Internal Erosion of a 0–5 mm Crushed Sand in a Rigid Wall-Permeameter: Experimental Methods and Results

ABSTRACT The article deals with internal erosion in 0–5 mm crushed sand, a material used in urban facilities. Several methods, designed for natural materials, predicted a risk of internal erosion for the crushed sand. The real risk was evaluated in clear-wall, watertight and airtight, rigid-wall permeameters with lateral piezometers. Each test started with a fully saturated specimen, verified by a mass-and-volume method. In their initial condition, the specimens were homogeneous, as directly confirmed by three methods. Each erosion test included successive seepage steps at a constant mean gradient. Migration of fine particles was observed through the clear wall. Dyed water colored three or four preferential seepage paths inside the specimen, not along the wall, confirming that the preparation method was correct. The predictive methods correctly forecast the size of mobile particles but overestimated the gradient that triggered the internal erosion process. The tests were the first ones to include nonreactive tracer tests, which gave the effective porosity value at different gradient steps. This value decreased with internal erosion, a novel result, which quantified the fact that more seepage concentrated in preferential seepage paths. Nonreactive tracer tests can thus be used to quantify the formation of preferential seepage paths.