Power Grid Stability Identification Based on extreme Gradient Boosting

Abstract

With the development of economy, the power grid structure becomes more and more complex, and many new business scenarios and new requirements has emerged. The power system stability analysis method based on physical model was unable to accurately model the uncertain factors in the power system, which make difficulty to analysis the system. In this paper, a method of grid stability identification based on XGboost method has been proposed, and the state of power grid was determined from the grid multidimensional operation data. Firstly, the decision tree method was introduced. Secondly, introduced the principle and characteristics of classification and regression tree, and then introduced the XGboost principle and characteristics. Finally, the IEEE39 node system has been used to verified the correctness of the XGboost.