Online distributed neurodynamic optimization for energy management of renewable energy grids

An online distributed neurodynamic optimization-based energy management method is proposed to accommodate the integration of intermittent renewable energy in smart grids. The proposed approach takes advantage of the online prediction method to deal with uncertainties from renewable energy generators, the distributed consensus algorithm for information exchange, and neurodynamic optimization to manage coupling constraints. The proposed distributed optimization neurodynamic approach utilizes a one-layer neural recurrent network without auxiliary variables and enables parallel computation, significantly alleviating the data calculation burden compared with the traditional centralized methods and requiring no auxiliary variables, in contrast to most of the existing distributed methods. The convergence, optimality and robustness to communication failures of the proposed method are verified by various case studies with a modified IEEE 33-bus distribution system and a modified IEEE 123-bus distribution system.