A Data-Driven Model of Cable Insulation Defect Based on Convolutional Neural Networks

The insulation condition of cables has been the focus research in power systems. To address problem that electric field is not easily measured under operating 10 kV transmission with defects, this paper proposes a data-driven cable defect model based on convolutional neural network approach. data during operation obtained by finite element calculation, and multi-dimensional input feature quantity set strength as output are constructed. A algorithm applied to construct model. used cloud map distribution operation. Comparing results method, overall accuracy 94.3% calculation time 0.025 s, which 360 times faster than calculation. show can quickly laying foundation for digital twin structure cables.