Abstract
Multidelity modeling representing the high-fidelity physics is a new technique introduced into nondestructive testing (NDT) systems. This work proposes a novel multifidelity modeling method combining polynomial chaos expansion and Cokriging. Specifically, the least angle regression method is used to select the most correlated orthogonal polynomial of the training data and subsequentially used bases as the trend functions of Cokriging model. This proposed polynomial chaos-based Cokriging (PC-Cokriging) is demonstrated for model-assisted probability of detection of NDT systems. The approach is demonstrated on an analytical test case, as well as on an ultrasonic testing case. The results show that PC-Cokriging is capable of reducing the training sample cost to around half of the training cost for the conventional Cokriging method.
How to Cite:
Du, X., Leifsson, L. . & Nagawkar, J., (2019) “Model-assisted reliability analysis of nondestructive testing systems using polynomial chaos-based Cokriging”, Review of Progress in Quantitative Nondestructive Evaluation .
Downloads:
Download PDF