|
Detection of High Impedance Faults in Electrical Distribution Systems Based on the RF-ICEEMDAN Approach
|
Mohammad Sadegh Attar *   |
| Department of Electrical Engineering, Tafresh University, Tafresh, Iran |
|
|
Abstract: (18 Views) |
One of the most critical faults in electrical distribution networks, often leading to hazardous incidents such as fires, is the high impedance fault (HIF). Due to the low magnitude of the fault current, this type of fault typically remains undetected by conventional protection devices, such as overcurrent relays. Moreover, the amplitude variations and waveform of HIF currents closely resemble those of other phenomena, such as variations in linear and nonlinear loads. In this paper, a Random Forest classification algorithm combined with an improved noise-assisted Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN) is employed solely for the identification of high impedance faults. In the proposed RF-ICEEMDAN method, the dataset consisting of measured current signals from high impedance faults and load variations is first collected. The ICEEMDAN method is then used to decompose the signals and extract features from each sample. The extracted features are fed into the Random Forest classification algorithm. The proposed approach is implemented on the IEEE 34-bus distribution test system. The results demonstrate high accuracy and effectiveness in detecting high impedance faults. For simulation and evaluation purposes, EMTP-RV software is used for modeling the power system, and Python programming in the Google Colab environment is utilized for feature extraction and implementation of the Random Forest algorithm. The obtained results confirm the high performance and reliability of the proposed approach. |
|
| Keywords: High impedance fault, Random forest, Load switching, Fault current, ICEEMDAN method |
|
|
Full-Text [PDF 1478 kb]
(8 Downloads)
|
Type of Study: Research |
Subject:
Special
Received: 2025/03/15 | Accepted: 2025/05/28 | Published: 2025/05/31
|
|
|
|
|
|
|
| Send email to the article author |
|
|
|
| Add your comments about this article |
|
|
|
