Research On Forming Of Multi-layer Single-pass Weld Of Medium And Heavy Plate Based On Laser Vision Sensing

In recent years,China has made great progress in industrial fields such as construction machinery,space shuttles,and rail transit.Therefore,the demand for medium and heavy plate processing parts is also increasing,and multi-layer welding is a commonly used processing method for medium and heavy plate processing parts.In today’s industrial manufacturing,manual welding and teaching and reproducible robot automatic welding are the common welding methods for medium and heavy plate welded parts.In traditional manual welding,the work intensity is high and the working environment is harsh;the teaching robot cannot recognize the errors generated in the manufacturing and installation process,which leads to the deterioration of the welding seam quality.Therefore,it is one of the prerequisites to ensure the automatic welding quality of medium and heavy plates to correct the welding torch trajectory according to the change of the welding seam position and intelligently adjust the welding process parameters according to the change of the space size of the welding bead.This paper takes the unilateral V-shaped groove multi-layer single-pass welding of the ladle flange as an example.The medium and thick plate welding parts are manually cut by acetylene flame,etc.,and the phenomenon of variable groove angle will occur.Therefore,a set of multi-layer single-pass welding seam forming system for medium and heavy plates based on laser vision sensing is researched to respond to the real-time welding deviation and incomplete penetration problems of layer single-pass welding.It is of great significance to conduct on-line detection of information,adjust welding heat input in a timely manner,and evenly control the amount of deposition.The main research contents are as follows:A set of laser vision-based multi-layer single-pass welding seam tracking system for medium and heavy plates is built.The system can detect the welding seam deposition size and metal deposition amount in real time during the V-shaped welding seam deposition process of medium and heavy plates.The upper computer calculates and compares the detection results and the set values,makes independent decisions on the welding deposition process,and adjusts the welding deposition parameters in real time.An algorithm for the detection and correction of welding seam forming size based on laser vision point cloud data is proposed,which realizes the detection and correction of the welding seam forming size and the adjustment of the inclination angle of the welding torch.Welding test was carried out by orthogonal experiment method,and the collection and analysis of sample data were completed.Based on experience,experimental conditions and experimental requirements,the welding current,welding voltage,welding speed,and wire dry elongation were selected as the main test influencing factors,and the experimental indicators were residual height,deposition amount,and fusion width.The orthogonal table is designed,and the welding test is carried out and the experimental sample data is collected as a guide.The selected welding process parameters are analyzed and calculated by the method of variance analysis,and their respective influences on the welding forming are obtained.A welding seam forming controller based on fuzzy neural network is designed.The neural network algorithm and fuzzy control algorithm in traditional PID controller and modern intelligent control method are introduced,and on the basis of combining the advantages of the two algorithms,a fuzzy neural network-based multi-layer single-pass welding seam forming controller for medium and heavy plates is proposed..The deposition experiments of multi-layer single-pass welding of V-shaped welds of medium and heavy plates with constant width and variable width were carried out.Comparing the open-loop control and closed-loop control of the multi-layer single-pass welding seam forming size and deposition amount of the medium and heavy plate,it can be seen that the open-loop control welding forming part weld height and metal deposition amount errors are accumulated layer by layer,which cannot meet the medium thickness.Plate welding requirements.The closed-loop control of the welded forming parts,the height deviation of the welding seam is less than 0.4mm,the welding deposition amount is moderate,and the welding seam forming is beautiful,which verifies the effectiveness of the welding seam forming controller based on the fuzzy neural network.