The rapid proliferation and widespread adoption of microgrids (MG) necessitate the
development of new methodologies to holistically model all the active components
within MGs. It’s crucial to focus on specific islanding requirements, especially when
the primary grid power is unavailable. In order to ensure a high level of reliability
in an interconnected MG network, this dissertation presents an optimal scheduling
model designed to minimize the day-ahead costs of the MGs while taking into account
the existing operational constraints.
This problem is thoughtfully decomposed using Bender’s Decomposition method
into two key operating conditions: grid-connected and resilient operations. The ultimate
goal is to ensure that each MG within the network maintains sufficient online
capacity in the event of an emergency islanding situation, such as during extreme
weather events. These events often come with uncertainties regarding their timing
and duration, necessitating the consideration of multiple potential islanding scenarios
for each event.
The primary objective of this thesis is to establish optimal scheduling that guarantees
the feasibility of islanding for all conceivable scenarios of such events, with
load shedding as a last resort. The optimization model has been put into practice
across different layouts of the modified IEEE 123-bus test system, encompassing various
events over a 24-hour period. In addition to proposing a day-ahead scheduling
approach oriented towards resiliency for multiple MGs, a comprehensive cost analysis
and comparisons among all the test cases are also offered. The results convincingly
demonstrate the utility of the proposed day-ahead scheduling algorithm, particularly
for MG owners looking to foster collaborations with neighboring MGs. Lastly, after
iv
comparing with the traditional Single Stage MILP approach, the proposed method
has proven to be computationally faster for practical usage. It has been shown that
decomposing the problem using the proposed model makes it possible to combat real
life events with thousand scenarios, where the single stage approach may fail.