Supervisory control design for microgrids energy management optimization based on renewable generation and consumption forecasting
Document typeMaster thesis
Rights accessOpen Access
Solar-based electricity production has become an essential part of the general energy production in the recent years with the will to use more renewable sources. The one issue that appears is the uncertainty of the solar irradiation. It is then more complicated to predict the energy generated in the future times. The Energy Management System used on the grid schedules the energy exchanges between the devices based on the prediction of the state of the system in the next time interval. The Model Predictive Control forecasts the power produced as well as that of the energy demand from the load and defines the state of the system. In order to minimize the corresponding cost function, this forecast should be as accurate as possible, with the minimum prediction error. To address these forecasting needs, we will extract some data from a database using an algorithm directly connected to the server. And we will compute the remaining values using an accurate forecasting method, the Simple Average. Then, for this information to be even more precise, we use the Rolling Horizon approach, that enables a regular updating of the forecast. Simulation results and experiments confirm the influence of some parameters on the prediction error and hence on the cost function.