An Effective Kriging-based Approach for System Reliability Analysis with Multiple Failure Modes

Kriging-based surrogate model is widely adopted in the area of component reliability analysis, thanks to its computational efficiency. However, surrogate models employed in component reliability analysis can not be employed directly in system reliability analysis. In this paper, an effective kriging-based system reliability analysis approach is proposed based on AK-SYS. In the proposed method, the best next samples are selected from the safe region or failure region, and samples in the areas that have little contribution to the composite performance function are avoided. The efficiency and accuracy of the proposed method are illustrated via a numerical example.