TL;DR
This paper presents a novel control framework that combines the slow thermal inertia of buildings with the fast response of energy storage systems (ESS) to provide secondary frequency control (SFC) services to the grid. This coordinated approach significantly enhances grid flexibility and reduces operational costs for buildings by up to 46% without compromising occupant comfort, marking a significant experimental demonstration in heterogeneous demand-response coordination.
What Problem Does This Paper Address?
The stability of power grids relies on maintaining a precise balance between electricity generation and consumption, which is constantly challenged by unpredictable fluctuations. Ancillary services, such as secondary frequency control (SFC), are crucial for compensating these deviations and maintaining grid frequency within acceptable limits. Traditionally, these services have been provided by expensive and less environmentally friendly power plants. While energy storage systems (ESS) offer fast response capabilities, their high cost and limited energy capacity can be a bottleneck. Conversely, buildings, with their inherent thermal inertia, can store significant amounts of energy but typically have slower response times and power limitations. This paper addresses the challenge of effectively utilizing these complementary resources – the fast but energy-limited ESS and the slow but energy-rich buildings – to provide SFC services more economically and reliably.
Methods
The authors propose a multi-level, multi-time scale control architecture. A Planning Module determines the day-ahead baseline power profile and available flexibility by solving an economic optimization problem. A Tracking Module then coordinates the resources in real-time. This module operates at two different timescales:
- High-Level Controller (HLC): Operates at 15-minute intervals, calculating setpoints for the building's HVAC system and deciding on intraday energy market transactions to manage the baseline. It uses a model predictive control (MPC) approach to balance comfort, ESS constraints, and tracking performance.
- Low-Level Controller (LLC): Operates at 4-second intervals, computing the power setpoint for the ESS to precisely track the residual AGC signal, accounting for the building's actual power consumption.
A key innovation is the use of a data-based AGC predictor that captures the time-correlation and daily patterns of the AGC signal, providing probabilistic forecasts to improve the MPC's performance. The control strategy is designed to split the AGC signal's frequency components, allowing the ESS to handle high-frequency components and the building to manage slower, more energy-intensive ones.
Datasets
The experiments utilized historical AGC data from the Swiss TSO (Swissgrid) for the year 2013/2014. This data was split into two subsets: one for constructing the uncertainty set for the Planning Module and another for randomly picking AGC signal realizations for closed-loop operation. Weather data (outside temperature, solar radiation) was collected from local weather stations.
Key Insights and Ideas
- Complementarity is Key: The core idea is that the limitations of fast, power-limited ESS and slow, energy-limited buildings are complementary. By intelligently combining them, a more robust and flexible ancillary service provider can be created.
- Multi-Time Scale Coordination: Effectively managing resources with different response times requires a hierarchical control structure that assigns tasks based on their capabilities.
- Data-Driven Prediction for Robustness: Using a data-driven AGC predictor that captures temporal correlations significantly improves the controller's ability to anticipate future grid demands, leading to better tracking performance.
- Economic Benefits: The synergy between resources not only enhances grid flexibility but also leads to substantial cost reductions for the building by enabling participation in energy markets and optimizing its operation.
- Comfort Preservation: The proposed control strategy successfully maintains occupant comfort within acceptable limits, demonstrating that providing grid services does not necessitate sacrificing user well-being.
Key Results
- SFC Provision: The system successfully provided SFC service for nine continuous days, fully complying with Swiss regulations.
- Enhanced Flexibility: The combined ESS and building system demonstrated significantly augmented flexibility compared to either resource operating alone.
- Cost Reduction: Buildings providing SFC service experienced an average reduction in operational costs of up to 46%.
- Comfort Maintained: The provision of SFC did not negatively impact occupant comfort, as measured by the "Ashrae Likelihood of Dissatisfaction" (ALD) metric.
- Accurate Prediction: The data-driven AGC predictor showed improved prediction accuracy compared to basic methods, especially for shorter horizons (below 30 minutes).
Limitations and Future Work
- ESS Cost: The paper acknowledges that the high capital cost of ESS remains a significant impediment to widespread deployment, even with optimized usage.
- Building HVAC Variability: The diverse nature of HVAC systems in commercial buildings presents a challenge for generalization, though the proposed method aims to be adaptable.
- Experimental Setup: While the experiments were extensive (9 days), they utilized an ESS emulator. Future work could involve testing with a real ESS.
- Cost Function Complexity: The paper mentions that more complex cost functions could be incorporated into the bidding problem, suggesting avenues for further economic optimization.
- Generalization: While the framework is presented as general, further validation across different building types and grid regulations would be beneficial.
Summary updated