Optimal planning of microgrids for resilient distribution networks

As severe weather events disrupt the power system more frequently and more harshly, the concern is growing around the ability of future grids to recover from such natural disasters. Recently, a major research focus has been on microgrids (MGs) as a potential source of resiliency. While most of the works done so far center on how to benefit from existing MGs through operation schemes, this study focuses on the planning of MGs to strengthen the network against severe faults. In this regard, three solution approaches are proposed aiming to determine the optimal nodes for the connection of MGs as well as the capacity of the dispatchable generation units deployed within MGs. These algorithms satisfy the power balance of MGs and the main grid as well as the operational and topological constraints. A computationally-efficient heuristic method is developed in two stationary (S-HM) and time-dependent (T-HM) versions. The concept of the heuristic approach, which was first introduced by the authors and is matured in this study, is based on a multi-stage search algorithm that efficiently reduces the undesirable restoration strategies and utilizes the original power flow equations. The other approach is a multiobjective mixed-integer linear programming (MO-MILP) that strives to find the globally-optimal solution in a time-dependent scheme. The validity of the outputs of these methods is assessed using an exhaustive search algorithm (ESA), capable of finding the globally-optimal solution. The MG model constitutes renewable and dispatchable generation units, energy storage systems, and local loads. The uncertainty of intermittent energy resources is tackled through robust optimization formulation based on the worst-case scenario. The performance of the proposed methods are evaluated by the IEEE 37and IEEE 123-bus test systems under several severe fault scenarios.