Location and Recognition System for Lightning Fault of Transmission Line Based on Data-driven Technology

Abstract

In order to solve the complex and difficult identification problems of overhead transmission line fault diagnosis, and to improve the accuracy of classification effectively, a new method of fault diagnosis for overhead transmission line is proposed in this paper. Firstly, the collected traveling wave signals are processed by HHT (Hilbert-Huang Transform) to realize joint feature extraction in time-frequency domain. And a data-driven lightning strike warning model for transmission lines is adopted. The model includes PCA (principal component analysis), data acquisition and preprocessing, data analysis and prediction, and model online correction. For eliminating the influence of noise and singularity on fault diagnosis; then input training set and production rules to train the intelligent classification method, by which exact fault diagnosis model was obtained. Finally, apply the algorithm to the intelligent lightning traveling wave monitoring system of an actual 500 kV transmission line, the experimental results show that the proposed method can not only calculate the exact location of fault points, but also accurately classified them that classified both single fault and multi-fault, which opens up a new approach for overhead transmission line to intelligent fault diagnosis.