Optimal day-ahead scheduling of a hybrid electric grid using weather forecasts

Abstract

The compromise between the stability of a hybrid electric grid (HEG) and the total operating cost can be reached by accurately anticipating the future renewable power productions. This thesis suggests the use of weather forecasts to establish day-ahead operating schedules for a grid that include the operating plan of dispatchable fuel-based generators, the charge or discharge of energy storage units, and the energy to exchange with the commercial grid if the configuration of the HEG allows it. The weather forecasts used as a key factor to establish the optimal plan are subject to uncertainty. In order to mitigate this problem, multiple weather forecast scenarios are used in the optimization. This thesis alters the optimization model to represent various configurations of the HEG and optimizes over a variety of weather forecasts. It then tests the operating plans suggested by the model using particular weather scenarios representing actual observed weather conditions. Finally, this thesis gives an illustration of how to run the optimization model with the rolling horizon method using updates of weather forecasts.