上海科技大学知识管理系统

Traveling waves within the distribution network will be produced by faults. In practical distribution network, the wavefront of traveling wave is hard to be captured, as the measurement signal contains a mass of noises. The sources of noises are not only from the measurement devices, but also from traveling waves caused by other events such as load changes in the distribution systems. In addition, the wave speed of the traveling wave needs to be accurately determined to ensure fault location. This paper collects and analyzes field data from a real power distribution system. The zero sequence current signals are captured by sensors placed at each line terminal, with the high sampling rate of 50 MHz. First, the natural frequency method is adopted to derive the true velocity of the traveling wave. Second, the arrival time of the current traveling wave is captured by the tool of continuous wavelet transform (CWT). The high frequency components of CWT are utilized to avoid the frequency dependent effect of the cable line. With the arrival time and velocity of traveling waves, the accurate fault location is determined. Field data experiments show that the proposed method can accurately capture the wavefront with large background noises. The estimated fault locations are close to the actual values, proving the effectiveness of the method in practical distribution cables. © 2023 IEEE.