Distributed Real-Time Energy Management of Multi-microgrid for Incomplete Information Scenarios

The interconnected microgrids in a certain area would meet power supply demands in the area through internal energy transactions. Due to the uncertainty of renewable energy output, forecast error will affect the economics of operation. This paper develops a distributed real-time dispatching strategy for multi-microgrid to avoid the large error from a long time forecast. Aiming at main difficulties of real-time scheduling of energy storage systems, a variable cost model is proposed. Each microgrid exchanges information of tradable energy and trading price with its neighboring microgrid and optimizes output of local components and adjusts the selling price to minimize the local cost. A modified alternating direction method of multipliers (ADMM) algorithm is presented for the optimization problem. Numerical results of a 3-MG system demonstrate the advantage of this dispatching strategy in incomplete information scenarios which including inaccurate predictions and communication interruptions.