Enhancing laser surface texturing with driving training-based optimization: A metaheuristic approach

Abstract

This paper investigates the capability of Laser Surface Textruing (LST) to induce texture on Ti-6Al-4V, aiming on optimizing process parameters viz. average power, pulse frequency, scanning speed, and gas pressure using the Driving Training-based Optimization (DTBO) algorithm. Both single and multi-objective optimizations are conducted to determine optimal parametric settings. The study systematically examines the impression of these LBM process parameters on various responses. Comparative analyses was performed with five other metaheuristic algorithms such as Ant colony optimization, Particle swarm optimization, Differential evolution, Firefly algorithm, Teaching-learning-based optimization, and Artificial bee colony. Furthermore, statistical validation via paired t-tests confirms the unique effectiveness of the DTBO algorithm. Detailed examination through developed box plots and convergence diagrams consistently demonstrates DTBO superior performance in terms of accuracy, minimal variability in optimal solutions, and reduced computational effort. The DTBO achieves a higher MRR by 35.7 %, 20 %, 11.9 %, 54.7 %, and 33.3 % compared to ABC, ACO, FA, DE, and TLBO, respectively. Simultaneously, DTBO also achieves a lower ATW by 13.6 %, 14.8 %, 3.02 %, 15.9 %, and 16.1 % compared to the same algorithms. These results underscore DTBO's superior performance in achieving improved MRR values and reduced ATW values across the considered optimization algorithms. Hence, The DTBO algorithm demonstrates robustness and applicability in optimizing LBM processes in context of laser texturing, which may enhance manufacturing efficiency and product quality significantly.

Keywords

• DTBO algorithm
• Laser surface texturing
• Metaheuristic algorithms
• Optimization