Improved spatiotemporal decomposition algorithm for real-time decentralized optimal scheduling of MG with energy storage

The real-time optimal scheduling of microgrid (MG) with energy storage is usually formulated as a nonlinear problem. In order to solve it while safeguarding the information privacy and decision independence of distributed generators, an improved spatiotemporal decomposition (STD) algorithm is proposed in this paper. Under the framework of STD, the original problem is decomposed into multiple subproblems in both spatial and temporal dimensions, wherein value functions are used to describe the interactions among subproblems. Compared with the existing STD algorithm, the proposed improved spatiotemporal decomposition (ISTD) algorithm extends the piecewise approximate value function from the linear form to the quadratic form, which helps to describe the nonlinear characteristics among different decomposed subproblems. Different from the existing STD algorithms, which needs to call commercial solvers repeatedly to solve a number of subproblems, the proposed ISTD algorithm directly obtains their closed-form solutions according to the algebraic expressions. Such that the algorithm can be substantially accelerated, thereby facilitating the real-time implementation. Numerical simulations in different MG systems validate the efficacy of the proposed algorithm.