Examination Committee

Prof Chin-Tau LEA, ECE/HKUST (Chairperson)
Prof Danny H K TSANG, ECE/HKUST (Thesis Supervisor)
Prof Ling SHI, ECE/HKUST

Abstract

Demand-response is an increasingly valuable resource option that plays a significant role in the reliable operation of electric grid by modifying the consumers' electricity usage especially during peak periods. In this thesis, we develop system models for smart buildings that involve reduction of energy consumption for acceptable levels of occupants' comfort.   

We first propose a framework for the simultaneous control of temperature, illumination and window roller blind position in a building and use Model Predictive Control (MPC) as the control strategy. We further study an aggregation of buildings and consider demand-side flexibility in providing the frequency regulation service with a particular focus on the thermal systems. A hierarchical demand-response market is proposed with a three-step algorithm to model the interactions between the three entities: Independent Service Operator (ISO), aggregators, and end-users. A robust optimization approach is examined to improve the user's decision making subject to the electricity price uncertainty and a bi-level optimization problem is solved to model the interaction between ISO and aggregators. Each aggregator allocates its successful trading reserve among end-users based on their performance score, which is a good scheme to motivate participant resources to respond accurately to the real-time frequency regulation signal.