Transient-Stability-Aware Frequency Provision in IBR-Rich Grids via Information Gap Decision Theory and Deep Learning

TL;DR

This paper presents a novel framework combining deep learning and information gap decision theory to enhance transient stability and frequency provision in inverter-rich power grids, proactively preventing system collapse under worst-case scenarios.

Contribution

It introduces a risk-averse dispatch strategy that uses early post-fault predictions to improve virtual inertia scheduling in high-IBR grids, a novel integration of DL and IGDT.

Findings

Prevents system collapse in high-IBR scenarios.

Ensures frequency stability with only 5% cost increase.

Validated on IEEE 39-bus system with 70% IBR penetration.

Abstract

This paper introduces a framework to address the critical loss of transient stability caused by reduced inertia in grids with high inverter-based resource (IBR) penetration. The proposed method integrates a predictive deep learning (DL) model with information gap decision theory (IGDT) to create a risk-averse dispatch strategy. By reformulating the conventional virtual inertia scheduling (VIS) problem, the framework uses early predictions of post-fault dynamics to proactively redispatch resources, ensuring the system's center of inertia remains stable under worst-case contingencies. Validated on the IEEE 39-bus system with 70% IBR penetration, the proposed approach prevents system collapse where a conventional VIS strategy fails, ensuring frequency stability at a cost increase of only 5%.