Dissertation
FREQUENCY REGULATION SERVICE PROVISION FROM DISTRIBUTION CONNECTED BATTERY ENERGY STORAGE SYSTEMS
Washington State University
Doctor of Philosophy (PhD), Washington State University
01/2022
DOI:
https://doi.org/10.7273/000004531
Handle:
https://hdl.handle.net/2376/123412
Abstract
The integration of high levels of intermittent renewable energy sources and distributed energy resources (DERs) introduces significant challenges for the power system operator in maintaining secure operating conditions. For instance, the intermittency of these resources causes frequency deviations and over-voltage or under-voltage conditions that may result in partial power grid collapse. Traditionally, operators rely on specially equipped generators to provide various ancillary services such as voltage support, load shifting, frequency regulation, and grid stabilization that ensures stable grid operations. Recently, there is an abundance of interest on the use of battery energy storage systems in providing ancillary services due to their almost instantaneous response time, especially, in frequency regulation services where the regulation instructions are highly stochastic and occur at a very high time resolution. With rapid decrease in battery cell manufacturing cost and the changes in market rules that has lowered the barriers for batteries to provide regulation services, battery energy storage systems (BESS) are proving to be an effective solution in providing frequency regulation services to the bulk grid. However, there are several concerns for the transmission/distribution system operators (TSO/DSO) with the frequent dispatching of the distribution-connected fast-responding storage systems. First, the existing decoupled models for transmission and distribution (T&D) simulations are unable to capture the complex interactions between the two systems. Second, distribution aggregators controlling large numbers of energy storage devices with individual constraints in accordance with the strict rules of frequency regulation markets and network level constraints is challenging. Given these limitations, in this work, we developed a framework for co-simulation of T&D model, real-time control algorithms, and mathematical and deep reinforcement learning (DRL) based optimization methods to allow for efficient utilization of the BESS in the frequency regulation market while maintaining the distribution network level constraints. This thesis also presents methods of imitation learning (IL), an improvement to the developed DRL algorithms that solves time complexity and scalability issues. The large-scale applicability and minimal customization of these algorithms led to the development of an open-source power grid-RL wrapper for researchers interested in exploring the applications of RL algorithms for power distribution systems.
Metrics
1 File views/ downloads
38 Record Views
Details
- Title
- FREQUENCY REGULATION SERVICE PROVISION FROM DISTRIBUTION CONNECTED BATTERY ENERGY STORAGE SYSTEMS
- Creators
- Gayathri Krishnamoorthy
- Contributors
- Anamika Dubey (Advisor)Anjan Bose (Advisor)
- Awarding Institution
- Washington State University
- Academic Unit
- Voiland College of Engineering and Architecture
- Theses and Dissertations
- Doctor of Philosophy (PhD), Washington State University
- Publisher
- Washington State University
- Number of pages
- 158
- Identifiers
- 99900882927501842
- Language
- English
- Resource Type
- Dissertation