Open AccessModeling and Analysis of Machining Parameters and Responses of Wirecut Electric Discharge Machining of Al2124/SiCp using Response Surface Methodology and Soft Computing Techniques
Author(s) -
B Sridhar Reddy,
A. B. Koteswara Rao,
G Ranga Janardhana
Publication year - 2019
Publication title -
international journal of recent technology and engineering
Language(s) - English
Resource type - Journals
ISSN - 2277-3878
DOI - 10.35940/ijrte.b3556.078219
Subject(s) - electrical discharge machining , response surface methodology , machinability , particle swarm optimization , machining , surface roughness , design of experiments , voltage , central composite design , servo , materials science , multi objective optimization , control theory (sociology) , computer science , mechanical engineering , algorithm , mathematical optimization , mathematics , engineering , statistics , composite material , machine learning , artificial intelligence , electrical engineering , control (management)
In this work, Wirecut Electric Discharge Machining (WEDM) of Al 2124/ SiCp metal matrix composite material is studied to evaluate the influence of input parameters on response characteristics namely, kerf, Material Removal Rate (MMR), Surface Roughness (SR), Recast Layer Thickness (RCT), and Surface Crack Density (SCD). Central composite design, a technique from design of experiments is used to conduct 31 experiments. The input parameters selected for estimation of machinability are pulse on time (Ton), pulse off time (Toff), current (IP), and Servo Voltage (SV). Analysis of variance (ANOVA) is carried out to know the effect of influence parameters on responses. The regression models are developed in Response Surface Methodology (RSM)and are used in soft computing techniques as input equations for optimizing the single and multi-response optimization of response parameters. Desirability approach is used in single and multi-objective optimization of response parameters. Single objective optimization is carried out by RSM, the Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and Firefly Algorithms (FA). Confirmation experiments are conducted on the adequacy of the mathematical models developed in RSM and it shows good agreement between experimental and predicted values. The variation of predicted responses from different optimization techniques for single objective optimization is found to be less than 1%. From the results it is also observed that for single objective optimization all evolutionary algorithms are found to be suitable for WEDM