Multi-Objective Optimization and Predictive Modeling of Surface Roughness and Material Removal Rate in Turning Using Grey Relational and Regression Analysis

Abstract

This study applies Taguchi's design of experiment methodology and regression analysis for optimization of process parameters in turning AISI 1040 steel using coated carbide insert under dry environment. Experiments have been carried out based on L9 standard orthogonal array design with three process parameters namely cutting speed, feed and depth of cut for surface roughness and material removal rate. Based on the S/N analysis, the optimal process parameters for surface roughness are as follows: Cutting speed at level 3, feed at level 1 and depth of cut at level 3 i.e. v3-f1-d3 considering smaller-the-better approach. Similarly, the optimal process parameters for material removal rate are as follows: Cutting speed at level 3, feed at level 3 and depth of cut at level 3 i.e. v3-f3-d3 considering larger-better approach. Results of the main effect plot indicate that cutting speed is the most significant process parameter for surface roughness and material removal rate followed by feed. The depth of cut is found to be least affecting parameter for both the responses. The mathematical models have been developed for individual responses using regression analysis. Regression models proposed are statistically significant and adequate because of higher R2 value. The normal probability plot vs. residuals of model shows that the residuals lie reasonably close to a straight line implying that the terms mentioned in the model are significant. At the same time, the predicted value from the developed model and experimental value are very close to each other showing significance of models developed. For simultaneous optimization of responses, Grey relational analysis combined with Taguchi method has been proposed. It is observed that there is good agreement between the estimated value (0.779) and experimental value (0.821) and the improvement of grey relational grade from initial parameter combination (v2-f2-d2) to the optimal parameter combination (v3-f3-d2) is found to be 0.284. This indicates the improvement through the proposed methodology.