摘要
针对城市轨道交通车辆牵引电机存在的滤网堵塞问题,为降低售后服务人员故障排查的难度并提高排查的即时性,文章提出一种基于循环神经网络的电机滤网堵塞预警方法。其根据电机原理并结合数据驱动方法构建模型,利用实际运营数据对模型进行训练,基于列车现有数据对电机温度进行实时预测,根据预测值与实际值的残差关系实现电机滤网堵塞预警。将该方法用于某地铁运营车辆,结果表明,其可以准确地对电机滤网堵塞问题进行预警,指导精准维修,从而有效地促进状态修进程。
Aiming at the problem of filter clogging in traction motor of urban rail transit vehicles,in order to reduce the difficulty and improve the instantaneity of after-sales service personnel in troubleshooting,a warning method of motor filter clogging based on recurrent neural network is proposed.The model is constructed based on the combination of motor principle and data-driven method,and actual operation data is used to train the model.By this method,motor temperature is predicted in real time based on the existing data of train,and motor filter clogging warning is realized according to the residual relationship between predicted values and actual values.Application results of a subway operation vehicle show that this method can accurately warn the clogging problem of motor filter,guide precise maintenance,and effectively promote the process of state maintenance.
作者
张士强
戴计生
徐海龙
詹彦豪
ZHANG Shiqiang;DAI Jisheng;XU Hailong;ZHAN Yanhao(CRRC Zhuzhou Institute Co.,Ltd.,Zhuzhou,Hunan 412001,China)
出处
《控制与信息技术》
2021年第5期97-101,共5页
CONTROL AND INFORMATION TECHNOLOGY
关键词
故障预警
电机滤网堵塞
循环神经网络
散热模型
智能运维
fault warning
motor filter clogging
recurrent neural network(RNN)
heat dissipation model
intelligent operation and maintenance
