Fault Detection Algorithm for Transmission Lines Based on Random Matrix Spectrum Analysis

Considering the lack of anti-interference ability of fault detection algorithm based on the fundamental wave phasor system and similarity of time-domain waveform, a novel transmission line fault detection algorithm based on random matrix spectrum analysis is proposed. Firstly, using the almost reverse characteristics of waveforms for the current fault components at both ends of the line, the original matrix is formed by a fixed time window that consists of current fault component on one side and opposite current fault component on the other side. A random matrix is constructed by copying, translating, and adding noise, its mean spectral radius is calculated and spectrum analysis is taken, which brings about two criteria of fault detections for lines. Then, the single ring theorem of the theory for random matrix is used to set the threshold and the influence of various factors on the proposed algorithm is analyzed. Finally, a large number of simulation examples in the IEEE-39 bus system prove the effectiveness of the algorithm. The simulation results show that the algorithm is not affected by the fault location, non-full-phase operation and heavy load and it will not cut off the normal line when the system oscillates. It also has a strong ability of resisting the transition resistance and does not require high information of synchronization. © 2019 Automation of Electric Power Systems Press.