Study On Expert System For Drawbead Design Based On Objected-oriented Method

Along with the development of the automobile industry, especially the passenger cars, the forming and the mold design of the large automobile panels are increasingly paid attention by academician and car manufacturer. Recently, computer has been more popular in automobile field, and also brought huge benefit on economy.During the forming process of the automobile panels, the resistance on binder surface is uneven, so we can layout bead to control it. But until now people do not quiet know how the bead resistance being and how do some factors Influence it. So, the design of the bead depends on the experience of the designer not the theory. So the original design had always been reworked a lot during the process of mold debugging, and this take more time and money, and sometimes, even made the whole mold discarded as useless.After comprehend the application of the FEM in automobile panels' forming, aim at the drawbead design which is included in the stamping part design, accomplish where does the bead lay and how to design the parameters of the bead,in order to apply boundary controlling for FE simulation.Combined with local investigation in the automobile company, author analyses the character of the automobile panels' forming, and grasp the key of bead design: pay attention to circumjacent field of an automobile panel, and take the part character in mind. Object-Oriented method and feature technology are introduced into define the arrangement according to the bead design. And combined with the character of the bead design knowledge (fuzzy), I adopt the knowledge representation which includes both rule and artificial neural network, and all the knowledge about the bead design is distributed in the feature (part) class. Besides, the knowledge management system is provided, which can manage the knowledge type like rule and artificial neural network. Database system is provided too, as the exterior deposit.A mapped connection between parameters of the bead and length of flown-in material was set up with artificial neural network in single bead or double beads. Based on the restriction of bead setting, used genetic algorithm, transformed parameters of the bead to chromosome, and the restriction to single adaptation function, then by selection, gene recomposition and variation, searched the best chromosome.