High-Speed Voltage Control in Active Distribution Systems with Smart Inverter Coordination and Deep Reinforcement Learning
Mohammad Golgol, Anamitra Pal
TL;DR
This work tackles voltage regulation in distribution systems with high renewable penetration by removing reliance on full real-time system data. It combines a ML-based high-speed DSSE, usable with feeder-head measurements, and a Deep Deterministic Policy Gradient (DDPG) controller to coordinate inverter reactive power for voltage support, without modifying traditional devices. By formulating voltage control as an MDP and training offline with diverse scenarios, the approach learns a policy that maintains voltages near $V_{nominal}$ while minimizing PV curtailment, and executes actions online with sub-second latency. Validation on a renewable-rich IEEE 34-node feeder shows accurate state estimation, effective voltage regulation, and practical viability for real-time distribution system operation. The methodology offers a scalable, infrastructure-light path toward resilient active distribution networks under high PV penetration.
Abstract
The increasing penetration of renewable energy resources in distribution systems necessitates high-speed monitoring and control of voltage for ensuring reliable system operation. However, existing voltage control algorithms often make simplifying assumptions in their formulation, such as real-time availability of smart meter measurements (for monitoring), or real-time knowledge of every power injection information(for control).This paper leverages the recent advances made in highspeed state estimation for real-time unobservable distribution systems to formulate a deep reinforcement learning-based control algorithm that utilizes the state estimates alone to control the voltage of the entire system. The results obtained for a modified (renewable-rich) IEEE34-nodedistributionfeeder indicate that the proposed approach excels in monitoring and controlling voltage of active distribution systems.