Artificial Neural Network-based Fault Detection

摘要

Undoubtedly, the fault diagnosis role is vital in monitoring technological processes. Regarding the modern technology ever-growing complexity, the research community has spent huge efforts to adapt diagnosis with today's systems requirements. From this sight, the present work develops an intelligent Artificial Neural-Network (ANN)-based diagnosis algorithm. Indeed, the ANN is a widespread technique in the “Artificial Intelligence” area. It is adjusted in this proposal to ensure fault-detection task. As a free-model technique, the suggested method present very promising perspectives and great convenience to a large scale of systems. Otherwise, since it is considered as a typical experimental mechanism, the inverted pendulum (IP) is selected to be our case of study. Instead of using a real IP, a model describing this system is built on Matlab/Simulink. The results of the established fault-detection method have proved its accuracy and high efficiency.