Error Correction for Aggregation Model of Wind Farms Considering LVRT Characteristic

Abstract

Accuracy and simplicity issues of the widely-used aggregated model of wind farms focus on minimizing the error during the transient characteristics. In order to reduce the deviation between the detailed and reduced-order model, this paper presents a data-driven system identification method to discover a mathematical structure from the error data that possibly improve the established model. In this work, we devise a library of candidate dynamics tailored to wind farms for regression algorithm. This framework only requires fairly little data and is robust to noise, having good performance for the response to fast system variation. The simulation result illustrates the accuracy of the improved aggregation model of a wind power plant under low voltage ride-through (LVRT) mode.