Abstract: In order to enhance thermal performance of a U-shaped channel equipped with delta-shaped vortex generators， a meta-model based design optimization and data mining method is proposed for the thermal performance optimization design of the delta-shaped vortex generator. The method combines a Kriging-based global optimization algorithm with data mining technique of analysis of variance (ANOVA)， parameterization method for delta-shaped vortex generator and 3D Reynolds-Averaged Navier-Stokes (RANS) Solver technique. Upon numerical validation， design optimization of a U-shaped channel equipped with delta-shaped vortex generators is carried out for the maximization of its overall thermal performance. After optimization， both the intensity and the space of the down-wash vortex pairs generated by the delta-shaped vortex generators increased， and thus the overall thermal performance of the optimal design increased by 14.5% when compared to that of the reference design. Furthermore， knowledge discovery is carried out to detect the design variables which have significant effects on the thermal performance of U-shaped channel in prescribed design space. The parameter interactions among significant design variables and objective functions are analyzed in detail. It is indicated that the length of delta-shaped vortex generator has most significant effect on heat transfer and friction loss of U-shaped channel， while the width of delta-shaped vortex generator has most significant effect on the thermal performance of U-shaped channel.

Key words: Delta-shaped vortex generator；Thermal performance；Global optimization design；Knowledge discover

摘要： 为了强化布置三角形涡发生器的U型通道综合换热性能，耦合基于子元模型的全局优化算法、三角形涡发生器参数化方法、三维RANS方程求解技术与基于总变差分析的知识挖掘技术，建立了高效的三角形涡发生器综合换热性能优化与知识挖掘方法。在验证本文数值方法正确性的基础上，以综合换热性能最优为目标函数，对布置于U型通道内的三角形涡发生器进行设计优化与知识挖掘。优化后，最优设计的三角形涡发生器诱导产生的下洗涡对的强度和间距增加，使得U型通道的综合换热性能相对提高了14.5%。同时对设计空间进行知识挖掘，筛选对综合换热性能影响显著的设计变量，分析显著变量对目标函数的影响机制。结果表明:三角形涡发生器的高度对通道换热能力和阻力损失的影响最为显著，而三角形涡发生器的宽度对通道综合换热性能的影响最为显著。

关键词: 三角形涡发生器；综合换热性能；全局优化设计；知识挖掘