On the Pore-pressure Generation and Movement of Rainfall-induced Landslides in Laboratory Flume Tests

Abstract

Using a small flume, a series of tests was conducted to trigger rainfall-induced landslides. Based on monitoring of sliding distance and pore pressures, the process of pore-pressure generation in relation to sliding distance was examined. By performing tests on sands of different grain sizes (silica sand no. 7 (D 50= 0.14 mm) and no. 8 (D 50= 0.057 mm)) at different initial dry densities or different thicknesses, the effects of these factors on pore-pressure generation and failure behavior of a landslide mass were analyzed. Results from tests of different initial densities showed that for each sample there was an optimal density index, at which the pore pressure build-up after failure reached its maximum value. This optimal density index varied with the thickness of sample and the grain size of samples. Moreover, observed failure phenomena showed that the failure mode also depended greatly on the grain size and sample thickness. In fact, flowslides were initiated in the tests on finer silica sand (no. 8), whereas retrogressive sliding occurred in the tests on silica sand no. 7. Results of tests on mixtures of silica sand no. 8 with different contents of loess by weight showed that the existence of fine-particle soil (loess) could significantly change the flow behavior of a landslide mass during motion: the flow behavior of soils with 20 and 30 percent loess was different from these two silica sands and the mixture with 10 percent loess, showing greater velocity without deceleration. This suggests the existence of a mechanism that maintains high pore pressures during motion for these soils. In addition, by rotating saturated samples in a double-cylinder apparatus, a mechanism was examined in which pore pressure in saturated soils during motion was maintained. The results showed that the pore pressure of the saturated mixture increased with velocity because of the “floating” of sand grains that accompanied the movement for each test. In addition, the sample with finer grain sizes or greater fine-particle (loess) contents floated more easily, and high pore pressure could be maintained during motion. The floating ratios of grains reached a high value (>0.85) at a very slow velocity for samples with 20 and 30 percent loess. Based on these test results, it is concluded that grain size and fine-particle contents can have a significant impact on the mobility of rainfall-induced landslides.