A Modelica-based framework for modeling and optimization of microgrids

Abstract: Microgrids have lately drawn much attention due to their considerable financial benefits and the increasing concerns about environmental issues. A solution that can address different engineering problems - from design to operation - is desired for practical reasons and to ensure consistency of the analyses. In this thesis, the capabilities of a Modelicabased framework is investigated for various microgrid optimization problems. Various sizing and scheduling problems are successfully formulated and optimized using nonlinear and physical component models, covering both electrical and thermal domains. Another focus of the thesis is to test the optimization platform when varying the problem formulation; performance and robustness tests have been performed with different boundary conditions and system setups. The results show that the technology can effectively handle complex scheduling strategies such as Model Predictive Control and Demand Charge Management. In sizing problems, although the platform can efficiently size the components while simultaneously solving for the economical load dispatch for short horizons (weekly or monthly), the implemented approach would require adaptations to become efficient on longer horizons (yearly).