Facing the challenge of high heat flux in electronic devices, effective heat dissipation with low consumption has drawn great attention. Based on the constructal theory, the optimal structures of fractal heat sinks are obtained by genetic algorithm (GA) optimization. The optimal branching level is 4. Furthermore, the Taguchi experimental design is employed to conduct the CFD simulation, and the grey analysis and variance analysis are used to investigate the influence rank and contribution of impact factors (branching levels, fractal number, heat flux, and flow rate) on the performance parameters (maximum temperature, pressure drop and comprehensive performance). It is found that the heat flux exhibits the highest contributions of 74.75% and 79.26% on the performance value of maximum temperature and comprehensive performance, respectively. For the pressure drop, the fractal number show the highest contribution of 67.99%. This work provides a novel method and guidance for the design and analysis of heat sinks.

面对电子设备高热通量的挑战，低消耗的有效散热引起了人们的极大关注。基于构造理论，通过遗传算法(GA)优化得到分形散热器的最优结构。最优分支级数为4。此外，采用田口实验设计进行CFD模拟，通过灰色分析和方差分析研究影响因子（分支级数、分形数、热通量、和流量）对性能参数（最高温度、压降和综合性能）的影响。发现热通量对最高温度和综合性能的性能值的贡献最大，分别为74.75%和79.26%。对于压降，分形数的贡献最高，为 67.99%。这项工作为散热器的设计和分析提供了一种新的方法和指导。