AUTHOR=Wang Yanwen , Sun Yanying , Dan Yangqing , Li Yalong , Cao Jiyuan , Han Xueqian 

TITLE=Online load-loss risk assessment based on stacking ensemble learning for power systems

JOURNAL=Frontiers in Energy Research

VOLUME=Volume 11 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2023.1281368

DOI=10.3389/fenrg.2023.1281368

ISSN=2296-598X

ABSTRACT=Power systems faces significant uncertainty during operation owing to the increased integration of renewable energy into power grids and the expansion of the scale of power systems, these factors lead to higher load-loss risks; therefore, realization of a fast online load-loss risk assessment is crucial to ensuring the operational safety and reliability of power systems. This paper presents an online loadloss risk assessment method for power systems based on stacking ensemble learning. Four different machine learning models, including support vector regression (SVR), extremely randomized trees (ET), extreme gradient boosting (XGBoost) and elastic network (EN) were used to form a stacking ensemble learning model to extract the relationship between the feature variables and the risk index. Moreover, to further improve the model performance, recursive feature elimination using cross validation (RFECV) and particle swarm optimization (PSO) algorithms were used for feature selection and parameter optimization, respectively. The application of the proposed method on IEEE test systems demonstrated that the proposed method was more accurate than methods based on individual machine learning models, from which the stacking was designed, while still maintaining a significant advantage in terms of runtime compared to the traditional risk assessment method.2 This is a provisional file, not the final typeset article consuming, is as follows: First, probabilistic modeling of each part of the system, such as loads, wind farms, and units is conducted. Then, using enumeration or Monte Carlo method, a large number of possible states of the system are generated. This is followed by quantifying the severity of each possible state based on the power flow analysis. Finally, the risk index of the system, based on the probabilities of all the possible states and their severity values, is calculated (