Study on Prediction Model of Magnetic Field Intensity of Submarine Power Cable Based on LSTM

Abstract

After the submarine power cables are laid, the parameters such as distance between cables and surface of the sea (H), distance between two submarine power cables placed in parallel (d) are constant. It is difficult to study the magnetic field intensity distribution of submarine power cable under different H and d, because the submarine cable can't be placed randomly. In order to solve this problem, firstly the influence of different groups of H and d on the distribution of submarine cable magnetic field intensity is studied, and the magnetic field intensity of submarine cable generated by different groups of H and d are used as training samples and testing samples. Then the training samples are used to construct the submarine power cable magnetic field prediction model based on Long Short Term Memory (LSTM) neural network, and the submarine power cable magnetic field is predicted with the testing samples as the input. Finally, the simulation results show that there is good prediction ability of LSTM prediction model for submarine cable magnetic field data.