1 AUTOMATION IN BUILDING DESIGN WITH SPATIAL INFORMATION Tang-Hung Nguyen Assistant Professor, Department of Civil Engineering, North Dakota State University, 120D CME Building, Fargo, ND 58105 – Tel.: (701) 231-8646 – Email address: hung.nguyen@ndsu.nodak.edu. Amr A. Oloufa Associate Professor and Director of Construction Engineering Program, Department of Civil and Environmental Engineering, University of Central Florida, P.O. Box 16245, Orlando, FL 32816 – Tel.: (407) 823-3592 - Email address: aoloufa@mail.ucf.edu Abstract: While current CAD (Computer-Aided Design) systems provide a variety of data representation schemes (e.g. wire-frames, surface, and solid modeling) and data exchange protocols, they fail to address automation issues in retrieval of building information to be used in different AEC (Architecture, Engineering, and Construction) applications. In these CAD systems, topological information (also known as spatial relations) describing the three- dimensional spatial relationships between building components is conventionally represented in a manual fashion into data models. The manual data representation, however, inherently is a complex and challenging task due to the wide variety of spatial relationships. This paper outlines a computer-based building design framework with emphasis on the engine capable of automatically deducing topological information of building components, which support various aspects in building design such as constructability analysis, construction planning, and building code compliance checking. Keywords: Computer-Aided Design, Topological Information, Automated Building Design Systems, Solid Modeling. Introduction Presently, most CAD systems provide two major tools for representing data of building components: the geometric modeling system dealing with spatial abstractions and the database management system dealing with functional information. The spatial information describes the building components’ geometry ((i.e. dimensions and locations) and topology (i.e. spatial relations among the components), while the functional data represents all the other discipline-specific properties of the components (e.g. structural, thermal properties, design constraints, and building regulations). Formal representation of spatial data in general and topological data in particular of building components is a complex and challenging task in developing building design systems. The complexity is in part due to the fact that each professional usually utilizes his/her own representation of topologies and dimensionalities to express spatial information of building components. Furthermore, different design tasks require different types of topological information. For example, information about the walls surrounding a particular space (e.g. fire zone) is needed for code compliance checking of that space, details of connections between individual structural members of a reinforced concrete frame should be provided for reasoning about constructability, and information about the adjacency among floors of a high-rise building is required for planning the sequence of construction activities. The variety of topological information to be used throughout the design and construction process may contribute to complexity in data definition, representation and retrieval. Moreover, the topological information of building/product elements is conventionally structured directly in product data models, in which designers are prompted to manually specify spatial information of interest to different disciplines. Such a manual representation is usually subject to data inconsistency, incompleteness, and prone to error,