Economic predictive control for isolated microgrids based on real world demand/renewable energy data and forecast errors

Abstract

In this work, the operation of microgrids is studied, using real data for demand and renewable energy sources. A mixed integer nonlinear program to operate the microgrid is proposed, following a model predictive control methodology, which allows to enhance the economic performance by means of a predictive strategy. A comparison with other techniques without predictive feature nor forecast mismatches is made, showing that our method outperforms them by adapting the current control decision to future costly issues. The case study shows savings of more than 10%, analysing the qualitative aspects of the proposed strategy.