Scalable enumeration approach for maximizing hosting capacity of distributed generation

Abstract

At the stage of planning distributed generation (DG) for a distribution network, the network configuration is a key factor in increasing the DG hosting capacity. The determination of a configuration that maximizes the hosting capacity is a highly complex, nonlinear combinatorial optimization problem. No existing method can yield the global optimal solution for practical-scale networks. Therefore, this paper proposes a scalable optimization method. Specifically, the proposed method enumerates all optimal configurations while simultaneously considering optimal DG placement. The proposed method first optimizes the DG placement for possible partial networks using a second-order cone programming technique. Next, it enumerates possible combinations of the partial networks while avoiding a combinatorial explosion using a highly compressed data structure. Finally, it finds the optimal configurations by exploring solutions over the data structure. In experiments involving a large-scale network containing 235 switches, our enumeration method obtained 1.49×1018 global optimal configurations in 17.1 h. Another powerful feature of our method is that it enables distribution system operators to select the preferred optimal configuration interactively.