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

Mauro Caporuscio,Antoine Dupuis,Welf Löwe

2024-02-23

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increase in renewable energy penetration at the distribution level introduces a multi-directional power flow that outdated traditional fault location techniques.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.

Systems and Control,Artificial Intelligence,Machine Learning,Signal Processing

What problem does this paper attempt to address?

The focus of this paper is on the problem of grounding fault location in distribution systems under the context of distributed generation (DG). Traditional methods have become outdated due to bidirectional power flow, thus new approaches are needed to ensure fast and accurate fault location for enhanced reliability of power systems. The paper proposes a data-driven grounding fault location method.

Specifically, the researchers simulate a grounding fault in an 11-node 20kV power system using Matlab/Simulink. They analyze the three-phase voltage during the fault using discrete wavelet transform, and then extract statistical features to train an artificial neural network (ANN) for establishing the mapping relationship between voltage characteristics and faults. The ANN is divided into three parts, predicting the faulted phase, faulted branch, and the distance to the fault point from the substation. According to the results, this method achieves a total relative error of only 0.4% in fault distance prediction.

The paper also compares the proposed method with traditional methods and other data-driven methods, highlighting its flexibility in adapting to specific distribution networks. Additionally, the robustness of the method is tested by applying it to a dataset with unknown system states.

In conclusion, this paper aims to address the issue of traditional fault location techniques becoming ineffective due to the increase in distributed generation, and how to develop and apply new data-driven methods to improve the accuracy and efficiency of grounding fault location in distribution systems.