Jiacan Xu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Jiacan Xu has a total of 36 co-authors. Here are the top co-authors by number of shared publications:
- Dazhong Ma
Northeastern University
5 shared publications - Donglin Li
College of Electrical Engineering, Shenyang University of Technology, Shenyang, CHINA
5 shared publications - Peng Zhou
School of Mechanical Engineering, Shenyang Jianzhu University, Shenyang, China
5 shared publications - Zinan Wang
School of Mechanical Engineering, Shenyang Jianzhu University, Shenyang, China
5 shared publications Peng Zhou
4 shared publications- Yuxian Zhang
Shenyang University of Technology
3 shared publications
Latest Publications
- 1 October 2026
- Published by IOP Publishing in Journal of Instrumentation
- Vol. 21 (10) , P10025
A multi-label dynamic decoupling method for compound faults of rolling bearings based on contrastive learning
Rolling bearings are core components of rotating machinery and power-transmission systems, and accurate decoupling of compound faults is crucial for equipment safety and efficiency. Compound fault vibration signals are a nonlinear coupling mixture of multiple single fault characteristics in the time-frequency domain. Without explicit measurement of feature similarities and differences, coupling interference arises in the embedding space during feature extraction and decoupling, causing feature confusion and severely limiting compound fault decoupling accuracy. This paper proposes a multi-label dynamic decoupling method for compound faults of rolling bearings based on contrastive learning (CL-MCFDD). Firstly, a multi-level convolutional embedding module is adopted to extract the deep feature representation of the vibration signal, thereby capturing multi-scale fault information from local details to the global structure. And the relation module is utilized to construct a cosine similarity matrix (CSM) among samples, explicitly modeling the similarity relationships between different fault categories. Meanwhile, the contrastive learning (CL) strategy is introduced to enhance the aggregation of similar fault features and the separability of dissimilar features. Secondly, to balance the separability of features and the classification performance, a joint loss function is proposed, which integrates the triplet contrastive loss (TCL) and the mean square error loss (MSEL) to adjust and optimize the objective. Finally, an attention mechanism is introduced into the decoupled classifier to construct a multi-label attention decoupled classifier (ML-ADC). The key fault features are strengthened through the dynamic feature weighting mechanism, and multi-label probabilities are output to achieve decoupling of compound faults. Experiments on PU and laboratory datasets yield decoupling accuracies of 98.59% and 97.39%, respectively. The results show that this method improved the robustness and generalization ability of the decoupling of compound faults in rolling bearings, providing a new idea for decoupling compound faults.
- 1 October 2026
- Published by IOP Publishing in Journal of Instrumentation
- Vol. 21 (10) , P10025
A multi-label dynamic decoupling method for compound faults of rolling bearings based on contrastive learning
Rolling bearings are core components of rotating machinery and power-transmission systems, and accurate decoupling of compound faults is crucial for equipment safety and efficiency. Compound fault vibration signals are a nonlinear coupling mixture of multiple single fault characteristics in the time-frequency domain. Without explicit measurement of feature similarities and differences, coupling interference arises in the embedding space during feature extraction and decoupling, causing feature confusion and severely limiting compound fault decoupling accuracy. This paper proposes a multi-label dynamic decoupling method for compound faults of rolling bearings based on contrastive learning (CL-MCFDD). Firstly, a multi-level convolutional embedding module is adopted to extract the deep feature representation of the vibration signal, thereby capturing multi-scale fault information from local details to the global structure. And the relation module is utilized to construct a cosine similarity matrix (CSM) among samples, explicitly modeling the similarity relationships between different fault categories. Meanwhile, the contrastive learning (CL) strategy is introduced to enhance the aggregation of similar fault features and the separability of dissimilar features. Secondly, to balance the separability of features and the classification performance, a joint loss function is proposed, which integrates the triplet contrastive loss (TCL) and the mean square error loss (MSEL) to adjust and optimize the objective. Finally, an attention mechanism is introduced into the decoupled classifier to construct a multi-label attention decoupled classifier (ML-ADC). The key fault features are strengthened through the dynamic feature weighting mechanism, and multi-label probabilities are output to achieve decoupling of compound faults. Experiments on PU and laboratory datasets yield decoupling accuracies of 98.59% and 97.39%, respectively. The results show that this method improved the robustness and generalization ability of the decoupling of compound faults in rolling bearings, providing a new idea for decoupling compound faults.
- 3 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 75 (00189456) , 3516110
Edge-Oriented Bearing Fault Diagnosis via Triple-Lightweight Network With Adaptive Pruning
Real-time diagnosis of motor bearing faults is essential for fault detection and timely maintenance, minimizing operational risks. However, most existing approaches often depend on cloud platforms or high-performance local servers, which introduce latency and potential security vulnerabilities due to raw data transmission. To overcome these limitations, this paper presents an adaptive pruning triple-lightweight network (APTL-net), which achieves model lightweighting for edge deployment across three stages: model construction, feature extraction, and structural pruning. First, a lightweight fault diagnosis framework optimized for edge environments is proposed, enabling parallel training and recursive inference to eliminate the traditional trade-off between training efficiency and inference performance. Second, a frequency-domain decomposition-based, weight-sharing multiscale convolution module is designed to efficiently extract multiscale features from periodic signals while harmonizing scale variations. Moreover, a fully automated adaptive pruning strategy is proposed to dynamically group and remove redundant structures during training without manual intervention. Finally, APTL-net is deployed on the edge device and validated on the physical platform equipped with real hardware. Experimental results indicate that APTL-net achieves a diagnostic accuracy of 99.54% while reducing parameters by 47.82%, FLOPs by 49.98%, and edge inference latency by 20.59%, thereby delivering superior performance over existing lightweight fault diagnosis networks.
- 3 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 75 (00189456) , 3516110
Edge-Oriented Bearing Fault Diagnosis via Triple-Lightweight Network With Adaptive Pruning
Real-time diagnosis of motor bearing faults is essential for fault detection and timely maintenance, minimizing operational risks. However, most existing approaches often depend on cloud platforms or high-performance local servers, which introduce latency and potential security vulnerabilities due to raw data transmission. To overcome these limitations, this paper presents an adaptive pruning triple-lightweight network (APTL-net), which achieves model lightweighting for edge deployment across three stages: model construction, feature extraction, and structural pruning. First, a lightweight fault diagnosis framework optimized for edge environments is proposed, enabling parallel training and recursive inference to eliminate the traditional trade-off between training efficiency and inference performance. Second, a frequency-domain decomposition-based, weight-sharing multiscale convolution module is designed to efficiently extract multiscale features from periodic signals while harmonizing scale variations. Moreover, a fully automated adaptive pruning strategy is proposed to dynamically group and remove redundant structures during training without manual intervention. Finally, APTL-net is deployed on the edge device and validated on the physical platform equipped with real hardware. Experimental results indicate that APTL-net achieves a diagnostic accuracy of 99.54% while reducing parameters by 47.82%, FLOPs by 49.98%, and edge inference latency by 20.59%, thereby delivering superior performance over existing lightweight fault diagnosis networks.
- 31 March 2026
- Published by IOP Publishing in Journal of Neural Engineering
- Vol. 23 (2) , 026022
Feature alignment and enhancement network with guided tuning for non-stationary EEG classification
Objective. Electroencephalogram (EEG) signal variability caused by external factors and subject differences limits the adaptation of motor imagery (MI) classification models in brain–computer interfaces (BCIs). Existing domain alignment methods often inadequately utilize critical source and target domains information, leading to negative transfer problems. This paper proposes a Feature Alignment and Enhancement Framework for cross-domain MI-EEG classification to address these limitations. Approach. First, by aligning the covariance matrices of the source and target domains, the spatial distributions of the two domains are preliminarily aligned, establishing a consistent foundation for feature mapping. Second, a conditional domain adversarial network optimizes cross-domain representations, reducing distribution discrepancies while enhancing discriminability. Finally, this paper introduces an EEG feature-based guided tuning method. This method extracts high-confidence features from both the source and target domains and generates centroid features to construct cross-domain feature banks. The input feature representations are dynamically optimized by attending to the relationships between centroid features, thus enhancing the model’s adapt-ability to target domain tasks. Main results. Experimental data show that in the four-class MI task of the BCI Competition IV-2a dataset, the cross-session and cross-subject model classification accuracies were 76.89% and 57.91%, respectively. The model achieved accuracy rates of 84.61% and 82.78% on the BCI Competition IV-2b datasets and the High Gamma Datasets, respectively, as well as 84.09% and 70.81%. Significance. The proposed framework effectively mitigates cross-domain variations, providing a reliable solution for cross-session and cross-subject MI-EEG classification.
- 31 March 2026
- Published by IOP Publishing in Journal of Neural Engineering
- Vol. 23 (2) , 026022
Feature alignment and enhancement network with guided tuning for non-stationary EEG classification
Objective. Electroencephalogram (EEG) signal variability caused by external factors and subject differences limits the adaptation of motor imagery (MI) classification models in brain–computer interfaces (BCIs). Existing domain alignment methods often inadequately utilize critical source and target domains information, leading to negative transfer problems. This paper proposes a Feature Alignment and Enhancement Framework for cross-domain MI-EEG classification to address these limitations. Approach. First, by aligning the covariance matrices of the source and target domains, the spatial distributions of the two domains are preliminarily aligned, establishing a consistent foundation for feature mapping. Second, a conditional domain adversarial network optimizes cross-domain representations, reducing distribution discrepancies while enhancing discriminability. Finally, this paper introduces an EEG feature-based guided tuning method. This method extracts high-confidence features from both the source and target domains and generates centroid features to construct cross-domain feature banks. The input feature representations are dynamically optimized by attending to the relationships between centroid features, thus enhancing the model’s adapt-ability to target domain tasks. Main results. Experimental data show that in the four-class MI task of the BCI Competition IV-2a dataset, the cross-session and cross-subject model classification accuracies were 76.89% and 57.91%, respectively. The model achieved accuracy rates of 84.61% and 82.78% on the BCI Competition IV-2b datasets and the High Gamma Datasets, respectively, as well as 84.09% and 70.81%. Significance. The proposed framework effectively mitigates cross-domain variations, providing a reliable solution for cross-session and cross-subject MI-EEG classification.
- 13 August 2025
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (8) , 086141
Zero-shot gear fault diagnosis based on attribute consistency embedding
In industrial scenarios, as a core component of mechanical transmission systems, gears are prone to various faults, such as wear and fatigue after long-term operation. Diagnosing its operating status is of vital importance for ensuring production safety. However, there are many difficulties in actually obtaining gear fault samples, such as the difficulty in collecting rare fault samples and the high cost of collecting some faults. To alleviate this problem, this paper proposes a zero-shot gear fault diagnosis method based on Attribute Consistency Embedding (ACE). The method constructs semantic links between fault samples and attributes by designing semantic vectors of length-width-depth attributes with gear surface faults. The consistent mapping of cross-modal semantics has been achieved. The embeddedreconstruction structure is adopted to jointly map the fault samples and attribute vectors to a unified semantic embedding space. The correlation-consistent embedding matrix with an orthogonal constraint optimization strategy is introduced to strengthen the clustering of similar faults and the separability from dissimilar faults. Experiments were conducted on the dataset collected on the gear test bench. The results show that the average diagnostic accuracy rate of the method proposed in this paper in the unseen fault category reaches 93.65%. The fault features and attribute vectors can be precisely aligned in the semantic embedding space through the correlation-consistent embedding matrix and feature visualization. Enable the samples of the same type of faults to be highly clustered in the feature space and the faults of different types to be significantly distinguishable. The Grad-CAM heat map further indicates that the key fault characteristics such as high-frequency shock in the vibration signal can be effectively retained through the embedding-reconstruction structure. It is conducive to providing a visual verification basis for the fault diagnosis model.
- 13 August 2025
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (8) , 086141
Zero-shot gear fault diagnosis based on attribute consistency embedding
In industrial scenarios, as a core component of mechanical transmission systems, gears are prone to various faults, such as wear and fatigue after long-term operation. Diagnosing its operating status is of vital importance for ensuring production safety. However, there are many difficulties in actually obtaining gear fault samples, such as the difficulty in collecting rare fault samples and the high cost of collecting some faults. To alleviate this problem, this paper proposes a zero-shot gear fault diagnosis method based on Attribute Consistency Embedding (ACE). The method constructs semantic links between fault samples and attributes by designing semantic vectors of length-width-depth attributes with gear surface faults. The consistent mapping of cross-modal semantics has been achieved. The embeddedreconstruction structure is adopted to jointly map the fault samples and attribute vectors to a unified semantic embedding space. The correlation-consistent embedding matrix with an orthogonal constraint optimization strategy is introduced to strengthen the clustering of similar faults and the separability from dissimilar faults. Experiments were conducted on the dataset collected on the gear test bench. The results show that the average diagnostic accuracy rate of the method proposed in this paper in the unseen fault category reaches 93.65%. The fault features and attribute vectors can be precisely aligned in the semantic embedding space through the correlation-consistent embedding matrix and feature visualization. Enable the samples of the same type of faults to be highly clustered in the feature space and the faults of different types to be significantly distinguishable. The Grad-CAM heat map further indicates that the key fault characteristics such as high-frequency shock in the vibration signal can be effectively retained through the embedding-reconstruction structure. It is conducive to providing a visual verification basis for the fault diagnosis model.
- Open Access
- 17 July 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (14) , 7983
Identification and Experimental Study of Sand Gravel Formations Driven by an Earth Pressure Balance Shield Machine Based on GTNet
The earth pressure balance shield machine (EPB) is an important piece of engineering equipment used in tunnel excavation and plays an important role in large underground tunnel projects. This article takes the sand and gravel formation as the research object, while discrete element simulation is utilized to study the correlation between cutterhead torque and thrust and other parameters. The EPB tunneling experiment was carried out by setting up formations with different sand and gravel contents. The reliability of the simulation model was verified by the experimental data, which provided the data samples for the training of the excavation formation identification network. Finally, a GTNet (gated Transformer network) based on the formation identification method was proposed. The reliability of the network model was verified by contrasting the model used with other network models and by analyzing the results of experiment and visualization. The effects of different parameters were weighted using the ablation study for tunneling parameters. The proposed method has a high accuracy of 0.99, and the cutterhead torque and thrust have a great recognition feature, the weight of which is over 0.95. This paper can provide significant guidance for the torque and thrust analysis of cutterheads in tunnel construction.
- 17 July 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (14) , 7983
Open AccessIdentification and Experimental Study of Sand Gravel Formations Driven by an Earth Pressure Balance Shield Machine Based on GTNet
The earth pressure balance shield machine (EPB) is an important piece of engineering equipment used in tunnel excavation and plays an important role in large underground tunnel projects. This article takes the sand and gravel formation as the research object, while discrete element simulation is utilized to study the correlation between cutterhead torque and thrust and other parameters. The EPB tunneling experiment was carried out by setting up formations with different sand and gravel contents. The reliability of the simulation model was verified by the experimental data, which provided the data samples for the training of the excavation formation identification network. Finally, a GTNet (gated Transformer network) based on the formation identification method was proposed. The reliability of the network model was verified by contrasting the model used with other network models and by analyzing the results of experiment and visualization. The effects of different parameters were weighted using the ablation study for tunneling parameters. The proposed method has a high accuracy of 0.99, and the cutterhead torque and thrust have a great recognition feature, the weight of which is over 0.95. This paper can provide significant guidance for the torque and thrust analysis of cutterheads in tunnel construction.
- Open Access
- 1 January 2025
- Published by Institution of Engineering and Technology (IET) in IET Cyber-Physical Systems: Theory & Applications
- Vol. 10 (1)
APHformerNET: A Gear Fault Diagnosis Model Based on Adaptive Prototype Hashing Optimisation Algorithm
Fault‐diagnosis methods based on deep learning technology have been widely applied in gear fault diagnosis. Gearboxes often operate under complex and harsh conditions, which can lead to faults. Therefore, monitoring the condition of gearboxes and diagnosing faults are crucial for ensuring the reliability and safety of the system. In response, this paper proposes a gear fault diagnosis model based on the adaptive prototype hashing (APH) optimisation algorithm for diagnosing faults in rotating machinery. This method combines the advantages of adaptive prototype hashing with transformers to improve the accuracy of fault diagnosis. The model utilises an adaptive prototype selection mechanism to dynamically select the most representative samples as prototypes and employs the transformer model to extract feature representations of the input data. In classification tasks using two datasets, the model achieved an accuracy of 98.11% under normal conditions. In experiments with added white noise and a smaller sample size, the accuracies reached 96.81% and 86.41%, respectively. Additionally, we conducted ablation experiments with advanced transformer models, where the APHformer model incorporating the APH layer achieved fault diagnosis accuracies exceeding 97%, significantly outperforming other combinations. Furthermore, T‐SNE visualisation results indicate that the method performs well in feature representation. This study provides important insights into the field of gear fault diagnosis based on deep learning and has potential practical application values.
- 1 January 2025
- Published by Institution of Engineering and Technology (IET) in IET Cyber-Physical Systems: Theory & Applications
- Vol. 10 (1)
Open AccessAPHformerNET: A Gear Fault Diagnosis Model Based on Adaptive Prototype Hashing Optimisation Algorithm
Fault‐diagnosis methods based on deep learning technology have been widely applied in gear fault diagnosis. Gearboxes often operate under complex and harsh conditions, which can lead to faults. Therefore, monitoring the condition of gearboxes and diagnosing faults are crucial for ensuring the reliability and safety of the system. In response, this paper proposes a gear fault diagnosis model based on the adaptive prototype hashing (APH) optimisation algorithm for diagnosing faults in rotating machinery. This method combines the advantages of adaptive prototype hashing with transformers to improve the accuracy of fault diagnosis. The model utilises an adaptive prototype selection mechanism to dynamically select the most representative samples as prototypes and employs the transformer model to extract feature representations of the input data. In classification tasks using two datasets, the model achieved an accuracy of 98.11% under normal conditions. In experiments with added white noise and a smaller sample size, the accuracies reached 96.81% and 86.41%, respectively. Additionally, we conducted ablation experiments with advanced transformer models, where the APHformer model incorporating the APH layer achieved fault diagnosis accuracies exceeding 97%, significantly outperforming other combinations. Furthermore, T‐SNE visualisation results indicate that the method performs well in feature representation. This study provides important insights into the field of gear fault diagnosis based on deep learning and has potential practical application values.
- Open Access
- 1 January 2025
- Published by Wiley in Shock and Vibration
- Vol. 2025 (1)
Effect of Structural Parameters on the Vibration of Slewing Bearing Under External Loading
As the main bearing component of the tunnel boring machine (TBM), the three‐row roller slewing bearing is required to withstand a heavy load and overturning moment in the tunnelling process. Therefore, a dynamic model of a slewing bearing considering the external load and internal structural parameters is established. First, the influence of the complex load on the internal structural parameters (clearance, radial displacement, roller tilt and skew) of the slewing bearing is analysed. Then, by using the four‐stage Runge–Kutta method, the dynamic model is solved, and the influence of different parameters on the dynamic performance and stability is obtained. Finally, the proposed model is verified via experimental research, and the accuracy error of the model is less than 7%. The state of the inner ring’s motion progressively changes from stable to unstable as the rotational speed increases. The dynamic characteristics of a slewing bearing can be improved by reducing the radial clearance and the number of rollers while it is subjected to excessive eccentric loading. The results of this study provide important guidance for the design of slewing bearing structures.
- 1 January 2025
- Published by Wiley in Shock and Vibration
- Vol. 2025 (1)
Open AccessEffect of Structural Parameters on the Vibration of Slewing Bearing Under External Loading
As the main bearing component of the tunnel boring machine (TBM), the three‐row roller slewing bearing is required to withstand a heavy load and overturning moment in the tunnelling process. Therefore, a dynamic model of a slewing bearing considering the external load and internal structural parameters is established. First, the influence of the complex load on the internal structural parameters (clearance, radial displacement, roller tilt and skew) of the slewing bearing is analysed. Then, by using the four‐stage Runge–Kutta method, the dynamic model is solved, and the influence of different parameters on the dynamic performance and stability is obtained. Finally, the proposed model is verified via experimental research, and the accuracy error of the model is less than 7%. The state of the inner ring’s motion progressively changes from stable to unstable as the rotational speed increases. The dynamic characteristics of a slewing bearing can be improved by reducing the radial clearance and the number of rollers while it is subjected to excessive eccentric loading. The results of this study provide important guidance for the design of slewing bearing structures.
- 28 October 2024
- Published by SAGE Publications in Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science
- Vol. 239 (4) , 1395-1408
Research the lubrication characteristics on the modified main thrust roller of the slewing bearings
The load capacity and lubrication performance of the slewing bearing are fairly sensitive to edge effects at both ends of the main thrust rollers. Therefore, quantitative investigations of the edge effect at both ends of slewing bearing is practically important. In this study, an elastic fluid lubrication model of the slewing bearing that considered the external load and internal structural parameters is established. A modification function of rollers is included in a finite long-line contact elastic fluid lubrication model to perform the dynamic analysis. The simulated oil film’s thickness and pressure distributions are compared under different working conditions. And the edge effect is eliminated by modifying the rollers’ construction. The results indicate that under the large eccentric load, the speed of the slewing bearing must be decreased to ensure elastic fluid lubrication performance. A genetic algorithm that used to discover the optimal modification length shows the best busbar of the main thrust roller is 12 μm, which could decrease the vibration amplitude of the unmodified roller by 8.13%. This result gives a significant way for slewing bearing’s structural design.
- 28 October 2024
- Published by SAGE Publications in Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science
- Vol. 239 (4) , 1395-1408
Research the lubrication characteristics on the modified main thrust roller of the slewing bearings
The load capacity and lubrication performance of the slewing bearing are fairly sensitive to edge effects at both ends of the main thrust rollers. Therefore, quantitative investigations of the edge effect at both ends of slewing bearing is practically important. In this study, an elastic fluid lubrication model of the slewing bearing that considered the external load and internal structural parameters is established. A modification function of rollers is included in a finite long-line contact elastic fluid lubrication model to perform the dynamic analysis. The simulated oil film’s thickness and pressure distributions are compared under different working conditions. And the edge effect is eliminated by modifying the rollers’ construction. The results indicate that under the large eccentric load, the speed of the slewing bearing must be decreased to ensure elastic fluid lubrication performance. A genetic algorithm that used to discover the optimal modification length shows the best busbar of the main thrust roller is 12 μm, which could decrease the vibration amplitude of the unmodified roller by 8.13%. This result gives a significant way for slewing bearing’s structural design.