A Fault Diagnosis Model Based on Hierarchical Belief Rule Base for Complex Mechatronics System

Currently, the fault diagnosis model of complex mechatronics systems is mostly based on monitoring data. The effect of expert knowledge in the fault diagnosis for complex mechatronics systems is completely ignored. Belief rule base (BRB) is an excellent hybrid fault diagnosis model that combines monitoring data and expert knowledge. However, the input of multiple features can cause rule explosion in BRB. Therefore, this study proposed the model of fault diagnosis based on hierarchical belief rule base (HBRB) for complex mechatronics systems. Specifically, the HBRB model proposed consists of multiple progressive BRBs compared with the traditional BRB. Which allows multiple features to be input step by step, thus solving rule explosion caused when multiple features are input at the same time, then the P-CMA-ES is selected as the optimization algorithm for training the initial parameters. Finally, the proposed model’ effectiveness was verified by the fatigue endurance experiment on the automotive mirror system, the accuracy is 97.5%, Mse is 0.025, which is better than other models.