3 March 2025 in Reliability Engineering & System Safety
23 January 2026 in IEEE Transactions on Instrumentation and Measurement
Data-driven approaches are widely employed for diagnosing open-circuit faults in inverters. However, their performance often deteriorates significantly under unknown operating conditions due to reliance on known training domains. To address this, this paper proposes a cross-domain differential attention network (CDDAN). Firstly, a wavelet principal component feature extraction strategy is introduced. This method precisely locates sensitive frequency bands through wavelet transformation, then employs principal component analysis to compress them into compact, discriminative features, effectively suppressing noise and redundancy. Subsequently, the CDDAN module utilizes a differential attention mechanism to conduct parallel comparative learning across multiple known operating conditions. This mechanism explicitly captures feature differences, thereby extracting key fault characteristics insensitive to operating condition variations. To further enhance the model's generalization capability, a multi-scale differential loss function was designed. This function synergistically optimizes multiple objectives, simultaneously improving diagnostic accuracy and cross-operating-condition adaptability. Experimental results demonstrate that this method can effectively diagnose all 22 open-circuit faults under unknown operating conditions, outperforming existing mainstream approaches.
18 November 2022 in The International Journal of Advanced Manufacturing Technology
13 August 2025 in Measurement Science and Technology
Bearing reliability is fundamental to guaranteeing the secure and efficient function of equipment. As industrial settings become increasingly intricate, fault diagnosis methods for bearings must evolve to offer greater precision and adaptability. Currently, the bearing fault diagnosis domain faces several challenges, including a shortage of failure data, high levels of environmental noise, and models that offer limited interpretability. Recently, Digital Twin (DT)—an innovative approach that synergistically combines various advanced techniques—has increasingly attracted attention as a promising solution to these issues. This paper begins by examining the evolution and practical applications of bearing fault diagnosis technologies. It first outlines the fundamental concepts of DT technology and fault diagnosis, detailing their historical development and the critical techniques involved. Next, this paper provides a comprehensive review of mainstream approaches to bearing fault diagnosis from two key perspectives: signal processing techniques and data-driven methods, outlining their respective strengths and limitations. To tackle the inherent challenges of these methods, existing DT systems for bearing fault diagnosis are categorized from a modeling standpoint. In addition, a practical DT implementation framework suited to real-world industrial applications is proposed. The paper further analyzes critical research challenges and offers targeted strategies and future directions. Overall, this work aims to offer both theoretical insights and practical guidance for advancing the integration of DT technology into intelligent bearing fault diagnosis.
9 December 2025 in Engineering Research Express
Disturbances frequently occur at metro stations with large passenger flows, leading to elevated energy consumption of metro trains and an increased incidence of stranded passengers on the platform. For this, this paper aims at developing a deep reinforcement learning (DRL) based redundant time distribution approach for the real-time bi-objective metro train timetable rescheduling problem so that the energy consumption of metro trains and the stranded passenger number can be minimized. In the first layer, a multi-particle metro train operation model is formulated to obtain speed profiles in the redundant running time of each segment for the train. In the second layer, a bi-objective metro train timetable rescheduling model is developed, considering the objectives of minimizing both the pure energy consumption and the stranded passenger number. A deep Q network based redundant time distribution approach is introduced for model calculation. Finally, the practical case study focuses on Nanning Metro Line 5 for analysis and examination. Compared to the existing method, the proposed approach can realize a 5.1% reduction in energy consumption and a 5.77% decrease in the stranded passenger number, with an average calculation time of only 0.184 s. Further experiment demonstrate the effectiveness of the DRL approach.
22 January 2025 in Engineering Research Express
Remaining useful life (RUL) prediction of rolling bearings prevents sudden mechanical failures and reduces equipment maintenance costs. Due to the strong performance in time series forecasting tasks, the temporal convolutional network (TCN) has become a mainstream model for RUL prediction. However, existing TCN-based prediction models struggle to fully capture both long-term and global dependencies in complex data. To address these issues, a temporal convolutional long short-term memory network integrated with multi-head attention mechanism (TCLSTM-MA) is proposed to predict the rolling bearings’ RUL. Firstly, the time-domain and frequency-domain features are extracted from the acquired raw vibration signals to form a complete degradation feature. Secondly, we enhance the traditional TCN by combining it with LSTM and introducing a multi-head attention mechanism. This integration allows the model to effectively capture both global degradation information and local context information. Additionally, a time-weighted t-MSE loss function is employed throughout training to make the model focus more on data close to failure points. Finally, the trained TCLSTM-MA model is used for RUL prediction. Extensive experiments were conducted on two authoritative rolling bearing datasets and compared with other methods. The experimental results demonstrate that the proposed method exhibits good accuracy and generalization capability.