Neural network modeling for the prediction of texture evolution of hot deformed aluminum alloys

Commercial aluminum rolling mills operate under very restricted thermomechanical conditions determined from experience and plant trials. In this paper we report results for four-stand tandem mill rolling simulations within and beyond the thermomechanical conditions typical of a rolling mill by plane strain compression (PSC) testing to assess the effect of deformed conditions on the texture of the hot deformed aluminum strip after annealing. A neural network modeling study was then initiated to find a predictive relationship between the observed texture and the thermomechanical parameters of strain, strain rate, and temperature. The model suggested that temperature is the prime variable that influences texture. Such models can be used to evaluate optimal strategies for the control of process parameters of a four-stand tandem mill.