Neural Network Analysis of the Electromagnetic Field of Cable Systems with Insulation Made of Polymer Materials

The chapter is devoted to the study of the electromagnetic field (EMF) in the electrical insulating material of power cable lines (PCL). Models are constructed for calculating and analyzing the voltage distribution of inhomogeneous electric fields in a dielectric medium with inclusions of different areas and with different electrophysical parameters (filling). Modeling and analysis of the distribution of the electric field strength in the defect area for various types of inclusion filling was carried out and a pattern was established that can be a diagnostic parameter of the quality of the insulation of the PCL. Within the framework of the conducted research, a neural network (NN) model describing the architecture of a cyber-physical system (CPS) for predicting the resource of EM cable electrical networks was built. An algorithmic solution for automatic parameter selection and NN training with subsequent forecasting has been synthesized, which makes it possible to increase the reliability of the CPS by reducing the time to create an optimal NN configuration. The obtained NN model can be effectively used for the analysis of thermal fluctuation processes occurring in the control object—PCL, and the prediction of the behavior of the object.