Surface roughness prediction using Taguchi-fuzzy logic-neural network analysis for CNT nanofluids based grinding process

The present study highlights the Taguchi design of experiment techniques proved to be an efficient tool for the design of neural networks' surface roughness to predict in the grinding process, where CNT mixed nanofluids are used as dielectric for machining AISI D3 Tool steel material. Empirical model for the prediction of output parameters has been developed using regression analysis and the results are compared for with and without using CNT nanofluids. Analysis of variance and F test is used to determine the significant parameter affecting the surface roughness which is the crucial parameter for any grinding process. Feedforward artificial neural networks are used to train the experimental values with the Levenberg-Marquardt algorithm; the most influencing factors are determined. The predicted surface roughness for without using CNT based cutting fluid is 11.3 % and with CNT is 10.37 %. Further, a fuzzy logic system is used to investigate the relationship between the machining process parameters' accuracy and to determining the efficiency of each parameter design with Taguchi design of experiments.