Collaborative modeling-based improved moving Kriging approach for low-cycle fatigue life reliability estimation of mechanical structures

Abstract

To effectively estimate the reliability level of low-cycle fatigue (LCF) life of mechanical structures, a novel method of collaborative modeling-based improved moving Kriging (IMKCM) approach is proposed by fusing improved moving Kriging method and collaborative modeling strategy. In this method, the improved moving Kriging model is developed by the Kriging model, moving least squares method and artificial hummingbird algorithm, in which the Kriging model is selected as the basis function, the artificial hummingbird algorithm is used to define the optimal radius of compact support region and determine the effective modeling samples, and the moving least squares method is adopted to resolve the unknown coefficients; the collaborative modeling strategy is developed from the decomposition and coordination method, in which the mathematical models of the decomposed sub-objectives are synchronously built by the cell array theory, and the mathematical model between the objective and the sub-objectives is established by the improved moving Kriging model. Furthermore, a numerical example (i.e., nested nonlinear function) and an engineering example (i.e., turbine blisk LCF life reliability estimation) are employed to verify the effectiveness of the IMKCM approach. The results show that the proposed IMKCM method holds excellent abilities in modeling features and simulation performance.