Font Size: a A A

Research On Narrow Gap Seam Tracking System Of K-TIG Welding

Posted on:2020-02-07Degree:MasterType:Thesis
Country:ChinaCandidate:W C LiuFull Text:PDF
GTID:2381330590460835Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
Welding technology,which is of importance in manufacturing,is widely applied to various manufacture field.However,in the traditional welding area,the welding torch is actually manipulated by welder or by robot that is initially taught and playback,leading to the low level of the automation of welding and the shortage of welding efficiency.Therefore,it is meaningful to accomplish welding automation.So as to satisfy this purpose,this paper develop a seam tracking system for K-TIG welding.By means of taking frames of welding area from high dynamic camera,computing the welding deviation according to the frames and finally drive the robot move along with the weld seam,the system finish the seam tracking task.In the system this paper developed,we do the research as followed:In order to detect the weld pool edge and weld seam edge of the weldment with interferences including surface texture and noise,this paper present an algorithm with the composition of wavelet transform,k-means and random forest,to detect the edges in frames of welding area.On this basis,this paper present a method to compute the welding deviation.Firstly,apply eclipse fitting method to the weld pool edge detected,and see the center of the eclipse as point on the weldment which weld torch center project to.Secondly,magnify the fitting eclipse and find the intersecting point of magnified eclipse and weld seam edge as the next moving target.With the space relation ship of projected point and moving target,the welding deviation can be computedTo satisfy the real-time demand of weld seam tracking system,this paper parallelize the computing of wavelet transform,K-means and random forest with the help of CUDA programming,according the characteristic of commended algorithm.After the algorithm parallelization,the operation time on GPU is 62.34 ms,is faster 123.79 times compared to operating on CPUFinally,this paper carry out weld seam track precision experiment on straight line weld seam weldment,poly line weld seam weldment and curve line weld seam weldment to verify the tracking precision and robustness of the system.As the experiment result show,tracking deviation is in the range of ±0.12 mm on straight line weld seam weldment,±1.65 mm on poly line weld seam weldment,and ±1.76 mm on curve line weld seam weldment...
Keywords/Search Tags:seam tracking, edge detection, CUDA programming, K-TIG
PDF Full Text Request
Related items