Feature-based modeling is an accepted approach to include high-level geometric information in product models as well as to facilitate a parameterized and constraint-based product development process. Moreover, features are suitable as an intermediate layer between a product's knowledge model and its geometry model for effective and efficient information management. To achieve this, traditional feature technology must be extended to align with the approach of knowledge-based reasoning. In this paper, based on a previously proposed unified feature modeling scheme, feature definitions are extended to support knowledge-based reasoning. In addition, a communication mechanism between the knowledge-based system and the feature model is established. The methods to embed and use knowledge for information consistency control are described.Keywords: Unified feature; Knowledge-based reasoning; Feature-based modeling
INTRODUCTIONHistorically, computer-aided tools, such as CAD, CAPP and CAM systems are developed to support the corresponding product lifecycle phases. Product geometry takes up a significant position in these systems. Feature technology is mainly used to provide high-level geometric representations to facilitate parameterized and constraint-based product geometric design process. However, traditional CAD systems usually assume that a designer has already finished the conceptual design as well as the concept-to-geometry mapping [17]. Therefore, only geometric modeling functions are provided in these CAD systems. Due to this limitation, some information, such as the design intent and the process patterns, is lost during the design process. Explicitly embedding knowledge-based reasoning processes within the traditional engineering systems can significantly enhance the systems' reusability, scalability and flexibility. Knowledge-oriented techniques can support more complex tasks with natural human-oriented intelligent processes. They are more acceptable by human-beings than dataoriented techniques. They can also represent and deal with inaccurate and incomplete information as well as rule-ofthumbs, which are usually difficult to be described mathematically. Due to the different natures between the knowledge information entities and the geometric entities, to interface the knowledge-based reasoning processes with the geometric modeling processes, an intermediate information layer is necessary. Feature-based technology can provide such a bridge. However, the current feature modeling technology has to be extended in two directions. First, feature definitions need to be extended to support knowledge-based reasoning processes. The purpose of this extension are checking and maintaining feature validity in the view of design intent. In other words, each feature model should be consistent to its specific knowledge bases. Second, the communication mechanism between the knowledge bases and the feature models must be established. This paper is intended to address the key issues for enabling knowledge-based reasoning in...