ORIGINAL CONTRIBUTION An Accelerated Particle Swarm Optimization Algorithm on Parametric Optimization of WEDM of Die-Steel V. Muthukumar • A. Suresh Babu • R. Venkatasamy • N. Senthil Kumar Received: 26 March 2014 / Accepted: 25 July 2014 / Published online: 5 December 2014 Ó The Institution of Engineers (India) 2014 Abstract This study employed Accelerated Particle Swarm Optimization (APSO) algorithm to optimize the machining parameters that lead to a maximum Material Removal Rate (MRR), minimum surface roughness and minimum kerf width values for Wire Electrical Discharge Machining (WEDM) of AISI D3 die-steel. Four machining parameters that are optimized using APSO algorithm include Pulse on-time, Pulse off-time, Gap voltage, Wire feed. The machining parameters are evaluated by Taguchi’s L 9 Orthogonal Array (OA). Experiments are conducted on a CNC WEDM and output responses such as material removal rate, surface roughness and kerf width are determined. The empirical relationship between control factors and output responses are established by using linear regression models using Minitab software. Finally, APSO algorithm, a nature inspired metaheuristic technique, is used to optimize the WEDM machining parameters for higher material removal rate and lower kerf width with surface roughness as con- straint. The confirmation experiments carried out with the optimum conditions show that the proposed algorithm was found to be potential in finding numerous optimal input machining parameters which can fulfill wide requirements of a process engineer working in WEDM industry. Keywords Meta-heuristic APSO  Taguchi’s L 9 OA  Regression models  Surface roughness  Kerf width Introduction Wire electrical discharge machining process is a non-tra- ditional machining alternative in which material is removed by generating a series of discrete sparks between electrode and work piece immersed in a liquid dielectric medium. WEDM is widely used in various fields like mould-making tool and die industries, medical, aerospace applications, automotive and electronics industries to make intricate shapes and profiles in extremely hard materials like tita- nium, zirconium, molybdenum, tool and die steel etc. WEDM is a highly non-linear and complex machining process controlled by a large number of process parameters such as pulse-on time, pulse-off time, pulse current, gap volt- age, wire speed, wire feed, wire tension. Any slight deviation in the process parameters can affect the performance character- istics which include material removal rate, dimensional accu- racy, surface roughness and cutting speed, kerf width [1]. Several authors have attempted to improve the WEDM performance characteristics [2–6], but the effective utiliza- tion of this process is not completely solved because of its complex nature and many number of variables involved in the process. Traditionally, selection of appropriate machin- ing conditions was carried out by relying heavily on the operator’s experience or data provided by the WEDM manufacturers. Therefore, the selection of appropriate machining parameters for machining different new materials is still a challenging task. Some researchers have [7] inves- tigated the effects of various process parameters of WEDM process, such as pulse on-time (Ton), pulse off-time (Toff), gap voltage (SV), peak current (IP), Wire Feed (WF) and V. Muthukumar (&)  R. Venkatasamy Saveetha Engineering College, Chennai, Tamil Nadu 602105, India e-mail: profmuthukumarmech@gmail.com A. Suresh Babu Department of Manufacturing Engineering, Anna University, Chennai, Tamil Nadu 600025, India N. Senthil Kumar Adhiparasakthi Engineering College, Melamaruvathur, Tamil Nadu 603319, India 123 J. Inst. Eng. India Ser. C (January–March 2015) 96(1):49–56 DOI 10.1007/s40032-014-0143-z