基于钢轨焊缝提取的轨道不平顺数据里程误差修正

Trackage error correction of track irregularity based on railway welds extraction

为修正轨检仪历史数据里程误差以保证数据具有可对比性,基于轨道结构特征——呈规律性分布的钢轨焊缝,采用轨检仪轨面高低数据提取焊缝特征:通过动态标准差和滑动平均预处理原始高低数据以加强焊缝特征,建立3次样条插值寻峰模型以提取焊缝特征位置,以真实焊缝里程进行轨检仪历史数据里程误差修正。试验结果表明:在不额外增加设备的情况下,经过里程修正后的历史数据之间误差绝对值积分降低50%,相关系数能够达到0.90。研究成果可为后期轨道养护和预测不平顺提供技术支持。

The premise of comparative analysis is to correct the mileage error of the historical data of the track detector to ensure the data be comparable.Based on the characteristics of the track structure—the rail welds with regular distribution,this paper adopts the vertical profile of the rail surface to extract weld features of the rail inspection instrument:firstly,the original high and low data were preprocessed by dynamic standard deviation and smooth function to strengthen the weld characteristics,then established the finding peak model of cubic spline interpolation to extract the weld feature position,finally the mileage error correction of the track data of the track tester was performed with the actual weld mileage.The test confirms that the absolute value of the error between the historical data after mileage correction is reduced by 50%without any additional equipment,and the correlation coefficient could reach 0.90.The research results provide technical support for the later track maintenance and prediction irregularity.