Semantics for Smart Building Systems:Towards Portable Demand Side Management Applications in Buildings

Demand Side Management (DSM) strategies enhance building energy efficiency by optimizing energy consumption to align with grid capacity. By employing smart meters, automated systems, and demand response programs, buildings can regulate demand in real-time, thereby balancing supply and demand. These high-level software applications must leverage varied data streams to make analytical or optimal control decisions. Nonetheless, integrating disparate systems into a unified application presents challenges due to system heterogeneity and the absence of standardized data access. Moreover, the unique nature of building systems renders highly customized applications difficult to adapt and scale to other buildings addressing similar issues. These limitations hinder the widespread adoption of energy-efficiency applications in buildings. Therefore, this thesis seeks to address the following research questions:•Q1: How can diverse data streams, such as those from BMS systems, IoT streams, and BIM models, be integrated to more efficiently create analytical and control applications using ontologies and semantic web technologies?•Q2: What is the appropriate reference system architecture for a smart building to facilitate the integration of heterogeneous information systems in support of energy efficiency applications?•Q3: How can service-oriented architecture and semantic metadata schemes assist in designing more portable and reusable energy efficiency applications?•Q4: How can the design of a Model predictive Control-based DSM application be standardised to enhance its portability? The introductory content and study background are contained in Chapters 1 and 2 of this thesis. To answer the above research questions, Chapter 3 introduces a reference system architecture for data integration in smart buildings. Data from heterogeneous sources are integrated and made accessible via a standard interface. Data-driven applications interact with this standard interface to query data. Three proof-of-concept studies were conducted using the proposed data integration method based on two university buildings and an office building used as a living lab. The results demonstrate the effectiveness of the proposed data integration method in integrating BMS sensor data, semantic BIM data, and IoT data streams. The proposed service-oriented architecture facilitates the integration of heterogeneous information systems and acts as a blueprint for a smart building architecture. To create DSM applications in a portable manner, Chapter 4 introduces a modular service called the semantics-driven portability service, which analyses the application metadata schemas and building metadata graphs to appropriately configure the application for the specific building, thus facilitating portability. In Chapter 5, employing the above-proposed metadata schemas and modular service-oriented design, a Model Predictive Control-based DSM application for an EV- and PV-integrated building is conceptualized, developed, and implemented within a real smart building. Chapter 6 concludes the thesis and Chapter 7 explores the limitations, and avenues for further research. The main contribution of this thesis is the incorporation of semantic web technologies in DSM control applications to increase their portability. The results suggest that building energy systems can be effectively combined to support the creation of portable DSM applications through a proposed architecture that is service-oriented, integrates data via ontologies, and uses a modular application design, thereby fostering widespread adoption of energy-efficient building applications.