Convolutional Physics-Informed Neural Networks for Fast Prediction of Core Losses in Axisymmetric Transformers

CC BY 4.0Brendel, PhilippPhilippBrendelMedvedev, VladVladMedvedevRoßkopf, AndreasAndreasRoßkopf2025-11-122025-11-122024https://publica.fraunhofer.de/handle/publica/499188https://doi.org/10.24406/publica-628510.1109/TMAG.2024.343170310.24406/publica-62852-s2.0-85199393441In this article, solutions of the magnetostatic Maxwell's equation on parametric axisymmetric transformer geometries are approximated by a convolutional approach on physics-informed neural networks (ConvPINNs). The trained ConvPINN is capable of predicting magnetic vector potentials (MVPs) and magnetic flux densities in a matter of milliseconds for a range of geometries described by a total of 18-20 degrees of freedom (DoFs). The combination of ConvPINN with an existing framework for core loss prediction yields a super fast workflow for the approximation of core losses on a wide range of geometric setups. The combination of inference speed and accuracy enables new orders of magnitude for the optimization of transformer designs going forward.entrueConvolutional neural networkscore lossesmagnetostaticsphysics-informed neural networks (PINNs)transformersConvolutional Physics-Informed Neural Networks for Fast Prediction of Core Losses in Axisymmetric Transformersjournal article