Abstract
This study outlines a numerical integration for evaluating the failure probability of structural systems with non-normal distributions based on the hypersphere division method. The structural failure probability is formulated as a sum of the failure probability contributions of several concentric hyperspheres. In order to deal with the non-normal distributions efficiently, the center of the hypersphere is shifted to the mean value points of the failure domains and the range of integration of the radial distance is limited within the effective distance range determined according to the size of standard deviations of the failure domains. These statistical parameters are estimated by a preliminary simulation. Numerical examples of assessment of the structural failure probability are presented to show the validity of the proposed method.
