Enhancement in manufacturing systems using Grey-Fuzzy and LK-SVM approach
T. P. Latchoumi, G. Kalusuraman, J. Faritha Banu, T. L. Yookesh, T. P. Ezhilarasi, K. Balamurugan
Abstract
Magnesium with 10% of SiC is prepared through the stir casting process. An integrated Taguchi-Fuzzy model for L27 orthogonal array for Airjet machining parameters is developed. Optimum machining condition predicted by Grey Fuzzy Relational Grade (GFRG) is similar to Grey Relational Grade (GRA) and the optimum machining conditions for low Erosion Rate (ER) and Vickers Hardness Number (VHN) are Impact Angle (IA) = 30o, Erodent Velocity (EV) = 90 m/sec and Discharge Rate (DR) = 4 gms/min. For the experimental observations, the regression equation developed from GFRG has a good correlation. Further, an attempt has been to reduce the mathematical complexity (offline mode) by using the advantages of the Linear Kernel Support Vector Machine (LKSVM) method and Agglomerative Hierarchical Clustering-based (AHC) algorithm, L27 Orthogonal Array (OA) observations are classified into 3 classes. Regression equations developed for ER and VHN particularly for class having more observations (Class 2) alone has found to have excellent correlation with experimental observations.