Evaluation Of Surface Irregularities Of High-speed Railway Tracks Based On Fractal Theory And Machine Learning

The continuous operation of the railway will cause a certain degree of wear and tear of the track,which may lead to very serious operational safety problems,so the evaluation of the track condition is of great significance.In order to dynamically evaluate the wear status of the track,this paper introduces the fractal dimension as an index for evaluating the wear status of the railway track based on the fractal theory,and uses machine learning technology to predict the fractal dimension of each section of the track.Compared with the track quality index,the fractal dimension can identify the finer structure of the track surface and obtain a more accurate track wear condition.The actual measured data of the two routes from Leshan to Chengdu East and Beijing South to Shanghai are analyzed by example.The research content of the paper is as follows:(1)In order to calculate the fractal dimension of the track irregularity,three different calculation methods are used to calculate it.Comparing the obtained results with the actual survey situation,it is found that the box dimension method is the best method to calculate the fractal dimension of track irregularity.(2)In order to obtain the main parameters affecting the wear state of the track,the fractal dimension of the track geometric parameters is reduced by principal component analysis,and the main parameters are the high and low fractal dimension and the orbital fractal dimension.(3)In order to predict the fractal dimension of track irregularity,three supervised learning algorithms—neural network,linear regression and RSM regression are used,and four performance indicators are used to evaluate the prediction results of the three algorithms.Comparing the predicted value with the real value of fractal dimension,it is found that the prediction effect of the RSM regression algorithm is the best,and its prediction accuracy is approximately 85%.Based on the above research,the variation of the fractal dimension of the track irregularity in each section of the above two lines with time is obtained.Then,the sections of the track where abnormal conditions may occur in the future are obtained,and suggestions are provided for the inspection and maintenance of railway-related labor departments.