Harnessing energy flexibility of buildings and communities for decarbonization and resilience using advanced controls and machine learning
The electrical grid experiences peak and off-peak demands on a daily and seasonal basis. In the context of electrification and integration of renewable energy, the grid is confronting more challenges to handle its ever-growing peaks, as well as risks to balance its supply and demand. Buildings, as major electricity consumers, can play a significant role in facilitating this energy transition by providing flexibility services to the grid. The capability of buildings to adjust their electrical demand in response to grid requirements, either known in advance or anytime, can be broadly defined as energy flexibility of buildings. Untapping this flexibility potential of buildings, or communities on a larger scale, requires better control strategies and technologies. This project aims to investigate how advanced controls, such as model predictive control, reinforcement learning or other machine learning methods, can assist to harness the energy flexibility of buildings and communities. The control strategies developed will be evaluated in a simulation environment based on key performance indicators for flexibility. For the numerical simulation, both data-driven and physical energy models will be developed.
Required knowledge
Knowledge of building energy modelling techniques and HVAC systems; Knowledge of programming languages such as Python, Modelica, or MATLAB; Familiarity with machine learning and data science.