Fault Location Identification in Power Islands With Inverter Interfaced Distributed Generators

Accurate fault location identification (FLI) helps in improving system reliability and provides faster system restoration. This article proposes a voltage-dependent data-driven FLI scheme for inverter interfaced distributed generator (IIDG)-based radial distribution networks. The scheme utilizes post fault one cycle three-phase IIDG terminal voltage ( $V_{oabc}$ ) to locate the fault from its end, therefore, it does not depend upon the measurement from smart meters and PMUs placed at various location via high bandwidth communication links. A trained and tested Gaussian process regression (GPR) model, implemented in the local controller of IIDG, predicts the fault distance from respective IIDG using $V_{oabc}$ information. Compared to the existing methods, proposed scheme is found to be immune to fault levels, inverter control structure, network loading, and fault type. The scheme is tested on 33 bus radial distribution network and IEEE 123 node test feeder for all types of shunt faults, at various fault locations with different fault resistances, loading conditions and inception angles. Accuracy during line parameter and load variation, transformer tap change, noisy conditions, network reconfiguration, and DG voltage variation is evaluated. The proposed scheme is validated with real time data obtained from Opal-RT-based real time simulator. The performance of the proposed method is also compared with the existing communication-based FLI scheme. The accuracy of the proposed schemes in various fault and no-fault scenarios proves its effectiveness for practical applications.