Feature Extraction and Recognition of Transmission Line Faults Based on Multifractal and 1DCNN

The accurate and rapid identification of high voltage transmission line fault types is crucial for the safe and stable operation of the power grid. When the high voltage transmission line fails, the relay protection devices at both ends of the line will collect fault recording data. The fault recording data contain rich fault information. The research results show that the voltage signals of the three phases in fault recording data exhibit multifractal characteristics. In this paper, a high voltage transmission line fault model is established through Simulink and generates simulated fault data. The left and right boundary points and peak points of the multifractal spectra of three-phase voltage signals in the fault data are extracted as features through multifractal theory, and the proposed features are input into a one-dimensional convolutional neural network (1DCNN) model for training. Finally, the trained model is used for fault type recognition. Through the verification of simulated and actual data, the method proposed in the paper can effectively identify the types of faults in high voltage transmission lines, especially for the identification of two-phase short circuit faults and two-phase grounding faults, with high accuracy.