Fault line selection algorithm for single-phase grounding faults in distribution networks based on multi-feature fusion of multi-level wavelet decomposition coefficients

This paper introduces a novel method for determining fault lines, addressing the time-consuming nature of traditional steady-state quantity methods and the susceptibility to errors in transient quantity methods. A multi-feature fusion technique, involving multi-level wavelet decomposition coefficients, is employed. This involves the extraction of zero-sequence current wavelet transform coefficients with the Discrete Wavelet Transform (DWT), and the formation of a similarity function through the S-transform. Subsequently, an energy measure function is derived by integrating these coefficients. A function representing faultiness, aligned with various fault lines, is then developed and used as the primary criterion for identifying the faulted line. Simulation data from a single-phase grounding fault model in a distribution network confirms the efficacy of the method in choosing fault lines.