A Physical Investigation of the Flow Structure and Mobility Behaviour of Collisional Granular Landslides

Abstract

A fundamental understanding of the mechanisms affecting the mobility of granular landslide flow is critical for the management of risk to life and infrastructure below unstable slopes. Modelling methods used in practice are frequently semi-empirical and require calibration against high-quality physical data. Large scale physical landslide modeling in the Queen’s University landslide flume was conducted with the primary objective of quantifying the effect of collisional flow on the mobility of granular landslides. A series of monodisperse landslide tests was conducted with a constant volume of material, using various nominal grain sizes. A second series of monodisperse tests was conducted using a constant nominal particle size, while varying the size of the source volume. Flow velocities and flow thicknesses were quantified using high frame rate digital image data captured from top-down and side-profile viewing angles. Shear rates and volume fractions were calculated using data from the side-profile viewing camera. Highly instrumented “Smart Rock” sensors embedded within the landslide flow experiments were used as a means to measure the accelerations within the flows. Terrestrial laser scanning provided a method of accurately capturing deposit shape results, used for mobility analyses. Highly collisional landslide flows were produced when flow thicknesses were small relative to the particle size of the constituent landslide materials. Collisional flows were more dilute, travelled faster, were sustained for shorter durations, and resulted in larger measurements of acceleration magnitude. Landslide flows with similar collisional activity were observed to exhibit similar mobility behaviour. The flows with largest observed differences in collisional activity produced a change in the Fahrböschung of nearly 15 %. Collisional flows exhibited increased mobility and spreading compared to more frictional flows.