Multi-requirements ontology engineering for automated processing of document-based building codes to linked building data properties

Abstract

The lifecycle of a building is characterized by precise compliance with regulatory specifications for both traditional planning as well as digital planning methods such as Building Information Modeling (BIM). Currently, this regulatory information is mostly available in the unstructured and non-machine-readable form of guidelines, regulations, and standards. The acquisition and conversion of unstructured data into structured semantic knowledge bases for the use in the BIM processes are increasingly being automated using natural language processing in ongoing research. Since ontologies provide such techniques that allow raw data to be formally transformed into domain knowledge, in this paper an ontology is developed based on the data format for data catalogs, properties, and groups of properties described in the ISO 23386. The contribution of this paper is a systematic requirements analysis based on existing literature, standardization, and existing ontologies and the implementation of an ontology accordingly. Applied to a feature of interest, an exemplary property validation using the SPARQL Protocol and RDF Query Language (SPARQL) and generated Shapes and Constraint Language (SHACL) shapes is performed to show how the collected data can be used for automatic constraint checking.