T-S Fuzzy $H_{i}/H_\infty$ Optimization Based Fault Detection Method for Traction Motor in High-Speed Train

This paper focuses on the problem of speed sensor fault detection (FD) for traction motor in high-speed train subject to multi-source interference. An observer-based fault detection filter (FDF) is employed to generate a residual signal for speed sensor fault described by a Takagi-Sugeno (T-S) fuzzy model. The robustness to multi-source interference is measured by the
$H_{\infty}$
-norm of the interferences to the residual, while the sensitivity to fault is measured by the
$H_{\infty}$
-norm or the
$H\_$
criteria of the fault to the residual. As a result, the design of the FDF is formulated as an
$H_{i} /H_{\infty}$
optimization problem, aiming to achieve an optimal balance between the sensitivity to fault and the robustness to interference. A unified solution is proposed to obtain the observer gain matrix and post-filter. To illustrate the effectiveness of the developed method in this paper, two types of faults are considered in the simulation, including sudden-change and slow-change faults. The simulated results are shown that the developed method in this paper can achieve effective detection of speed sensor fault for traction motor in high-speed train.