ABSTRACT
Fault diagnosis of mechanical systems plays an increasingly important role in reducing costly breakdowns and avoiding catastrophic accidents in modern industry. With the development of artificial intelligence, especially deep learning (DL) approaches, the application of DL-enabled methods to diagnose potential equipment malfunctions has gone through tremendous progress with verified success in both academia and industry. In this chapter, we will first determine the significance of prognostics and health management (PHM) and intelligent fault diagnosis (IFD) for mechanical systems. Then, we will introduce the basic concepts of deep neural network (DNN) and provide a brief review of the technologies used for DL-enabled IFD. Then we will provide a roadmap for the following chapters. With our in-depth research on advanced topics, such as autoencoders, deep belief networks, convolutional neural networks, data augmentation, multisensor fusion, unsupervised deep transfer learning, neural architecture search, self-supervised learning, and reinforcement learning, we aim to reveal the potential of these methods in IFD and their advantages and contributions. We are committed to fundamentally changing the essence of IFD, thereby improving the efficiency, safety, and reliability of mechanical systems in different industrial fields.
