Research on load management strategies based on Bi-kmeans clustering algorithm

Aiming at the current problem that the formulation of orderly power consumption program does not fully consider the differences in user load characteristics, a load management strategy based on load time series clustering is proposed. Firstly, the typical load curve pattern of a single user is extracted; then the load time series data of different users are clustered by Bikmeans clustering algorithm, so as to divide multiple users into different types of electricity consumption; finally, the load characteristics of different types of users are formulated according to their load characteristics. Finally, load management strategies are formulated according to the load characteristics of different types of users. The actual load data of a regional power grid is analyzed as an example. The results show that the proposed method can more accurately categorize users with different load characteristics than the traditional method, thus more effectively guiding the orderly power consumption management strategy.