Reliability sensitivity analysis of automotive disc brake with multiple failure modes correlation

The failure of the automotive braking system may lead to brake failure and even severe traffic accidents. Traditional single failure mode reliability analysis can no longer meet current design requirements, while the Gumbel Copula model can measure the correlation between different failure modes and evaluate the risk of failure combinations. Therefore, the model is adopted in this paper. However, in engineering practice, the reliability prediction methods for thermal fatigue and complex modal multiple failure modes have encountered problems such as computational inefficiency and solution difficulties. To solve these problems, a new method, Multi-Failure Mode Reliability Sensitivity Analysis Method (MF-RSAM), is proposed. Based on friction vibration theory and control theory, a performance function is established with the damping ratio of the most unstable mode of the system as the response. Based on the Manson-Coffin model, a performance function is established with the maximum strain value of the brake disc surface as the response. Subsequently, the optimal Gumbel-Copula function is selected and introduced into the reliability analysis model. The Monte Carlo support vector regression method is adopted for the vibration-fatigue gradual reliability prediction of automotive braking systems under multiple failure modes. Finally, the reliability sensitivity under thermal fatigue and complex modal correlated failure modes is analysed. The results show a significant difference between the results of parametric reliability prediction considering correlated failure and single failure modes, with a maximum reduction of about 7.5% in reliability under correlated failure modes as the parameters are changed and a change in the parameter sensitivity order.