Research on Building Cluster Identification Based on Two-stage Clustering and BP Neural Network

Abstract

As a common demand resource in urban environment, buildings have huge demand response potential. In order to accurately grasp the power consumption characteristics and categories of demand resources during the implementation of demand response, consider the peak cutting and valley filling capacity of demand side building resources and cluster them for identification. Combining the advantages of system clustering method and fuzzy clustering method, the two-stage clustering method is introduced to extract the corresponding load characteristic parameters, classify the demand side resources from the category dimension and obtain the category tag value, and use neural network to identify the load of unknown categories collected subsequently. The results show that the algorithm is simple and effective, and can select potential users who participate in peak shaving and valley filling of demand response, and can assist the implementation of demand response of urban resources in many aspects, such as user selection, response capability evaluation, etc.