基于深度学习的列车制动盘剩余使用寿命预测研究

打开文本图片集

中图分类号：U270.35；TP183 文献标志码：A

Research on Prediction ofRemainingUseful Life of Train Brake Disc Based on Deep Learning

Zhu Haiyan， Xu Jinhua， Xu Chenzhao，Li Xiangkun， Zhou Shengtong （SchoolofMechatronics& Vehicle Engineering，EastChina JiaotongUniversity，Nanchang33oo13，China）

Abstract： To achieve accurate prediction of the remaining useful life （RUL）of brake discs，ensure train braking safety，and optimize economical maintenance，this paper proposesa prediction model based on the fusion of selfattention mechanism and long short-term memory network （BiLSTM-SA）， which takes the crack propagation lifeas the divisionbasis.Firstly，the test data ofbrake discs arecollectedand the working conditions arecalibrated，andathermal-mechanical coupling finiteelement model is established toobtain the simulation dataset.Secondly，a Time-GAN neural network isconstructed， which enhances data througha double-layer LSTM generator and a physical constraint discriminator.Itsdistribution similarity，root meansquare errorandcoefficientof determination are significantly beter than traditional models.Finally，the BiLSTM-SA fusion prediction model is proposed，which uses bidirectional LSTMand self-atention mechanism to capture temporal dependencies and key features. In the prediction of single expanding cracks，the RMSE is reduced by 49.8% and 46.5% compared with the traditional LSTMand TCN-LSTM，respectively.In complex working conditions，the RMSE and Score are optimized by 25.5% and 51.1% ，respectively， significantly improving the prediction accuracy and robustness.

This study can provide areliable technical solution for condition monitoring and preventive maintenance of highspeed train brake discs.

Key words： brake disc; fatigue crack;remaining life prediction; time series generation adversarial network; selfattentionmechanism

Citation format： ZHU HY，XUJH，XU C Z，et al. Research on prediction of remaining useful life of train brake disc based on deep learning[J]. Journal ofEast China Jiaotong University，2025，42（4）： 48-61.

深度学习技术作为早期机器学习演化的一种分支，相较于机器学习技术具有能够自动提取特征数据的优势[1]。（剩余16916字）