GMAW welding procedure expert system based on machine learning

In order to simplify the robot preparation before welding and improve the automation of the whole welding process, an intelligent expert system for Gas Metal Arc Welding is designed in this paper. In the system, the user inputs the initial welding information and the output interface displays suitable welding procedure parameter schemes. The user can choose the schemes according to the actual requirements or directly generate the welding procedure specification required by the enterprise format for direct use. In addition, the system also combines the database technology and XGBoost algorithm in the field of machine learning, migrates the model trained on the data set to predict the welding raw data, accumulates more data for daily use to optimize the model, which makes the whole system more systematic and intelligent, and achieves the goal of more accurate use.