Identification and Stabilization of Critical Clusters in Inverter-Based Microgrids

TL;DR

This paper introduces a novel stability assessment method for inverter-based microgrids using spectral analysis of the weighted admittance matrix to identify critical clusters affecting system stability.

Contribution

It proposes a spectral clustering approach that links eigenvalues and eigenvectors of the admittance matrix to system stability and critical inverter groups, providing a stability certificate based on grid parameters.

Findings

Eigenvalues of the admittance matrix correlate with unstable clusters.

The stability threshold depends only on the grid's R/X ratio, not topology.

The method guides adjustments to inverter settings for stability improvement.

Abstract

A new method for stability assessment of inverter-based microgrids is presented in this paper. It leverages the notion of critical clusters -- a localized group of inverters with parameters having the highest impact on the system stability. The spectrum of the weighted network admittance matrix is proposed to decompose a system into clusters and rank them based on their distances from the stability boundary. We show that each distinct eigenvalue of this matrix is associated with one cluster, and its eigenvectors reveal a set of inverters that participate most in the corresponding cluster. The least stable or unstable clusters correspond to higher values of respective eigenvalues of the weighted admittance matrix. We also establish an upper threshold for eigenvalues that determines the stability boundary of the entire system and demonstrate that this value depends only on the grid type (i.e. $R/X$ ratio of the network) and does not depend on the grid topology. Therefore, the proposed method provides the stability certificate based on this upper threshold and identifies the lines or inverter droop settings needed to be adjusted to restore or improve the stability.