Quality characteristics optimization in CNC end milling of A36 K02600 using Taguchi’s approach coupled with artificial neural network and genetic algorithm

Abstract

This study investigates surface roughness and energy consumption in a CNC end milling of plain low carbon steel (mild steel) A36 K02600 using a carbide end mill cutter. Taguchi’s L9 orthogonal array was adopted for designing the experimental runs considering several process parameters viz. cutting velocity, Feed rate, spindle speed and cutting depth in order to study their influence on the quality characteristics. Optimal control parameter combinations for minimizing surface roughness and energy consumption were evaluated from signal to noise ratio. Analysis of variance revealed the contribution of control factors on the quality characteristics. Numerical predictive models using linear regression and artificial neural network were developed to envisage the responses accurately. Multi-objective Genetic Algorithm optimization was exploited in order to obtain a specific set of control parameter which would optimize both the responses simultaneously. This study concludes that spindle speed (68.24% contribution) and feed rate (92% contribution) are the most responsible variables for surface quality and energy consumption respectively. The outcome of artificial neural network model and genetic algorithm confirm that both quality characteristics can be optimized simultaneously and Taguchi’s robust design approach is a successful tactic for optimizing machining parameters to achieve desired surface quality at low energy consumption. Improvement in surface quality and reduction in energy consumption were found to be 27.79% and 30% respectively. Low carbon steel is extensively accepted by the industries for its wide variety of which make this study physically viable.