Investigation of Surface Roughness during Machining of Al-Composite on EDM using Regression Analysis and Genetic Algorithm

Abstract

Al-composites have several applications in lightweight structures. But due to the high hardness particulates reinforced in the Al matrix, it isn't easy to machine it via conventional machining. Therefore, a non-conventional machining method was used to process the aluminium matrix composite (AMC). Out of all the available methods, electric discharge machining (EDM) was considered one of the viable options for processing Al-composite. Thus, EDM was used to machine the AMC at different machining parameters and with different tool electrodes. Brass and copper electrodes of 12 mm diameters were used. The other machining parameters were pulse on-time (Pon), pulse off-time (Poff) and voltage (V). A Taguchi-based L18 mixed array was used for the planning of experiments. The mean roughness value (Ra) was measured for the corresponding experimental setting. The regression coefficients were computed using regression analysis and finally empirical model has been developed. The developed empirical model was solved using a genetic algorithm. The optimization was completed using an integrated approach of Regression and Genetic algorithms. The optimized setting suggested by Regression-GA is tool: brass, Pon: 30µs, Poff: 90 µs and V: 8V and the corresponding value of Ra is 2.99 µm. The confirmation experiments were conducted on the suggested optimized setting and a close agreement has been found between the predicted value (predicted by integrated approach of regression analysis and genetic algorithm) and experimental value.