Investigation of MRR in Face Turning Unidirectional GFRP Composites by Using Multiple Regression Methodology and an Artificial Neural Network

Abstract

In this manuscript, an experimental study is prepared to optimize the cutting parameters for the material removal rate in face turning of UD-GFRP composite under a dry environment using a polycrystalline diamond tool of 0.8 mm tool nose radius. An ANN and MRM models are developed and compared to predict material removal rate of the face-turned part surface. Graphical user interface of MATLAB is adopted to use a multi-layer feed-forward artificial neural network by the investigational values as input output pairs. The effectiveness of ANN with various transfer functions and number of neurons in intermediate layers is checked and an optimum transfer function and number of hidden neurons are selected. During the tests, speed, feed, DOC, and approach angle are varied. The trial arrangement was based on Taguchi's L27 OA. From the results, the most favorable collection of cutting parameters for a large MRR is a cutting speed 550 rpm, feed rate 0.1 mm/rev, DOC 1.0 mm, and approach angle of 75°. The feed is found to be major variable affecting MRR, followed by other variables. It is clear from the results that the predicted MRR matches with the experimental data and the correlation coefficient is found to be more than 0.9. The mean absolute percentage inaccuracy in a neural system model is −13.68%, while the same for a regression model is found to be −17.90%.