Artificial intelligence for electric machine fault diagnosis

Electric machines (EMs) play an important role in industrial production, and their reliability is directly related to enterprise safety and economic benefits. With the continuous development of science and technology, the application fields of electric machines have been widely expanded, such as electric vehicles, wind power generation, rail transit, etc. State monitoring and fault diagnosis can provide a reliable guarantee for the normal operation of industrial systems.Common EM faults can be divided into two main categories: electrical faults and mechanical faults [1]. According to research data from the American Electric Power Research Institute (EPRI) [2] on EM faults, approximately 53% of motor faults are due to mechanical faults, such as bearing faults, unbalance faults, etc.; and approximately 47% of motor faults are electrical faults, of which about 10% are due to the rotor, such as rotor casting defects, rotor broken bars, and other factors caused by the air gap magnetic field imbalance. In addition, about 37% of the faults are due to turn-to-turn and phase-to-phase short circuits and other faults caused by damage to the insulation of the stator windings.