Software for the Identification of a Load Model by Arrays, Measured by Passive Experiment

Abstract

This paper demonstrates the software for an identification of voltage static load characteristic's model by passive experiment data arrays. The software is able to provide the full cycle of the big data processing, measured by the passive experiment in power systems. The high-level, statically typed programming language C# with the automatic memory management, and, also, with the statistical data manipulation library Accord.NET is used. The dial-in capability to the real time data base of operative-information complexes is implemented. This capability is programmed for receiving data in a real time and following reduction of arrays of mode parameters to the unified time base. The software includes unique software codes for clustering by EM-algorithm in combination with the randomized Suger-James algorithm for the autodetection of cluster's numbers in arrays. It also consists of the program code for taking into account the external power grid influence on mode parameters. These facts make the software different from existing ones. Apart from this, the software provides ellipse of dispersion building based on the code for their displaying with fixed parameters. This paper also presents the operating result of the software work when the data arrays for the industrial consumer is processed and the voltage static load characteristic's mode is identified. The verification with the results after the active experiment shows that the error is in the acceptable boundaries.