锅炉悬吊管焊接机器人初始焊位识别及电弧跟踪研究  

Research on the Robotic Welding of Boiler Suspension Tube Based on the Recognizing of the Initial Welding Position and Arc Tracking

作　　者：张巧琳[1] 李杰[2] 叶震[1] 陈华斌[1] 沙万华[2] 陈善本[1] 

机构地区：[1]上海交通大学材料科学与工程学院焊接所,上海200240 [2]上海锅炉厂有限公司,上海200245

出　　处：《锅炉技术》2012年第6期52-56,共5页Boiler Technology

摘　　要：针对超超临界锅炉悬吊管由于装配误差、焊接热变形等导致实际焊缝偏离示教轨迹的问题,搭建了一套基于视觉和电弧传感的自动焊接系统。进行了摄像机标定和手眼标定,识别初始焊位,采用"一目双位"的方式对目标点进行三维重建,实现初始焊位的导引。焊接过程中实时采集电流信号,进行滤波、傅里叶变换等信号处理,获得焊缝偏差量,发送给机器人控制器,实现焊缝跟踪。研究结果表明:初始焊位导引精度可达1mm,焊缝跟踪精度为1.3mm,焊缝成形良好。In the welding process of the ultra-supercritical boiler suspension tube,the assembly error and thermal deformation would lead to the actual weld deviating from teaching track.In order to solve this problem,an automatic welding system based on visual sensing and arc sensing had been developed.Camera and hand-eye calibration should be the first step and then the initial welding position was recognized.The ‘one eye-double position’ mode was applied to complete the stereo reconstruction of object point so as to achieve the guidance. During the process of welding,welding current was acquired and processed.In the aspect of signal processing,the Butterworth filter and Fourier transform was used to get the deviation which was sent to the robot controller in order to realize the real-time seam tracking.Finally,a welding experiment of boiler suspension tube was done.The result proved that a good formation of seam could be acquired while the precision of guidance and seam tracking was less than 1mm and 1.3mm respectively.

关 键 词：视觉传感 图像匹配 立体视觉 电弧跟踪 

分 类 号：TG409[金属学及工艺—焊接]