A statistical inference of hydraulic shear and clogging in internally unstable soils

Abstract

Water-retaining earth structures often experience cyclic hydraulic flow patterns, which are believed to alter the soil fabric. Gap-graded soils are identified to be internally unstable. These soils exhibit a change in their hydraulic conductivity when the critical hydraulic gradient is exceeded. The changes beyond the critical hydraulic gradient, however, are yet to be understood. This research study examines the unsteady nature of clogging and unclogging process in saturated gap-graded soils subjected to hydraulic pulses. A set of 60 min flow (suffusion) tests was conducted using a positive displacement pump to subject a cylindrical specimen to controlled hydraulic flow pulses. The global hydraulic gradient across the specimen was continuously monitored. The grain-size distribution of the soil was re-evaluated after the test. The hydraulic shear stresses induced by the flow pulses were estimated numerically and normalised over the maximum value. Similar trends of the hydraulic conductivity and the applied flow rates do not hold when the duration of the flow pulse is increased. A simple procedure based on statistical inference is proposed to visualise the temporal variation of the clogging and unclogging process subjected to the frequency and the phase shift parameters of the flow pulse.