Electron beam welding (EBW) of thin-walled aerospace components demands precise control during beam entry and exit to avoid defects such as porosity, undercut, and excessive heat-affected zones. The paper presents a unified control framework that integrates transient thermal modeling with multi-objective optimization to automate beam ramp-up and ramp-down profiles. During entry, three candidate current ramps—linear, quadratic-increasing, and quadratic-decreasing—are evaluated via a moving-source heat equation solution, identifying the minimal time \(t_{\textrm{ust}}^{\textrm{opt}}\) required to reach melt temperature at the weld zone. For exit, a Pareto-based genetic algorithm (GA) optimizes piecewise-constant beam-current segments by balancing two competing objectives: minimizing weld-zone temperature deviation from melting point ( \(C_{\textrm{depth}}\) ) and limiting overheating outside the seam ( \(C_{\textrm{outside}}\) ). The resulting control laws are implemented on a commercial EBW system with real-time feedback and dynamic beam deflection. Experimental validation on 2 mm-thick titanium-alloy specimens demonstrates a 20–25% reduction in entry/exit defects, a 58–65% decrease in repeated beam-pass length, and penetration and seam-width deviations under 5%—meeting aerospace quality standards. These findings represent a substantial step toward more reliable and efficient EBW procedures for critical applications, aligning with the demanding standards of aerospace component fabrication.
