Numerical and Experimental Comparison of Flow Stress and the Effect of Microcylinder Topography in the Barreling Test: An Artificial Neural Network Model

In any case, macroscale shaping cannot be transferred to microscale due to the size effect. There are different parameters to evaluate small-scale forming. In this article, the effect of flow stress and topography in two states of lubrication and without lubrication for T2 copper micro cylinder has been investigated using the numerical solution and prediction of a trained neural network. An artificial neural network model has been presented to investigate the flow stress and the effect of friction with the miniaturization of the copper microcylinder. The results show that the flow stress decreases with decreasing the initial specimen diameter in both lubrication conditions, and the flow stress decreases by 30 MPa with the initial specimen diameter decreasing from 8 mm to 1 mm. The friction factor increases obviously with decreasing the initial specimen diameter in the case of lubricating with castor oil, and the friction factor increases by 0.11 with the initial specimen diameter decreasing from 8 mm to 1 mm. However, the tribology size effect is not found in the case without lubrication. The reasons for the flow stress and tribology size effects were also discussed. The good matching of outputs and objectives in the regression graphs shows that the response of the neural network is satisfactory.