A Data-driven Approach for Identification and Detection of Intermittent Faults

The main focus of this paper is to analyze and design the data-driven fault detection (FD) approach for systems with intermittent faults (IFs). To this end, the systems with IFs are modeled by switched systems first, then based on this framework, the schemes of the design are proposed for a stochastic process is proposed, and the data-driven method is utilized to generate residual signals. The effectiveness and performance of the theoretical results are verified and demonstrated through numerical examples.