Active power ancillary service provision of commercial building energy storage resources

Abstract

The aim of this thesis is to propose new control strategies of building energy storage systems and analyze their effects on operations of power grids and electricity markets. Specifically, novel control schemes of plug-in electric vehicles (PEVs) and variable speed heat pumps (VSHPs) are proposed to improve grid frequency regulation (GFR) and day-ahead (DA) electricity market clearing. The feasibility and effectiveness of the proposed control methods are evaluated using small-signal analysis, simulation case studies, and experimental verifications. An alternative power system for commercial buildings is designed using steady-state and dynamic models of power converters and corresponding controllers. A dynamic model of a VSHP is also presented for real-time simulation studies, while considering the operational characteristics such as the heat rate and coefficient of performance. Using the simulated models, new GFR schemes of PEVs and VSHPs, responding to direct load control (DLC) signals, are proposed and analyzed. First, a small-signal analysis is carried out using transfer functions that represent the aggregated dynamic responses of generators and DLC-enabled PEVs and VSHPs. The closed-loop properties of the proposed GFR scheme are then analyzed using Bode and pole-zero plots. Simulation case studies are then performed using a test grid with various penetrations of PEVs and VSHPs that respond to primary (PFC) and secondary frequency control (SFC) signals. The test grid is implemented using an experimental laboratory-scale microgrid, and its control centers communicate with the hardware units to provide real-time control. In addition, a closed-loop model of an independent system operator (ISO) and a commercial building aggregator (CBA) is presented where the CBA determines optimal energy consumption and reserve deployment of VSHPs and PEVs in response to locational marginal prices (LMPs), while satisfying distribution network (DN) operational constraints. DA market clearing and price-based demand response (DR) are modeled with stochastic optimization problems. Simulated case studies are then performed to estimate variation in the operational costs of the ISO and the CBA, as well as in the LMPs under various conditions, as determined by the temperature control methods, the building energy storage resource penetrations, and the DN operational constraints.