A novel approach to create historical powerflow cases from PI data using open source software

This paper discusses a novel procedure by which historical powerflow cases (HPCs) are created using open-source tools to mitigate deficiencies in previous documented methods. The procedure described in this paper is completely automated and performed entirely in Python to retrieve data from the historian of the author's utility (PI System^TM by OSIsoft^®) and perform all commands in the powerflow software of the author's utility (PSS^® E by Siemens PTI). Regular expressions are used to dynamically assign PI tags from the author's PI server to each relevant component in a PSS^® E powerflow case. Validations of the retrieved PI data and the HPC itself are automatically performed by the procedure. This paper compares solved branch powerflows with the PI data which they are supposed to represent by running the procedure 100 times on random timestamps in one calendar year. 95% of the branch errors are less than 25.0 MW and the mean branch error is less than 1.0 MW. 70 seconds of computation time is required to create an HPC using the approach described in this paper which includes all PSS^® E manipulations, downloads from the PI server, and validations.