Predictive Energy Management Based on Solar Forecast and Applied to a University Campus
Abstract
Energy Management Systems (EMS) are essential for the optimal operation of energy systems powered by solar renewables like photovoltaics (PV). They increase local renewable energy use while ensuring affordable energy prices for end users. Predictive EMS, based on solar forecasts, promise to be more efficient than ruled-based EMS commonly implemented in operational systems such as microgrids. However, the selection and integration of the most relevant solar forecasts in the optimization process of EMS are rarely addressed in the literature. This work proposes to compare different EMS and state-of-the-art solar forecasts in order to minimize the operating cost of a campus microgrid. In the case study considered, the predictive EMS achieves a microgrid operating cost 10% lower than that of a standard rule-based EMS. The higher the quality of the forecasts used as input of the predictive EMS, the lower the operating cost. However, with a microgrid self-sufficiency of 84%, the rule-based EMS offers a good compromise between technical and economic performance. These results will provide useful information to microgrid managers for the development and operation of their systems.
Keywords
microgrid, ruled-base and predictive EMS, solar forecasts, deterministic, probabilistic