基于RFID标签的铁路车辆轴温检测预警方法的研究  

Research on railway vehicle axle temperature detection and early warning method based on RFID tag

作　　者：秦春林 赖远桥 李天宝 Qin Chunlin;Lai Yuanqiao;Li Tianbao(China Energy Railway Equipment Company Limited,Baotou Branch, Inner Mongolia Baotou, 014060, China;Shenzhen Invengo Information Technology Co., Ltd., Guangdong Shenzhen, 518052, China)

机构地区：[1]国家能源集团铁路装备有限责任公司包头车辆维修分公司,内蒙古包头014060 [2]深圳市远望谷信息技术股份有限公司,广东深圳518052

出　　处：《机械设计与制造工程》2021年第11期71-75,共5页Machine Design and Manufacturing Engineering

摘　　要：轴温是识别列车故障的关键指标之一,实时监测轴温变化并对故障及时预警对车辆检修、机车行车安全具有重要意义。在温度采集方面,采用射频识别(RFID)标签实现无源温度监测,分别设置数字信号处理芯片以及射频芯片控制信号的收发,实时、准确采集货运机车的轴温。在预警算法方面,采用回归系数算法进行车轴故障预警,结合采集的轴温变化数据,分析影响温度的成分因子,并经过特征降维处理后与温差实现多元线性回归拟合运算,获得回归系数后,使用改进的支持向量机模型进行故障预警。实验结果显示,该预警算法可以结合机车的运行速度、工况环境等,实现对故障的综合判断并预警,预警准确率可达97.67%。Axle temperature is one of the key factors to identify train faults.Real time monitoring of axle temperature changes and timely warning of faults have an important impact on vehicle maintenance and locomotive production and operation.In terms of temperature acquisition,RFID tag is used to realize passive temperature monitoring,and digital signal processing chip and RF chip are set respectively to control the sending and receiving of signals,so as to collect the axle temperature of freight locomotive in real time and accurately.In terms of early warning algorithm,the regression coefficient algorithm is used for axle fault early warning.Combined with the collected axle temperature change data,the component factors affecting the temperature are analyzed.After feature dimension reduction processing,the multiple linear regression fitting operation is realized with temperature difference.After obtaining the regression coefficient,the improved support vector machine model is used for fault early warning.The experimental results show that the early warning algorithm can realize the comprehensive judgment and early warning of faults in combination with the running speed and working condition environment of the locomotive,and the early warning accuracy can reach 97.67%.

关 键 词：货运机车 射频识别技术 特征降维 回归系数 轴温监测 故障预警 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]