摘要
Underwater welding is developing fast because of the exploration of marine resources, and underwater wet welding automation is urgently needed because of the rigorous environment. To control the welding process automatically, the model of the process should first be built to predict the current welding process status. In this paper, arc and visual sensors were used simultaneously to obtain the electrical and visual information of underwater wet welding, and support vector machines （SVM） were used to model the process, experiment results showed that the method could effectively use the information obtained and give precise prediction results.
Underwater welding is developing fast because of the exploration of marine resources, and underwater wet welding automation is urgently needed because of the rigorous environment. To control the welding process automatically, the model of the process should first be built to predict the current welding process status. In this paper, arc and visual sensors were used simultaneously to obtain the electrical and visual information of underwater wet welding, and support vector machines （SVM） were used to model the process, experiment results showed that the method could effectively use the information obtained and give precise prediction results.
基金
This work was supported by the National Natural Science Foundation of China under the Grant （No. 51105103 ）, China Postdoctoral Science Foundation under the Grant （ No. 2012M510945, No. 2013T60362） , Project（ HIT. NSRIF. 2015115 ） supported by Natural Scientific Research Innovation Foundation in Harbin Institute of Technology.
