Abstract
Our traditional electricity system is experiencing a drastic change as microgrids are fast spreading due to their inherent advantages of energy resilience, prosperity, and sustainability for the communities. The microgrid concept can be defined as the integration of distributed energy sources, energy storage systems and controllable loads into localized energy systems. However, non-linearities in the operation of microgrids, variations in load demand, and uncertainties in power generation from renewable energy sources pose significant challenges to determine the optimal microgrid operation planning. Here, timely planning, analysis, and control of all the components in a microgrid play a crucial role to achieve the resilient, sustainable, and secure smart energy infrastructure. To this end, a dynamic data driven operation control and optimization framework is introduced for addressing significant challenges in operation planning and design of microgrids, the economic and environmental unit commitment and load dispatch problems, and timely monitoring and adaptation of microgrid operations to real-time system fluctuations and contingencies to increase the power network resilience and energy surety. The proposed framework incorporates new and advanced optimization models and algorithms for operation and control including a comprehensive optimization model and a decomposition algorithm for the operation of off-grid AC and DC microgrids, and multi-scale adaptive simulation models. Initially, a dynamic data driven multi-objective optimization framework was developed for the economic and environmental (near) real-time load dispatching considering demand side management decisions for DC microgrids. The framework has been tested and validated via a synthetic microgrid and as the numerical analysis revealed, it is capable of adopting the uncertainties in load demand by minimizing the operation cost and GHG gas emission. In addition to DC microgrids, a comprehensive microgrid operation models and advanced decomposition algorithm to solve the models are proposed for radial and mesh structured AC microgrids. Here, a stochastic network constrained AC microgrid unit commitment problem is first modeled. This model is the first model that presents a mixed-integer linear programming formulation for the network-constrained AC unit commitment problem including the energy storage systems and then proposes an efficient two-stage Benders’ decomposition approach to solve this problem under load and solar power uncertainty. The performance of the algorithm is demonstrated through the numerical results from two case studies on the IEEE-18, and IEEE33 radial test systems. Moreover, a new comprehensive security-constrained operation planning topology reconfiguration optimization model and a computationally efficient decomposition algorithm to solve the developed optimization model for the operation of off-grid AC microgrids are introduced. To the best of our knowledge, this model is the first in the literature that integrates network security constraints, branch flow equations, and transmission switching decisions into a multi-period operation planning problem. The capabilities and performance of the proposed approaches are tested on IEEE-9, IEEE-30, and IEEE-118 test systems. The numerical analysis has shown quite promising results in optimizing the design and operation planning such that the algorithm provides a (near-) optimal solution in a few minutes. Lastly, multi-fidelity simulation models are designed for resilient back-up power smart energy systems. Here, a low-fidelity battery load simulation model is developed for the estimation of the system states and assessment of potential power mismatch in power supply and demand while a high-fidelity component level simulation model is used for the analysis of power losses and inefficiencies in the system. The simulation models were illustrated and validated via data obtained from the City of Coral Gables and the experiments indicated, it is quite capable of determine the battery and PV capacities that are subject to uncertainties in the power supply and demand, and assessing of potential power mismatch in a smart microgrid by computing the amp-hour capacities of batteries in worst case scenarios.