Prediction of workpiece surface roughness using soft computing

A study is presented to model surface roughness in end-milling using soft computing (SC) or computational intelligence (CI) techniques. The techniques include the artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS). ANFIS combines the learning capability of ANN and the effective handling of imprecise information in fuzzy logic. Prediction models based on multivariate regression analysis (MRA) are also presented for comparison. The machining parameters, namely, the spindle speed, feed rate, and depth of cut, were used as inputs to model the workpiece surface roughness. The model parameters were tuned using the training data maximizing the modelling accuracy. The trained models were tested using the set of validation data. The effects of different machining parameters, number, and type of model parameters on the prediction accuracy were studied. The procedure is illustrated using the experimental data of end-milling 6061 aluminium alloy. Although statistically all three models predicted roughness with satisfactory goodness of fit, the test performance of ANFIS was better than ANN and MRA. In comparison with MRA, the performance of ANN was better in training but similar in test. The results show the effectiveness of CI techniques in modelling surface roughness.