Abstract:To address the issues arising from the integration of a high proportion of distributed generation (DG) into the distribution network, which has led to the transition from traditional single-source to multi-source distribution systems, resulting in increased complexity of the distribution network topology and difficulties in fault localization, this paper proposes a fault localization method based on graph convolutional networks (GCNs) for distribution networks with a high proportion of distributed generation. By abstracting busbars and lines into graph structure nodes and edges, GCN captures spatial coupling relationships between nodes, using key electrical quantities such as node voltage magnitude, current magnitude, power, and phase angle as input features to construct a fault localization model. A multi-type fault dataset is generated using the Matpower toolbox, and model training is evaluated using K-fold cross-validation. The training process is optimized through early stopping mechanisms and learning rate scheduling. Simulations are conducted based on the IEEE 33-node distribution network benchmark, with photovoltaic generation, wind generation, and energy storage systems connected at specific nodes, validating the model's fault localization capability under various fault types (single-phase ground fault, phase-to-phase short circuit, and line open circuit). Experimental results demonstrate that the proposed model can effectively locate fault nodes in complex distribution networks with high DG integration, achieving an accuracy of 98.5% and an AUC value of 0.9997. It still shows strong robustness in noisy environments and is significantly higher than convolutional neural networks and other methods in terms of model localization accuracy, training time, F1 score, AUC value, and single fault detection inference time, which has good potential for practical application.

Multi-Source Data Fusion Outage Location in Distribution Systems via Probabilistic Graph Models

Fault Location in Active Distribution Networks Using Multiple Measurement-Based Bayesian Learning

Fault Location in Active Distribution Networks Using Block Sparse Bayesian Learning

Block-Sparse Bayesian Learning Method for Fault Location in Active Distribution Networks with Limited Synchronized Measurements

A Fault Branch Location Approach for Partially Observable Distribution Network

Fast risk assessment of distribution grid with iterative inference on probabilistic graph

A Bayesian Network Model for Predicting Outages of Distribution System Caused by Hurricanes

Synthetic Active Distribution System Generation Via Unbalanced Graph Generative Adversarial Network.

A Generic Bayesian Method for Faulted Section Identification in Distribution Systems against Wind-induced Extreme Events

Real-Time Topology Estimation for Active Distribution System Using Graph-Bank Tracking Bayesian Networks

Intelligent Location Method With Limited Measurement Information for Multibranch Distribution Networks

Research on Power Outage Sensing Technology of Medium and Low Voltage Distribution Network Based on Deep Learning of Multi-source Data Fusion

PPGN: Physics-Preserved Graph Networks for Real-Time Fault Location in Distribution Systems with Limited Observation and Labels

Fault location method for distribution networks based on multi-head graph attention networks

Distribution Network Fault Recovery Strategy Considering Multi-Time Sections and Multi-Source Collaboration

Quick Line Outage Identification in Urban Distribution Grids via Smart Meters

A Method for Fault Localization in Distribution Networks with High Proportions of Distributed Generation Based on Graph Convolutional Networks

A Novel Hybrid Method for Short-Term Probabilistic Load Forecasting in Distribution Networks

A robust distribution network state estimation method based on enhanced clustering Algorithm: Accounting for multiple DG output modes and data loss

Distribution Network Topology Identification Using Smart Meter Data and Considering the Same-Bus-Different-Feeder Condition

A BN-based risk assessment model of natural gas pipelines integrating knowledge graph and DEMATEL