Traction motor bearings, serving as a critical component in trains, have a significant impact on ensuring the safety of train operations. However, there is a scarcity of sample data for bearing failures during train operations, and the complex and variable operating conditions of train bearings result in significant differences in domain distribution. Traditional cross-domain fault diagnosis methods are no longer adequate for addressing train bearing faults. Therefore, this study proposes a novel adversarial domain-adaptation meta-learning network (NADMN) for the purpose of diagnosing train bearing faults. Firstly, a deep convolutional neural network is proposed, which enhances the model’s feature extraction capability by incorporating attention mechanisms. Moreover, by employing domain adversarial adaptation learning strategy, it effectively extracts domain-invariant features from both source and target domains, thereby achieving generalization across different domains. Three experiments of bearing fault diagnosis are carried out, and the superiority of NADMN is proved by charts, confusion matrix and visualization techniques. Compared with the other five methods, NADMN showed obvious advantages in diagnostic scenarios characterized by significant changes in domain distribution.
The train signal system is crucial to ensuring the safe operation of trains. However, with long-term operation, various signal system components degrade to different degrees, leading to potential faults that pose risks to railway operations. Among them, switch machine, track circuit, axle counter, signal lamp, and power equipment are the most common and critical signal devices. This paper systematically analyzes these key signal devices’ fault causes and diagnosis methods, introducing the fundamental principles and applicable conditions of different diagnostic techniques. Furthermore, this paper explores the application of multi-source information fusion technology in train fault diagnosis, highlighting its advantages in integrating sensor data from multiple sources to enhance fault identification and localization. Finally, the study summarizes the challenges of current train signal system fault diagnosis. It outlines future research directions, emphasizing the need for more intelligent, automated, and data-driven diagnostic systems to ensure railway safety and reliability.
The safety of subway trains is paramount as a public transit. Due to the frequent start and stop, it is challenging to diagnose the bearing fault of the subway train at variable speeds. Consequently, the study proposed an approach to detecting subway train bearing faults under variable speeds. Firstly, a novel Speed-guided adaptive chirp mode decomposition (Speed-guided ACMD) algorithm is presented to obtain the optimal single-component speed signal. Secondly, a framework is designed to extract the exact speed curve. Eventually, the precise speed curve cumulative integral was used to realize the order tracking and complete the fault diagnosis of train bearings. Two cases are employed to verify the proposed approach. The experimental results show that the Speed-guided ACMD can extract the accurate speed curve and achieve accurate order tracking, which is better than other comparison techniques. The proposed procedure can provide a new solution for the fault detection of train bearings at a variable speed.
31 July 2025 in Engineering Applications of Artificial Intelligence