Traditional approaches of intelligent fault diagnosis in rotating machinery rely heavily on prior knowledge of signal processing and diagnostic expertise. Manual feature extraction is traditionally a key step for obtaining satisfactory diagnosis of certain fault types. This limits the application of these approaches to diagnosing other fault types and in new machinery. Deep neural networks have been successfully utilized in the applications of computer vision, natural language processing, speech recognition and bioinformatics. Deep neural networks based approaches of fault diagnosis are proposed and discussed. Representative features can be leant automatically from the raw signals. The effectiveness of the developed methods is evaluated using datasets of typical rotating machinery e.g., gearboxes and ball bearings. The results show the superior diagnosis performance of the proposed methods compared with traditional approaches.
Min Xia received his bachelor’s degree from Southeast University and master’s degree from University of Science and Technology of China. He is currently working towards the PhD degree in the Department of Mechanical Engineering at University of British Columbia. His research interest includes machine fault diagnosis, artificial intelligence, deep neural networks, etc. He is guest editor of Mobile Networks and Applications, Special Issue on “Cloud-assisted Cyber-Physical Systems for the Implementation of Industry 4.0” and IEEE Access, Special Section on “Key Technologies for Smart Factory of Industry 4.0”. He serves as reviewer of IEEE/ASME Transaction on Mechatronics, IEEE Transaction on Industrial Informatics, IEEE Communications Magazine, Computer Networks, Mobile Networks and Application, IEEE Access, etc.