You are on page 1of 17

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/257564370

Undrained Shear Strength of Normally Consolidated and Overconsolidated


Clays from Pressuremeter Tests: A Case Study

Article  in  Geotechnical and Geological Engineering · October 2013


DOI: 10.1007/s10706-013-9675-x

CITATION READS

1 1,127

1 author:

Ali Soleimanbeigi
University of Wisconsin–Madison
27 PUBLICATIONS   97 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Mechanically Stabilized Earth (MSE) Wall Backfill Water Infiltration View project

All content following this page was uploaded by Ali Soleimanbeigi on 05 October 2014.

The user has requested enhancement of the downloaded file.


Undrained Shear Strength of Normally
Consolidated and Overconsolidated Clays
from Pressuremeter Tests: A Case Study

Ali Soleimanbeigi

Geotechnical and Geological


Engineering
An International Journal

ISSN 0960-3182
Volume 31
Number 5

Geotech Geol Eng (2013) 31:1511-1524


DOI 10.1007/s10706-013-9675-x

1 23
Your article is protected by copyright and all
rights are held exclusively by Springer Science
+Business Media Dordrecht. This e-offprint
is for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

1 23
Author's personal copy
Geotech Geol Eng (2013) 31:1511–1524
DOI 10.1007/s10706-013-9675-x

ORIGINAL PAPER

Undrained Shear Strength of Normally Consolidated


and Overconsolidated Clays from Pressuremeter Tests:
A Case Study
Ali Soleimanbeigi

Received: 22 March 2013 / Accepted: 3 July 2013 / Published online: 10 July 2013
Ó Springer Science+Business Media Dordrecht 2013

Abstract Undrained shear strength (su) of founda- a factor of 1.5 for NC clay and 2.2 for OC clay. The
tion soil of Marquette interchange near Milwaukee, results of finite element analysis considering Cam-
Wisconsin was evaluated from the results of a number Clay soil model and finite length for pressuremeters
of pressuremeter tests conducted on normally consol- suggest the undrained shear strength of 63 ± 7 kPa for
idated (NC) organic silts and overconsolidated (OC) NC organic silt and 259 ± 68 kPa for OC silty clay.
silty clay. The su-values were interpreted from tradi-
tional closed-form methods. The pressuremeter geom- Keywords Pressuremeter test  Undrained shear
etry and test sequence as well as response of the soil strength  Finite element  Cam-Clay  Foundation
profiles were also simulated using axisymmetric finite
element (FE) method with Cam-Clay soil model. The
Cam-Clay model parameters were estimated from
laboratory tests on undisturbed soil samples. Results 1 Introduction
show that the su estimated from the rate of cavity
pressure change with volumetric strain (referred to as Over the past 20 years there has been increasing
direct traditional method) is almost twice the su reliance on in situ testing rather than laboratory testing
estimated from an indirect traditional method that due to increasing availability of more reliable in situ
estimates su from shear modulus, in situ horizontal test interpretation methods that make in situ testing
stress, and ultimate cavity pressure obtained from the less time-consuming and more economical than lab-
cavity pressure curves. The su-values predicted from oratory testing (Salgado 2008). Different available
the FE models are lower than those estimated from the in situ tests based on their perceived applicability,
traditional methods and shows that the assumption of different ground conditions and relevant geotechnical
infinite pressuremeter length in traditional methods information obtained from each test were categorized
results in overprediction of undrained shear strength by and based on the ratings assigned to each in situ test
method, the pre-bored pressuremeter test is the only
method applicable to most ground conditions varying
from soft soil to hard rock (Robertson 1986). The
pressuremeter device is used to estimate shear
A. Soleimanbeigi (&) strength, deformation characteristics, in situ horizon-
Department of Civil and Environmental Engineering,
tal stress and permeability of the soil (Ménard 1956).
University of Wisconsin-Madison, 2243 Engineering Hall,
1415 Engineering Drive, Madison, WI 53706-1691, USA However, similar to other in situ test procedures, soil
e-mail: ali.soleiman@gmail.com properties estimated from pressuremeter test data are

123
Author's personal copy
1512 Geotech Geol Eng (2013) 31:1511–1524

affected by several sources of errors including (1) soil the cavity pressure and volumetric strain for undrained
disturbance (collapse, erosion, or softening of bore- clay (Gibson and Anderson 1961). As the cavity
hole walls) associated with boring and equipment pressure increases, an expanding annulus of plastic soil
installation technique (2) equipment calibration asso- develops around the cavity. During this phase the cavity
ciated with test procedure (e.g., volume change pressure is obtained using the following relationship:
applied pressure), and (3) method of interpretation
pca ¼ rh0 þ su ½1 þ lnðG=su Þ þ su ½lnðDV=V Þ ð1Þ
including models for soil response, initial and bound-
ary conditions (Baquelin et al. 1978; Mair and Wood where pca is the applied pressure by the pressuremeter;
1987). The traditional method used to interpret the rh0 is the in situ horizontal stress; G is the shear
undrained shear strength (su), models the soil as an modulus, and DV=V is the volumetric strain. The
elastic-perfectly plastic material in plain strain condi- limiting conditions at which all the surrounding soil
tion (i.e., infinite length for pressuremeter probe) to deforms plastically is assumed to be reached when
obtain a simplified closed-form solution. However, the DV=V ¼ 1 (Gibson and Anderson 1961), therefore the
interpreted su-values are often considerably larger than third expression in Eq. (1) is omitted and the limiting
su-values obtained from other in situ or laboratory test cavity pressure (pL) is obtained by:
results (Carter et al. 1979; Wroth 1982; Mair and
pL ¼ rho þ su ½1 þ lnðG=su Þ ð2Þ
Wood 1987; Yu 1990; Houlsby and Carter 1993; Yu
et al. 2005). No studies have been conducted to specify By substitution of Eq. (2) into Eq. (1), we obtain:
the ratio of the su-values interpreted from closed-form
pca ¼ pL þ su lnðDV=VÞ ð3Þ
solutions to the su-values from a method which takes
into account the finite pressuremeter length and more From Eq. (3), if pca is plotted versus lnðDV=VÞ, the
precise soil models in the light of field measurements. plastic phase of the curve lies on a straight line with
In this paper undrained shear strength of foundation gradient su as shown in Fig. 1c. In this paper, this
soil in Marquette Interchange located near Milwaukee, approach of estimating su proposed by Gibson and
Wisconsin was evaluated from the results of a number Anderson (1961), is referred to as direct traditional
of pressuremeter tests. The su-values were estimated method. The limiting pressure pL is intercepted as pca
using the traditional closed-form solution. The soil corresponding to lnðDV=V Þ ¼ 0 as illustrated in
profile, pressuremeter test sequence, and geometry Fig. 1c.
conditions were also simulated using axisymmetric Alternatively, su can be obtained from Eq. (2).
finite element (FE) method and Cam-Clay soil model Rewriting Eq. (2) for su gives:
to evaluate the effect of geometry condition and the
soil model on the predicted undrained shear strength. pL  rho
su ¼   ð4Þ
The undrained shear strengths interpreted from the 1 þ ln sGu
traditional method were compared to those estimated
from the FE method. The su-value from Eq. (4) can be obtained from trial
and error after estimating values of pL, G, and rho . This
alternative method is referred to as indirect traditional
method. The pL-value can be estimated from the
2 Background y-intercept of horizontal asymptote of pca in the cavity
pressure curve (Fig. 1b) or from the pca  lnðDV=V Þ
The basis of the pressuremeter test is radial expansion of curve (Fig. 2b). The G-value is obtained from the
a long cylindrical membrane installed in a borehole as unload–reload loop in the cavity pressure curve using
illustrated in Fig. 1a. The pressure is applied inside a the following relationship (Palmer 1972; Ladayani
cylindrical membrane and volume change of the 1973):
expanding membrane is measured. The in situ properties
1 dpca
of the soil are then interpreted from the pressure versus G¼ ð5Þ
2 dec
volumetric or radial strain curve (Fig. 1b). Cavity
expansion problem in an elastic-perfectly plastic where ec is the cavity (radial) strain and is obtained
medium was solved to obtain a relationship between from:

123
Author's personal copy
Geotech Geol Eng (2013) 31:1511–1524 1513

Fig. 1 Schematic of Control unit and Cavity pressure, p ca


pressuremeter test device pressure source
(a), typical cavity pressure pL
curve (b), and cavity
pressure versus ln(DV/V) to Borehole
estimate undrained shear
strength (c)

Measuring cell Guard cells


A

Volumetric or radial strain,εc


(a) (b)

pL
Cavity pressure, p ca

su
1

ln(ΔV/V) 0
(c)

ec ¼ ð1  DV=VÞ1=2  1 ð6Þ other three boreholes, PB-4, PB-5, and PB-6 were
located in over-consolidated (OC) silty clay soil.
The G-value is therefore half of the slope of Figure 2a, b show sample soil profile, soil physical
unloading–reloading loop in the pca–ec curve. The rh0 properties, depths of water table, and the pressureme-
can be estimated from the point on the pca–ec curves ter test spots in PB-3 and PB-6 respectively. The soil
where the elastic response of the soil starts after the profiles at other borehole locations are shown in
initial flat part (point A in Fig. 1b). The ratio sGu ¼ Ir is Fig. 12 in the ‘‘Appendix’’ section. The soil descrip-
termed the rigidity index of the soil. The pca–ec curves tion and depth of the pressuremeter tests at each
were also predicted using finite element method to borehole are summarized in Table 1. Corrections for
back-calculate parameters of Cam-Clay soil model. pressure and volume were applied to the pressureme-
Finite element models will be discussed later in this ter test results (Roscoe and Schofield 1963). Radial
paper. cavity strains (ec) were calculated from the measured
volumetric change of each pressuremeter using
Eq. (6). Figure 3 shows the pca–ec curves from the
3 Pressuremeter Test Reslts pressuremeter tests in boreholes PB-3 and PB-6. The
cavity pressure curves for the rest of the pressuremeter
The Menard type pressure-controlled pressuremeter tests are shown in Fig. 13 in the ‘‘Appendix’’ section.
tests were performed in six boreholes in Maquette The initial flat portion of each curve indicates the soil
Interchange site near Milwaukee, Wisconsin. Three disturbance during borehole preparation and pressure-
boreholes, PB-1, PB-2 and PB-3, were located in meter installation. Therefore, unload-reload cycles are
normally consolidated (NC) organic clay soil and the performed to remove the effect of pressuremeter

123
Author's personal copy
1514 Geotech Geol Eng (2013) 31:1511–1524

Fig. 2 Boring profiles for pressuremeter tests: a PB-3, and b PB-6. (x = in situ water content, eo = in situ void ratio, Gs = specific
gravity, c = in situ unit weight of soil)

installation disturbance and measure the undrained Alternatively, the su-values of NC organic silts and
shear modulus. OC silty clays were obtained by indirect traditional
To estimate su of NC organic silts and OC silty method using Eq. (4) by trial and error. The best
clays using the direct traditional method (Gibson and estimate of rh0, G, and pL, obtained from the pca–ec
Anderson 1961), the pca  lnðDV=V Þ curves were curves were used to estimate the su-values from
generated from cavity pressure curves and plotted in Eq. (4). The G-values were calculated from half of the
Fig. 4. The unload–reload cycles were omitted from slope of unload-reload loop in each pca–ec curve using
the curves. Initial flat part of the curves reflects the soil Eq. (5). The approximate in situ horizontal stresses
disturbance during boring and pressuremeter installa- were estimated from the point on the pca–ec curves
tion in the boreholes and does not affect the slope of where the elastic response of the soil starts after the
the straight portion of the curves (su). The estimated initial flat part. The pL-values were estimated from the
su-values obtained from the slope of the best fitting- y-intercept of horizontal asymptote of pca in the pca–ec
line over the data points after the initial flat part of the curves. The estimated rh0, G, and pL at different
pca  lnðDV=V Þ curves are summarized in column 8 pressuremeter test locations are summarized in col-
of Table 1. The su-value for the NC organic silt is umns 5–7 of Table 1.
184 ± 106 kPa and for the OC silty clay is The estimated values of su from the indirect
1,164 ± 406 kPa. traditional method are summarized in column 9 of

123
Table 1 Soil description and engineering properties from pressuremeter tests
Soil type Boring Average Description G (kPa) rho (kPa) pL (kPa) su (kPa)— su (kPa)— k j eo su (kPa)—FE
name depth (m) direct indirect method
method method
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)

NC organic silt PB1 3.3 OL-very loose 3,360 65 380 67 63 0.20 0.010 1.55 58
c = 20 kN/m3 4.8 OL-very loose, LL = 48, PI = 14 1,760 130 460 170 81 0.35 0.035 1.17 62
Geotech Geol Eng (2013) 31:1511–1524

6.0 OL-very loose 3,400 95 650 239 130 0.35 0.0175 1.17 71
11.7 OL-very soft 5,800 205 590 163 71 0.35 0.070 1.29 76
PB2 4.8 OL-very loose 2,100 95 490 168 97 0.35 0.0175 2.24 60
6.2 SM-very loose 2,090 105 490 171 94 0.35 0.01 0.46 59
10.3 OH-very soft, LL = 88, PI = 21 2,580 200 480 83 58 0.35 0.0175 2.86 56
PB3 4.8 CL-medium stiff 6,120 150 850 452 148 0.03 0.003 0.46 70
6.3 OH-soft 2,350 140 550 178 98 0.20 0.020 2.24 65
8.6 OH-soft, LL = 56, PI = 18 2,140 165 600 150 109 0.15 0.025 1.65 56
PB4 1.1 HF/ML-medium dense to dense 8,170 140 2,400 1,234 636 0.04 0.005 0.32 390
OC silty clay 3.3 CL-medium stiff to stiff 5,820 105 2,000 1,010 570 0.08 0.01 0.49 270
c = 21 kN/m3 5.1 CL-medium stiff to stiff 9,360 165 1,800 807 392 0.08 0.004 0.49 240
8.5 CL-medium stiff to stiff 1,0340 225 3,050 1,310 791 0.08 0.004 0.49 330
Author's personal copy

PB5 3.3 ML-medium dense 9,830 160 3,150 1,610 874 0.03 0.005 0.40 295
4.8 CL-medium stiff 11,000 250 2,100 1,328 438 0.06 0.004 0.49 143
6.6 ML-loose 9,000 190 2,600 1,182 670 0.07 0.0035 0.48 230
12.1 CL-very stiff to hard 8,950 325 3,800 2,042 1,133 0.02 0.007 0.48 313
PB6 1.1 HF/CL/ML-stiff 6,950 70 1,400 538 328 0.03 0.002 0.41 173
5.3 CL–ML-soft to medium stiff 1,9850 190 2,400 923 465 0.04 0.003 0.59 249
7.4 CL–ML-soft to medium stiff 1,1680 215 2,200 739 472 0.12 0.007 0.59 276
13.6 CL–ML-soft to medium stiff 9,750 280 2,900 1,243 729 0.12 0.01 0.54 258
1515

123
Author's personal copy
1516 Geotech Geol Eng (2013) 31:1511–1524

Fig. 3 Measure and predicted cavity pressure curves from


pressuremeter tests: borehole PB-3 (a) and borehole PB-6 (b) Fig. 4 Cavity pressure versus natural log of volume increment
ratio for pressuremeter tests: borehole PB-3 (a), and borehole
PB-6 (b)
Table 1. The interpreted su-value for the NC organic
silt is 95 ± 29 kPa and for the OC silty clay is
579 ± 176 kPa.The average su-values from the indi- flat part in pca  lnðDV=V Þ plane (Fig. 4b). This
rect traditional method are lower than those estimated method is commonly followed in design practice.
from the direct traditional method. Additionally, the However, the slope of the majority of the pca 
indirect traditional method results in lower standard lnðDV=V Þ curves decreases with increasing the cavity
deviation about the average su. Figure 5 compares the strain which shows lower shear strength at larger
su-values obtained from the two methods. The data deformations. Less uncertainty surrounds the deter-
points show a linear trend with coefficient of deter- mination of the apparent large strain shear strength [or
mination (R2) of 0.95. The su-values from the indirect ultimate strength, su,ult in Fig. 4b given by the slope of
traditional method are approximately half of the the pca  lnðDV=V Þ curve at larger deformations as
su-values obtained from the direct traditional method. the influence of possible disturbance during pressure-
The su-value from the direct traditional method meter installation and uncertainty concerning the true
determined in this case study represents the slope of reference condition is smaller (Mair and Wood 1987).
the best fitting-line over the data points after the initial The average su-values obtained from the slope of the

123
Author's personal copy
Geotech Geol Eng (2013) 31:1511–1524 1517

determined from either the slope of the pca 


lnðDV=V Þ curves at large strain (su,ult) or from the
indirect traditional method. With the incorporation of
three additional soil properties to estimate the su, (i.e.,
shear modulus, in situ horizontal stress, and ultimate
cavity pressure), the indirect traditional method is
considered more reliable and therefore suggested for
estimation of undrained shear strength from the
pressuremeter test results.

4 Cam-Clay Soil Model

The elastic-perfectly plastic soil model used in


traditional method does not take into account the
Fig. 5 Comparison of undrained shear strength between effect of stress history on the soil behavior. Critical
traditional methods state models are capable of considering effects of
stress history and therefore more appropriate for
best fitting-line is 28 % higher than the average modeling behavior of overconsolidated soils (Roscoe
su,ult-values obtained in this study. Previous studies and Schofield 1963; Roscoe and Burland 1968; Wood
also reported that the interpreted su-values from the 1990; Yu 2000). The modified Cam-Clay model is a
direct traditional method (Gibson and Anderson 1961) volumetric hardening–softening model based on the
are often considerably larger than su-values obtained critical state concept and has been widely used to
from other field and laboratory test results (Carter et al. simulate behavior of NC and OC clay (Roscoe and
1979; Wroth 1982; Mair and Wood 1987; Yu 1990; Burland 1968; Wood 1990). Figure 6 shows the
Houlsby and Carter 1993; Yu et al. 2005). Mair and schematic of the modified Cam-Clay model and the
Wood (1987) recommended the su-values be model parameters. The yield surface of the model in

Fig. 6 Elliptical yield


surface for Cam-Clay model
in p0 –q plane (a) normal
compression line and
unloading–reloading line in
compression plane (b,
c) (After Wood 1990)

123
Author's personal copy
1518 Geotech Geol Eng (2013) 31:1511–1524

Table 2 Range of Cam-Clay model parameters from labora- Axis of symmetry Δu=0
tory tests
NC organic clay OC silty clay

k 0.069–0.630 0.004–0.083
j 0.003–0.074 0.0004–0.007 Δu=0
Mc 1.2 1.2
OCR 1–2 2–15
eo 0.81–3.78 0.28–0.72

po

Fig. 8 Typical finite element model of borehole PB-6 for each


pressuremeter test

NC organic silt and OC silty clay specimens taken


from the foundation of the Marquette Interchange site.
One-dimensional consolidation tests following ASTM
2435 were conducted to obtain k and j. The initial
void ratios (eo) were obtained following ASTM D
Fig. 7 Stress points at failure measured from triaxial compres- 4959. Table 2 summarizes the range of k and j for NC
sion tests
organic silts and OC silty clays. The critical state
friction angle (/c) of the soil was estimated from
consolidated undrained (CU) triaxial compression
p0 –q (deviator stress-mean effective stress) plane has tests on undisturbed soil samples with pore-water
an elliptical shape. After soil yielding, the yield pressure measurement following ASTM D 4767. The
surface expands to model stress hardening (increase of deviator stress versus mean effective stress values
deviator stress after yielding) behavior of NC soil but were plotted in Fig. 7. The slope of the best-fitting line
shrinks to model stress softening (reduction of devi- over the p0 –q points is Mc = 1.2. The /0c of both NC
ator stress after yielding) of OC soil. The basic model organic silts and OC silty clays is 31o as obtained
parameters include size of the elliptical yield surface, from:
p0c (i.e., the effective overconsolidation pressure);
slope of the virgin compression line, k; slope of the
Mc ¼ 6sin/0c =ð3  sin/0c Þ ð7Þ
swelling line, j; and the slope of the critical state line
(CSL), Mc, in p0 –q plane. Poisson’s ratio (l) and shear
modulus (G) are used to predict elastic response of soil The initial void ratios were determined from
in the modified Cam-Clay model. Details of the model measurement of in situ water content, specific gravity
is described in (Wood 1990). and total unit weight at each borehole and summarized
The model parameters were obtained from consol- in Table 2. The modified Cam-Clay model parameters
idation and triaxial compression tests on undisturbed were further adjusted to match the predicted cavity

123
Author's personal copy
Geotech Geol Eng (2013) 31:1511–1524 1519

pressure curves to those measured form pressuremeter


test data.

5 Development of Finite Element Models

Axisymmetric finite element (FE) analyses were


conducted using ABAQUS software (ABAQUSTM
2008) to model the boreholes and pressuremeter tests
in NC organic silts and OC silty clays. Figure 8 shows
a sample FE model. Only one half of each borehole
section was modeled due to symmetry. The radius of
each borehole is 0.05 m. The length of the expanding
pressuremeter probe is 0.75 m. The width of each FE
model is 3.0 m after trials to minimize the end effects
on the stresses induced by the probe pressure. The
Fig. 9 Typical time-dependent pressure boundary applied to
height of each model is equal to depth of each soil by pressuremeter
pressuremeter test conducted in the borehole plus
an appropriate extra distance to minimize end
effects on the induced stresses. The initial condi- duration time for each pressure level is 60 s after
tions are in situ vertical and horizontal stresses. To which the volume change of the probe is recorded. The
estimate the in situ horizontal stress, the coeffi- duration time to increase the pressure from a previous
cients of lateral earth pressure at rest (K0) for NC level is 10 s. The type of the finite element is 8-noded
organic silt and OC silty clay at each borehole quadrilateral axisymmetric with pore pressure. The
were obtained from the following equation (Mayne element size was refined near the pressuremeter test
and Kulhawy 1982): location and becomes coarser with increasing dis-
tance. The size of the elements along the test intervals
0
K0 ¼ ð1  sin /0 ÞOCRsin / ð8Þ was adjusted based on the convergence of the numer-
ical predictions.
Boundary conditions applied to each FE model The Cam-Clay model parameters measured from
include: (1) zero rotation about vertical axis and about the laboratory test results (Table 2) were assigned
the vector normal to the plane on the axis of symmetry; to soil profile in each FE model. For each FE
zero horizontal displacement on the axis of symmetry model, the time-dependent pressure of each pres-
and on the right side of the model; zero horizontal and suremeter test was applied to the cavity wall at the
vertical displacement on the bottom side; (2) no test location and the cavity strains during the test
displacement constraints along wall of the borehole; were predicted.
(3) zero pore water pressure on the ground surface and
on the borehole wall except the depth interval over
which the pressuremeter test is conducted, and (4) the 6 Finite Element Analysis Results
time-dependent pressure of each pressuremeter test as
illustrated in Fig. 9. The probe pressures at each step The predicted cavity pressure curves from the FE
were normalized with respect to the maximum applied models were compared to those measured from
pressure during the test. The pressures were corrected pressuremeter tests. The Cam-Clay model parameters
to account for the volume change and pressure loss were slightly adjusted to calibrate the FE models such
within the pressuremeter system (Ménard 1956; that for each pressuremeter test, the predicted cavity
Baquelin et al. 1978; Mair and Wood 1987). The pressure curves match the measured ones. Columns

123
Author's personal copy
1520 Geotech Geol Eng (2013) 31:1511–1524

Fig. 11 Comparison between traditional method and FE


method

qf
su ¼ ð10Þ
2
where qf is the maximum deviator stress determined
from the q–ec curves. The predicted cavity pressure
curves for the pressuremeter tests conducted in
boreholes PB-3 and PB-6 are shown in Fig. 3a, b by
the dotted curves. The predicted cavity pressure curves
approximately match the measured ones. Similar
cavity pressure curves were predicted for the remain-
der of the pressuremeter tests (PB-1, PB-2, PB-4, PB-
5, and PB-5) as shown in Fig. 13 in the ‘‘Appendix’’
Fig. 10 Deviator stress versus cavity strain at the cavity wall in section. The deviator stress at failure and thus
borehole PB-3 (a) and PB-6 (b) su-values were determined from the deviator stress-
cavity strain (q–ec) curves as shown in Fig. 10. The
predicted su-values for NC organic silt and OC silty
10–12 in Table 2 contain k, j and eo, assigned to soil clays from the pressuremeter tests in six boreholes are
profile in FE models. In the cavity expansion problem, summarized in column 13 of Table 1. The predicted
the radial stress (rr) and circumferential stress (rh) at su-value for the NC organic silt is 63 ± 7 kPa and for
the cavity wall are the maximum and minimum the OC silty clay is 259 ± 68 kPa.
principal stresses respectively. Values of rr and rh The su-values predicted from FE models are smaller
during each simulated pressuremeter test were pre- than those interpreted from the traditional methods.
dicted from the FE analyses. The deviator stress is Figure 11 compares the su-values for NC organic silt
defined as: and OC silty clay interpreted from indirect traditional
method to those predicted from the FE method. The
q ¼ rr  rh ð9Þ
su-values of NC organic silt interpreted from the
The deviator stress versus cavity strain (q–ec) traditional method is 50 % higher than those predicted
curves during each test were plotted and the su-values using FE method. Similar results were reported by
were obtained from: Houlsby and Carter (1993). They performed finite

123
Author's personal copy
Geotech Geol Eng (2013) 31:1511–1524 1521

element analysis of pressuremeter tests with finite undrained shear strength of the foundation soil in
length in the soil following elastic-perfectly plastic Marquette Interchange site. Traditional closed-form
Tresca criterion and concluded that the assumption of solution was used to interpret the undrained shear
infinite pressuremeter length results in overestimation strength from the test results. The pressuremeter test
of the su by a factor of approximately 1.25–1.43 for procedure and soil profiles were also simulated using
typical clays. axisymmetric finite element method with Cam-Clay
The overprediction of su-values interpreted from model for soil response. The following findings were
the traditional method is higher for OC silty clay. concluded from the results of this study:
Figure 11 shows that the overprediction of su-values
1. The undrained shear strength interpreted from the
for OC silty clay due to assumption of infinite length
slope of cavity pressure versus logarithm of
for the pressuremeter is about 70 % higher than that
volumetric strain rate (referred to as direct
for NC organic silt. The results obtained from analysis
traditional method), is almost twice the undrained
of the pressuremeter test results in this case study
shear strength interpreted from the indirect tradi-
agree with the previous understanding that the inter-
tional method, where in situ horizontal stress,
preted shear strength using conventional elastic-per-
shear modulus and ultimate cavity pressure mea-
fectly plastic methods is higher than the shear strength
sured from the cavity pressure curve are taken into
measured from other in situ or lab test results due to
account.
simplifying assumption of infinite pressuremeter
2. The su-value interpreted from the direct tradi-
length in traditional method (Mair and Wood 1987;
tional method is 184 ± 106 kPa for NC organic
Wroth 1982; Yu 1990; Houlsby and Carter 1993; Yu
silt and 1,164 ± 406 kPa for the OC silty clay.
et al. 2005; Yeung and Carter 1990; Shuttle and
The su-value interpreted from the indirect tradi-
Jefferies 1995).
tional method is 95 ± 29 kPa for NC organic silt
The ratio of su interpreted from the traditional direct
and 579 ± 176 kPa for the OC silty clay.
method to su predicted from FE method is 2.2 for NC
3. The undrained shear strength predicted from finite
organic silt and 3.7 for OC silty clay. The ratio from
element method is lower than that from the
the traditional indirect method is 1.5 for NC organic
traditional methods and shows that the assump-
silt and 2.2 for OC silty clay. The results illustrate the
tion of infinite pressuremeter length in traditional
significance of pressuremeter length and soil behavior
methods results in overprediction of undrained
in prediction of undrained shear strength of clay from
shear strength by a factor of 1.5 for normally
pressuremeter test data. The results obtained herein are
consolidated clay and 2.2 for overconsolidated
specific to the experimental data in this case study. To
clay.
obtain more robust reduction factors and to generalize
the findings in this study, a comprehensive investiga- The results obtained herein are specific to this case
tion which involves a larger data base is suggested. study. More comprehensive study is recommended to
obtain a range of overprediction for different types of
soils.
7 Conclusions

In this study, data from pressuremeter tests conducted Appendix


on normally consolidated organic silt and overconsol-
idated silty clay were analyzed to estimate the See Figs. 12 and 13.

123
Author's personal copy
1522 Geotech Geol Eng (2013) 31:1511–1524

Fig. 12 Boring profiles for pressuremeter tests: PB-1 (a), PB-2 (b), PB-4 (c), and PB-5 (d)

123
Author's personal copy
Geotech Geol Eng (2013) 31:1511–1524 1523

Fig. 13 Measure and


predicted cavity pressure
curves from pressuremeter
tests: borehole PB-1 (a), PB-
2 (b), PB-4 (c), and PB-5 (d)

References Palmer AC (1972) Undrained plane-strain expansion of a


cylindrical cavity in clay: a simple interpretation of the
ABAQUS Theory manual, version 6.8 (2008) Hibbit, Karlsson pressuremeter test. Geotechnique 22(3):451–457
and Sorensson, Inc Robertson PK (1986) In-situ testing and its application
Baquelin F, Jézéquel JF, Shields DH (1978) The pressuremeter to foundation engineering. Can Geotech J 23(4):573–
and foundation engineering. Trans Tech Publications, 584
Clausthual Roscoe KH, Burland JB (1968) On the generalised stress-strain
Carter JP, Randolph MF, Wroth CP (1979) Stress and pore behavior of wet clay. In: Heyman J, Leckie FA (eds)
pressure changes in clay during and after the expansion of a Engineering plasticity. Cambridge University Press,
cylindrical cavity. Int J Numer Anal Meth Geomech Cambridge, pp 535–609
3:305–322 Roscoe KH, Schofield AN (1963) Mechanical behavior of an
Gibson RE, Anderson WF (1961) In situ measurement of soil idealized ‘‘wet clay’’. In: Proceedings of European con-
properties with the pressuremeter. Civil Eng Public Works ference on soil mechanics and foundation engineering.
Rev 56(658):6l5–6l8 Wiesbaden, pp 47–54
Houlsby GT, Carter JP (1993) The effects of pressuremeter Salgado R (2008) The engineering of foundations. McGraw-
geometry on the results of tests in clay. Géotechnique Hill, New York
43(4):567–576 Shuttle DA, Jefferies MG (1995) A practical geometry correc-
Ladayani B (1973) Expansion of a cavity in a saturated clay tion for interpreting pressuremeter tests in clay. Géotech-
medium. J Soil Mech Found Eng Div ASCE 89(SM4):127– nique 45(3):549–553
161 Wood DM (1990) Soil behavior and critical state soil mechan-
Mair RJ, Wood DM (1987) Pressuremeter testing: methods and ics. Cambridge University Press, Cambridge
interpretation. CIRIA Ground Engineering Report: In-Situ Wroth CP (1982) British experience with the self boring pres-
Testing. Butterworths, London suremeter. In: Proceedings of symposium on the pres-
Mayne PW, Kulhawy FH (1982) Ko-OCR relationships in soils. suremeter and its marine applications, pp 143–164
J Geotech Eng ASCE 108(6):851–872 Yeung SK, Carter JP (1990). Interpretation of the pressure-
Ménard L (1956) An apparatus for measuring the strength of soils meter test in clay allowing for membrane end effects
in place. Master of Science Thesis, University of Illinois and material non-homogeneity. In: Proceedings of 3rd

123
Author's personal copy
1524 Geotech Geol Eng (2013) 31:1511–1524

international symposium on pressuremeters, Oxford, Yu HS (2000) Cavity expansion methods in geomechanics.


pp 199–208 Kluwer Academic Publishers, London
Yu HS (1990) Cavity expansion theory and its application to the Yu HS, Charles MT, Khong CD (2005) Analysis of presure-
analysis of pressuremeters. PhD Thesis, University of meter geometry effects in clay using critical state models.
Oxford Int J Numer Anal Meth Geomech 29:845–859

123

View publication stats