Modelling and Control of Building Energy Systems using Transfer Learning

Building operations contribute to approximately one-third of global CO² emissions. A promising strategy to reduce these emissions by about 30% is to equip buildings with advanced control systems that optimize energy consumption in relation to renewable energy generation. However, an advanced control system requires an accurate mathematical model that captures the underlying dynamics of the building. Due to the unique characteristics of each building, a new model must be developed for every building. In my dissertation, I employ AI to leverage knowledge from existing buildings to model unknown buildings. This approach facilitates energy-efficient control across a wide range of buildings.