Online prediction and correction control of static voltage stability index based on Broad Learning System

Abstract

With the development of artificial intelligence technology, more advanced algorithms can be used to predict the static voltage stability index more quickly and accurately. However, when the predicted stability index is beyond the set safety range, how to make the static stability index quickly get back into the safety range has brought challenges to the development of the smart grid. In this paper, the Broad Learning System (BLS) is used to predict the static voltage stability index online, and the feasibility of applying BLS to the voltage stability index is analyzed. It is tested on the IEEE-14 bus system and verified on the IEEE-118 bus system. The results show that the BLS algorithm is superior to the machine model in RMSE and MAPE. At the same time, the prediction performance of the BLS model when N-1 fault occurs in the system is considered. In order to solve the problem of how to optimize the system to make the voltage stability index return to the safe operation index when the stability index exceeds the safe operation index, this paper proposes a correction control model based on SHAP-XGBoost-BLS, which is composed of SHAP-XGBoost and BLS models. SHAP-XGBoost-BLS can easily calculate the approximate value of the sensitivity of each characteristic to the voltage stability index under any operation mode, the sensitivity is used to correct the power system, which is verified on the IEEE-14 bus system. The results show that when the voltage stability index exceeds the safety index, the model can give the system optimization scheme so that the stability index can run in the safe range.