Investigation of Surface Roughness in Machine using Artificial Intelligence Techniques

Carbon fiber and carbon fiber composites have become more widely used in a multitude of sectors, including defensive line, military, as well as industries. Surface quality is also given careful consideration, as machineries rely on matching parts to work. The Carbon Fiber Reinforced Polymer (CFRP) turning process composites is explored in this research by adjusting three critical cutting variables: cutting speed, depth of cut and feed. Artificial Neural Networks (ANN) and Response Surface Methodology (RSM) have been used to develop an experimental model for assessing surface roughness. The three cutting variables were assessed as experimental design input criteria for these models. In the context of ANN methodology, the traditional backpropagation technique was found as the best option for training the model. Analysis of variance which is referred as ANOVA was used to determine the consequence of cutting parameters on roughness of respective surface. $R^{2}$, RMSE and MEP were computed as 99.9%, 0.016 and 2.17 respectively based on RSM modelling results