Research Of Funtional Knowledge Clustering Methodology To Support Mechanical Product Conceptual Design

Conceptual design is a crucial task in mechanical product development cycle. Researchers have pointed out that conceptual design is extremely important in computer-aided design but it is difficult to carry out. Conceptual design has massive impact on manufacturing and product quality. Many manufacturing processes (such as molding, casting and machining) are predetermined indirectly by conceptual design stage. The concept generated at the conceptual design stage will influence basic product shape and material selection. In the following detail design, it is impossible to compromise or correct any poor design that comes from previous stage. In nowadays industrialized society, resources and equipment are geographically distributed as well as knowledge and expertise. To keep up the competency of enterprises, shorten product development timelines and develop creative product for the market are of critical importance. Traditional case based design methodology has largely increased the design efficiency. The drawback of case based design is that it lacks effective support for design creativity due to its single disciplinary nature. During case retrieval and case modification, the work is done within one area and one area alone. Thus it is necessary to develop a new design method that could ultilize multidisciplinary knowledge.My research is grounded on the characteristics of conceptual design, utilizing related technology of knowledge-based engineering to propose a computer-aided intelligent design approach based on functional knowledge clustering. Major research achievements are as follows: 1. Propose a constrained functional knowledge model (CFKM) for function knowledge in conceptual design. The model is based on the information obtained in conceptual design stage such as design requirements, physical constrains, design experience and etc. The knowledge modeling focuses on two major design elements which are function and structure; apply OWL ontological language to establish the mapping relations between them as well as constrains.2. Propose a functional semantic clustering approach for CFKM. Build up a semantic model for function predicates in CFKM; develop a semantic clustering algorithm based on semantic similarity of function predicates. Construct fuzzy linguistic computation model to resolve terms conflict.3. Propose a collaborative clustering approach of structural constrains for CFKM. Define collaborative similarity of structural constrains by analyzing different constrain types and their corresponding data categories. Propose a collaborative fuzzy clustering algorithm to deal with multi-constain occasions. The algorithm is demonstrated on small-scale data sets.4. Propose a conceptual design solution space reduction approach. For the clustering sets obtained from previous steps, initial conceptual design solution space is formed by structure combination operation. Structure compatibility analysis is carried out on a quality and quantity bases. Fuzzy multi-objective decision making approach is adapted as quality analysis and the concept of Compatibility Index (CI) is proposed to be the norm of quantity analysis.5. Design case validation and system application. A CFKM-based conceptual design system is developed based on the methodology proposed in the paper. The conceptual design process of multi-mode molecular imaging device is demonstrated to prove the effectiveness of the system and the methodology proposed in the current research.