Prediction and Optimization of Surface Roughness in a Turning Process Using the ANFIS-QPSO Method

Mahdi S Alajmi; Abdullah M Almeshal

doi:10.3390/ma13132986

Prediction and Optimization of Surface Roughness in a Turning Process Using the ANFIS-QPSO Method

Materials (Basel). 2020 Jul 4;13(13):2986. doi: 10.3390/ma13132986.

Authors

Mahdi S Alajmi¹, Abdullah M Almeshal²

Affiliations

¹ Department of Manufacturing Engineering Technology, College of Technological Studies, The Public Authority for Applied Education and Training, Safat 13092, Kuwait.
² Department of Electronic Engineering Technology, College of Technological Studies, The Public Authority for Applied Education and Training, Safat 13092, Kuwait.

Abstract

This study presents a prediction method of surface roughness values for dry and cryogenic turning of AISI 304 stainless steel using the ANFIS-QPSO machine learning approach. ANFIS-QPSO combines the strengths of artificial neural networks, fuzzy systems and evolutionary optimization in terms of accuracy, robustness and fast convergence towards global optima. Simulations revealed that ANFIS-QPSO results in accurate prediction of surface roughness with RMSE = 4.86%, MAPE = 4.95% and R² = 0.984 for the dry turning process. Similarly, for the cryogenic turning process, ANFIS-QPSO resulted in surface roughness predictions with RMSE = 5.08%, MAPE = 5.15% and R² = 0.988 that are of high agreement with the measured values. Performance comparisons between ANFIS-QPSO, ANFIS, ANFIS-GA and ANFIS-PSO suggest that ANFIS-QPSO is an effective method that can ensure a high prediction accuracy of surface roughness values for dry and cryogenic turning processes.

Keywords: ANFIS-QPSO; ANN; adaptive neuro-fuzzy inference system; machine learning; quantum particle swarm optimization; surface roughness; turning process.