⚡ MATLAB Model: Fault Detection, Classification & Location in IEEE 13-Node Test Feeder using ANN - (ANN-based fault detection classification and location in IEEE 13 Node)
🧩 Overview
This MATLAB/Simulink model demonstrates an Artificial Neural Network (ANN)–based intelligent fault diagnosis system for the IEEE 13-Node Test Feeder.
The project performs fault detection, classification, and location estimation under various abnormal conditions in an unbalanced distribution system.
It integrates data collection, ANN training, and model validation to deliver a complete end-to-end solution for smart grid fault analysis.
⚙️ System Configuration
Test System: IEEE 13-Node Distribution Feeder
Software Platform: MATLAB/Simulink (R2018a or above)
Control Technique: Feed-forward back-propagation ANN
Input Parameters: Voltage and current (ABC phase signals)
Outputs:
Fault detection (faulted / non-faulted)
Fault classification (SLG, LL, DLG, 3-phase)
Fault location (nearest bus number)
🔍 Working Principle
Fault Simulation
Different types of faults are created at various bus locations (e.g., 632–671).
Each fault condition—single line-to-ground, line-to-line, double line-to-ground, and three-phase—is simulated to generate training data.Data Acquisition & Pre-Processing
Three-phase voltage and current waveforms are collected from each node and stored in .mat files.
These files form the input dataset, while binary target labels indicate fault type and location.ANN Training
The datasets are trained using MATLAB’s Neural Network Toolbox (nnstart).
The ANN is optimized to minimize Mean Square Error (MSE) and maximize correlation (R) during training, validation, and testing.Simulink Integration
Trained ANN blocks for both fault detection/classification and fault location are exported to Simulink.
Real-time analysis is achieved by feeding system outputs (V, I) into the ANN models.Result Visualization
The system identifies:Whether a fault occurred
Type of fault (e.g., BCG fault)
Approximate location (e.g., near Bus 671)
Results are displayed on the simulation panel in real time.
📈 Performance Highlights
Parameter | Result |
---|---|
Training Accuracy | > 97 % |
MSE (Validation) | < 0.002 |
R-value (Regression) | ≈ 0.99 |
Execution Time | < 3 s per simulation |
Fault Location Error | ± 1 bus |
📂 Package Includes
IEEE 13-Node Feeder Simulink Model (.slx)
ANN Training MATLAB Script (.m)
Pre-trained Network .mat Files
Sample Fault Dataset
Demonstration Video Link - https://www.youtube.com/watch?v=XIj4eSOwE9E
ANN Based Fault Detection, Classification and location in IEEE 13 Node Test Feed
Simulink Super Sale
Demonstration Video Link - https://www.youtube.com/watch?v=XIj4eSOwE9E