Research on Power Behavior Analysis Based on Clustering

This research is based on data clustering to analyze power users' electricity consumption behavior, and analyzes the load curve to obtain the users' electricity consumption characteristics and classify users according to the characteristics of electricity consumption behavior. It is of far-reaching significance to the power industry and socio-economic development. This paper introduces the principle and flow of the main clustering algorithm K-means, fuzzy clustering and neural network clustering algorithm. The data preprocessing method is given. The clustering algorithm and the optimal clustering number are determined by defining the clustering volatility and clustering accuracy index. Then the power consumption behavior of the user is analyzed through the load curves and characteristic parameters obtained through clustering.

[1]  James C. Bezdek,et al.  Fuzzy c-means clustering of incomplete data , 2001, IEEE Trans. Syst. Man Cybern. Part B.

[2]  J. A. Hartigan,et al.  A k-means clustering algorithm , 1979 .

[3]  Teuvo Kohonen,et al.  The self-organizing map , 1990, Neurocomputing.

[4]  G. Chicco,et al.  Support Vector Clustering of Electrical Load Pattern Data , 2009, IEEE Transactions on Power Systems.

[5]  Habtom W. Ressom,et al.  Adaptive double self-organizing maps for clustering gene expression profiles , 2003, Neural Networks.

[6]  J. Bezdek,et al.  FCM: The fuzzy c-means clustering algorithm , 1984 .

[7]  Erzsébet Merényi,et al.  Exploiting Data Topology in Visualization and Clustering of Self-Organizing Maps , 2009, IEEE Transactions on Neural Networks.