Abstract:Despite the numerous benefits of underground cabling in modern power distribution systems, the persistent challenge of fault detection and localization remains a critical concern. The reliability of underground power systems hinges upon swift and accurate fault detection to facilitate rapid grid restoration. Addressing this challenge, this study introduces a pioneering approach that integrates deep learning with conventional power signal analysis techniques, with a specific emphasis on frequency domain analysis. Key to the proposed method is the exploitation of the resonance frequency behavior exhibited by underground power cables in response to signal fluctuations. These resonance frequencies, inherently linked to cable parameters, offer a unique insight into fault detection and location. By leveraging this feature, the proposed methodology enables precise identification and location of faults. This study introduces a pioneering approach that integrates long short-term memory networks with conventional power signal analysis techniques with a specific emphasis on frequency domain analysis. Central to the approach is the utilization of the system's normal resonance frequency as a benchmark for training the deep learning network. Through the induction of simulated faults at various positions along the cable network and subsequent analysis of resulting resonance frequency variations, the network is effectively trained to discern faults accurately. Validation of the proposed methodology is conducted through extensive MATLAB simulations on a representative network. The outcomes validate the method's effectiveness and precision, highlighting its capacity to improve the dependability and efficiency of subterranean power distribution networks.

Enhanced incipient fault identification in underground distribution cable implementing Random Forest classifier

Characterization and Identification of Electrical Tree Growth Stages Inside High-Voltage Cable Insulation.

Iterative Online Fault Identification Scheme for High-Voltage Circuit Breaker Utilizing a Lost Data Repair Technique

Cable Incipient Fault Identification Method Using Power Disturbance Waveform Feature Learning

A Novel Arc Fault Detection Method Integrated Random Forest, Improved Multi-scale Permutation Entropy and Wavelet Packet Transform

A Hybrid Technique for Fault Classification and Location in a Jointed Overhead-Underground Distribution Line

Intelligence Scheme for Fault Location in a Combined Overhead Transmission Line & Underground Cable

A novel graph search and machine learning method to detect and locate high impedance fault zone in distribution system

Deep learning-based fault detection and location in underground power cables using resonance frequency analysis

Fault classification using Artificial Neural Network in combined underground cable and overhead line

Underground Power Cable Fault Detection and Localization Based on the Spectrum of Propagation Functions

Cable Insulation Fault Identification Using Partial Discharge Patterns Analysis

Incipient Fault Detection in Power Distribution System: A Time–Frequency Embedded Deep-Learning-Based Approach

An empirical wavelet transform based fault detection and hybrid convolutional recurrent neural network for fault classification in distribution network integrated power system

Effectiveness of Wavelet Scalogram on Partial Discharge Pattern Classification of XLPE Cable Insulation

A Novel Method for Identifying Cable Defect and Improving Location Accuracy Based on Frequency Domain Reflectometry

Effective fault diagnosis and distance calculation for photovoltaic‐based DC microgrid using adaptive EWT and kernel random vector functional link network

Fault Diagnosis for Power Cables Based on Convolutional Neural Network With Chaotic System and Discrete Wavelet Transform

Fault classification on the hybrid transmission line system between overhead line and underground cable

An Intelligent Recognition Method of Partial Discharge of Power Cables Based on Data Augmentation

Transmission Line Fault Classification Based on the Combination of Scaled Wavelet Scalograms and CNNs Using a One-Side Sensor for Data Collection