摘要
转辙机是重要的铁路信号基础设备,在保障铁路列车运行安全中起到了重要的作用。基于铁路信号集中监测系统采集的道岔动作电流及功率曲线数据,利用CNN(卷积神经网络)对数据故障特征提取及LSTM(长短时记忆网络)对时间序列数据处理的优势,建立基于CNN-LSTM的故障诊断模型。实验仿真表明,基于CNN-LSTM的故障诊断模型对转辙机的各类故障具有较好的识别效果,准确率达93.32%,能有效地识别道岔故障类型。
The switch machine is the important railway signal basic equipment,which plays an important role in ensuring the safety of railway trains.The study collects turnout action current and power curve data based on railway signal centralized monitoring system,uses CNN to extract the characteristics of data fault features,uses LSTM to play the advantages of time series data processing,and constructs fault diagnosis model based on CNN-LSTM.The experiment simulation shows that the fault diagnosis model based on CNN-LSTM has a good recognition effect on all kinds of faults of switch machine.The accuracy rate reaches 93.32%.It can effectively identify the types of switch faults.
作者
张胜
韩承桓
李孟娇
李德威
孙彤
Zhang Sheng;Han Chenghuan;Li Mengjiao;Li Dewei;Sun Tong(Cangzhou Jiaotong College,Cangzhou 061100,China)
出处
《黑龙江科学》
2023年第8期113-115,118,共4页
Heilongjiang Science
基金
沧州市重点研发计划指导项目(204102012)
沧州交通学院校级科研项目(CJ202301002)
作者简介
张胜(1983-),男,硕士研究生。研究方向:控制理论与技术;通信作者:韩承桓(1995-),男,硕士研究生。研究方向:控制理论与技术。
