Application of low-discrepancy sampling method in structural reliability analysis

This study introduces and investigates various low-discrepancy sequences and then develops a new procedure in which the low-discrepancy sequences are combined with the importance sampling technique to estimate the failure probability. This proposed low-discrepancy sampling method is based on the concept that the deterministic low-discrepancy sequences of points can significantly improve the accuracy of the classical Monte Carlo (MC) method over purely random sampling. Different benchmark examples verify that the proposed method is more accurate with the same number of samples and has a faster rate of convergence in order to achieve a given accuracy when compared with the MC method. Therefore, the low-discrepancy sampling method shows great potential for improving the accuracy and efficiency of the MC-based simulation method for structural reliability analysis.