Fast and accurate fault detection and classification in transmission lines using extreme learning machine

Goni, Md Omaer Faruq, Nahiduzzaman, Md, Anower, Md Shamim, Rahman, Md Mahabubur, Islam, Md Robiul, Ahsan, Mominul, Haider, Julfikar ORCID: https://orcid.org/0000-0001-7010-8285 and Shahjalal, Mohammad (2023) Fast and accurate fault detection and classification in transmission lines using extreme learning machine. e-Prime: Advances in Electrical Engineering, Electronics and Energy, 3. p. 100107. ISSN 2772-6711

Preview

Published Version Available under License Creative Commons Attribution. Download (5MB) | Preview

Abstract

To provide stability and a continuous supply of power, the detection and classification of faults in the transmission lines (TLs) are crucial in this modern age. It is required to remove a faulty section from a healthy section to provide safety and to minimize power loss due to the fault. In the contemporary world, machine learning (ML) is extensively used in every aspect of life. In this study, a spontaneous fault detection (FD) and fault classification (FC) system based on ML has been proposed. MATLAB Simulink was employed to simulate two different TLs and to generate normal and fault data (Per unit voltage and current) of ten different types. TL-1 consisted of a single generator and a single load whereas TL-2 consisted of two generators and three loads. Upon normalizing the data, an extreme learning machine (ELM) algorithm was used as the classifier. Two different ELM models were developed for FD and FC purposes through training. The method achieved fault classification accuracies of 99.18% and 99.09% for the TL-1 and TL-2 respectively. On the other hand, fault detection accuracies of 99.53% and 99.60% were achieved for the TL-1 and TL-2. The proposed ELM model compared to a traditional artificial neural network (ANN) model demonstrated relatively a shorter processing time and reduced computational complexity. In addition, the proposed method outperformed the existing state-of-the-art methods.