Study On Expert System For Stamping Process Planning Based On Neural Network

The stamping process planning for automotive panels builds a close connection between the products design and the die manufacture of automotine panels. It has been people's common ground to adopt computer aided process planning (CAPP) technique to change the behind situation and to improve the level of the process design. The form of automotive panel is so complex that the stamping process planning depends seriously on the background of experience. Therefore the failings of the conventional system must be overcome, such as in the illegible recognition and the inference method, in order to develop a stamping process planning system with higher intelligent level. In this paper, the expert system (ES) technology is applied to the field of traditional stamping process design, and the knowledge representation and implementation of the expert system is particularly discussedOn the basis of analyzing and summarizing of characteristics of application for expert system on stamping process planning, an integrated expert system is designed by using knowledge representation and knowledge acquisition based on artificial neural network (ANN). And this system possesses the capability of simulating the human experts' ideation, owing to the abilities of model recognition, creation and association that the ANN model has.Based on feature technique and the theory of object-oriented planning, the feature classification of the automotive panel with its detailed data structure is proposed, and the coding rule of the features of automotive panel is established, which creates favorable conditions for the input of automotive pressing process information.The expert system model, proposed in this paper, not only carries forward the virtues of the conventional system, such as using natural language, having the ability of explaining, offering favorable user's interface, but also integrates the technical advantages of the model in the illegible recognition and the reasoning method, therefore, the mixed system model proposed in this paper has both important scientific significance and obvious technical advantages.