Abstract:Detecting high-impedance faults (HIF) in distribution networks poses a significant challenge for conventional relay devices due to low fault currents and various characteristics like weaker fault features, distortion offset, and background noise interference. This paper introduces a novel and streamlined method for HIF detection, which ingeniously integrates frequency-band energy curve (FBEC) analysis and Gaussian smoothing to extract trend changes, thus enhancing the precision and effectiveness of HIF detection. The proposed method utilizes continuous wavelet transform (CWT) to extract the time-frequency spectrum from the zero-sequence current. By analyzing the feature-band energy, the FBEC is computed. To mitigate noise interference and enhance the periodic change pattern, a Gaussian filter is applied for smoothing. Distinguishing between HIF and normal operations, including low impedance fault (LIF), capacitor switching (CS), inrush current (IC), and ferromagnetic resonance (FR), is achieved by analyzing the peak points of FBEC. The proposed method’s performance was extensively validated through simulations and field data. The performance of the proposed method was extensively validated through a series of simulations and detailed analysis of real-world field data. The results demonstrated an excellent detection performance on field data, with an impressive accuracy rate of 86.5% and F1-score of 0.87. Moreover, we examined the method’s resilience against noise using data with a Signal-to-Noise Ratio (SNR) of 20dB, resulting in a detection accuracy of 77.3% and F1-score of 0.79. These findings underscore the method’s clear physical meaning, strong interpretability, and versatility, establishing its effectiveness and practicality for real-world applications.

A CNN Based Transfer Learning Method for High Impedance Fault Detection

Fault Prediction for Distribution Network Based on CNN and LightGBM Algorithm

A High-Impedance Fault Detection Method for Active Distribution Networks Based on Time–Frequency–Space Domain Fusion Features and Hybrid Convolutional Neural Network

Deep learning for high-impedance fault detection and classification: transformer-CNN

High-impedance Fault Detection Method Based on Feature Extraction and Synchronous Data Divergence Discrimination in Distribution Networks

Unsupervised High Impedance Fault Detection Using Autoencoder and Principal Component Analysis

A Data-Driven Approach for High-Impedance Fault Localization in Distribution Systems

High Impedance Fault Detection Based on Arc Changing Tendency and Nonlinear Least Squares for Active Distribution Networks

A Fault Branch Location Approach for Partially Observable Distribution Network

Deep-Learning-Based Fault Classification Using Hilbert–Huang Transform and Convolutional Neural Network in Power Distribution Systems

Detection of single-phase-to-ground faults in distribution networks based on Gramian Angular Field and Improved Convolutional Neural Networks

Physics-Informed Learning for High Impedance Faults Detection

High Impedance Fault Detection Method Based on Variational Mode Decomposition and Teager–Kaiser Energy Operators for Distribution Network

Fault classification and location of a PMU-equipped active distribution network using deep convolution neural network (CNN)

Detection of High-Impedance Fault in Distribution Networks Using Frequency-Band Energy Curve

High-impedance fault detection method based on sparse data divergence discrimination in distribution networks

Fault detection and classification on insulated overhead conductors based on MCNN‐LSTM

Double convolutional neural network for fault identification of power distribution network

High Impedance Fault Detection in Overhead Distribution Feeders Using a DSP-Based Feeder Terminal Unit

Deep Learning for High-Impedance Fault Detection: Convolutional Autoencoders