The invention discloses a
wavelet transform and energy entropy-based cable insulation defect identification method and device, and relates to the field of cable insulation defect identification, and the method comprises the steps: collecting three-phase wire core current signals of a to-be-detected high-
voltage cable, and carrying out the
discrete wavelet transform to obtain a multi-level high-frequency detail coefficient of each phase current
signal; and a high-frequency detail coefficient corresponding to the highest
decomposition level is extracted as a target high-frequency energy feature, so that fine high-frequency energy
mutation can be accurately captured. A
continuous current signal is segmented, the
wavelet energy spectrum of each window is calculated, and the multi-window
energy spectrum is normalized to eliminate differential interference, so that the defect phase can still be stably positioned even if the cable arrangement mode and the like are changed. The
wavelet energy entropy of each phase of current
signal is calculated, and insulation defects and external disturbance are effectively distinguished. According to the wavelet
energy spectrum, the normalized energy spectrum and the energy entropy, an insulation defect identification result of the to-be-detected high-
voltage cable is determined, a decision
closed loop is formed through multi-dimensional
verification, and the accuracy and robustness of insulation defect identification are greatly improved.