摘要
无缝线路的轨道不平顺严重影响车辆的运行平稳性和乘客的乘坐舒适性,甚至威胁行车安全。针对上述问题,文章基于不同传感器采集的多维特征量,以深度神经网络强大的特征提取功能为基础,建立了钢轨焊接接头不平顺故障诊断模型,同时进行了大量样本训练及实际线路测试。结果表明,该方法诊断准确率达到83.3%,可以满足实时性和准确性的要求,为焊接接头不平顺故障诊断提供了新的解决办法。
The track irregularity of jointless line seriously affects the vehicle stability and comfort,and even threatens the running safety.To solve the above problems,based on the multi-dimensional feature data collected by different sensors,and based on the strong feature extraction function of deep neural network,a model which can diagnose the irregularity of welded joints on the rail is established,and simultaneous interpreting with a large number of samples and field tests is carried out.The diagnostic accuracy of this method is about 83.3%,which meets the requirements of real-time and accuracy,and provides a new solution for the fault diagnosis of welding joint irregularity.
作者
李强
刘铁生
崔霆锐
黄玉
张志亮
LI Qiang;LIU Tiesheng;CUI Tingrui;HUANG Yu;ZHANG Zhiliang(Beijing Metro Operation Co.,Ltd.,Beijing 100044,China;Beijing Tangzhi Science&Technology Development Co.,Ltd.,Beijing 100097,China)
出处
《铁道车辆》
2022年第6期69-72,共4页
Rolling Stock
基金
国家重点研发计划项目(2020YFB1600700)
北京市地铁运营有限公司科研项目(202000050100000501)。
关键词
焊接接头
不平顺
故障诊断
深度神经网络
welding joint
irregularity
fault diagnosis
deep neural network