Double-Layer K-Means++ Clustering Method for Evaluation of Dispatchable Potential of Massive Regional 5G Base Stations

Abstract

5G base stations (BSs), which are the essential parts of the 5G network, are important user-side flexible resources in demand response (DR) for electric power system. However, a 5G BS has little and difference dispatchable potential, how to make massive 5G BSs participate in DR conveniently is an urgent problem to be solved. Clustering is an effective solution. Aiming at the special requirements of big data analysis and dispatching difficulties brought by the massive and ultra-dense distribution of 5G BSs, this paper proposes a double-layer K-means++ clustering method for regional 5G BSs considering the main characteristics of 5G BSs. The proposed clustering method considers both the geographical location and power consumption characteristics of the regional 5G BSs to partition the 5G BSs into appropriate number of clusters. Besides, the comprehensive charging and discharging dispatchable potential evaluation indices are established based on the clustering results to evaluate the dispatchable potential of each cluster. The effectiveness of the proposed method and the dispatchable potential evaluation indices are verified through a case of 5000 5G BSs in Jiaxing, China. The simulation findings reveal that the performance of the proposed clustering method is better than classical clustering algorithm and the dispatchable potential of the 5G BSs clusters can be accurately analyzed by the proposed comprehensive indices, which lay the foundation for friendly and fine interaction between 5G BSs and the power grid.

Keywords

• 5G base station
• K-means++ clustering
• dispatchable potential evaluation
• big data analysis
• demand response