Enhancing Accuracy of Surface Roughness Model Using Box-Cox Transformation in Surface Grinding AISI 5120 Alloy Steels

The aim of this paper is to present a research study on enhancing the model accuracy of surface roughness when surface grinding. The used material is this study is AISI 5120 alloy steel and the grinding wheels are made of CBN. The experimental matrix is prepared according to response surface method in terms of Central composite design. Four cutting parameters, i.e. workpiece speed, feed rate, depth of cut, and dressing depth, are selected to investigate their influences on the surface roughness. Moreover, the interactions between the cutting parameters are also considered. Regression models for predicting surface roughness are suggested. Box – Cox transformation is applied to transform the non-normal distribution data to the new ones exhibiting normal distribution shapes. The results show that the new model (using Box – Cox transformation) predicts surface roughness better than the old model (without using Box – Cox transformation) does.