Preprocessing and Golomb-Rice Encoding for Lossless Compression of Phasor Angle Data

Abstract

Phasor measurement units (PMUs) are being increasingly deployed to improve monitoring and control of the power grid due to their improved data synchronization and reporting rates in comparison with legacy metering devices. However, one drawback of their higher data rates is the associated increase in bandwidth (for transmission) and storage requirements (for data archives). Fortunately, typical grid behavior can lead to significant compression opportunities for phasor angle measurements. For example, operation of the grid at near-nominal frequency results in small changes in phase angles between frames, and the similarity in frequencies throughout the system results in a high level of correlation between phasor angles of different PMUs. This paper presents several methods for preprocessing of phasor angles that take advantage of these system characteristics, including a new method—frequency compensated difference encoding—that is able to significantly reduce angle data entropy. After the preprocessor stage, the signal is input to an entropy encoder, based on Golomb–Rice codes, that is ideal for high-throughput signal compression. The ability of the proposed methods to compress phase angles is demonstrated using a large corpus of data—over 1 billion phasor angles from 25 data sets—captured during typical and atypical grid conditions.