Study On Magnetic Properties On Iron Cores Under Harmonic And DC Bias Excitation

The rapid development of DC transmission technology and the wide application of power electronic equipment cause DC bias and an increase of excitation harmonics in the winding current core and winding core magnetic density.,which makes it difficult to get accurate core loss.In order to calculate the core loss caused by complex excitation with higher accuracy,the magnetic properties of silicon steel sheets under sinusoidal,harmonic and DC bias are measured in this thesis firstly.The effects of frequency,harmonic characteristics and DC bias on the magnetic properties of silicon steel sheets are analyzed.Based on the Bertotti loss model,the high-order terms of the magnetic induction intensity in the hysteresis loss is introduced to improve the fitting accuracy,and three kinds of core loss calculation models are used to fit the sinusoidal losses at different frequencies.Deriving the improved Bertotti loss model into harmonic form,fitting harmonic losses and calculating errors.Then,considering the difference with the actual transformer laminated core and silicon steel sheets in the magnetic characteristics,a product-level laminated core model was built in this thesis,and its loss under sinusoidal,harmonic and DC biad excitations were measured.Simulation of core hysteresis loop under DC bias by BP neural network.Finally,The magnetic field finite element software MagNet was applied to simulate the flux density distribution,and then the field quantity of each mesh in the simulation model was extracted and the core loss was calculated by the improved Bertotti loss model under sinusoidal and harmonic excitation,and calculated by neural network under DC bias excitation.The results show that the improved Bertotti loss model and BP neural network are more accuracy than the MagNet software.