Font Size: a A A

Research On Knowledge-Based Intelligent System And Key Technologies Of Cold Extrusion Process Planning

Posted on:2006-05-26Degree:DoctorType:Dissertation
Country:ChinaCandidate:J GaoFull Text:PDF
GTID:1101360152481244Subject:Materials Processing Engineering
Abstract/Summary:PDF Full Text Request
Metal forming is a kind of process to form billets into parts with proper shapes and mechanical properties, under huge energy and load by utilizing the plasticity of metal materials. Not only can parts with complicated shapes be formed, but the mechanical properties can be promoted after metal forming. So metal forming plays a great role in national economy, and it is widely used in industries. Nowadays, in order to reduce the subsequent machining and the product cost, metal forming is developed forwards net forming or near-net forming. Cold extrusion process is one of the most important processes of net or near-net forming. The industrial competition is becoming more and more drastic. High quality and efficiency, low cost of product and energy, net forming or near-net forming, environment protect and durative development have been becoming the key factors for the manufacturer to win the competition. Cold extrusion process can meet the requirement of economy, technique and society in a certain degree. Therefore, how to establish the cold extrusion processes with high quality and efficiency, and how to design and manufacture high quality extrusion dies, have become the pivotal cores of the cold extrusion processes. As we all know, process design is the basis of die design. Multi-stage extrusion is often needed in forming complicated parts, so it is the key work to design proper cold extrusion processes. Basically, the cold extrusion process design undergoes three main phases: traditional handcrafted design, computer-aided design and intelligentized design. now the center of research is transferring to the intelligentized design (knowledge-based design). Research on intelligentized design is interrelated to the development of computer hardware and software, artificial intelligence and database techniques. It has two main problems to be solved in the process of applying artificial intelligence techniques into process design, the first is domanial knowledge expression, second is to intelligentize the reasoning tactics. This paper mainly focuses on some researches of knowledge-based intelligent system and key technologies of cold extrusion process design.First of all, the knowledge expression methods, which are rule-based andfeature-based knowledge expression, were researched. The knowledge base of cold extrusion process design system was established at the platform of VC++, the capability and efficiency of knowledge expression were improved because of object-oriented technique. Management and maintenance system of the knowledgebase was developed, knowledge can be added, amended or deleted in the knowledge base, but only the one who has license of accessing the knowledge base can do it. Knowledge editing system was developed to acquire domanial knowledge. Conflict resolution tactic and reasoning strategy was established. The databases of materials and extrusion equipments, which play great roles in establishing the intelligent knowledge-based cold extrusion process planning system, were also developed.The ANN-based cold extrusion process planning system (ICEP2S) was set up, so the intelligence of system was promoted. In this part, the selection of Bp network was researched on, the mapping network was discussed, and the centralized feedback mapping structure was selected in order to realize ANN mapping for the process design. The neurons in the input, hidden and output layers were selected according to reasons including the influencing factors of cold extrusion process. The training samples and some problems that may be generated in the training process were solved on the basis of researches of Bp arithmetic, and the Bp arithmetic was improved. Feature classification was analyzed, and a new kind of binary encoding method was brought forward for the first time. Different encoding manners were discussed according to shape feature, process feature, precision feature and material feature, and all kinds of feature codes were explained properly. In the end of this part, typical cold extrusion parts were selected to test the int...
Keywords/Search Tags:cold extrusion process designing, knowledge-based expert system, artificial neural network, fuzzy logic, numerical simulation
PDF Full Text Request
Related items