Comparison of some selected time-domain fatigue failure criteria dedicated for multi input random non-proportional loading conditions in industrial components

This paper presents time-domain procedure to evaluate High-Cycle Fatigue (HCF) life of industrial components with complicated geometries and subjected to multi-input random loading conditions. To this end, the authors discuss some well-known equivalent stress criteria which can consider the orientation of critical plane changes with the time in non-proportional loading conditions to estimate the time history of equivalent stress. This task is successfully accomplished by knowing time histories of local stress tensor components in a critical area of the part that is prone to failure. Next, two different categories were used to predict fatigue life of component: 1- Linear Damage Accumulation (LDA) rule by employing Rain-flow cycle counting and 2- probabilistic approach by utilizing level crossing counting method. To apply these methodologies and awareness of their challenges, a case study was performed on the automotive steering knuckle. Finally, the predicted lives were compared to experimental results, and the accuracies of different fatigue life assessment techniques were reported. The obtained results show that combination of Rain-flow cycle counting with the Liu-Zenner criterion and especially, with the Energy-based instant fatigue damage tracing proposed by Shariyat lead to the highest accuracies. However, using the probabilistic approach, we will get the response in a much shorter time compared to another category. Therefore, the choice which method to use also depends on the parameters of the research situation, and the interaction of the two factors of prediction accuracy and prediction time in the industry.