Intelligent fault diagnosis for robotic systems

摘要

Robotic systems are widely used in industry. Preventive maintenance of electrical machine systems plays a very important role in the industrial life. This requires monitoring their operations on-line which can detect a fault as it occurs and diagnosing the malfunction of a faulty component. In this paper, we present a fault diagnosis method for robotic systems. First, the fault monitoring is designed to enable the system to detect a fault occurrence based on the residual generator and parameter convergence. Subsequently, the fault isolation algorithm is designed based on known fault types. If the isolation scheme is not successful, the fault diagnosis incorporating neural network information is activated. Finally, case study is given to illustrate the effectiveness of the proposed method.