基于循环神经网络的动车组温度数据预测研究
Research on Temperature Data Prediction of EMU Based on Cyclic Neural Network
杨永 1王瑞锋2
作者信息
- 1. 蒙冀铁路有限责任公司 内蒙古自治区 呼和浩特 010050
- 2. 中国铁路呼和浩特局集团公司 包头车辆段,内蒙古自治区 包头 014000
- 折叠
摘要
采用循环神经网络建立了基于CRH5A型动车组温度类数据的预测模型,对影响预测结果的影响因子、模型层数及神经元个数进行了明确的界定,对CRH5A型动车组实车开展持续性追踪分析,采集动车组运行真实数据,进行积累和培养.在利用神经网络预测模型对数据进行训练后,CRH5A型动车组变压器温度峰值预测模型精度可达94.2％,牵引电机温度峰值预测模型精度可达93.8％,齿轮箱温度峰值预测模型精度可达95.3％,轴箱温度峰值预测模型精度可达92.7％.动车组温度数据预测结果的精确度可满足实际应用需求,预测模型在提高列车检修效率、节支降耗方面有着重要的作用.
Abstract
The increasing modernization of China high-speed railway provides a data base and application plat-form for the application of information technology and big data technology on multiple units.A prediction mod-el based on CRH5A EMU temperature data is established by using a cyclic neural network.The influence fac-tors,number of layers and number of neurons that affect the prediction result are clearly defined.Continuous tracking analysis is carried out on CRH5A EMU real vehicle,and real data of CRH5A EMU operation is col-lected to accumulate and cultivate.After training the data with the neural network prediction model,the accu-racy of the temperature data prediction results can meet the practical application requirements.The precision of the CRH5A type EMU transformer temperature peak prediction model can reach 94.2％,the precision of the CRH5A type EMU traction motor temperature peak prediction model can reach 93.8％,the precision of the CRH5A type EMU gearbox temperature peak prediction model can reach 95.3％,and the precision of the CRH5A type EMU axle box temperature peak prediction model can reach 92.7％,which plays an important role in improving the efficiency of train maintenance and reducing costs.
关键词
循环神经网络/动车组/温度数据/预测模型Key words
cyclic neural network/EMU/temperature data/prediction model引用本文复制引用
基金项目
中国国家铁路集团有限公司重大课题(K2021J009)
出版年
2024
NETL
NSTL
万方数据