Grid partitioning method for medium voltage distribution network based on improved k-means algorithm

Using principal component analysis method to reduce the dimensionality of medium voltage distribution network data. Using the farthest distance algorithm to solve the length between two pairs of data in the medium voltage distribution network, applying the improved k-means algorithm to cluster all data into multiple clusters based on distance measurement, achieving clustering of distribution network data. Based on this, grid positioning is carried out for the medium voltage distribution network to evaluate the power flow value of the distribution network. Finally, design the grid division process for the distribution network. The experimental results show that the medium voltage distribution network structure divided by the research method is reasonable. Under the application of this method, the load in different areas is relatively balanced, and there is no cross between the matching line segment and the main supply line segment in the research method division results, which meets the condition of non cross power supply of the line.