Summary |
Hazard mitigation against natural terrain landslide requires the understanding of its mobility and impact behaviour. In this research, the flow mobility and impact behaviour of granular materials were investigated. Flume model tests and numerical simulations using the sled model, the dynamic analysis (Hungr, 1995) and Particle Flow Code in Three Dimensions (PFC3D) were conducted. Two flumes were used in the physical model tests, namely the runout flume and the barrier flume. The runout flume included an inclined channel of 1.4m long, 45° slope angle and 0.4m wide. The inclined channel was connected to a horizontal channel of 1.2m long. The travel angle was measured in each runout tests. The barrier flume consisted of an inclined channel of 3.8m long, 40° slope angle and 0.4m wide. The confinement angle was adjustable for both flumes. An aluminium barrier with a finite stiffness was adopted on the barrier flume. A high speed camera was used to record the impact behaviour. Leighton Buzzard sand fraction C & E, as well as completely decomposed granite (CDG) were used in physical model tests. Measured data were compared and analysed with computed results using the sled model, the dynamic analysis (Hungr, 1995) and PFC3D. Parametric studies using the dynamic analysis (Hungr, 1995) with different flow resistance models and PFC3D were conducted. Flume data revealed the existence of critical water contents. The critical water contents were around 20% and 25% for fraction C sand and CDG respectively. The travel angle decreased with mean particle size in a granular mixture. A critical confinement angle was observed between 0° and 8° using fraction C and fraction E sands respectively. Reverse segregation was observed using CDG soil. Back analysis of measured data shows that the sled model fails to model the effects of mass on travel angle. By including side resistance and dynamic drag in calculation using the dynamic analysis, the critical confinement angle could be captured. The deposition process has a longer duration by adopting the Bingham model than the friction model and the Voellmy model. The duration of impact was longer for fraction E sand than that of fraction C sand. No peak force was measured and progressive deposition was observed for both sands. The duration of impact was shorter for wet fraction C sand. The measured impact force reduced with the confinement angle. Parametric study using PFC3D shows that the maximum impact force for dry granular material is contributed mostly by static load. The computed impact force on a vertical barrier is larger than a barrier which is perpendicular to the inclined flume. By incorporating a deposition zone of 0.5m in front of the vertical barrier, the computed static load is reduced by around 80%. |