Introduction
As residential solar photovoltaic (PV) systems and battery energy storage become more prevalent, households are transitioning into "prosumers"—entities that both consume and produce energy. While this decentralization offers benefits, it poses significant challenges for Distribution System Operators (DSOs), particularly in radial distribution networks. In these networks, voltage levels are harder to manage at the "end of the line," leading to potential spikes or drops that threaten grid stability. To mitigate this, DSOs mandate specific local inverter control rules, such as Volt-Var and Volt-Watt.
The Problem: Locational Unfairness
The primary issue with current local inverter rules is their inherent "locational disparity." Prosumers located at the end of a long feeder experience much higher voltage volatility than those near a substation. Consequently, their inverters are forced to throttle (curtail) active power more frequently to maintain voltage limits. This prevents them from fully participating in energy arbitrage—the practice of storing energy when prices are low and selling or using it when prices are high. This creates an economic "unfairness" where a prosumer's profit is dictated more by their physical location on the grid than by their energy management strategy.
Methods and System Architecture
The researchers propose a two-timescale framework to address this disparity. On a slower timescale (every 15 minutes), an energy management system (EMS) uses Linear Programming (LP) to optimize battery usage and flexible loads (e.g., HVAC) based on time-varying electricity prices. On a faster timescale (every minute), a local controller manages the inverter's active (P) and reactive (Q) power.
The core of the study compares three control policies:
- Positive Reinforcement Control (PRC): An aggressive mode where the inverter actively corrects voltage deviations.
- Avoiding Negative Reinforcement Control (ANRC): A "do no harm" approach where the inverter ensures its actions do not worsen existing voltage issues.
- Hybrid Policy: The proposed solution, which applies ANRC to active power (P) and PRC to reactive power (Q). By prioritizing "do no harm" for the profit-generating active power while remaining helpful with reactive power, the system balances grid needs with economic fairness.
Key Results
The framework was tested using real-world price data from the New York Independent System Operator (NYISO) and a stylized 4-bus radial network model.
- Restoring Fairness: Under standard PRC rules, end-of-line prosumers faced a 43.35% Loss of Consumer Gain (LCG) compared to those at the start of the feeder. The Hybrid policy slashed this disparity to just 1.4%.
- Efficiency and Curtailment: The Hybrid policy reduced solar PV curtailment by 91.7% compared to restrictive base cases.
- Computational Performance: The LP-based optimization is extremely fast, with a full day's simulation taking less than 0.02 seconds, proving its viability for real-time home energy management.
Conclusion
This research demonstrates that locational disparity in the power grid is a quantifiable and solvable problem. By intelligently separating the control strategies for active and reactive power, DSOs can maintain grid stability without unfairly penalizing prosumers based on their geographic location. Future work will focus on incorporating forecast uncertainties and comparing these distributed methods with centralized control schemes.
Summary updated