Neurofuzzy based surface roughness modelling for ball end milling operation

Abstract

Now a day a manufacturing system is oriented towards higher production rate, quality, and reduced cost. Surface roughness is an index for determining the quality of machined products and is influenced by the cutting parameters. In die manufacturing industries surface roughness of dies are considered as a vital quality characteristic. For the complex shapes of a die, three dimensional machining is done by ball end mill in the most cases. In this study the average surface roughness (Ra) for a die material AISI 4340 namely EN24 and Hot Die Steel have been measured after ball end milling operation. Before conducting the experiments a design of experiment was done with Fractional Box-Behnken Design of Experiment. 49 experiments have been conducted varying Cutter axis inclination angle (φ degree), Tool diameter (d mm), Spindle speed (S rpm), Feed rate (fy mm/min), Radial depth of cut or Feed along X-axis (fx mm) and Axial depth of cut (t mm) in order to find Ra. These 49 data have been used for training purpose and more 25 data have been collected with random selection of input parameters and used as testing dataset. The training dataset has been used for train different ANFIS, ANN and RSM models for Ra prediction. And testing dataset has been used for validate the models. Better ANFIS architecture has been selected for minimum value of root mean square error (RMSE) and better of ANN architecture has been selected based on Root Mean Squared Error (RMSE) and Absolute Percentage Error (MAPE). The Selected ANFIS model has been compared with theoretical model, ANN model and RSM. This comparison was done based on RMSE and MAPE. The comparison shows that the selected ANFIS model gives better result for training and testing data for both the die materials, EN24 and Hot Die Steel. Proposed ANFIS model for EN24 composed of 2 two-sided Gaussian curve built-in (gauss2MF) membership functions for each of the six input functions and a linear output function. And ANFIS model proposed for Hot Die Steel composed of two Gaussian Membership Functions (gaussMF) for each Input and Linear Membership Function for Output. So, these ANFIS models can be used further for predicting surface roughness of a commercial die material (AISI 4340 and Hot Die Steel) after ball end milling operation. Correlation test shows that only cutter axis inclination angle and feed along X-axis (radial depth of cut) have positive correlations with surface roughness.