Skip to main content
Open Access Publications from the University of California

Cyclic threshold strains in clays versus sands and the change of secant shear modulus and pore water pressure at small cyclic strains

  • Author(s): Mortezaie, Ahmad Reza -
  • Advisor(s): Vucetic, Mladen
  • et al.

When fully saturated soils are subjected to cyclic loading in undrained conditions involving moderate and large cyclic shear strain amplitudes, c , their stiffness and strength decrease and the pore water pressure changes permanently with the number of cycles, N. Such cyclic degradation of stiffness and pore water pressure change are among the most fundamental and important phenomena in soil dynamics. Fully saturated sands are most susceptible to cyclic degradation and significant pore pressure buildup. During cyclic loading, they can completely lose their stiffness and strength while large excess pore pressure develops and they can eventually liquefy. On the other end of the spectrum are fully saturated clays of high plasticity that under significant cyclic loading can lose only a fraction of their original stiffness and strength while the pore water pressure change is relatively small. Moreover, in overconsolidated clays of high plasticity pore pressure can decrease instead of increase with N while cyclic degradation is even smaller.

As opposed to that, when fully saturated soils are subjected to very small cyclic shear strains, c, soil`s structure practically does not change and, consequently, neither cyclic degradation nor permanent cyclic pore pressure change occur. The amplitude c below which there is no cyclic degradation and above which the degradation occurs is known as the threshold shear strain for cyclic degradation , td. The amplitude c below which there is no permanent pore water pressure change with N and above which the permanent pore pressure is recorded after every cycle is known as the threshold shear strain for cyclic pore water pressure change, tp.

For sands, tp has been extensively investigated but td has been hardly investigated at all. For clays, both tp and td were investigated to quite a limited extent. For example, tp in overconsolidated clays has practically not been investigated.

The thesis describes laboratory investigation focused on td in a clean sand and tp in two laboratory-made normally consolidated (NC) and overconsolidated (OC) clays, kaolinite clay having PI=28 and kaolinite-betonite clay having PI=55, and on the comparison and connection between different thresholds. The Norwegian Geotechnical Institute (NGI) type of direct simple shear device (DSS) was employed for the constant-volume equivalent-undrained cyclic testing. Two types of tests were conducted, single-stage cyclic strain-controlled test with constant c throughout the test, and the multi-stage cyclic strain-controlled test in which c was constant in each stage but larger in every subsequent stage. The magnitude of c covered the range from 0.003% to 2%.

In the context of investigating td in sand and tp in clays, the following tasks were also performed: (1) the NGI-DSS device was adapted for small-strain cyclic testing and a procedure for eliminating false loads and deformations from test records was developed, (2) the effect of the vertical consolidation stress, vc', and frequency of cyclic straining, f, on cyclic degradation and pore water pressure change in clays was tested, and (3) cyclic stress-strain behavior and the change in secant shear modulus, GSN, with N in sands at very small cyclic strains was investigated.

The following conclusions are derived and results obtained: (1) small-strain cyclic testing at c as small as 0.003% can be conducted in the standard NGI-DSS device and the results can be used in soil dynamics analyses if the device is properly modified and the false loads are eliminated from the test records, (2) cyclic degradation in clays is affected moderately to significantly by vc' and frequency, f , (3) in sands modulus GSN increases with N at very small c below tp (sand is stiffening) while the cyclic pore water pressure does not develop; (4) in sands at small to moderate c above tp modulus GSN first increases and then decreases with N while the cyclic pore water pressure continuously increases, (5) because of such behavior in which soil stiffness starts degrading after certain number of cycles, td in sand cannot be defined like td in clays, (6) in both clays tested td is not visibly affected by overconsolidation ratio, OCR, (7) in kaolinite clay tp is not affected by OCR, (8) in kaolinite-bentonite clay with PI=55 and overconsolidated to OCR=4 and 7.8 the pore water pressures between c=0.003% and 0.3% did not change in a consistent manner and tp could not be evaluated, (9) the thresholds tp tested show increase with PI just like in the previous studies, but their values are at or below the lower bound of published tp-PI trends, and (10) the tested td thresholds in clays do not follow the trend of increase with PI like in the previous studies and they are smaller than those published earlier for similar soils.

Main Content
Current View