The theoretical developments of data-driven fault diagnosis methods have yielded fruitful achievements and significantly benefited industry practices. However, most methods are developed based on the assumption of data balance, which is incompatible with engineering scenarios. First, the normal state accounts for the majority of the equipment’s lifespan; second, the probability of various faults varies, both of which result in an imbalance in the data. The consequence of data imbalance in intelligent fault diagnosis methods has attracted extensive attention from the research community, and a significant number of papers have been published. Nevertheless, a comprehensive review of achievements in this field is still missing, and the research perspectives have not been thoroughly investigated. To end this, we review and discuss all the research achievements in fault diagnosis under data imbalance in this paper, based on our best knowledge. First, the existing imbalanced learning methods are classified into three categories: data processing methods, model construction methods, and training optimization methods. Then, the three methodologies are introduced and discussed in detail: the data processing method is to optimize the inputs of the intelligent fault diagnosis model so that the imbalance rate of the sample set involved in training is reduced; the model construction method is to design the structure and the features of the intelligent fault diagnosis model so that the model itself is resistant to the effects of imbalance; the training optimization method is an optimisation of the training process for intelligent fault diagnosis models, raising the importance of the minority class in the training. Finally, this paper summarizes the prospects of the imbalanced learning problem in intelligent fault diagnosis, discusses the possible solutions, and provides some recommendations.

In contrast to fault diagnosis relying solely on a single sensor, the method of multi-sensor information fusion for fault diagnosis (MSIFFD) broadens the spectrum of available information sources. It is renowned for its high accuracy and reliability, traits that have attracted growing attention within the research community and led to the generation of a substantial body of publications. However, there is currently a lack of a comprehensive and systematic review in this domain. Thereby, this review aims to thoroughly explore all research achievements in the field of MSIFFD. At the outset, an analysis is undertaken to delineate the fusion level of multi-sensor information, with a specific focus on its location and types of inputs and outputs within the information flow. Subsequently, an examination of the six primary fundamental operations for amalgamating multi-sensor information is undertaken to clarify the motivations and processes involved in information fusion across various operations. Following this, factors influencing fusion diagnostics and strategies aimed at improving their performance are examined to explore approaches for enhancing fusion accuracy. The subsequent sections delve into the analysis of sensor types and application scenarios, providing a reference guide for practical applications. Finally, this review outlines potential future challenges in the field of MSIFFD and provides a range of recommendations and possible solutions for consideration.

A single-vibration signal is no longer adequate to fulfill the requirements of intelligent fault diagnosis (IFD) of bearings in complex systems. With the rapid advancement of the industrial Internet of Things, IFD methods based on multimodal information fusion have gained popularity. Acoustic signals are noninvasive, easily captured, and have a wide monitoring range. Therefore, acoustic-vibration fusion IFD (AVFIFD) holds promising application prospects. Nevertheless, current AVFIFD methods suffer from two limitations that lead to reduced accuracy: insufficient consideration of both local and temporal features during the feature extraction process, and inadequate emphasis on the correlation between acoustic and vibration features. To overcome these limitations and enhance the accuracy of AVFIFD, we propose the convolution and cross-fusion transformer (CCFT), which combines convolution and transformers to enhance local and temporal feature extraction and introduces cross-fusion transformers to improve the correlation between acoustic and vibration features. Finally, fault type identification is accomplished through a fusion classification module. In two case studies, CCFT outperforms other fusion methods. Additional visualization analysis illustrates that the cross-fusion transformer can improve the correlation of fault information by progressively minimizing the discrepancies between acoustic and vibration feature representations at each layer.

Rolling bearings are prone to failure due to the complexity and serious operational environment of rotating equipment. Intelligent fault diagnosis based on convolutional neural networks (CNNs) has become an effective tool to ensure the reliable operation of rolling bearings. However, interference caused by environmental noise and variable working conditions can affect the data. To solve this problem, we propose an improved fault diagnosis method called deep convolutional neural network based on multi-scale features and mutual information (MMDCNN). In our approach, a multi-scale convolutional layer is placed at the front end of a 1D_CNN to maximize the retention of the multi-scale initial features. Meanwhile, the key fault features are further enhanced adaptively by introducing a self-attention mechanism. Then, the composite loss function is constructed by maximizing mutual information as an auxiliary loss based on cross-entropy loss; thus, the proposed method can extract robust fault features with high generalization performance. To demonstrate the superiority of MMDCNN, we compared the performance of our scheme with several existing deep learning models on two datasets. The results show that the proposed model successfully achieves bearing fault diagnosis with interference from noise and variable working conditions, possessing a powerful fault feature extraction capability.