A Projection-based Binary Classification Method for Fault Detection of T-S fuzzy Systems with Multiplicative Faults

The main objective of this paper is to address the detection issues of multiplicative faults in T-S fuzzy systems. To this end, the orthogonal projection, as an effective tool to deal with the problem of fault detection and isolation, is introduced first. Then, the image subspace and kernel space of T-S fuzzy systems is introduced, and then formulated in the time-varying framework. Based on it, the time-varying fuzzy projection-based binary classification scheme is proposed. The core of the method is to construct the orthogonal projectors based on the dynamics of the nominal system and faulty system respectively, and develop a binary fault detection logic. Finally, a numerical example is given to illustrate the effectiveness of the proposed method.