Systematic Approaches for PI System™ Data Compression Tuning

The consolidation of automation in the process industry has brought a challenge: translate the large volume of data in useful information. The algorithms for plant data compression emerge as an alternative to reduce the space demanded by such information. On the other hand, they should preserve the important features that data hold. PI™ (Plant Information, OSIsoft®) is a virtual tool structured to automate the collection, historization and displaying data of companies. Nevertheless, the efficiency of the compression algorithm used in this system is often not satisfactory, because of poor tuning. This paper presents two methodologies to automate the tuning of process parameters for data compression of PI™, based on a representative dataset extracted from the measurements. The applicability of the proposed techniques is corroborated by several case studies.