My research goal is to design efficient data-driven strategies to control and optimize complex energy systems. For example, large industrial buildings consume a large portion of global energy. On the other hand, building systems encompass complex dynamics of multiple electrical, mechanical, meteorological and control systems. These complexities are challenging to classical model-based optimization and control methods, often requiring significant manual labor for model building and ad-hoc design for individual buildings.
My proposed research applies advanced data-driven artificial intelligence to modeling and control of large-scale, dynamic energy systems. Starting from buildings, by leveraging rich volumes of sensor data, we replace the conventional physical model with data-driven descriptions. We then advance the state of art in combining machine learning and control algorithms to allow operators to reduce energy consumption while maintains the comfort level and normal working conditions. We then extend our methodology to other systems such as microgrids.
Advisor: Baosen Zhang – Electrical & Computer Engineering