkes, J. P. (1992) Investigations of
Deteriorated Concrete from the Liner of the Waste Isolation Pilot Plant Waste
Shaft. Sandia National Lab. Rep. SAND91-2733, Albuquerque.
NWWA (1975) Manual of Water Well Construction Practices. EPA-570/9-75-001,
National Water Well Association Committee on Water Well Standards, Office
of Water Supply, Washington, DC.
Nye, J. D. (1987) Abandoned wells: how one state deals with them. Water Well J.
September.
Odom, E. E. (1986) Na/Ca montmorillonite: properties and uses. Soc. Mining Engin-
eers of AIME, 282, 1893~901.
O'Driscoll' M. (1988) Bentonite: overcapacity in need of markets. Industrial Min-
erals, No. 250, pp. 43~63.
Ogden, F. L. and Ruff, J. F. (1989) Axial Shear Strength Testing of Bentonite Water
Well Seals. Rep. Department of Civil Engineering, Colorado State Univ.
Olsen, H. W. (1962) Hydraulic flow through saturated clays, in Proc. 9th National
Con! on Clays and Clay Minerals, Vol. 9, Pergamon Press, New York,
pp. 131~61.
Olsen, H. W. (1966) Darcy's law in saturated kaolinite. Water Resources Res., 2,
287~95.
Olsen, H. W. (1969) Simultaneous fluxes of liquid and charge in saturated kaolinite.
Soil Sci. Soc. Amer. Proc., 33, 338~44.
Olsen, H. W., Nichols, R. W. and Rice, T. L. (1985) Low gradient permeability
measurements in a triaxial system. Geotechnique, 35, 145~57.
Olson, R. E. and Daniel D. E. (1981) Measurement of the hydraulic conductivity of
fine-grained soils. Permeability and Groundwater Contaminant Transport, ASTM
STP 746, American Society for Testing and Materials, Philadelphia, pp. 18~64.
Organization for Economic Cooperation and Development, Nuclear Energy Agency
(1982) Geological Disposal of Radioactive Waste Research the aECD Area,
OECD, Paris.
Oriard, L. L. (1982) Blasting effects and their control, in Underground Mining
Methods Handbook SME/AIME, New York, pp. 1590~603.
O'Rourke, M. and Wang, L. R. L. (1978) Earthquake response of buried pipeline, in
Earthquake Engineering and Soil Dynamics, Proc. ASCE Geotech. Eng. Div. Spec.
Con!, Pasadena, Vol. 2, pp. nO~31.
Otounye, F., Konya C. J. and Skidmore, D. R. (1983) Effects of stemming size dis-
tribution on explosive charge confinement: A laboratory study, Mining Engin-
eering, 35, 1205~8.
Ouyang, S. and Daemen, J. J. K. (1989) Crushed Salt Consolidation. US Nuclear
Regulatory Commission Rep. NUREG/CR-5402, Washington, DC.
Ouyang, S. and Daemen, J. J. K. (1990) Performance of bentonite/crushed tuff back-
fill for nuclear waste repositories, in Proc. Waste Management '90 Sym., Tucson,
Vol. 2, pp. 605~ 11.
Ouyang, S. and Daemen, J. J. K. (1991a) Design of bentonite/crushed rock seals.
Geotech. Geol. Eng., 9, 63~ n.
Ouyang, S. and Daemen, J. J. K. (1991b) Yield stress and flow of bentonite in
Bibliography 315
mixture sealants. Geotechnical Engineering Congress, Boulder, Vol. 2,
pp. 1244-55.
Ouyang, S. and Daemen, J. J. K. (1992) Sealing Performance of Bentonite and Benton-
ite/Crushed Rock Borehole Plugs. US Nuclear Regulatory Commission Rep.
NUREG/CR-5685 Washington, DC.
Ouyang, S., Su, Y. -H. and Daemen, J. 1. K. (1991) Bentonite/crushed rock seals for
hazardous waste containment. Proc. 6th Can! on Waste Management Technology
in Republic of China, pp. 191-9.
Owen, G. N. and Scholl, R. E. (1981) Earthquake Engineering of Large Underground
Structures, Federal Highway Administration and National Science Foundation,
Rep. FHWA/RD-80/195, NTIS, Springfield, VA.
Oyler, D. C. (1984) Use of Sodium Silicate Gel Grout for Plugging Horizontal Meth-
ane-Drainage Holes. US Bureau of Mines Rep. 8843.
Pandian, N. S. and Nagaraj, T. S. (1990) Critical reappraisal of colloidal activity of
clays. ASCE J. Geotech. Eng., 116, 285-96.
Pao, Y. H. and Mow, C. C. (1973) The Diffraction of Elastic Waves and Dynamic
Stress Concentrations, Crane and Russak, New York.
Papadopulos, S. S., Bredehoeft, J. D. and Cooper, H. H. (1973) On the analysis of
'slug test' data. Water Resources Res., 9, 1087-9.
Paterson, M. S. (1983) The equivalent channel model for permeability and resistivity
in fluid-saturated rock-a re-appraisal. Mechanics of Materials, 2, 345-52.
Peach, C. J., Spiers, C. J., Tankink, A. J. and Zwart, H. J. (1987) Fluid and Ionic
Transport Properties of Deformed Salt Rock. Commission of the European
Communities Rep. EUR10926, Luxembourg.
Penman, A. D. M., and Charles, J. A. (1979) The influence of their interfaces on the
behavior of clay cores in embankment dams. Int. Congr. on Large Dams, 13th
Trans., Vol. 1, New Delhi, pp. 695-714.
Peterson, E. W. (1987) Fluid Flow Measurements of Test Series A and B of the Small
Scale Seal Performance Test. Sandia National Lab. Rep. SAND87-7041,
Albuquerque.
Peterson, E. W. and Kelkar, S. (1983) Laboratory Tests to Determine Hydraulic and
Thermal Properties of Bentonite-Baclifill Materials. Sandia National Lab. Rep.
Peterson, E. W., Lagus, P. L. and Lie, K. (1987) WIPP Horizon Free Field
Fluid Transport Characteristics. Sandia National Lab. Rep. SAND87-7164,
Albuquerque.
Pettman, E. R. (1984) Tunnel plugs in recent H. E. C. practice, 5th Australian Tunnel-
ing Corif-, Sydney, pp. 207-12.
Pfeifle, T. W. (1991) Consolidation, Permeability, and Strength of Crushed Salt/Ben-
tonite Mixtures with Application to the WIPP. Sandia National Lab. Rep.
Pfeifle, T. W. and Brodsky, N. S. (1990) Swelling Pressure, Water Uptake and Permea-
bility of 70/30 Crushed Salt/Bentonite. ReSpec. Inc. Rep. RS 1-0378, Rapid City.
Pfeifle, T. W., Mellegard, K. D. and Senseny, P. E. (1983) Constitutive Properties
of Salt from Four Sites. Office of Nuclear Waste Isolation Rep. ONWI-314,
Columbus.
Philip, J. and Daemen, J. J. K. (1988) High-level waste (HLW) repository sealing.
Proc. Int. Sym. Underground Engineering, New Delhi, Vol. 1, pp. 497-508.
Philip, 1. and Daemen, J. J. K. (1990) Initial results of tuff borehole sealing experi-
ments. Proc. Waste Management '90 Sym., Tucson, Vol. 2, pp.643--49.
Picking, L. W. and Wilton, D. E. (1985) Testing the hydraulic characteristics of low
permeability carbonates, Palo Duro, Texas. Proc. 17th Int. Congr. Int. Assoc.
Hydrogeologists. Tucson, pp. 30-43.
316 Bibliography
Piper, A. M. (1969) Disposal of Liquid Wastes by Injection Underground-Neither

Myth nor Millenium. US Geological Survey Cir. 631.
Powers, T. C. (1958) Structure and physical properties of hardened portland cement
paste. J. Am. Ceramic Soc., 41, 38-48.
Pratt, H. R, Stephenson, D. E., Zandt, G. et al. (1979) Earthquake damage to under-
ground facilities, in Proc. 1979 Rapid Excavation and Tunneling Conf, Atlanta,
Vol. 1, pp. 19-51.
Puchala, R. J., Hawrylewicz, B. M., Kjartanson, B. H. and Gray, M. N. (1989) Devel-
opment of water jetting equipment for excavating large-diameter boreholes in
granite, in Proc. 5th American Water Jet Conf, National Research Council of
Canada and the US Water Jet Technology Association, pp. 27-38.
Pusch, R. (1980) Swelling Pressure of Highly Compacted Bentonite. KBS Project
15:05 Rep. 80-13. Univ. of Lulea.
Pusch, R. (1983) Borehole sealing for underground waste storage. J. Geotech.
Eng.-ASCE, 109,113-9.
Pusch, R (1987) Highly compacted Na Bentonite as Buffer Substance. KBS Rep. 74,
Lulea.
Pusch, R (1988) Rock Sealing-Large Scale Field test and Accessory Investigation.
SKB Rep. 88-04, Stockholm.
Pusch, Rand Alstermark, G. (1985) Experience from preparation and application of
till/bentonite mixtures. Eng. Geol., 21, 377- 82.
Pusch, R. and Borgensson, L. (1989) Bentonite sealing of rock excavations. Sealing
of Radioactive Waste Repositories. Braunschweig, OECD Nuclear Energy
Agency and Commission of the European Communities.
Pusch, R, Borgensson, L. and Ramquist, G. (1987a) Final Report of the Borehole,
Shaft, and Tunnel Sealing Test- Volume II: Shaft Plugging. Stripa Project
Rep. 87-02.
Pusch, R., Edstrom, M. and Borgensson, L. (1987b) Piping and Erosion Phenomena
in Soft Clay Gels. SKB Technical Rep. 87-09, Stockholm.
Pusch, R., Jacobsson, A. and Bergstrom, A. (1980) Bentonite-based buffer substances
for isolating radioactive waste products at great depths in rock. Proc. Under-
ground Disposal of Radioactive Wastes, Vol. 1, International Atomic Energy
Agency, Vienna.
Pusch, R Borgensson, L., Fredrikson, A. etal. (1988) Rock Sealing-Interim Report
on the Rock Sealing Project (Stage 1). SKB Rep. 88-11, Stockholm.
Quirk, J. P. (1968) Particle interactions and soil swelling. Israel J. of Chem., 6,
213-34.
Radhakrishna, H. S. and Chan, H. T. (1985) Strength and Hydraulic Conductivity of
Clay-Based Buffersfor a Deep Underground Nuclear Fuel Waste Disposal Vault.
Atomic Energy of Canada Limited Rep. TR-327, Pinawa.
Radhakrishna, H. S. and Lau K. -C. (1992) Analysis of Buffer/Container Test I-Re-
vision 2, Final Report to AECL Research, Ontario Hydro Research Division,
Toronto.
Radhakrishna, H. S., Chan, H. T., Crawford, A. M. and Lau, K. -C. (1989) Thermal
and physical properties of candidate buffer-backfill materials for a nuclear fuel
waste disposal vault. Can. Geotech. J., 26, 629-39.
Radhakrishna, H. S., Crawford, A. M., Kjartanson, B. H., and Lau, K. -co (1992)
Numerical modelling of heat and moisture transport through bentonite-sand
buffer. Can. Geotech. J., 29, 1044-59.
Randolph, M. F. and Wroth, C. P. (1978) Analysis of deformation of vertically
loaded piles. ASCE J. Geotech. Eng. Div., 104, 1465-88.
Rennick, G. E., Pasini, J. III, Armstrong, F. E. and Abrams, J. R. (1972) Demonstra-
tion of Safety Plugging of Oil Wells Penetrating Appalachian Coal Mines. US
Bureau of Mines Technical Progress Rep., July.
Bibliography 317
Resendiz, D. (1976) Relevance of Atterberg limits in evaluating piping and breaching

potential. Dispersive Clays, Related Piping and Erosion in Geotechnical Projects,
79th Annual Meeting of ASTM, ASTM-STP-632, pp. 341-53.
Rhoderick, J. E. and Buck, A. D. (1981a) Examinations of Simulated Borehole Speci-
mens. US Army Engineers Waterways Experiment Station, SL-81-7, Vicksburg.
Rhoderick, J. E. and Buck, A. D. (1981b) Petrographic Examination of Several Four-
Year-Old Laboratory Developed Grout Mixtures. Office of Nuclear Waste Isola-
tion Rep. ONWI-342, Columbus.
Roscoe Moss Company (1990) Handbook of Ground Water Development, Wiley-
Interscience, New York, NY.
Rosenquist, I. T. (1955) Investigations in the Clay- Electrolyte- Water System.
Norwegian Geotechnical Institute Pub. No.9, Oslo.
Rosewell, C. J. (1977) Identification of susceptible soils and control of tunnelling
failure in small earth dams. Dispersive Clays, Related Piping, and Erosion in
Geotechnical Projects, ASTM STP 623, American Society for Testing and
Materials, Philadelphia, pp. 362-9.
Ross-Brown, D. M., Trent, B. C. and Wahi, K. K. (1981) Numerical simulations
of seismic loading of deep underground tunnels. SME-AIME Fall Meeting,
Denver, CO, Preprint No. 81-408.
Roy, D. M. and Grutzeck, M. W. (1980) PSUjWES Interlaboratory Comparative
Methodology Study of an Experimental Cementitious Repository Seal Material.
Roy, D. M. and Langton, C. A. (1982) Longevity of Borehole and Shaft Sealing
Materials: Characterization of Cement-Based Ancient Building Materials. Office
of Nuclear Waste Isolation Rep. ONWI-202, Columbus.
Roy, D. M., Grutzeck, M. W. and Wakeley, L. D. (1983) Selection and Durability of
Seal Materials for a Bedded Salt Repository: Preliminary Studies. Office of
Nuclear Waste Isolation Rep. ONWI-479, Columbus.
Roy, D. M., Grutzeck, M. W. and Wakeley, L. D. (1985) Salt Repository Seal
Materials: A Synopsis of Early Cementitious Materials Development. Office of
Nuclear Waste Isolation Rep. BMIjONWI-536, Columbus.
Roy, D. M., Brindley, G. W., Biggers, U. V. et al. (1977) Borehole Cement and Rock
Properties Studies. Office of Waste Isolation Rep. ONWIjSUB-4091j6,
Columbus.
Saulnier, G. J. and Avis, J. D. (1988) Interpretation of Hydraulic Tests Conducted in
the Waste-Handling Shaft at the Waste Isolation Pilot Plant (WIPP) Site. Sandia
Sawyer, W. D. II and Daemen, J. J. K. (1987) Experimental Assessment of the Sealing
Performance of Bentonite Borehole Plugs. US Nuclear Regulatory Commission
Rep. NUREGjCR-4995, Washington, DC.
Scheetz, B. E., Licastro, P. H. and Roy, D. M. (1986) A Full-Scale Borehole Sealing
Test in Salt Under Simulated Downhole Conditions, Vol. 2. Office of Nuclear
Waste Isolation Rep. BMIjONWI-573(2), Columbus.
Schmid, W. E. (1957) The permeability of soils and the concept of a stationary
boundary-layer. Proc. Am. Soc. Testing and Materials, 57, 1195-218.
Schofield, A. N. (1981) Dynamic and earthquake geotechnical centrifuge modeling.
Proc. Int. Corif'. Recent Advances in Geotech. Earthquake Eng. and Soil Dynamics,
Rolla, MO, Vol. 3, pp. 1081-100.
Scott Blair, G. W. and Crowther, E. M. (1929) The flow of clay pastes through
narrow tubes. J. Phys. Chern., 33, 321-30.
Scully, L. W. and Rothman, A. J. (1982) Repository and Engineering Barriers Design.
US Department of Energy Rep. DOEjNWTS-30.
318 Bibliography
Seaber, P. R. and Vecchioli, J. (1966) Use of soil-consolidation test data to determine

permeability of clays. US Geological Survey Water Supply Pap. 1822, pp. 105-12.
Selvadurai, A. P. S. (1991) Hydrothermal behaviour of a clay barrier developed for use
in a nuclear waste disposal vault-Characterization of hydrothermal behaviour,
Department of Civil Engineering Report, Carleton University, Ottawa.
Shainberg, I., Bresler, E. and Klausner, Y. (1971) Studies on NajCa montmorillonite
systems: 1. The swelling pressure. Soil Sci., 111, 214-9.
Sherard, J. L., Decker, R. S. and Ryker, N. L. (1972) Piping in earth dams of disper-
sive clay. Proc. ASCE Special Conference on Performance of Earth and Earth-
Supported Structures, Vol. 1, pp. 589-626.
Shor, A. J., Baes, C. F. and Canonico, C. M. (1981) Consolidation and Permeability of
Salt in Brine. Oak Ridge National Lab. Rep. ORNL-5774, Oak Ridge.
Simi, N. and Harsulescu, A. (1979) The use of bentonites for sealing earth dams. Bull.
Int. Assoc. Eng. Geol., 20, 222-6.
Simmons, G. R. (1990) Operating phase experiments planned for Atomic Energy of
Canada Limited's Underground Research Laboratory. Proc. Int. Sym. on
Unique Underground Structures, Denver, Vol. 2, pp. 6711-9.
Simmons, G. R. and Baumgartner, P. (1994) The Disposal of Canada's Nuclear Fuel
Waste: Engineering for a Disposal Facility. Atomic Energy of Canada Limited
Rep. AECL-10715, Pinawa.
Singh, S. K. (1982) Chemical, physical, and engineering characterization of candidate
backfill clays and clay admixtures for a nuclear waste repository- Part I. Scien-
tific Basis for Nuclear Waste Management, Materials Research Society Symposia
Proceedings, Boston, Vol. 6, pp. 413-33.
Singhal, R. K. (1974) Determination of stress in evaporites. Can. Mining J., August,
pp.35-47.
Sitz, P. (1981) Querschnittsabdichtungen untertagiger Hohlraume durch Damme und
Pfropfen (Sealing the cross section of underground excavations with dams or
plugs), Freiberger Forschungshefte A643, VEB Deutscher Verlag fur Grundstoff-
industrie, Leipzig.
Sitz, P. and Forster, W. (1989) Design of sealing systems for shafts, boreholes and
access tunnels-basis for gas storage in abandoned mines and rock caverns.
Proc. Int. Con! on Storage of Gases in Rock Caverns, Trondheim, pp. 193-200.
Sitz, P., Koeckritz, V. and Oellers, T. (1989) Shaft sealing for nuclear waste reposito-
ries, Institution of Mining and Metallurgy Shaft Engineering Con!, Harrogate,
England.
Sjaardema, G. D. and Krieg, R. D. (1987) A Constitutive Modelfor the Consolidation
of WIPP Crushed Salt and Its Use in Analyses of Backfilled Shaft and Drift
Configurations. Sandia National Lab. Rep. SAND87-1977, Albuquerque.
Skempton, A. W. (1944) Notes on the compressibility of clays. Q. J. Geol. Soc.
London, 100C, 119-35.
Smiles, D. E. and Rosenthal, M. J. (1968) The movement of water in swelling ma-
terials. Aust. J. Soil Research, 6, 237-48.
Smith, D. K. (1976) Cementing, Society of Petroleum Engineers of AIME, New
York.
Smith, D. K. (1986) Primary cementing, in Drilling Practices Manual, Penn Well
Publishing, Tulsa, pp. 400-60.
Smith, D. K. (1990) Cementing, Society of Petroleum Engineers of AIME, Dallas.
Smith, D. K. (1992) Cementing, 2nd edn, Society of Petroleum Engineers of AIME,
Richardson.
Smith, G. E., Smith, W. R., Littleton, D. J. and Simmons, J. (1993) Recent improve-
ments in state regulatory programs and compliance practices. 1st SPEjEPA
Joint Con!, San Antonio, pp. 51-60.
Bibliography 319
Smith, I. M. (1977) Numerical and physical modeling, in Numerical Methods in
Geotechnical Engineering, McGraw-Hill, New York, pp. 556-88.
Smith, 1. B. and Browning, L. A. (1993) Proposed changes to EPA class II well
construction standards and area ofreview procedures. 1st SPE/EPAJoint Conf.,
San Antonio, pp. 269- 77.
Smith, M. 1., Anttonen, G.l., Barney, G. S. et. al. (1980) Engineered Barrier Develop-
ment for a Nuclear Waste Repository in Basalt: An Integration of Current
Knowledge. Rockwell Hanford Operations Rep. RHO-BWI-ST-7, Richland,
Washington.
Smith, S. A. (1994) Well & Borehole Sealing: Importance, Materials, Methods, and
Recommendationsfor Decommissioning. Ground Water Publishing Co., Dublin.
Smith, S. A. and Mason, C. (1985) Bentonite grouting. Water Well J., 39, 35-7.
Snow, D. T. (1968) Rock fracture spacings, openings, and porosities. J. Soil Mech.
Found. Div. ASCE, 94, 73-91.
Snow, D. T. (1969) Anisotropic permeability of fractured media. Water Resources
Res., 5, 1273-89.
South, D. L. and Daemen, 1.1. K. (1981) Laboratory experiments of borehole plug
performance. Proc. 22nd US Rock Mechanics Sym., Boston, pp. 116-20.
South, D. L. and Daemen, J. 1. K. (1982) Laboratory studies of fluid flow through
borehole seals. Proc, 6th Int. Sym. Scientific Basis for Nuclear Waste Manage-
ment, Boston, Vol. 15, pp. 703-10.
South, D. L. and Daemen, 1.1. K. (1986) Permeameter Studies of Water Flow Through
Cement and Clay Borehole Seals in Granite, Basalt and Tuff. US Nuclear Regula-
tory Commission, Tech. Rep. NUREG/CR-4748, Washington, DC.
South, D. L. Cobb, S. L. and Daemen, 1.1. K. (1981) Laboratory experiments of
borehole plug performance. Proc. Waste Management '81 Sym., Tucson, Vol. 2,
pp.881-94.
Sowers, G. F. (1976) Dewatering rock for construction. Proc. Rock Eng. Foundations
and Slopes, Geotechnical Engineering Div. ASCE, Boulder, Vol. 1, pp. 200-15.
Spane, F. A. lr and Thorne, P. D. (1985) Effects of drilling fluid invasion on hydrau-
lic characterization of low permeability basalt horizons: a field evaluation. Proc.
7th Int. Congo Int. Assoc. Hydrogeologists, Tucson, pp. 30-43.
Spiers. C. 1., Urai, 1. L. and Lister, G. S. (1988) The effect of brine (inherent or added)
on rheology and deformation mechanisms in salt rock. Proc. 2nd Con! on
Mechanical Behavior of Salt, Federal Republic of Germany, pp. 89-102.
Stephens, D. B. (1979) Analysis of constant head borehole infiltration tests in the
vadose zone. PhD Dissertation, Univ. of Arizona, Tucson.
Stetzenbach, K. 1. and Thompson, G. M. (1983) A new method for simultaneous
measurement of CI, Br, SCN, and I at sub-ppm levels in groundwater. Ground
Water, 21, 36-41.
Stevens, P. R. (1977) A Review of the Effects of Earthquakes on Underground Mines.
US Dept. of the Interior, Geological Survey, Open-File Rep. 77-313, Reston.
Stormont, 1. C. (1983) Mechanical Strength of Borehole Plug. MS Thesis, Univ. of
Arizona, Tucson.
Stormont, 1. C. (1986) Development and Implementation: Test Series A of the Small-
Scale Seal Performance Test. Sandia National Lab. Rep. SAND85-2602,
Albuquerque.
Stormont, 1. C. (1987) Small-Scale Seal Performance Test Series A-Thermal/Struc-
tural Data Through the 180th Day. Sandia National Lab. Rep. SAND87-0178,
Albuquerque.
Stormont, 1. C. (1988) Preliminary Seal Design Evaluation for the Waste Isolation
Pilot Plant. Sandia National Lab. Rep. SAND87-3083, Albuquerque.
Stormont, 1. C. (1990a) Discontinuous behaviour near excavations in a bedded salt
formation. Int. J. Mining Geol. Eng., 8, 35-56.
320 Bibliography
Stormont, J. C. (1990b) Summary of 1988 WIPP Facility Horizon Gas Flow Measure-
ments. Sandia National Lab. Rep. SAND89-2497, Albuquerque.
Stormont, J. C. (1990c) Gas Permeability Changes in Rock Salt During Deforma-
tion. PhD Dissertation, Univ. of Arizona, Tucson.
Stormont, J. C. and Arguello, J. G. (1988) Model Calculations of Flow Through Shaft
Seals in the Rustler Formation. Sandia National Lab. Rep. SAND87-2859,
Albuquerque.
Stormont, J. C. and Daemen, J. J. K. (1983) Push-out tests on borehole plugs. Proc.
Waste Management '83 Sym., Tucson, Vol. 2, pp. 265-70.
Stormont, J. C. and Daemen, J.J.K. (1992) Laboratory study of gas permeability
changes in rock salt during deformation. Int. J. Rock Mech. Min. Sci & Geomech.
Abstr., 29, 325-42.
Stormont, J. C. and Fuenkajorn, K. (1994) Dilation-induced permeability changes in
rock salt. Int. Association for Computer Methods and Advances in Geomech.,
Morgantown pp. 1269-73.
Stormont, J. C. and Howard, C. L. (1986) Development and Implementation: Test
Series B of the Small-Scale Seal Performance Tests. Sandia National Lab. Rep.
Stormont, J. C. and Howard, C. L. (1987) Development, Implementation and Early
Results: Test Series C of the Small-Scale Performance Tests. Sandia National
Lab. Rep. SAND87-2203, Albuquerque.
Stormont, J. C., Daemen, J. J. K. and Desai C. S. (1992) Prediction of dilation and
permeability changes in rock salt. Int. J. Num. Anal. Meth. Geomech., 16,
545-569.
Stormont, J. c., Howard, C. L. and Daemen, J. J. K. (1991) Changes in rock salt
permeability due to nearby excavation. Proc. 32nd US Rock Mech. Sym., Univ.
of Oklahoma, pp. 899-907.
Stormont, J. c., Peterson, E. W. and Lagus, P. L. (1987) Summary of and Observa-
tions about WIPP Facility Horizon Flow Measurements through 1986. Sandia
Strobel, G. S. (1993) Impedance-Computed Tomography for Monitoring Moisture
Content in Clay Based Buffer. Atomic Energy of Canada Limited Tech. Rec.
TR-586, Pinawa.
Stroup, D. E. and Senseny, P. E. (1987) Influence of Bentonite Content on Consolida-
tion and Permeability of Crushed Salt from the WIPP. Sandia National Lab. Rep.
RSI-0309, Albuquerque.
Suman, G. O. and Ellis, R. C. (1977a) Cementing oil and gas wells-part 1. World
Oil, 184, 43-51.
Suman, G. O. and Ellis, R. C. (1977b) Cementing oil and gas wells -part 2. World
Oil, 184, 69-76.
Suman, G. O. and Ellis, R. C. (1977c) Cementing oil and gas wells-part 3. World
Oil; 184, 48-57.
Suman, G. O. and Ellis, R. C. (1977d) Cementing oil and gas wells-part 4. World
Oil, 184, 69-77.
Suman, G. O. and Ellis, R. C. (1977e) Cementing oil and gas wells-part 5. World
Oil, 185, 117-25.
Suman, G. O. and Ellis, R. C. (1977f) Cementing oil and gas wells-part 6. World
Oil, 185, 43-50.
Suman, G. O. and Ellis, R. C. (1977g) Cementing oil and gas wells-part 7. World
Oil, 185, 87-95.
Suman, G. O. and Ellis, R. C. (1977h) Cementing oil and gas wells-part 8. World
Oil, 185, 79-92.
Suro, R. (1992) The Toxic Legacy of Dead Oil Wells, New York Times, May 31.
Bibliography 321
Sutherland, H. J. and Cave, S. P. (1980) Argon-gas permeability of New Mexico rock

salt under hydrostatic compression. Int. J. Rock Mech. Min. Sci. & Geomech.
Abstr., 17, 281-8.
Tavenas, F., Jean, P., Leblond, P. and Leroueil, S. (1983a) The permeability of
natural soft clays. Part II: permeability characteristics. Can, Geotech. J., 20,
645-60.
Tavenas, F., Leblond, P., Jean, P. and Leroueil, S. (1983b) The permeability of
natural soft clays, I, methods of laboratory measurement. Can, Geotech. J., 20,
629-44.
Taylor, C. L., Anttonen, G. J., O'Rourke, J. E. and Allirot, D. (1980) Preliminary
Geochemical and Physical Testing of Materials for Plugging of Man-Made
Accesses to a Repository in Basalt. US Department of Energy Rep. RHO-
BWI-C-66, Washington, DC.
Taylor, C. L., O'Rourke, J. E., Allirot, D. and O'Connor, K. (1980) Preconceptual
Systems and Equipment for Plugging of Man-Made Accesses to a Repository in
Basalt. Rockwell Hanford Operations Rep. RHO-BWI-C-67, Richland.
Taylor, D. W. (1984) Fundamentals of Soil Mechanics, John Wiley & Sons,
New York.
Theis, C. V. (1935) The relation between the lowering of the piezometric surface and
the rate and duration of discharge of a well using groundwater storage. Am.
Geophys. Union Trans., 16, 519-24.
Thomas, H. R., Rees, S. W., Kjartanson, B. H. et al. (1993) Modelling in situ water
uptake in a bentonite-sand barrier-preliminary results, in Proc., Engineering
Geology of Waste Storage and Disposal, University of Wales, Cardiff.
Thompson, H. P. (1988) Review and Comment on the US Department of Energy Site
Characterization Plan Conceptual Design Report. Nuclear Waste Project Office
Rep. NWPO-TR-009-88, Nevada.
Thomson, I., Bodimeade, c., Kapteyn, P. and Pianosi, J. (1994) Placement of ben-
tonite seals in boreholes with multiple piezometers using an injector device.
Geotechnical News, 12, 79-83.
Tillerson, J. R., Fernandez, J. A. and Hinkebein, T. E. (1989) Uncertainties in sealing
a nuclear waste repository in partially saturated tuff. Proc. NEA/CEC Workshop
on Sealing of Radioactive Waste Repositories, Paris, pp. 174-86.
Timoshenko, S. P. and Gere, J. M. (1961) Theory of Elastic Stability, 2nd edn,
McGraw-Hill Kogakusha Ltd, Tokyo.
Todd, D. M., (1980) Groundwater Hydrology, John Wiley & Sons, New York
Torres, T. M., Howard, C. L. and Finley, R. E. (1992) Development, Implementation,
and Early Results: Test Series D, Phase I of The Small-Scale Seal Performance
Tests. Sandia National Lab. Rep. SAND91-2001, Albuquerque.
TRB (1991) Transportation Research Board Special Summer Meeting on Geotech-
nical Exploratory Holes. Albany, New York, May 28.
Trimmer, D. (1982) Laboratory measurements of ultralow permeability of geologic
materials. Review of Scientific Instrumentation, 53, 1246-54.
Trimmer, D., Bonner, B., Heard, H. C. and Duba, A. (1980) Effect of pressure and
stress on water transport in intact and fractured gabbro and granite, J. Geophys.
Res., 85, 7059- 71.
Unger, K. W. and Howard, D. C. (1986) Drilling techniques improve success in
drilling and casing deep overthrust belt salt. AIME, SPEDE Trans., 1, 183-92.
US Bureau of Mines (1994) Mine Closure Guidelines. Technology News, No. 435, US
Department of Interior, Washington, DC.
US Bureau of Public Roads (1962) Aggregate Gradationfor Highways. US Bureau of
Public Roads, Washington, DC.
322 Bibliography
US Department of Defense (1971) Instrumentation of Earth and Rock-Fill Dams

(Groundwater and Pore Pressure Observations), Part 1, Engineer Manual
1110-2-1908.
US Department of Energy (1988) Site Characterization Plan, Yucca Mountain Site
Nevada Research and Development Area, Nevada. Office of Civilian Radioactive
Waste Management, Washington, DC.
US Department of Interior (1977) Ground Water Manual, 1st edn, US Government
Printing Office, Washington, DC.
USEPA (1975) Manual of Well Construction Practices. US Environmental Protec-
tion Agency Rep. EPA-570/9-75-001.
USEPA (1977) The Report to Congress: Waste Disposal Practices and Their Effects
on Ground Water. US Environmental Protection Agency Rep. PB-265-081,
January.
USEPA (1987) Report to Congress on Management of Wastes from Exploration,
Development, and Production of Crude Oil, Natural Gas, and Geothermal Energy.
US Environmental Protection Agency Rep. EPA/530-SW-88-003, December.
USEPA (1988) Guide to Technical Resources for the Design of Land Disposal Facili-
ties. US Environmental Protection Agency Rep. EPA/625/6-88/018.
USEPA (1991) Design and Construction of RCRA/CERCIA Final Covers. US Envi-
ronmental Protection Agency Rep. EPA/625/4-91/025.
USGAO (1989) Drinking Water, Safeguards are not preventing Contamination
from Injected Oil and Gas Wastes. US General Accounting Office Rep.
GAO/RCED-89-97, July.
US Nuclear Regulatory Commission (1983) Disposal of High-Level Radioactive
Wastes in Geologic Repositories. Final Rule 10 CFR 60, Federal Register, Vol.
48, No. 120, June 30.
US Nuclear Regulatory Commission (1985) Disposal of High-Level Radioactive
Wastes in Geologic Repositories. Final Rule, Unsaturated Zone Amendment,
Federal Register, Vol. 50, No. 140, pp. 29641-8, July 22.
van der Kamp, G. and Gale, J. E. (1983) Theory of earth tide and barometric effects
in porous formations with compressible grains. Water Resources Res., 19,
538-44.
Van Everdingen, A. F. and Hurst, W. (1949) The application of the Laplace trans-
formation to flow problems in reservoirs. Petro. Trans. AIME, 186, 305-24.
Van Sambeek, L. L. and Stormont, J. C. (1987) Thermal/Structural Modeling of Test
Series A of the Small Scale Seal Performance Tests. Sandia National Lab. Rep.
Wahi, K K and Trent, B. C. (1982) Analysis of seismic response of repository
tunnels. Proc. 4th Int. Con[. Numer. Meth. in Geomech., Edmonton, pp. 625-34.
Wahi, K. K, Trent, B. c., Maxwell, D. E. et. al. (1980) Numerical Simulations of
Earthquake Effects on Tunnels for Generic Nuclear Waste Repositories.
Savannah River Laboratories, E. I. du Pont de Nemours & Co. Rep. DP-1579,
Aiken, Sc.
Wakeley, L. D. and Walley, D. M. (1986) Development and Field Placement of an
Expansive Salt-Saturated Concrete (ESC) for the Waste I solation Pilot Plant
(WIPP). Sandia National Lab. Rep. SL-86-36, Albuquerque.
Wakeley, L. D., Harrington, P. T. and Weiss, C. A. Jr (1993) Properties of Salt
Saturated Concrete and Grout After Six Years In Situ at the Waste Isolation Pilot
Plant. Sandia National Lab. Rep. SAND93-7019, Albuquerque.
Wakeley, L. D., Roy, D. M. and Grutzeck, H. W. (1981) Experimental Studies of Seal
Materials for Potential Use in a Los M edanos- Type Bedded Salt Repository.
Bibliography 323
Walls, J. D., Nur, A. M. and Bourbie, T. (1982) Effects of pressure and partial water
saturation on gas permeability in tight sands: experimental results. J. Petro.
Tech., 34, 930-6.
Walton, W. C. (1962) Selected Analytical Methods for Well and Aquifer Evaluation.
Illinois State Water Survey Bull. No. 49.
Warner, D. L. and Lehr, J. H. (1977) An Introduction to the Technology of Subsurface
Wastewater Injection. US Environmental Protection Agency, EPA-600/
2-77-240.
Wawersik, W. R. and Hannum, D. W. (1980) Mechanical behavior of New Mexico
rock salt in triaxial compression up to 200C. J. Geophys. Res., 85, 891-900.
Wenzel, L. K. (1942) Methods for Determining Permeability of Water-Bearing
Materials with Special Reference to Discharging-Well Methods. US Geological
Survey Water Supply Pap. 887.
Wheelwright, E. J., Hodges, F. N., Bray, L. A. et al. (1981) Development of Backfill
Material as an Engineered Barrier in the Waste Package System. Pacific North-
west Lab. Rep. PNL-3873.
White, E. L., Sheetz, B. E., Roy, D. M. et al. (1979) Permeability measurements on
cementitious materials for nuclear waste isolation. Scientific Basis for Nuclear
Waste Management, Materials Research Society Annual Meeting, Boston, Vol. 1,
pp.471-8.
Wilkinson, W. B. (1968) Constant head in situ permeability tests in clay strata.
Geotechnique, 18, 172-94.
Williams, J. R. and Daemen, J. J. K. (1987) Sealing Performance of Bentonite/
Crushed Basalt Borehole Plugs. US Nuclear Regulatory Commission Rep.
Winterkorn, H. F. (1975) Soil stabilization. Foundation Engineering Handbook, Van
Nostrand Reinhold Co., New York, pp. 312-36.
Woodcock, D. R., Everitt, R. A., Chernis, P. J. and Good, D. H. (1991) Geology of the
Buffer/Container Experiment Room 213 of the 240 Level. Atomic Energy of
Canada Limited Tech. Rec. TR-538, Pinawa.
Wroth, C. P. and Wood, D. M. (1978) The correlation of index properties with some
basic engineering properties of soils. Can. Geotech. J., 15, 137-45.
Wu, T. H. (1976) Soil Mechanics, 2nd edn, Allyn and Bacon, Boston.
Yong, R. N. and Warkentin, B. P. (1975) Soil Properties and Behaviour, Elsevier
Scientific Publishing Co., Amsterdam.
Yong, R. N., Boonsinuk, P. and Wong, G. (1986) Formulation of backfill material
for a nuclear fuel waste disposal vault. Can. Geotech. J., 23, 216-28.
Yost, F. G. and Aronson, E. A. (1987) Crushed Salt Consolidation Kinetics. Sandia
Zeigler, T. W. (1976) Determination of Rock Mass Permeability. US Army Corps of
Engineers, Waterways Experiment Station Rep. AD/A-021 192, Vicksburg.
Zeuch, D. H. (1990) Isostatic hot-pressing mechanism maps for pure and natural
sodium chloride: applications to nuclear waste isolation in bedded and domal
salt formations. Int. J. Rock Mech. Min Sci. & Geomech. Abstr., 27, 505-24.
Zeuch, D. H. and Holcomb, D. A. (1991) Experimental and Modeling Results for
Reconsolidation of Crushed Natural Rock Salt Under Varying Physical Condi-
tions. Sandia National Lab. Rep. SAND90-2509, Albuquerque.
Zoback, M. D. and Byerlee, J. D. (1975) The effect of microcrack dilatancy on the
permeability of Westerly granite. J. Geophys. Res., 80, 752-5.
ZoBell, C. E. and Molecke, M. A. (1978) Survey of Microbial Degradation of Asphalts
with Notes on Relationship to Nuclear Waste Management. Sandia National Lab.
Index
Aquifer 141 shrinkage limit 67

Artesian 140 sodium 281,282-5
Asphalt 210-11 structure 282-4
swelling 5, 65-6, 73, 74, 88, 209, 273, 274
Backfill 277, 278 tablets 206-9,289-91,292
see also Bentonite; Buffer void ratio 80, 87, 89
Bailer 2, 179, 205, 278 water content 84,85-7,90-1
Ballast material 67,68,209 workability 289
Basalt 13, 20, 25 yield stress 83-6,95
Bentonite Bond
as annular sealant 286-7 failure 37
anisotropy 6, 93 hydraulic 25, 269
Atterberg limits 90 see also Hydraulic conductivity;
block 218 Permeability; Plug-rock interface
bonding 289 mechanical 37-8,269,273,274
as borehole sealant 285 see also Concrete-rock interface; Plug-
calcium 281 rock interface; Push-out testing;
cation exchange 66-7 Stress strength 5,36-8,117-18,
chips 289-91,292 211-12
compaction 6,70-1,93-4,96,278 Borehole
crack 297 . applications 268-9
crushed rock mixtures 65,67-9,70-1,93, conditions 270-1
272,276 damage 24-5,271,278
cis granular 66-7 emplacement 104, 107, 117-18
dehydration 296, 297 hydrogeology 225
diffuse double layer 67, 94 inspection 270-1
as drilling fluid 291,292 logging 157,270-1
drying 99 wall 270-1
electrolyte 90 see also Well
erosion 65, 93, 272 Bridging 205, 277, 278, 286, 295
expansion 74,273 British Gypsum Ltd 254-5
flowability 289 Buffer 96-7,102,107-9,110,112-13,
granular 281,291,293-4 117-19,122-3
heterogeneity 6, 93 see also Bentonite
high-solid 292 Bulkhead 242
hydration 70, 74 see also Concrete ; Plug
hydraulic conductivity 5, 7, 96-7, 285
liquid limit 66-7,90-3 Case study 254-64, 294- 7
origin 282 Casing 189,277, 278, 286
pellets 201, 210 Cement
permeabilities 65,67-79,86-92,209, additives 13
218,285 class A 13-14,42
piping 5,65-6,78-80,82,93-5,272 class C 214
plastic limit 66-7, 90-2 class H 207
porosity 75,81,88 durability 207-8
sand mixtures 6, 96-7 grout 4,5,6,7,13,19,29,32,40
sedimentation 70 longevity 207-8
self-sealing 99 preparation 13-14
shrinkage 5 see also Cement plug
326 Index
Cementation Mining Ltd 257-8 strength 118
Cement plug temperature 119
axial load 28-9 thermal effect 248
axial strength 36-9 transporting 246-7
bond strength 29,37-8 see also Cement; Cement plug
crack 26,52 Concrete-rock interface 251
drilling effects 24-5 see also Plug-rock interface
drying 4,5,26-7,42-3,45-52,60 Consolidation
under dynamic loading 40, 52-8 granular salt 201-5
expansion 6, 13-14,25-6, 52, 62 plug 227, 258
hydraulic conductivity 42 Coring, see Drilling
interface effects 25-6 Crack, see Fracture
length 275 Critical pressure gradient 85- 6
length-to-radius ratio 32, 33, 34
peak shear strength 38 Damage
permeability 14,15,20,25 by casing removal 278
resaturation 27, 49 around drill hole 24, 271
shearing resistance of 31-6,275 around excavation 103,104,
shrinkage 4, 26, 52 196-9,228, 248
strength 28, 232-3 Darcy's law 88-9,130,132,151-2,196-9
stress distribution 18-19,20 see also Flow analysis
see also Stress Design
stress variation effect 20-4 of bulkhead door 242
swelling 13-14 calculation 233-46
Chemical compatibility 272, 277 considerations 218-23,270-4
Chemical equilibrium 207 criteria 39, 274-6
see also Chemical compatibility under dynamic load 56-60
Clay 135, 209-10, 280 factors 227-33
see also Bentonite; Buffer; Montmorillonite location 228
Colbert Project 294,295 steel load transfer 240-1
Compaction Diffusivity 96-8, 135-7, 153-4
backfill 278 Dried cement 26, 56, 60
bentonite 6, 70-1, 93-4, 96 see also Fracture
bentonite-crushed rock mixture Drift 7
70-1,276-7 Drill cutting 277,285
bentonite-sand buffer 107-8 Drillhole, see Borehole
buffer 107-8,112-13 Drilling
granular salt 201-6 damage 24-5,271
in situ 107-8 fluid 291,292
laboratory 70-1 methods 24-5
testing 70-1 mud 25
Compressive storage 143, 150, 169-70 oftest holes 13, 158, 178-9
Concrete Dry density 107-8,118
batching 246-7 Dye testing 50-2, 175
crack 251 Dynamic
expansion 120-1 acceleration 43, 52
hardening 119 compaction 205, 206
mix design 118,247-8 deformations 58-9
performance 120 experiments 42-4, 52-6
permissible stress 242-5 frequency 52
placing 246- 7 loading 40-1,52-6,57-8
plug 118 scaling factor 59-60
plug construction 233,246-9
plug displacement 121 Earth pressure 120
plug length 230-1,252-4 Earthquake 40
plug strength 232-3 see also Seismic
plug stress 232- 3 Effective porosity 151,152
porous 250 see also Porosity
quality control 247 Emplacement
shrinkage effect 248, 251 cementitious seal 278
.shrinkage strain 120 concrete plug 246-9
Index 327
granular salt 205 - 6 Grout
nuclear waste 96-9 cement 206-9
salt seal 205-6 procedure 249-50
see also Bailer; Compaction PureGold 296
Evaporite 3, 6 see also Cement
see also Rock salt
Excavations 227-8 Halite, see Rock salt
see also Drift; Mine; Ramp; Shaft; Heater 102, 109, 117, 124
Storage cavern; Tunnels Heater-buffer interaction 99
Host media, see Rock
Factor of safety 118,245-6,253 Host rock, see Rock
see also Design Hydraulic bond, see Bond
Field site, see Oracle Ridge mine; Hydraulic conductivity 4,6,7, 15, 118,
Superior Arizona; Underground 158,170-1
Research Laboratory; Waste Isolation see also Permeability
Pilot Plant Hydraulic gradient 110,122,130,151
Field testing, see Flow testing; In situ buffer Hydrodynamic dispersion 126,152
compaction; In situ experiment; Hydrogeology 1l0, 122, 141
Isothermal experiment
Filler materials 209 Impulse variation 144
Finite element analysis In situ buffer compaction 107-8, 112-13
on flow 15-20,132-5 In situ experiment 100-1
on stress 18-19,31-6 In situ testing, see Flow testing; In situ buffer
Flow analysis compaction; In situ experiment;
axisymmetric 15-18,134-5, 138-40, Isothermal experiment
147-50,162-5,171 Interbed 189
one dimensional 128,136-8,142-7, Isothermal experiment 101, 117-18
165-6,173
Joffe effect 204
see also Finite element analysis
Flow calculation, see Finite element Kozeny-Carman equation 88,90-3
analysis; Flow analysis
Flow net 15-16 Laboratory testing, see Compaction; Dye
Flow testing testing; Dynamic; Flow testing; Push-
with brine 218 out testing
with dye 50-2 Leakage
with gas 214 at cement plug-rock interface
head-buildup 140-50,154,168-71 25-6,52
high temperature 78-9 at concrete-rock interface 251
longitudinal 76 through concrete plug 250-1
one-side-configuration 127, 134-5, through rock fracture 25
155, 183 through surrounding rock 251
polyaxial 77-8 Limestone 158
steady constant-head 152-3 Load
steady-state flow 131-5,159 axial 28
tracer 150-2,154,171-6,217 see also Push-out testing
see also Dye testing thermal 273
transient constant-head 153-4 transfer 240
transient flow 135-40, 166-8 seismic 273-4
two-side configuration 127,155,182 see also Dynamic
with water 14-15 see also Stress
Fluid barrier 209
Flyash 207 Material selection 276-7
see also Cement see also Design
Fractional density 202, 203 Mechanical bond, see Bond
see also Consolidation Mine
Fracture 11,25,216,217 abandonment 126
Fuller-Thompson equation 68,69 closure 4
discharge 4
Geochemistry 227 flooding 4
see also Chemical compatibility safety 126
Granite 13,20,42,46-7,102,117-18 waste 126
328 Index
Montmorillonite 66-7,96,280-2 shearing resistance 275
see also Bentonite see also Bond; Push-out testing
size effect 59-60
National Coal Board North Selby strength 232-4, 239
Mine 258-64 stress 18-19,232
Nuclear waste repositories 2,28,41,61 see also Stress
see also Waste disposal taper 237, 238-40
Openings, see Excavations temporary 227,258
Oracle Ridge mine 158 thermal effect 248, 273
Overcoring 183 types 226-7
Overpack, see Buffer width 233
see also Bentonite; Cement plug, Concrete;
Performance testing, see Case study; Field Plug-rock interface
testing; Laboratory testing Plugging, see Sealing
Permeability Plug-rock interface
basalt 20 aperture 26
bentonite 65,67-79,209,218,285 axial strength 31
bentonite-crushed rock mixture flow 25-6,50-1,52
75-80 interaction pressure 25-6
calculation, see Finite element analysis; leakage 52, 251
Flow analysis shear stress 30-1
cement plug 15-18,20-5,46-50, Plug-rock permeability ratio
175-8,180 17-18,22-3
granite 20,46-50,118 see also Finite element analysis
granular salt 202-3,206 Pneumatic packer 159,168,
limestone 158 178,212-13
rock salt 186, 194-6 Porosity 75, 88, 199,206
tuff 20, 178-9 Portland cement, see Cement
Permeameter 11-12,131 Pulse test 140-2
Placement 2,7,275,277-8 see also Flow testing
see also Emplacement Pump 11-12,159-60
Piezometer 101-2,110,122,158 Push-out testing 29-31,211
Plug see also Bond
annular 26
axial strength 37-8
bearing resistance 234-5 Radioactive waste repository, see Nuclear
boundary 228 waste repositories; Waste disposal
confining pressure effect 20-4, 273 Ramp 7,97
concrete 118,213,246-9 Recovery variation 142-3
consolidation 227, 258 Recrystallization 204
construction 246-9 Rock, see Basalt; Granite; Limestone; Rock
control 226-7 salt; Tuff
deformation 120 Rock bridge 10, 14
design 233-46,249-54,274-6 Rock salt
displacement 121 constitutive model 193-4
drilling effect 24-5 creep 192-4
drying effect 26 dilation 199
emergency 227 disturbed zone 196-9
failure 228 granular 201-6,215-17,221
fracture effect 25 healing 199-201
granular salt 215-17 mechanics 191-2
see also Compaction; Consolidation microcrack in 196-7
leakage 250-1 permeability 194-6
length 230-1,252-4,275 triaxial compression 191-2
location 228 Rock specimen, see Basalt; Granite;
placement 277-8 Limestone; Rock salt; Tuff
see also Emplacement Salt, see Rock salt
precautionary 226,257-8 Seal
seismic effect 273-4 annular 286-7, 295
see also Dynamic bentonitic 215-18
shape 228-31 see also Bentonite
Index 329
cementitious 221,275 Stress
see also Cement; Cement plug analysis 18-19,33-6
compressive storage 143 axial 11-12, 30-1, 33-4
concrete 213-15 concrete 242-5
see also Concrete confining 11-12
design, see Design in situ 273
in situ stress effect 273 interface shear 30-1,33-4
see also Stress invariant 22
installation 179-80 permissible stress 242 - 5
see also Compaction; Emplacement; plug 18-20,232-3
Placement radial 19
interference from nearby rock 20-4,34-6,231-2
activities 272-3 shear 30,33-4,233
length 272,275 steel 245, 246
location 274 tangential 19
performance 10,211-12 tensile 34-6
see also Performance testing Suction 99,106,122-3
under pressure 272 Superior Arizona 29, 178-9
removal 175 Swelling clay 280
seismic effect 272-3 see also Bentonite; Montmorillonite
see also Dynamic; Seismic
system 274,276 Temperature gradient 119
thermal effect 273 Temperature variation 26
see also Plug Testing, see Field testing; Flow testing;
Sealing Laboratory testing
design, see Design Test site, see Oracle Ridge mine; Superior
location 270,274 Arizona; Underground Research
long-term 190 Laboratory; Waste Isolation Pilot Plant
objectives 268-9 Thermal conductivity 98,114
options 189-90 Thermal expansion 112,116
permanent 190,269 Tuff 13,19,20,29,34-5,67,178-9
plan 270-1 Tunnels 7,28,96-7, 126
practice 2,186,188-9,276 Type curve 136-7, 138, 139, 144-5,
regulations 1-2,188-9,269 149-50,154
research 2-3 see also Flow analysis
short-term 189-90
Underground excavation, see Excavations
Seal-rock interface 211,212
Underground Research Laboratory 99-100
see also Concrete-rock interface; Plug-
Uniformity coefficient 276
rock interface
Seismic Ventilation 228
activity 273-4 Viscosity 130
hazard 40-1 Void ratio 89
load 61 Volume strain 202
see also Dynamic; Earthquake
Shaft 7,28,225,258-64 Waste disposal
Shaking table 43,44, 52 Canada 97-9
see also Dynamic Sweden 3,96
Site characteristics 270-4 United States 186
see also Design see also Waste Isolation Pilot Plant
Smectite 280, 282 Waste Isolation Pilot Plant
see also Bentonite; Clay; Montmorillonite 186,212,213
Specific discharge 130 Water head 231,253
Specific storage 129, 130, 135-7, 153 Well
Stiffness 275 abandonment 269
Storage cavern 41,187,272 monitoring 286-7, 294-6
Strategic petroleum reserve 187-8 oil and gas 2, 186-7

J. J. K. Daemen Auth., K. Fuenkajorn, J. J. K. Daemen Eds. Sealing of Boreholes and Underground Excavations in Rock

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

J. J. K. Daemen Auth., K. Fuenkajorn, J. J. K. Daemen Eds. Sealing of Boreholes and Underground Excavations in Rock

Uploaded by

Copyright:

Available Formats

Sealing of Boreholes and Underground

JOIN US ON THE INTERNET VIA WWW, GOPHER, FTP OR EMAIL:

CHAPMAN & HALL

ISBN-l3: 978-94-010-7173-4 e-ISBN-13: 978-94-009-1505-3

~ Printed on acid-free text paper, manufactured in accordance with

2. Laboratory Performance of Cement Borehole Seals . 9

3. Strength Parameters of Cement Borehole Seals in Rock . 28

3.5 Push-Out Test Results and Discussion . 36

4. Dynamic Loading Impact on Cement Borehole Seals . 40

6. In situ Performance of a Clay-Based Barrier. 96

7. In situ Hydraulic Performance Tests

7.4 Conclusions.. 152

8. In situ Hydraulic Performance of Cement Borehole Seals . 157

9. Sealing Boreholes in Rock Salt . 184

10. Design of Underground Plugs. 225

11. Design of Borehole Seals: Process, Criteria

12. Sodium Bentonite as a Borehole Sealant. 280

Index .... 325

Sealing of boreholes and underground excavations has not received much

reference source for design professionals and researchers in many disciplines

Sealing of underground openings has assumed a growing importance with

1.2 SEALING REQUIREMENTS-RULES AND REGULATIONS

Sealing of Boreholes and Underground Excavations in Rock.

In recent years many countries and states have introduced or strengthened

1.3 CURRENT PRACTICE

1.4 RECENT BOREHOLE SEALING RESEARCH-AN

1.5 SEALING OF SHAFTS, TUNNELS, MINE ADITS,

Sealing of excavations larger than boreholes, such as mine shafts or adits, or

1.6 BOOK SUMMARY

Chapter 3 briefly summarizes a study of bond strength of cement grout

comparable to those of bentonite only. Compaction is a significant variable

1.7 THE FUTURE

Borehole sealing technology is well developed in a number of areas, particu-

of methods to detect abandoned holes, the development of non-invasive

The editors wish to acknowledge the support of the US Nuclear Regulatory

This chapter presents an experimental method to assess the performance

Sealing of Boreholes and Underground Excavations in Rock.

2.2 CONCEPTUAL APPROACH

The conceptual approach used to evaluate borehole seals in this work is to

Fig. 2.1 Specimen and permeameter configuration.

2.3 EXPERIMENTAL APPARATUS

2.3.1 Permeameter design

is applied by pressurizing water in the annulus between the specimen and

2.3.2 Pump design

2.3.3 Data acquisition system

2.4 EXPERIMENTAL PROCEDURES

2.4.1 Sample preparation

Cement borehole plug

1 Load permeameter, check seal

2.5.1 Flow calculation

calculated as a function of plug-to-rock permeability ratio (kp/k R , as shown

2.5.2 Stress analysis

Fig.2.5 Normalized flow rate as a function of plug-to-rock permeability ratio

2.6 SUMMARY OF TEST RESULTS

1.2 1.5 0.5 0.8 0.9

1.05 1.1 1.151.21.3 1.6 0 0.60.8 0.9

Table 2.2 Summary of rock and borehole plug permeability

Charcoal granite 52-85 <52

2.6.1 Effects of stress variation

6 Axial and Confining

Tap Pressure, PT , (MPa)

10- 8 10-7 em/min

Fig.2.10 Permeability as a function of first stress invariant for a Charcoal granite

Fig.2.11 Flow rate plotted as a function of plug-to-rock permeability ratio for a