Decomposed Stochastic Model Predictive Control for Optimal Dispatch of Storage and Generation

Abstract

This paper proposes a stochastic model predictive control (SMPC) approach to optimally dispatch energy storage and dispatchable generation in the electric energy system under uncertainties introduced by variable energy sources as well as demand. The objective is to minimize the expectation of the sum of the production and ramping costs for generators while satisfying all the system constraints. The uncertainties are represented by scenarios, resulting in a large-scale and computationally demanding optimization problem. We use the optimality condition decomposition (OCD) to decompose the SMPC problem into subproblems which can be solved in parallel thereby reducing the computation time. Both a scenario-based decomposition and a temporal-based decomposition are formulated and numerically evaluated in terms of speed and convergence using a modified IEEE 39-bus test system. Simulation results indicate that the scenario-based decomposition scheme achieves a better trade-off between convergence speed and subproblem size for the considered optimal dispatch problem.