It proposes an intelligent recognition method based on the feature extraction of partial discharge pulse signals for addressing the challenge of operational reliability assessment in the maintenance and testing of distribution cable lines.By employing high-speed sampling hardware to perform broadband sampling of cable partial discharge current pulses

it extracts 13 multi-dimensional time-domain features to establish a multi-dimensional time-domain feature dataset for cable faults.Subsequently

it utilizes the support vector machine(SVM) to analyze these time-domain feature data

enabling intelligent identification of two or more insulating conditions of the cable.It uses this method for fault diagnosis by extracting complete time-domain feature signals.The test results show that its diagnostic accuracy rate increases from 56.3% to 97.6%.It significantly reduces the issues of false alarms and misses reports in warning information

enhancing the operational reliability and safety of cable lines.