Data integration for digital twins in the built environment based on federated data models

Abstract

Improving the efficiency of operations is a major challenge in facility management given the limitations of outsourcing individual building functions to third-party companies. The status of each building function is isolated in silos that are controlled by these third-party companies. Companies provide access to aggregated information in the form of reports through web portals, emails or bureaucratic processes. Digital twins represent an emerging approach to returning awareness and control to facility managers by automating all levels of information access (from granular data to defined key performance indicators and reports) and actuation. This paper proposes a low-latency data integration method that supports actuation and decision making in facility management, including construction, operation and maintenance data, and Internet of things. The method uses federated data models and semantic web ontologies, and it is implemented within a data lake architecture with connections to siloed data to keep the delegation of responsibilities of data owners. A case study in the Alan Reece Building (Cambridge, UK) demonstrates the approach by enabling fault detection and diagnosis of the heating, ventilation and air-conditioning system for facility management.