Answers are grounded in this paper and cite their sources.
Demand Side Management (DSM) applications in buildings rely on heterogeneous information systems, with data originating from different sources. Semantic Web technologies allow for connecting these disparate data sources by standardising metadata based on ontologies. Recent research focuses on designing portable control applications that can run across buildings. Although the first step towards realising portable control actions is standardising the metadata of the buildings, significant gaps still exist in the literature when it comes to creating portable Model Predictive Control (MPC)-based DSM applications. Many existing portable applications are simple rule-based programmes and the principles of semantic portability are not clear in the areas like with MPC systems for DSM. This paper proposes a combination of modular services and metadata standardisation towards making MPC systems more portable across buildings. The method consists of (1) decomposing the MPC system into modular and reusable services and exposing their data using already standardised web interfaces, (2) extending metadata schemes to formalise the information requirements of the modular services, and (3) devising a modular semantic-driven portability service to query, validate, and configure the MPC system. Although full portability remains a challenge due to heterogeneity in building systems and metadata modelling styles, our approach demonstrates the feasibility of using standardised ontologies, semantic validation, and modular services to partially automate configuration and integration, thereby a step towards portable MPC applications. The proposed workflow is implemented, tested, and validated in a MPC system as part of a DSM strategy for controlling Electric Vehicle (EV) charging behaviour in an office microgrid system.
No takes yet. Share an insight, caveat, or question.
Chamari et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: