[Paper Review] Power System Dynamics as Primal-Dual Algorithm for Optimal Load Control

Changhong Zhao, Ufuk Topcu+2 more|arXiv (Cornell University)|May 2, 2013

Power System Optimization and Stability18 references27 citations

TL;DR

This paper proposes a decentralized optimal load control (OLC) mechanism in power systems where local frequency deviations drive load adjustments via a primal-dual algorithm embedded in swing dynamics and branch power flows. The system converges to the global optimal solution despite multiple equilibria, enabling fully distributed, communication-free control with improved transient performance.

ABSTRACT

We formulate an optimal load control (OLC) problem in power networks where the objective is to minimize the aggregate cost of tracking an operating point subject to power balance over the network. We prove that the swing dynamics and the branch power flows, coupled with frequency-based load control, serve as a distributed primal-dual algorithm to solve OLC. Even though the system has multiple equilibrium points, we prove that it nonetheless converges to an optimal point. This result implies that the local frequency deviations at each bus convey exactly the right information about the global power imbalance for the loads to make individual decisions that turn out to be globally optimal. It allows a completely decentralized solution without explicit communication among the buses. Simulations show that the proposed OLC mechanism can resynchronize bus frequencies with significantly improved transient performance.

Motivation & Objective

• To develop a decentralized optimal load control (OLC) mechanism for power networks that avoids centralized coordination.
• To prove that power system swing dynamics and branch power flows naturally implement a primal-dual algorithm for OLC.
• To demonstrate convergence to the global optimal solution even with multiple equilibrium points.
• To show that local frequency deviations convey sufficient global information for optimal load decisions.
• To validate the transient performance improvement of the proposed OLC mechanism through simulations.

Proposed method

• Formulates the OLC problem as minimizing aggregate cost subject to network power balance constraints.
• Models the system using swing dynamics and branch power flows, with frequency-based load control as the dual update mechanism.
• Establishes that the system dynamics correspond to a distributed primal-dual algorithm for solving the OLC problem.
• Analyzes equilibrium points and proves convergence to the optimal solution regardless of initial conditions.
• Uses Lyapunov stability theory to establish convergence despite multiple equilibria.
• Simulates the system to evaluate transient response and synchronization performance.

Experimental results

Research questions

• RQ1 Can power system dynamics inherently implement a primal-dual algorithm for optimal load control?
• RQ2 Does the system converge to the global optimal solution even with multiple equilibrium points?
• RQ3 Can local frequency deviations provide sufficient information for globally optimal load adjustments without explicit communication?
• RQ4 What is the transient performance of the proposed decentralized OLC mechanism?
• RQ5 How does the proposed mechanism compare to conventional control in terms of synchronization and stability?

Key findings

• The power system dynamics with frequency-based load control implement a distributed primal-dual algorithm for optimal load control.
• The system converges to the global optimal solution even though multiple equilibrium points exist.
• Local frequency deviations contain sufficient information for loads to make globally optimal decisions in a decentralized manner.
• The mechanism operates without explicit communication between buses, enabling full decentralization.
• Simulations demonstrate significantly improved transient performance in frequency resynchronization.
• The approach achieves optimal power balance using only local measurements and dynamics, without requiring centralized coordination.