Teaching-Learning-Based Optimization for Ground Motion Selection

Abstract

This research introduces a teaching-learning-based optimization (TLBO) algorithm to select seismic ground motions (GMs). The target spectrum is the uniform hazard response spectrum (UHRS). The selected GMs undergo scaling using scalar values within a user-defined range, ensuring no alteration to the phase or shape of the earthquake ground motion response spectra. The proposed approach efficiently explores thousands of earthquake records and recommends a subset that closely aligns with the target design spectrum. Utilizing TLBO with 11 or 7 pairs of records and appropriate scaling factors expedites this process, providing recordings that closely match the target spectrum with minimal manipulation. Furthermore, the authors present error measures comparing the target spectrum to the predicted spectrum derived from the mean of the selected records (7 and 11 records) for a downtown Memphis location. The results affirm the effectiveness of the proposed model in obtaining suitable GM record sets for time history analyses.