电工技术学报2025，Vol.40Issue(4)：997-1008,12.DOI:10.19595/j.cnki.1000-6753.tces.240061
基于替代模型的旋转磁耦合器磁心结构优化
Optimization of the Core Structure for the Rotary Magnetic Coupler Based on Surrogate Model
摘要
Abstract
During the design process of rotary magnetic couplers,many magnetic cores are typically employed to achieve magnetic constraint,enhance coupling performance,and reduce magnetic leakage.However,decreasing the cost,self-weight,and volume of rotary magnetic couplers is challenging,and power density needs to be improved.In some special applications,it may also affect the reliability of the equipment.Current research mainly relies on engineering experience and trial-and-error methods for magnetic core structure design,causing issues such as imprecise parameter design and extensive simulation workload.This paper proposes a magnetically optimized method based on surrogate modeling to meet the lightweight design requirements of magnetic core structures for rotary couplers.This method can reduce the usage of magnetic cores while ensuring stable transmission performance. The structural parameters of the rotary magnetic coupler are first determined,a Latin hypercube sampling method is employed to sample different structural parameters randomly,and a dataset is constructed.A surrogate model is constructed using the extreme random forest algorithm to predict system coupling coefficients for different magnetic core structures.The predictive accuracy of the surrogate model is validated using a test dataset.Finally,combining the surrogate model with a multi-objective particle swarm optimization method,an optimization approach for the magnetic core structure of the rotary magnetic coupler is presented.This approach provides optimal magnetic core design solutions considering the usage of the magnetic core. Simulation results indicate that the maximum error between the simulated coupling coefficient and the output results of the surrogate model is 0.001 38,with a maximum relative error of 0.161% and an average relative error of 0.096% .The parameter set based on this surrogate model decreases magnetic core volume by 32.15% .The volume of the magnetic core significantly decreases when the coupling coefficient is allowed to fluctuate during the design process of the magnetic core structure.When the coupling coefficient decreases by 1.74%,the volume of the magnetic core decreases by 66.63% .In the experiment,with the same energy efficiency design objectives,the proposed optimization method decreases the utilization volume of magnetic cores by 29.47% compared to traditional methods.The experimental setup achieves a power transmission of 3 kW with an efficiency of 97% .In addition,the rotary magnetic coupler designed in this paper maintains almost constant output power and efficiency under angular variation.Therefore,the rotary magnetic coupler exhibits stable output performance under rotational conditions. The main contributions of this paper are as follows.(1)A surrogate model for the magnetic core structure and coupling coefficient of the rotary magnetic coupler is constructed based on extreme randomized trees(ERT),achieving an R2 value of 0.999 65.This surrogate model accurately captures the nonlinear relationship between the magnetic core structure parameters and the coupling coefficient.(2)A magnetic core structure optimization method is proposed for rotary magnetic couplers,considering magnetic core usage volume and coupling coefficient,which effectively reduces the usage of magnetic cores and maintains the system's energy efficiency.关键词
无线电能传输/旋转磁耦合器/磁心结构优化/极端随机森林/多目标粒子群优化Key words
Wireless power transfer/rotary magnetic coupler/magnetic core structure optimization/extreme random forest/multiple objective particle swarm optimization分类
信息技术与安全科学引用本文复制引用
丰宇宸,孙跃,邓德强,胡宏晟,胡韩..基于替代模型的旋转磁耦合器磁心结构优化[J].电工技术学报,2025,40(4):997-1008,12.基金项目
国家自然科学基金面上资助项目(62073047). （62073047）
