Towards Better Co-Design with Disciplinary Ontologies: Review and Evaluation of Data Interoperability in the AEC Industry

Diellza Elshani, Thomas Wortmann, Steffen StaabLDAC@ESWC 2022 · OpenReview

Introduction

The Architecture, Engineering, and Construction (AEC) industry faces a persistent challenge: data interoperability. Early conceptual decisions in building design significantly influence long-term construction costs and environmental performance. Achieving optimal outcomes requires "Co-Design," where architects, structural engineers, and environmental analysts integrate their data from the project's inception.

However, different disciplines conceptualize building elements in fundamentally different ways. For instance, an architect views a column as a volumetric 3D object defining space, whereas a structural engineer abstracts it into a 1D linear "bar" for load analysis. Bridging these distinct disciplinary views is currently hampered by rigid or semantically thin data exchange methods.

Interoperability Paradigms in AEC

Current approaches to data interoperability generally fall into three categories:

  1. Centralized (e.g., Industry Foundation Classes - IFC): This paradigm relies on a single, monolithic data schema. While it aims for universality, its rigidity often stifles unconventional workflows and evolves too slowly to accommodate emerging disciplines.
  2. Distributed (e.g., Speckle): Tools in this paradigm exchange data directly, often focusing on geometric translation. While efficient for moving data between software like Rhino and Revit, these methods often lack a shared semantic understanding, leading to data redundancy and loss of meaning.
  3. Federated (e.g., Building and Habitats object Model - BHoM): This approach allows disciplines to maintain their own specific representations while using a central connector to map data between them.
Comparison of data interoperability paradigms: (a) point-to-point distributed, (b) centralized schema, and (c) federated shared ontology.

A Federated, Semantic Approach

While the BHoM framework successfully enables disciplinary representations (e.g., distinct "Physical" and "Structural" models), it lacks a formal shared schema to link them semantically. This paper proposes extending the federated approach by integrating BHoM with Semantic Web technologies.

By translating BHoM object models into the Web Ontology Language (OWL) and utilizing Resource Description Framework (RDF) standards, teams can create modular disciplinary ontologies. Instead of a single massive schema, disciplines maintain lightweight, specialized ontologies linked by "alignment modules" that explicitly define relationships between them.

Bridging Disciplinary Views with Semantic Mapping

The power of this approach lies in the ability to use semantic relationships like owl:sameAs. This allows a computer to recognize that an architect's 3D column and an engineer's 1D bar represent the same physical entity.

Graph-based mapping between an architectural 'Column' and a structural 'Bar' using shared geometric nodes.

The proposed framework demonstrates how these independent objects can be linked at the instance level. By mapping the start and end nodes of a structural "bar" to the specific endpoints of a physical "column" using RDF/OWL vocabularies, the system ensures data consistency across diverse software ecosystems without forcing all disciplines into a one-size-fits-all model.

Conclusion and Future Work

This conceptual framework offers a path toward more flexible and semantically rich Co-Design. By moving away from monolithic schemas and toward modular, linked ontologies, the AEC industry can better support the multidisciplinary nature of modern building design.

Future work will focus on the active development of these modular ontologies within the BHoM framework and investigating automated methods to convert BHoM’s C#-based data models into OWL. This will facilitate seamless integration with existing Linked Building Data (LBD) standards, such as the Building Topology Ontology (BOT).

Summary updated

BibTeX
@inproceedings{elshani2022towards,
  author = {Diellza Elshani and Thomas Wortmann and Steffen Staab},
  title = {{Towards Better Co-Design with Disciplinary Ontologies: Review and Evaluation of Data Interoperability in the AEC Industry}},
  booktitle = {LDAC@ESWC},
  year = {2022},
  url = {https://openreview.net/forum?id=U7aghae7uI},
}
Figure