电工技术学报
论文 |
综合考虑材料热各向异性与多种传热方式的磁性元件热阻网络精准模型
郭轩, 肖云昊, 李驰, 郑泽东
电力系统及大型发电设备安全控制与仿真国家重点实验室(清华大学)北京 100084
An Accurate Thermal Resistance Network Model for Magnetic Elements Considering Thermal Anisotropy of Materials and Various Heat Transfer Ways
Guo Xuan, Xiao Yunhao, Li Chi, Zheng Zedong
State Key Laboratory of Control and Simulation of Power System and Generation Equipment Tsinghua University Beijing 100084
全文: PDF (1739 KB)  
输出: BibTeX | EndNote (RIS)      
摘要 

磁性元件正在向高频化、小型化、高功率密度的方向发展,随着功率密度的提高,散热成为影响器件可靠运行的关键因素,对器件的热分析提出了更高的要求。传统的热分析模型存在运算时间长、传热方式单一等问题。该文引入已有研究中提出的三轴九热阻网络模型,在此基础上添加热对流与热辐射两种传热方式,综合考虑材料热各向异性与热传导、热对流、热辐射。并通过精细划分元件温度计算子区域、实际器件损耗场分析、热电耦合迭代提高了模型的计算精度,提出一种更加完善的三轴十五热阻网络模型,基于该模型对实际磁性器件建模,验证了模型的准确性,并于最后提出了一种通用的针对EE、EI、UU等典型磁心构成的磁性元件的热阻网络建模方法。

服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
郭轩
肖云昊
李驰
郑泽东
关键词 磁性元件材料热各向异性多种传热方式实际损耗分布集总参数热阻网络    
Abstract

The development trend of magnetic components is higher frequency, smaller volume and higher power density. With the increase of power density, heat dissipation becomes a key factor affecting the reliable operation of magnetic components, which puts forward higher requirement for thermal analysis of magnetic components. The traditional thermal analysis models have some problems such as long calculation time and single heat transfer way. In addition, the thermal anisotropy, different distribution of loss density in magnetic core and interaction effect between temperature and loss are usually ignored. In order to meet the calculation requirements of a large number of data points in the magnetic components optimization design and match the actual working condition with complex heat dissipation ways, a precise and generalized analytical thermal modeling method is needed. The inductor made of EE type magnetic core is taken as an example, and three-axis nine-thermal-resistance network model proposed in previous studies for solving problem of heat conduction with thermal anisotropy being considered is introduced. A three-axis fifteen-thermal-resistance network model considering multiple heat transfer ways, thermoelectric coupling, material thermal anisotropy and actual loss distribution of magnetic core was proposed. For multiple heat transfer ways, the influence of heat conduction, heat convection and heat radiation should be incorporated into the model because high power density magnetic components are often used with water cooling, air cooling or other cooling structures. And the influence of heat convection and heat radiation has been considered in the model in the form of air thermal resistances. The distribution of magnetic field has influence on the distribution of loss density in each area. And the loss distribution of magnetic core is calculated by the 2D finite element simulation of the actual magnetic field in order to match the actual condition. The loss of winding and magnetic core requires iterative calculation because temperature has influence on the iron loss density of magnetic core and electrical conductivity of copper while the winding loss and magnetic core loss both influencing the temperature. In addition, the thermal anisotropy is considered in the model. The conduction thermal resistances of different axis in Cartesian coordinate system are calculated respectively by different thermal conductivity due to thermal anisotropy. At the frequency of 50 kHz, three working conditions of Vp=350 V, Vp=460 V and Vp=590 V were selected for experiment. The temperature measuring results and calculating results are given and the accuracy of model was verified. The results show that the max relative error for the calculation of magnetic core temperature is no more than 14% and the max relative error in the highest temperature area of magnetic core is no more than 6% under three working conditions. In the comparison with other thermal resistance network model, the precision of thermal resistance network model can be improved by considering the material thermal anisotropy, thermoelectric coupling and the actual distribution of core loss. The single calculation time of the model can be reduced from several hours in 3D finite element simulation to almost one millisecond in thermal resistance network, and the total calculation time of thermal resistance network model can meet the time requirement of calculating a large number of design points for the optimization of a specific structure magnetic core. Based on the comprehensive thermal resistance network model, a general thermal modeling method is summarized for magnetic components composed of EE, EI, UU and other typical magnetic cores. The thermal equivalent modeling of air gap, the influence of edge effect and leakage flux on flux density near the air gap can be considered in the model in the future. A more comprehensive analytical analysis of the temperature field can be carried out and more precise temperature field calculation results can be obtained, providing a more reliable reference for the heat dissipation design of magnetic components.

Key wordsMagnetic device    thermal anisotropy of the material    multiple heat transfer modes    actual loss distribution    lumped-parameter thermal resistance network   
收稿日期: 2023-01-18     
PACS: TM47  
通讯作者: 郑泽东 男,1980年生,副教授,博士生导师,研究方向为电力电子与电气传动。E-mail:zzd@mail.tsinghua.edu.cn   
作者简介: 郭 轩 男,1996年生,博士研究生,研究方向为磁性元件和电力电子变压器的建模及优化。E-mail:guo-x18@mails.tsinghua.edu.cn
引用本文:   
郭轩, 肖云昊, 李驰, 郑泽东. 综合考虑材料热各向异性与多种传热方式的磁性元件热阻网络精准模型[J]. 电工技术学报, 0, (): 141-141. Guo Xuan, Xiao Yunhao, Li Chi, Zheng Zedong. An Accurate Thermal Resistance Network Model for Magnetic Elements Considering Thermal Anisotropy of Materials and Various Heat Transfer Ways. Transactions of China Electrotechnical Society, 0, (): 141-141.
链接本文:  
https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.230072          https://dgjsxb.ces-transaction.com/CN/Y0/V/I/141