Abstract:Structural Health Monitoring, Ahead of Print. Due to the challenges associated with acquiring fault data from high-speed train axle box bearings operating under parameter fluctuation working conditions, the diagnostic accuracy of data-driven fault diagnosis methods often suffers from insufficient fault information. To address this issue, we propose an intelligent fault diagnosis strategy utilizing a novel continuous cohomology-based meta-transfer learning (CCMTL) framework. This approach aims to enhance fault identification accuracy by developing comprehensive feature representations despite limited data availability. Initially, acoustic emission data are transformed into a barcode chart using the continuous cohomology mechanism, from which fault topology features are derived. Subsequently, these features are input into a specially designed meta-transfer model for multi-task learning, thus forming a pre-trained model. This pre-training model undergoes further refinement both externally and internally to optimize its ability to adapt quickly to few-shot tasks. According to carry out the abovementioned steps, the proposed method can effectively overcome the challenges (e.g., data scarcity, difficulty in fault feature extraction, low robustness, and generalization ability) in the fault diagnosis of high-speed train axle box bearings under parameter fluctuation working conditions. Compared with existing fault diagnosis methods, the proposed CCMTL algorithm has three obvious advantages: (1) improve the fault diagnosis robustness of high-speed train axle box bearings, (2) reduce the cost of data annotation, and (3) adapt to a wide range of working conditions and parameter fluctuations. Experimental findings validate the effectiveness and feasibility of the proposed approach in scenarios requiring few-shot fault diagnosis. Specifically, when compared with current advanced diagnostic methods, the CCMTL method demonstrates superior diagnostic accuracy for fault diagnosis in high-speed train axle box bearings under parameter fluctuation working conditions. In conclusion, the practicality and generalization ability of the proposed CCMTL-based fault diagnosis approach are thoroughly demonstrated through comprehensive ablation studies. On the other hand, the designed CCMTL framework is not limited to specific machinery or operational conditions. With proper training and adjustments, the model can be applied to other industrial equipment with similar fault patterns and data characteristics. In summary, although this article emphasizes a specific application scenario (e.g., parameter fluctuation working conditions), the proposed method possesses broad applicability and can be extended to other similar industrial fault diagnosis tasks.

Integrating Misidentification and OOD Detection for Reliable Fault Diagnosis of High-Speed Train Bogie

An uncertainty-informed framework for trustworthy fault diagnosis in safety-critical applications

A Fault Diagnosis Method for High-speed Train Bogie Based on Multi-source Data Fusion

Novel Multistate Fault Diagnosis and Location Method for Key Components of High-Speed Trains.

Multidimensional Feature Fusion and Ensemble Learning-Based Fault Diagnosis for the Braking System of Heavy-Haul Train

Hybrid physics-embedded recurrent neural networks for fault diagnosis under time-varying conditions based on multivariate proprioceptive signals

Fault Diagnosis of High-speed Train Bogie Based on Spectrogram and Multi-channel Voting

An Open-Set Method for Diagnosing Unknown Bogie Bearing Faults Using a Hybrid Open Score Relation Network

Out-of-distribution detection-assisted trustworthy machinery fault diagnosis approach with uncertainty-aware deep ensembles

Data-driven fault diagnosis for dynamic traction systems in high-speed trains

An effective framework using identification and image reconstruction algorithm for train component defect detection

Density-Based Affinity Propagation Tensor Clustering for Intelligent Fault Diagnosis of Train Bogie Bearing

Intelligent fault diagnosis of high-speed train axle box bearings under parameter fluctuation working conditions using continuous cohomology-based meta-transfer learning framework

An Online Classification Method for Fault Diagnosis of Railway Turnouts

HC-OFD Net: Heterogeneity-Captured Network Towards Out-of-Distribution Generalization in Fault Diagnosis.

A BRB-Based Effective Fault Diagnosis Model for High-Speed Trains Running Gear Systems

Just-in-time Learning-aided Nonlinear Fault Detection for Traction Systems of High-speed Trains

Fault Diagnosis of Multi-Railway High-Speed Train Bogies by Improved Federated Learning

A Simulated-to-Real Transfer Fault Diagnosis Method Based on Prototype Clustering Subdomain Adversarial Adaptation Network for HST Bogie Bearing

Bi-TLLDA and CSSVM based fault diagnosis of vehicle on-board equipment for high speed railway

A Global and Joint Knowledge Distillation Method with Gradient-Modulated Dynamic Parameter Adaption for EMU Bogie Bearing Fault Diagnosis