Research On Key Technologies For Digital Stamping Die Structure Design Based On AI Planning Theory

The target of modern die design research field is to improve digital and intelligent design for die structure. The theory, architecture and key technologies need to be further researched. In this dissertation, AI planning, advanced theory of AI field, is induced to the stamping die structure design. The key technologies, such as mapping relationships between blanking process planning module and die structure design module, intelligent planning design for die gross structure, intelligent planning design for die's components are mainly studied.The traditional die CAD system didn't pay attention to the mapping relationships between blanking process planning and die structure design. The deep research job about this area is obtained in this dissertation. According to AI planning theory, the EDI denotes external driver which effects die structure design. Meanwhile, CI is the interior driver which expresses the complex and indirect affections. Modularization and standardization can be utilized to reduce the difficult degree of design process. The new theory is put forward to deal with the mapping relationships. Integrating planning for variety (PFV) theory into AI planning research field, the scope of AI planning is enlarged and the design efficiency is enhanced.There are some problems in the traditional stamping die CAD system, such as shortage of full intelligent design methods and lacking some important design types. In order to overcome these difficulties, the advanced theory of study planning and recovery planning are introduced in the stamping die structure design system and applied to the design process, such as transform from parts to models, development of character model bases, recovery from character models to working parts. The architecture of the AI planning is refreshed, and some new contents are added into the architecture. The foundation of stamping die structure platform is established firstly. The AI planning theory of knowledge based HTN(Hierarchical Task Network)is integrated with traditional stamping die structure design theory of top-down method for the first time in this dissertation. The dissertation demonstrates method to accomplish the design process for die design variety. The original design task is divided into hierarchical tasks in the design system. The design scope becomes larger for more die types. Intelligent planning operator can improve the design intelligence for die gross structure. Some special design projects can also be achieved by new AI planning design methods. AI planning distinguish method is introduce in the stamping components intelligent design module. The intelligent design results of cutting edges clustering can improve design for components grouping. The repeated design job is avoided, the design complexity is reduced, and efficiency of stamping die components is improved. The partial-order planning theory is utilized to generated new components intelligently. The successful conversation from initial state to goal state can be achieved to accomplish the components design process step by step. The conversation process can be controlled by modify intelligent planning operators. The GRAPHPLAN theory is modified to reach the design goal for assembling the components with gross structure which was generated before. The human expert's knowledge and experience is integrated into the AI planning design system. The dissertation puts forward a new technical way to combine AI planning based system and knowledge-based system together.On the basis of the above-mentioned key technologies, using JAVA API on the 3D-CAD platform I-DEAS 9.2, the intelligent stamping die CAD software system is established and already applied in the industry field. The feasibility and validity of the theory and application mentioned in the dissertation are validated.