Optimization design of short-circuit test platform for the distribution network of integrated power system based on improved K-means clustering

Short-circuit test platform can verify the correctness of the distribution network simulation model and the effectiveness of short-circuits protection. It is necessary to simplify the structure of the distribution network into a typical branch and build an equivalent test platform due to the system's complexity. The equivalent system based on empirical classification has particular subjectivity and randomness, which will reduce the representativeness of the short-circuit test platform. Therefore, this paper adopts the improved K-means clustering algorithm to cluster the short-circuit impedance parameters from the typical short-circuit point of the distribution network to the power supply into different clusters. Each cluster center is used as the equivalent system's characteristic impedance to realize the complex system's simplification. The feasibility of the improved K-means clustering algorithm is verified by comparing the short-circuit simulation results with the original and the empirical classification equivalent system.