Flawed Decompression of Swinging Door Algorithm in PMU Historians

Currently, swinging door algorithm is being used by some utilities in United States for compressing phasor measurement units (PMUs) data to reduce storage cost. The algorithm compresses the PMU data by discarding a subset of the raw PMU measurements. Thus, the compressed data becomes a subset of the raw PMU measurements. During decompression, samples in the discarded subset are interpolated from samples in the compressed subset. In this paper, differences between raw and decompressed PMU measurements are investigated in both time and frequency domains for particular data sets from the Western Electricity Coordinating Council system. When compared with original raw measurements, the decompressed data is found to contain some artifacts. Further investigation revealed that the artifacts are from errors in both the decompression algorithm and phase wrapping. Using reverse engineering, plausible causes of these artifacts are hypothesized mathematically. Based on the hypothesis, a novel algorithm is proposed which corrects the decompressed data. Results of the algorithm suggests that utilities who have PMU data corrupted by the flawed decompression algorithm and/or flawed phase wrapping can use the algorithm to remove the artifacts from the corrupted decompressed data. Ideally, in the future the utilities would correct the errors in the decompression code.