基于格子玻尔兹曼方法的功率模块自动布局优化  

Power Module Automatic Layout Optimization Based on Lattice Boltzmann Method

作　　者：回晓双 宁圃奇[1,2,3] 崔健 HUI Xiaoshuang;NING Puqi;CUI Jian(Institute of Electrical Engineering,Chinese Academy of Sciences,Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 100049,China;Key Laboratory of Power Electronics and Electric Drive,Institute of Electrical Engineering,Chinese Academy of Sciences,Beijing 100190,China)

机构地区：[1]中国科学院电工研究所,北京100190 [2]中国科学院大学,北京100049 [3]电力电子与电气驱动重点实验室(中国科学院电工研究所),北京100190

出　　处：《电源学报》2025年第1期236-242,共7页Journal of Power Supply

基　　金：国家重点研发计划资助项目(2021YFB2500600);中国科学院青年交叉团队资助项目(JCTD-2021-09)。

摘　　要：针对在传统的功率模块自动布局优化算法中,方案的电气评估效率低、占用大量计算时间的问题,提出利用格子波尔兹曼方法LBM(lattice Boltzmann method)代替传统评估方法,由于LBM不需要进行多个可逆矩阵的求解,可以更快地进行电气互连合理性判断及电压/电流计算。首先,在基于遗传算法自动布局设计程序的基础上,采用D2Q4格子类型建立二维布局评估方法;然后,通过ANSYS Q3D软件仿真验证了布局方案评估结果的准确性;最后,在Python3.10环境中进行对比测试,结果表明LBM平均缩短了75.4%的方案电气评估总时间,且评估方案中回路数量越多,LBM的计算优势越大。In the traditional power module automatic layout optimization algorithm,the electrical evaluation is inefficient and takes up a lot of computing time.To solve this problem,lattice Boltzmann method(LBM)is used to replace the traditional evaluation method.Since LBM does not need to solve multiple invertible matrices,it can quickly judge the rationality of electrical interconnection and calculate the voltage/current.With the program of automatic layout design based on the genetic algorithm,an evaluation method of two-dimensional layout is established by using a D2Q4 lattice type,and the accuracy of the evaluation result under the layout scheme is verified by ANSYS Q3D software simulation.A comparative test was conducted in Python 3.10,and results show that LBM reduces the total time of scheme evaluation by 75.4%on average.Moreover,the more the number of loops in the evaluation scheme,the greater the computing advantage of LBM.

关 键 词：功率模块 遗传算法 格子波尔兹曼方法 阻抗评估 

分 类 号：TM571[电气工程—电器]