ScholarWorks@성균관대학교

초록

Building Energy Management Systems (BEMS) collect and monitor building operational data for energy efficiency but face challenges such as complex data management, reliability issues, and a lack of holistic approaches to energy efficiency. To address these limitations, this study proposes Digital Twin-enabled BEMS (DT-BEMS), integrating Digital Twin technology with BEMS to enhance functionality. DT-BEMS uses the Brick schema to create a BEMS ontology and incorporates a virtual building model defined as a mathematical representation of building physical behavior. This model includes three behavioral models: the energy imputation model, the Predicted Mean Vote (PMV) virtual model, and the occupancy model. Holistic Operational Signature (HOS) analyzes building operation data, identifying patterns and relationships for comprehensive efficiency evaluation. In DT-BEMS, a rule-based HOS classifier enables real-time operational signature identification, generating energy efficiency insights. The process of energy efficiency information generation was continuously updated throughout the building operation stages to ensure reliability and accuracy. The proposed system was validated using one month's data from an office space, with data expanded at three-day intervals. DT-BEMS identified eight operational signature types, achieving a 97.9 % accuracy rate when using just 12 days of data compared to a full month. During the study, 23.5 % of operations were identified as inefficient, with energy waste during unoccupied periods being the most prevalent (19.9 %). The virtual model improved identification accuracy by 4.6 % of the data. Real-time operational signature identification was implemented in the Dynamo environment, highlighting DT-BEMS as a practical tool for enhancing building energy management, reducing inefficiencies, and supporting sustainable building operations. © 2025 Elsevier Ltd

키워드