Prediction of Material Spreading in Hot Open‐Die Forging

Prediction of spreading during the hot open-die forging process is a key issue in the optimal control of forging technology. Exact measurements of the cross-sectional shape and dimensional changes during forging in industry are very complicated and demand large financial investment, but on the other hand, their contribution to optimization during the forging phase is not essential. Hence the only reasonable choice is to make optimization during the technological planning stage, i.e. before the actual forging. The use of physical simulations and results from forging in the laboratory has confirmed that the ratio between the contact area length and work piece width is not the only influence on material spreading. Additionally, two geometrical parameters, the ratio between the width and the height of the cross-section and the relative deformation, as well as the temperature of the forged piece must be considered. It is important to treat each material in the model separately. For prediction of spreading two approaches were chosen, i.e. description in the form of a function in which the coefficients are specific for each grade of steel, and prediction by the application of the neuronal network approach (CAE NN - Conditional Average Estimator, Neural Networks). Each of the approaches has its own characteristics: prediction with CAE NN is more accurate than the description in the form of a function, but the latter is more suitable for modelling hot open-die forging technology.