Neural network modeling and optimization of semi-solid extrusion for aluminum matrix composites

Problems such as the difficulty of the selection of technical parameters and the large quantity of experimental work, exist in the forming of composite tubes or bars by the semi-solid extrusion process. In order to deal with these existing problems, on the basis of experimental investigation, the genetic algorithm (GA) was applied to the optimal design of technical parameters in the semi-solid extrusion process. The optimized model of a semi-solid extrusion system for composites was established by adopting the artificial neural network and the GA cooperatively. By dealing with the key techniques in the GA realization, for example, the coding mechanism, the generation of the initial population, the mapping from objective function to fitness function, the adjustment of the fitness and etc., the appropriate genetic parameters were determined and the optimized technical data about a semi-solid extrusion processes for composites were provided. Conducting experiments according to the derived data, satisfactory results were achieved with the deforming force of semi-solid extrusion being reduced significantly, indicating the feasibility of the proposed method.