Optimizing laser powder bed fusion parameters for enhanced hardness of Ti6Al4V alloys: A comparative analysis of metaheuristic algorithms for process parameter optimization

This study investigates the influence of Laser Powder Bed Fusion (LPBF) processing parameters of high-performance Ti6Al4V components on mechanical properties and microstructural uniformity. Experimental results indicate that lower laser power (200–203 W) and moderate scan speeds (600–604 mm/s) optimize the hardness and the Hatch distance (0.10–0.11 mm) and layer thickness (0.04–0.05 mm) significantly impact hardness, with specific parameter combinations yielding superior results. A comparative assessment of six metaheuristic algorithms, such as the JAYA algorithm, Cohort Intelligence (CI), Particle Swarm Optimization (PSO), Teaching-Learning-Based Optimization (TLBO), Genetic Algorithm (GA), and Simulated Annealing (SA), was performed for LPBF parameter optimization. JAYA and CI exhibited the fastest convergence, achieving peak fitness values within the first five iterations and demonstrating high stability with minimal oscillations. The results show a significant improvement in hardness, consistently ranging between 560 and 570 HV. JAYA and CI reached comparable maximal hardness measurements at 570 HV, which agreed with the laboratory-reported 428.6 HV value. The HYV values obtained by PSO and TLBO algorithms were competitive at about 569.5 HV yet displayed reduced accuracy in their results. GA and SA showed moderate to slow convergence speed, which led to unreliable precision in their application. JAYA and CI demonstrate better performance through their effective balancing capability of exploitation and exploration processes that enable rapid and stable convergence. Given its simplicity alongside its accuracy and robust performance, the JAYA algorithm proves the most appropriate method for LPBF parameter optimization. Real-time applications benefit from CI as a promising solution.