Study On The Test System Of Rail Welding Head Flatness

In recent years,the high speed railway obtained the swift and violent development in our country.For safe operation of high-speed,rail track put forward higher requirements of production.The straightness of the welding line which comes from the production progress and the abrasion comes from the using are directly related to the safety and stability during the running of train and the comfort of passengers,so the detection of the flatness of the rail welding head is of great importance during the production and daily inspection and maintenance for the long rail.After fully research of the current situation of the development of rail testing equipment at home and abroad,For the straightness on the surface of rail,this issue designs a kind of machine vision based detection system which is used to detect the rail surface parameters combined with computer and laser.Put forward the improvement in the detection accuracy and speed.There are two LMI Gocator 2340 laser sensor moving along the rail with the help of the servo system based on Ether CAT bus leads,and the three dimensional coordinate information on the detected surface of the rail is obtained.In view of the phenomenon that the vibration of mechanical structure can cause interference on the data collected by sensor,build a mathematic model of the system and use the input shaping algorithm based on finite time Laplace transform to suppress the vibration of mechanical structures.After the coordinate is transformed and the noise is denoised,Analyze the corresponding relationship of two point cloud data sets,and combine with the characteristics of the cluster point cloud data,the improved iterative closest point algorithm is adopted to carry out splicing and reconsitution for the original data,by which we can get full of rail head contour.According to the test standard of rail welding head,the rail coordinate data of three test points are extracted and the flatness test is carried out.For the wrong number,use a special color to line out them from the right ones and indicate its relative position to the initial point of the rail.In the course of this research,the solution to different problems mainly involves movement control,the collection and processing of point cloud data,which is as follows:(1)In view of the mechanical vibration when the sensors are moving,design and research the input shaping suppression algorithm which based on the finite time of Laplace transform to suppress the vibration of mechanical structures.(2)For the point cloud data collection from the surface of the rail,research the algorithm to process the point cloud data such as coordinate transformation and bilateral filtering algorithm,an improved iterative closest point algorithm based on KDtree.(3)Choose the suitable PLC,servo motor,laser outline sensors and other hardware equipment,build the detection equipment,design the servo motion plan and control procedures.Design PC program,build a multithreaded architecture,implement the collection,analysis and display of the data from the two laser outliner sensors.This system has realized the dynamic testing for the parameters of track surface and the straightness of welding line,and it can execute the task of analysis,data processing and the 3D display of the parameters of track surface at the same time precision is 0.034 mm and the detection frequency can reach 700Hz.