A Voltage Sag Source Localization Method Combining VMD and XGBoost

With the widespread access of various new energy power generation equipment and power electronic interface equipment, voltage sag seriously affects the normal work of all kinds of equipment in the distribution network, and causes huge economic losses to the power network. The accurate positioning of voltage sag sources has an important impact on the stable, reliable, and economical operation of the power system. In terms of voltage sag source localization, in order to overcome the defects of using signal processing methods or neural network methods alone, we combine VMD and XGBoost fusion method to determine the specific line number where voltage sag occurs. The powerful of this method was tested through IEEE 33 model simulation, and compared with GRU, LSTM, LSTM+GRU and Lightgbm networks, indicating that this method has high testing accuracy and profound guiding significance for future power system adjustments.