Method for Estimating the Composition Ratio of Substation Industry Based on K-Means++ Clustering and Sparse Coding

In order to match and fuse the information collected by energy management system and electricity consumption information collection system, a substation industry composition ratio estimation method based on sparse coding and K-means++ clustering is proposed to achieve the purpose of estimating the load industry composition ratio based on energy management system data and past electricity consumption information collection system data in the absence of electricity consumption information collection system data. Taking the actual grid data as an example, the K-means++ algorithm is firstly used to cluster the bottom-level customer electricity consumption data classified by industry to obtain the typical daily load curve of each industry; then, based on the clustering results, the electricity consumption data of upper-level substations are decomposed; finally, the industry composition ratio of upper-level substations is calculated based on the decomposition results. The results show that the estimated industry composition ratio of the upper substation is within the acceptable range of error from the actual load situation.