Answers are grounded in this paper and cite their sources.
The transition from passive distribution systems to actively managed microgrids requires planning frameworks that account for islanded operation under network contingencies. In radial microgrids with inverter-interfaced, droop-controlled distributed generation units (DGs), a single-feeder outage partitions the system into two islands with different generation and load compositions. Under these conditions, DG allocation and sizing, combined with priority-driven load shedding, determine the loadability and curtailment needed to satisfy voltage, frequency, and line-current limits. This paper develops a loadability-driven planning framework for islanded radial microgrids that minimizes the total priority-weighted load shed across all single-feeder contingencies. The planning problem is formulated as a double-stage mixed-integer nonlinear program (MINLP). The outer stage selects DG locations and discrete ratings under a fixed capacity budget using a bound-guided search, while the inner stage determines priority-level loadability factors for each island through deterministic droop-based nonlinear power-flow (NLPF) evaluations subject to DG capability, voltage, frequency, and line-current constraints, with voltage- and frequency-dependent loads. On the IEEE 33-bus microgrid, the selected layout fully serves high-priority loads in all single-feeder contingencies and requires shedding in only two contingencies, with a weighted shedding objective of 0.2013 pu. Under a common NLPF evaluation, it reduces weighted shedding by 91% relative to a layout selected by mixed-integer second-order cone programming (MISOCP), while maintaining operational constraints on voltages, frequency, and line ampacity. Additional studies on droop representation, installed capacity, DG spatial distribution, load prioritization, loss-oriented planning, DG firm-capacity availability, and 141-bus scalability examine the effectiveness and practicality of the proposed contingency-driven planning framework.
No takes yet. Share an insight, caveat, or question.
Shabar et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: