A genetic algorithm approach for the identification of microgrids partitioning into distribution networks

In this paper a Genetic Algorithm (GA) is used to partition a distribution network with the aim to minimize the energy exchange among the microgrids (i.e. maximize self-consumption) in presence of distributed generation. The proposed GA is tested on the IEEE prototypical network PG & E 69-bus. The microgrid partitioning is tested over a period of one year with hourly sampled data of real household consumption and real distributed generation data. The proposed GA approach is compared with a Tabu Search (TS) method already presented in the scientific literature. Results show that both GA and TS lead to the identification of equivalent microgrids. However, the GA based approach achieves better convergence results allowing for a reliable network partitioning with less CPU effort. Moreover, the histograms of the power unbalances of the microgrids show unimodal and skewed distributions offering an interesting starting point for the appropriate deployment of storage and control systems.