The Identity and Geometry of Building Elements in Multidisciplinary Co-Design of Buildings

Weiße Schrift

THE IDENTITY AND GEOMETRY OF BUILDING ELEMENTS IN MULTIDISCIPLINARY CO-DESIGN OF BUILDINGS

This project sets out research methodologies to manage the identity of building elements in models of buildings and to integrate the geometry of building elements with their semantics. In the previous project, RP 20-1, we have developed a methodology and a tool to connect ontologies (i.e. vocabularies for knowledge representation) from different Architecture, Engineering and Construction (AEC) disciplines and showed how some constraint violations can be detected at earlier stages of the design process. In this project we will investigate two shortcomings in current methodologies: 

1. The lack of consistent management of design component identities. Currently, when an architect modifies a column (e.g. changes its design from round to square, or moves it), a new file is created, with a new identifier for the column. It is unclear when the column should have the same identity or when a new identity should be created. Moreover, the rule for the architectural discipline may not be the same as for the structural engineer calculating the column load. 
2. The limited expressiveness of semantic languages with respect to geometric constraints. Semantic relationships, such as the columns that support a beam, are not explicit in architectural models. Although these relationships can be inferred from the geometry of building elements, this inference is not automatic. This lack of integration of semantics and geometry prevents the detection of violated design constraints until it is too late, i.e. after construction has started. 

To address these issues, we will extend the knowledge-driven framework and design methodology developed in the first part of this project, to support reasoning methods that involve the identity of design components and their geometries. We will develop an ontology and rules for reasoning about the component identity, and we will select a set of core geometric operators and map them to semantic relations that can be integrated into knowledge representation and planning languages. We will evaluate the developed methodologies by applying them to selected building designs and/or parts of the demonstrator developed in the IntCDC Cluster of Excellence. 

PARTICIPATING RESEARCHERS

Prof. Dr. Steffen Staab 
Institute for Artificial Intelligence (KI), University of Stuttgart 

Tenure-Track Prof. Dr. Thomas Wortmann
Institute for Computational Design and Construction (ICD), University of Stuttgart 

TEAM

n .n.