A Review of Fault Diagnosis in Mechatronics Systems

Abstract

Fault diagnosis gets increasing importance in today’s production environment and with the advancements in the field of artificial intelligence, researchers look for new ways to keep a system away from faults, that would interrupt the production. In recent years, many papers were written regarding this subject especially regarding predictive maintenance and fault diagnosis. This paper presents recent works that expose new methods for intelligent fault diagnosis. This step is important for future research in order to have a better understanding of state of the art algorithms and look for ways to improve the existing fault diagnosis approaches. The focus will be on electrical systems and actuators and manipulating systems (robots), production experience showing that the mechanical parts are the most exposed to production-ending faults. That’s why most of the observer systems are using vibrations as the main data for their algorithms, but also other measurements can provide useful information about the condition of a system.