Research on Fast Fault Identification Method for Transmission Line Traveling Wave Data Based on Machine Learning Algorithms

• Published in: 2024 IEEE 3rd International Conference on Electrical Engineering, Big Data and Algorithms (EEBDA) pp. 449 - 456
• Main Authors: Chengjun, Ren, Yuhui, Peng, Yingpu, Xie, Min, Wu, Mengfei, Lei, Lin, Zhou, Yadi, Zhang, Zhuowen, Li, Haonan, Zhu, Yongzeng, Ji
• Format: Conference Proceeding
• Language: English
• Published: IEEE 27.02.2024
• Subjects: Distributed databases; Distributed devices; Fault diagnosis; Feature extraction; machine learning; Machine learning algorithms; Power transmission lines; Traveling wave data; Wavelet analysis; Wavelet packet; Wavelet packets
• DOI: 10.1109/EEBDA60612.2024.10485761

The different types of faults on transmission lines seriously threaten the safe and stable operation of the system. Currently, the fault diagnosis of transmission lines involves very little fault information, only focusing on the zero sequence current information in the fault recording. Therefore, this article proposes a fast fault identification method based on machine learning algorithms for transmission line traveling wave data. The method is mainly based on the obtained traveling wave data of distributed online monitoring devices for transmission lines, and integrates the actual effective fault waveform database and labeled information database of transmission lines. The time-domain function, FFT, and wavelet packet analysis are used to extract the time-domain, frequency-domain, and the CNN-LSTM-SE machine learning algorithm is used to complete fault diagnosis of transmission lines based on time-frequency domain fault eigenvalues. And based on historical trip line fault data, the correctness and effectiveness of the method proposed in this paper were verified through practical examples.