Regression analysis for seismic slope instability based on a double phase viscoplastic sliding model of the rigid block

Semi-empirical models based on Newmark's sliding block permit the estimation of expected co-seismic displacements in relation to one or more parameters which characterize the ground motion that theoretically caused them. Taking this into consideration, a regression analysis, based on a double-phase viscoplastic (DPV) model, was developed using 96 Italian ground motion accelerograms for a total of 1,448 combinations obtained for different parametric conditions of the indefinite slope model. Repeated stability analysis, performed by means of the DPV model, allows for the assessment of the seismic instability of a slope in relation to different reached behaviour levels, as well as seismically induced permanent displacements. At these behaviour levels, co-seismic increases and possible subsequent decreases of viscoplastic shear strengths are associated. This implies that the post-seismic persistent mobility (collapse) of the slope can be obtained from the computation. On the other hand, coherently with the increasing of shear resistances during fast sliding displacements in clay soils, the seismic-forced displacements result substantially lower than corresponding values obtained by means of the rigorous Newmark's sliding block. In addition, in relation to some seismic ground motion parameters, regression and functional border and separation curves were obtained with the aim of providing an expeditious seismic slope stability evaluation in reference to the co-seismic and post-seismic behaviour of clayey slopes. Regarding this, the real behaviour of two historical landslide events is discussed in the light of the results of the regression analysis outlined in this work.