Optimal data compression techniques for Smart Grid and power quality trend data

The article concludes on various options for lossless compression algorithms (LCA) of archives of smart grid or power quality data series - the main archive aggregated from a typical continuous monitoring - which contains huge amount of different quantity readings. We have benchmarked a range of available LCA and we proposed the LZMA algorithm as a best performing one. In this article we also describe several methods which further improve the performance of this compression algorithm. We propose to process the input data with relatively simple prediction models to utilize common features in the data sets. We select aggregating model in combination with differential encoding and we develop an interval selection optimization technique specially fitted to define interval distributions to achieve better compression ratio. As a side effect the evaluated modeling technique does also provide an abstraction of recorded data and can be directly used in the superset analytical modules which are so often mentioned in relation with the emerging “Smart Grid” technology.