Research On Dynamic Measurement Method Of Rail Wear Based On Laser Displacement Sensors

With the rapid development of the railway network in China,rail wear nondestructive dynamic measurement technology based on machine vision,has replaced the traditional static contacting measuring methods,become an important means of improving the railway maintenance efficiency.During the dynamic measuring process,however,the poor and complicated outdoor environment,multidimensional random vibration of car body,and congenital defect of the traditional methods,have brought serious challenges for the accurate measurement of rail wear.How to realize the fast and accurate detection under a series of disturbances has been widely concerned by universities and railway research institutes at home and abroad.In this paper,by constructing the experiment platforms utilizing the 2D and 1D laser displacement sensors,we carried overall thorough research on the several practical problems which restrict the high-precision dynamic measurement of rail wear.The main contributions are listed as follows:(1)By conducting a lot of experiments and analyzing the experimental results,we found that the regular methods faced up with a series of problems in actual test,and understood the influences on the accurate measurement of rail wear.Thus laid a solid foundation for the subsequent research.(2)In actual test,the rail profile is faced up with two serious challenges: one is that the shining area located on the rail surface and the unrelated area located on the roadbed and fastening cause a lot of outliers on the profilogram,and the other is that the rail diversity may bring the invalid profi le.To solve the problems,an effective method for the outlier and profile validity detection based on profile coarse registration is proposed in this paper.First,the profile is fragmented according to the point spacing,and most sparse outliers are removed fast based on the density feature.Then,according to the principle of proximity,the fragmets belong to one area are reassembled again,and the underlying railhead and rail waist area are located correctly based on the profile shape feature.Third,the railhead is employed for realizing the coarse registration between the underlying areas and the standard profile,and the rail waist similarity is calculated to serve as the index of profile validity.Finally,if the profile is valid,its original version is matched with the standard one again,and all of the outliers are detected and removed effectively in this step based on the relative Euclidean distance.(3)During the process of dynamic test,the vehicle vibration may make the light plane be nonperpendicular to the rail longitudinal axis.Thus causes the profile affine distortion.To solve the problem,two different methods are proposed here.The first method found the difference between normal profiles and distorted ones are mainly focused on triple primitives including rail jaw,railhead inner line,and rail waist curve,and used a novel curve similarity measurement method,i.e.,point set mapping(PSM),to evaluate the difference.Then,by optimizing the profile matching error with the modified PSO(MPSO),the rail profile under affine distortion is calibrated precisely.In essence,this method belongs to the category of regional matching.By contrast,the second method hits the mark by seeking the corresponding feature points.By constructing the convex hull on the rail waist curves of the reference profile and the measured one,respectively,we obtain a lot of affine invariant feature vectors,and realize the fast extraction of feature points and the solution of affine parameters.Meanwhile,considering the manufacturing error,this method employes the measuring normal profile to serve as the reference profile for registration,which makes the accuracy of distortion calibration is improved further.(4)Traditional chord-based models are faced up with two severe challenges in actual applications: one is that none of traditional single chord models cannot cover the required broadband from 30 mm to 60 m individually,and the other is that the common frequency sampling method(FSM)is unavailable to design the FIR inverse filters for asymmetrical chord and multi-points chord models.To solve the two problems,a practical combination-chord model(CCM)is proposed in this paper.For the first problem,the model is constructed on the existing three-point asymmetrical chord platform.Through the flexible combination of three sensors,CCM integrates the advantages of long wave measurement from two-point chord model and shortwave measurement from three-point asymmetrical chord model,and realizes the corrugation broadband measurement in one platform successfully.For the second problem,the least square(LS)is employed to design the inverse filter.Compared with FSM,the design process is more unlimited and easier for the user.More important,even with the same output filter order,the filter performance designed by LS is better.(5)During the process of dynamic measurement of rail corrugation,with the existence of the curve rail and the snaking motion of car body,the positions of laser measuring points on the rail surface cannot be kept in the effective range the whole time.This induces the error of corrugation measurement results.To solve the problem,an auxiliary system for position monitoring and correction of chord-based measuring points is constructed in this paper.The proposed system comprises the position monitoring subsystem(PMS)based on machine vision and the position correction subsystem(PCS)based on stepping motor.The PMS monitors the positions of measuring points with given sampling interval,and calcula tes the position deviations with respect to the railhead centerline.Once the deviation exceeds the preset range,the PCS makes corresponding operations to pull the measuring points back to the effective range.Indoor static test and simulated dynamic test show that the system can monitor and correct the position deviations of chord-based measuring points effectively,and make the rectified results approximate the true state of rail corrugation better.