Event-Triggered Islanding in Inverter-Based Grids

The decentralization of modern power systems challenges the hierarchical structure of the electric grid and requires the implementation of automated schemes that can overcome adverse conditions. This work proposes an adaptive isolation methodology that can segregate a grid topology in autonomous islands that maintain stable and economic operation in the presence of deliberate (e.g., cyberattacks) or unintentional abnormal events. The adaptive isolation logic is event-triggered to avoid false positives, improve detection accuracy, and reduce computational overheads. A measurement-based stable kernel representation (SKR) triggering mechanism inspects distributed generation controllers for abnormal behavior. The SKR notifies a machine learning (ML) ensemble classifier that detects whether the system behavior is within acceptable operational conditions. The event-triggered adaptive isolation framework is evaluated using IEEE RTS-24 bus system. Simulation results demonstrate that the proposed framework detects anomalous behavior in real-time and identifies stable partitions minimizing operating costs faster than traditional islanding detection techniques.