Influence of input motion and surface layer properties on seismic site response: A stochastic simulation method–based MLR model

Seismic site response analyses are simulations in which the effects of geological conditions on seismic waves are examined. The uncertainties that make these analyses crucial are defined as the source of motion, the travel path of seismic waves, and geological conditions. In this study, a series of non‐linear seismic response analyses were performed using data from site investigation studies. The results of non‐linear analyses performed under earthquakes with different characteristics based on real soil data were investigated in terms of acceleration time histories and response spectra. The relationship between site response with the surface layer properties and input motion properties used in the simulations was examined in a parametrical manner. Based on the results obtained, Monte Carlo simulation, which is a stochastic data simulation method, was performed. Additionally, multiple regression and variance analysis (ANOVA) was performed on the dataset created by both site response analysis and stochastic simulations. The MLR model displayed highly accurate results with a coefficient of determination, R2 of 0.9763, and a standard error of 0.109. The efficiency level of the independent variables used as input parameters in the simulations on the dependent variable, AF was examined. It was revealed that the coefficient of lateral earth pressure at rest (Ko), earthquake motion (i.e., PGA of input motion) and surface layer thickness (dsurface) from the soil properties had the highest effect on the amplification factor.