A Stochastic-Robust Approach for Resilient Microgrid Investment Planning Under Static and Transient Islanding Security Constraints

TL;DR

This paper introduces a hybrid stochastic-robust model for microgrid investment planning that accounts for static and transient islanding security constraints, enhancing resilience against both rare catastrophic and frequent operational uncertainties.

Contribution

It develops a novel three-stage solution for a complex min-max-min problem incorporating dynamic transient response constraints in microgrid planning.

Findings

Model effectively captures HILF and LIHF uncertainties.

Proposed algorithm successfully solves the complex optimization problem.

Application on CIGRE network demonstrates practical viability.

Abstract

When planning the investment in Microgrids (MGs), usually static security constraints are included to ensure their resilience and ability to operate in islanded mode. However, unscheduled islanding events may trigger cascading disconnections of Distributed Energy Resources (DERs) inside the MG due to the transient response, leading to a partial or full loss of load. In this paper, a min-max-min, hybrid, stochastic-robust investment planning model is proposed to obtain a resilient MG considering both High-Impact-Low-Frequency (HILF) and Low-Impact-High-Frequency (LIHF) uncertainties. The HILF uncertainty pertains to the unscheduled islanding of the MG after a disastrous event, and the LIHF uncertainty relates to correlated loads and DER generation, characterized by a set of scenarios. The MG resilience under both types of uncertainty is ensured by incorporating static and transient islanding constraints into the proposed investment model. The inclusion of transient response constraints leads to a min-max-min problem with a non-linear dynamic frequency response model that cannot be solved directly by available optimization tools. Thus, in this paper, a three-stage solution approach is proposed to find the optimal investment plan. The performance of the proposed algorithm is tested on the CIGRE 18-node distribution network.