Robotic Deep Penetration K-TIG Welding System Based on Weld Penetration Detection

doi:10.3901/JME.2019.17.014

Abstract

Abstract: A deep penetration K-TIG welding system based on penetration detection is developed to realize the automatic adjustment of the contact tip-to-work distance(CTWD) and the on-line detection of the penetration state. Based on the corresponding relationship between the arc voltage and the CTWD, the arc voltage signal collected by the arc voltage sensor in real time is converted into the welding height value. Then, the difference between the actual height value and the pre-set height value is fed back to the robot control system through the OLE for process control(OPC) communication system, so that the robot can realize the automatic adjustment of the welding height. Besides, a penetration detection system is constructed. A charge coupled device(CCD) camera is used to collect the images of the molten pool and key-hole, whose characteristics are extracted. Then the correlation between the pool-hole characteristics and the penetration state is established. Based on this, a back propagation(BP) neural network for identifying the penetration state is established to realize the indirect detection of the penetration state under different welding heights, and finally the optimal welding height is obtained according to the detection result. After testing, the whole system is proved to be reliable and stable, and can accurately detect the penetration state of the weld, realizing the automatic adjustment of the welding height which improves the welding quality.

Key words: robotic welding, K-TIG welding, welding height control, online welding penetration detection

CLC Number:

TG156

ZHANG Baori, GU Shengyong, SHI Yonghua. Robotic Deep Penetration K-TIG Welding System Based on Weld Penetration Detection[J]. Journal of Mechanical Engineering, 2019, 55(17): 14-21.

References

[1] FENG S,TERASAKI H,KOMIZO Y,et al. Development of evaluation technique of GMAW welding quality based on statistical analysis[J]. Chinese Journal of Mechanical Engineering,2014,27(6):1257-1263.
[2] 马可,薛龙,黄军芬,等. GMAW自动焊熔透影响因素分析及多元回归预测[J]. 机械工程学报,2018,54(18):55-61. MA Ke,XUE Long,HUANG Junfen,et al. Analysis of influencing factors and multivariate regression prediction of GMAW automatic welding penetration[J]. Journal of Mechanical Engineering,2018,54(18):55-61.
[3] 宋凯,曾琼,何智成,等. 基于超声参数化和熵模型的汽车焊点质量识别[J]. 机械工程学报,2016,52(16):86-92. SONG Kai,ZENG Qiong,HE Zhicheng,et al. Quality identification of automotive solder joints based on ultrasonic parameterization and entropy model[J]. Journal of Mechanical Engineering,2016,52(16):86-92.
[4] ZHANG H,HOU Y,ZHAO J,et al. Automatic welding quality classification for the spot welding based on the Hopfield associative memory neural network and Chernoff face description of the electrode displacement signal features[J]. Mechanical Systems & Signal Processing,2017,85:1035-1043.
[5] ZHANG Wan,JIA Minping,ZHU Lin,et al. Comprehensive overview on computational intelligence techniques for machinery condition monitoring and fault diagnosis[J]. Chinese Journal of Mechanical Engineering,2017,30(4):782-795.
[6] BUFFA G,CAMPANELLA D,PELLEGRINO S,et al. Weld quality prediction in linear friction welding of AA6082-T6 through an integrated numerical tool[J]. Journal of Materials Processing Technology,2016,231:389-396.
[7] 张勇,周昀芸,王博,等. 基于声发射信号的铝合金点焊裂纹神经网络监测[J]. 机械工程学报,2016,52(16):1-7. ZHANG Yong,ZHOU Yunyun,WANG Bo,et al. Neural network monitoring of aluminum alloy spot welding crack based on acoustic emission signal[J]. Journal of Mechanical Engineering,2016,52(16):1-7.
[8] 张朴,彭齐治,孔力. 基于虚拟仪器技术的激光焊接质量实时监测系统设计[J]. 焊接学报,2005,26(11):24-26. ZHANG Pu,PENG Qizhi,KONG Li. Design of laser welding quality real-time monitoring system based on virtual instrument technology[J]. Transactions of the China Welding Institution,2005,26(11):24-26.
[9] 韩喆. 基于LabVIEW的焊接质量检测系统的研究[D].武汉:武汉理工大学,2007 HAN Zhe. Research on welding quality inspection system based on LabVIEW[D] Wuhan:Wuhan University of Technology,2007.
[10] WU Chuansong. Proactive control effect of arc and weld pool behaviors by an external magnetic field in high speed GMAW[J]. Journal of Mechanical Engineering,2016,52(2):1.
[11] WANG Z,ZHANG D. Progressive switching median filter for the removal of impulse noise from highly corrupted images[J]. IEEE Trans. Cricuits & Syst.ii,1999,46(1):78-80.
[12] 蔡敏. 铝合金GTAW熔池区视觉特征检测及焊缝成型控制[D]. 上海:上海交通大学,2013. CAI Min. Visual characteristics detection and weld forming control of aluminum alloy GTAW pool area[J]. Shanghai:Shanghai Jiao Tong University,2013.
[13] CANNY J. A computational approach to edge detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,1986,8(6):679-698.
[14] 李克海,何德孚,王国苹,等. 脉冲TIG焊熔池几何参数的计算机视觉检测[J]. 焊管,2000(6):29-32,36. LI Kehai,HE Defu,WANG Guoping,et al. Computer vision detection of geometric parameters of pulsed TIG welding pool[J]. Welded Pipe,2000(6):29-32,36.
[15] XIE Rong,WANG Xinmin,LI Yan,et al. Research and application on improved BP neural network algorithm[C]//Industrial Electronics and Applications (ICIEA),2010 the 5th IEEE Conference on. IEEE,2010:1462-1466
[16] KOHAVI R. A study of cross-validation and bootstrap for accuracy estimation and model selection[C]//International Joint Conference on Artificial Intelligence. 1995:1137-1143.
[17] KLINE D M,BERARDI V L. Revisiting squared-error and cross-entropy functions for training neural network classifiers[J]. Neural Computing & Applications,2005,14(4):310-318.