Multispectrum Joint Localization Method for Cable Faults Based on Broadband Impedance Spectrum and Morlet Wavelet

The precise diagnosis of cable fault type and location before the fault occurs is crucial. To address the issues of pseudo-reflection interference and insufficient fault feature extraction accuracy in traditional cable fault detection methods, this paper proposes a multi-spectrum joint localization method based on broadband impedance spectroscopy (BIS) and Morlet wavelet. By constructing a cable distributed parameter model, the impedance spectrum characteristics of four fault types—open-circuit, high-impedance, low-impedance, and short-circuit—are analyzed, including amplitude shift, phase mutation, and resonance pattern changes. A ridge smoothness optimization algorithm based on Morlet wavelet transform is proposed to effectively suppress interference from non-fault points, such as “pseudo-peaks” and ripple oscillations. Furthermore, a multi-spectrum joint analysis mechanism combining the real part, imaginary part, and amplitude is established, and the particle swarm optimization algorithm is applied to dynamically optimize the weight combination of the spectra, overcoming the limitations of traditional single-spectrum analysis. Experimental results show that when a typical fault is set at 70 m of a 100 m coaxial cable, the relative localization error is within 0.7%, which is more than 60% more accurate than the traditional impedance method. Additionally, a fault type recognition criterion is proposed, which differentiates fault types based on impedance spectrum feature differences, and the effectiveness and accuracy of this method have been verified.