Research on Quantitative Method of Power Network Risk Assessment Based on Improved K-Means Clustering Algorithm

Chong Ren,Zhenyu Sun,Xiaoteng Li,Guoqi Fan,Chao Huo,Lin Cheng

DOI: https://doi.org/10.1109/aeees48850.2020.9121429

2020-01-01

Abstract:The risk assessment of the power grid is helpful to guide decision makers to take corresponding measures to improve the security and stability of the power grid operation. Existing risk assessment methods cannot comprehensively and accurately quantify grid risks. To this end, this paper proposes a quantification of grid risk assessment method based on an improved K-means clustering algorithm. The improved K-means clustering algorithm is used to classify key state quantities of the grid. The actual operating state quantity is matched with the classification situation to determine the risk of power grid abandonment and power restriction. Next, determine the best classification number and intra-class data of the data. According to the magnitude of the peak-valley difference, select the data of the photovoltaic and hydropower peak-valley difference and the small peak-valley difference when discarding the light, and verify the load probability of abandoning light by fitting the load data. In the load limit verification, select the load with large peak-to-valley difference and photovoltaic and hydropower data with small peak-to-valley difference, and perform load-limit verification by fitting the hydropower data. Finally, based on the actual data of the Northwest Power Grid, the quantitative comparison and analysis of the scheduling deviation between the fitted data and the actual data, and the regional power grid risk without probability improvement, verified the effectiveness and superiority of the method.