Three-Point Synchrophasor Estimator With Better Off-Nominal Frequency Leakage Reduction

Roman Novak

doi:10.1109/access.2021.3069043

Abstract

Synchronous phasor measurements have long been an important tool in power flow modelling and network state estimation as well as in many other tasks that support power grid observability. Classic phasor estimation algorithms are often preferred over complex alternatives. They are based on a variant of the Discrete Fourier Transform (DFT) while relying on a sampling frequency locked to the system nominal frequency. The spectral leakage caused by the small off-nominal frequency excursions in such settings introduces errors that are handled by the estimation algorithms to a different extent. Well-known three-point averaging filter (3P) is a low-cost spectral leakage reduction technique in one of the simplest estimators based on a single-bin DFT. However, it only partially removes the oscillating components. When an independent standard-based frequency tracking is available, the errors can be further compensated with minimal cost. We propose a frequency-corrected three-point estimator (F3P) to eliminate the induced spurious spectrum progressively with the increasing sampling rate while building on the analytical expression of the remaining error. We show performance benefits of the F3P under finite sampling rates when compared to the baseline 3P and to the classical Interpolated DFT (IPDFT) in static conditions. The F3P removes most of the leakage without the need of data resampling. In less ideal conditions with additive 50-dB noise, up to a 10-fold reduction of the 3P error is achieved. The error ratio of the reference IPDFT over the F3P is approximately 1.4 under the same conditions and largely independent of the frequency excursion. The presented simulation results prove the algorithm's superior performance in most of the steady-state tests defined by the relevant Standard, whereas the dynamic performance can be brought to an acceptable level by the use of a single-cycle observation window, with the response times comparable to those of the IPDFT.

Highlights

Power grid observability is based on the increasingly sophisticated measurement instruments and techniques
EVALUATION We assess the accuracy of the F3P with respect to the static and dynamic performance metrics outlined in the synchrophasor measurements standard [4]
The performance measure used in the evaluation is the Total Vector Error (TVE), which is the normalized difference between the reference and the assessed phasor, sensitive to both the amplitude and the phase angle error

Summary

Introduction

Power grid observability is based on the increasingly sophisticated measurement instruments and techniques. Voltage or current phasor representation, i.e., magnitude and phase angle at a selected measurement point, is the key parameter, among many other uses, for modelling power flow, network state estimation, wide are control, or for critical event detection. Phasor Measurement Units (PMUs) are commonly deployed over the transmission networks. They are gradually installed in the distribution grids as well [1], where the more dynamic operational parameters raise stringent accuracy requirements. Low voltage differences constrain the tolerable error levels of the synchronous measurements, with the misalignment of the external clock references further penalizing the measurement accuracy. The latter is caused by the use of diverse clock synchronization techniques

Results

Discussion

Conclusion