Study on the On-line Support System for Welder

Abstract

In recent years, the shortage of experienced welder has become a problem. Therefore, we are developing online welder support system so that we can demonstrate the performance of experienced welder by providing in-formation to non-skilled welder during welding. This system can be applied to the actual work of welding as well as training only. In this system, the pre-diction and optimization for welding condition is provided to the welder as acoustic, visual and haptic information during actual work by analyzing the measurement results of welder’s behavior and welding condition with molten pool shape, etc. For data analysis, the prediction of penetration shape and optimization of torch movement can be also carried out using numerical simulation by mol-ten pool model. In this paper, in order to develop the on-line support system for welder, a basic investigation was carried out. As a result of the welder gaze measurement test using the smart glass with the proper filtering, it was confirmed that the eye tracking of the skilled welder was more stable than the non-skilled welder. Also, in the TIG manual welding, based on the measurement results of the torch movement by the camera for skilled welder and beginner, the situation of the backbead by numerical simulation was compared. The simulation results agree with the trend of actual welding. It was shown that the application of numerical simulation based molten pool model to simulation manual welding is effective.