Forecast-based optimal operation of islanded microgrids

Abstract

In this thesis, we provide an introduction to microgrids and associated concepts. We look at the central challenges to their operational management, and provide a discussion on the classi?cation of microgrids based on a number of characteristics and how this influences the planning process. In the second part, we describe a forecast-based centralized planning approach for the (economically) optimal operational management of (islanded) microgrids. Using mathematical programming techniques, we formulate four models: (1) a simple MILP-formulation of our problem, (2) an extension adding safety constraints, (3) a scenario-based two-stage model and (4) a scenario-based multi-stage model. Additionally, we describe an algorithm for the generation of the scenario tree needed for the fourth model. The performance of the models is compared on three test instances: two are based on an existing microgrid in the Netherlands, and one is arti?ficially constructed. The economic advantage of the described planning approach is not unequivocally shown for all described instances, but it is shown to have its advantages.