9500009 Kim Automated design methodology research in injection molding will be conducted on an actual production part provided by the industrial partner. Prior research has demonstrated the predictive capability of neural network learning algorithm applied to injection molded parts. Building on that experience, this research will optimize the location of weldline and minimize warpage using statistical methods, combining both Box complex and Tagnchi experimental designs. The design space will cover part, mold, machine set up and resin. The simulation will be carried out at the University and actual injection molding will be conducted at the site of the industrial partner. This is a case study approach with a specific practical, high volume part. Such a demonstration enables industrial production to replace the trial and error process optimization with an automated design. Injection molding is a widely applied process for the manufacturing of numerous plastic parts. Process parameters are often setup by trial and error. Successful completion of this project has the potential to improve the consistency and quality of injection molded products. By integrating product design with process simulation, products can be designed for manufacture thus reducing the developmental time. As a result of cooperative research between the University and industry, manufacturing research relevance to the real world scenario will be maintained as a result of the cooperative research.
