DYNAMIC CONTROL AND OPTIMIZATION OF DISTRIBUTED RESOURCES IN MICROGRID

Abstract

Existing economic dispatch mechanisms fails to operate e ectively in case of microgrid with high penetration of renewable energy resources (RESs). An intelligent control technique is required that can tackle both variability and unpredictability of RESs while satisfying time varying load demands without violating operation constraints. In this paper, we propose a MPC scheme for islanded microgrid for hourly economic dispatch of generators, storage units and RESs. MPC, subjected to constraints and forecasts, aims at minimizing running cost of microgrid along with rewarding renewable power infeed. In order to have higher forecasting accuracy, SVM approach is adopted for day ahead hourly load forecasting. Multi-layer perceptron based neural network is used for PV array output forecast. System is modelled using MILP. MPC uses load and solar output forecast data and solves optimal control problem subjected to constraints and generates optimal power dispatch plan at each time instant.First generated sequence is applied and horizon is shifted to next time instant after executing the rst step of the previously determined schedule thereby dynamically adjusting and self-correcting itself for future time steps. The results depict the e ectiveness of this metho