A Comparative Analysis of Regression Coefficient, Mse and Mape for Abrasive Slotting Machine

Abstract

This study presents an experimental investigation on abrasive air jet machining on alumina reinforced with zirconium composite materials. The aim is to optimize the machining process parameters using a Taguchi and neural network techniques. The experiment was designed based on the L27 orthogonal array of the Taguchi method, considering four input parameters: pressure, abrasive flow rate, stand-off distance and type of materials. The response variables, material removal rate, and surface roughness were measured and analyzed. Regression analysis was performed to evaluate the goodness of fit (R2) between experimental and predicted values. Additionally, Mean Absolute Percentage Error (MAPE) and Mean Squared Error (MSE) were calculated to assess the accuracy of predictions using both Taguchi and neural network models. The combined approach of Taguchi and neural network techniques offers a robust method for optimizing abrasive air jet machining parameters for alumina-zirconium composites. The findings provide valuable insights for enhancing machining efficiency and surface quality in various industrial applications