A Method for Predicting Typical Characteristics of Voltage Sags Based on TSO-XGBoost Algorithm

A typical feature prediction method for user side voltage sags based on the optimization of the Extreme Gradient Boosting (XGBoost) algorithm using the Tuna Swarm Optimization (TSO) algorithm is proposed to address the issue of utilizing grid side voltage sag monitoring data and considering the propagation characteristics of voltage sags. Firstly, based on the network topology of the distribution network system, a feature set that can reflect the propagation characteristics of voltage sags is defined. Secondly, from the perspective of parameter optimization, built TSO optimized XGBoost parameters. Then establish a prediction model based on the TSO-XGBoost algorithm to analyze and predict the characteristic quantities of voltage sags on the user side. The final analysis and validation results indicate that the proposed method has higher prediction accuracy and better noise resistance compared to other commonly used prediction model algorithms, and is not easily overfitting during training