Method for division of urban load power supply district based on cluster analysis

Abstract

The division of urban power supply district is an essential issue in the medium voltage distribution network plan. This study develops a method for division of an urban load power supply district, integrating the open source data into the distribution network planning, at first, raw georeferenced point of information data is crawled by crawler program based on location retrieval service interface (Place API), the buildings' data of the planned urban district is extracted and power load estimation are introduced in detail, the dataset of the low-voltage load spatial distribution is set-up; secondly, the clustering algorithm selects both the local density of samples ρ i and the distance between samples δ i as criteria to form the clusters, cluster centres are recognised from the binary pair in `decision map', as load density peaks. Thirdly, the spatial distribution dataset of the low-voltage users is taken as the data points in the clustering algorithm; the result of clusters corresponds to division in the power distribution with certain capability. Consequently, the methodologies proposed are verified on one example district of ~71.7587 hectares, the division scheme can provide theoretical guidance for the location and sizing of power distributors in the urban distribution network.