Accelerated Distributed Hybrid Stochastic/Robust Energy Management of Smart Grids

The uncertainties of renewable energy, load and electricity price pose significant challenges to the economical and secure energy management of smart grids. In this paper, a hybrid stochastic/robust (HSR) optimization method is developed to minimize the overall cost of all units. The proposed approach takes advantage of stochastic programming, robust optimization and distributed optimization methods while considering various system constraints. First, electricity price scenarios are selected by the Latin hypercube sampling method. Second, the uncertainties of renewable energy generation and loads are managed by the proposed robust optimization method under each price scenario. Then, an improved distributed optimization method is proposed to solve the formulated problem, which considerably enhances convergence with the accelerated gradient method. Numerical case studies on both small-scale and large-scale power systems demonstrate the accuracy, effectiveness and scalability of the proposed distributed HSR approach.