表面响应法在埋置型大功率多芯片微波组件热布局中的应用研究  

Applications of Response Surface Methods in the Embedded High-Power Microwave Components Chip Placement

作　　者：陈品[1] 吴兆华[1] 黄红艳[1] 张生[1] 毕唐文[1] 赵强[1] 

机构地区：[1]桂林电子科技大学机电工程学院,广西桂林541004

出　　处：《电子质量》2011年第1期40-43,共4页Electronics Quality

基　　金：广西研究生教育创新计划资助项目:三维微波多芯片组件的热-电磁特性分析(2010105950802M06)

摘　　要：该文将有限元模拟与基于统计试验的表面响应法(RSM)相结合,应用于特定需求的埋置型大功率多芯片微波组件热布局分析中,先通过ANSYS温度场分析,得出大功率芯片布局是影响整体温度和芯片结温的关键因素,再对一含有四个大功率芯片的微波组件模块进行了表面响应分析,得到了关于芯片坐标的线性回归方程,利用该方程可预测坐标组合下芯片的结温和进行芯片坐标的寻优。同时,评价了RSM响应模型的精度,采用有限元仿真验证表明:线性回归方程的预测值与仿真值误差仅为0.0623℃,研究成果对埋置型大功率多芯片微波组件热设计具有指导意义。In this paper,the analysis method of the combination of the finite element analogy and response surface method is applied to the analysis of thermal distribution for special use of embedded high-power chips microwave component.We can get the result that the placement of High-power chips is the key factor to the whole and the chip junction temperature through the temperature analysis by ANSYS.Then,the Microwave component module which concludes four High-power chips is analysed by the response surface method,and get the linear regression equation about the chips coordinates,we can predict the device junction temperature over the assemble of coordinates and look for the best chip coordinates use this equation.The accuracy of the RSM model is evaluated by the best coordinates,finite element simulation results show that the error in the predictive value of the linear regression equation with the experimental value is only 0.0623℃.It turns out that the chip placement optimization approach proposed in this work can be effective and robust in providing thermal optimal design of chip placement in embedded high-power chips microwave component.

关 键 词：热分析 微波组件 回归方程 表面响应法 热布局 

分 类 号：TN454[电子电信—微电子学与固体电子学]