Fault Detection and Classification in Smart Grid Using Machine Learning Approach

Variability in power system loads across different times makes it challenging to develop precise mathematical models for fault detection. Most protection devices in power grids rely on approximate models, which can lead to issues with accuracy when identifying and diagnosing faults. The proposed work aims to address this challenge by employing machine learning algorithms for detecting and classifying faults in an electrical power network. Artificial intelligence (AI) and machine learning (ML) techniques are highly effective in analyzing and classifying large datasets due to their superior efficiency. In this study, three-phase current and voltage values at the receiving end serve as inputs. Multiple algorithms, including K-Nearest Neighbors (KNN), Decision Tree, Random Forest, and XGBoost, are used for fault detection and classification, along with phase-by-phase analysis. The work focuses on analyzing faults in a transmission line using MATLAB/SIMULINK, where a 300-kilometer transmission line model was developed. Various fault types-such as single-phase-to-ground (L-G), two-phase-to-ground (LL-G), three-phase-to-ground (LLL-G), phase-to-phase (LL), and three-phase (LLL)-were simulated to assess grid performance. Simulation results demonstrated that the machine learning-based approach efficiently detects and classifies all transmission line fault types in the power system network.