5 March 2024 in IEEE Internet of Things Journal
Buildings consist of systems that have to interact and coordinate with various systems to function smoothly. It is essential to have models and representations of buildings in forms that are easy to read for effective simulation, control, and optimization of building systems. It is also necessary for integrating and creating novel applications and functions. A Building Energy Management System (BEMS) is a common feature of most commercial buildings. It contains models of some aspects of the building and its systems. However, these models in the BEMSs do not entirely model all systems and subsystems and their relationships because their primary function is to control heating, ventilation, air-conditioning (HVAC), and lighting. Project Haystack and Brick have made significant progress in modelling buildings for operational purposes by adopting a metadata approach. They offer machine and human-readable representations of buildings, systems, and their relationships. However, tags and tagsets in the metadata approach have some limitations that stem from a weak structure in defining entities, their properties, and relationships. In this study, we identify seven problems with the metadata approach to modelling buildings and address these problems with an object-oriented metamodel: Metamodel for Energy Things (MetamEnTh). Using an object-oriented modelling technique to establish structure and constraints, MetamEnTh produces a model that portrays a building and its systems. MetamEnTh adheres to the same naming convention of entities as other projects like Project Haystack and Brick. We accomplish a UML representation of the core structure of MetamEnTh and validate the representation through three different case studies.
29 May 2026 in ACM Transactions on Cyber-Physical Systems
Access to Building Management Systems (BMS) and their data is essential for developing applications and conducting research aimed at improving building energy efficiency. The management layer of BMS includes Human-Machine Interfaces (HMIs), such as building management software, and in some cases Human-Device Interfaces (HDIs), which enable direct interaction with individual devices (e.g., thermostats). These interfaces are accessible via Application Programming Interfaces (APIs) that facilitate programmatic access to subsystems such as HVAC, lighting, and security. However, vendor-specific APIs present heterogeneous representations that pose significant interoperability challenges. To address this, we analyse APIs from five major BMS vendors—Siemens, Honeywell, Schneider Electric, Johnson Controls, and Delta Controls—and categorise their capabilities into 12 feature groups. We go beyond descriptive analysis by evaluating these APIs against 85 common applications, providing a systematic framework to measure their suitability for real-world building solutions. Recognising the role of semantic ontologies in interoperability, we further examine how Project Haystack and Brick can represent the data exposed by BMS APIs within these feature groups, highlighting strengths and gaps in existing approaches. Finally, we demonstrate a practical use case that leverages Siemens Desigo APIs to instantiate standardised models through a metamodel. This metamodel supports control algorithm integration, enabling energy-efficient building operation. Our work advances the state of the art by empirically linking BMS API capabilities to application requirements, semantic models, and practical control integration—providing both researchers and practitioners with actionable insights for deploying interoperable building solutions.