A load clustering algorithm based on discrete wavelet transform and fuzzy K-modes

In order to study the power consumption modes of users under the background of smart grid, a fuzzy K-modes clustering algorithm based on discrete wavelet transform is proposed considering the deficiencies of existing clustering algorithms. The load curves in the time domain are converted to the frequency domain by the discrete wavelet transform, so that the different features of load curve can be isolated at different frequency domain levels. The effective component curves of the primitive curve are selected by the idea of lower order approximation. The selected component curves are coded and the continuous load data are translated into discrete attribute data. The initial clustering condition is determined based on average density and the shapes of curves are clustered by the fuzzy K-modes clustering algorithm, based on which, the load curve forms are obtained. The effectiveness of the proposed algorithm is verified by comparing it with the traditional K-means algorithm and the hierarchical clustering algorithm. © 2019, Electric Power Automation Equipment Press. All right reserved.