Scholarlink

<p>As the core component of the power system in high-speed trains, the traction motor directly determines the power output efficiency and operational safety of the train. Consequently, real-time monitoring and accurate prediction of its temperature constitute a critical link in ensuring the safe and stable operation of high-speed trains. However, traditional prediction models generally suffer from insufficient capability to capture long-range dependencies when processing temperature time series data, resulting in prediction accuracy that fails to meet the requirements of practical applications. To address this issue, this paper proposes a CNN-Bi-GTU-WAVE-based temperature prediction model for high-speed train traction motors, with its practical application value realized by relying on HPC technology. Bi-GTU, the core module of the proposed model, is developed based on the bidirectional gated recurrent unit and integrated with temperature trend information and train operational state information, which enhances the ability to capture long-range dependencies. Meanwhile, a dilated convolution feature extraction module is designed to expand the receptive field flexibly, thereby improving the sensitivity to multi-scale change features of temperature. Additionally, a weighted average layer is introduced to dynamically optimize feature weights, strengthening the bidirectional network’s ability to learn key information in the middle of the sequence and further enhancing prediction accuracy. Experimental validation using actual operational data from CR300BF electric multiple units demonstrates that, supported by HPC, the proposed model not only achieves a single prediction latency that meets the needs of real-time monitoring, but also exhibits significantly superior performance in terms of prediction accuracy and stability compared to traditional methods. Particularly, it shows outstanding performance in the scenario of abnormal temperature prediction for traction motors. This study not only provides reliable technical support for traction motor fault early warning, but also verifies the core role of HPC technology in promoting the application of complex data-driven models in high-speed railway safety-critical scenarios.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A high-speed train traction motor temperature prediction model based on improved Bi-GTU

  • 期刊论文

  • 发表日期：

    2025年9月22日

查看全文
  • Yunming Wang,
  • Jinbo Tang,
  • Menghua Chen,
  • Xianwu Chu

摘要

As the core component of the power system in high-speed trains, the traction motor directly determines the power output efficiency and operational safety of the train. Consequently, real-time monitoring and accurate prediction of its temperature constitute a critical link in ensuring the safe and stable operation of high-speed trains. However, traditional prediction models generally suffer from insufficient capability to capture long-range dependencies when processing temperature time series data, resulting in prediction accuracy that fails to meet the requirements of practical applications. To address this issue, this paper proposes a CNN-Bi-GTU-WAVE-based temperature prediction model for high-speed train traction motors, with its practical application value realized by relying on HPC technology. Bi-GTU, the core module of the proposed model, is developed based on the bidirectional gated recurrent unit and integrated with temperature trend information and train operational state information, which enhances the ability to capture long-range dependencies. Meanwhile, a dilated convolution feature extraction module is designed to expand the receptive field flexibly, thereby improving the sensitivity to multi-scale change features of temperature. Additionally, a weighted average layer is introduced to dynamically optimize feature weights, strengthening the bidirectional network’s ability to learn key information in the middle of the sequence and further enhancing prediction accuracy. Experimental validation using actual operational data from CR300BF electric multiple units demonstrates that, supported by HPC, the proposed model not only achieves a single prediction latency that meets the needs of real-time monitoring, but also exhibits significantly superior performance in terms of prediction accuracy and stability compared to traditional methods. Particularly, it shows outstanding performance in the scenario of abnormal temperature prediction for traction motors. This study not only provides reliable technical support for traction motor fault early warning, but also verifies the core role of HPC technology in promoting the application of complex data-driven models in high-speed railway safety-critical scenarios.