Survey of Real-Time Fault Diagnosis Techniques for Electromechanical Systems

Abstract : This paper deals with a review of research works which have been recently carried out on the topic of real-time fault diagnosis of electromechanical systems. The proposed approaches rely mainly on the implementation of fault diagnosis algorithms in digital signal processors and field programmable gate arrays as well as usage of real-time simulators applied formerly in the context of hardware-in-the-loop for prompt and reliable fault diagnosis of electromechanical systems. This covers three groups of fault diagnosis namely, signal-based, modelbased and data-based methods. The main challenge is to optimize both the utilization of hardware resources and computation time in addition to having a prompt and reliable fault diagnosis. A single processing unit for both fault detection and control algorithms will be the ultimate aim which leads to a fault tolerant configuration ensuring continuous operation of any electromechanical system and its related electrical drive.