Predictive model for seismic displacements of flexible sliding block subjected to near-fault pulse-like ground motions

Abstract

A post investigation of landslides induced by strong earthquake events in recent years shows that near-fault pulse-like ground motions have a considerable impact on slopes. Accurate evaluation of the stability of slopes under near-fault earthquakes has become a crucial issue in seismic-prone areas. Permanent displacement is widely recognized as an effective index for evaluating the stability of slope. Based on the coupled analysis, an empirical earthquake-induced permanent displacement prediction model was developed. A displacement database of 318 pulse-like ground motions from 50 earthquake events was established. The proposed predictive model is a function of the maximum seismic coefficient-time history (kmax), maximum velocity coefficient of velocity-time history of the k-time history (kv-max), yield acceleration (ky), and period ratio (ratio of the natural period of the slope to the mean period of input motion, Ts/Tm). The dynamic responses (kmax and kv-max) are represented by two different functions. At small period ratios (Ts/Tm≤1.8), it is a function of peak ground acceleration (PGA), peak ground velocity (PGV) and ky. At large period ratios (Ts/Tm > 1.8), it is a function of PGV, Ts and ky. It successfully captured the velocity pulse and long period characteristics of pulse-like ground motions. Compared to the existing model, it is found that the model is more accurate when considering pulse-like ground motions. This prediction model can be used not only for the preliminary evaluation of slope stability but also for probabilistic seismic demand analysis of flexible slopes considering pulse-like ground motion in near-fault regions.