A WT-ResNet based fault diagnosis model for the urban rail train transmission system

TL;DR

This work targets fault diagnosis for urban rail transmission systems by integrating wavelet-based time-frequency feature extraction with Residual Networks (WT-ResNet). The approach leverages Morlet wavelets to transform sensor signals into discriminative 64×64 feature images and uses CNN/ResNet architectures to classify faults across multiple types. Experimental results with ResNet50 and ResNet34 on WT-transformed data show competitive class-wise performance and reasonable overall accuracy, highlighting strengths in capturing transient fault features and robust learning, while also revealing areas for improved generalization. The proposed framework has implications for proactive maintenance, reduced downtime, and enhanced reliability in urban rail operations, with potential extensions to real-time diagnostics and broader fault repertoires.

Abstract

This study presents a novel fault diagnosis model for urban rail transit systems based on Wavelet Transform Residual Neural Network (WT-ResNet). The model integrates the advantages of wavelet transform for feature extraction and ResNet for pattern recognition, offering enhanced diagnostic accuracy and robustness. Experimental results demonstrate the effectiveness of the proposed model in identifying faults in urban rail trains, paving the way for improved maintenance strategies and reduced downtime.

Figures (10)

• Figure 1: Convolutional neural network structure diagram
• Figure 2: Structure diagram of the residual block
• Figure 3: Resnet-101 network diagram
• Figure 4: The overall framework of this work
• Figure 5: Resnet 50 Validation Accuracy over Epochs