Predictive Modelling and Optimisation of Surface Roughness in Turning of AISI 1050 Steel Using Polynomial Regression

Abstract

An investigation has been conducted to address the surface integrity optimisation and prediction issue by applying the polynomial regression method for a variety of experiments and cutting conditions. A higher correlation coefficient (R2) was obtained with a cubic regression model, which had a value of 0.9480 for Ra. The use of the response surface optimisation and composite desirability show that the optimal set of machining parameters values are (250m/min, 0.2398 mm/rev and 2.3383 mm) for cutting speed, feed and depth of cut, respectively. The optimised surface roughness parameter and productivity are Ra =2.7567 µm and Q = 95.341*103 mm3/min, respectively. Results show that the models developed can accurately predict the roughness on the basis of measured cutting conditions as input parameters, and can also be used to control the surface roughness by making a comparison between measured and estimated values. Furthermore, operators can benefit from the proposed models if the aim is the reverse determination of the cutting conditions corresponding to the requested roughness profile.