A time-frequency transformation method for cable defect localization integrating step and impulse responses

To overcome the limitations of traditional time domain reflectometry (TDR) in long-distance cable defect localization and the inability of frequency domain reflectometry (FDR) to assess degradation severity, this paper proposes a power cable defect localization method based on time–frequency transformation. First, a time–frequency analysis approach based on double integration was employed, combined with low-frequency reflection coefficient prediction, to derive the time domain representations of step and impulse responses from the frequency domain. Subsequently, the characteristic impedance variation patterns are analyzed, and a MATLAB simulation model covering multiple defect scenarios was established to investigate the effect of spectrum reconstruction on defect characterization performance. Experimental results show that, under laboratory conditions, the relative localization error at the defect initiation point does not exceed 0.07%. Further analysis shows that, due to the uncertainty in electromagnetic wave propagation velocity, the maximum error in estimating the defect length was approximately 0.16%. On-site tests verify its effectiveness under different defect types. Compared with existing FDR techniques, the proposed method not only localizes defects but also assists in identifying defect types and assessing degradation severity.