摘要
轨道质量指数(TQI)检测数据通常为非等时距,采集的数据如缺乏修正和分类去噪处理会极大影响最终的预测精度,目前关于此类研究较少,尤其是去噪方法研究。结合TQI轨道质量指标,在对动静态检测方法进行比较的基础上,选取京广下行线开行轨检列车进行轨道不平顺动态检测。针对收集到的轨道病害检测数据,提出不良数据筛选原则及修正方法,引入卡尔曼(Kalman)滤波法对数据进行去噪,并在京广下行线验证去噪处理效果。结果表明:Kalman滤波效果较好,实时性强,能够有效去除高斯噪声及非高斯噪声,以Kalman滤波去噪为导向的线路精调后所得TQI指标优于未采用该方法的线路指标。最后,针对京广线检测到的轨道病害提出可行整修措施,为现场线路维护工作提供理论指导与依据。
The time interval of track quality index(TQI)data is usually non-equal.If there’s no correction and classification for collected data,the denoising will greatly affect the final prediction accuracy.There are few studies on this problem currently,especially on denoising methods.In combination with rail quality TQI,and on the basis of comparison of the dynamic and static test methods,the track inspection train is selected to be operated on Beijing-Guangzhou downlink for dynamic test.Selection principles and correction methods of bad data are proposed for the track disease detection data collected,and the Kalman filter method is introduced to denoise the data.The denoising effect is verified with experiments on Beijing-Guangzhou downlink.The experimental result shows the Kalman filtering has preferable effect,and better real-time performance,which can effectively remove Gaussian and non-Gaussian noise.Meanwhile,compared with the line without Kalman filtering denoising,the line has better TQI quality after maintenance.Finally,in allusion to the track diseases detected on Beijing-Guangzhou line,feasible treatment measures are put forward,which can provide theoretical guidance and basis for the track line maintenance.
作者
胡哨刚
孔祥芳
叶海波
肖志锋
冯博
旷利平
HU Shaogang;KONG Xiangfang;YE Haibo;XIAO Zhifeng;FENG Bo;KUANG Liping(Hunan Technical College of High-speed Railway,Hengyang 421000,China;China Railway Guangzhou Bureau Group Co.,Ltd.,Guangzhou 510000,China)
出处
《交通科技与经济》
2021年第6期58-63,共6页
Technology & Economy in Areas of Communications
基金
国家自然科学基金项目(71771218)
湖南省自然科学基金项目(2020JJ7025)。
关键词
轨道平顺
Kalman滤波法
TQI
铁路
轨道检查仪
track smoothness
Kalman filter method
track quality index(TQI)
railway
track checker