Detailed Small-Signal Stability Analysis of the Cigr\'e High-Voltage Network Penetrated by Grid-Following Inverter-Based Resources

TL;DR

This paper conducts a detailed small-signal stability analysis of a modified Cigré high-voltage network with an inverter-based resource, introducing an adaptive sampling method and SVM classifier to efficiently estimate stability regions.

Contribution

It introduces an adaptive sampling approach and SVM classifier to accurately predict stability regions in complex power networks with inverter-based resources.

Findings

Stability regions are conservatively estimated by The9venin equivalents.

The adaptive sampling improves the accuracy of stability boundary detection.

Full network analysis provides more precise stability insights than simplified models.

Abstract

This paper presents a detailed small-signal stability analysis of a modified version of the Cigr\'e European high-voltage network, where one of the synchronous generators is replaced by a grid-following inverter-based resource (IBR). The analysis focuses on the influence of the parameters defining the grid-following IBR control scheme on the stability of the system. Given a set of potential grid configurations and the value of the IBR control parameters, stability is verified by the direct eigenvalue analysis of a high-detailed linearized model of the overall Cigr\'e network. Starting from this procedure, we propose an adaptive sampling method for training a support vector machine classifier able to estimate the probability of stability of the power system over a domain defined by candidate intervals of the considered parameters. The training of the classifier is refined to identify with more accuracy the boundaries of the parameters' stability regions. The obtained results are then compared with those obtained by representing the grid with the classical Th\'evenin equivalent. Results suggest that, when the Th\'evenin equivalent is accurate, the predicted stability region is conservative yet contained within that of the full network.