Research And Application Of Multi-pass Welding Of Medium And Heavy Plates Based On Laser Vision Sensor

Welding of medium and heavy plates is common in many fields such as aerospace,shipbuilding,large industrial vessel manufacturing,bridge steel plates and automobile beam steel plates and so on.Among them,multi-pass welding technology is one of the key technologies to realize medium and thick plate welding.Aiming at the inconsistent geometric characteristics of weld grooves caused by accuracy errors in the processing of large quantities of workpieces in multi-pass welding,the slow process parameters of adapting to different weld beads,and the offline programming method of traditional manual teaching cannot be adapted the path planning problem of multi-pass welding with different weld groove characteristics.This paper deeply researches the overall scheme design of multi-pass welding system and welding seam feature extraction,SPSS-based welding bead prediction ternary linear regression model,multi-pass welding path planning and multi-pass welding program design.(1)Multi-pass welding system overall scheme design and welding seam feature extraction.Obtain the weld groove geometry information through the laser vision sensor.First,the laser vision sensor is built with a one-line laser and a CMOS industrial camera,plus a reasonable design of the sensor housing,and then processed the collected weld images with one-line laser stripes to extract the feature points of the weld groove,and calculate the geometric information of the weld groove.Finally,the accuracy of the feature points and geometric information acquisition algorithm of the weld groove is verified through experiments,and the automatic detection of the geometric information of the weld groove is realized,and the model parameters are provided for the subsequent planning of the weld bead of the medium and thick plate.(2)A ternary linear regression model for weld bead prediction based on SPSS.First,with the base material of Q235 steel plate as the experimental carrier,a single-pass cladding test of mixed gas shielded welding with 80%CO₂+20%N₂was designed.Then use the data collected in the experiment to establish a prediction model between the welding process parameters and weld bead forming parameter,including the ternary linearity regression prediction model between the cross-sectional area,melting width and excess height and welding current,welding speed and swing amplitude.Finally,the independent variable significance test,regression equation fit test,and regression equation linear significance test were performed on the model.At the same time,the correctness of the model was verified by the comparison between the predicted data and the real measurement data,which solved the problem of rapid adaptation to different welding,and it provides a theoretical basis for the selection strategy of welding process parameters in the subsequent multi-pass welding path planning.(3)Multi-pass welding path planning.First,through the weld groove geometry information and the weld bead forming parameter prediction model,the custom weld bead filling strategy is adopted to adaptively plan the number of weld beads and the welding current,speed,torch angle,swing size and torch of each bead.position.Then perform finite element simulation on different weld bead schemes to obtain the best weld bead sequence.In addition,a compensation strategy considering welding defects is proposed to improve arc craters and undercut defects.Finally,by considering the two three-layer six-pass welding physical test experiments before and after welding defects,the rationality of the proposed multi-pass welding path planning strategy is verified,and the problem of low efficiency in multi-pass welding path planning is solved.(4)Multi-pass welding program design and application.Mainly completed the design of the relevant operation interface,and realized the function control of the laser vision sensor,including white balance control,exposure gain control,ROI setting,image optimization,frame rate control,image acquisition,image feature point extraction and multi-pass welding planning Features.At the same time,the use of multi-threaded programming technology greatly improves the real-time performance of image processing and display,and improves the convenience of multi-pass welding path planning.