Statistical Estimation of Propagation Constants Using Stepped-Frequency Waveform Reflectometry

Stepped-frequency waveform reflectometry (SFWR) can simultaneously detect fault locations and estimate the propagation constant of a power cable. The estimated propagation constant is used to assess the degradation of the cable insulation. However, propagation signals experience distortion over distance and do not maintain constant magnitude and phase within the time duration of the signals. This degrades the performance of fault detection and insulation assessment. To solve this problem, we optimized the conventional propagation constant estimation results using a kernel density estimation (KDE) and validated the performance of the proposed method through the signal detection results of the inverse discrete Fourier transform (IDFT).