Ontology-based knowledge representation for industrial megaprojects analytics using linked data and the semantic web. (October 2020)

• Record Type: Journal Article
• Authors: Zangeneh, Pouya; McCabe, Brenda
• Is Part Of: Advanced engineering informatics. Volume 46(2020)
• Publication Date: 2020-10
• Subjects: Industrial megaprojects -- Knowledge representation -- Project analytics -- Risk analysis -- Ontology -- Semantic web; Computer-aided engineering -- Periodicals; Engineering -- Data processing -- Periodicals; 620.00285
• DOI: 10.1016/j.aei.2020.101164
• Languages: English
• ISSNs: 1474-0346

Abstract

Abstract: The fourth industrial revolution has affected most industries, including construction and those within the delivery chain of megaprojects. These major paradigm shifts, however, did not considerably improve the track record in predicting project outcomes and estimating required resources. One reason is the lack of unified data definitions and expandable knowledge representation across project lifecycle to represent megaprojects for analytics. This paper proposes and evaluates a unified ontology for project knowledge representation that facilitates data collection, processing, and utilization for industrial megaprojects through their lifecycle. The proposed Uniform Project Ontology, or UPonto, provides a data infrastructure for project analytics by enabling logical deductions and inferences, and flexible expansion and partitioning of the data utilizing linked data and the semantic web. The ontology facilitates cost normalization processes, temporal queries, and graph queries using SPARQL, while defining universal semantics for a wide range of project risk factors and characteristics based on comprehensive research of the empirical project risk and success literature augmented by practical considerations gained through expert consultations. UPonto forms the basis for a project knowledge graph to utilize unstructured data; it as well provides semantic definitions for smart IoT agents to consume project risk data and knowledge.