With the continuous expansion of the power grid scale, how to analyze the behavior of power users and intelligent clustering methods has become an urgent problem to be solved by power grid companies. In order to solve the problems described above, the traditional DTW algorithm is first optimized and improved with the help of similarity algorithm, so that it meets the requirements of power user behavior analysis. After that, the knowledge graph is used to preprocess the power user behavior, store it in the form of dataset, and realize the intelligent clustering of power user behavior through the clustering analysis of Gaussian mixture model. Build the experimental environment, set the comparison algorithm and evaluation indexes, and use the data analysis software to verify and analyze the intelligent clustering scheme of power user behavior based on KGEG algorithm. In the data set A~J, the data of the three indicators of this paper’s algorithm is much better than the other three comparison algorithms, and the distribution of the data of the three indicators is in the range of 0.6~0.9, which confirms the application effectiveness of this paper’s user behavioral intelligent subgrouping research program, so as to improve the level of development of the smart grid.
