Distributed optimization for scheduling energy flows in community microgrids

With the increasing development of grid-connected microgrids predominantly powered by renewable energy sources, their negative impact on the distribution grid cannot be ignored. Whilst this burden is borne by the distribution system operator (DSO), microgrid-users can contribute in grid congestion management to maintain a stable grid connection by providing flexibility on the DSO's request. This paper uses Jacobi-alternating direction method of multipliers to optimize power exchange between a microgrid and the grid to assist in congestion management. The algorithm decomposes the optimization problem into sub-problems solved locally and in parallel using fitted Q-iteration. The local optimization plans the operation of heat pumps and batteries to provide the required flexibility. The performance of the proposed framework is evaluated using real-world data from thirty residential prosumers. Simulation results show that solving the sub-problems with fitted Q-iteration leads to feasible control policies within acceptable computation times while providing the required flexibility for grid congestion management.