Improving the evaluation efficiency of failure probability for large and complex systems using an innovative active set strategy

Abstract

Reliability-based design optimization (RBDO) is a well-known design procedure for solving uncertainties and seeking better design, but evaluation of its failure probabilistic constraints involves an unbearable computational burden. In this study, the innovative active set strategy (ASS) for the reliability index approach is proposed to improve the efficiency and robustness of evaluating probabilistic constraints for RBDO by establishing an ASS and defining an update factor. The innovative ASS is defined using inequality to effectively perform reliability analysis and obtain the feasible most probable failure point in the inner loop. Meanwhile, a defined update factor is applied to eliminate the unnecessary probability constraints and renew the active constraints in the outer loop. The inner loop and outer loop are synchronized to enhance computational efficiency during the whole optimization process. Five commonly used examples are presented, demonstrating that the ASS is a superior efficient and robust method for a large-scale system.