Robust Optimal Microgrid Management using a Virtual Battery System for Application to Unit Commitments

A virtual battery (VB) model is proposed in this paper to facilitate in the management of a smart microgrid. A cluster of controllable load and generations (CLGs) is referred to as VB, and it operates similarly as the battery storage that reduces the size of the energy storage device (ESS). Aggregators use the VB to distribute power among CLGs when the grid has excess or shortage capacity. The management of a microgrid with renewable energy instability and VB is a difficult problem; as a result, this proposed study introduces adaptive robust optimization (ARO) for unit commitment (UC) problems in microgrids that takes VB into account. Three separate UC cases are depicted in the proposed scheme. Case 1 treats only optimization and deterministic UC problem for the day ahead forecasted load; Case 2 discloses the UC with treating robust optimization, which considers forested reference day ahead load utilizing robust constraints. Case 3 emphasizes real-time load considering the worst uncertainty/prediction error of renewables and virtual battery for polishing the actual load. A neural network is adopted to forecast variable renewable generations (VRGs) output power. MATLAB
^®
simulation with Mixed Integer Linear Programming (MILP) has been utilized for solving the ARO UC problem with VB.