Research on Performance Degradation Estimation of Key Components of High-Speed Train Bogie Based on Multi-Task...

Area of Science:

• Engineering
• Artificial Intelligence
• Signal Processing

Background:

• High-speed train bogie component degradation impacts safety and requires accurate estimation.
• Traditional methods like information theory struggle with complex vibration signals.
• Existing deep learning approaches often focus on single components, not the bogie system.

Purpose of the Study:

• To propose a multi-task deep learning model for simultaneous performance degradation estimation of key high-speed train bogie components.
• To address the limitation of current research focusing on individual components rather than the bogie as a whole system.

Main Methods:

• A multi-task and multi-scale convolutional neural network (CNN) leveraging soft parameter sharing is developed.
• The model incorporates multi-scale convolution to capture diverse signal features.
• Soft parameter sharing enables feature sharing across tasks, improving information utilization.

Main Results:

• The proposed multi-task and multi-scale CNN effectively estimates the performance degradation states of key bogie components.
• Experimental results demonstrate the model's effectiveness and superiority over existing methods.
• The approach provides a feasible scheme for enhancing bogie component performance estimation.

Conclusions:

• The developed deep learning framework offers a robust solution for holistic bogie system health monitoring.
• This method improves the accuracy and efficiency of performance degradation estimation for high-speed train bogies.
• The research contributes to ensuring the safe and reliable operation of high-speed trains.