Professional Documents
Culture Documents
Excavations in Rock
Edited by
K. Fuenkajorn
Rock Engineering International,
Tucson, USA
and
J. J. K. Daemen
Mining Engineering Department,
University of Nevada, USA
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
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
Bibliography . 298
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
K. Fuenkajorn
J. J. K. Daemen
CHAPTER ONE
Introduction
J. J. K. Daemen
1.1 BACKGROUND
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.
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
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
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.
ACKNOWLEDGEMENTS
Laboratory Performance of
Cement Borehole Seals
D. L. South and K. Fuenkajorn
2.1 INTRODUCTION
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
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.
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
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
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.
18 Laboratory performance of cement borehole seals
2
10
10
-1
10
-2 -1
10 10 10
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.
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
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"
20 Laboratory performance of cement borehole seals
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'.
'2
~ 3
Q
Ji
:!
"
II:: 2
~
Ii:
,
./
2 4 6 8 10
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.
22 Laboratory performance of cement borehole seals
( 'O.OMPa
,7.0 MPa
c10~2 ~3.5MPa
1
:1
"
<r
~IO-3
CG-I02
3.3
4.2 ROCK BRIDGE
4.6
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)
3.5MPa
'C10- 2
~u
oS
!
"
a:
~ I _ _ _ _ _ _ _ _ _--r=--..::::::::7.0MPa
f!.IO- 3 t-
1 - - - - - - 3.5 MPa
CG-I02
CEMENT 1 PLUG
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.
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.
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.
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).
Rock
sample
Dial gage
(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).
3.4.1 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).
32 Strength parameters of cement borehole seals in rock
,----,
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.
Finite element analysis and discussion 33
0.8
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).
34 Strength parameters of cement borehole seals in rock
'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.
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
36 Strength parameters of cement borehole seals in rock
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
2.6 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.
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
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
Fig.3.7 Applied axial stress vs. axial plug displacements of a push-out sample.
38 Strength parameters of cement borehole seals in rock
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
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).
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
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
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.
44 Dynamic loading impact on cement borehole seals
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.
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.
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
~
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.
46 Dynamic loading impact on cement borehole seals
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'
g
of
8
of
Iii
of
~
'0 120 lID 240
8
~
Iil
~
~
m
Iil
;
8
of
Iil
of
~
'0 60 120 180 240 300
(d) ELA'SEO TIME lOAYS'
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
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.
50 Dynamic loading impact on cement borehole seals
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.
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).
52 Dynamic loading impact on cement borehole seals
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.
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
~
i'a::!
'"
'"~
0 0
~
0
II?
0
0
0
(a)
-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.
54 Dynamic loading impact on cement borehole seals
-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)
-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
'.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.
56 Dynamic loading impact on cement borehole seals
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.
Tension Compression
(a)
Negative Curvature
(b)
(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.)
58 Dynamic loading impact on cement borehole seals
nearly normal to the tunnel axis. One effect of the hoop deformation is the
distortion of the cross-sectional shape (Figure 4.13c).
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.
Design considerations 59
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.
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
60 Dynamic loading impact on cement borehole seals
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.
4.7 DISCUSSION
4.8 CONCLUSIONS
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
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.
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.
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
0 100
10 1 0.1
GRAIN SIZE (mm)
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.
--~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.
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.
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 )
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.
1()-8..--_ _ _ _ _ _ _ _---.
I
I I
~1()-8
E
.!:!- I
~
II
::::;
ffi
;J
II
w
~ I I
II:
~ 1010
I
I I
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%.
~
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
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.
~ 10.8
-!!!
l
I' I I i
~
::::;
iii
.--..
<
w
:;
a:
w
10.9
I I
Ilf
0.
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
80 Bentonite and bentonite/crushed rock borehole seals
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
Equation (5.3) gives a better fit and therefore is more suitable for estimating
the permeability of bentonite.
Permeability tests and results 81
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.
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
~ 103
~
W
II:
!ii
c
...J
W
;;: 102
Fig.5.6 Yield stress of bentonite pastes vs. bentonite wt%. Triangle points are
from Marsland and Loudon (1963).
or
log('I) = 3.841 - 0.00214w (5.7)
.
I
2'f
=- (5.13)
c,p Rm
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.
(5.16)
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
and
(5.20)
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.
5.6.1 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
94 Bentonite and bentonite/crushed rock borehole seals
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
This work was part of research effort sponsored by the US Nuclear Regula-
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 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.
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.
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
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.
6.5.4 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
104 In situ performance of a clay-based barrier
Buffer/Container
Experiment
P~
Isothermal
T
",# A
~~4
-=:J
o 20m
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.
Buffer/container experiment 105
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.
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.
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.
80
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
Fig 6.7 Temperatures measured in the buffer and the rock in three different direc-
tions, illustrating axial symmetry.
110 In situ performance of a clay-based barrier
........ _-----
_____ ~~~_~_'--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
Scale:
r""'1"iii""
too
50
or-~~-----------------------------------1
Fig 6.10 Pore pressure distribution in the rock at the heater mid-height.
112 In situ performance of a clay-based barrier
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.
.~
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.
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
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
Fig 6.14 Calculated moisture content variations across the buffer annulus.
116 In situ performance of a clay-based barrier
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'-". . . .
-----.
/' \
-100OL--L.--1-00L.----J--200...L...-...L.....--:3~00---L.--:-40~0:--......L..--'---'
(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.
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.
120 In situ performance of a clay-based barrier
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.
122 In situ performance of a clay-based barrier
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.
Rock
Concrete Plug
High Suction
As-Compacted
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
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
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.
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
Packer
Head = H--==---,
Grout seal
Head - H, _-==-""1
(H,<H)
0 0 - - Shut-in valve
Packer
Head = H --t:::==--'
Grout seal
Head = He ---==--.
(H,<H)
Packer
[x-l--Shut-in valve
~
Q, = collection rate
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.
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)
ding Walton (1962) and Neuzil et al. (1981). A detailed explanation of the
method is provided by Wenzel (1942).
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).
132 In situ hydraulic performance test of borehole seals
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)
r1 r2
K. Kr r5
ra
r7 rs
Fig.7.3 Flow field for axisymmetric analysis of steady constant-head seal test.
134 In situ hydraulic performance test of borehole seals
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.
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.
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
136 In situ hydraulic performance test of borehole seals
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.
One-dimensional analysis
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)
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.
138 In situ hydraulic performance test of borehole seals
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~
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.
Axisymmetric analysis
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:
(7.12)
and
(7.13)
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).
140 In situ hydraulic performance test of borehole seals
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).
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).
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).
One-dimensional analysis
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:
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
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)
144 In situ hydraulic performance test of borehole seals
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.
10
Fig.7.6 Type curves (hjH vs. tdJ) for recovery variation of one-dimensional head-
buildup test.
1.0
~ 0.5
.r:.
10
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).
~ 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
1.0
~ 0.5
.c:
10
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).
Axisymmetric analysis
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.
148 In situ hydraulic performance test of borehole seals
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)
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~.
150 In situ hydraulic performance test of borehole seals
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.
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.
One-dimensional analysis
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)
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
One-dimensional Axisymmetric
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.
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.
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
This 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 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.
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.
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.
160 In situ hydraulic performance of cement borehole seals
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
Seal test at Oracle Ridge Mine 161
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.
162 In situ hydraulic performance of cement borehole seals
Axisymmetric analysis
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
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
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.
164 In situ hydraulic performance of cement borehole seals
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
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.
Seal test at Oracle Ridge Mine 165
10~c---------------------------------,
i 10.1 L-~---
';.:
10. 11 0~-'-'--'--'-'-"---'---':-2000~--'-'-'--'-'--'--'-4'-=OO':c0:'-'-.J-1--'-'-'--"--':--:'6000
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.
One-dimensional analysis
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
166 In situ hydraulic performance of cement borehole seals
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.
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.
One-dimensional analysis
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
Seal test at Oracle Ridge Mine 167
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)
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
I-'E=-----if7\ Bourdon-tube
vpressuragage
A
B Valve
Fig.8.7 Collection zone and system for monitoring pressure for the head-buildup
test.
Seal test at Oracle Ridge Mine 169
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.
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 .
One-dimensional analysis
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)
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))
Axisymmetric analysis
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.
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
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.
One-dimensional analysis
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)
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.
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
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.
Porous stone
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).
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.
ACKNOWLEDGEMENTS
This 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 NINE
Sealing Boreholes
in Rock Salt
J. C. Stormont and R. E. Finley
9.1 INTRODUCTION
9.2.1 Applications
Some of the important applications which require sealing of boreholes
which penetrate halite deposits are briefly described below.
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.
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.
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
190 Sealing boreholes in rock salt
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.
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.
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).
194 Sealing boreholes in rock salt
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.)
196 Sealing boreholes in rock salt
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.)
Rock salt properties relevant to borehole sealing 199
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
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
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
Materials for sealing boreholes in rock salt 203
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
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.
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.
9.4.3 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
210 Sealing boreholes in rock salt
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.
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).
Pressure and
Flow Meters
00
Receiving Packer (Optional)
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.
214 Sealing boreholes in rock salt
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.
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).
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)
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.
I
I
I
I
Borehole Borehole Time Geochemical
Drilled Sealed (Dimensionless) ~radation (years)
-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.
Design considerations 221
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.
224 Sealing boreholes in rock salt
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
10.1 INTRODUCTION
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.
---
Possible water leakage paths Possible water leakage paths
through strata through strata
, ; - - .........
*- ....
* ~~..~...~~.
JllIYii~II!l. ".' ~~
. !.
..
Water
pressure
..
Water
pressure
...Water
--
pressure
(e) (d)
(I (I. .. o' 0
.0'
O. D ,'o~' ~ ...0' o.
lIS
(e) (f)
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
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
232 Design of underground plugs
1= pbh (1O.1b)
2(b + h) Ppe
For a square cross-section, Equation (1O.1b) becomes
1= pb/4p pe (lO.1c)
giving
(1O.2b)
-Compression component
ex t of force on plug
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 which
[= pbh (10.8b)
(b + h) Pbe tan IX
236 Design of underground plugs
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 .
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.
Design calculations 237
(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.
(1O.14b)
(1O.14c)
(1O.14d)
238 Design of underground plugs
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
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.
Design calculations 239
(10.21b)
sive stress situation. Care should be taken to ensure that the calculated
combined compression and bearing stress,
(1O.23c)
(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
where Pms is the permissible steel bending stress and tf is the flange thick-
ness.
Evaluating the effective membrane thickness for both the cylinder wall
and the flange gives the effective membrane thickness per unit length,
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.
P1tr~f
f bs = 1t(r~ _ rf) ~ Pbs
(10.30)
(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).
242 Design of underground plugs
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.
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
a Based on a factor of safety = 4; h British Standards Institution (1969b) (withdrawn); C British Standards Institution (1985).
Design calculations 245
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.
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.
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).
materials were incorporated to minimize the thermal effects which are dis-
cussed in more detail below.
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
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.
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.
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.
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.
Plug sealing and resistance to leakage 253
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
254 Design of underground plugs
.. '
. . . Rising main
6>~~'
q~ borehole
.o~~r to surface
--------.~~~~----~~~~
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
,\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
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).
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
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"-"~"
".'-~": ~~~'(/#:~ ;~~-==
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.
260 Design of underground plugs
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
~
- -
(e) (d)
<a>
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.
262 Design of underground plugs
fa Bl
Visqueen Strata grouting Steel angle
membrane standpipes (48 No.) support frames
rc Cl
Gravel be<
(a)
(c)
(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.
264 Design of underground plugs
were sleeved and two water bars were incorporated, the inner one protected
and the outer one sacrificial for plug sealing.
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
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
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
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
270 Design of borehole seals
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.
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.
11.3.5 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.
11.3.8 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.
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.
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.
276 Design of borehole seals
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
Sodium Bentonite as
a Borehole Sealant
1. E. Papp
12.1 INTRODUCTION
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.
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.
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
o Oxygens
Hydroxyls
Aluminum
Silicon
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.)
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.
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.
Concrete
Borehole ------
Sand or
gravel pack
I=~r"'-'----Silotted well
screen
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.)
Types of bentonite sealants 289
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).
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
Types of bentonite sealants 291
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.
Fig. 12.8 Diagram of a multilevel well installation sealed with the sand/bentonite
injector.
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
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
Rep. NUREG/CR-4641, Washington, DC.
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
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
Rep. NUREG/CR-4919, Washington, DC.
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.
SAND83-1293, Albuquerque.
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
National Lab. Rep. SAND87-7155, Albuquerque.
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
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
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.
SAND86-1329, Albuquerque.
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
National Lab. Rep. SAND87-0176, Albuquerque.
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
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.
NUREG/CR-4983, Washington, DC.
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
National Lab. Rep. SAND87-0264, Albuquerque.
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.
Rep. SAND78-1371, Albuquerque.
Index