High‐Impedance Fault Detection in Distribution Networks Based on Support Vector Machine and Wavelet Transform Approach (Case Study: Markazi Province of Iran)
Open Access
- 19 February 2025
- journal article
- Published by Wiley in Energy Science & Engineering
- Vol. 13 (3) , 1171-1183
Abstract
High impedance faults (HIFs) can lead to crucial damage to the utility grid, such as the risk of fire in material assets, electricity supply interruptions, and long service restoration times. Due to their low current magnitude, conventional protective equipment, such as overcurrent relays, cannot detect these faults. Alternatively, the waveform and variation range of current in HIFs are similar to other phenomena, such as linear and nonlinear load changes and capacitor banks. This paper employs a support vector machine (SVM) classification algorithm that demonstrates reliable accuracy and discrete wavelet transform (DWT) in HIF detection. First, the data set containing measured current signals of HIFs is collected to implement this approach. Then, DWT decomposes it to extract the features of each sample in the data set. The extracted features from this part are used as input to the SVM classification algorithm. The proposed idea is initially implemented on the IEEE 34‐bus distribution test network. The proposed method achieves high capability and accuracy in detecting high‐impedance faults. The proposed method is also applied to a real power distribution network in Markazi Province of Iran, yielding satisfactory results. EMTP‐RV simulation software is used to simulate and evaluate the proposed method for power network modeling. Moreover, MATLAB software is used for feature extraction, and Python programming language in Google Colab and Spyder environment is applied to implement the SVM algorithm. The simulation results confirm the high accuracy of the suggested method. The main criteria obtained by the proposed method include accuracy, sensitivity, specificity, precision, F‐score, and Dice, which are 99.581%, 98.684%, 100%, 100%, 99.338%, and 99.338%, respectively, for the test network, and 97.94%, 93.45%, 100%, 100%, 96.614%, and 96.618%, respectively, for the real power distribution network.Keywords
This publication has 36 references indexed in Scilit:
- High impedance fault location in power distribution systems using deep learning and micro phasor measurements units ($\mu$PMU)Multiscale and Multidisciplinary Modeling, Experiments and Design, 2023
- A new method based on the discrete time energy separation algorithm for high and low impedance faults detection in distribution systemsElectric Power Systems Research, 2023
- A novel graph search and machine learning method to detect and locate high impedance fault zone in distribution systemEngineering Reports, 2022
- High impedance fault detection method in distribution network based on improved Emanuel model and DenseNetEnergy Reports, 2022
- High-impedance fault detection in power distribution grid systems based on support vector machine approachElectrical Engineering, 2022
- Deep learning for high-impedance fault detection and classification: transformer-CNNNeural Computing and Applications, 2022
- Fuzzy based relaying scheme for high impedance faults in DG integrated distribution systemElectric Power Systems Research, 2022
- High impedance fault detection based on current harmonic analysis using Chirp Z transformEnergy Systems, 2021
- High-impedance fault detection in medium-voltage distribution network using computational intelligence-based classifiersNeural Computing and Applications, 2019
- Fault classification in power systems using EMD and SVMAin Shams Engineering Journal, 2017