TL;DR
This paper introduces a novel hierarchical framework that coordinates buildings with both thermal and battery energy storage to act as a collective "price influencer" in the electricity market. By using a game-theoretic approach, it optimizes energy usage for both building owners (minimizing costs) and the distribution system operator (maximizing profit), leading to significant peak load reduction and cost savings.
What Problem Does This Paper Address?
The increasing demand for electricity, coupled with the rise of renewable energy sources, presents challenges for grid stability and affordability. Buildings, responsible for a large portion of electricity consumption and peak demand, offer a significant opportunity for flexibility. Behind-the-meter (BTM) energy storage systems, like thermal energy storage (TES) and battery energy storage (BES), can help manage this flexibility. However, effectively aggregating and coordinating these diverse BTM resources across multiple buildings to influence grid operations and market prices, while also considering distribution system constraints, remains a complex challenge. Existing approaches often treat aggregators as price-takers or focus on individual building optimization, failing to capture the full potential of coordinated, price-influencing behavior at the distribution level.
Methods
The paper proposes a hierarchical, bi-level optimization framework that models the interactions between the Distribution System Operator (DSO) and aggregators.
- Upper Level (DSO): The DSO acts as the leader in a Stackelberg game. It optimizes its power procurement from the wholesale market to minimize costs, considering distribution system constraints (power flow, voltage). This optimization determines the wholesale power price, which then influences the local retail prices offered to aggregators.
- Lower Level (Aggregators): Aggregators, representing groups of buildings with TES and BES, act as followers. They optimize the dispatch of their buildings' energy storage assets to minimize their operational costs (energy and demand charges) based on the prices set by the DSO.
- Game Theory: A Stackelberg game is employed to model the sequential decision-making between the DSO (leader) and aggregators (followers). An iterative optimization process is used to find an equilibrium where both parties achieve their objectives.
- Physics-Based Modeling: The paper utilizes physics-based models for building energy assets, including chillers and thermal energy storage, to accurately capture their nonlinear behaviors and operational constraints.
Datasets
- IEEE 33-bus system: A standard radial distribution network is used as the testbed for simulations.
- EnergyPlus: Building simulation models are generated using EnergyPlus to provide realistic electric and thermal demands for office buildings.
- Typical Meteorological Year (TMY) weather data: Summer day weather data from Chicago, IL, is used for the simulations.
Key Insights and Ideas
- Aggregators as Price Influencers: By coordinating TES and BES across multiple buildings, aggregators can move beyond being mere price-takers and actively influence electricity prices, creating a more dynamic and efficient market.
- Synergistic Operation of TES and BES: The study demonstrates that TES and BES can operate synergistically, with one system compensating for the limitations of the other to meet demand and optimize costs. For instance, when TES is charging, BES might be used more, and vice-versa.
- Location Matters: The location of aggregators within the distribution network significantly impacts local power prices due to factors like power losses.
- Hierarchical Coordination is Crucial: A bi-level approach is essential to capture the interplay between grid-level optimization (DSO) and building-level optimization (aggregators), leading to more holistic benefits.
Key Results
- Peak Load Reduction: The proposed approach achieved a significant reduction in peak electric load, ranging from 25-29% compared to the base case.
- Cost Savings: The optimal strategy resulted in approximately 27% total cost savings for building customers, primarily driven by substantial reductions in demand charges.
- Price Influence Demonstrated: The study showed that local power prices at aggregator buses increased when demand was high and decreased with coordinated energy storage operation, indicating the aggregators' influence.
- Improved Performance over Rule-Based Control: The optimal strategy outperformed a simple rule-based control (charging/discharging at fixed times) in terms of peak load reduction and overall cost savings.
Limitations and Future Work
- Business Model of Aggregators: The paper acknowledges that the business model of the aggregator was not considered, which could influence real-world implementation.
- Forecasting Accuracy: The real-time optimal dispatch relies on predictions of future costs and loads, and the accuracy of these forecasts can impact performance.
- Incorporating Renewables: The authors suggest that the framework is adaptable to incorporate renewable energy sources like photovoltaics (PV) as additional decision variables, which could further enhance grid operation and sustainability.
- Generalizability to Other Building Types: While office buildings were used as an example, the framework's adaptability to different building types with varying energy system models is highlighted as a future direction.
Summary updated