Perspectives of Transfer Learning on the Diagnosis of Faults in Electrical Machines, Power Electronics, and Drives

Fault Diagnosis and Failure Prognosis on Electrical Machines, Drives, and Power Electronics has seen rapid advancement in the last decades. Nevertheless, there is a pressing need for rapid and data efficient training of algorithms that can be applied across platforms and under varying operating conditions and environments. Contemporary machine learning methods based on transfer learning and semantic consistent identification techniques have found wide acceptance in the processing of audio, video and other applications, while requiring minimal training. This review discusses the available techniques and the applications to-date to the diagnosis and identification of faults, and failure prognosis.