Scholarlink

<p>As the rail transit comes to the intelligent maintenance stage from the massive construction stage in China, the operation condition of the train bogie is more and more complicated due to the high speed and heavy load, and new sensing devices and data fusion algorithms are required for structural health monitoring (SHM) and fault detection of high-speed trains. A flexible dual functional sensor (FDFS) with piezoelectric and hot-film sensor units is designed for structural vibration and aerodynamic signal measurement of the train bogie. It is a conformal contact with the curved surface of bogie for improving the measurement accuracy of the airflow and vibration signal. A dual-channel one-dimensional Residual Network (1D ResNet) model based on the structural vibration and air flow signal is proposed for the operation condition recognition of the train bogie. Experiments have validated that the seven typical operation conditions of the bogie have been successfully recognized by the dual channel 1D ResNet model and FDFS with 92.3% average recognition success rate. The dual channel 1D ResNet has great advantage in the condition recognition of train bogie compared to dual channel 1D convolutional neural network (CNN), multilayer perception (MLP) and long short-term memory (LSTM) algorithms. The dual channel 1D ResNet model with FDFS has great potential to SHM and intelligent maintenance in the rail transit system.</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

Dual channel 1D residual network for condition monitoring of high-speed train bogie with flexible dual functional sensors

  • 期刊论文

  • 发表日期：

    2025年9月2日

查看全文
  • Wentao Dong,
  • Zanhua He,
  • Kun Xiong,
  • Lin Yang

摘要

As the rail transit comes to the intelligent maintenance stage from the massive construction stage in China, the operation condition of the train bogie is more and more complicated due to the high speed and heavy load, and new sensing devices and data fusion algorithms are required for structural health monitoring (SHM) and fault detection of high-speed trains. A flexible dual functional sensor (FDFS) with piezoelectric and hot-film sensor units is designed for structural vibration and aerodynamic signal measurement of the train bogie. It is a conformal contact with the curved surface of bogie for improving the measurement accuracy of the airflow and vibration signal. A dual-channel one-dimensional Residual Network (1D ResNet) model based on the structural vibration and air flow signal is proposed for the operation condition recognition of the train bogie. Experiments have validated that the seven typical operation conditions of the bogie have been successfully recognized by the dual channel 1D ResNet model and FDFS with 92.3% average recognition success rate. The dual channel 1D ResNet has great advantage in the condition recognition of train bogie compared to dual channel 1D convolutional neural network (CNN), multilayer perception (MLP) and long short-term memory (LSTM) algorithms. The dual channel 1D ResNet model with FDFS has great potential to SHM and intelligent maintenance in the rail transit system.