A distributed newton method for optimal operation of microgrid clusters

Integrated together in a self-organized manner, a cluster of microgrids can be connected to the active distribution grid as an autonomous entity, which significantly enhances reliability and reduces operational cost. In this paper, we present a fully distributed method for optimal operation of microgrid clusters without central coordination. Based on peer-to-peer communication, microgrids need only exchange information between neighbors over a sparse communication network and hence the communication burden is very light. By combining the Newton-type second-order algorithm and a consensus-based information exchange, the method can achieve minimum operation cost through a few iterations, and is very resistant to contingencies. Simulation results indicate that the method has faster convergence rate than those subgradient based algorithms. Besides, influence of communication network is also discussed in numerical tests, which demonstrate the method is more reliable, scalable and flexible than centralized schemes.