Abstract: Due to the complex operating environments and stringent de-energized maintenance requirements of distribution network cables, there is an urgent need for accurate and effective live detection technologies. We investigated the induced current of the metal shielding layer in 10 kV single-core cables. By deriving theoretical formulas and employing modeling simulations, we examined the evolution patterns, intrinsic mechanisms, and influencing factors of harmonic distortion in the induced current. Simulations reveal that magnetic field distortion at insulation defect locations will facilitate the occurrance of abnormal harmonic components in the induced current of the metal shielding layer, primarily consisting of the 3rd, 5th, and 7th harmonics. Further full-scale experiments indicate that the primary harmonic components are influenced by load current, insulation defect types, and through-flow aging time of the insulation material. Notably, the 3rd harmonic content of metal floating defects is approximately 140.2% that in normal cables. Finally, by applying the K-means clustering algorithm to the measured harmonic data, effective classification of insulation defects in distribution network cables is achieved. With a conductor current of 300 A, the water drop defect achieves a defect recognition rate of 84.09%, providing a theoretical foundation for timely warning and live diagnosis of insulation defects.