16-Civ-B3 Geotechnical Design · May 2017

Question 2 of 9: Choice of triaxial or direct shear test for three field problems

Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)

Question 2: Choice of triaxial or direct shear test for three field problems (7 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

The test is chosen by matching the drainage condition and the strain level in the laboratory to the drainage condition and the strain level in the field. Two questions settle almost every case. First, will the pore pressures generated by the field loading have dissipated by the time the critical condition is reached? If yes the analysis is drained and the parameters must be effective-stress parameters; if no, the analysis is undrained and a total-stress parameter is appropriate. Second, is there a pre-existing shear surface or a fabric that prevents the peak strength from ever being mobilised simultaneously along the whole failure surface? If so, the operational strength is the residual, not the peak.

Part (i) — long-term stability of a clay foundation beneath an embankment. The critical condition here is reached years after construction, once the excess pore pressures generated by the embankment load have dissipated and the clay has consolidated under it. The analysis is therefore an effective-stress analysis and the parameters required are c' and phi'. I would specify a consolidated undrained triaxial test with pore-pressure measurement (the CU-bar test, ASTM D4767) on undisturbed samples, consolidated isotropically or, better, anisotropically to the estimated in-situ stresses, then sheared slowly enough for pore pressure to equalise across the specimen. This is preferred over the fully drained CD test for a practical reason: in a clay of low permeability a CD test must be sheared over days or weeks to avoid a pore-pressure gradient, whereas the CU-bar test measures the pore pressure and so obtains the same effective-stress envelope in a fraction of the time, while additionally giving the pore-pressure parameter A that is needed to estimate the excess pore pressures during construction. Three specimens at consolidation pressures bracketing the in-situ vertical effective stress define the envelope. The embankment should also be checked for the end-of-construction case, for which the same undisturbed samples yield an unconsolidated undrained strength, so in practice both sets are run on the same tube.

Part (ii) — short-term stability of a footing on saturated clay. A footing is loaded quickly relative to the consolidation time of a saturated clay, so at the end of construction essentially none of the excess pore pressure has dissipated and the clay behaves undrained. The appropriate test is the unconsolidated undrained (UU) triaxial test (ASTM D2850) on high-quality undisturbed samples at their in-situ water content, giving the undrained shear strength cu with an apparent friction angle phi_u = 0. Because the specimen is not allowed to consolidate, the measured strength is independent of cell pressure and the failure envelope is horizontal, which is precisely the condition assumed by the phi = 0 bearing-capacity solution qu = 5.14 cu + q. The unconfined compression test is an acceptable quick screening variant but underestimates cu when the sample has fissures or has lost suction, so it should not be the sole basis. Field vane tests, corrected by Bjerrum's plasticity factor, are the standard in-situ cross-check and are strongly recommended alongside the laboratory work. The direct shear box is not suitable for this case: it cannot be run fast enough to guarantee undrained conditions in a specimen only 20 mm thick with drainage stones top and bottom, and it cannot control or measure pore pressure at all.

Part (iii) — long-term stability of a slope in fissured, expansive clay. Two features rule out the peak strength. The fissures are pre-existing discontinuities along which the clay has already been sheared, and an expansive clay in a slope swells, softens and loses its cemented structure as it absorbs water over years. In addition, a slope fails progressively: strain is not uniform along the potential slip surface, so where the peak has been passed the strength has already fallen to residual while elsewhere it is still climbing, and the average operational strength is well below peak. The correct laboratory test is therefore a drained ring-shear test (Bromhead apparatus, ASTM D6467), which can impose the very large displacements — hundreds of millimetres — needed to reach a true residual condition with the platy clay particles fully aligned. Where a ring shear apparatus is unavailable, a multiple-reversal drained direct shear test on a pre-cut specimen is the accepted substitute, sheared slowly enough that no pore pressure is generated. The parameters obtained are phi'_r with c'_r taken as zero. The fully-softened strength, measured in a drained direct shear or CD triaxial test on a normally consolidated remoulded specimen, is the appropriate upper bound for a first-time slide in the same material, and it is good practice to report both. A conventional CU-bar triaxial test on an intact specimen would give a peak c' and phi' that are real for the intact clay but unsafe for the slope, because the specimen is too small to contain a representative fissure spacing.

The three cases together illustrate the underlying rule: (i) drained loading, peak effective-stress strength; (ii) undrained loading, total-stress strength; (iii) drained loading, residual effective-stress strength. The drainage condition selects the test type; the presence of discontinuities and the strain history select the point on the stress–strain curve that may be relied upon.