Artificial neural network-based optimization of operating parameters for minimum quantity lubrication-assisted burnishing process in terms of surface characteristics

Roller burnishing is an alternative approach to enhance surface properties under plastic deformation and most investigations focused on optimizing process parameters. However, the impacts of operating parameters of the minimum quantity lubrication (MQL) system on the surface properties have not been considered. The roller burnishing process is widely applied to fabricate high-pressure bushings and crankshafts with superior quality, while the MQL system is extensively employed to facilitate different machining operations for saving lubricant usage, decreasing negative impacts on the environment, and protecting worker’s health. The purpose of this investigation is to select the optimal MQL system factors, including the nozzle diameter (D), impingement angle (I), flow rate (Q), and air pressure (P) for decreasing the maximum profile peak height of the roughness (MAR) and improving Vickers hardness (VH) for the roller burnishing process. The optimal artificial neural network (ANN) model was proposed to render the relations between the optimizing inputs and burnishing responses. An efficiently evolutionary technique entitled multi-objective glowworm swarm optimization (MOGSO) was utilized to produce a set of feasible solutions. The VIKOR method was employed to determine the best optimal solution. The results revealed that the 4–10–2 architecture of the developed ANN models efficiently described the burnishing performances and precisely predicted the response values. The optimal outcomes of the D, I, Q, and P were 1.5 mm, 45 deg., 130 ml/h, and 0.6 MPa, while the improvements in the MAR and VH were 17.0% and 14.0%, respectively, as compared to the common values used. The proposed approach comprising the ANN, MOGSO, and VIKOR could be considered as a powerful technique to deal with the complicated optimizing issue for the roller burnishing operation. The obtained finding could be expected as a significant contribution to enhancement in the machining quality for the roller burnishing process under the MQL condition. © 2021, The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature.