Optimizing Ti6Al4V Alloy Wire Edm Parameters using Regression Analysis and Metaheuristic Algorithm

Abstract

This research is aiming to identify optimal WEDM parameters for Ti6Al4V alloy, which is a Ti-based alloy widely used in the aviation field, by using a combination of regression analysis and meta heuristic algorithms. The main task is to make the process of the WEDM more effective, which would mean that materials removing rate (MRR) and surface roughness (SR) would also be increased for Ti6Al4V alloy. The experimental setup procedure includes applying Charmilles Wire-EDM technique on Ti6Al4V material with brass wire electrode submerged in de-ionized water. To obtain the results there were whole series of experiments were conducted and the datasets was submitted to the regression methods. It finally turned out that the Random Forest regression model turned out to be the best one, with the largest R2 values for both MRR (0.8555) and SR (0.8887), thus showcasing its distinctive capabilities by the means of capturing the intricate cause-effect patterns within manufacturing process. Specifically, the data is processed by multiple meta-optimization algorithms including D-ORCA, PSO, Firefly, Cuckoo, MOTLBO, GWO and Salp Swarm which evolve at the same time for optimization. Among the three regression models, D-ORCA presented with the highest combined score to indicate its strongest suit in terms of finding the most suitable parameters for the WEDM process. Nevertheless, PSO and Firefly algorithms delivered a combination of optimized MRR and SR. Through this wider research aim is to enrich the existing understanding and practicability of WEDM technology for Ti6Al4V alloys and provide a new light to the optimization of output parameters to achieve a companion of optimized MRR and SR. By solving identified challenges from the available literature, this study comes in handy, not only in improving efficiency.