Session12

Soil lmprovement- State-of-the-ArtReport

Améliorationdes Sols

J.K. MITCHELL Professorand Chairmanof Civil Engineering,Universityof California,rBerkeley,

CA, USA

SYNOPSIS Because of the increasinq need to utilize marginal sites and because many soils
can be made into useful construction materials if properly treated, soil improvement has become a
part of many present day civil engineering projects. In this state-of-the-art report the princi-
ples, applications, and design procedures for soil improvement using different methods are Presented.
Soil improvement methods reviewed include in-situ deep ccmpaction of cohesionless soils, precom-
pressíon with and without vertical drains, injection and grouting, admixture stabilization, thermal
methods, and soil reinforcement. Comprehensive recent references on each topic are listed.

INTRODUCTION report is on practical aspects of soil improve-

ment, to include such considerations as soil
yost of manrs construction is done on, in, or types best suited for treatment, effective
with soil. As the availability of suitable con- treatment depths, properties of treated soils,
struction sites decreases, the need to utilize major applications, and relatir/e costs. Because
pcor soils for foundation support and earthwork of the extensive scope of the subject the in-
construction increases. In addition, it is clusion of detailed case histories is not
becoming increasingly necessary to strengthen possible.
the ground under existing structures to insure
stability against adjacent excavation or tunnel-
ing, or to improve resistance to seismic or
other special loadings. Furthermore, many hun- METHODSAND SCOPE
Creds of recent successful projects have shown
rhat through the use of suitable reinforcement The Conference organizing Cornmittee has identi-
:raterials and systems, the uses of nature's fied the following topics for discussion in
nost abundant constructi-on material--soil--can Session 72, so they have been chosen for cover-
De greatly extended. Tt is not'surprising, age in this state-of-the-art report.
t,herefore, that the general area of soil improve-
nent and reinforcement has been of great inte- 1. Compaction, especially heavy tamping and
rest and rapid development in the past several blasting. Emphasis is on in-situ deep
i'ears. The inclusion of this topic for the densification of cohesionless soils, and
Sirst time as a main session in an International compaction in thin layers is excluded.
:onference of our Society is clear recognition
cf this fact. 2. Consolidation by preloading and,/or vertical
drains, eLectro-osmosis.
The basic concepts of soil improvement; namely,
Crainage, densification, cementation, rein- 3. Grouting, excluding control of groundwater
:orcement, drying, and heating, were developed flow and seepage.
hundreds or thousands of years ago and remain
valid today. The coming of machines in the 4. Soil stabilization using admixtures and by
lgth century resulted in vast increases in both ion exchange (chemical stabilization) .
rhe quantity and quality of vrork that could be Emphasis is on new developrnents and applica-
Cone. Among the most significant developments tions other than subgrades and base courses
of the past 50 years are the introduction of for roads and airfields.
,.'ibratory methods for densification of cohesion-
Iess soj.ls, new injection and grouting mater- 5. Thermal stabilization.
:.als and procedures, and new concepts of soil
re inforcement . 6. Reinforcement of soil. Emphasis is on the
inclusion during construction or installa-
The purpose of the present report is to synthe- tion in-situ of both tensile and compres-
sj.ze the present state-of-the-art of soil sion reinforcement elements. Geotextiles
improvement into a form suitable for direct used as reinforcement are included.
application by practicing engineers. At the
same time the author has attempted to identify The report concludes with a tabular summary of
sufficient references that interested readers the methods discussed and consideration of
r¡i11 be able to locate more detailed informa- factors governing the choice of a method for any
rion, case hi.stories, etc. Emphasis in this given case.

509
 taining fines and many waste-fit1s, densifica-
DEEP COMPACTIONOF COHESIONLESSSOILS
tion iá mainly by collapse of the soil structure
and escape of gas from the voids' The process
Introduction by impact com-
is much lhe same as densification
Thick deposits of loose cohesionless soils may paction as commonly done in the laboratory'
require improvements in order to eliminate the
suüsequent- development of excessive total and Densification accompanying ground treatment by
itiffeiential settlements and to minimize the pos- these methods occuri rápidIy. settlement of the
"iUifity for liquefaction under dynamic loading' giound surface is essentially complete by.the
Suitablé improvém.ent can be achieved in many énd of treatment. Improvement in properties, as
-ases by densificationi however, the needed neasured, for example, by penetration tests or
ttensifióation cannot ordinariJ-y be achieved Dressuremeter tests, may continue over extended
using preload surcharge fills or.compaction at Lime periods. This latter point may be of con-
the éuiface. rn-situ densification of loose, "idérá¡i" practical importance in the evaluation
cohesionless soil layers is usually done by of ground treatment.
dvnamic methods. In many methods dynamic load-
iig is accompanied by displacement in the form Experience has indicated that it is often easier
ot-tfr. inseriion of á probe and/or construction to'densify to a specified hiqh relative density
of a sand or gravel column in-situ. from a loóse initial condition than from an
interrnediate relative density. This is because
Methods that are used for the in-situ deep den- the initial structure of the loose material is
sification of cohesionless soils include blast- easier to break down.
ing, vibrocompaction, and heavy tamping'
viÉiocompactión is used herein to refer col- Soil Type Considerations
lectivel! to alL those methods involving tle +n:
sértion óf vibrating probes into the ground.with Vibrocompaction rnethods are best suited for den-
of a backfill material' sificatión of clean, cohesíonless soils'
ár without the addition
n"pátié"." has shown that they are generally
Compaction piles are also considered in this when the percentage by weight of
"ullqoty. ihe ability of any of these methods ittLftecti.t.
iines (particles finei than 200 mesh sieve or
to aócornplish the needed improvement- in proper- rhis is
ties depénds on several factors, including: ó:ót¡ "* diameter) exceeds 20 t'o 25'
úá."""á the permeability of materials contain-
1. Soil type, especially its gradation and ins greater percentaqes of fines is too low to
fines content. .fíot tn" rapid drainage of pore v'ater that is
requireil for densification following liquefac-
tión under the action of the vibratory forces'
2. Degree of saturation and v¡ater table
.rrd ba."o"" the structure may be more difficult
location by the
lá aisrupt owing to cohesion contributed
3. Initial relative densitY fines.

Some soils containing greater amounts of fines;

4. Initial in-situ stresses
á.f.,-"á*" silty sanás and loess, can be densi-
soil structure, including the iiéa'uy blasting and/or heaw tamping, both of
5. Initial
effects of age, cementation, etc' which impart laige amounts of energy alf at once
and causé large ground displacements-' More
of the method specific consideiations concerning the in-
6. Special characteristics
Fio.tt"." of soil type are presented in the dis-
used. methods. For preliminary
.""Ái"" of particulár
olanninq, hówever, it may be considered that th€
Mechanism of Densification size distributíons shown by
iár,g" oi-particle
Densification of cohesionless soil layers with rig. f. wiit ¡e best suited for densification by
accompanying improvement in mechanical proper- deep in-situ methods.
ties ieqüirés first that the initial soil struc-
ture be-broken down so that particles can be Evaluation of Treated Ground
mávea to new packing arrangements. In saturated
Measurement of the effectiveness of deep com-
cohesionless materiáls this is most readily
by means - pá.ii"" can be nade using one or more of several
accomplj.shed by inducing liquefaction
loadings. In the case of fr"lir"a". Techniques thai have been used includc
ot aynamic and cyclic
meth;ds such as blasting and heavy tamping the
I. Surface settlement markers
compression htave generated by the sudden large
.tt.igy release can give an immediate-bui1d-up in
the 2. Volume of soil added to fill craters, to
pore water pressure which greatly-reduces or to carry out a
lhear sttength. This wave is followed by a form compaction piles,
shear wave wt¡ict is responsible for failure of vibrocomPaction Process
ih. *""". After passagé of these waves the soil Tests (SPT)
Darticles settle into nel.¡ and, ultiriately' more 3. Standard Penetration
'sta¡le
positions. Vibrocompaction methods are (cPT)
4. cone Penetration Tests
eftective in much the same way, except that the
energy Per event is many times- smaller' the
5. Pressuremeter Tests (PMT)
viuráiións continue over a much lonq¡er period,
and the effects are felt to distances from the determination¡
6. Seismic shear wave velocity
é""tgy source of one to two meters instead of
up tó- 10 m or more as is the case with blasting driving resistances
and heavY tamPing. 7. Pife

soils, including some con- 8. Plate load tests

For partly saturated

510
 !
€
Ol
(¡,
=
\
a
\
(t) Most DesirobleSizeRonge
c
l.¡-

!
c(¡)
(J
\
q.)
o-

) 5 2 t.0 0.5 0.2 0.

0.2 0.tt 0.05 0.02
0.05 0.0t
0.02 0.0 o n5
t 0.ffi5 0.00t
P or ticle Size - m m
Fig. I Range of Particre size Distributions suitable for Densification b y Vibrocompaction

9. Down-hole density meters. Approximate correlations between equivalent

penetration resistance, sand density and proper-
Settlement measurements and SpT, CpT, or pMT are ties pertinent to the assessment of foundátion
the most commonly used methods. Th.e CpT is stability are given in Table r.
particularly useful because it provides a con-
tj.nuous record of p e n e t r a t i o n resistance with In a few cases instrumentation has been used to
depth, it is fast, and it is well-suited for use monitor conditions during densification. Tot,al
in sands. Penetration tests for evaluation of pressure cells and piezometers have provided data
improved ground are usually done at locations useful for developing improved
j.ntermediate understanding of
between probe points in order to the densification and property improvement
provÍde the most conservative estimate of process.
improvement.
Blasting
Penetration resistance values, both before and
after ground improvement, are often Deep compaction by detonation of buried explo-
converted
to relative densities using one or more of sives can provide a rapid, lohr cost means for
several correlations that have been developed soil improvement in some cases. The general
for particular conditions. Desiqn criteril procedure consj-sts of:
and
specifications are many times developed in terms
of relative density. A direct conversion of a
l. fnstallation of pipe by jetting, vibration,
penetration resistance to relative densitv or other means to desired depth of charqe
is
uncertain, however, because penetration placement
résis-
tance depends on factors other than density.
The correlations are not independent 2. Placement of charge in pipe
of soil
type. Increased lateral pressure, increased
time under pressure, increased stability 3. Backfilling the hole
of
structure, and prior seismic strains lead to
increased penetration resistance (Seed, l-9] 9) . 4. Detonation of charges according to a pre-
Fortunately, these latter factors also lead to established pattern.
corresponding increases in resi.stance to settle-
nent and liquefaction, and it is the penetration In some cases the pipe is withdrawn prior to
resistance values themselves that are important, detonation of the charges. In others it is
not the actual relative dénsity. It has been reclaimed after the blast, a new sectj_on is
found convenient for some applications, welded to the bottom. and it can be used again.
however,
to work with an " e g u i v a l e n t relative The explosives used include dynamite, TNT, and
density,'
which is the true relative density a sand Ceposit anmonite. Detailed descriptions of blasting are
',¡ould have to possess to exhibit given by Prugh (1963), rvanov ( 1 9 6 7 ) , Mitchell
the measured
penetration (1970), Litvinov (I973, I976), Damitio (1970-
resistance if ít were freshlv
deposited and normally consolidated. 72), Donchev (1980), and others.

511
 TABLE I

Penetration Resistance and Sand ProPerties

30-35 > 5 0
SPT N-value 4-10 l0- 30
( b 1 o w s , / 0 . 3m ) *
150-200 >200
CPT cone resistance 50-100 r00-r 50
(kg/cmzl *
35-65 65-85 8s-100
Equivalent Relative l5-35
óensity (t) **

1 4- 1 6 I6- I8 r8-20
Dry Unit Weight
(kN,/m")
32-35 35-38
(") 30-32
Friction angle

Cyclic Stress Ratio 0.04-0.10 0.r0-0. 35 > 0 .3 5

óausinq Liquefaction
(r/oo') ***

overburden pressure of 100 kPa'

*At an effective vertical
**Fresh1y deposited, normally consolidated sand'

***From Seed (19?9), fig. 6(a).

6. Surface settlement: 2 to 10t of layer

for den-
Saturated, clean sands are well-suited thickness '
b v b r a s t i n s ' s u c c e s s in anv case
;i;i;;;i;;
of the shock wave sene- can be used
;;;;;;-;; lÉe auiritv
down the initiar The maximum depth to which blasting known'
IIM-iv-ti'á-¡tast tá'¡rear tor soil compaction i" lgt
condition successfully a
structure, an¿ creaie a liquefaction to The author is currently associated-with
period to enable Particles have
for a sufficient rt "r"llli'-i; which charsés of up to 30 ^ks
;;;r;J;;-ttrárnserves in a denser packins' aéóÉt's more than 40 m below
t"áí"á!.á""1éá-"t
follows, therefore, that the stronger the sand sié'niticant surrace-settlements
be ;;;";-;;;;;;.
i;;;r;líy; the larser the charses.that wilr ;;á-i;p;;;;^á"t i" Éhe equivar:"t l:l?ti"'
densifióaliont. Thus' the than
required for effective rs á;;";¿; áf 1oo"" zones,a€ q"plh:-g1:iter feature
"iÉIilt-tÁ; depth to which densification 30 m have ueen acntá""d' An interesting
the initiar equivalent that although
X";;;;-#-tr'á'ttist'"r of this work is ttre-observation
;r;¿i"; density, the greater the explosive settlement is irimediate' which means
;;r;;;; im¡nediate'
energy required. i; ;i;" essentially
that densificatÍon indi-
o f c o n e p " t " l t á l i o " tests-do'not
accepted theo- results density
There appear to be no generally by cate an Íncrease t"-tq"i""ftnt relative
retical design pto.tdoít" for áensification
il-ih."t.s"ir.a t'"ioe-iár severtr weeks' reflec-
;;;;;;;r'-;ná riera trials are usuarlY used itá"ij'ne effect folrowins dis-
A number of fier¿l ;;";";"-;;i"g-"r and rormation
;;i;;l¿'p;áá,'.ti"" urá"li"g' i"lii"ii ;;-;ñ" ini[iái átructure
;;":-¡.át' ",ttt*tized Év rvanov (1967) in-
:;;;; "'o* tn'"' of a new one.
Lo.20
;;i;;";-¿;.atment 4-"oll" up
general T.:..
guidelines maximum depth for
experiences tne tojtbwing It would be expected that the
emérge, treatment wiII be limitecl by the prac-
effective
tical itifficurty oi piácing concentrated charges
1. Charge size: <I to 12 kg
of sufficierrt *tgttitüae to create a shock wave
íiquerv the initial-soil struc-
;;.;;-;;;;;t¡ to
2. Depth of burial: >1/4 ilepth to bottom ;;-¿tre deptÉ-inóreases' "o q?-th' errec-
i;;;:
of l a y e r t o - b e t r e a tea¡ L/2 Lo 3/4 of ii"!'"ttá""á" .,'á strensth' .1:"9'9'-is1Y':l:
stréss-Yi11-i":.:?ase' ano
dePth cornmon iá"iii.á áisruptive
of influence will decrease'
ñ:-"?;;";rve iaaius Pmax
Charge sPacing in Plan: 5-15 m wave pressure'
3. The maqnitudes of the shock pá' Yl-i!-?""' r' in
l-5 with 2-3 usual' iii=*ü7Zii;"';;áLh.l;pii"é (196'71as forlows
4. Number of coverages: kq-seé/cm-, ^t. gtt"i"'[v-i""it""
number of
Each coverale coásists of a
inaiv:-¿uaf óharges' Successive coveraqes
are usually-sepárated by hours or days' (r )
nr.* = kr ( )ut
5. Total .explosive use: 8 - I 5 o g m ' l m 3' 1 0 -
30 gm,/mi tYPical

512
 I = n,(*l-)('€)" (2) Attempts have been made to compact using surface
explosions, because of sirnplicity, 1ow cost, and
speed. Because of energy loss above ground,
in which c = size of charge (k9 of TNT) lack of confinement, and the formation of surface
= distance from center of depressions, however, this method has been of
R
limiteil ef fectiveness.
charge (m)
U 1r U 1 = e m p i r i c a l c o e f f i c i e n t s from A hydro-blasting technique has been used very
Table II. successfully and economically for cornpaction of
collapsible loess deposits (Litvinov, L973, 1976¡
These relationships can be used for comparative Donchev, 1980). Although collapse of the loess
studies of the probable influences of charge can often be accornplíshed by flooding alone, it
size, charge spacing, and sand saturation. It has been found that more uniform results can be
may be seen that the presence of even smaLl achieved more quickly and economically by this
amounts of gas leads to significant danping of method. The procedure, which is illustrated in
the P-wave pressure. Fig. 2, consists of first cutting a contour
trench 0.2 m to 0.4 m wide and several meters
It has been possible by blasting to densify deep around the perimeter of the area to be
sands to equivalent relative densities of 75 to densified. Boreholes spaced a few meters apart
30 percent. In some cases, however, the results in a grid pattern are then used to pump r{tater
:ray be erratic, initially dense zones may be into the loess, over a period of several days,
).oosened, and the method is not likely to be ideally until the water content is increased to
effective in the upper one or two meters below above the liquiil limit. Slurry walls or plastic
:he ground surface. Typical behavior may be membranes can be installed to prevent lateral
s'¡mmarized as follows. migration of the r.rater and softening of adjacent
ground.
-. Almost immediate settlement of the ground
surface, with little further settlement Explosive charges of about 5 kg each are then
with tine. inserted down tubes installed at spacings of
three to six meters in grid patterns and deton-
-. Initially loose zones show little immediate ated. Surface settlements of up to 10 percent
change in penetration resistance. Penetra- of the layer thickness and reduction in porosity
tion resistance increases slovrly with tirne of several Dercentase points are not unconmon.
until after several weeks the material A r e a s o f 1 0 - 0 0m 2 t o ' 1 0 ; o o o m 2 i n v o l v i n g 1 0 , 0 0 0 -
indicates a marked improvenent in properties I O O , O O Om 3 o f l o e s s c a n b e t r e a t e d a t o n e t i m e .
compared to its initial condition.
Successful cornpaction of saturated sand and
. zones which are initially very dense may be collapsible loess has been accomplished in the
permanently loosened or weakened by the USSR using high energy, high voltage electrical
blastt however, ttre resultant condition is discharges from probes inserted in the ground
sti1l likely to be satisfactory. (Lomize et al., L963, 1973). Each discharge,
which may release 50 to 100 kJ of energyr has an
. 'Jl.timateIy,
an effective blasting program effect similar to that of an explosion of com-
results in a deposit in which all the init- parable magnitude. A number of discharges are
:a11y loose zones have been suitably released spaced at intervals of several seconds
:mproved. at each level as the probe is moved upwards from
the bottom. It appears that use of this method
. t:. á series of coverages is used, the surface has not yet been widely adopted.
-----lement accompanying each coverage is usually
-:s chan the one precedinq.

TABLE TI

Parameters for Estimating B1ast Pressures and lmpulse Values

(from Ivanov, 1967)

Gas Content
t*l
kt u1 kz

Sand below water table 0 600 r.0s 0.080 1.05

0.05 450 r.5 0.075 I .10

I 250 2.0 0.045 r.25

4 4J 2.5 0.040 1.40

Moist sand (8-10t water) 7.5 3.0 0.03s 1.50

(2-4* water) 3.5 0.o32 1.50

513
 Conlour Trcnchag 15 Hz and a vertical amplitude of 10-25 nm.
About 15 probes per hour can be done at
spacings of l- to 3 m. It is of marginal
eifectiveness in the upper 3 to 4 m of the
o aa o o aa zone to be densified.
t Droin Jfclb
" d o o Vibro-rods developed by Saito (1977), Fi9. 3,
are also driven using a vibratory pile driv-
/ Elotling Holcs
. t o o t ing hammer. Several cycles of insertion and
withdrawal are used in the densification
process.

t 3 O O ¡
Uncompoclcd
Soirs compoclad Arao
i
A
tgl'--::ll:ÉÉ= ll - :ll = :ll: - ll =--ll- ill--1
tl'.----il lt =ll- --ll- s{lF; ll. --ll.¡.$
!>K{l-H,|l.l:€
[=_;]59=1,
llrF ='l
r\\.¡,

il-.
l\,-24+-'stu'¡,ping
1 . . - _ -
Soil tin Flooded Zoni I t

Section A-A
:1]

lI lr5oo

Eig. 2 Loess Compaction by Hydro-Blasting

/nI l.."
ffif
y.ibrocompaction and
These methods for deep compaction of cohesion-
less soils are characterized by the insertion
of a cylindrical oi torpedo-shaped probe into
the ground followed by compaction by vibration
during withdrawal. In a number of the methods
a granular backfill is added so that a comPacted
sand or gravel column is left behind within a
volume of sand compacted by vibration.
of the probe to the desired treatment
Sinking
depth is L-"rs-l
Zh
usually accomplished using vibratory methods,
often supplemented by water jets at the tip.
lnjection of air at the same time has been found

a7
to facilitate penetration to large depths. Up-
ward directed $tater jets along the sides has
also been found helpful in some cases. Soil
gradations suitable for densification by vibro-
óompaction are indicated in Fig. 1. Compaction
formed by these methods (o) DoubleTube Rod (b) Rod Wilh Proleclrves
piles of sand and gravel
áre also used in soft cohesive soils, in which
case they function as compression and shear
reinforcement, as discussed in a later section
Fiq. 3 Vibro-Rods Used for Sand Densificat:c:.
of this report. Ground treatment depths of 20 n
can be achieved routinely by these methods. (from Saito, l9--
Depths in excess of 30 m can be attained in some
cases.

A brief description of some of the more exten- 2. Vibroflotation

sively used vibro-compaction methods is given
below. This method was developed in Germany alrrcs:
50 years ago, and its development has cc:'.-
1. Vibrating Probes tinued there and in the U.S.A. \^¡here it Y.s
introduced in the 1940's. The equipment
The Terraprobe method, developed in the consists of three main parts: the Vj.bra--:: '
(Anderson, L974), uses a Foster Vibro- extension tubes, and a supporting crane. -¡
U.S.A.
driver pile hamrner on toP of a 0.76 m dia. schematíc diagram of the equipment and
open tubular probe (pipe pile) that is 3 to process is given in Fiq. 4. The vibrar'c: :¡
5 m longer than the desired penetration a hollow steel tube contai.ning an ecceri,t:::
depth. The unit operates at a freguency of weiqht mounted on a verticaf axis in the

514
 (i) ( ¡i) (¡ii) (iv) (v) (vi) (vii)

5!PPtY OF
GIA¡ULAR
sorL rAft

f,

aj.g. 4 Vibroflotation Eguipment and process

Fig. 5 Construction of Compaction Piles by

lower part so as to give a horizontal the compozer system
vi-bration. Vibrator diameters are in the
range of 350 to 450 mn and the Lenqth is
about 5 m including a special flexible vertical vibration of a vibratory driving
coupling. One vibrator weighs about 20 kN. hammer and the horizontal vibration of a
Units developing centrifugal forces up to ViJ.ot depth compactor. The Vilot is a
160 kN and varj_able vibration amplitudes of special probe of about the same size as
up to 25 ¡nm are avail_able. Most usuaL vibroflot units. Sand backfill is used, but
operating freguencies are 30 Hz and 50 Hz. water is not used in either the sinkinq or
The extension tubes have a slightly smaller compaction process.
liameter than the vibrator and a length
Sependent on the depth of penetration The gradation of both the in-situ soil and the
reguired. backfil1, which may or may not be the same
'.':broflot material, influence the leve1 of improvement that
sinking rates of I to 2 m,/min and may be obtained. Coarse sands give greater
;:thdrawal,/compaction rates of aboui densification than fine sands, evidently because
,'.3 rn,/min are typical. water pressures of the coarser material is better able to transmit
j.: to 0.8 MPa and flow rates up to 3,OOO vibrations. Brown (1977) has defined a suit-
:'min may be used to facilitate penetration. ability number for vibroflotation backfills that
j¡:ld backfill i s c o n s u m e d a t a r á t e ñ r rq rY ñ +L v^
'
is given by
i -32- ¡"-r-^-*;- ; h e :c: o
; : ;m
^ ;p
; ^a
: ction ::::,v+
process. The
z:ne of improved soil extends from 1.5 m to -ñil;;;---, ,--
{ ::t from the vibrator, = 1.7 +?6i.T, +#tT, (3)
-,-.'pe and vi_brof 1ot por4rer.
depending upon soil /l5:-T,
r.-50, r"2O' ,"10,
Additional details
r:e presented by Baumann and Bauer (1974),
:eL1 ( 1 9 7 5 ) , and Brown (I9'?7), among others. in which D5O, D2O, and D16 are the 50, 20, and 10
percent size grain diameters in mm. Correspond-
: - g!!r9:ggrpgsr_-{9999 ing suitability numbers and backfill ratings are
:.-.is sand compaction pile method was deve- - 10
0 ExcelLent
-::ed by Murayama in Japan in 1958 (Murayama, - 20
10 Good
-J;8). The apparatus and procedure used in 20 - 30 Fair
::e compozer system are shown schematically - 50
30 Poor
-:. Fig. 5. A casing pipe is driven to the > 50 Unsuitable
:i-srred depth by a vibrator at the top. A
::::3 charge is then introduced into the The lower the suitability number the faster the
: ::e, the pipe is withdrawn part way while vibroflot can be withdrawn whife stil1 achievinq
::::ressed air is blown down inside the acceptable compaction,
-rs:rg to hold the sand in place. The pipe
-: '.'ibrated down to compact the sand pile The influence of fines on the level of improve-
.-.i enlarge its diameter. the process is ment that can be obtained by vibrocompaction is
:::eated until the pipe reaches the ground shown clearly by the data in Fig. 6 (Saito,
:^j::ace. The resulting pile is usually 600 19'7'7\. The data provide
- - 800 mm in diameter. excellent support for
The actual diameter the rule of thumb that vibrocompaction is in-
'::. be estimated from the sand volume dis- effective in soils containing more than 20 per-
::.¿:sed into the ground. cent fines. Bhandari (1977) describes a case in
which compaction piles 1ed to soil densification
i. Soj,1 Vibratory Stabilizing Method only in zones where the fines content was less
than 20 percent.
:::s method, termed both the SVS method and
':.! Toyomenka method, comb.ines both
the

515
 40 ¡n(r + eo)
s = r.or \t" u
u (s)
\_-=- )
for piles in a triangular pattern,.Fig' 7(b) ' in
a
whici¡ d is the sand pile diameter (up to 800 mm)'
.After iml)rovement An influence curve method that can be used to
o,30 estimate vibroflotation probe spacings to give
! a specified minimum relative density is given by
a\t
srown (L97?). Brown's curves are based on the
2 o procedure developed by D'Appolonia (1953) '

8 2 0
3
o \
I

= t
ta\
O 1- @ I
@
I

o
o
-
ao oa
o . \.t . tI 3ond pilo
l - { - a-a
-r---t-,-----: L
T @ @-
(o) Squorc Pollorn (b) TriongulorPotlarn
\ l
Before improvement
0 20 Usual VibrocomPaction and
7, ELg. 7
Compaction Pile Patterns
Finerfraction: <74tt

Empirical. design curves for vibrocompaction have

F i ^ A Effect o f F i n e s Content on Increase
in Penetration Resistance by Vibro- beán developeét-by Thorburn (1975) . Fig' 8 is a
relationship Uetween the relative density of a
cornpaction
(from Saito, 1977)

In some instances penetration resistances are so

hiqh followinq denéification by vibrocomPaction
thát relative densities greater than 100 percent
are indicated according to conventional corre-
]ations. In reality, however, this result is in c
most ínstances causéd by the increased lateral o
ct
pressure induced by the vibration Process.
in-situ lateral préssure measurements reported o
q
by Saito {[g77) yietaea values of Ko as high as I

6 in some zones after densificationi

at
Vibrocompaction of large areas is done in a grid c
pattern, either triangular or rectangular, with q,
1.5 m to o
-3r o b e s p a c i n g s u s u a l l y i n t h e r a l g e o f
p
m on éenters. The actual spacings depend on o
the soil and backfill types, probe type and .:
energy, and level of improvement required' o
althóüqh field tests are usually done to fina- o
lize désigns, there are some guidelines that can É,
be useful in PreliminarY studies. o
.o
If it ís desired to increase the average density o
of loose sand from an inítial voiti-TáEi6 e o t o a l¡
e, and if it is assumed that instal- o
void ratio
Iation of a sand pile causes compaction only in o-
a lateral direction, the pile spacings may be
determined using

/¡(r +.^)\r/2
'=\*q=;) d (4) 2 3
Spocingsof Cenfersof Vibrotion-n

Fío. 8 Relative Density of Clean Sand at ?.::'J

for sand piles in a square Pattern, Fig. 7 (a) , Midway Between Centers of Vibratic:: lr I
and Funct-ion of Probe Spacing (Thorbun, l9'!

516
clean sand at points nidway bet!,¡een the centers ffhen coarse sands or gravels are used as back-
of probe points and probe spacing. Any relation- fill ¡naterials for vibrocompaction, the resulting
ship such as this will depend to some extent on pil-es can act as drains because of their high
soil characteristics, especially particle size, hydraulic conductivity relative to the sur-
water table conditions, and the amplitude and rounding sand. As such, they nay serve to dis-
frequency of the vibrator. A relationship sipate pore pressures from potentially lique-
betr^reen allowable bearing pressure to limit fiable deposits and thereby prevent liquefaction.
settlements to 25 mn and probe spacing is given A method for designing gravel drain systems for
in Fig. 9. Here again it must be realized that this purpose is given by Seed and Booker (L977).
for the same reasons such a relationship is It is important to note in such systems that the
suitable only for preliminary design. use of drains to prevent liquefaction under
seismic loading does not eliminate the potential
for settlement.

600 6 Heavy Tamping

N
c
Soil compaction by heavy tamping involves
z repeated dropping of heavy weights onto the
s ground surface. The method is also termed
I
c dynamic compaction, dynamic consolidation, or
c pounding. The technique as developed in its
|'f. present form for improvement of large areas to
N
itepths of up to 30 m was pioneered by Ménard
I (Ménard, 19i4; uénard and Broise, Lgisl. when
applied to partly saturated soils, the densifi-
E 4oo cation process is essentially the same as that
o
for impact (Proctor) compaction in the labora-
(n
o) tory. In the case of saturated cohesionless
(\l
soils liquefaction can be induced' and the
E \ densification process is similar to that
o and vibrocompaction. The
J c accompanying blasting
o effectiveness of the method in saturated, fine-
o
o
grained soils is uncertaini both successes and
failures have been reported. It vrould appear
that in such materials a breakdown in the soil
(D ¿UU
F structure, the generation of excess pore water
pressures, and the formation of drainage
ol Heavy
channels by fissuring may be required.
- tamping has been especially effective for com-
q) paction of hraste and rubble fills. Among
(D
recent references which provide details of the
o method and case histories are Bhandari (1977) ,
Á BoZko et aI. (1976), Charfes (1981), Hansbo
,
o
( L 9 7 7 , ! 9 7 8 , , K r u t o v e t a I . ( 1 9 7 8 ), L e o n a r d s
e t a I . ( 1 9 8 0 ) , L u k a s ( 1 9 8 0 ) , M é n a r d ( L 9 7 4 ),
o Ménard and Broise (f975), Minkov et aI. (1980),
t 2 3 Zhenget aI. (1980).
S p o c i n g so f C e n t e rosf V i b r o l i o n - m

The pounders used for heavy tamping may be

iig. 9 Allowable Bearing Pressure as a concrete blocks, steel plates, or thick steel
Function of Probe Spacing for wj.th concrete or sand and may
shells filled
Footings Having Wi¿lths Varying
range from one or two up to 200 tons in weight.
from one to Three Meters Drop heights up to 40 m have been used. The
(Thorburn,1975)
pounders are usually square or cj.rcular in plan
and have dimensions of up to a few meters
depending on weight required, material, and the
-:<e blasting, vibrocompaction can loosen very dynamic bearing capacity at the surface of the
:e:se zones and break up weakly cemented layers. ground to be treated. More streamlined shapes
r.---hough the strength of such zones will there- have been used for underwater tamping.
:::e be decreased, they will usually still be
-.:if :.cient1y dense and strong for the project at
-::.j. Also, as observed for sands densified by For large area compaction several repetitions at
r -3sting and by heavy tamping, the penetration points spaced several meters apart in a grid
rs:Stdnc€ may increase wlth tlme after treat- pattern are applied. A typical treatment will
''.:.--, even though densificatlon and surface set- iesult in an áverage of 2 lo 3 blows/m2. An
'-¿:ent are essentially complete at the end of illustration of a typical grid pattern and
':.t:::ent. Experience at the Kwinana Termlnal ln representative equipment is shown in Fig. 10.
r.:s'-ern Australia (Contracting and Construction Tvro or three coverages of an area may be
:-::::eer, 1974) is a good case in point. At required, separated by tirne íntervals dependent
.':s site 12,500 vibroflotation probes were on the rate of dissipation of excess pore \úater
-:i: full depth into a sand deposit about 25 m pressure and strength regain. The general
:-.-:k. Cone penetration resistance values response of the ground as a function of time
--::eased by I0 to 15 percent over the three after a coverage is shown in Fig. 11. The time
.e¿..s f ollor^¡ing treatment. interval reguired between coverages may range

517
 , l
),¡;'-uÁ1
s\
t , , I

o Applied energy in tm,/m'

l^¡ith time
@ Volume variation
O Ratio of pore-pressure to initial
effective stress
Fig. 10 Illustration of Heavy Tamping Variation of bearinq caPacitY
@
Time between passes varies from one to four r^teeks
coarse sands to according to the soil tYPe.
from days for freely draining
weeks for finer-grained soils. The ground sur-
face is usually levelled betl^teen coverages. To Fi9. 11 Ground Response with Time After
insure uniformity anil high density in the near Successive Coverages of DYnamic
"ironing" is used. SmaII
surface zone, surface Consolidation
impacts by the pounder are made over the entire
(tqénard and Broise, 1975)
surface. Surface settlements may be from two to
five percent of the thickness of the zone being
densified per coverage.
achieved. ¡¡énard and Broise (1975) proposed
"effective
Zheng et al. (1980) determined the that
deformation," defined as the volume of crater
less the volume of adjacent heave from dis- o = lñtt- (6)
placed soil, for successful dynamic consolida-
clay and a uni- D = maximum depth of influence, m
Lion of a soft to medium stiff where
forn fine sand. In the clay the deformation was w = falling weight, metric tons
30 percent effective, and in the sand j-t \"tas H = height of drop, m.
62 percent effective.
Leonards et al. (1980) analyzed seven cases and
When heavy tamping is used to prepare ground for concluded
support of relatively light (low rise) struc-
treatment is some- o=t./zffi (7)
turés on shallow foundations,
times made only at footing locations. This can
be an economical and effective means for mini- vras more appropriate, and Lukas (1980) concluded
mizing total and differential settlements. that
when this method is being D = (0.65 ro 0.80) ,6-H- (8)
Of particular interest
considered are the depth of influence and the
level of property improvement that may be

518
eas best suited for the eight cases studied by results of heavy tamping on soils ranging from
:im, silts to clean sands to rubble fills. The
point referenced to Guyot and Varaksin (1980) is
llearly, the depth of influence should depend on for a strip mined waste pile of clay, silty
:actors in addition to the impact energy. Soil clay, sandy cIay, and boulders of limestone and
--ype might be expected to be the most important. shale. From these results it appears that the
¡. crane drop is less efficient than a free drop. use of equation (7) would provide a conservative
:he presence of soft layers has a damping in- estimate of the effective depth of dynamic com-
:iuence on the dynamic forces. Definition of paction achieved, in most cases.
:epth of influence is itself subjective and
:epends both on the method of measurement and The amount of soil improvement that develops in
:re engineer's definition of what constitutes a any case depends on soil type, water conditions,
-easurable ground improvement. Within a homo- and input energy per unít area. Finer-grained
:eneous soil layer the amount of ground improve- soils cannot be strengthened to the same level
-ert decreases with depth as shown, for example, as can coarser materials. Soft layers of clay
:. ELg. r¿- and peat inhibit high compaction of adjacent
cohesionless material because of their flexi-
bility. A review of available cases suggests
that there may be a definable maximum level of
F R E N C H R I V E R A A I R P O R TI N N I C E irnprovement. Leonards et aI. (1980) suggest
IMPROVEMENTOF IHE RECLAIMEDLANO Fffi
T H E P R O P O S E ON E W R U N - W A Y
that this level may be a^cone penetration resis-
tance of about 150 kq/cm'. A study of the data
associated with the cases plotted in Fig. 13
E M o d u l u s( t s f )
shows maximum values of cone penetration resis-
tance of 180 kg,/cm2, standard penetration
resistance of 45 blows/0.3 m, pressuremeter
limit pressure of 3 MPa, and pressuremeter
modulus of 25 MPa for clean sands. Finer-
3'mulor F¡ll grained more compressible soils may have maximum
-rh 30f¡€
t . poctelS values that are less than half of these. Values
decrease from a maximum near the ground surface
to the original in-situ value at alepth D.
tO / mer
vud Lino
An additional concern relative to heavy tamping,
and blasting as wel1, is whether damage may
occur to facilities located beyond the edges of
the area being densified because of the large
impact energies. Measurements of vibration
frequencies have given vafues in the range of
2-20 Hz. Particle velocity is related to scaled
energy factor in Fig. 14. The lines for wet
sand, dry sand, and clay were given by Vliss
(f967) for prediction of particle velocities
resulting from pile driving operations. The
line for building rubble (construction debris)
compaction by heavy tamping was developed by
Lukas (1980), and the additional points are for
heavy tamping of sand. These latter points plot
near the l-ine for rubble and suggest that a
given scaled energy produces a somewhat lo\ter
particle velocity than the same energy input by
pile driving. Fig. 14 may be used to estimate
the distance from point of impact where damage
L09sno could occur.
o Befor6 dynomic conlolidot¡on
. Aftar 3 posle3 of dynomic consolidotion
Progressive Liquefaction
(€och Dos! cdr3i3f3 of 6 blow3 of o In many cases the volume of soil densified by
t7O ton hommsr tulling f rom 23 m) deep compaction lies within a potentially lique-
fiable deposit of much larger areal extent. The
question aríses then concerning whether, if in
" :. 12 Variation in Pressuremeter Modulus and an earthquake the surrounding soil liquefies,
Limit Pressure with Depth at Nice there will be the possibility of loss of stabi-
Airport lity in the densified zone. Conceivably, the
(Courtesy M. Gambin, development of high pore pressures in the lique-
Techniques Louis Menard) fied zone could generate higher pore pressures
in the densified zone with consequent loss of
strength. To guard against this possibility it
--¡er of field experiences have been sunma- should be sufficient to extend the zone of soil
:':: :n Fj.g. I3. Because considerable judgment
improvement Iaterally outhrard from the required
: :e;uired in assessing the depth of ground foundation area a distance equal to the thick-
::'.'enent in a number of cases and because ness of the layer being densified. More precise
::e:ent methods for evaluating improvement evaluations can be made using a computer
::' :s€d on different projects, the data in analysis of pore pressure generation for the
:. ^f are not precise. Nonetheless, the trend given conditions as described by Booker et al.
---ear. The plotted points represent the (1976). Alternatively, the method proposed by

519
 30 o Predominontlycohesionlesssoil
o S i l t y s o i l s ,f i l l s , r u b b l e t/to
t Full depth of improvedzonenot determined

6,'
\¿/

E 20 .7 )t

.9
I

(I, /
o ,/o3
c
o
References
f
I Leonordset ol. (1980)
c
z Bhondori(1977].
o 3 Gombin(1979)
c ¿rtu^rr"
o |_ 4 Gombín(l9T9FPersonol Cor
o
o ro ,9'.' a' s
6
GuyoiondVoroksin(tgBO)
Voroksin(1980)
.,, ){ .
z . J u l t i é( 1 9 7 8 )
Cr)'d^j. e Minkover oL(1980)
e L ü k o s( 1 9 8 0 )

0
o 20 60 80
- (t-m)E

Fj-q. 13 Depth of Influence as a Function of Impact Energy for Heavy Tamping

Nandakurnaran et al. (1977), which appears guite conclude that evaluations of the ground shortly
conservative, might be considered. after the completion of deep densification will
give conservative results.
time Egectg.
More and more evidence is becoming available to BY PRECOMPRESSION
SOIL TMPROVEMENT
indicate that tn many sands time-dependent in-
creases in strength and decreases in compressi- Introduction
bility develop after densification by any of the
deep compaction methods. Because these effects The strenqthening and Preconsolidation of weak
continue over periods of many weeks or months, and compressible soils by preloading prior to
they cannot be explained in terms of pore pres- construction is one of the oldest and most
sure dissipation, which continues only for peri- widely used methods for soil improvement. It i3
ods of several mlnutes at the most in the case particularly well-suited for use with soils that
of clean sand. The aging effect has been shorrn undergo large volume decreases and strength
to give substantial increase in the strenqth of increases under sustaíned static loads and r¡hen
sands under cyclic loading, as may be seen in there is sufficient time available for the re-
Fig.15. quired compressions to occur. Surcharge loads;
i.e., loads in excess of those to be applied by
Although a number of hypotheses have been a permanent fill or structure can be used to
advanced to explain this behaviori e.9., thixo- accelerate the process. When the anticipated
tropic hardening, chemical cementation, the time for compression is excessive, vertical
effects of dissolved gases, the mechanism is not drains may be used to shorten the time required
yet comptetely clear. From a practical stand- provided the compression is of the primary con-
point, however, it would be reasonable to solidation type. The soil tyPes best suited for

520
 o G i g o n( 1 9 7 7 ) fully to improve the foundation soils for build-
ings, embankments, highways, runways, tanks, and
elol.(f98O)
a Leonords, o Lukos (1980) abutments.
2fr
Several recent comprehensive treatments of pre-
t25 compression and topics related to its applica-
P l o s l e r c r o c k s( r e s i d e n c e s ) tion, with and without drains, are available,
¡ q, including those by Johnson (1970a, 1970b),
I
o R e c o m m e n d e ds o f e l e v e l o Bjerrum (1972), U.S. Navy (1971), Pilot (L977),
(residences) E Schlosser and .furan (1979) , Akagi (1977, 1979) ,
g E Hansbo (1979). The state-of-the-art by Pilot
25
; (1977) is especially complete and well-documented
= and contains an extensive list of case histories.
t ? . 5 ^9
o
Because of these references, detailed treatment
o
o of all topics important to precompression is not
q)
l'
'rt F unpláosonl g
given here, Emphasis is on recent developments
./ It and practical considerations that influence
9 ^-t'^.^'f
*
noircüue ; design and performance of precompression systems.
- . /tttf
- .rt .r'-
-t/_ 2.5
rt ./t liior',T turcepfibre Types of Preloads
Although earth fills are the most commonly used
type of preload, any system that leads to drain-
age of pore water and compression of the soil
may be suitable. water in tanks has been used
o.25 to preload small areas, and water in lined ponds
can be used to preload larger areas. vacuum
I ro 40 preloading by pumping from beneath an impervious
/Eñ4il?ii6i/ (ft)
Drstonce membrane placed over the qround surface can pro-
duce surcharge loads up to 60 to 80 kPa (Holtz
and Wager, 1975; Pilot, L977). Anchor and jack
systems can be devised for special cases.
Fig. 14 Particle Velocity as a Function of Groundwater lowering provides an increase j.n
Scaled Energy Factor consolidation pressure equal. to the unit weight
of water times the drawdown distance. Consolida-
tion by electro-osmosis is the same in many
respects as consolidation under externally
applied stresses, except that the driving force
T i m a o ft 3 r ' D a p o 3 ¡ i i o-ny a o r ! for drainage is induced internally by an electri-
t o
I to too to@ cal field.
EIE
7l¿1
o t >
&tü O Loborolorylall dolo - lúoloray tlo.O roñd Preloading by vacuum, water table lowering, and
E t o
otx
A Hyd.oulic aond fill lrco l.ñr SonFcrñondooqr electro-osmosis offer the advantages that there
O Hydroulic¡ond till trñ UpDo.SonF!.nondoDom
s0 tuat E S@lh l.ror ro¡d are no stability problems, and large volumes of
V Son ilol.o ¡ond surcharge fill are not required. on the other
dt8 hand, they are more complex in execution than
Bon.ri...lol. (1979)
EIE (olc.l4rglcm , K.l.o, +.sf %l the other methods.
-tó
atA e In¡liol l¡quafoctioñin lO cyclG
I
i A Basis for Design
Pl^
O l 3 % o ¡ i o l s l r o i n i n l Oc y c l . ¡ Theoretical
!t$
=tr
The usual objective of design of surcharge loads
and duration of their application is to reduce
-els the magnitude of settlement after construction.
5tF Settlement at any time may be expressed as
:Ió '
,l@
8l: "t=si*Ú".on"*"" (9)
;. 't t oE
olo
El-s = settlement
J l o
in which "t at time t
l,l c
-doy3 = immediate settlement (from 9as
Tim¿ of ler Deposition f
compression, shear distortion)

ú = average degree of consolida-

Fig, 15 Tnfluence of Period of Sustained
tion
Pressure on Stress Ratio Causing
Peak Cyclic Pore Pressure Ratio = final
-sc o n s consolidation settlement
of 100t
s, = secondary compression settl.e-
s
ment.
::provement by precompression are saturated soft
:i.ays, compressible silts, organic clays, and
:eats. Vertical drains are of qreatest The objective may be either: (1) to determine
¿Ífectiveness in inorganic clayé and silts that the magnitude of surcharge pressure (po) re-
¿xhibit little secondary compression. Preload- quired to insure that the total settleñent
::.g and precompression have been used success- ánticipated under the final pressure (p.) will

521
 indefínitely, then a time-settlement curve could
be complete in a given length of time, or (2) to as shown
de:ermine the length of time reguired to achieve be obtained- in the same way and plotted
by the lower curve in Fig..16. If the surcharge
a given amount of settlement under a given sur-
charge load. wére left in place until time tsR, then the
layer will have settled an amount equal tÓ that
It is convenient to consider primary consolida- to be expected under the permanent fill alonei
tion and secondary compression as separate i.e., s¡i = s¡. At this time the layer will
phases, even though both types of volume change have reáóhed án average degree of consolidation
may be in progress concurrently. rf the time of dr* given bY
loading is short relative to the total consolida-
S5
tion period, then the asswnption of instantan- ii- S R (10)
eous application of the total load at the end of sf+s
one half the loading period will not lead to
serious error. If not, then more exact analy-
As pointed out by Aldrich (1965) and extended by
ses may be made.
iórtitson (1970a) the distribution of effective
Consolidation settlement estimates for permanent and excess pore lrrater pressures before and after
fill and structure loads are made in the usual surcharge rámoval will be as shown in Fig' 17
rday. rn those cages where soil properties and/ for a ciay layer drained at both boundaries '
or stress conditions vary with depth, it may be Thus, a sübstántial portion of the layer may
necessary to analyze the profile as a series of uncteigo further consólidation after removal of
sublayers. the súrcharge load, I^thereas the remainder will
be unloaded. although the unloading of the
zones near drainage boundaries will generally
The time rate of settlement for primary consoli-
not lead to signiiicant heave, the additional
dation in one dimension may be determined using
the classical Terzaghi theory. In those cases consolidation in the central portion may be
where the straín profile with depth is not important.
essentially constant, as assumed in the Terzaghi
solution, a non-linear stress-strain method,
such as that presented by Duncan and Buchignani
(1976) based on Janbu (1965) may give better
results. The effect of decreasing vertical
strains with clepth is to increase the rate of
settlement compared to the Terzaghi solution.
The stress vs. time and settlement vs. time consol¡doted
relationships for a soil layer under stress in- fu r€moved
crease p¡ would be as shown in Fig. 16 in the
abgence óf secondary compression.

If a surcharge load that applied an additional Undarcoñsol¡dot8d

stress ps were also used and left in place ofter p3 removed

o
a
o
o,
()
l
c a;+ %'* p¡ co'f Pl.,

E
o (l) Poreprs!3uradi3lribut¡onof t co¡rssPonding
to U¡R
q €) fure prsssursdi3lributionroguircdfo prevenl
furthorsetllem€ntif p, removgd

Fig. 17 Pore Pressure Distribution after

Removal of surcharge Load at Time
+-S R
G
(¡,
E
(u To eLiminate further prirnary consolidation
following the removal of the surcharge, the
(l) surchargé should be left in place until the pore
a pre"suré at the most critical point-; i.e', the
point that is last to consolidate, has itself
ieacheil a consolidation ratio U, given bY

P¡
/' " " r' f *r" \ = - (Ir)
Pf * P"

Fig. r6 Compensation for Primary Compression

Using Surcharge Loading

522
Times corresponding to USR and Uo are found Vertical Drains
using the Terzaghi theory and thé coefficient of
consolidation c__ according to In many cases the time reguired for surcharging
v
is excessive, the surcharge required for the
Tuz time available is too great, or the rate of
t = : (L2') strength gain is too slow to permit rapid fill
v placement. For soils whose compression is domi-
nated by primary consolidation, vertical drains
where T is the appropriate value of dimension- may be used to accelerate the rate of settlement,
less time factor. because consolidation times vary as the square
of the drainage path length, and because most
Because this approach requires that the sur- deposits have greater permeability in the hori-
charge be left in place until all points are zontal than in the vertical direction. A sche-
fully consolidated under po, overconsolidation matic diagram of a typical vertical drain
will devefop in most of thé soil layer. Thus installation is shown in Fig. 18, Vertical
it is conservative, and the actual precompres- drains are ineffective in peats, organic clays,
sion settlement will be significantly greater and other soils whose settlement behavior is
than s€. A further complicating factor arises dominated by secondary compression.
in that equation (11) is stress-based; whereas
c., is usually determined using a strain-based
procedure. As the stress-strain behavior of the cl.rt
seTfLcuthT
soil is non-linear, there is some uncertainty in
ftUDae.cT
the predicted times and surcharge loads required. t.eC¿ rttt
this problem is presently under study at the
ccet.J!¡tf
University of California, Berkeley, and preli- Ít,L

mj.nary indications are that in its present form

the method gives conservative results¡ i.e.,
higher values of p. or tine of surcharging are
predicted than actüally required.

Secondary compression may represent a very signi-

iicant portion or the total compression of some
soi.Is, especially organic clays and peats. pre-
,-ompression using surcharge loadings may be ''ae5
effective for minimizing the effects of subse- tt tl
SOt.l ¡ COr¡P¡t3SrlLt SrOü Oear¡,¡c
;uent secondary compression under permanent
loads. The concept is to estimate the total
settlement under p¡ as the sum of that due to
:rimary consolidation and that due to secondary
:.:mpression, ssec, anticipated to occur during
::e life of the structure. s"." is determined
::cording to
) l l Ftet SOtL
rsec = Co Hp log (t,/tp) (t > tp) (13) SANO OP/lN \
Dc,

-:. which Co is the vertlcal strain per 1og cycle Fig. 18 Typical Vertical Drain Installation
-r.:rease in time subsequent to the end of prl-
-::1' consolidation at tpr and Hp is the layer
'-.:ckness at time tp. The analógous equation
' r (II) for the cri-tical point The theory for consolidation by radial drainage
for thls case l-s
and by combined radial and vertical drainage is
welL developed (Barron, 1948¡ Carillo, l - 9 4 2 ).
The results are summarized in Figs. 19 and 20.
s€ + C^ H^ 1o9(t,/t^)
ft is i¡nportant to note that drain spacing is a
I f r
' " z tI f + s -
- s

s -
f y
(r4) far more significant parameter in the determina-
r+s tÍon of consolidation time than is drain diarneter.
Olson et al. (L974') have developed a finite dif-
-.e nature of secondary compression is such that ference computer program that can be used to
---e time after surcharge pe is removed, secon- investigate the effects of time-dependent load-
r:'/ compression will reappéar under P¡. This ing, variable consolidation coefficients, non-
::ect is smaIl, however, and can usually be linear soil stress-strain behavior, the in-
.¡;lected (Johnson, 1970a). fluences of larqe strains, and the installation
of drains after one-dimensional settlement has
-e increase in undrained shear strength which begun.
:c::panies preloading may be the most important
::ect in many cases. The undrained strength Until a few years ago vertical drains of sand,
':er a given duration of preloading may be esti- typically 200 to 500 mm in diameter and spaced
::e3 using the principles o f t h e S H A N S E Pp r o - anywhere from 1.5 to 6.0 m on centers, were
::re (Ladd and Foott, 1974) together with an widely used. Installation was accomplished using
:-i'sis to estimate the increase in effective a variety of techniques of both the displacement
-.sclidation pressures withín the preloaded and non-displacement type. Displacement drains,
while generally less expensive and faster to

523
 install, can disturb the surrounding soils' The
r.--j------{ a. ,l iésultiirg "smear" zone can impede,drainage, and
¿ wELL a t-- soil may be weakened. These
tñc¡¡G the disturbed
effects may not be as detrimental as earlier
,O FLO,
:--r-r
I believed, irowever, owing to the possibilities of
.cEoss ¡ H (1) reconsolidation to a higher strength than
oUfEP -_ t the original, and (2) the opening of.cracks and
SUPF'CE EX l^'ith sand during installation
-\ fissureá tha¿ fill
and thereby increase the effective drainage area'
.{ I Akagi (197i , Ig79l concluded that reliable data
a
-¡'bfl are lacking to establish whether non-displacement
n =d¡5 *ra-_ drains are indeed more effective than the dis-
d r = l . O SS
placement tYPe.
s. PLAN OF DRAIN tttLL PAfTERN I sEcrto?{ A-A
Present indications are that conventional sand
drains instalfed for the acceleration of con-
solidation may soon be things of the past, as a
variety of préfabricated drains are coming into
vERTICALcoatOLlOaTlo. : wide uie. band-shapeil drains of the order of
100 mm wide by 1 to 7 nm thick are produced by
TrHl several manufácturers. These drains can be
.- Y - {t r il . c" o-r - rH 7
l r vr , * t c * rapidly installed to depths up to 50.m by
rn.óftirré" with special mandrels. Drain spacings
ái tn" order of- 1 m are typical. Both dynamic
I and static methods of installation a¡e used'
RADIALCOi9OLlDATl0rl
ci;.;i;;-;táiabricatecl drains or wicks comPosed
T¡dl of sand within cylindrical fabric contáiners are
kh
c,+'tf
il - ?or clT¡
i, oR | ' c,* also used. A detailed discussion of these new
árains, their specific features, and methods of
instaliation is beyond the scope of this report'
Hansbo (1979) has presented a comprehensive
CO¡Eltl€ORADIALArO VERTICaLFL(t' using prefabricated
overview of consolidation
A f A X YT l r E r band-shaped drains. Only a few general char-
acteristics are summarized here'
- /l\ fl) ^T ^ PorxT
PoeEI \". / "-n \ u o l " \uo,/¡ K'iellmanrs cardboard wick, introduced in 1937,
ExcEss wás the first of the prefabricated drains '
f^rER PRESSURE
)
RATros Presently there are several types-on.the market
a v E R ^ vc ^EL u E s under suóh names as Alidrain, Geodrain, castle
(;)"-,'(;), (;), The
) Ááita, Colbond, Mebradrain, and PVc Drain'
clrainá usually consist of a core of plastic and
a filter sleeve of paper, fibrous material, or
u . l -u o' AT A POII{T porous plastic. The cross section design pro-
v:.aes ¡ór a system of vertical channels for
water floh¡. lhe Colbond Drain differs from the
ú-I--l AvERAGEVALUE in that it is a non-rdoven fabric through-
uo áth.r"
ár-ri, .tt¿ it is signif icantly wider (30 cm) than
the others. The mandreLs used for drain
t9 Summary of Consolidation Theory installation are of cross section considerably
Fig. Fig' 2l is an
for Vertical Drain Design larger than the drain itself.
exaáple. Thus installation of the drain neces-
sariiy causes some disturbance of the surround-
i -^ ¡¡arr¡A
¿¡¡Y Yrvs.¡Y.

f
o
6
l

ñ
E .n'lO
0 7
o
3
t
() n = 6"/dr1
o - --
o Vert¡colf low
o -
(tr Rodiot f tow
o
6
o,ol o.l
Time Foctor, T" ond T,

Fiq.2t Cross Section of a Plastic Drain

Fig.20 Consolidation Solution for Radial Flot¡t
and Equal Vertical strain at Ground and Mandrel
surface (earron, 1948)

524
The filter sleeve surrounding the plastic core dimensional deformations and drainage conditions
must satisfy several criteria: are more complex than the one-dimensional pro-
cedure described herein. Cornputer analyses can
l. Its permeability should not be signlficantly be developed for special cases. Some solutions
less than that of the surrounding soil. are available for determination of approximate
or limiting cases.
2. Fine soil particles should be retained to
prevent clogging of flow channels in the If the compressible layer thickness H is large
core. relative to the wiilth of the loaded area B, then
3. The filter should be stiff enough so as there \.úi11 be an immediate settlement sr in
not to be pressed into core channels by Equation (9) . Its magnitude can be estimated
high lateral soil pressures and strong using elastic theory. The consolidation settle-
enough not to be damaged during instal- rnent, saorr" in these cases may be reduced depend-
lation. ing on the magnitude of the pore pressure coeffi-
4. The filter should not undergo physical, cient Á. Estimation of consolidatíon settlements
chemical, or biological deterioration according to the procedure of Skempton and
during the intended Iife of the drain. Bjerrum (1957) may be appropriate. Lateral
dliainage becomes important in accelerating the
Theoretical determination of required drain rate oi consolidation when the value of H/B
spacing can be made in the same lttay as for sand exceeds 0.25 to 1.0, depending on the shape of
Crains; i.e,, with the aid of Figs. 19 and 20. loaded area and whether the compressible layer
.\Iternatively, if consolidation due lo vertical is singly or doubly drained. The rate of settle-
:low is negligible compared to that due to ment will be even greater if the horizontal
radj-al flow, which is often the case, then the permeability, and therefore c'-¡, is greater
:ime of consolidation t can be determined than the vertical permeability and crr-.\r. A
accordinq to quantitatíve estimate of both of these effects
óan be made using the solutions developed by
d-tu Davis and Poulos (L972).
r
+- = _:- -'-
|n -----i-- (rs)
8".r_h I _ ün vühen vertíca1 drains are used, the coefficient
of vertical consolidation due to horizontal
flow, co,-¡, controls the rate of consolidation.
.:.ere Ú¡ is the average degree of consolidation.
::e parámeter l, is given, to a good approxima- Both special laboratory tests on large samples,
::,on, in terms of n(=d"rzd.) bV e.g., Hansbo (1960), Rowe (1964), Berry and
w i Í k i n s o n ( 1 9 6 9 ) , a n d P a u t e ( 1 9 7 3 ), a n d i n - s i t u
tests, summarized by Mitchell and Gardner (L975),
u = ln (n) - 0.75 (16) have been used to determine c,,-". A major dif-
ficutty in laboratory testing'fór deter¡nination
of c,,-" is that very large samples are required
::r the values of n (>12, except for colbond,
(>0.8 m) used in practice. in máni cases if results representative of the
3) and drain spacing
diameter do, of a band-shaped true in-situ stratification are to be obtained.
::e equivalent
t is taken as
An approach that has yielded reasonable results
::ain of wiilth b and thickness and compute cn-¡
is to measure k¡ in the field
using a compressibility value determined by con-
- 2(
= --;- b+r) (17) ventional láboratory tests. A typical condition
o*
is that .rr_h/.rr_., = 2 to 10.

:::h well resistance and disturbance during Precompression by Electro-osmosis

-:.stallation may cause the actual tirnes for con-
Because vtater can be made to flow through fine-
;:!rdation to be greater than predicted by
grained soils from anode to cathode in a direct
:--i. 20 or Eq. (15). Both of these considera-
éurrent electrical field; i.e., by electro-
::cns are discussed by Hansbo (1979). osmosis, consolidation will result if water is
removed at the cathode but not replaced at the
. :efabricated band-shaped drains can be in- soil conditions and limited
anode. For certain
;:alled at orientations other than vertical, may be an economi-
soil volumes, electro-osmosis
..-.:ch enables their use for special applica-
means for consolidation. ff,
ca1 and effective
::.ons such as under-drainage and on slopes.
je3ause they can tolerate displace-
at the same time as water is being removed at
significant chemicals are injected
-e::ts without prefabrícated dralns are
the cathode, stabilizing
rupture,
-:: due
at the anode, soil improvement by electro-
as susceptible to loss of effectiveness can be achi-eved. Electro-
kinetic injection
i r shear displacements as can occur in the case
-: sand drains, kinetic inJection is discussed in a later section'
The mechanism of electro-osmosis has been elabo-
rated by Gray and Mitchell (1967) . The theory
, ,: : e Practrcai by electro-osmosis has been
for conlolidation
.:: precompression without vertical drains the developed by Esrig (1958), Wan and Mitche1l
:r:e of consolidation is controlled by the (1976), anil Mitchell and !üan (1977) . Recent case
:eÍíicient of consolidation for vertical flow, histories are sumrnarized by Pilot (1977) .
Both laboratory and fiefd tests may be
-ieá for its determination, and there is exten- The water flow rate, 9n, in a one-dimensional
.:'.'e literature on the methods, their accuracy direct current field is initialtY
r:.: limitations.
q h = k e i e A ( m 3/ s e c ) (18)
e:tlement predictions for two" and three-

525
where k. = I x lO-e to 7 x 10-s m/sec per volt/m where kn = the hydraulic conductivity

ie = electrical potential gradient (volt,/n) Y' = the unit weiqht of water

t^t
A = cross-sectional area (m2) V = voltage (a function of position)

Alternatively, the flow rate can be expressed by The amount of consolidation associated with this
effective stress increase is obtained from a
void ratio vs. effective pressure relationship
c¡, = k. I ( m3 , / s e c ) (19)
a for the soil determined in the usual manner.
Strength increases, in the absence of electro-
where k, = water flow per unit time per ampere chemical hardening effects, can be estimated in
a the same way.
{m37sec7anp)

I = current (amps) The rate of consolidation is governed by the

same relationships that apply to consolidation
The coefficients k" and kr are related by the under directly applied loading. The time t for
specific electricaL conduótivity, o, a given degree of consolidation is

k, = k.,/o (20) . TL2 (23)

"v

values of o range from about 0.02 mhor/m for low

salt content soils to 0.30 mho/rn for high salt where T = tirne factor for the specified degree
content soils. Values of k¡ for vtater contents of consolidation and boundary condi-
in the range of 50 to 1OO pércent are given in tions
Table III. The power consumption P is given by
L = efectrode spacing
P - qh AVlk. x 10-3 (kwh) (21) c = coefficient of consolidation.
v
values of T for different degrees of consolida-
where Av is the voltage drop. tion for the case of parallel plate electrodes
with a linear variation in voltage betrreen them
are given in Table Iv. Measurements by Johnston
and Butterfield (1977) indicate that rather than
TABLE III a linear variation in voltage betvteen electrodes,
an instantaneously infinite electrical gradient
Va1ues of Electro-Osmotic Water develops at the anode initially which decreases
Transport coefficient in a consistent way, to a uniform gradient at
(water content range 50-100t) the completion of consolidation. values of T for
these conditions are also listed in Table IV. It
may be seen that consolidation occurs more raP-
Pore Water idly for the latter case.
SoiI Type Salt Concen- *,2"!*2.*p
tration (N)
TABLE IV
Silty clay,
lo- 3 I x l o - s- 5 x 1 0 - 7
kaolinite Time Factor for Various Degrees of
Consolidation by Electro-osmosis
Silty clayr
10-2 5xlo-s-lxlo-7 Between Paralle1 Plate Electrodes
kaolinite

Clay
( illitic
IO 3xio- I -6x10- I Degree of
) Time Factor, T
Consolidation
Clay 2xlo-8-3x10-8
(illitic) (r) Linear v Infinite Initial
U Variation Gradient

0 0 n
10 0.05 0.001
During consolidation, the water flow rate de-
creases with tine. It ceases when a hydraulic 20 0.10 0.007
gradient, caused by a decrease in pore r,rater 30 0.16 0.017
pressure at the anode relative to that at the 0.020
40 0.22
cathode, causinq flow from cathode towards
anode, exactly balances the electrically in- 50 0.29 0.05
duced hydraulíc gradient causing flow from anode 0.38 0.07
60
towards cathode. At this condition the increase
in effective stress, Lo' , from that at the 70 0.50 0.10
start of treatment is 80 0.66 0.14
= (k./k¡) yrv (22) 90 0.95 0.20
¡s'

526
 ActuaL electrode arrangements are an array
of consolidation under a.preload or surcharge fill.
rods or.pi-pes, spaced typically at 2 !o 4 Analysis methods for tñese conaitións are pre_
meters in patterns, rather than parallel flat
plates. sented by Wan and Mitchell (1976).
In addition, variations-in pioperties,
especially in the ratio ke,/kh, that áé"éf"p Recent Developments
during consolidation lead to deviations from the
theory (Mitchell and vüan, Lg77). Several recent studies have been directed at the
rñus tfre
values in Table fV can be used onlv as .r, development of nerd approaches to aná appfici_----
approximation. for prediction tions for precornpreséion.
of coisolidation
E.rme. AnaJ.ysis of the electrical flows for dif-
ferent electrode arrangements, Fig. 22, shohrs Some studies have been made (Leong, L977) of
that a hexagonal arrangement is efficient consolidation induced by inflatabíé cylindrical
in membranes in vertical
terms of _(f) power consumption, (2) average bóreholes. One can imagine,
voltage (the higher the aizerage voítage ttre for- example, a hexagonal pattern of expandablé
gl:a,ter the.average amount of-consoliáation membranes in lieu of the ánodes in rig. ZZ wfrich
that induce an excess pressure in the surróunding
can De oDtalned for_ a given applied voltage),
and (3) anode to cathoáe ratió. soil and, hence, water flow towards drain wélls
at positions shown by cathodes j.n the figure.
The nagnitude of coniolidation that can 6e pro_
duced is dependent-on the excess pore pressure
that can be induced, which, in tuin, dépends
the inflatable '"p".i"g--on
Fq membrane dia¡neter, fróf"
F¡.i
+ o + o + o + _ r . and configuration, soil strength, and soil
o o o o o o + o + o + o - J . stress-strain properties. Reóent studies of
+ o . o + o +
o o oloio o t o , ¡ , o . o
consolidation using both verticaL and radial
+ or..-\o
* o <lo.f o. cyclic loads have been summarized by Schlosser
o . i o ) .
o o o o o o + o'v'o + o
+ o t.-.1 o
and Juran ( L 9 7 9 ) .
+ o + o + o + + i o + t
Wt frn... ,qt *.&r S¡qrrí Densely_packed guicklirne piles in soft,
ñrwt hrú. satu_
Q Car&. . Ha rated clay have been used. Four effects are
Lú/e: ht z s2
I 25, ?6Sa beneficial:
a5/s2 o5/t,
/aar/uat h.o o 5/s2
aJt/st I. Water is taken from the soil to hydrate the
o5/s2
I
45,
Lime.
a5v
o avlpse oeo v/2s2 acovlpst 2. Hydration of guicklime is accompanied by a
z oiTpst 2 a7?pt. r ro7/pst
specific aravlty decrease ir"*-5.3 to 2.2.
This leads to expansion against the bore_
v ' atth.d yotof
?,e.r,rh,t, hole and development of póre pressure in
the native soil- Recent expeliments by
Kuroda et a1. (1980) showed that expanaion
Fl-g. 22 Characteristics of Different pressures in excess of 2 Mpa could be
Electrode developed for conditione
Patterns in Electro-osmosis of 1ow porosity and
restrained expansion.
3. The heat of hydration helps to further reduce
the water content.
Increasing the number of anodes relative to
c-athodes is generally beneficial 4. Slow diffusion of lime into the surrounding
for tlvo reasons. soiL ]eads to cementation and qround
Vle11 points are often used for cathádes,
where_ strengthening.
as, reinforcing bars or aluminum roás are used
as anodes. Hence, anodes cost less. fron and
aluminum anodes decompose during trÁatment The need for densification of fine-grained
participate and sJ.urries, sludges, and slimes has bécome a major
in electro-chemical-haraáning ró_-
actions that give strength increasás beyónd concern in the containment and storase of dis_
those attributable soleiy to consolidation. posal materials and the subsequent u[.itization
of disposal sites. oewaterinf and densification
Tn application it is known that electro_osmosj.s of these materials by special-drainage and con_
may be effective and economj.cal under the solidation systems can be among the ñost effec_
following conditions: tive technigues available (,rohñson et aI ., Lg77) .
Seepage consolidatign; i.e., surface ponaíng
1. Saturated siLts or silty without surface membranes but with unáerdraín_
clay soils
age, can be especially useful because of the
2. Normally consolidated conditions greater effective stresses that can be developed.
3. Loe¡ pore water electrolyte Prefabricated drains can be utilized effectivily
concentration. both within and beneath the materials to be
cas generation and drying and fissuring densified. Because-of the large volume changes
at the and strains
el-ectrodes can.impair thé efficiency involved, new consólidation theoiies
of the have been developed for analysis of these mate_
method and Limit the magnitude of cónsolidation
pressures that develop. Treatment results r:.als. The work of Imai (f979) and Somogyi
non-uniform changes in properties
in et al. (L981) is useful in this regard. -
between
erectrodes, because the induced consolidation
depends on the voltage, Finally, precompression with vertical drains is
and the voltaqe varies now being used in conjunction
between anode and caihode. Thus reveisat with other methods
of of ground improvement and reinforcement. Tr{ro
el"ectrode.polarity may be desirable to achieve recent. examples may be cited. At the Changí
a more uniform stress condition. Electro_ Airport in Singapore a loose sand fill was
osmosis may also be used to accelerate the pracect over a thick layer of soft clay.

527

,rd
 to a depth INJECTION AND GROÜTING
Prefabricated drains were installed
áccereraie consolidation.of the crav'
;;-;;-;-¿; the sand Introduction
tamping \tas tñán-usea to densify repaired
Heavv engineer Charles Bériguy
fill- ( H a n s b o , 1 9 7 8 ) . tr lta, *,"t*h a grout
by injecting
a scouring sluice t!-óitppt. since
embankment to lr"Iil"-ái¿-ñváiá"ri' lim; beáeath it'
At the site of the east apProach into the -ground has
Bridge curréntlv-under construc- then injectr-on or *áteiiafs
;;.;;ttb;;¿on
enct o! Ene - developed into a *tááiy used method. for soil
tion across the southern -lfrá Because of
Bay, sofc mud- foundation soil IE"iiiír"dó"-""a sto"i.'d improvement' to
san Francisco consolida- cost, groúling is üsually limited
a shear strengtñ--iess than 5kPa' its hiqh 'ttor''¡ne áncl. to special
has would have t.r"iii"iv-"^árr
;;;.;-¿;
tion under the proPosed embankment problems that cannot-Ue sotved by other methods
'
to 40 years' special measures were were
;;k;"10 applications of
itiiárát.-.equiied to próvide embankment supPort ñ"Ii-á?-*tá- early -groutinq these
and
settle¡nent of for groundw"t"t "ottttol or shut off'
and to accelerate thé ánticipated today'
i n - r i s : consisted continue to ue very ilnportant 3pplications
sorution, á h á w n
i " i . - - r i t .-u-gáálá"ti1á
reinforcement
?1'.
to distribute ñ;;-;;;";tlv injeótio-ns have been used ror
át"'irl sro"'a movement control'
;*b;;it";"É loading and reduce differential ;;;;"a-;¿;;"átt'"ái"ü-"'á
that are of primarv
rirt-(sawdust) to iláT; ;;-iilá;; appíicatió"s
:;üi;;;;,-iil riqht"eisht
(3Í pretá¡ricated verticar concern in this report'
rEáil;'i."áds, time to less
;;-;;eü¿.'tire'consolidation are possible' as shown
á;;il
( 4 ) a filter to Three modes of injection
ü;;";";-y;"i,-á"a seotextile in Fig. 24:
"iá"."t cóntamination'of á drainage blanket'

I
I n.' o,'o^-.", E t , ¡ r ¡ ^^gr ' 0 ¡ Ñ ñ r
I f
Eñboñ¡ñanl Slrul¡
o¡ RaQu'rad I
\
Sorlu5t F'rl ,^ Ct'l'col ltaot

Q.¡ñlotc,.q Foátta -./ t 9

> s
; : cro'¡¡
-
t-- 1-vc't'co'r'c¡
80, xud

--72-<--__-.--^--

F'tfr Mol"'ol

Eí9. 23 GeneralizedcrosssectionShowingstabilizationMeasuresatt.hesiteofthe
ó"m¡árto" Bridge Approach FiIl (Hannon, 1980)

b) Displocement c) EncoPsulotion
o) Permeot¡on

surrol¡nds
Groul fills voids Grout bulb disPloces Grouffluid of soll
surrounding soil intoct chunks

Fíg. 24 TYPes of Grouting

528
 1. Permeation grouting in which the grout fills why other chemicals could not be used.
soil pores. There is essentially no change
in the volume or structure of the original Particulate Grouts
ground. This type of grouting can generally
only be accomplished in soils coarser than Neat cement and soil-cement grouts are the most
fine sands and in fissured rocks. commonly used particulate grouts, although soil-
water grouts have been used in some cases. In
2. Displacemeng-gríltí^g in which a stiff vtater-cement grouts water:cement ratios of 0.5:L
mixture fills voids and compresses the sur- to 6:1 have been used. l{ith low urater:cement
rounding soil. ratios there is less segregation and filtering,
3. Encapsulation in which naturally fragrmented and higher strengths are obtained, but they are
ground or ground fractured hydraulically harder to inject than grouts with a higher water
under high grout fluid pressures is in- content. Chemical additives are sometimes used
jected by grout which coats but does nor to facilitate penetration, to prevent cement
permeate the individual chunks of soil. A flocculation, and to control set times. Set
lens structure in the form of a cardhouse times can be as short as 30 seconds or very long.
is formed.
In soil-cement grouts a soil volume of four to
Recent references that present comprehensive six times the loose volume of cement is common.
reviews of injection and grouting include ASCE Water volumes from one third to twice the soil
( 1 9 8 0 ) , B o w e n ( 1 9 7 5 ), C a m b e f o r t ( 1 9 7 3 ) , C a r o n volume per bag of cement are used. The 1ow water
et al. (1975), Herndon and Lenahan (1976), content mixes are typical of high viscosity dis-
Kirsch and Samol (1978), and Lenzini and Bruss placement grouts. Zero slump compaction grouts
(1975). Although successful grouting may be with 30 to 60 second gel times can be made using
complicated, it is more than an art. Some cement, c1ay, and flyash mixes with an alkaline
underlying concepts have been defined, and prín- accelerator. ff bentonite is used, expanded
ciples guiding its success have been estab- particles may collapse if the groundwater has a
lished. ft is t-hese that are the focus of thi's high salt content. Care should be taken in the
sect10n. use of cement in the presence of sulfate-bearing
soils or groundwater.
Applications
Particulate grouts cannot be injected as permea-
A number of applications of grouting for soil tion grouts into soils finer than medium to
improvement apart from seepage control may be coarse sands. Some "groutability ratios" that
noted, including: have proven useful are

1. Void filling to prevent excessive settlement. /' " 1

n 5 \'
Forsoils: N =rñ_-T- soil
2. Ground strengthening under existing struc- t"85'qrout
tures to prevent movements during adjacent
excavation, pile driving, etc.
N > 24 : grouting consistently possible
3. Ground movement control during tunneling
operatÍons. Tan and Clough (1980) present a N < 11 : grouting not possible
design method for determining the required
size and strength of stabilized zones around (D1o)soi1
tunnels for effective qround movement ,"" _
control. l5gsEro,rt
4. Soil strengthening to reduce lateral support
requirements. Nc > 11 : grouting consistently possible
5. Soil strengthening to increase the lateral N < 6 : grouting not possible
load resistance of piles.
6. Stabilization of loose sands aqainst lique-
faction. -- -' R Width of fissure
For rock: (D95)
7. Foundation underpj-nning. oro.rt
Slope stabilization, > 5 : grouting possible
N. consistently
K
9. Volume change control of expansive soils N. < 2 z grouting not possible
through pressure injection of lime slurry. K

This technique is controversial and likely

to be effective only under special condi- AdditionaL gruidelines relating to particulate
tions ( I n g l e s and Nei1, 1970; Wright, 1973; grout types and particle size are:
Blacklock, 1975¡ Thompsonand Robnett, 1976)
Types I and II Portland cement are suitable for
soils coarser than 0.60 Íun.
Grout Types
Type III Portland cement is suitable for soils
?ermeation grouts are of tr¡ro types. Particulate coarser than 0.42 m¡n.
grouts are made up of cement, soil, or clay and
:rixtures of these. Chemical grouts are composed Bentonite is suitable for soils coarser than
cf various materials in solution. Displacement 0.25 mm.
or compaction grouts are stiff, low slump (0 to
)0 mm) mixtures of cement, soil, and,/or clay and Chemical crouts
?ater. Lime slurries are the most commonly used
encapsulation-type groutst however, there is no Chemical grouts offer the advantages over parti-
inherent reason, except perhaps for economics, culate grouts that they can penetrate smaller

i s Volyme 4
529
pores, as may be seen in Fig. 25, they have a systems. Premixed combinations of sodium sj'Ii-
lower viscosity, and there is a better control cáte and appropriate catalysts and activators
On the other hand their technology are formuláled to give specified design strength
of set time. gas
is more complex and costs are high. Soils con- and setting times. A two-shot silicate--Co2
taining less than 10 percent fines (<74 um par- system has been developed and is. used- in the
ticlesi can usually be permeation grouted with lóess soiLs of the USSR(Sokolovich, 1973).
chemicats. If the fines content is greater than Ammonia gas injection has also been found effec-
15 percent effective chemical grouting may be tive in Íoess vrherein only modest strengthening
difiicult. For fines content greater than 20 is required to prevent collapse. The reaction
percent permeation grouting will-not be possible, is to replace aásorbed Ca++ by NH4+. The libe-
Lut chenical grouts can be distributed along and rated Cai+ combines with water to-form Ca(oH)2,
through hydraulic fractures. which acts as a cementing stabilizer.
crouts containing 25 to 30 percent silicate are
The most common chemical grout classes are sili- where
cates, lignins, resins' acrylamides, and typical for waterproofing applications.
formulations higfr strength is.required for structural appli-
urethanes. Hundreds of different
have been developed within these classes. Of cations, silicate concentrations of 40 to 60
them, however, the silicates account for more percent are used. Unique loca1 chemical conCi-
tions may influence reactivity. Temperature
than 90 percent of Present chemical grout use,
the others belng linited for reasons of cost and effects, dilution by groundwater, catalyst adsor-
toxicltY. ption by sand, and other factors may influence
áel time and character of the resultant grout '
"Tr,ro-shot" silicate systems, in which a first Áccordinqly, prellminary testing is desirable'
injection of sodir¡m silicate is followed by a
se¿ond injection of a material such as CaC12 to The properties of chemically grouted,sands cover
precipitated silica a wiáe iange. Some typical values of viscosity
cause formation of insoluble grout
geI, have largely given way to one-shot silicate and compressive strength for different

(rlot"r',
Portlond Cemenl
I vi,scosity,'rp,, lll | | |
| | [com7ress¡ve strengl

S¡licotcs
t5/roo-,r""ollll
)r,)r, Eo)

ftd t.4, o035, (4d s, 60)30 [>3000J

Resins

EU il, IqOOJ

Acrylomides 00t.3, Eo)1.7,(5d2.5

o'l o.o/ o.ool

Soil Porlicle Size'mm

Fig. 25 Soil particle Sizes for Different Grout Types and Grouted SoiI Properties

\ 530
types and the particle size range for which they
are suited are shown in Fig. 25. The unconfineá
compressive strength of chemically grouted sand
may range from 0.3 to more than 10 Mpa. The
results of triaxial testsf e.g., perez et aI.
(1981), indicate that grouting produces an in-
crease in the cohesion intercept but has little
effect on the friction angle.

The stiffness of silicate qrouted sand increases

with confinj.ng pressure in about the same way as
an ungrouted sand. Stress-strain behavior is
non-Iinear. Strength and stiffness decrease as
loading rate decreases, creep occurs under con-
stant load, and creep rupture rnay deveLop under
high stress levels. These aspects of behavior
are important when grouting is used in under-
pinning, tunneling, and open cut construction.
The results of recent studies on these aspects
of grouted soil properties are given by Clough
e t a 1 . ( J - 9 7 9 1c' a r t u n g a n d K a n y ( 1 9 ? 5 ) , a n d
Perez et aI. ( f 9 8 1 ) .

Syneresis, i.e., a time-dependent tightening of

the gel structure resulting in fissuring, may
occur in silicate-grouted sands, and the Q¡-
"
Snulttocous
seriousness of the resulting problems increases dr'lling.nd
wj.th increasing silicate concentration. In 9?outtng
cases where sealing these cracks, as well as
6-er*t.g,"
others that may have formed by hydraulic frac- succattira lifta
turing during the injection process, is required, f.om tha boloñ
of th. grouthol¿
a second injection stage using a cement grout - c"oifi tr-*oust
@ "tubat
may be needed. ¿ ñaDchafr.s^

Equipment and Techniques

The design of a successful grouting program Fí9. 26 Tnjection Methods
requires not only the selection of a suitabte (Caron et a1., L9751
grout material, but also the correct drilling
equipment, procedures, and grouthole patterns.
It is essential that the pipes and injection grout takes in very pervious materials. Pumping
ports be in the right place. It is more impor- time to ge1 time ratios of I0 are common. The
tant to ensure that the fuII design soil volume disadvantage of fast ge1 times, of course, is
is permeated with grout rrhen the objective is that human or mechanical failure can lead to
water cutoff than erhen the objective is to grout set up in the eguipment.
improve mechanical properties.
A coÍunon rul-e of thumb for maximurn injection
Both batch and continuous flow systems can be pressure for open pipe grouting is about 20 kPa/m¡
used effectively. Batch systems give better i.e., a value equal to the over-burden pressure,
mixing of components, but require longer geI unless grouting under a heavy structure or in a
times than do continuous flow systems. Gel situation with greater confinement. Values two
times of 50 to 90 min. are conrmon for batch to three times as great may be used with fast gel
systems. Continuous flow systems may give time systems. Because of the need to open the
better placement control but are more complex. rubber sleeve and to break through the plastic
grout sheath with the tube á machette method,
Hole spacings of about 1, 3 m to 2.5 m are typi- high pressures may be required initially. High
cal. Costs become excessive for smaller spac- initial pressures may also be useful if it is
ings, and grout placement in desired locations desired to open up the formation by hydraulic
can't be ensured if larger spacings are used. fracturing.
Both open-pipe and sleeve-pipe (tube á manchettes)
injection methods are used, as shown in Fig. 26. Electrochemical In j ection
Although the sleeve-pipe method is more expen-
sive, it is becoming widely used because it Just as direct electrical- current can be used in
gives much better control of grout location. lieu of a physical loadíng to consolidate a
The recent test program at Locks and Dam No. 26 compressible soi1, it can be used also to move
described by Perez et aI. (1981) is an excellent solutions into and through a porous material in
comparative study of the thro methods. At this lieu of a hydraulic injection pressure. Because
site multiple-stage sleeve-pipe grouting was the coefficient of electro-osmotic permeability,
much more effective than open-bottom pipe k., is insensitive to particle size and general-
grouting. ly fal1s wlthin a narrow range of about t x 10-s
to 7 x I0-' cm/sec/volt/cm, a unit electrical
The use of gel times less than the pumping time, gradient (l volt/cm) can be more effective than
termed fast gel- times, has the advantages of a unit hydraulic gradient for moving fluids
grout .Location control- in flowing ground water through finer grained soi1s. Hence, electro-
and in stratified soils. It also will limit kinetic injection might be considered for use in

531
silty soils which cannot be injected using ordi- Baker, f980) . Grout bulbs srere inserted via
nary grouting techniques. Electro-kinetic 76 rnmdiameter pipes placed 3 m apart to provide
injection might be useful also where lack of support between the tunnel crown, after the
confinement prevents grouting in the usual way. shield passed, and the ground surface. The
scheme is shown schematically in Fig. 27. Pump-
Chemical stabilizers are introduced at the anode ing pressures averaged about 2 MPa' injection
and carried toward the cathode using electrical depths were about 12 m below ground surface, and
gradients of the order of 50 to 100 volts per work was carried out about 2 m behind the shield.
meter. Several experiences with electrochemical
injection up to the mid-1970's are summarized by
Pilot (1977). orBannon et a1. (f976) describe
the electrochemical hardening of expansive c1ay.
oncescu and Ba1ly (L977) used an electro-kinetic
injection of sodium silicate to strengthen the
loess under the foundations of a theater.
"Piles" of treated soil 3 to 4 m long and 1.6 to
2.2 m in diameter were formed.

The electro-kinetic stabilization of a poten-

tially liquefiable sand has been investigated
by Yamanouchi and Matsuda (1975). The concePt
was to fill the voids of the loose sand with a
ge1 or coll-oidal material and thus prevent
collapse under cyclic loadings. Silicate solu-
tions, bentonite, and aluminurn hydroxide were
investigated as i.njection materials. The ben-
tonite and aluminum hydroxide, as negatively
charged colloids, were injected at the cathode
and moved into the pores by electro-phoresis.
The results demonstrated a marked increase in
resistance to liquefaction after treatment.
funnel ond
Segall et al. (1980) report the results of a o d v o n c i n gs h i e l d
large number of chemical analyses on water
leached both hydraulically and by electro-
osmosis from dredged soil. It was found that
because of the electrochemical reactions at Eíq. 27 Compaction Grouting During Tunneling
electrodes, the electro-osmosis water became very to Prevent Settlements
alkaline (pH = I3.4). As a result organic
materials went into solution. Heavy metals were
desorbed and carried out at the cathode, as htere Jet GroutÍng
also pesticides. Deposits of iron oxide, mag-
nesium hydroxide, and calcium carbonate were A new grouting technique termed jet grouting was
formed in the soil near the anode. These introduced in Japan several years ago (Yahiro
results suggest a potentially useful approach to and Yoshida, 1973¡ Miki, 1973; Miki et al., 1980).
cleaning up hazardous waste-contaminated soil. The basis for jet grouting is a special high
speed water jet acting under a nozzLe pressure
Displacement (Compaction) Grouting oi 15 to 70 MPa. The native soil may be mixed
in place with a suitable stabilizer as shown in
Highly viscous soi1, cement, and water displace- Fig. 28. Alternatively, poor soils can be re-
ment or compaction grout acts as a radial rnoved by in-situ excavation and replaced by a
hydraulic jack which compresses the surrounding mortar grout to form hard, impervious columns,
soil. The hardened grout mixture is a bulb of panels or sheets' as shown schematically i-n
strong, relatively incompressible material. itg. zs. Jet grouted columns up to 3 m in dia-
Displacement grouting can be used in partly meter are possible. The use of air jetting in
saturated soil masses and loose materials con- conjunction with grout jetting can yield dia-
taining void spaces. rt is used to correct dif- met¿rs up to twice as great, for a given jet
ferential settlements or to provide under- pressure, as the grout jet alone.
pinning and ground strengthening adjacent to
open excavation or tunneling activities. Avail- The method offers the advantages of both close
able equipment can develop up to 2.5 to 3.0 MN,/m2 control over the zones being treated and appli-
punping pressure, and zero slump grout can be cability to clays as well as sands. Miki et al.
pumped distances in excess of 30 m. To be (1980) report an increase in unconfined compres-
effective, compaction grouting should not be sive strength of an originally soft cfayey soil
undertaken at depths less than I Lo 2 m unless to I.5 to 4 MPa, some 30 times the ori-ginal
there is an overlying structure to provide con- strength. The secant modulus at 50 percent of
finement. the fáilure stress was increased by a factor of
200.
Cornpaction grouting materials and procedures are
described by Warner and Bro\"rn (I974), Graf Evaluation of Effectiveness
( 1 9 6 9 ) , a n d l f a r n e r ( 1 9 7 8 ) , a m o n go t h e r s . R e c e n t -
ly, during construction of the Baltimore, Mary- Precise determination of exactly where all the
Iánd, U.SIa. subway near the Bolton HiII'Station, grout went in the ground is usually not possible.
trle projected cost of underpinning existing Ássessment of grouting effectiveness is usually
structures was so high that compaction grouting rnade on the basis of grout take records and the
\,/as used to prevent surface subsi-dence (Haward results of in-situ tests and laboratory tests on

532