Semantic rule checking of cross-domain building data in information containers for linked document delivery using the shapes constraint language

• Published in: Automation in construction 2023-12, Vol.156, p.105106, Article 105106
• Main Authors: Hagedorn, Philipp, Pauwels, Pieter, König, Markus
• Format: Article
• Language: English
• Subjects: AEC; Benchmark; BIM; ICDD; Information container; Linked building data; Rule checking; Semantic web; SHACL
• ISSN: 0926-5805; 1872-7891
• DOI: 10.1016/j.autcon.2023.105106

Information exchange is a crucial process throughout the lifecycle of built assets. Building Information Modeling (BIM) and standardized information delivery enable exchanging, storing, and linking cross-domain information bundled in containers such as the standardized Information Container for Linked Document Delivery (ICDD). While assessing the quality of the interlinking and metadata is essential to allow all parties to interpret an information container correctly, most research efforts have focused on code compliance of domain-specific models. In order to address this gap, this paper proposes an approach for validating ICDD information containers using the Shapes Constraint Language (SHACL), the W3C recommendation for Resource Description Framework (RDF) data validation. For optimized querying and validation, a case study applies SHACL inference rules to the linking structure of ICDD to infer predicates between linked entities in four ICDD containers in varying dataset sizes and different linking methods. Afterward, the validation of these containers is analyzed using three types of validation rules with increasing complexity for another SHACL implementation. Results show the feasibility of cross-domain building information validation in ICDD with SHACL and how performance enhancements can be obtained on an inferred dataset compared to the original dataset. •Examines the application of ICDD for Linked Building Data (LBD).•Focuses on the validation of links in ICDD using Shapes Constraint Language (SHACL).•Presents a novel linking approach using object properties and SHACL rule inference.•Performs micro-benchmarks of the property inference and the SHACL validation.•Provides recommendations for the use of link types and identifiers in ICDD.