A fatigue damage prediction model with multi-parameter correlation

Abstract

A fatigue cumulative damage prediction model with multi-parameter correlation, which introduces eight parameters in the action coefficients, including adjacent stresses and their corresponding lifespans (counting to four parameters), previous level of fatigue cumulative damage, S-N logarithmic slope of the material, ultimate stress, and life of the fatigue limit, is proposed in this study. The new model is not a simple nonlinear fitting. It establishes the correlation between multiple parameters and the fatigue damage model. To comprehensively evaluate the proposed model, we conducted numerous experiments to compare the proposed model with a group of baseline models. By analyzing the verification data that compared the statistical value and distribution of the difference between the fatigue prediction damage and test damage of each model, we verified that the fatigue damage prediction effect of the proposed model is the best overall. Additionally, the proposed model preliminarily demonstrated that the evolution of fatigue damage accumulation during multistage stress loading is a complex process with multiple parameters that are highly correlated.