Abstract: In order to improve the ability of non-axisymmetric end wall reducing the second flow，and further improve the performance of turbines，a novel optimization design method of non-axisymmetric end wall contouring was developed，which was also adopted for a high pressure(HP)turbine stator row. End wall parametrization and 3D Navier-Stokes(N-S)flow solver and genetic algorithms(GA)based on artificial neural network(ANN)were carried out to the non-axisymmetric end wall optimization. The effects of optimized non-axisymmetric end wall on the flow field of the HP turbine stator were analyzed. The results indicate that the optimized non-axisymmetric end wall can effectively improve the flow structures in the HP turbine stator，hence the formation of the passage vortex is delayed and the intensity of the corner vortex is decreased. Furthermore，the flow loss is reduced. And the total pressure loss coefficient at the stator exit reduces by 3.724%. In addition，the non-axisymmetric end wall contouring on hub has few influences on the flow field upper half of the blade.

Key words: High pressure turbine；Non-axisymmetric end wall；Optimization design；Total pressure loss coefficient

摘要： 为提高非轴对称端壁控制端壁处二次流的潜能，进一步提高涡轮性能，发展了一种新型的非轴对称端壁优化设计方法，并以高压涡轮导叶为研究对象，采用端壁参数化、三维N-S方程流场求解与基于人工神经网络的遗传算法相结合的方法进行非轴对称端壁优化，分析了优化后的非轴对称端壁造型对涡轮导叶流场的影响。结果表明:优化后的非轴对称端壁改善了涡轮导叶的流场，延迟了通道涡的生成和发展，削弱了角涡的强度，降低了导叶通道内的流动损失，涡轮导叶出口处的总压损失系数降低了3.724%。此外，非轴对称下端壁造型对高压涡轮导叶上半叶高流场的影响不大。

关键词: 高压涡轮；非轴对称端壁；优化设计；总压损失系数