Civil and Building Engineering.

Triaxial Compression Tests on Normally Consolidated Specimens
Drained Tests. Drained (S-type) triaxial compression tests on saturated specimens are performed in the same manner in which they are performed for
Figure 3.41 Angle of friction versus initial porosity for fine sand (from Bjerrum et al., 1961).

sand, with the allowance of greater durations of time that allow adequate dissipation of the excess porewater pressures generated during shear. For tests performed using a constant rate of deformation, the time to failure may vary from a minimum of about 6 hours to over 6 months, depending on the coefficient of permeability of the clay.

The type of Mohr-Coulomb diagram obtained from such tests is shown in Figure 3.43. Normally consolidated clays are consolidated from dilute suspensions that have essentially zero shear strength. Hence, the application of a very small confining pressure results in the development of a very small strength, and the failure envelope passes through the origin. The slope of the failure envelope is denoted by —the effective stress angle of internal fric- S tion determined using S-type tests.

The curves for compressive stress–axial strain and for volumetric strainaxial strain for drained tests on specimens of saturated, normally consolidated clay are similar to the curves for sand shown in Figure 3.39, except that the specimens of clay may undergo volume decreases of more than 10% during shear. The decrease in volume during shear means that the relatively loose soil structure is breaking down under the action of shearing deformation.
Figure 3.42 Angle of friction versus initial void ratio for fine sand (from Bjerrum et al., 1961).

Unconsolidated-Undrained Tests, 0 Case. If a saturated specimen of soil is subjected to a change in hydrostatic total stress under undrained conditions, such as by increasing the pressure in a triaxial cell without allowing the specimen to drain, the porewater pressure changes by an amount equal to the change in hydrostatic pressure. Direct observations of this behavior are possible. A soil specimen is assumed to be taken from the field and placed in a triaxial cell. An initial cell pressure in the specimen is measured. The
difference between the cell pressure and the porewater pressure is the effective stress. If the cell pressure is increased, the porewater pressure will increase by an equal amount, and no change in effective stress will occur if both the specimen and the measuring system are saturated. Thus, according to the
principal of effective stress, no change in shearing strength is possible. The confining pressure in the triaxial cell, therefore, has no effect on the shearing strength of the soil. A total stress failure diagram for a saturated soil is shown in Figure 3.44. The dashed circle is the Mohr’s circle for effective stress at failure. The circles shown with solid lines are the total stress circles of stress.
The shear strength is independent of the cell pressure, all these circles of stress have the same diameter, and the tangent to the failure envelope is horizontal; thus, 0. Actual experimental data have been published for both sands and clays showing that the angle of internal friction is zero experimentally as well as theoretically (Bishop and Eldin, 1950).
If an undisturbed sample of a saturated soil can be obtained, the strength of the soil in situ can be obtained by performing a Q-type triaxial test. If the test specimen remains saturated under zero gage pressure, the shear strength of the soil may be measured using an unconfined compression test because the diameter of the circle of stress passing through the origin is the same as the diameter of all other circles of stress. Thus, the 0 concept is the justification for using unconfined compression tests to measure the shear strength of a sample of saturated clay. The unconfined compression test is justified only if the samples remain saturated.
In practice, samples of soil often become slightly unsaturated under conditions of zero total stress and are subject to some sampling disturbance. Both factors cause a loss of strength that cannot be estimated without more detailed investigations.

Consolidated-Undrained (R-Type) Tests. For consolidated-undrained (R- type) triaxial compression tests on samples of saturated clay, the testing procedure differs, depending on whether or not porewater pressures are to be measured. If porewater pressures are not measured, the specimen is consoli- dated under the desired initial state of stress, the drainage connections are closed, and the specimen is sheared to failure over a time period that varies from a few minutes to perhaps half an hour. The time to failure varies, depending on whether or not complete stress-strain data are recorded. If porewater pressures are to be measured, it is absolutely necessary that the porewater pressure connections be saturated and free of air bubbles. The presence of a small amount of free air in the drainage lines or fittings will usually have a negligible effect on shearing strength but will cause the measured porewater pressure to be considerably smaller than the actual porewater pressure in the specimen. Two approaches are used in practice to saturate the drainage system.

The preferred method is to consolidate the specimen first, then simultaneously increase the confining pressure and porewater back pressure by the same amount. No change in effective confining stress and no change in strength ( 0 condition) occur because the porewater pressure and cell pressure are increased by the same amount. The second method used is to apply back pressure to the specimen first under a low effective confining stress, usually around 1 psi, then consolidate the specimen to the final effective confining pressure under the back pressure.

The two methods work equally well for saturated clays of low plasticity. The first method usually requires back pressures in the range of 20 to 40 psi and works best on specimens that are initially partly saturated and on clays of high plasticity that are expansive. The second method may require back pres- sures of up to 100 psi to be effective in saturating the test specimen. In no case does the second method work better than the first method for common geotechnical applications. The increased pressure in the porewater pressure
system dissolves any air bubbles that might be present and ensures saturation.
When the system is properly saturated, the application of an increment of
In another modification, either the lateral stress is increased with a constant axial stress until failure occurs or the lateral stress is maintained constant and the axial stress is decreased until failure occurs. Such tests are called extension tests. All stresses must be compressive in an extension test. An extension test is not a tensile test, it is a compressive test, with the lateral stress exceeding the axial stress.

For normally consolidated clay, the tendency of the soil structure to break down during shear leads to a transfer of part of the effective stress to the porewater; thus, positive porewater pressures are developed. A typical stressstrain curve for an R-type test is shown in Figure 3.46. The strain at the peak stress difference (often defined as the point of failure) may vary from less than 1% to 20%.
Figure 3.46 Stress-strain curve for the R-test with positive pore water pressure.

A comparison of the stress-strain properties of clays under drained and  undrained conditions is of interest. Under drained conditions, the soil structure is broken down during shear and the volume decreases, that is, the density increases during shear. The continuously decreasing volume (increasing density) causes the soil to gain strength as it is deformed. Thus, the soil may be considered a strain-hardening material under these conditions of drainage. Eventually, the applied stress exceeds the strength and the specimen fails. The
large strains at failure result from the continuous increase of strain.

In an undrained test on a normally consolidated specimen, the total confining pressure is constant. Thus, the continuously increasing porewater pressure (Figure 3.46) causes a reduction of the effective confining pressure
The strength of the soil is controlled by the effective stress, not the total ( ). 3 stress; thus, continuous reduction in effective confining pressure causes the soil to lose strength as shearing deformations are applied. Hence, the soil in an undrained test is strain weakening, and the strains at failure (at peak stress difference) are considerably smaller than those obtained from drained tests on initially identical specimens (this argument is restricted to specimens that decrease in volume during drained shear and have positive porewater pres- sures during undrained shear, which is generally true for normally or lightly overconsolidated clays).
disturbance, nonuniform stress and strain conditions in the triaxial cell, and other experimental problems also influence the results of these shear tests.
Investigations into the shearing characteristics of undisturbed cohesive soils are in progress, but the problems are so complex that a simple solution of the problem is unlikely.
shear strength with depth for a uniform deposit of clay in Drammen, Norway, is shown in Figure 3.50, where the undrained shear strength is shown to increase linearly with depth for depths greater than about 7 m. For shallower depths, the soil has been desiccated and has thus been strengthened. The remolded strengths plotted in Figure 3.50 were obtained by rotating the vane several times and then measuring the strength again.
In the early investigations, the vane strength was always greater than the strength determined using unconfined compression tests because all samples taken into the laboratory were disturbed by the sampling operation. As specialized sampling techniques were developed, the laboratory strengths became higher (due to less disturbance); recent studies have shown that the laboratory strengths exceed the vane strengths. As a result of variations of soil properties with depth, it is difficult to make accurate comparisons of vane and unconfined compression strengths, but one recent study suggested that the unconfined compression tests yielded strengths about 30% higher than the vane strengths. The cause of this discrepancy is not fully understood, but part of it may result from the vane device measuring shear strength on vertical planes
(the area of the end of the vane is small), the high rate of shearing in the zone of failure, and the degree of plasticity of the soil. The normal effective stress on a vertical plane is K0 times the normal effective stress on the horizontal plane, where K0 is termed the coefficient of earth pressure at rest and typically has a value of about 0.5 for a normally consolidated clay. If the strength is a function of the normal effective stress, then the vane strength should be lower than the strength on diagonal planes as measured using the unconfined compression apparatus.
In situ shear strengths may also be determined by pushing a cone down into the soil and measuring the resistance. Most of the research had been done for the Dutch cone penetrometer. Beginning in the early 1980s, pore pressure measuring equipment was added to Dutch cone equipment and was used to identify changes in stratigraphy in the soil profile. The cone penetrometer test
is favorable in that each cone sounding is a model pile load test. However, the use of shear strengths determined from cone soundings for other applications, such as slope stability, should be done with caution.
The c/p ratio has proved to be a useful parameter in foundation engineering in areas where deep deposits of soft clays are encountered. Correlations between the c/p ratio of normally consolidated soils and the plasticity index have been developed (Figure 3.51). Additional research has established the relationship between c/p ratio and overconsolidation ratio for a variety of soils (Figure 3.52).
Use of the c/p ratio for design purposes was formalized by Ladd and Foott (1974) in the SHANSEP (Stress History And Normalized Soil Engineering Properties) method. This method is a synthesis of the observations and methods of several researchers into a relatively straightforward approach that can be used if one has the benefit of a complete field investigation and highquality laboratory tests. The difficulty in using the SHANSEP method is in establishing the stress history, consolidation characteristics, and overconsolidation ratio of the soil. The SHANSEP method is expensive, but its use can be justified for large projects where the soil profile is relatively uniform over the site.

Sensitivity. The remolded strength is shown to be less than the undisturbed strength in Figure 3.51. A quantitative measure of the strength loss on remolding is the sensitivity, which is the ratio of the undisturbed strength to the remolded strength at the same void ratio. The symbol for this parameter
is St. The classification of the sensitivity of clays, as proposed by Skempton and Northey (1952) and modified by Rosenqvist (1953), is shown in Table 3.5.
Bjerrum (1954) has reported sensitivities as high as 1000 in some of the Norwegian clays. Because the undisturbed strengths of the Norwegian clays are low, such high sensitivities mean that the remolded clays turn into viscous fluids. If such clays are disturbed in situ, landslides may result in which the clay flows away like a viscous fluid. Many landslides of this type have taken place in Norway, in Sweden, and along the St. Lawrence River Valley in Canada. Bjerrum (1954) has shown that the sensitivity of Norwegian clays can be correlated with the liquidity index (Figure 3.53). This correlation is useful because it warns the engineer to use special precautions with deposits where the natural water content is equal to the liquid limit or higher. Apparently, special sampling techniques are required when such soils are encountered. Otherwise, the shear strengths measured in the laboratory may be only a small fraction of the strengths in the field.