气膜冷却是提升涡轮叶片热防护能力的关键技术，其冷却效率受结构参数与工况条件影响。针对锯齿状槽道气膜冷却结构多参数优化面临建模复杂、计算成本高的挑战，在锯齿夹角(22.5–60°)、槽道高度(0.0127–9.525 mm)和吹风比(0.5–2.0)的取值范围内生成660组CFD工况样本，获取冷却效率数据。利用条件生成对抗网络（CGAN）建立输入参数与冷却效率分布之间的映射关系，实现快速预测，预测相对误差在各工况下均小于5.5%。结合麻雀搜索算法(SSA)进行结构优化，得到最优设计参数组合为锯齿夹角(42.562°)、槽道高度(4.118 mm)、吹风比(2.0)，面积平均冷却效率达到65.6%，较原始工况提高20.3%。实验结果表明，CGAN-SSA框架在复杂气膜冷却结构的快速建模与智能优化中表现出高精度与高效率，为气膜冷却结构设计提供了一种可行的技术路径。

Film cooling is a key technology for enhancing the thermal protection capability of turbine blades, and Its cool-ing efficiency is significantly influenced by the structural parameters and operating conditions. To address the challenges of complex modeling and high computational cost in multi-parameter optimization of serrated chan-nel film cooling structures, 660 CFD cases were generated under the conditions of serration angle (22.5–60°), channel height (0.0127–9.525 mm), and blowing ratio (0.5–2.0) to obtain cooling effectiveness data. A condition-al generative adversarial network (CGAN) was employed to establish the mapping between input parameters and cooling effectiveness distributions, achieving rapid predictions with errors less than 5.5% across all cases. By further integrating the sparrow search algorithm (SSA) for structural optimization, the optimal design parameters were obtained as serration angle (42.562°), channel height (4.118 mm), and blowing ratio (2.0), with the ar-ea-averaged cooling effectiveness reaching 65.6%, representing a 20.3% improvement compared to the initial case. The results demonstrate that the CGAN-SSA framework achieves high accuracy and efficiency in rapid modeling and intelligent optimization of complex film cooling structures, providing a feasible and effective pathway for film cooling design.

国家自然科学基金资助项目(52306082, 62502243);中国博士后科学基金资助项目(2024T170268, GZC20230789); 河南省研究生教育改革与质量提升工程项目(YJS2026YBGZZ46)