New paper published in Mechanical Systems and Signal Processing
In collaboration with T. Zhou and Y. Peng from Tongji University, B. Sudret and S. Marelli published a new paper on a novel adaptive approach for the probability density evolution method for reliability analysis.
B. Sudret and S. Marelli published a new paper in collaboration with T. Zhou and Prof. Y. Peng from Tongji University on a novel adaptive meshing algorithm to improve efficiency of the recently published Kriging-enhanced probability density evolution method (PDEM) algorithm, AK-PDEM.
PDEM is a powerful reliability analysis algorithm that can be employed to effectively solve time-dependent reliability analyses. The recent introduction of Kriging and active learning in the PDEM framework (AK-PDEM), the main performance bottleneck of the method remains the choice of the size of the underlying PDEM mesh. In this paper we propose a strategy to adaptively refine the PDEM mesh, in addition to the Kriging metamodel used in AK-PDEM, greatly improving the accuracy of the method, at negligible additional computational costs.
A benchmark study of the newly introduced AK-PDEMi algorithm against AK-PDEM and other methods on three applications of increasing complexity demonstrates that the proposed approach is both accurate and efficient.
For more information, please follow external pagethis linkcall_made for the publication and this link for the associated report on our internal archive.