Modeling of aluminum rolling using finite elements and gray-box modeling technique

A semi-physical gray-box model has been developed for the prediction of material properties, such as the stress/strain and the recrystallisation behavior for aluminum alloys. The model input variables are the temperature, strain-rate and strain. It predicts the material behavior and properties under plain-strain compression tests and rolling using a hybrid structure of intelligent models (neuro-fuzzy) and constitutive equations. An integration process between the gray-box model and the finite element has been undertaken, where the gray-box model is utilized for calculating the flow stress distribution for rolling conditions. The paper discusses the model developments and its integration to the final finite element model. Selected rolling results are reported for Al-1% Mg alloy.