Research on Power System Evaluation Based on Hybrid Cluster Analysis

In order to quickly and accurately achieve power system loss assessment, it will be an important research point. A novel network loss assessment method based on hybrid clustering analysis is proposed using data mining and typical scenario simulation ideas. This method first determines clustering attributes for network loss assessment; Secondly, based on the numerical types of each clustering attribute, the original clustering problem is decomposed into two sub clustering problems. After fully considering the characteristics of power data, partition clustering algorithm and hierarchical clustering algorithm are selected for clustering analysis, and the clustering results of each sub problem are integrated; Finally, a typical operating mode set of the power grid is generated based on the mixed clustering results for network loss assessment. Taking a provincial power grid as an example to verify the effectiveness of the proposed method in network loss assessment, the results show that the network loss assessment based on the proposed method has high accuracy, good computational efficiency, and strong practicality in engineering practice.