PhD Oral Exam - Peter Yefi, Software Engineering

When studying for a doctoral degree (PhD), candidates submit a thesis that provides a critical review of the current state of knowledge of the thesis subject as well as the student’s own contributions to the subject. The distinguishing criterion of doctoral graduate research is a significant and original contribution to knowledge.

Once accepted, the candidate presents the thesis orally. This oral exam is open to the public.

Abstract

Buildings are complex systems integrating heating, ventilation, and air conditioning (HVAC), lighting, fire safety, and security subsystems, which collectively optimise resource use while maintaining occupant comfort and safety. These mechanical, electrical and plumbing (MEP) systems are typically managed through building management systems (BMS), which integrate data and control functionalities. Despite having BMS, many buildings remain inefficient in energy use and contribute significantly to carbon emissions. A key barrier to energy-efficient solutions is the heterogeneous, proprietary nature of BMS, which hinders interoperability and the transfer of building solutions.

While researchers and practitioners have used traditional RDF/OWL-based ontologies to model building systems to address the heterogeneity of BMS representations, ontologies lack mechanisms for enforcing constraints, modelling behaviour, and integrating control strategies. In this thesis, we introduce MetamEnTh, an object-oriented metamodel for modelling the operational phase of buildings. We adopt a grounded-theory approach, informed by surveys and interviews with researchers and industry practitioners, to develop MetamEnTh, an object-oriented metamodel for MEP subsystems. MetamEnTh employs structured classes with predefined relationships, methods, and validation rules to ensure accurate dynamic models, complemented by interfaces that enable integration of user-defined control logic.

In this thesis, we also evaluate the APIs of common BMS platforms, assessing their ability to expose building system data for interoperable access by MEP subsystems and to integrate control logic. We validate MetamEnTh through real-world use cases and a comparative evaluation against existing ontologies, demonstrating improved accuracy, more detailed modelling of building systems, constraint enforcement, and enhanced error prevention. MetamEnTh provides a practical, extensible metamodel for modelling MEP subsystems that supports data-driven, energy-efficient building management. It enforces constraints to ensure valid models, and its interfaces enable users to implement and seamlessly integrate control logic to improve building operations.

Future work will broaden and expand MetamEnTh by incorporating classes for occupancy modelling and building information model importers, as well as utility methods for energy-efficiency tasks. We will extend our BMS API evaluation to cover predictive control, demand response, and digital twin integration. To support practitioners without programming expertise, we will develop user-friendly editors while aligning MetamEnTh with community initiatives such as ASHRAE, Brick, and Project Haystack.

Beyond its technical extensions, MetamEnTh opens new avenues for validated and executable representations of building systems that bridge the gap between semantic modelling and operational control. It provides a foundation for future research on model-driven digital twins, self-adaptive control, and cross-domain interoperability in cyber-physical infrastructures. By unifying data semantics and control logic, it enables researchers to experiment with novel AI- and simulation-driven methods for optimising energy use in buildings.