Influence of ground-motion selection on the nonlinear response of curved surface slider isolation systems

The paper investigates how different ground-motion selection procedures influence the non-linear seismic response of base-isolated systems, with particular focus on curved surface slider used for safety-relevant structures. Four commonly used selection strategies were evaluated under identical site-specific hazard conditions for four return periods: database filtering, minimum mean-square-error (MSE) spectrum fitting, spectral matching, and conditional spectrum (CS)-based selection. A simplified non-linear time-history analysis framework was applied to quantify two key engineering demand parameters: peak isolator displacement and reduced transmitted response spectra. The results show that the different procedures can produce similar mean and median responses, while leading to different record-to-record variability and individual responses. Database filtering resulted in the lowest central displacement estimates but also produced significant outliers, suggesting potentially unconservative design. The MSE-based approach provided intermediate response levels with relatively stable dispersion, making it computationally efficient but less representative of true hazard variability. Spectral matching produced the lowest variability and the most stable response trends, but substantially suppressed natural ground-motion variability (unless it is subsequently considered for design). In contrast, the CS-based approach preserved hazard-consistent variability but led to the highest variability and highest displacement demand, especially at lower hazard levels. Overall, the study shows that good agreement with a target spectrum alone does not ensure a similar non-linear isolator response. The choice of record-selection procedure directly affects both the estimated engineering demand parameters and their associated uncertainty. Ground-motion selection procedure is found to be a governing component of curved surface sliders isolator assessment, with important implications for both design-oriented evaluation and hazard-consistent risk assessment.