Dual Visual Tracking And Monitoring Of Large Structure’s Ripple Line Fillet Seams Based On Advanced And Current

Assembly methods of welding corrugated sheet and base plate together can be used toincrease the structure rigidity of Containers, carriages of Railway passenger and rail buses,and other large structures. Therefore, this assembly method has been widely used in a varietyof large box-type structures. When the corrugated sheet and the based plate were assembled,spot welding was used to fix the position of assembly gap to be welded (abbreviated to AGW),and then the AGW was welded continuously by metal active gas arc welding (abbreviated toMAG welding). However, under the influence of errors due to factors of processing, stitching,assembly between corrugated sheet and based plate and so on, error existed between thethree-dimensional trajectory and width of the theoretical AGW and the actual AGW.Therefore, three-dimensional trajectory of AGW (abbreviated TDT-AGW)and the width ofAGW (abbreviated W-AGW) need to be monitored and tracked for adaptively adjusting thetrajectory of welding torch and welding process parameters (welding current, welding voltage,welding speed, etc.) to ensure the welding quality. In this context, this research begun fromstructural features of Large Structure’s Ripple Line Fillet Seams (abbreviated LSRLFS) andrequirements of welding, to study the precise apparatus and method of visual monitoring andtracking of LSRLFS（abbreviated VSMT） in high-speed MAG welding. In this paper, basedon the object of Container’s Ripple Line Fillet Seams, and Level of the structure of LSRLFSand VSMT, Double VSMT system of LSRLFS based on advanced-current vision wasresearched: the front VSMT system was used to track and monitor the TDT_AGW andW_AGW for planning the trajectory of the weld torch and the welding parameters; based onthis, the direct VSMT system was used to track and monitor the weld pool and the proximateassembly gap to be welded (abbreviated WP_PAGW) for obtaining the offset between theweld torch and the centerline of the proximate assembly gap (abbreviated PAGW) caused byadvanced error, thermal distortion and et al, and the change of PAGW width, and then transferthe parameters to the motion controller and the Adaptively adjust for improving the accuracyof monitoring and tracking in high-speed MAG welding.The main contents of this research include:1. Dual VSMT system of LSRLFS based on advanced-current visionThere existed the problem of machining and assembly errors before welding, thermaldeformation and the magnetic blow in welding process existed in LSRLFS of Containers,carriages of Railway passenger and rail buses. To solve this problem, the precise apparatusand method suitable for the LSRLFS was proposed. The dual visual system based advanced VSMT and current VSMT were used: advanced VSMT was used to capture the image of laserline projection that was distance away from the weld torch, and then the TDT_AGW andW_AGW were obtained according to imaging processing and analyzing. It was used to guidethe trajectory of the torch and plane the welding parameters; Double target imaging systemwas used in the current VSMT to obtain the image of WP_PAGW. According to processingand analyze the image, the offset between the weld pool and the centerline of PAGW and thewidth were obtained to adjust the trajectory of torch and the welding parameters online. Theresult of Simulation shows that this device and method can work effectively in LSRLFSenvironment to track and monitor the trajectory and the width of weld seam, and guide thetorch to move along with the AGW and adopt adaptive welding parameters to weld.2. VSMT mechanism of TDT_AGW and W_AGWCombined with automated welding of ripple line fillet seams in containers, the advancedVSMT system’s structure design of LSRLFS of TDT_AGW and W_AGW based on the singleLLP assisted, parameter optimization and calibration of vision system, the model between theimage features and DT_AGW and W_AGW, the image processing and analysis algorithms forhigh speed MAG welding were discussed in detail. It focuses on the components of theadvanced VSMT system of LSRLFS of TDT_AGW and W_AGW., optimizing the positionparameter between laser, camera I and the weld pieces, the calibration of advanced VSMTsystem, the model of TDT_AGW and W_AGW, the translation between the image features ofLLP and the TDT_AGW及W_AGW of ripple line fillet seams; In addition, the phenomenonthat endpoints of LLP modulated by corrugated sheet and based plate at the AGW wereirregular, the influence of surface scratches, positioning joints on affecting extraction ofFP_LLP, were analyzed, and a method based on improved adaptive thresh-holding,morphological filtering, and connected component labeling was proposed to realize theFP_LLP extraction.3VSMT mechanism of TDT_AGW and W_AGW based adaptiveThe TDT_AGW and W_AGW of LSRLFS tracking and monitoring can be realized bythe advanced VSMT system in high-speed MAG welding, but the error caused by advancederror, thermal deformation and others cannot be eliminated, so it is difficult to meet the needsof high-precise monitoring. The current VSMT system was used to monitor the WP_PAGWand correct the inexact matching between the torch and the PGWA, and between the plannedweld parameter and the PGWA width for realizing the precise welding. The focus was on theimaging method WP_PAGW in MAG, an image forming apparatus of WP_PAGW based on DTIS was proposed. Based on this，filtering, dimming， auxiliary light were used to ensurethe image quality. The algorithm model of conversion between the Image feature ofWP_PAGW and the offset of the torch, the width PAGW based on the registration of DTIS,image feature fusion（abbreviated IFF）Also the offset between the torch and centerline ofPAGW and the PAGW width were obtained according to process the WP_PAGW image thatimaging characteristics was significantly and interference was complex.Through processing and analyzing the WP_PAGW image that has obvious features andcomplex interferences, the center point of welding pool in WP_PAGW image and the edges ofPAGW extracted. Then he offset between the torch and centerline of PAGW and the PAGWwidth were obtained with image fusion and mathematical model.The results show that: the VSMT system of TDT_AGW and W_AGW that is a certaindistance away from the weld pool, can acquire image of little interferences and obviousfeatures; through processing and analyzing image, and with corresponding mathematicalmodel, TDT_AGW and W_AGW could be acquired; The welding parameters adaptive systembased on direct monitoring with DTIS tracked the WP_PAGW and got the deviation betweenthe torch and the center line of PAGW and the PAGW width. a reference for correction ofwelding trajectories and parameters was given. the image fusion of dual VSMT system basedon advanced-current vision provides a reliable method for precise monitoring and tracking inthe high-speed MAG welding of LSRLFS.