PROBABILISTIC FATIGUE CRACK GROWTH ANALYSIS FOR LIFE PREDICTION OF AUTOMOTIVE COMPONENTS

摘要

This research focuses on investigating the probabilistic fatigue crack growth models on W319 aluminum alloy which has vast applications in automotive parts products. The aim of this study is to determine the crack growth rate probabilistically and quantification of uncertainty of probabilistic models on estimation of damages (crack length) versus life (number of cycles) under fatigue loading in automotive parts. The models used in this paper include Walker and Forman correlations. The deterministic forms of these models are verified with AFGROW code and validated experimentally with fatigue data of W319 aluminum. After verifying the accuracy of deterministic models, the models are treated probabilistically by considering the models' parameters stochastic. Monte Carlo simulation is devised to investigate the models under stochastic conditions by drawing samples from these random variables. Finally the propagation of uncertainty is quantified by calculating standard deviations of crack lengths through propagation of the uncertainties via cycles. The results are useful for selecting a proper probabilistic fatigue crack growth model in specific applications and can be used in future studies in automotive industry to obtain more accurate and reliable conclusions.