Measurement and Elimination of Membrane Compliance Effects in Undrained Triaxial Testing

摘要

Undrained loading tests are widely used to investigate the susceptibility of soils to liquefaction. Changes in effective confining stress cause variations in the degrees of penetration of the confining membrane into peripheral sample voids. The phenomenon, known as membrane compliance, invalidates the fundamental assumption of constant volume during undrained testing. Membrane compliance can have a serious detrimental effect on the accuracy and validity of such tests, particularly for medium to coarse sands and gravels. An improved testing procedure for the elimination of compliance effects during testing, recently developed for conventional scale (2.8-inch diameter) samples of sand, was verified by comparison with results of large-scale (12-inch diameter) tests of similar materials, for which compliance effects were negligible. The technique was then further developed and implemented for large-scale testing of gravelly soils. The compliance mitigation procedure used in these studies involved predetermining the magnitude of volumetric compliance as a function of effective confining stress and soil parameters, and then using computer-controlled injection or removal of water to continuously eliminate membrane compliance effects. Monotonic and cyclic load tests were performed on uniformly-graded gravel with and without implementation of the computer-controlled compliance mitigation system.