Ninth Laurits Bjerrum Memorial Lecture: ''Small Is Beautifulm-The Stiffness of Soils at Small Strains

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

Ninth Laurits Bjerrum Memorial Lecture: ''Small is beautifulM-the stiffness of soils

at small strains

Department of Civil Engineering, Imperial College, London S W7 2BU, United Kingdotn


Received February 20, 1989
Accepted July 7, 1989
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

The purpose of this lecture is to show how field measurements of displacement around foundations and excavations
combined with numerical analysis and recent developments in laboratory techniques have together brought about
improvements in the understanding of the stiffness properties of a wide range of soils at the low levels of strain often
encountered in practice.
Key words: excavations, foundations, weak rock, stiff clay, field measurements, strain, yield, laboratory instrumentation,
in situ testing.

Le but de cette presentation est de montrer comment des mesures in situ de deplacement autour des fondations et
excavations combinees avec L'analyse numirique et les developpements recents dans les techniques de laboratoire ont
contribue a I'amelioration de la comprehension des proprietis de rigidite d'une large plage de sols aux faibles niveaux
de deformation souvent rencontres en pratique.
Mots clPs :excavations, fondations, roche molle, argile raide, mesures in situ, deformation, limite Clastique, instrumen-
tation en laboratoire, essais in situ.
[Traduit par la revue]
Can. Geotech. J . 26, 499-516 (1989)
For personal use only.

Introduction modelling the stress-strain behaviour of soil-admittedly of


This lecture is concerned with the stiffness of the ground a highly idealized variety. In essence, the approach unified
at small strains. During its preparation I became increasingly the work of Rendulic (1937) for normally consolidated (con-
aware of the extent to which the work described has been tractant) clays and of Hvorslev (1937) for overconsolidated
shaped by three projects with which I was involved during (dilatant) clays and coupled these to the concepts of work-
the first six, very intense years of my professional career. hardening plasticity.
The projects are summarized briefly in the following. The theoretical framework has been described in detail
elsewhere (e.g., Schofield and Wroth 1968; Roscoe and
Excavation in London Burland 1968; Atkinson and Bransby 1978) and I will only
My first job in 1961 was concerned with the design and point out a few key features (Fig. 3). Compression of a
construction of the foundations and retaining walls for normally consolidated clay takes place on a state boundary
Britannic House in the City of London. The 35-storey tower surface in q, p' , e space defined by Rendulic's contours of
block, the tallest in London at the time, is founded in a 20 m constant void ratio and most appropriately named the
deep excavation. The project has been described in detail Rendulic surface (although Atkinson and Bransby call it the
by Cole and Burland (1972). At the north end of the site, Roscoe surface). Any path lying on this surface (e.g., all-
the construction of the tower raft foundation entailed round compression, KO compression, drained triaxial test)
excavating a temporary 1: 1 slope in the London Clay (Fig. 1). gives rise to large irrecoverable (plastic) volume strains and
The slope had to remain stable for a period of 2 months, associated shear strains.
until the surrounding diaphragm walls could be propped When overconsolidated owing to unloading, the soil is
from the tower raft. Measurements of movement of the assumed to be elastic. Thus, any state path for an over-
diaphragm wall were made during and subsequent to excava- consolidated soil moves on a so-called elastic wall. Such a
tion. The top of the wall moved into the excavation by a wall intersects the Rendulic surface for lightly overconsoli-
maximum of 55 mm, with the bottom moving somewhat less dated states and the Hvorslev surface for heavily overconsol-
(Fig. 2). At the time, the magnitudes of the movements gave idated states along a "yield line." When sheared or
rise to some concern but the basement was successfully com- compressed, an overconsolidated soil can be expected to
pleted. Some years elapsed before these movements were exhibit high initial stiffness and small strains until the current
analysed in detail. "yield line" is approached, when the stiffness will reduce
markedly. The sharp increase in the compressibility of clays
C a mbridge under one-dimensional conditions as the vertical effective
In 1963, I went up to Cambridge to work under Professor stress passes through the preconsolidation pressure is a
K. H. Roscoe. This was the time when the concepts of work- special case of this elastic-plastic model. This paper is
hardening plasticity had just begun to be applied to soils. devoted largely to the stress-strain behaviour of soils for
It was an exhilarating period and the debates about the states lying beneath the state boundary surface on the so-
subtleties of various energy equations, plastic potential, etc. called "elastic wall."
were intense and often heated.
We were seeing for the first time the prospect of develop- The chalk at Mundford
ing complete, self-consistent constitutive relationships for After leaving Cambridge, and attracted by the prospect

Printed in Canada / Imprime au Canada


500 CAN. GEOTECH. J. VOL. 26, 1989
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12
For personal use only.

FIG. 1. Excavation for basement of Britannic House in London Clay (Cole and Burland 1972).

-
E
60-
I I I I I
+ 8 . 5 L d level
A
IY I
-
E
4

*
g LO- -
-8
:: -
6 20-

-2,
0
May June July August September October November April
1963 196L
FIG.2. Movement of north diaphragm retaining wall towards excavation-Britannic House (Cole and Burland 1972).

of making measurements of field behaviour, I joined the particularly impressed by his keen observations of the
Building Research Station, Watford, United Kingdom, in weathering structures of the chalk. He also made a profound
1966. By good fortune I was immediately involved in the impression on the drillers on two counts. First, unlike most
site evaluation for a 300 GeV nuclear accelerator at visitors to the site, he was prepared to descend a trial shaft
Mundford, Norfolk, on the chalk. The investigation has in a bosun's chair, and second, he showed a keen interest
been described in detail by Ward et al. (1968). It involved in the drilling equipment and questioned the drillers closely
a full-scale loading test by means of a 4.546 x lo6 dm3 about it. It is perhaps appropriate that this paper should
(lo6 gal) steel water tank (Fig. 4), plate loading tests at begin with a description of my work on the founding prop-
various depths in vertical shafts alongside the tank and erties of chalk.
elsewhere, and the detailed visual description and classifica-
tion of the weathered rock from over seventy 0.75 m Deformation properties of chalk
diameter shafts covering the 50 km2 site. Chalk outcrops over a large area of southeast England.
Laurits Bjerrum was the geotechnical assessor for the It is also found in Eu'rope, large parts of Asia, and in some
European Organisation for Nuclear Research (CERN) and places in Africa, the Americas, and Australia. It is a soft
my first meeting with him was when he visited the site. I was white rock consisting almost entirely of calcium carbonate
BURLAND 501

Rendulic surface ritical stote line joints are closely spaced, open, and often infilled with
weathered debris. Grade V chalk consists of small fragments
of unweathered chalk set in a matrix of remoulded
weathered chalk and grade VI (suggested by Wakeling 1969),
is very soft structureless chalk containing a few small
unweathered fragments.
Isotropic consol. lin
The Mundford test tank mentioned in the Introduction
was founded on the surface of grade V chalk, which steadily
Constant e contour
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

improved in grade with increasing depth to grade I1 at a


depth of about 12 m. The tank was 18.3 m in diameter and
of the same height, so that when filled with water it exerted
a uniform pressure on the ground of 180 kPa. In a series
of five shafts under and nearby the tank, precise instruments
were arranged to measure the vertical displacements between
successive levels to an accurary of k0.0025 mm. Precise
water-level gauges were mounted at the top of each shaft
Tension failure
and at other locations on the ground surface to measure the
vertical displacements relative to a datum some distance
/ '
away from the tank to an accuracy of ~ 0 . 0 5mm. Ward
et al. (1968) give a detailed description of the
FIG. 3. Part of the state boundary surfaces, elastic wall, and
instrumentation.
"yield line" of the Cam-clay theory.
Some typical results from a load, unload, reload test
in the form of 65-85% skeletal remains of microorganisms are shown in Fig. 5, which shows the relationship between
(known as coccoliths, 0.5-4 pm in size) and 15-35% coarse surface pressure and vertical strain for selected levels beneath
carbonate debris. In its unweathered state the material is and outside the tank. It can be seen that the strains were
weakly cemented owing to pressure solution at intergranular extremely small, nowhere exceeding 0.01 %. Moreover, the
boundaries. As a result of weathering and tectonic forces stress-strain behaviour was remarkably linear and recoverable
For personal use only.

the chalk mass usually exhibits a jointed structure. Because with a small amount of hysteresis-a surprising result for
of its grain size the chalk is particularly prone to frost action a weak weathered rock.
under periglacial conditions causing shattering and Burland et al. (1973) analysed the measured displacements
remoulding. at Mundford and found that the vertical displacements
As a result of the work at Mundford, Ward et al. (1968) beneath and outside the loaded area agreed very closely with
developed a visual classification or "grading" of the in situ elastic theory for linearly increasing stiffness with depth.
chalk. Grades I and 11-working from depth upwards- Figure 6 shows the comparison between measured and
refer to unweathered chalk with widely spaced and tightly predicted displacements beneath the centre and just outside
closed joints. Nearer the surface the spacing of the joints the edge of the tank. By means of this analysis a range of
reduces and they become open (grade 111). For grade IV the Young's modulus values was assigned to each chalk grade.

FIG. 4. View of the test tank in the forest clearing at Mundford (Ward et al. 1968).
502 CAN. GEOTECH. J. VOL. 26, 1989

Shoft 1 levels 3 - L
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

-First loading
t t level 6
----- First unloading
Second loading

Vertical strain (O/O)

FIG. 5. Typical relationships between surface load intensity and vertical strain during loading, unloading, and reloading of the test
tank (Burland et al. 1973).
Settlement (mm) Settlement Imm) Young's modulus (MPo ~ 1 0 ~ )
For personal use only.

I I analysts
Plate looding \
tests
\
Beneath Centre Beneath Edge te. onalysis ]
(a) (bl
FIG. 6. Comparison between measured and predicted vertical displacements beneath the centre and the edge of the test tank (Burland
et al. 1973).

A number of 0.86 m diameter plate loading tests were compared in Fig. 6 b with the values derived from the tank
carried out in shafts at various depths alongside the tank test results and the agreement is seen to be very satisfactory.
and at various other locations (Burland and Lord 1969). The results of the Mundford experiment showed that at
Great care was taken to properly bed the plate at the bottom low stresses the chalk was very much stiffer than had
of the shaft. Figure 7 shows a typical average pressure versus previously been assumed by engineers. A clear correlation
settlement curve from a plate test in grade IV chalk. The between degree of fracturing and stiffness had also been
behaviour is characterized by an approximately bilinear established, which has since proved of very considerable
response with a fairly well defined "yield" pressure value in evaluating in situ properties of chalk.
separating high stiffness at low pressures from much lower Some years after the Mundford work an opportunity to
stiffness at high pressures. It is important to note that the carry out some further work on the chalk was taken when
"yield" pressure is nowhere near failure and takes place at the Building Research Station was invited to instrument a
settlements equal to about 0.1 To of the plate diameter. The group of large silos founded on chalk. The silos are 50 m
low-stress values of Young's modulus deduced from the high, 20 m in diameter, and are founded on circular rafts
plate loading tests carried out alongside the tank are bearing on grade IV chalk that is approximately uniform
Load intensity (kPa) Avq. Fol~ n d a t ~ oPressure
n (kPo)
200 3W 402
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

SOLSO

FIG.8. Relationship between average foundation pressure


and average settlement for two silos on chalk.
R e l a t i v e Settlement (mm)
0 10 20 30

10 t
FIG.7. Average pressure versus settlement of 865 mm diameter
10
For personal use only.

plate test on grade IV chalk (Burland and Lord 1969).


to a depth of about 25 m (Burland and Davidson 1976).
Precision borehole extensometers were installed beneath the
centre of each raft (Smith and Burland 1976).
Figure 8 shows the measured pressure-settlement relation- 20 Silo 3
ships for two of the silos. It can be seen that the behaviour . q ,vg=210 kPa
is remarkably similar to the plate loading test results shown
in Fig. 7, with a distinct yield pressure at about 250 kPa. f . q = 250 kPa
This yield pressure is approximately half the value obtained h q =313 kPa
from the plate loading tests, indicating an important scale + q og, =353 kPa
effect and resulting in larger settlements than predicted.
The data obtained from the borehole extensometers pro- q,vg =383 kPa
vide some valuable insights into the yielding of cemented
materials. Figure 9 shows the measured distribution of settle-
ment with depth for various average bearing pressures
beneath silo 3. It is apparent that yield first occurred at a FIG.9. Relationship between settlement and depth beneath
depth of about 15 m and thereafter progressed upwards. centre of silo 3 for various average foundation pressures.
The settlement distributions in Fig. 9 can be converted
into vertical strains and Fig. 10 shows the associated relation- stress (AT,,,,) and change in mean normal effective stress
ships between local vertical strain and average bearing (Ap'). It can be seen that the yield points for various depths
pressure for successive points down the extensometer. The and for the two silos splot on a unique curve in stress incre-
onset of yield at each depth is marked by very sharp and ment space.
well-defined kinks in the various curves. It is clear that yield The grade IV chalk at the site of the silos may be considered
occurred between points 2 and 4 (20 and 14 m depth respec- to be a residual soil arising out of in situ weathering of the
tively) at much lower bearing pressures than for shallower parent rock. The observed behaviour appears to be typical
depth. Note also the very small vertical strains at which yield of many residual soils, the properties of which were discussed
occurred-generally less than about 0.03%. Almost identical in detail by Vaughan (1985). Yield is the result of the break-
behaviour was observed for silo 4. down of cementation bonds. The cementation bonding gives
The values of initial Young's modulus were found to be rise to yielding effects separate from those induced by stress
approximately constant with depth. Moreover, the rafts of history. Unlike stress history, cementation bonding may be
the two silos were observed to settle more or less uniformly uniform over considerable depths, as appears to be the case
up to yield. Hence, elastic theory for the stresses beneath for the chalk underlying the silos.
a rigid circular punch has been used to estimate the stress The properties of artificially bonded soils are being studied
changes at which yield occurred at various depths. Figure 11 at Imperial College. They are manufactured by mixing a
shows the yield points for various average depths beneath kaolin slurry with sand, then pouring the mixture into a
silos 3 and 4 plotted on axes of change in maximum shear mould, and air drying the soil so that the kaolin retreats
504 CAN. GEOTECH. J. VOL. 26, 1989

LOO r
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

V e r f ~ c o l stro~n (001 % per dlvlslon)

FIG. 10. Relationships between local vertical strain at various depths beneath centre of silo 3 and average foundation pressure.
The recognition that even deeply weathered chalk is far
-a 80 -
0
oSilo 3

Silo 4
stiffer than indicated by the A line in Fig. 13a has had a
profound influence on foundation design in the United
-.6 0 -
Y

Kingdom and has resulted in considerable economies. The


For personal use only.

-
N
- measurements on the silos indicate that some care is needed
b"
a
I LO - do' in extrapolating the yield pressure q, obtained from plate
'b loading tests to much larger diameters, as there may be
4 - // important size effects. In view of the similarity between
20 - '/\ weathered chalk and other residual soils, it is very probable
that the methods of visual classification and field testing
0 m ' ' ' ' " ' ' ' ' ' ' ' described here can be successfully applied to other materials.
20 40 60 80 XX) 120 140
( A @ " + ~ A ~ R ) I(kPa)
~ Foundations and excavations in London Clay
i Frc. 11. The stress changes causing yield beneath the centre of In the introduction it was stated that some years elapsed
silos 3 and 4. before the measured movements of the diaphragm retaining
walls of Britannic House were analysed. The stimulus to
carry out a detailed back analysis of the measurements was
towards the grain contact points. The specimen is then fired provided by the design of the deep excavation for the
in a furnace; the degree of bonding depends on the tempera- Barbican Arts Centre in the centre of London, where the
ture and the time. The composition of the sand may be varied nature and proximity of existing structures made it necessary
by mixing quartz sand with crushed fired kaolin. Figure 12 to limit ground movements (Stevens et a[. 1977).
shows the yield locus obtained from samples prepared from The analysis was carried out using the finite element
57% quartz sand, 30% fired kaolin, and 13% kaolin method and represents one of the earliest applications of
(Maccarini 1987). The shape of the yield locus is broadly the method in ground engineering. The successful application
similar to that obtained for the chalk beneath the silos. of the method required an appropriate choice of stiffness
The field studies carried out on the chalk have resulted parameters for the London Clay. It was concluded that the
in a considerable upgrading of its suitability as a founding most direct method of evaluating these deformation proper-
material. The line labelled A in Fig. 13a shows the relationship ties was to carry out a back analysis of the measurements
between standard penetration test (SPT) blow count and made at the nearby Britannic House. The results of careful
Young's modulus proposed by Wakeling (1965). The values laboratory tests on specimens taken from block samples of
of E are based largely on laboratory tests and small-diameter London Clay (Bishop et al. 1965) coupled with their careful
plate tests in which the plate could not be properly bedded. analysis by Wroth (1972) strongly suggested that it would
In Fig 136 the filled circles and open circles show the be appropriate to treat the London Clay as a porous elastic
preyield values of Young's modulus (E,) for the Mundford material. Cole and Burland (1972) describe the results of
tank test and the large-diameter plate tests. It can be seen the back analysis, which involved a trial and error procedure
that the values of E, are 5-20 times larger than those given in which the values of E, (undrained Young's modulus)
by line A. The crosses give the equivalent value of tangent were adjusted iteratively until the observed and predicted
Young's modulus after yield (E,) and even these lie above horizontal wall displacements were in agreement. Average
line A. The pre- and post-yield values of E for the silos are values of KO obtained by Skempton (1961) and Bishop
also shown and agree well with the results from Mundford. et al. (1965) were used to derive the initial horizontal effective
BURLAND
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

(k Pa)
100

I1 &/ 1 I I I J I 1 i

100 200 300 LOO 500

FIG. 12. Results of triaxial tests on artificially bonded soil showing a yield locus (Maccarini 1987).
FoundoI8on
Q p i a l e Test-Dborneler shown 1 1 (St. John 1975). In all cases Eu increases with p' . The cases
looor( ) x Undrolned Triaxtol Test
involving vertical loading give lower values of Eu than do
excavations where large changes in horizontal stress occur,
pointing to anisotropic stiffness properties with Eh greater
For personal use only.

than E,. The range of Eu values obtained from the


laboratory tests on block samples from Ashford Common
(Bishop et al. 1965) are also shown in Fig. 14. The large
discrepancies between the field and laboratory values are
0 10 20 30 L0
evident.
SPT As a result of the above work, I had, during the late
1970's, become convinced that the low stiffness values deter-
9 Tank Test mined from laboratory tests were due to sampling effects
(Burland 1979) and had therefore become a strong advocate
of in situ testing as the only reliable way of determining
in situ stiffness properties. Then, two important
developments, which took place almost simultaneously,
caused me to modify this view.
Certain anomalies had appeared when predicted ground
displacements based on the use of linear elasticity were
compared with field measurements. For example, the shape
of the predicted surface settlement profile around the New
I I t I I I I I Palace Yard car park differed significantly from the
0 10 20 30 LO
SPT
measured profile, as shown in Fig. 15. Simpson et al. (1979)
showed that by using a bilinear stress-strain law with a high
FIG. 13. Relationships between SPT blow count and Young's initial stiffness the agreement could be greatly improved,
modulus E for chalk: (a) from Wakeling (1965); ( 6 ) comparison as shown by the broken lines in Fig. 15.
of Mundford data with Wakeling's line A.
Simultaneously with Simpson's theoretical work,
laboratory experimental studies began at Imperial College
stresses. An important and unexpected finding from this work in which strains were measured locally on the samples. These
was that the deduced values of E, were at least five times gave much larger stiffnesses at smaller strains than were
greater than those derived from careful laboratory tests. obtained from the average strains across end plattens (Costa-
The back-analysed values of Eu from Britannic House Filho and Vaughan 1980). Figure 16 shows the results of
were used successfully in the design of the Barbican Arts some unconsolidated undrained tests on 38 mm diameter
Centre and subsequently for other foundations and excava- specimens of London Clay in which locally measured strains
tions in the London area. Two notable recent examples are are compared with average overall strains.
the underground car park at New Palace Yard, Westminster The two approaches pointed strongly towards markedly
(Burland and Hancock 1977), and the Queen Elizabeth I1 nonlinear stress-strain behaviour with high stiffness at axial
Conference Centre (Burland and Kalra 1986). Figure 14 strains less than about 0.1 Yo. Such low values of strain are
shows the distribution of undrained Young's modulus Eu beyond the accuracy of most routine commercial testing
with mean effective stress p' derived from seven field studies apparatus. The obvious question that has to be raised at this
506 C A N . GEOTECH. J. VOL. 26, 1989

Eu (MPo)
-E

- Class A ( Burland & Hancock )


--
' 0 120 2$
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

S ~ m p s o net al.

FIG. 15. Observed and predicted distributions of ground


surface displacement with distance from the edge of the excavation
at New Palace Yard (Simpson et al. 1979).
LOOr
For personal use only.

FIG. 14. Variation of undrained Young's modulus E,, with


mean normal effective stress derived from various sites on London
Clay (St. John 1975).

stage is, what levels of strain are developed in the ground


under working conditions?

Ground strains induced under working conditions


The determination of strain in the field requires the
measurement of differential displacement, for example,
along a borehole extensometer. There are relatively few
published records of measured differential displacements
within the ground associated with buildings and excavations. FIG. 16. Unconsolidated undrained triaxial compression tests
The following cases have been taken from the literature. on London Clay using local strain transducers (Costa-Filho and
Attewell and Farmer (1974) present some observations of Vaughan 1980).
ground movements resulting from the excavation of a 4.2 m Again, it is apparent that only locally does the vertical strain
diameter shield driven tunnel at a depth of 30.8 m in London exceed 0.1 %. Figure 20 shows the settlement observations
Clay beneath Green Park. The tunnel was lined with cast- for a tall residential building in northern Germany founded
iron segments (seven per ring) and was subsequently grouted. on boulder clay. It is evident that, apart from the top 2.9 m,
Figure 17 shows the measured maximum settlements at the strains in the ground are less than 0.1 %. The somewhat
various depths over the centre line of the tunnel and these larger strains in the top 2.9 m may be due to disturbance
have been converted into average direct strains. It can be and softening in the base of the excavation.
seen that it is only within a tunnel diameter above the crown Finally, Fig. 21 shows the measurements of settlement at
that the vertical strains exceed 0.1 Yo. Most of the ground various depths beneath a 3.1 m x 3.1 m test footing on the
is strained to less than 0.05%. crust overlying soft Champlain Clay in eastern Canada
Turning now to foundations, Kriegel and Weisner (1973) (Bauer et al. 1976). The footing was constructed in an
presented measurements of settlement at various depths excavation 0.66 m deep. Design working pressures in the
beneath some buildings in East Germany. Figure 18 shows range of 125-200 kPa were used. The settlements have been
graphs of settlement versus average bearing pressure for a converted to average vertical strains and it can be seen that
tall hotel in Berlin founded in 40-50 m of medium dense below a depth of 0.78 m the strains are generally less than
sand. The settlement measurements have been converted to 0.2% at a pressure of 200 kPa. The much larger strains in
average direct strains. Even though high bearing pressures the top 0.78 m are once again believed to be due to distur-
were applied, only locally do the strains exceed 0.1070, bance and softening caused by the excavation for the
reaching a maximum of about 0.3%. footing.
Figure 19 shows the results of settlement measurements It is evident from these few examples that under working
on a tall residential building founded on medium dense sand. conditions we can expect the large mass of the ground
BURLAND

Vertsal s t r a ~ n(%) Settlement (mm) Avg. Bearing Pressure (kPo1


0 200 LOO 600 800 1000

.-.
E
-
E
20 -
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

Avg. Bearing Pressure ( k P o 1


0 200 LOO 600 800 1000

FIG. 17. Measured distribution of vertical displacement and


strain above a tunnel in London Clay (Attewell and Farmer 1974).

beneath and around a structure, be it an excavation or foun- " I


For personal use only.

0 1
dation, to experience direct strains of less than 0.1 %. It is
only locally that strains are likely to exceed this value. It FIG. 18. Settlements and strains beneath a tall hotel building
on medium dense sand (Kriegel and Weisner 1973).
follows that if successful predictions of ground movement
are to be made, it is necessary to ensure that reliable
measurements of strain can be made to an accuracy of of gauges is used, mounted diametrically opposite each
0.01% or better. It was with this objective in mind that a other, so that any tilting of the sample can be measured and
programme was initiated at Imperial College to develop and compensated for.
refine methods of precise measurement of local strains in Continuing developments are taking place in the measure-
the laboratory. ment of local strains on samples. For example, Ackerly et al.
(1987)describe a device in which the electrolevel is replaced
Comparison of laboratory and field investigation of by a strain-gauged pendulum inclinometer. Clayton and
small-strain behaviour Khatrush (1986) describe a simple, compact device that
makes use of the Hall effect. Figure 23 shows a photograph
Previous work on the measurement of local strains on of the latest inclinometer gauge in use at Imperial College.
samples in the triaxial apparatus had involved the use of
linear variable differential transducers (LVDT's). These Laboratory measurements of small-strain stiffness
instruments, though very accurate, are cumbersome and Figure 24 shows the results of a typical unconsolidated
difficult to use, since they require accurate alignment and undrained test on an undisturbed sample of London Clay
are prone to jamming and damage when the sample from Canon's Park in north London. The measured effective
approaches failure. Hence a programme was initiated to stress path is plotted in Fig. 24a, with locally measured
develop simple and precise methods of measuring local axial values of axial strain indicated. Note that a mobilized shear
strains in the laboratory. strength of about 0.25Cu is required to achieve an axial
strain of 0.1%.
Laboratory instrumentation Figure 24b shows a comparison between the externally
Burland and Symes (1982)describe such a method, which measured and locally measured stress-strain curves. It can
makes use of electrolytic liquid levels to measure change in be seen that the external measurements, which have been
inclination. Figure 22 shows the principles of the device, in corrected for apparatus compliance, lead to apparently
which a change of height Ah is converted to a change in slope linear initial stress-strain behaviour. However, the local
A6. Pads C and D are glued to the membrane surrounding measurements show much stiffer nonlinear behaviour.
the sample. There are hinges at C, D, and E. The hinges Figure 24c shows the results of the internal measurements
at C and D are located close to the mounting points on the of stiffness plotted in terms of Eu/C,, versus log E , where
specimen. Jardine et al. (1984)report that these gauges can Eu is the secant undrained Young's modulus. The strongly
resolve displacements to an accuracy of about 1 pm over nonlinear nature of the sample response is evident, with
a range of 15 mm. Eu/Cu decreasing from about 1700 at 0.003% strain t o
These inclinometer strain gauges are easy and quick to about 150 at 1.0% strain. The broken lines in Fig. 24c shows
mount and, because the mechanism is fully articulated, they the range of Eu/Cu results obtained from Canon's Park
are not damaged when the sample is taken to failure. A pair over a depth of 10 m. A similar range was obtained from
508 CAN. GEOTEC:H. J. VOL. 26, 1989
Avg. Bearing Pressure (kPo) Avg. Bearing Pressure (k Pa)
LOO
I
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

Avg. Bearing Pressure (kPa)


Avg. Bearing Pressure (kPa)

FIG. 19. Settlements and strains beneath a tall residential


building on medium dense sand (Kriegel and Weiser 1973).

another site many miles to the north at Bell Common, FIG. 20. Settlements and strains beneath a tall residential
For personal use only.

Epping Forest. building on boulder clay (Kriegel and Weisner 1973).


The following conclusions can be drawn from the results Avg. Bearing Pressure (kPo)
given in Fig. 24:
(I] At axial strains less than about 0.1 % (i.e., the upper limit
of strain experienced by the large mass of the ground under
working stresses) local measurements give much higher stiff-
nesses than those determined from traditional external
measurements.
(iz] The stress-strain relationships are continuously nonlinear
from very small strains and there is no clear indication of
a well-defined threshold strain of the type postulated by
Simpson et al. (1979).
(iig The use of local strain transducers leads to a very signif-
icant improvement in the consistency and repeatability of
stress-strain measurements, as can be seen from the small
range of results for Eu/Cuversus log 6 , shown in Fig. 24c.
(Note that the nondimensional ratio Eu/Cu was employed
here, but Jardine et al. (1984) have argued that the ratio Avg. Bearing Pressure (kPa)
Eu/pd has many fundamental and practical advantages.)
(iv) For ease of presentation and comparison the secant
stiffness-strain characteristics can be represented by the two
parameters Eu(o,ol,, the undrained modulus at 0.01 % strain,
and L, the index of linearity defined as Eu(o.l)/Eu(,~ol).
The techniques of measuring the stress-strain properties
of soils in the laboratory having been improved, the key
question is, how relevant are these measurements to field
conditions?
Field measurements of small-strain stiffness
The direct measurement of in situ stress-strain behaviour
requires the measurements of strain of the type described
previously. Three such investigations have been carried out
in London Clay and have been analysed by Jardine et al.
(1985).
Marsland and Eason (1973) present the results of a plate
loading test in which displacements were measured at various FIG. 21. Settlements and strains beneath a test footing on
depths during loading at a constant rate of penetration of Champlain Clay (Bauer et al. 1976).
BURLAND
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

l eve1
Frc. 22. Principles of the inclinometer vertical strain gauge
(Burland and Symes 1982).
2.5 mm/min. The plate was 865 mm in diameter and the
test was carried out at a site in central London at the base
of a 900 mm diameter, 17.5 m deep shaft. Figure 250 shows
the observed relationships between average pressure and
settlement for points at various depths beneath the plate.
Figure 25b shows the associated average strains deduced
from the differences in settlement at each load. Note the
much larger strains immediately beneath the plate. The
secant Young's modulus El, corresponding to each value of
For personal use only.

average strain can be evaluated using elastic theory to


calculate the appropriate stress changes beneath the plate.
The undrained strength at the depth of the plate test was FIG. 23. Inclinometer gauge.
200 kPa. Figure 26 shows a plot of Eu/Cu versus log E , for
the four levels beneath the plate. Also shown is the range ments of vertical shear strain in the clay against the shaft
of values from laboratory tests taken from Fig. 24c. The surface were made at various depths by means of precise
values of Eu/Cu from the three deepest levels are in excellent horizontal inclinometers. Figures 29a and 296 have been
agreement with the laboratory values. The values of Eu/Cu taken from their paper. These data can be used to derive
from the top level are very low and reflect bedding errors the relationships between rVhand y,, at various depths
between the plate and clay surface and also disturbance from which secant values of the shear modulus GVhcan be
effects caused by the drilling operation. calculated. Figure 300 shows the relationship between Gvh
Recently, my advice was sought on the design of a slab and log yVh for three depths remote from the tip of the
foundation on the surface of London Clay at a site in pile.'
Bracknell where very stringent limits had been set on For undrained plane strain conditions and zero horizontal
permissible differential settlements. In view of the importance strain, the major principal strain E , is equal to yvh/2. If the
and sensitivity of the project, it was decided to carry out assumption of isotropy is made, then Eu = 3GVh.In figure
a field loading test on a 3 m x 3 m area with settlement 30b, values of 3Gvh/Cuare plotted against log E , , thereby
observations at various depths beneath the centre of the allowing comparison with the range of laboratory results
loaded area. One hundred tonnes of kentledge in the form of Eu/Cu taken from Fig. 24c. It can be seen that the field
of concrete blocks was stacked on a thin, lightly reinforced values of 3GVh/Cu lie above the laboratory measure-
3 m x 3 m concrete pad cast on the ground surface and ments-the values differing by a factor of about two. The
settlements were measured at depths of 0 , 1, 2, 3, and 5 m most likely explanation for this difference is that the process
beneath the centre. Figure 27 shows the observed settlements of installation and subsequent consolidation has resulted in
2.8 h and 28 days after commencement of loading together an increase in the mean effective stresses in the clay adjacent
with the associated average vertical strains. As in the to the pile shaft, with a consequent increase in Gvh. In
previous case, the values of Eu corresponding to the addition, London Clay is known to be anisotropic with
average strains after 2.8 h were calculated using elastic Eh > Ev and it seems highly probable that GVhwill be
theory. Figure 28 shows the resulting values of Eu/Cu greater than Eu,/3 (where E,, is the undrained Young's
compared with the laboratory values taken from Fig. 24c. modulus in the vertical direction). In view of these effects,
Once again, the agreement is excellent. the comparison between field and laboratory measurements
The two field tests just described involve the loading of given in Fig. 30b is considered to be very reasonable.
footings. Cooke et al. (1979)present data for a completely On the basis of these direct comparisons between
different type of test in London Clay. They tested a 4.5 m laboratory and field measurements of stiffness, I have been
long, 168 mm diameter tubular steel pile that had been jacked
in place. The pile was instrumented with load cells at various
depths so that the shear stress transfer distribution at each 'strictly, the subscripts vr should be used, since the problem is
applied load was accurately. known. Also, accurate measure- axisymmetric and G,, is not necessarily equal to G,,.
5 10 CAN. GEOTECH. J. VOL. 26, 1989
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

Ax101 strains
indicated In %

Strains determined from external


d~splocernents corrected for load
For personal use only.

cell compliance

Axial strain (%I

Axial s t r a ~ n (%)

F I G . 24. Typical London Clay unconsolidated undrained test data from Canon's Park at 7.3 r n depth: (a) undrained stress path;
(b) initial stress-strain behaviour, and (c) stiffness-strain characteristics (Jardine et al. 1985).
Avg. Pressure (k P a )
I I

-E 10
0
, 1500 -
-E 610mm

-
C
L57mm
-
W

C
a 20 305mm g -
-5
I000
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

-
+.
a,
"l
m 8 6 5 m m Oto Plate 152mm
* 3 0
'
500

I I I l l I I ,
0 0: 0 i 10
Avg. Pressure ( k P a )
V e r t ~ c o l S t r a ~ n (%I

FIG. 26. Deduced stiffness-strain characteristics from results of


plate loading test in London Clay given in Fig. 25 (Jardine et a/.
1985).
Vertical s t r a t n (70) Settlement (rnrn)
For personal use only.

FIG.25. Settlements (a) and strains (b) beneath a plate test at


the bottom of a shaft in London Clay (Marsland and Eason 1973).

compelled to modify the views I held in the late 1970's about


the inappropriateness of laboratory tests for the measurement
of stiffness-at least for stiff materials. If the .strains are
measured locally to a high accuracy, tests on high-quality 1
samples at the appropriate confining pressures give D e p t h (rn)
remarkably accurate and consistent measurements of the FIG. 27. Distributions of settlement and strain beneath the
in situ small-strain stiffness. centre of a 3 m x 3 m test footing on the surface of London Clay
at Bracknell.
Fundamental studies
The inclinometer strain gauges described previously have dimensional conditions. From these data it has been possible
been used during the last few years in the Imperial College to identify the regions in stress space surrounding a given
Soil Mechanics Laboratory to carry out fundamental studies initial stress state for which the strains are small and the stiff-
of the behaviour of resedimented clays. Much of this work ness large. These are termed "small-strain regions" or
was carried out by Jardine (1985) and has been summarized "high-stiffness regions."
by Hight et al. (1985). An outcome of the work has been Jardine (1985) has shown from cyclic tests on a low-
clarification of our understanding of the factors that plasticity resedimented clay that there are three regions of
influence small-strain stiffness and, in particular, the stress change to be considered:
influence of stress history. Region I-axial strains I0.01 %, which give rise to straight
For each clay, blocks of reconstituted soil were formed undrained stress paths and no plastic strains.
by one-dimensional consolidation from a slurry using a Region 11-axial strains from about 0.01% to about 0.1'7'0,
225 mm diameter oedometer; the clay was allowed to swell which give rise to approximately straight undrained stress
to an overconsolidation ratio (OCR) between 2 and 4. paths, but the strains are plastic and exhibit creep.
Specimens cut from these blocks were then reconsolidated Region 111-axial strains > 0.1-0.2%, which lead to the
one dimensionally in the hydraulic triaxial cell (Bishop and onset of volumetric yield with sharp kinks in undrained stress
Wesley 1975) to a vertical effective stress at least 1.75 times paths.
the maximum value applied in the oedometer to remove the Figure 31 shows the "small-strain regions" for a
effects of sampling (Gens 1982). Specimens were then resedimented low-plasticity clay for the following three stress
sheared undrained in triaxial compression and extension histories:
from a normally consolidated state and from various over- (i)The soil was normally consolidated to point A under KO
consolidated states produced by swelling under one- (one-dimensional) conditions.
512 CAN. GEOTECH. J . VOL. 26, 1989

regions" are time dependent and that a rest period (ageing)


causes them to expand, probably because of the development
of interparticle bonds.
The concept of kinematic "small-strain regions" provides
a powerful insight into the influence of stress history of the
stiffness of soils at small strains and in particular the influence
of sampling. For example, by comparing the regions around
points A and B in Fig. 31 it can be seen that a given reduction
in, say, (a: + ui)/2 (keeping (a, - ar)/2 constant) will give
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

rise to a much stiffer response from the soil at stress state A


than at stress state B. This is because for stress state A the
stress path has to traverse a much longer length before reach-
ing the 0.01 or 0.1% contour than for stress state B. On
the other hand, a given increase in (a, - q ) / 2 (keeping
(0; + ur7/2 constant) will give rise to a much stiffer
Vert~calstrain (%) response from the soil at stress state B than at stress state A .
FIG. 28. Deduced stiffness-strain characteristics from results of Thus we see that even a perfectly sampled specimen (stress
footing test given in Fig. 27 (Jardine et a/. 1985). state B) will have a significantly different low-strain response
from the in situ material (stress state A).
(ii) From point A the shear stress was removed under The precise definition of the small-strain region requires
undrained conditions such that effective stresses followed more study. Jardine (1985) uses the magnitude of the major
the path AB with point B being the equilibrium condition. principal strains to define a region, whereas Simpson et al.
This particular stress history reproduces the "ideal" sampl- (1979) use an arbitrary definition of generalized strain. I am
ing process. of the opinion that defining the region in terms of incre-
(iii) From point A the soil was swelled back one dimensionally mental strain energy has many advantages.
to point C, giving an overconsolidation ratio of about 4.
For each of the stress histories represented by points A,
For personal use only.

B, and C, undrained triaxial compression and extension tests Concluding remarks


were carried out as well as a limited number of other tests This lecture has described a variety of studies of the
following various stress paths. behaviour of various types of ground at levels of strain that
Around each of the three states A, B, and C are marked are small but nevertheless represent the majority of practical
points in stress space corresponding to axial strains of 0.01 working conditions.
and 0.1 Yo. Two contours have been sketched in around each It is important to note the interplay that has taken ;lace
initial stress point to represent regions I (axial between analysis and experiment-both developing in
strain 5 (0.01 Yo) and regions I1 (axial strains s 0.1 Yo). parallel. Much of the interpretation of the field measure-
It can be seen from Fig. 31 that as an element of soil ments could not have been carried out without the aid of
experiences large changes of effective stress, the current numerical analysis. It is also important to recognize the inter-
stress state carries around with it a region of small strain active roles played by field measurements, in situ testing,
or high stiffness. It can be seen that the shapes and sizes and laboratory studies.
of these regions change with stress history, as does, most The finding that at small strains the ground exhibits high
importantly, their location around the current stress. There stiffness and, frequently, nonlinear properties has far-
is evidence to suggest that the sizes of the "low-strain reaching practical and fundamental consequences. For
Shear stress (kPo) Shear strain x
0 10 15 20 25 30 35 800 1200 1600

FIG. 29. Measured distributions down a pile o f (a) shear stress at pile surface and (b) shear strains in the clay against the pile surface
(Cooke et al. 1979).
BURLAND

50
(a)
O lrn depth
a
+ 0 2m d e p t h
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

Shear strain Y V H (%)

Major principal strain (%I


FIG.30. Deduced stiffness-strain characteristics from results of pile test given in Fig. 29 (Jardine et al. 1985).
For personal use only.

FIG. 31. Regions of small strain and high stiffness at three states of stress history for a resedimented clay with low plasticity index.
A, normally consolidated under KO conditions; B, unloaded from A undrained (ideal sampling); C, KO swell from A to an OCR =
4 (Jardine 1985).
example, Jardine et al. (1986)have shown that nonlinearity displacement corresponding to any specified load factor was
of stress-strain response has very significant effects on soil- known. The apparent secant Young's modulus was then
structure interaction, stress distributions in the soil mass, evaluated from elastic theory, as is the traditional approach
and displacement profiles around loaded areas and excava- to interpreting field measurements and in situ tests. Finally,
tions. Unless small-strain nonlinearity is recognized, the for each problem, the apparent Young's modulus (E;) was
interpretation of field measurements and in situ tests may plotted against load factor. Figure 32 shows pIots of E,~'/c,
be confusing and misleading. versus load factor for five boundary-value problems that
To illustrate the above statement, Jardine et a/. (1986) can be compared with the relationship given by an undrained
have carried out numerical analyses of a wide range of prob- triaxial compression test.
lems (a footing, a pile, a propped cantilever excavation, and It can be seen that, for a given value of load factor, the
an expanding cylindrical cavity) using a nonlinear soil model value of apparent Young's modulus is highly problem depen-
having a fixed undrained stress-strain curve. The soil model dent. Consider, for example, a load factor of 0.3. The values
represents a low-plasticity clay that is lightly overcon- of E$ for the footing and the strutted excavation are close
solidated. Though stiffer than many soils, it is not particularly to the value obtained from a triaxial compression test. In
nonlinear. contrast, the expanding cylindrical cavity gives a very much
For each problem the complete load-displacement relation- lower value of E:, whereas the rigid pile and pile element
ship was derived up to failure. Thus, for each problem, the give rnuch higher values of E;.
514 CAN. GEOTECH. J. VOL. 26, 1989

5000 r I Younqls modulus (M Pa x lo3)


o O _ lI 2I 3I LI 5I 6I 7I

e Plate loading tests

---analysis
Finite element back

-0-- Seismic refraction


Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

FIG.33. The chalk at Mundford-a comparison between E


values derived from static tank and plate tests (Fig. 66) and E values
from in sitir shear wave refraction measurements (Abbiss 1979).
Load factor LLI,, . PPI,
For personal use only.

FIG. 32. Relationship between apparent undrained stiffness


and load factor for various boundary value problems (Jardine et a/.
1986). \ o E ,,(0.01)from triaxial tests

Footings, piles, and expanding cavities form the basis of


the majority of in situ loading tests. It is clear from Fig. 32
that great care is needed when deducing apparent elastic
modulus values from such tests and applying them in design.
Interpretation is greatly assisted if a full load-displacement
curve has been obtained or, preferably, local strains within
the ground are measured, as for the case records described
in the previous section.
The development of laboratory techniques for the precise
measurement of small strains has assisted in closing the gap
between the dynamic and static measurement of ground
stiffness. In the past, dynamic measurements of Young's
modulus, or shear modulus, have tended to give results that
are so much higher than static values determined in the
laboratory that the dynamic values have been discounted.
However, in a number of recent cases, the accurately deter-
mined static small-strain values of stiffness have been found
to be very close to the values measured by dynamic methods.
For example, Fig. 33 shows a comparison between the
values of E deduced from the static tank loading test on the
chalk at Mundford described earlier and the values derived
from in situ shear wave refraction measurements (Abbiss
1981). Similarly, Fig. 34 shows a comparison between values
of Eu(o,ol)from undrained triaxial tests on samples of FIG.34. Comparison of laboratory values of E,,@,,,,and field
London Clay from Canon's Park and dynamic values of dynamic values obtained by Abbiss (1981), for London Clay.
Eu (= 3 x G) obtained by Abbiss (1981). The dynamic
values of E, are only about 30% greater than the values of erties, improved understanding of damping and rate effects,
Eu(0.01). and important developments in studies of wave propagation
These results open up the way for a whole new area of and dynamic soil-structure interaction.
study linking dynamic and static deformation properties. In the past, laboratory techniques of measuring strain
Such studies would lead to a wider application of geophysical have been remarkably crude. The studies described here have
measurements for the determination of in situ stiffness prop- shown that improvements in the measurement of local strain
BURLAND 515
have resulted i n valuable practical a n d f u n d a m e n t a l infor- wall movements associated with a large excavation. Proceedings,
mation o n t h e stiffness a n d yielding properties o f t h e soils 5th European Conference on Soil Mechanics, Madrid.
investigated. T h e r e is e n o r m o u s scope f o r t h e development COOKE,R.W., PRICE,G., and TARR,K. 1979. Jacked piles in
o f improved instruments and techniques and f o r the exper- London Clay: a study of load transfer and settlement under
imental study o f stiffness properties o f soils a n d weak rocks working conditions. Geotechnique, 29: 113-147.
COSTA-FILHO,L.F., and VAUGHAN,P.R. 1980. Discussion on
a t small strains. A computer model for the analysis of ground movements in
I h o p e t h a t i n t h e course o f this lecture I have been suc- London Clay. Geotechnique, 30: 336-339.
cessful i n conveying m y belief that, insofar as strains a r e GENS,A. 1982. Stress-strain and strength characteristics of a low
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

concerned (and in contrast t o t h e length o f t h e lecture), plasticity clay. Ph.D. thesis, University of London, London,
S M A L L IS B E A U T I F U L . United Kingdom.
HIGHT,D.W., GENS,A., and JARDINE, R.J. 1985. Evaluation of
ABBISS,C.P. 1979. A comparison of the stiffness of the chalk at geotechnical parameters from triaxial tests on offshore clay.
Mundford from a seismic survey and a large scale tank test. Proceedings, International Conference on Offshore Site
Geotechnique, 29: 461-468. Investigation, Society for Underwater Technology, London,
1981. Shear wave measurements of the elasticity of the pp. 253-268.
ground. Geotechnique, 31: 91-104. HVORSLEV,M.J. 1937. Uber die Festigkeitseigenschaften
ACKERLEY, S.K., HELLINGS,J.E., and JARDINE,R.J. 1987. Gestorter Bindiger Boden. Ingerniorvidenshabelige Skriften,
Discussion on A new device for measuring local axial strains on no. 45, Danmarks Naturvidenshabelige Samfund, Kopenhavn,
triaxial specimens. Geotechnique, 37: 414-415. Denmark.
ATKINSON, J.H., and BRANSBY, P.L. 1978. The mechanics of soils. JARDINE, R.J. 1985. Investigations of pile-soil behaviour with
McGraw-Hill, Maidenhead, United Kingdom. special reference to the foundations of offshore structures. Ph.D.
ATTEWELL, P.B., and FARMER,I.W. 1974. Ground deformations thesis, University of London, London, United Kingdom.
resulting from shield tunnelling in London Clay. Canadian JARDINE,R.J., SYMES,M.J., and BURLAND,J.B. 1984. The
Geotechnical Journal, 11: 380-395. measurement of soil stiffness in the triaxial apparatus.
BAUER,G.E., SHIELDS,D.H., MCROSTIE,G.C., and SCOTT,J.D. Geotechnique, 34: 323-340.
1976. Predicted and observed footing settlements in a fissured JARDINE, R.J., FOURIE,A., MASWOSWE, J., and BURLAND, J.B.
clay. Proceedings, Conference on Performance of Building 1985. Field and laboratory measurements of soil stiffness.
Structures, Glasgow. Pentech Press, London, United Kingdom, Proceedings, 1lth International Conference on Soil Mechanics
For personal use only.

pp. 287-302. and Foundation Engineering, San Francisco, vol. 2, pp. 5 11-5 14.
BISHOP, A.W., and WESLEY,L.D. 1975. A hydraulic triaxial JARDINE, R. J., POTTS,D.M., FOURIE,A.B., and BURLAND, J.B.
apparatus for controlled stress path testing. Geotechnique 25: 1986. Studies of the influence of non-linear stress-strain
657-670. characteristics in soil-structure interaction. Geotechnique, 36:
BISHOP,A.W., WEBB,D.L., and LEWIN,P.I. 1965. Undisturbed 377-396.
samples of London Clay from the Ashford Common shaft; KRIEGEL,H.J., and WEISNER,H.H. 1973. Problems of stress-
strength-effective stress relationships. Geotechnique, 15: 1-3 1. strain conditions in subsoil. Proceedings, 8th International
BURLAND,J.B. 1979. Contribution to discussion-Session 4. Conference on Soil Mechanics and Foundation Engineering,
Proceedings, 7th European Conference on Soil Mechanics and vol. 1.3, pp. 133-141.
Foundation Engineering, Brighton, vol. 4, p. 137. MACCARINI, M. 1987. Laboratory studies of a weakly bonded
BURLAND, J.B., and DAvIDsON, W. 1976. A case study of crack- artificial soil. Ph.D. thesis, University of London, London,
ing of columns supporting a silo due to differential foundation United Kingdom.
settlement. Proceedings, Conference on Performance of Building MARSLAND,A , , and EASON, B.J. 1973. Measurements of
Structures, Glasgow. Pentech Press, London, United Kingdom, displacements in the ground below loaded plates in deep
pp. 249-267. boreholes. Proceedings, Symposium on Field Instrumentation,
BURLAND, J.B., and HANCOCK,R.J.R. 1977. Underground car London. Butterworths, London, United Kingdom, pp. 304-317.
park at the House of Commons, London: geotechnical aspects. RENDULIC, L. 1937. Ein Grundgesetz der Tonmechanik und sein
Structural Engineer, 55: 87-100. experimenteller Beweis. Bauingenieur, 18: 459-467.
BURLAND,J.B., and KALRA,J.C. 1986. Queen Elizabeth I1 ROSCOE,K.H., and BURLAND, J.B. 1968. On the generalised
Conference Centre: geotechnical aspects. Proceedings-the stress-strain behaviour of 'wet' clay. Symposium on Engineering
Institution of Civil Engineers, 80: 1479-1503. Plasticity. Cambridge University Press, Cambridge, United
BURLAND, J.B., and LORD, J.A. 1969. The load deformation Kingdom, pp. 535-609.
behaviour of Middle Chalk at Mundford, Norfolk: a comparison SCHOFIELD,A.N., and WROTH,C.P. 1968. Critical state soil
between full-scale performance and in situ and laboratory mechanics. McGraw-Hill, London, United Kingdom.
measurements. Proceedings, Conference on In-Situ Investigations SIMPSON, B., O'RIORDAN,N.J., and CROFT, D.D. 1979.
in Soils and Rock. British Geotechnical Society, London, United A computer model for the analysis of ground movements in
Kingdom, 3-15. London Clay. Geotechnique, 29: 149-175.
BURLAND, J.B., and SYMES,M. J . 1982. A simple axial displace- SKEMPTON, A.W. 1961. Horizontal stresses in an overconsolidated
ment gauge for use in the triaxial apparatus. Geotechnique, 32: Eocene clay. Proceedings, 5th International Conference on Soil
62-65. Mechanics and Foundation Engineering, Paris, vol. 1,
BURLAND, J.B., SILLS,G.C., and GIBSON,R.E. 1973. A field and pp. 351-387.
theoretical study of the influence of non-homogeneity on settle- SMITH, P.D.K., and BURLAND, J.B. 1976. Performance of a high
ment. Proceedings, 8th International Conference on Soil precision multi-point borehole extensometer in soft rock.
Mechanics and Foundation Engineering, Moscow, vol. 1.3, Canadian Geotechnical Journal, 13: 172-176.
pp. 39-46. STEVENS,A , , CORBETT,B.O., and STEELE,A.J. 1977. Barbican
CLAYTON, C.R.I., and KHATRUSH, S.A. 1986. A new device for Arts Centre: the design and construction of the substructure.
measuring local axial strains on triaxial specimens. Geotechnique, Structural Engineer, 55: 473-485.
36: 593-597. ST. JOHN, H.D. 1975. Field and theoretical studies of the
COLE,K.W., and BURLAND, J.B. 1972. Observations of retaining behaviour of ground around deep excavations in London Clay.
516 CAN. GEOTECH. J. VOL. 26, 1989

Ph.D. thesis, University of Cambridge, Cambridge, United tion methods against the results of a detailed geotechnical
Kingdom. investigation of Middle Chalk at Mundford, Norfolk.
VAUGHAN, P.R. 1985. General report; Session 2. Mechanical and Proceedings, Conference on In-Situ Investigations in Soils and
hydraulic properties of tropical lateritic and saprolitic soils, Rock. British Geotechnical Society, London, United Kingdom,
particularly as related to their structure and mineral components. pp. 17-22.
Proceedings, 1st International Conference on Geomechanics in WARD, W.H., BURLAND,J.B., and GALLOIS,R.W. 1968.
Tropical Lateritic and Saprolitic Soils, Brasilia, vol. 3, Geotechnical assessment of a site at Mundford, Norfolk for a
pp. 231-263. large proton accelerator. Geotechnique, 18: 399-431.
WAKELING, T.R.M. 1965. Foundations on chalk. In Chalk in WROTH,C.P. 1972. Some aspects of the elastic behaviour of over-
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 11/10/12

earthworks and foundations. Institution on Civil Engineers, consolidated clay. Proceedings, Roscoe Memorial Symposium.
London, United Kingdom. Foulis, Yeovil, United Kingdom, pp. 347-361.
-1969. A comparison of the results of standard site investiga-
For personal use only.

You might also like