Research On Roller Repair Welding Bench Based On Machine Vision

Since the reform and opening up,major breakthroughs have been made in the steel industry around the world.With the continuous progress of steel rolling methods and manufacturing,steel production is still completed by rolling methods.As one of the most important components in the steel production process,the roll roll directly determines the quality of the steel forming.However,due to its own material strength and working conditions,the roll is prone to cracks,spalling,fractures and other failures.Reuse of failed rolls after repair is in line with current sustainable manufacturing practices.At the same time,during the investigation,it was found that the current roll repair equipment is not intelligent enough,and repairing a roll often requires high labor and high cost.This paper develops a roll repair and welding experimental platform based on machine vision technology,combining visual technology with traditional mechanical equipment to make mechanical equipment have perception space and realize intelligent welding ability.First of all,the current domestic and foreign roll repair technology and machine vision technology were investigated.Combined with the field working conditions,the appropriate repair means were selected,proposed the main realization functions of the experimental platform,and designed each subsystem according to the functional requirements of the experimental platform,so as to provide theoretical support for the subsequent work.The main components were selected and designed in combination with the design parameters,and the three-dimensional model of the components was drawn in the Inventor software,so as to complete the overall construction of the roll repair welding test bench;ANSYS software was used in combination with the previous design and calculation results to respectively analyze the roll repair welding test bench.Static calculation and modal analysis were carried out for the main parts of V-shaped hydraulic cylinder rotating support slider,coupling,scissor hydraulic lifting device,and the accuracy of the experimental bench was demonstrated according to the conclusion.Based on the construction of the roll repair and welding test bench,combined with VS2017,MATLAB and the open source visual algorithm library Open CV4.5.4,the machine vision inspection system and image preprocessing technology were described.The machine vision inspection system mainly includes the selection and design of the CCD camera,the theoretical basis of the vision system,and the calibration of the binocular camera based on Zhang Zhengyou’s calibration method.The theoretical basis of the vision system involves the transformation relationship of the camera imaging coordinate system,the principle of camera imaging,the causes of imaging distortion,and the visual calculation of the binocular camera,which are the basis of the machine vision inspection system.The camera calibration can solve the internal and external parameters of the left and right cameras,providing a technical basis for image preprocessing.The image preprocessing technology includes image stitching,histogram equalization,smooth noise reduction,edge detection and morphological closing operation.Processing the images collected by the camera can obtain more accurate roll surface defect contours.Finally,the image was corrected by combining the obtained camera parameters to realize the correspondence between the left and right camera image features.The feature points were selected in the corrected image,and the spatial coordinates of the feature points were obtained,so as to meet the requirement of contour positioning and realize the intelligent welding repair of the waste roll.The roll repair and welding test bench designed and constructed in this paper is currently being constructed and commissioned in Benxi Iron and Steel Plant,which provides a certain theoretical reference for future roll repair methods.The paper has 92 figures,10 tables,and 85 references.