Seam tracking based on Kalman filtering of micro-gap weld using magneto-optical image. Issue 1 (March 2016)

• Record Type: Journal Article
• Authors: Gao, Xiangdong; Mo, Ling; Xiao, Zhenlin; Chen, Xiaohui; Katayama, Seiji
• Is Part Of: International journal of advanced manufacturing technology. Volume 83:Issue 1/4(2016)
• Page Start: 21
• Page End: 32
• Publication Date: 2016-03
• Subjects: Micro-gap weld -- Kalman filter -- Seam tracking -- Magneto-optical image; Manufacturing processes -- Periodicals; Production engineering -- Periodicals; 670.427
• DOI: 10.1007/s00170-015-7560-x
• Languages: English
• ISSNs: 0268-3768

Abstract

Abstract Seam tracking ability of a welding system is significant for welding process and obtaining good welds. It is necessary to realize weld seam detection and tracking for welding automation. Kalman filter (KF) is applied to get the optimal state estimation of micro-gap (whose width is less than 0.2 mm) butt joint weld position. A magneto-optical sensor was used to obtain the weld information. The weld position was detected by the maximum entropy segmentation method, and the weld position parameter from a magneto-optical image was extracted as a state eigenvector, which included the weld position at previous sampling time and the variation of weld position. The state equation based on the weld position parameter and the measurement equation for the weld position are established. Considering that the system dynamic noises were white noises, a traditional Kalman filtering algorithm was developed with white noises, and the optimal state estimation of the weld position was obtained. The influence of noise statistical uncertainty characteristics on Kalman filtering was analyzed. Experimental results show that the Kalman filter algorithm can effectively restrain the noise jamming, and the effect of Kalman filter is affected by noise statistical characteristics directly.