Multi-time Scale Optimal Dispatch for AC/DC Distribution Networks Based on Cluster Partition

A multi-time scale optimal dispatch model for AC/DC distribution networks based on cluster partition is proposed to deal with the increasing scale of variable renewable energy. Firstly, a novel concept of improved electrical distance is utilized as the measurement of the clustering algorithm, then AC/DC distribution networks is divided into several cluster using the spectral clustering method, which is the basis of multi-time scale optimization. Finally, the multi-time scale optimal dispatch is carried out. In the long-time scale optimization, the scheduling result is solved by minimizing the value of operation cost. In the short-timescale optimization, rolling optimization is adopted based on the result of long-time scale optimization to compute the output power of distributed generation in each node within each cluster. The effectiveness of the proposed model is verified by a 50-node system with distributed generation.