Ontology-based Digital Twins for informed decision-making: Scenario testing and optimization in hospital building operations

August Thomsen, Jakob Bjørnskov, Muhyiddine Jradi · 2025 · 4 citationsRead the paper

With the increasing urgency to achieve ambitious energy reduction targets in both existing and newly constructed buildings, Digital Twins have emerged as a transformative solution for optimizing building operations. By enabling data-driven decision-making, these virtual replicas help minimize energy waste while enhancing occupant comfort and overall building performance. This study explores the potential of ontology-based Digital Twins by demonstrating their scenario implementation and testing capabilities. A case study is conducted on a hospital building, which includes 12 conditioned thermal zones and an integrated HVAC system. The scenario development is performed by employing three strategy complexity levels: Flexible strategies, Rule-based strategies, and Preheating strategies. The scenarios are evaluated using occupant-centered KPIs for indoor comfort while determining energy consumption and cost of operation. The main aim is to limit thermal discomfort while maintaining a energy consumption. Different alternatives for lowering thermal discomfort are evaluated, providing building operators with the possibility of informed decision-making. By increasing CO 2 setpoints in each thermal zone while increasing the heating setpoints in all spaces, the thermal discomfort was lowered by 82% and related monetary energy costs by 22.7%. By increasing the supplied air temperature, the thermal discomfort decreased by 98%; however, this strategy resulted in a 129.7% increase in monetary energy costs. Preheating strategies proved relevant, lowering thermal discomfort by 15%, and ensuring air quality while marginally decreasing consumption and cost. The proposed work demonstrates how ontology-based digital twins can be used for informed decision-making by implementing and testing various operation scenarios as a service.

1 idea Seedlabs derived from this research