In this paper, finite volume method (FVM) is adopted to simulate irregular Al alloy profile extrusion process since the coefficient of extrusion (ratio of billet and profile cross section area) is too large (more than 25), the plastic deformation is quite severe that it is almost impossible to get satisfactory results using traditional lagrangian finite element method to simulate that kind of deformation. Calculations show it is a good way for the description of material flow, and mesh distortion could be avoided effectively. On the other hand, a computer aided optimization (CAO) model based on orthographic experiments, artificial neural network (ANN) and genetic algorithm (GA) is built and utilized to obtain optimal values of multi-parameters of Al alloy profile extruding processes and moulds, as it's a combination optimization problem and traditional arithmetic algorithm can hardly solve that. Experiment shows satisfactory results, such as process and mould parameters, are obtained. (c) 2007 Published by Elsevier B.V.
