Dynamic optimal management of a hybrid microgrid based on weather forecasts

Hybrid microgrids containing both renewable and conventional power sources are becoming increasingly attractive for a variety of reasons. However, intermittency of renewable power production and uncertainty in future load prediction increase risks of electric grid instability and, by consequence, restrict the portion of renewable power production in microgrids. In order, to prefigure the upcoming renewable power production, particularly, wind power and photovoltaic power, we suggest using weather forecasts. In addition to illustrating short term renewable power prediction based on ensemble weather forecasts, this paper focuses on optimizing the management of distributed power generation, power storage, and power exchange with the commercial electric grid. Establishing an optimal operating plan for the grid makes the integration of the predicted renewable production more efficient. The optimization problem is formulated as an integer linear program. The operating plan is initially optimized over the upcoming day and then regularly updated and incremented to cover a longer time horizon. After analyzing the obtained plans, we test and evaluate them relative to the observed weather conditions. Finally, we investigate some practical problems that had arisen and we apply a time cascade optimization technique to mitigate the effects of initial conditions and end-of-horizon effects, as well as to take advantage of updated forecasts.