Considered multi-objective optimization process, electron beam welding (EBW) using the experimental data obtained during the execution of welding passes on samples of austenitic steel. The application of statistical approach to the analysis of the geometric characteristics of the zones of penetration provides a fairly accurate model of the process EBW, which describe the dependence of the geometry of the welds on stainless steel from the parameters of welding. Solved a number of specific tasks related to improving the quality of welds in the EBW data steels. Developed a graphical user interface allowing the selection and optimization of process parameters EBW to obtain welds with specified performance characteristics. The resulting models are applicable in the computer expert system, which can be developed to support operators in the selection of welding modes in accordance with customer requirements and technical standards. Such systems can be used in education and training of operators and in forecasting and optimization techniques of electron beam welding.

electron-beam welding, electron beam, beam focusing, beam splitting, modeling, statistical methods, regression models, multi-objective optimization, graphic optimization.