Identification of the mostprobablepoint in original spaceapplications to structural reliability
Abstract
This paper presents a new iteration procedure to identify the mostprobablepoint (MPP) that has the highest probability density function in the failure domain without using probability transformation. The proposed iteration procedure constructs the mostprobablepointlocus (MPPL) of a linearized limitstate function starting from the mode and uses this MPPL to search the next linearization point. The iteration is converged to the MPP, where the contour of equal probability density function is tangent to the limitstate function. The conventional Uspace MPP is not the true MPP if the probability distribution is asymmetric. Because probability transformation is not used in the iteration, the proposed method provides an efficient way to identify the Xspace MPP.
 Publication:

AIAA/ASME/ASCE/AHS/ASC 34th Structures, Structural Dynamics, and Materials Conference
 Pub Date:
 April 1993
 Bibcode:
 1993ssdm.conf.2791L
 Keywords:

 Failure Analysis;
 Iterative Solution;
 Probability Density Functions;
 Structural Reliability;
 Interpolation;
 Linearization;
 Sensitivity;
 Structural Mechanics