融合SVM与T-S模糊观测器的磁流变阻尼器故障诊断方法OA

Magnetorheological(MR)damper is a damping device applied in semi-active suspension system of automobiles,and its failure or health state determines whether the suspension system can work properly or not.In this paper,a new magnetorheological(MR)damper fault diagnosis algorithm was proposed.Compared with the previous methods,this algorithm only relies on two accelerometers to diagnose whether the damper fails or not without the necessity of additional displacement sensors.Firstly,a Takagi-Sugeno(T-S)fuzzy unknown input observer was designed based on the Lyapunov stability theory.Then the damping force obtained from the observer and the fuzzy model was collected for preprocessing.Finally,a support vector machine(SVM)classifier was used to automatically identify whether the damper fails or not,without relying on experience to set the threshold.The experimental results show that both the SVM-based fault diagnosis accuracy and F0.5 metric are higher than the fixed-threshold fault diagnosis accuracy and F0.5 metric,and both the accuracy and F0.5 metric are higher than 0.94,which indicates that the proposed fault diagnosis method has a robust performance in MR damper fault diagnosis.