Ontological foundations for feature-based modeling

Abstract : Feature-based modeling is amongst the leading approaches for Computer Aided (CAx) product modeling. Its core benefit is the use of features to embed design intents into pure geometric product models in order to convey information based on experts' knowledge and applications requirements. Despite various attempts, the very notion of feature remains ambiguous and no promising approach has been proposed to disambiguate and possibly unify its various meanings under a common framework. As a consequence, feature-based models are tuned on specific applications, are hardly reusable across systems, and are scarcely transparent for human comprehension. The purpose of this paper is to present an ontological characterization of features that can act as backbone conceptual and computational structure to represent the meaning of feature classes in a clear manner. For this goal, the ontology formalizes the most general and fundamental properties that all features are required to satisfy. The ontology is built on previous works and integrates the notion of feature within a broader framework for product knowledge representation