Diagnosis And Process Optimization Of Injection Molding Products Defects Research And Development Of Expert Systems

The quality of injection molding part is relative deeply to material properties, mold design, plastic processing parameters setup, and environment condition in injection molding process. Of all the factors, the process setup has direct effect on the melt behavior and the final part quality. The conventional method diagnoses plastic part's flaws and optimize process variables manually, which is deeply depend on the operator's knowledge and experience. So it is significant to design and develop an expert system for part flaws diagnosis and process variables optimization of injection molding based on knowledge and experience.In this paper, the research on plastic part quality control and process optimization in injection molding process is mainly carried out. On the basic of the research, combining common principles and methods of expert systems development, an expert system focus on the part flaws diagnosis and process parameters optimization of injection molding is developed using GUI tool-Visual C++6.0 on windows 2000 platform.The knowledge base is the basis and premise of expert system to make reasoning. Knowledge database include theory, general and heuristic knowledge acquired from expert, injection handing book, experiment design and analysis, and the real process of injection molding. The frame representation method and the production representation method are adopted to express the fact knowledge and the rule knowledge,and fuzzy technique is used to deal with the uncertainty knowledge of part flaws and corresponding rules, to form the knowledge database and the relationship among knowledge database.Inference engineer is the organization and control construction of expert system and use the data offered by user to carry out the inference procedure following the determinate inference policy. On the basis of the structure of knowledge database and the characteristic of injection molding process, the system inference engineer is developed using the production rule inference and the fuzzy inference. The positive inference technique is used as the macroscopical inference policy and the heuristic hunting policy is used in the inference system.Modularization design is adopted in the system, which has strong portability and friendly interface. The expert system can diagnose 11 kinds of part flaws include short shot and sink mark, etc. Further more, it can help man to gain the adjustment values of the process parameters that reduce the intensity of flaws or completely erase them. At last it is proved that the injection molding part flaws diagnosis and process optimization expert system has a fine achievement through application on an example.