Application of RSM and ANN in Predicting Surface Roughness for Side Milling Process under Environmentally Friendly Cutting Fluid

Abstract

The paper presents a potential study on prediction of surface roughness in side milling by optimization techniques approaches. Two methods, response surface methodology (RSM) and artificial neural networks (ANN) were used for optimized prediction. The model of surface roughness was expressed as the main parameter in side milling term of cutting speed, feed rate and axial depth of cut. Rotatable central composite design (RCCD) is employed in developing second-order response surface mathematical model. The ANN model using a multi-layer feed forward, back propagation and training function Levenberg-Marquardt (LM) algorithm with a single hidden layer. Vegetable oils have often been recommended as sustainable alternative cutting fluid since the ecological and health impacts in the use of mineral oil have been questioned and also the rising cost of mineral oil. The advantages of oxidative stability of coconut oil as vegetable oil were utilized in this study to investigate surface roughness of low carbon steel. The machining of ferrous alloy like steel is sometimes a difficult task. This study used uncoated tool because it is suitable when turning and milling alloy. Flood condition was selected because it has been proved effective at low cutting speed. The analysis predicted by RSM and ANN models resulted a good agreement between the experimental and predicted values. The results indicated that the ANN model predict with more accurate compared with the RSM model.