COMPARISON OF ADAPTIVE NEURO FUZZY INFERENCE SYSTEM AND RESPONSE SURFACE METHOD IN PREDICTION OF HARD TURNING OUTPUT RESPONSES

Abstract

Hard turning has been used to replace cylindrical grinding to obtain high quality surface finish of complex parts with hardness above 45 HRC. Surface roughness is characterized among the most critical attributes in hard turning and it is important to the manufacturer to have accurate understanding of the machining process prior to its optimization process. The aim of this paper is to compare the capability of adaptive neuro-fuzzy inference system (ANFIS) model with response surface method (RSM) in developing the correlation of machining parameter and output responses. The input for both models are cutting speed (v), feedrate (f), and depth of cut (d), whereas the output responses are flank wear (V_b) and surface roughness (R_a, R_q and R_z).  The results indicate that the accuracy of predicted values of ANFIS model overwhelmed the predicted value of RSM model with up to 42% higher. At this level of accuracy, ANFIS model shows its applicability to be an objective function in evolutionary algorithm.