Examining the rank of Semi-definite Programming for Power System State Estimation

TL;DR

This paper investigates the rank properties of semi-definite programming relaxations in power system state estimation and proposes a new approach to reduce rank by integrating PMU data, enhancing the likelihood of physically meaningful solutions.

Contribution

It introduces a novel method to potentially lower the SDP relaxation rank in power system state estimation by supplementing measurements with PMU data.

Findings

Rank reduction achieved with PMU data integration

Improved likelihood of obtaining physically meaningful solutions

Validated on IEEE test systems

Abstract

In the power system, state estimation (SE) is important monitoring task for the reliable operation of the system. The optimal estimate from the SE is delivered to all EMS application such as fault analysis, automatic generation control. Hence, it is crucial to have good estimation before taking any critical actions. However, the SE problem is challenging problem due to nonconvexity of power flow equations in the nonlinear AC power flow model, which give us a usually local solution. To deal with this nonconvexity, some recent literatures applied the convex semi-definite (SDP) relaxation technique to relax the SE problem attaining or approximating a global solution. In this paper, we investigate the rank of this technique, which is critical to yield a physically meaningful solution with the five-bus test system and propose new approach to possibly reduce the rank by complementing the traditional set of measurement with PMU data. Numerical tests on the standard IEEE 14, 30, 57, 118-bus test system are presented for the demonstration.