Evaluation method of distribution network state based on IT-II-Fuzzy K-means Clustering Algorithm for Imbalanced Data under PIOT

In view of the problem of data over-matching or data loss when processing unbalanced data sets in traditional distribution network operating state evaluation methods, the evaluation results are inaccurate. A power distribution network operating state assessment method based on the Interval Type II Fuzzy (IT-II-Fuzzy) K-means clustering algorithm(K-means CA) for unbalanced data in the power Internet of Things (PIOT) environment is proposed. First, the overall architecture of PIOT is analyzed. A thematic database of historical data is constructed by applying PIOT to distribution network inspection and data collection. Then, the iterative center update formula of the traditional IT-II-Fuzzy K-means CA is optimized. Adaptive eliminates the influence of imbalanced data on clustering results. On this basis, an evaluation system for the Operation Status of the Distribution Network(OSDN) is constructed. Finally, the evaluation time, evaluation score and equipment failure prediction performance of the proposed method and the other two methods with 5 databases are compared and analyzed through simulation. The results show that the evaluation time is at least 1.68s, and the evaluation score and equipment failure prediction accuracy are 92.16 and 96.2% respectively. The performance is better than the other two comparison algorithms.