[Paper Review] Distributed Power Apportioning with Early Dispatch for Ancillary Services in Renewable Grids

Sourav Patel, Blake Lundstrom+2 more|arXiv (Cornell University)|Jul 22, 2020

Microgrid Control and Optimization17 references4 citations

TL;DR

This paper proposes a distributed finite-time consensus protocol for coordinating hundreds of distributed energy resources (DERs) to provide secondary frequency response (SFR) in renewable-rich distribution grids. By integrating a novel early dispatch mechanism with brown start, the framework enables faster-than-state-of-the-art response—achieving initial dispatch in under 5 seconds and ramping within 50 seconds—while prioritizing renewable generation and maintaining stability under communication delays and variable RES output.

ABSTRACT

This article develops a distributed framework for coordinating distributed energy resources (DERs) in a power network to provide secondary frequency response (SFR) as an ancillary service to the bulk power system. A distributed finite-time protocol-based solution is adopted that allows each DER in the network to determine power reference commands. The distributed protocol respects information exchange constraints posed by a communication network layer while being robust to delays in the communications channels. The proposed framework enables coordinated response and control of the aggregated DERs by apportioning the share of generation that each DER needs to provide towards meeting any specified global SFR command while allowing for adjustments due to variability in generation and demand in order to prioritize renewable energy sources in the network. A novel early dispatch mechanism with brown start is synthesized to achieve initial DER response to changing SFR commands that is faster than state-of-the-art distributed approaches. The proposed power apportioning protocol is validated using an end-to-end power hardware-in-the-loop configuration at a distribution system scale with 40+ physical hardware DERs, underlying 7-MW power system model, a 250-DER communication topology with physical and simulated distributed controller nodes, varied communication protocols, and an underlying real-world power system model. Experimental results demonstrate the efficacy of the proposed method toward distributed coordination of hundreds of DERs for providing fast response at SFR timescales.

Motivation & Objective

• To address the scalability and latency limitations of centralized control in ancillary service coordination for distribution systems with high renewable penetration.
• To enable fast, coordinated response of distributed energy resources (DERs) to secondary frequency response (SFR) commands under communication delays and variable renewable generation.
• To develop a distributed power apportioning protocol that prioritizes renewable energy sources while ensuring accurate and timely dispatch to meet aggregator commands.
• To achieve finite-time convergence and early dispatch capability in a decentralized framework robust to bounded communication delays and dynamic generation capacity.

Proposed method

• A distributed finite-time ratio consensus algorithm is used to compute individual DER power reference commands based on global SFR commands, ensuring convergence within a bounded time horizon.
• The protocol respects communication network constraints and is robust to bounded delays in message exchange between DERs.
• A novel early dispatch mechanism with brown start enables immediate response to new commands by pre-computing feasible dispatch trajectories before full consensus is reached.
• The framework dynamically apportions power among DERs based on real-time generation and load conditions, prioritizing renewable sources (PV and BESS) to minimize curtailment.
• The system integrates physical and simulated DER nodes in a real-time power hardware-in-the-loop (PHIL) testbed with a 7-MW power system model and 250-DER communication topology.
• The algorithm is implemented across 40+ physical DERs and 250 total nodes, including both real and simulated controllers, under varied communication protocols.

Experimental results

Research questions

• RQ1 Can a distributed consensus-based framework achieve finite-time convergence for power apportioning in a large-scale DER network with bounded communication delays?
• RQ2 How can early dispatch be achieved in a decentralized DER coordination system to improve response speed for SFR services?
• RQ3 To what extent can renewable energy sources be prioritized in a distributed dispatch framework without compromising system frequency regulation performance?
• RQ4 What is the achievable ramp response time of a distributed DER network using the proposed early dispatch and brown start mechanism under real-world communication and generation variability?
• RQ5 How does the proposed framework maintain voltage stability and system reliability when subjected to sudden changes in aggregator commands or solar irradiance?

Key findings

• The proposed framework achieved an initial dispatch response in less than 5 seconds, significantly faster than state-of-the-art distributed approaches.
• The aggregate DER network reached steady-state power output within 21–44 seconds after a command change, meeting SFR timescale requirements.
• Voltage deviations on the secondary feeder remained within ±1% of nominal voltage under normal operation and within ±5% during dynamic events.
• During periods of high solar irradiance, the system prioritized PV generation, reducing BESS dispatch to zero and minimizing energy storage usage.
• When solar generation dropped, BESS units were automatically dispatched to their maximum capacity of 1,200 kW to compensate for the deficit, demonstrating dynamic load balancing.
• The brown-start and early dispatch mechanism enabled rapid adaptation to changing aggregator commands and variable RES output, maintaining accurate power apportioning under real-time constraints.