A Fuzzy $H_{i}/H_{\infty }$ Optimization Approach to Fault Detection of High-Speed Train Traction Motor Systems

Abstract

In this article, an $\mathit {H_{i}/H_{\infty }}$ optimization approach to fault detection (FD) is proposed for high-speed train traction motor under complex environment and working conditions. Considering the inherent system nonlinearity, the dynamics of the traction motor are firstly described by a Takagi–Sugeno (T-S) fuzzy model subject to $\mathit {l}_{2}$ norm-bounded disturbances and additive faults. Then, a T–S fuzzy observer-based fault detection filter (FDF) is proposed as a residual generator, and, in order to enhance simultaneously the robustness of residual to disturbances and the sensitivity to fault, the design of the FDF is formulated as the maximization problem of finite horizon $\mathit {H_{-}/{H}_{\infty }}$ and $\mathit {H_{\infty }/{H}_{\infty }}$ indices. Moreover, an $\mathit {H_{i}/H_{\infty }}$ optimization approach is developed to find a solution of the T–S fuzzy FDF, which can achieve an optimal tradeoff between the sensitivity to fault and the robustness to disturbances. It shows that the optimal solution is not unique, and the feasible solutions including static and dynamic postfilter are obtained by recursive computing of Riccati equations. Finally, a case study of traction motor in CRH5 EMUs is presented to exhibit the efficacy of the developed FD approach.