Prediction of fatigue crack growth life based on a new fusion model

Abstract

A novel method based on filtering algorithm is proposed to predict fatigue crack growth life of metal structures under repeated loading. Three parts, including improvement of the two-parameter exponential formula, the improved two-parameter exponential formula parameter estimation and residual life prediction, are involved in the proposed method. The improved two-parameter exponential formula is discretized to establish a parameter estimation physical model of state space. Then, the uncertainty parameters α, β and crack length a in the state space physical model are determined according to the crack propagation data obtained by monitoring/detection and the filtering algorithm, to reduce the influence of uncertain factors such as environment and operation conditions on the prediction of fatigue crack growth life of in-service metal structures. Finally, the discretized improved two- parameter exponential formula are employed to predict the growth life of fatigue crack. An example is presented to test the validity of the proposed method. The result demonstrates that the proposed method is available to well deal with the uncertainty of fatigue parameters in the process of fatigue crack growth with the improved two-parameter exponential formula, and the particle filter algorithm can accurately predict the growth residual life of fatigue short crack.