Artificial Intelligence-Based Method for Constructing the Temperature Field of Power Cables

Power cable temperature is an important index for determining current-rated capacity, evaluating operational state and identifying defects. A key challenge lies in quickly capturing the temperature distribution of a cable section and improving digitalization efforts. Current numerical computation methods are time-consuming and artificial neural networks alone cannot be explained. This study introduces a fast calculation method for cable temperature field estimation via a network incorporating physical information. First, cable model is established, and a dataset of cable temperature field variations under different load currents is calculated. Second, a neural network based on physical information is constructed, and the heat conduction governing equation and boundary conditions are incorporated into the loss function of the network as residuals to constrain and optimize the model training process. This method provides a new method for calculating the cable temperature in emergencies and evaluating the ageing of different cable material layers.