Application of Fuzzy Rule-Based Model to Predict TiAlN Coatings Roughness

Abstract

In this work, an approach for predicting the roughness of Titanium Aluminum Nitride (TiAlN) coatings using fuzzy ruled-based model was discussed. TiAlN coatings were produced using magnetron sputtering process. Tungsten carbide (WC) was selected as the substrate and titanium alloy was used as the material to coat the cutting tool. The sputtering power, substrate bias voltage and substrate temperature were selected as the input variables while roughness of the TiAlN coatings was considered as the response variable. A statistical design of experiments method known as centre cubic design (CCD) was selected to collect the data for developing the fuzzy rules. The prediction performances of the fuzzy rule-based model with respect to percentage error, mean squared error (MSE), co-efficient determination (R2) and model accuracy were compared with the response surface regression model (RSM). The result shown that the fuzzy rule-based model has much better predicting capability compared to the RSM.