Data-Driven Fault Diagnosis Methods

By mining and analyzing equipment maintenance, repair, and operation data, potential product failures can be diagnosed in a timely manner, maintenance needs can be generated, and intelligent predictive maintenance as shown in Fig.  1.2 can be proactively executed to achieve intelligent operation services. This chapter mainly introduces the feature extraction methods of non-stationary non-linear time series signals, and data-driven fault diagnosis methods based on convolutional neural networks, ensemble learning, and transfer learning.