Modelling of Initial Magnetic Curves of Non-Oriented Laminated Steels Using Artificial Neural Networks

The article discusses modelling of the initial static magnetic curve of non-oriented laminated steels using artificial neural networks (ANN) for usage in electrical machine calculations. ANN enables modelling of nonlinear characteristics, which is crucial for developing electrical machines that meet application requirements. The initial magnetic curve must satisfy certain boundary conditions to be suitable for utilization in electromagnetic calculations. It is important to consider the measurement uncertainty in calculations that can be in case of ANN achieved by introducing noise injection. In conclusion, the significance and effectiveness of ANN in improving calculations and simulations in the field of electrical engineering are emphasized. This opens new possibilities for developing and understanding the magnetic properties of materials, emphasizing the significant impact and effectiveness of ANN in improving calculations and simulations.