Stochastic seismic analysis of geosynthetic-reinforced soil slopes using the probability density evolution method

Abstract

This paper presents a stochastic seismic analysis of geosynthetic-reinforced soil slopes (GRSS) using the probability density evolution method (PDEM). The combination of the uncertainties of seismic ground motions (SGM), soil properties and reinforcement parameters was considered. The stochastic SGM was generated from a random function based on the spectral representation method. The present method is proved to be more precise than the pseudo-static method and pseudo-dynamic method due to the consideration of the uncertainty of frequency spectrum of seismic excitations, and more efficient than the Monte Carlo simulation method for reducing the number of samples by 39 times. Without involving the stochastic earthquake, the result of the estimated performance would incur large errors since it greatly depends on the selected earthquakes, which may be one of the major reasons for GRSS failure. The slope displacement and the reinforcement strain occur at a relatively narrow region at the beginning, then increase and fluctuate intensively with large uncertainty, and finally become steady in the earthquake decay phase. The reinforcement length has a more substantial influence on the failure probability of the reinforcement than other parameters in the studied ranges. The reinforcement strain increases but the slope displacement decreases with the reinforcement length because of the variation of failure mechanisms.