Study On Electromagnetic-Thermal Multi-field Models And Optimal Design Of High-Power Medium/High-Frequency High-Voltage Transformer

With the maturity and application of wide-band-gap and high-power semiconductors,the emerging applications based on high-power electronics conversion are becoming increasingly popular.They have found extensive use in fields such as electric traction,renewable power generation and transmission,power electronic transformers,and electrostatic precipitators,yielding significant economic and societal benefits.The high-power,medium/high-frequency,and high-voltage transformer is the critical component of high-power power electronic conversion.It performs many functions,such as voltage conversion,electrical isolation,resonant component,and current limiting protection.Therefore,excellent transformer modeling and design can significantly improve the overall performance of power conversion and reduce system costs.Currently,modeling methods on transformer windings,magnetic cores,distributed parameters,and thermal calculations do not fully consider real-world conditions in practical applications.Various multiphysics models struggle to fully meet on-site requirements regarding accuracy,simplification conditions,application area,complexity,and computational speed.In this context,this thesis conducts research on specific issues in multiphysics parameter modeling and optimization design of high-power medium/high-frequency high-voltage transformers,mainly covering six aspects:1.In the modeling of magnetic winding losses,to address the issue of shield winding losses in high-frequency transformers,a universal model for shield winding losses is proposed.This model is extended to various shield structures,including copper foil,round wire,Litz wire,multilayer round wire,interleaved primary-secondary shield,and coaxial transformer shield.The universality and accuracy of the loss model are verified through simulation and experimental comparisons.Furthermore,a shield optimization design method is proposed to both reduce shield losses and suppress EMI levels.This thesis combines case studies to provide design steps for commonly used copper foil shield windings in high-frequency transformers and validate them.2.In the modeling of magnetic core losses,addressing the impact of encapsulated magnetic core losses under different encapsulation thicknesses and locations in real operating conditions,a high-frequency transformer magnetic core loss test platform is established.By testing the losses of nanocrystalline cores under different encapsulation methods,it is found that the impact of small encapsulation thickness on core losses can be neglected,while losses gradually increase with the increase of encapsulation thickness.Based on this,a method for modifying the magnetic core loss model considering the effects of encapsulation processes is proposed.3.In the modeling of electric parasitic capacitance,addressing issues related to voltage/current resonance and EMI caused by parasitic capacitance between ports such as windings,magnetic cores,enclosures,ground,etc.,this thesis proposes a method for extracting critical parasitic capacitance in transformers based on parallel plate capacitors and their correction coefficients.This includes primary-side capacitance,secondary-side capacitance,inter-winding capacitance,capacitance between windings and enclosures,and capacitance between windings and magnetic cores.Finally,a comprehensive multi-port capacitance network for transformers is presented.4.In the modeling of magnetic leakage inductance,addressing the need for leakage inductance in high-power,high-frequency converters to provide resonance or current-limiting components,a transformer structure with large leakage inductance is proposed.This design can reduce the need for additional series inductances or even eliminate them in the scheme.Based on the design method,two specific design cases are provided and compared with discrete transformer and inductance design solutions.Simulation and experimental results confirm that the new design can reduce magnetic component volume,cost,and quantity effectively.5.In thermal modeling,to address the multiple heat transfer mechanisms such as heat conduction,convection,and radiation,a thermal network structure for transformers is established.Based on this,a temperature calculation method for transformers considering multiphysics coupling is proposed.Combining the structural characteristics of dry-type transformers,a method for simplified thermal modeling is proposed and verified with two types of prototypes.6.With theoretical modeling and verification,a multi-objective optimization design process for transformers based on artificial intelligence algorithms is developed.A transformer prototype is designed as a case study for an electrostatic precipitator power supply-matched 72 k V/10 k Hz/160 k VA transformer.Various design optimization and constraint conditions are considered,including distributed parameters,losses,cost,volume,temperature rise,insulation,etc.This thesis presents simulation validation based on the optimization design results,and the optimization design scheme is also validated through experimental measurements.Finally,the transformer prototype is put into on-site operation to demonstrate further the accuracy and effectiveness of the modeling,design,and measurement results.