Data-Driven Ground-Fault Location Method in Distribution Power System With Distributed Generation

TL;DR

This paper tackles the challenge of locating ground faults in distribution power systems with high distributed-generation penetration, where traditional fault-location methods struggle due to multi-directional power flow. It proposes a data-driven approach that converts time-domain fault voltages at the substation into informative features via discrete wavelet transform, then uses a set of artificial neural networks to predict the faulted phase, fault distance, and faulted path. The method is validated on an 11-bus 20 kV DPS with varied fault scenarios and DG levels, achieving a mean distance error of 0.40% and faulted-phase accuracy of 100%, with path accuracy around 75%. While promising, the study notes robustness and dataset-reliability challenges, recommending larger, balanced datasets and more complex models, and points toward digital-twin implementations to address real-world disturbances.

Abstract

The recent increase in renewable energy penetration at the distribution level introduces a multi-directional power flow that outdated traditional fault location techniques. To this extent, the development of new methods is needed to ensure fast and accurate fault localization and, hence, strengthen power system reliability. This paper proposes a data-driven ground fault location method for the power distribution system. An 11-bus 20 kV power system is modeled in Matlab/Simulink to simulate ground faults. The faults are generated at different locations and under various system operational states. Time-domain faulted three-phase voltages at the system substation are then analyzed with discrete wavelet transform. Statistical quantities of the processed data are eventually used to train an Artificial Neural Network (ANN) to find a mapping between computed voltage features and faults. Specifically, three ANNs allow the prediction of faulted phase, faulted branch, and fault distance from the system substation separately. According to the results, the method shows good potential, with a total relative error of 0,4% for fault distance prediction. The method is applied to datasets with unknown system states to test robustness.

Figures (11)

• Figure 1: Distribution power system test model structure.
• Figure 2: Three-phase voltages recorded at the system substation under various fault scenarios and different fault locations.
• Figure 3: Discrete wavelet transform with multi-level decomposition where $j=1,2,3$.
• Figure 4: Detail wavelet coefficients of phase A voltage for a phase A-to-ground fault at 4500 m along path 1.
• Figure 5: Correlation between faulted phase and standard deviation.