Predicting Surface Roughness of AISI 4140 Steel in Hard Turning Process through Artificial Neural Network, Fuzzy Logic and Regression Models

Abstract

In this study, the average surface roughness values obtained when turning AISI 4140 grade tempered steel with a hardness of 51 HRC, were modeled using fuzzy logic, artificial neural networks (ANN) and multi-regression equations. Input variables consisted of cutting speed (V), feed rate (f) and depth of cut (a) while output variable was surface roughness (Ra). Fuzzy logic and ANN models were developed using Matlab Toolbox. Variance analysis was conducted using MINITAB. The predicted values of mean squared errors (MSE) were employed to compare the three models. Results showed that the optimum predictive model is the fuzzy logic model. With small errors (e.g MSE = 0.0173166), the model was considered sufficiently accurate.   Key words: Hard turning, surface roughness, artificial neural network, fuzzy logic model, multi regression model.