Exploiting Network-induced Correlation for Efficient Compression of PMU Data

Power systems across the world are deploying large numbers of high sampling rate synchronized phasor measurement units at the bulk transmission level. As PMUs generate large volumes of data, transmitting and storing this data for future studies becomes challenging. An attractive solution to this problem is to compress the data using efficient compression algorithms. This paper presents a novel lossy compression algorithm which utilizes the characteristics of power system measurements to improve the compression. Specifically, ideas from image compression are adapted to exploit the spatio-temporal correlations induced by the network in the measurement data. The effectiveness of the compression algorithm is judged based on the space savings and accuracy of data reconstruction. The results presented show that compression ratios as high as 95:1 can be achieved using the algorithm.