GEN-Y: people-driven GenAI for energy management
Bolon Chen (Zhuhai Chen)
This project investigates how people-driven generative AI tools can support local energy management. Building on the GEN-Y project, it examines how GenAI can help users and community stakeholders understand energy information, identify opportunities and risks, and make better-informed decisions without replacing technical models. My focus is community-level EV charging, where residents, site managers and local actors face fragmented data, infrastructure constraints, privacy concerns and different levels of digital literacy. The project aims to define practical, reliable and user-centred GenAI use cases for planning, guidance and awareness raising.
Supervisor 1: Professor Daphne Tuncer, Institut Polytechnique de Paris
Supervisor 2: Dr. Fei Teng, Department of Electrical and Electronic Engineering
GEN-Y: people-driven GenAI for energy management
Melts Vlachou Economou
Her thesis investigates the use of a people-driven generative AI interface for residential energy management. This interest connects to her wider motivation to make energy technologies more accessible, practical, and meaningful for everyday users. As generative AI becomes increasingly embedded in society, Meltsia is particularly interested in how it can be used more responsibly and efficiently. While its growing use raises concerns around environmental impact, she believes that the focus should not only be on limiting its use, but on helping people engage with it critically and effectively so that it can support better decision-making within the energy transition.
Supervisor 1: Professor Daphne Tuncer, Institut Polytechnique de Paris
Supervisor 2: Dr. Fei Teng, EEE
Prototype for GPT-Powered Data Assistant for Net Zero Decision-Making
Mohamad Zuhair Taraboulsi
Building energy decisions increasingly depend on data that facilities managers rarely have the time or technical background to interpret. Rooftop solar PV and battery storage offer clear decarbonisation potential, yet recommendations often hide how incomplete data weakens their reliability. This project develops a conversational AI assistant that guides non-technical users through PV and battery decisions using real half-hourly meter data from Imperial's Silwood Park campus. It maximises on-site self-consumption within grid export constraints while communicating uncertainty transparently through graded confidence labels, so users can judge how far each recommendation can be trusted.
Supervisor 1: Dr. Salvador Acha, Department of Chemical Engineering
Smart Shadows: Developing a new optical technique for clean coolant flows
Shengkai Ji
Air conditioning contributes about 4% of global emissions, largely through refrigerant leakage. Supercritical CO2 is a promising clean alternative, but its flow and heat transfer near the pseudocritical point remain poorly understood, partly because it is difficult to measure: the extreme density gradients bend light so strongly that conventional optical methods break down. This project turns that difficulty into a measurement principle. Using CFD-generated temperature fields and nonlinear ray tracing, it develops BOSSIA, a background-oriented schlieren algorithm that reconstructs the flow from the bending of light itself, supporting the design of future CO2-based cooling systems.
Supervisor 1: Professor Christos Markides, Department of Chemical Engineering
Exploring the Benefits of Thermal Energy Storage in Data Centre Operations: A Case Study in the Johor-Singapore Region
Shi Hong Teng
Tropical regions, particularly Southeast Asian countries, are experiencing a significant surge in data centre construction. However, the high temperature and humidity in these regions place considerable strain on cooling systems. This project develops a data centre Heating, Ventilation, and Air Conditioning system simulation model integrated with Ice Thermal Energy Storage, aiming to optimize operation under varying objectives such as cost-effectiveness and carbon reduction. Using the Johor-Singapore region as a case study, the project evaluates how thermal storage integration can ease cooling system tension while improving financial and environmental performance.
Supervisor 1: Professor Graham Hughes, Department of Civil and Environmental Engineering
Supervisor 2: Dr. Po-Heng Lee, Department of Civil and Environmental Engineering