Abstract: In view of the design difficulty of global optimization design for turbine blades， such as the long computing time and large sample space， a feasible optimization design method is put forward. Seventeen parameters controlling the blade profile are used as optimization variables， and the second generation of multi-objective genetic algorithm is applied for global optimization automatically. Based on this method， a global optimization design platform for turbine blades is built. By using this design platform， the optimal design of a supersonic turbine blade is carried out using two modes: fixed axial solidity and free optimization. Numerical results show that at the design condition， the total pressure loss coefficient of the optimization blades is decreased by 19.5% and 10.0% compared to the reference blade， and the intensity of shock wave is reduced obviously for the two optimized profiles obtained from two methods respectively. And the aerodynamic performance at all operating conditions is also improved. The flow field and shock wave structure show that the outer trail shock wave has a greater influence on the total loss than the inner trail shock wave. The outer trail shock wave can be reduced by decreasing expansion turning angle or the inner trail shock wave turning angle.

Key words: Global optimization；Supersonic turbine blades；Shock-wave attenuation；Pneumatics performance

摘要： 针对涡轮叶型全局优化设计计算时间长、样本空间大等难点提出一种可行的优化设计方法，该方法将控制叶型的17个参数作为优化变量，采用第二代多目标遗传算法进行全局自动寻优。基于此方法，搭建了涡轮叶型全局优化设计平台。利用此平台，分别采用轴向稠度固定和自由优化两种方式对超声速涡轮叶型进行了优化设计。数值计算结果表明，两组优化设计叶型在设计工况下总压损失系数比参考叶型分别低19.5%和10.0%，流道中的激波强度更弱，且在变工况条件下都具有较好的气动性能。深入分析流场与激波结构后发现，外尾激波相比于内尾激波对总损失的影响更大，通过减小气流膨胀转折角或内尾激波气流转折角能够有效削弱外尾激波强度。

关键词: 全局优化；超声速涡轮叶片；激波弱化；气动性能