A Novel Optimal Scheduling Framework for Hybrid Microgrids Based on Alternating Direction Method of Multipliers

Abstract

Optimal operation and management of hybrid AC–DC microgrids are challenging issues due to the in-equivalent voltage level/type as well as heterogeneous distributed generation and load types in the AC and DC parts. Such a structure requires a powerful and distributed optimal scheduling framework to avoid central control mechanism used in the conventional microgrids. In order to get into this concept, this article investigates the optimal operation and management of hybrid AC–DC microgrids in a distributed multi-agent framework. The proposed framework first splits the hybrid microgrid into several agents and then uses a distributed formulation to solve the optimal scheduling problem. Different agents can exchange limited information with each other until the consensus happens through several iterations. Using the alternating direction method of multipliers, it is proved that the final consensus operating point is the global optimal solution for the hybrid microgrid cost operation. Due to the high-nonlinearity and -nonconvex characteristics of the augmented Lagrangian formed in the proposed framework, a new optimization algorithm based on modified bat algorithm (MBA) is also proposed to solve the proposed augmented function. The proposed MBA is strengthened by a new modification method to escape from the many local optima and avoid premature convergence. In order to model the uncertainty effects, point estimate method of 2m-scheme is employed to model the forecast error in wind turbine and photovoltaics output power, load demand and the market price. The feasibility and satisfying performance of the proposed method are examined on an IEEE test.