Knowledge discovery from simulation and its application in optimization of metal forming process

The finite element method, an artificial intelligence technique, and a gradient-based optimization algorithm are combined and applied to the optimization of an extrusion-forging process. The,finite element simulation is performed according to the scheme of design of experiment (DOE). Then the rough sets-based data mining technique is applied to extract field knowledge from simulation results. Based on the acquired knowledge, the whole design space is reduced to two small sub-spaces. In the reduced design sub-spaces, the optimal design result can be easily obtained through sequential quadratic programming algorithm. The result shows that extracting field knowledge from simulation by artificial intelligence technology is very meaningful and feasible to extend-filed knowledge source. In addition, the optimization process can be remarkably simplified based on the extracted knowledge.