Prediction and Optimization of Residual Stresses on Machined Surface and Sub-Surface in MQL Turning

Residual stress in the machined surface and subsurface is affected by materials, machining conditions, and tool geometry and can affect the component life and service quality significantly. Empirical or numerical experiments are commonly used for determining residual stresses but these are very expensive. There has been an increase in the utilization of minimum quantity lubrication (MQL) in recent years in order to reduce the cost and tool/part handling efforts, while its effect on machined part residual stress, although important, has not been explored. This paper presents a hybrid neural network that is trained using Simulated Annealing (SA) and Levenberg-Marquardt Algorithm (LM) in order to predict the values of residual stresses in cutting and radial direction on the surface and within the work piece after the MQL face turning process. Once the ANN has successfully been trained, an optimization procedure, using Genetic Algorithm (GA), is applied in order to find the best cutting conditions in order to minimize the surface tensile residual stresses and maximize the compressive residual stresses within the work piece. The optimization results show that the usage of MQL decreases the surface tensile residual stresses and increases the compressive residual stresses within the work piece.