No abstract

The objective of this paper is to present a comprehensive review of the contemporary techniques for fault detection, diagnosis, and prognosis of rolling element bearings (REBs). Data-driven approaches, as opposed to model-based approaches, are gaining in popularity due to the availability of low-cost sensors and big data. This paper first reviews the fundamentals of prognostics and health management (PHM) techniques for REBs. A brief description of the different bearing-failure modes is given, then, the paper presents a comprehensive representation of the different health features (indexes, criteria) used for REB fault diagnostics and prognostics. Thus, the paper provides an overall platform for researchers, system engineers, and experts to select and adopt the best fit for their applications. Second, the paper provides overviews of contemporary REB PHM techniques with a specific focus on modern artificial intelligence (AI) techniques (i.e., shallow learning algorithms). Finally, deep-learning approaches for fault detection, diagnosis, and prognosis for REB are comprehensively reviewed.

In the era of the fourth industrial revolution (Industry 4.0) and the Internet of Things (IoT), real-time data is enormously collected and analyzed from mechanical equipment. By classifying and characterizing the measured signals, the fault condition of mechanical components could be identified. However, most current health monitoring techniques utilize time-consuming and labor-intensive feature engineering, i.e., feature extraction and selection, that are carried out by experts. This paper, on the contrary, deals with an automatic diagnosis method of machine monitoring using a convolutional neural network (CNN) with class activation maps (CAM). A class activation map enables us to discriminate the fault region in the images, thus allowing us to localize the fault precisely. The goal of the paper is to demonstrate how CNN and CAM could be employed to real-world vibration video to characterize the machine's status, representing normal or fault conditions. The performance of the proposed model is validated with a baseexcited cantilever beam dataset and a water pump dataset. This paper presents a novel industrial application by developing a promising method for automatic machine condition-based monitoring.INDEX TERMS Convolutional neural network, class activation maps, discriminative region, fault detection, mechanical component, explainable AI

No abstract

Fault diagnosis of rotor systems is important to prevent unexpected failures. Recently, deep learning (DL) methods, such as a convolutional neural network (CNN), have been utilized in many research areas, including fault diagnosis. DL has gained significant attention thanks to its ability to efficiently learn proper features from input data. It is possible to learn enriched hierarchical features by making the DL architectures deeper; therefore, many studies have been conducted to stack the neural networks, which are the basic building blocks of DL, deeper. However, it becomes difficult to comprehensively train neural network architectures as they become deeper, due to problems in the flow of gradient information during the training phase. In this paper, a direct connection based CNN (DC-CNN) method is proposed to significantly improve training efficiency and diagnosis performance. DC-CNN connects feature maps of different layers within a CNN to improve the gradient information flow over the layers. These additional connections, however, can increase the number of trainable parameters within the network. To prevent problems that might be caused by an increased number of parameters, dimension reduction modules are also developed. Moreover, to consider the anisotropic characteristics inherent in rotor systems, the vibration images containing both spatial and temporal information are generated and utilized. The effectiveness of DC-CNN is validated using experimental data from a rotor testbed. The experimental results indicate that the proposed method outperforms other conventional approaches with a smaller number of parameters. Also, visualizations of the learned features indicate that the proposed method can learn much more effective and significant features. Furthermore, the proposed method outperforms other approaches under conditions of insufficient or noisy data.