A major problem in product design is that the design process is not clear to designers. Therefore, every time designers develop a new product, they face difficulties in determining the order in which the product attributes should be determined, especially in the case of large, complicated products. This problem mainly occurs due to two reasons. First, the knowledge about past product designs is not well arranged and thus there is no way to utilize it. Therefore, this research focuses on developing a design support system that proposes a design process in which the designer can easily reflect the important attributes of a product while facing less difficulties in completing the design; this is done using a topdown design support system. In a top-down design system, the designer expresses the product knowledge using elements such as entity, attribute, constraint, interface, etc. Further, five types of knowledge are expressed in this system. They are: knowledge about product structure, knowledge about product entity, knowledge about product function, knowledge about product constraint, and knowledge about product design process. Since this research focuses on the design process, extracting knowledge about the product design process is very important. To extract this knowledge, we first compare the template of past products and the product currently being designed. Next, we calculate the consistency of the two models. Then, based on the results of the consistency calculation, we select and extract the available knowledge. We create a new process by using this extracted knowledge from the design template. It is possible to produce more than one process by combining the knowledge from more than one template. Finally, we evaluate the process from three perspectives: whether it is easy to reflect the customer requirements, whether the design conflict difficulty is small, and whether the design loop difficulty is small. Based on the evaluation result, the designers can select a process to design a new product. In this research, the ocean thermal energy conversion (OTEC) system is used as an example. Further, a process that can easily design the important attributes with a smaller possibility of breakdown than the existing process is chosen based on the results of applying a model proposed by this research. A well-organized design process has been achieved in the OTEC example. Future works must focus on improving the evaluation of the design process and the method for expressing the design knowledge as a template.
Construction of Robot Structure Design Support System by Constraint Logic Programming.