Online Monitoring Method of Cable Insulation Based on Digital Intelligent Power Disturbance

This article proposes a cable insulation online monitoring method based on digital intelligent power disturbance to address the frequent faults caused by cable insulation aging or damage in the power system. Firstly, by collecting and preprocessing power disturbance signals, accurate characteristic data reflecting the insulation status of cables is obtained; next, the signal processing technique of wavelet transform is used to extract features from the perturbed data, in order to effectively identify possible insulation anomalies; then, combined with machine learning algorithms based on Transformer architecture, the feature data is classified and predicted to achieve accurate monitoring and early warning of cable insulation status. The experimental results show that the classification accuracy of the Transformer model based on healthy, slightly aged, and severely aged states is 98.5%, 94.2%, and 91.7%, respectively. The average response time of the system is 14.9 milliseconds, and the detection rate remains above 90% in most experiments. Overall, it exhibits high detection accuracy and real-time performance, effectively improving the reliability and efficiency of cable insulation monitoring, and has significant engineering application advantages.