Optimal Management of Microgrid Under PV Forecast Uncertainty Using Stochastic Programming

Probabilistic PV forecasting plays an important role in microgrid management to improve the predictive control process. The first contributions of this paper are the development of a method to obtain empirical distribution from PV probabilistic forecast and the determination of the best copula type to model temporal correlation between PV production random variables. Then a procedure to get reduced scenario sets with the least loss of information is described. Finally, an optimal planning method with an ancillary service reducing daily amplitude grid withdrawal and performing peak shaving based on stochastic programming is proposed. This approach is compared with the determinist method and the results show the performance of the proposed solution.