Routine effective stress triaxial tests

in which the units are as stated above.

Its accuracy depends on the reliability of the value of cvi. The result will not be realistic
if its value is that of a silt or greater (i.e. more than 10–8 m/s). If side drains are used the value
of k should not be quoted.

19.7.3 Shear strength CU test

Calculated axial stress
Symbols used to illustrate the calculations for shear strength are summarised in Table 19.6.
Each line of data is calculated as follows:
ΔL
Axial strain ε% = ×100%
L

Axial force P = C r × R N

The load ring calibration Cr might not be constant, but can vary with ring deflection
(see Volume 2, Section 8.4.4).
The area of cross-section (A) of the specimen, allowing for barrelling deformation
(Volume 2, Section 13.3.7) is equal to
100A c
mm 2
100 − ε%
The axial stress in the specimen (the measured deviator stress) induced by the force P is equal
to P/A N/mm2, i.e. (P/A)1000 kPa. Hence from Equation (13.7) (Volume 2, Section 13.3.7)

( σ1 − σ3 )m = AP ×1000 × 100100
− ε%
kPa
c

P
= × 10 (100 − ε% ) kPa (19.16)
Ac

If deformation occurs by failure on a definite surface (‘single plane slip’), the area correction
has to be modified as described in Section 18.4.1.

Membrane and drain corrections

The calculated deviator stress must now be corrected to allow for the effects of the rubber
membrane, and side drains if fitted. These corrections are discussed in Section 18.4.2. and
18.4.3), and are deducted from (σ1–σ3)m, i.e.

( σ1 − σ3 ) = ( σ1 − σ3 )m − σmb − σdr
The membrane correction for barrelling, σmb, for the appropriate strain is obtained from the
graph in Figure 18.16 (Section 18.4.2), which applies to a 38 mm diameter specimen in a

245
 Manual of Soil Laboratory Testing

0.2 mm thick membrane. For any other specimen diameter (D mm) and membrane thickness
(t mm) this value is multiplied by
38 t
×
d 0.2

to obtain the membrane correction to use.

The drain correction, σdr for several specimen diameters is given in Section 19.7.4.

Corrected stresses
The fully corrected deviator stress is used for the calculation of the principal stresses, and other
parameters, from each set of recorded data as follows. The symbols are those given in Table 19.6.
Minor effective principle stress σ′3 = σ3–u
Major effective principal stress σ′1 = (σ1–σ3) + σ′3
Principal stress ratio = σ′1 /σ′3
–
Pore pressure coefficient A = (u–u0) / (σ1–σ3)
–
In a fully saturated soil, A = A
The stress path parameters, s’ and t’, in terms of effective stress, may be calculated from
the equations derived in Section 16.1.4(2))
σ '1 + σ '3 (16.3)
s'=
2

σ '1 − σ '3 (16.2)

t =
2

The above calculations are plotted in Figure 19.22(d).

Graphical plots
The calculated data are used for plotting the following graphs:
Deviator stress against strain.
Pore pressure against strain, with the value at the start of compression emphasised
to show the datum to which pore pressure changes are related.
Principal effective stress ratio against strain, starting from unity at zero strain.
Pore pressure coefficient? against strain.
Stress path plot of t against s’.

Failure criteria
From the curve of deviator stress against strain, the maximum value (at the ‘peak’ point or
‘failure’ condition) is located and denoted by (σ1–σ3)f. The corresponding values of strain
(ef) and pore pressure (uf) are also read off, as shown in Figure 19.22(c). This can also be
achieved by reading the maximum value from the relevant columns of computer-generated
data, although such values should be reviewed against the graphical output. The curve might
indicate that the maximum value lies between two sets of readings, in which case a sensible

246
 Routine effective stress triaxial tests

interpolation should be made. Values of σ′3, σ′1, σ′1 / σ′3 and A (denoted by Af) corresponding
to the ‘failure’ condition are calculated as described above.
Similar sets of values corresponding to the point of maximum principal stress ra-
tio, or the ‘critical state’ condition (constant deviator stress and pore pressure) are read
off from the curves and calculated, if either of these alternative criteria for failure is
appropriate.

Stress paths
Stress path plots of t=0.5(σ1−σ3) against s’=0.5(σ′1+σ′3), or q against p’ (Section 16.3) are
also drawn. These plots convey more information on soil behaviour during the test than
Mohr circles.

Mohr circles
The sets of values of σ′1 and σ′3 obtained as described above enable Mohr circles of effective
stress to be drawn to represent the relevant criterion for failure.
Mohr circles of total stress are not relevant to the derivation of a strength envelope. They
are not required by BS 1377, but they are in ASTM D 4767.

Set of tests
Results from a set of three specimens taken from one sample can be presented on one test
report sheet by grouping the graphs together as shown in Figures 19.22a–d.
The stress path plots of t against s’ for the set of specimens are presented in Figure
19.22d. The point on each plot corresponding to the failure criterion of maximum deviator
stress is marked and the line of best fit is drawn through the set of points. The angle of slope
of the line to the horizontal axis is denoted by θ, and its intercept with the t (vertical) axis is
denoted by t0.
The shear strength parameters (c’, φ′) are calculated from the following relations derived
in Section 16.2.3)
sin(φ′) = tan(θ) (16.9)

t0 (16.10)
c'=
cos ( φ' )

Alternatively, three Mohr circles from the set, representing the appropriate failure criterion,
can be plotted as shown in Figure 19.22a), and the envelope of ‘best fit’ is drawn to them.
The slope of the envelope gives the effective angle of shear resistance,φ′ and the intercept
with the vertical axis gives the effective cohesion, c’.

19.7.4 Shear strength CD test

Calculations and corrections
The symbols used to illustrate shear strength calculations are the same as for the CU test
(Table 19.6, Section 19.7.3).

247
 Manual of Soil Laboratory Testing

Figure 19.23 Graphical data from a set of consolidated-drained CD triaxial compression tests.
(a) Summary data and Mohr's circle plots.

248
Figure 19.23 (b) Saturation and consolidation stages; and

249
Figure 19.23 (c) Shearing stage and specimen photographs

250
 Routine effective stress triaxial tests

Axial strain (ε%) and axial force (P N) are calculated as for the CU test. In calculating
the deviator stress ((σ1–σ3)m kPa), the change in volume of the specimen due to drainage is
an additional factor to be taken into account, as explained in Section 18.4.1.
The volumetric strain is equal to ευ(ΔV/Vc) × 100% and the deviator stress is calculated
from the equation
P 100 − ε
( )
σ1 − σ3 =
m
×
A c 100 − εν s
× 1000 kPa (19.17)

which is derived from Equation (18.7).

If there is no significant change in pore pressure there is little to be gained from calculat-
ing the principal stress ratio because, when plotted, it will give a curve identical in shape to
the stress–strain curve.

Membrane and drain corrections

The calculated deviator stress is corrected to make allowance for the effects of the mem-
brane, as described in Section 19.7.3 for the CU test.
If vertical side drains are fitted, the correction to apply for barrelling is constant for
strains exceeding 2%. The correction depends on the specimen diameter (see Section 18.4.3)
and is summarised as follows
Specimen diameter (mm) 38 50 70 100 150
Drain correction σdr (kPa) 10 7 5 3.5 2.5
Up to 2% strain it can be assumed that the correction increases linearly to the values
stated above, as explained in Section 18.4.3.
No correction is applied if spiral drains are used.
For single plane slip the corrections for area change, membrane and side drains de-
scribed in Sections 18.4.1 to 18.4.3 may be applied.
Details of the corrections applied for area change, membrane and side drains, and their
magnitudes, should always be clearly shown in the test report. This is particularly important
for single plane slip corrections which can be of appreciable magnitude.

Graphical plots
The following graphs are plotted: deviator stress against axial strain; and volumetric strain
against axial strain. Pore pressure may also be plotted against strain if it changes signifi-
cantly.

Failure
Read off the peak deviator stress (σ1–σ3)f and the corresponding axial strain (εf) from
the deviator stress–strain curve, as for the CU test. Also read off the volumetric strain
and the actual pore pressure (uf) coresponding to εf (see Figure 19.23c). Corresponding
values for the condition of constant deviator stress and constant volume are obtained if
relevant.
Calculate the stresses corresponding to failure in the same way as for the CU test
(Section 19.7.3).

251
 Manual of Soil Laboratory Testing

Mohr circles
The values of σ′3f , σ′1f are used for plotting stress paths or Mohr circles of effective stress
at failure under drained conditions. If several specimens are tested from one sample, they
are plotted together as a set and the failure envelope is drawn. The drained shear strength
parameters cd,φd are determined in the same way as for the CU test (see Figure 19.22). For
most practical purposes the drained parameters can be considered as being identical to the
undrained parameters c’, φ′, and the latter symbols are normally used.

19.7.5 Reporting results

The following data are reported, together with the relevant graphical plots as indicated. The
data are similar for both CU and CD tests except where separately identified. Items marked
* are additional to those listed in Clauses 7.6 and 8.6 of BS 1377.

General
Sample identification, reference number and location
Type of sample
Soil description
Type of specimens and initial dimensions
Location and orientation of test specimens within the original sample
Method of preparation of each specimen
Moisture content determined from trimmings*
Date test started

For each specimen:

Initial conditions
Moisture content
Bulk density, dry density
Voids ratio and degree of saturation*
Thickness of membrane*
Whether side drains were fitted, and if so what type

Saturation stage
Saturation procedure
Cell pressure increments if by incremental pressures
Differential pressure if by incremental pressures
Pore pressure and cell pressure after saturation
Value of B achieved
Degree of saturation reached
Graphical plot of B value against pore pressure (or cell pressure)*

Consolidation stage
Cell pressure
Back pressure

252
 Routine effective stress triaxial tests

Initial pore pressure

Final pore pressure
Percentage pore pressure dissipated
Coefficient of consolidation cvi*
Coefficient of volume compressibility mvi*
Calculated time to failure
Graphical plot of volume change
against square-root time

Compression stage (CU test)

Cell pressure
Initial pore pressure
Initial effective confining pressure
Rate of strain applied
Failure criterion used
Data at failure: axial strain, deviator stress, pore pressure, effective major and minor
principal stresses, effective principal stress ratio
Sketch of specimen at failure, illustrating mode of failure
Details and magnitude of membrane, drain and other corrections applied
Actual time to reach failure*
Final density and moisture content
Graphical plots: deviator stress against strain, effective principal stress ratio against
strain, pore pressure against strain
Mohr circle of effective stress representing failure
Stress path for effective stresses with s’ plotted against t
Statement that the test was carried out in accordance with Clauses 4, 5, 6 and 7 of BS
1377: Part 8: 1990.

Compression stage (CD) test

Cell pressure
Back pressure
Initial pore pressure
Effective confining pressure
Rate of strain applied
Failure criterion used
Data at failure: axial strain, deviator stress, volumetric strain, pore pressure, effective
major and minor principal stresses, effective principal stress ratio
Sketch of specimen at failure, illustrating mode of failure
Details and magnitude of membrane, drain and other corrections applied
Actual time to reach failure*
Final density and moisture content
Graphical plots:deviator stress against strain, volume change against strain, pore pres-
sure changes (if significant) against strain
Mohr circle of effective stress
Statement that the test was carried out in accordance with Clauses 4, 5, 6 and 8 of BS
1377: Part 8: 1990.
253
 Manual of Soil Laboratory Testing

For a set of CU specimens

Numerical data as listed above, grouped together as a set
Graphical plots as stated above, including stress path plots, grouped on common axes
Envelope to points representing failure on stress path plots, indicating derivation of
shear strength parameters c’,φ′
Set of Mohr circles corresponding to failure, with failure envelope indicating slope (φ′ )
and intercept (c’).

For a set of CD specimens

Numerical data as listed above, grouped together as a set
Graphical plots as stated above, including stress path plots, grouped on common axes
Envelope to points representing failure on stress path plots, indicating derivation of
shear strength parameters c’, φ′
Set of Mohr circles corresponding to failure, with failure envelope indicating slope (φ′)
and intercept (c’).

References
ASTM Designation D 4767–11, Standard Test Method for Consolidated-Undrained Triaxial
Compression Test on Cohesive Soils. American Society for Testing and Materials, Phila-
delphia, USA.
Arulanandan, K., Shen, C.K. and Young, R.B. (1971) Undrained creep behaviour of a coastal
organic silty clay. Géotechnique, Vol. 21(4), p. 359.
Bishop, A.W. and Henkel, D.J. (1962) The Measurement of Soil Properties in the Triaxial
Test, 2nd edn). Edward Arnold, London (out of print).
Bishop, A.W. and Little, A.L. (1967) The influence of the size and orientation of the sample
on the apparent strength of the London Clay at Maldon, Essex. In: Proceedings of the
Geotechical Conference, Oslo, Vol. 1, pp. 89–96.
BS 1377: Parts 1 to 8: 1990, British Standard Methods of Test for Soils for Civil Engineering
Purposes. British Standards Institution, London.
Case, J. and Chilver, A.H. (1959) Strength of Materials: an Introduction to the Analysis of
Stress and Strain. Edward Arnold, London.
Hight, D.W. (1980) Personal communication to original author.
Hight, D.W. (1982) Simple piezometer probe for the routine measurement of pore pressure
in triaxial tests on saturated soils. Technical Note. Géotechnique, Vol. 32(4), p. 396.
Lumb, P. (1968) Choice of strain-rate for drained tests on saturated soils. Correspondence.
Géotechnique, Vol. 18(4), p. 511.
Olson, R.E. (1964) Discussion on the influence of stress history on stress paths in undrained
triaxial tests on clay. In: Laboratory Shear Testing of Soils. ASTM Special Technical Pub-
lication No. 361, American Society for Testing and Materials, Philadelphia, pp. 292–293.
Sandroni, S.S. (1977) The strength of London Clay in total and effective shear stress terms.
PhD Thesis, Imperial College of Science and Technology, London.
Sangrey, D.A., Pollard, W.S. and Eagan, J.A. (1978) Errors associated with rate of undrained
cyclic testing of clay soils. In: Dynamic Geotechnical Testing. ASTM Special Technical
Publication No. 654. American Society for Testing and Materials, Philadelphia.
Wesley, L.D. (1975) Influence of stress path and anisotropy on behaviour of a soft alluvial
clay. PhD Thesis, Imperial College of Science and Technology, London.
254
 Chapter 20

Further triaxial shear strength tests

20.1 Scope
This chapter describes several triaxial test procedures for the measurement of the shear
strength of soils, additional to the basic tests described in Chapter 19. Some of these proce-
dures are not often used, but may be appropriate in particular circumstances. The procedures
are generally indicated in outline only, and make reference to Chapters 18 and 19 for many
of the details. The apparatus required is the same as that already described for the tests given
in Chapter 19.
There are numerous other ways of determining soil shear strength properties for particu-
lar applications, most of which need more advanced procedures requiring additional equip-
ment. Such procedures are beyond the scope of this book.

20.2 Multistage triaxial test

20.2.1 Introduction
A multistage quick-undrained triaxial test, in terms of total stress, on a single specimen of
soil such as ‘boulder clay’, was described in Volume 2, Section 13.6.5. Multistage effective
stress tests have been reported by Lumb (1964), Kenney and Watson (1961), Watson and
Kirwan (1962) and Ruddock (1966). These tests were shown to give effective shear strength
parameters (c′, φ′) that are practically indistinguishable from those obtained by the normal
procedure using a set of three specimens for many types of soil. However, other factors such
as compressibility, dilatancy, pore pressure changes and voids ratio changes do not compare
so well with conventional tests. Multistage triaxial tests on partially saturated, fairly perme-
able soils were described by Ho and Fredlund (1982). The use of stress paths in multistage
testing was described by Janbu (1985).
Some form of stress path plotting is necessary for multistage tests because it is essential
to keep the stress conditions and soil behaviour under observation as the test proceeds, espe-
cially immediately before the peak deviator stress is reached. Stress paths and their use are
described in Chapter 16.
The main advantage of multistage testing is in the saving of time and material when only
one large specimen is available for testing and when small specimens are impracticable.
This applies particularly to soils containing relatively large-scale features such as gravel in
boulder clays, or discontinuities such as fissures, which cannot be adequately represented in
a small specimen. Multistage tests should, in general, be restricted to soils such as these that
are also of low sensitivity and have a stable structure, and require a relatively small strain

255
 Manual of Soil Laboratory Testing

and volume change to induce failure. Tests are performed using consolidated-undrained pro-
cedures, usually on specimens that are first saturated, or have a high degree of saturation
initially. The consolidated-drained procedure is not recommended for multistage testing.
Three stages of loading are normally applied, the final stage being extended beyond the point
of peak deviator stress.
An essential factor in multistage testing is the selection of an appropriate criterion for
‘failure’ at each intermediate stage. This should take into account the type of soil and its
behaviour, and practical in situ considerations, especially those relating to limitations of
strain. Excessive deformation should be avoided during the first two stages of a three-stage
test, and these stages should be discontinued just before reaching peak deviator stress. Some
general guidance is given in Section 20.2.4, and each type of soil should be treated on its
merits rather than following a rigid procedure.
Multistage triaxial testing in general should be regarded, not as standard practice, but as
an expedient to fall back on for certain types of soil when there is no practicable alternative
due to the limited number of specimens available.
Multistage consolidated-drained test procedures have not been described in the follow-
ing sections as the test is not recommended unless the person commissioning the test is fully
conversant with its limitations. Volume changes occur within each stage of testing, so each
stage of the test is performed on a material in a different state. Unless allowances can be
made for the change in state, the results obtained can be of little value (Hight, personal com-
munication, 2011). The criteria for determining failure described in Section 20.2.4 should
also not be regarded as trustworthy when applied to drained tests. Sharma et al. (2011)
describe a procedure for determining failure of consolidated-drained multistage tests on
artificially cemented samples of silty sand using the criterion dεv/dεh = 0 for termination of
the earlier stages of the test. They report an average error of 6% and 5% in c′ and φ′, respec-
tively, when compared with the results of single-stage tests, but further work is needed to
confirm the validity of this procedure.

20.2.2 Test procedures

General
The following outline procedures relate to CU triaxial tests on undisturbed specimens,
except where the test procedure requires special treatment. Remoulded or compacted speci-
mens may also be used if enough time is allowed between preparation and testing to enable
pore pressures within the sample to equalise.
With a large specimen of clay soil of low permeability, drainage from both ends is
desirable to halve the drainage path (without having to use side drains), thereby reducing
the theoretical consolidation time by a factor of four. To allow double drainage during
consolidation preceding an undrained test, the arrangement referred to in the following is
necessary.

Procedural steps
1. Set up the specimen as for a normal CU test, as shown in Figure 19.1, Section
19.2.3. Valve a is connected to the pore pressure transducer. Valves b and d are con-
nected to a common back pressure system, or if a back pressure is not necessary, to
open burettes with the water levels at the specimen mid-height. With this arrange-
ment pore pressures can be measured during compression (valve a open, valves b

256
 Further triaxial shear strength tests

and d closed) after allowing drainage from both ends during consolidation (valve a
closed, b and d open) without having to change over any connecting lines.
2. Saturate the specimen by one of the procedures given in Section 19.6.1 to obtain an
acceptable B value. Alternatively, apply a cell pressure to confirm that the sample
initially has an acceptable degree of saturation.
3. Consolidate to the first (lowest) of the effective confining pressures selected for the
test (Section 19.6.2).
4. Calculate the time to failure (Section 19.6.2, step 9) and estimate the appropriate
rate of strain, depending on the type of test and drainage conditions.
5. Start the compression stage in the usual way. As the test proceeds, calculate data
for plotting the following against strain (%), applying all the necessary corrections
(Section 19.7.3).
• Deviator stress (kPa)
• Pore pressure (kPa)
• Principal stress ratio (σ′1 / σ′3)
Also calculate data for making the stress path plot of t against s′, where
t = (σ′1 – σ′3 )/2
s′ = (σ′1 + σ′3 )/2
(see Section 16.1.4(2)). These graphs can be shown on the computer using soft-
ware for plotting in real time, but if necessary, when drawing them manually, run
the test a little slower than normal to enable these graphs to be kept up to date as
the test proceeds.
6. When failure is approached according to the appropriate criterion, stop the com-
pression machine. Impending failure may be indicated by the development of a
surface of failure within the specimen, or by flattening of the stress–strain curve, or
by one of the other criteria discussed in Section 20.2.4.
7. Immediately reverse the motor to reduce the axial force rapidly until it just reaches
zero, and record the corresponding axial strain reading. Sketch the configuration of
the specimen. Allow enough time for the pore pressure to reach equilibrium before
proceeding further. Removal of the axial force prevents ‘creep’ of the soil structure
between loading stages and allows a small elastic recovery of strain.
8. Raise the cell pressure to the level required for the next stage and wait for the pore
pressure to stabilise. Maintain the back pressure at the same level as previously un-
less it is essential to lower it to achieve the next effective confining pressure.
9. Consolidate the specimen as before. Calculate the new consolidated dimensions.
10. Determine the time to failure for the next stage of loading from the new consolida-
tion data and calculate the new rate of strain. Failure is usually approached at a
much smaller strain than in the first stage; for example, if failure in the first stage
was almost reached at between 5 and 10% strain, only 1–2% strain may be needed
for a subsequent stage. Furthermore, as the effective pressure is increased, the val-
ue of cvi usually decreases. The machine speed for stages 2 and 3 is therefore likely
to be considerably slower than for stage 1.
11. Wind up the machine platen to remake contact between the cell piston and the top
cap, then reset the strain dial gauge and the load ring dial gauge to zero or to a new
datum reading.

257
 Manual of Soil Laboratory Testing

12. Resume application of the axial load at the appropriate rate of strain, with
simultaneous calculation and plotting as in step (5). Continue until ‘failure’ is again
approached as in step (6).
13. Calculations are made as follows:
Strain: use the new specimen length to calculate strain referred to the new zero
datum, and add this to the strain at the end of unloading to obtain cumulative
strain. This gives strain due to application and removal of deviator stress only.
Deviator stress: use the new specimen area and volume and strain measured
from the new datum to calculate the (uncorrected) deviator stress. To apply the
membrane correction, use the cumulative strain.
Pore water pressure: plot the recorded values. The pore pressure at the start
of each stage is usually equal to, or a little above, the back pressure used for
consolidation, which should be unchanged wherever possible.
14. Repeat steps (7) to (13) to apply a total of three stages of loading (if practicable). A fourth
stage may sometimes be possible. Continue the final stage beyond peak deviator stress.
15. Reduce the applied load and pressures to zero and remove the specimen for sketch-
ing, final weighing and measurements, as for a normal triaxial test. Complete the
usual calculations.

20.2.3 Graphical plots

The graphical data to be presented are similar to those for a normal set of CU tests, except
that the curves for the second and subsequent compression stages start from the strain cor-
responding to the unloaded condition of the previous stage. A typical set of report sheets for
a multistage CU test are shown in Figures 20.1 to 20.4. From the stress–strain curves for
the first two stages the ‘failure’ conditions are extrapolated. These values, together with the
actual failure data for the final stage, can be used for constructing the Mohr circles to which
a failure envelope is drawn for determining the values of c′, φ′. However, the stress path plot
described in Section 20.2.4(6) is generally more useful.

20.2.4 Failure criteria

In a typical triaxial compression test the maximum total applied strain should not exceed
about 25%. The available length of the cell piston is one limiting factor, but the severe
distortion of the specimen itself at a strain much in excess of 20% can result in dubious
values of calculated axial stress. In a three-stage test the axial strain in each stage should
not therefore exceed about 8%, but it may be permissible to extend the first stage to a larg-
er strain because stiffening of the soil due to further consolidation often results in smaller
strains to failure in subsequent stages.
The failure criterion for each specimen should be selected on the basis of soil type and in
situ conditions. Some of the criteria which might be applied are as follows.
1. Visible failure, e.g. the development of a slip surface in the specimen.
2. Flattening out, or reaching the peak, of the deviator stress–strain curve (with all
corrections applied).
3. Reaching a predetermined strain, such as 10, 15, 20% for three successive stages.
For a plastic type of soil, strains of 16, 18 and 20% may be more appropriate
(Anderson, 1974).

258
 Further triaxial shear strength tests

0.0
Shear Stress (kPa)

RLW

Figure 20.1 Graphical data from multistage CU triaxial compression test – specimen data and Mohr circles

259
 Manual of Soil Laboratory Testing

485

548

%
%
Volume Change (cm³)

RLW

Figure 20.2 Graphical data from multistage CU triaxial compression test – saturation and consolida-
tion stages

260
 Further triaxial shear strength tests

Deviator Stress (Measured) (kPa)

Principal Stress Ratio
Pore Water Pressure (kPa)

RLW

Figure 20.3 Graphical data from multistage CU triaxial compression test – shearing stage

261
 Manual of Soil Laboratory Testing

t' (kPa)

RLW

Figure 20.4 Graphical data from multistage CU triaxial compression test – stress paths and specimen
photographs after test

262
 Further triaxial shear strength tests

4. The condition of maximum principal stress ratio (σ′1 / σ′3) provides a useful
indication. In overconsolidated soils it occurs before maximum deviator stress.
This ratio should be plotted against strain as the test proceeds.
5. The change in pore water pressure can also be used as a guide. The maximum
excess pore pressure (in a non-dilating soil) occurs at about peak principal
stress ratio.
6. In addition to the plots referred to in (2), (4) and (5), the stress path method of
plotting (described in Chapter 21) is particularly useful for undrained effective
stress tests. For an overconsolidated soil such as boulder clay, it was demonstrated
by Watson and Kirwan (1962), using a plot of σ′1 against σ′3, that the slope of the
failure envelope can be estimated from a test at a single cell pressure after the
maximum effective stress ratio has been reached, as illustrated in Figure 20.5,
where test data are plotted on the MIT stress field (Section 16.1.4(2)). In stage
1 the maximum stress ratio is reached at point S and the succession of readings
from S to U lies along the failure envelope. However, it is difficult to derive the
slope of the envelope from these closely grouped points. Normally, the first stage
is terminated at a point such as T, before the specimen has been subjected to a
deformation large enough to reach the ‘peak’ deviator stress. A second stage is
needed to enable the failure envelope to be drawn, and this is also terminated at
a similar point T. A third stage is applied, and this can be extended far enough to
achieve failure and define the peak strength envelope more positively, as shown in
Figure 20.5.

It is usually desirable to make more than one type of plot to provide enough evidence for
deciding when to terminate one stage and proceed to the next.

Figure 20.5 Stress path plot of multistage CU triaxial compression test. sin φ' = tan θ = 0.374; φ' = 22°

263
 Manual of Soil Laboratory Testing

20.3 Undrained compression tests

20.3.1 General
Types of test
This section briefly describes relatively simple triaxial compression test procedures which
vary from the BS procedures, but are considered to be good practice. The procedures
described here are based on methods described by D. W. Hight (personal communications,
1996, 2010), which call for the use of a pore pressure probe at the mid-height of a specimen
in addition to the usual pore pressure transducer at the base. Use of the mid-height pore
pressure probe offers the advantage that pore pressure readings are not subject to end
restraint and the derived effective stress conditions are more representative of the specimen
as a whole. Indication of pore pressure changes is more rapid and so tests can be carried out
at a faster rate than when relying on base pore pressure measurements. The use of the mid-
height pressure probe has increased significantly in recent years and preparation of a test
specimen in this way is described in the following.
The procedures described in this section are:
1. Unconsolidated, undrained test, with measurement of pore water pressure, referred
to as the UUP test;
2. Consolidated-undrained test for 100 mm diameter specimens of undisturbed clay,
referred to as the CUP test.
The use of mid-height local strain gauges and bender elements in conjunction with
any of the preceding tests is also described briefly. Such instrumentation is more usually
incorporated into tests on anisotropically consolidated specimens and as such is considered
to be beyond the scope of this book.
The testing described in this section is normally carried out on 100 mm diameter specimens
of rotary core or pushed thin walled tube samples. A particular form of sample disturbance
of rotary core samples is the penetration of the cores by the drilling fluid. This produces a
wetted zone on the perimeter of the core, which should be removed before any testing for
undisturbed parameters. This must be undertaken as soon as the core is removed from the
ground, as part of the sealing and preservation of any samples for testing.

Applications
The UUP test outlined in Section 20.3.2 is applicable to undisturbed specimens in which
no change in moisture content from the in situ value can be permitted and is used to
determine the effective stress parameters of a sample of clay during undrained shear. It can
also be used for the determination of effective shear strength parameters for compacted
fill materials.
The CUP test outlined in Section 20.3.4 is intended primarily for near-saturated or
saturated clay to measure the effective stress parameters of a sample of clay following
isotropic consolidation or swelling. This procedure avoids the usual back pressure saturation
process in which the soil has access to water at very low effective stresses.
The success of these test procedures depends on the skill and experience of the technicians
when setting up the test specimens and apparatus and also requires close liaison between the
laboratory and the supervising engineer. It is recommended that where the laboratory or
technician is new to the procedures set out here, practice tests are undertaken on ‘dummy’
samples to ensure their effectiveness.

264