摘要
针对超临界机组末级长叶片的设计特点,采用遗传算法和人工神经网络,提出对长叶片典型截面叶型进行分区优化设计思想,并对原型与改型进行了多工况点的数值计算,结果表明,将叶型吸力侧后半段由直线型改为内凹型,能够显著降低超声速叶型在超声速工况范围内的叶型损失。对叶型前缘以及压力侧的局部优化设计能够改善超声速叶型在临界马赫数工况下的气动性能。优化设计最大程度地减小了样本空间,提高了优化效率。
Based on genetic algorithm and neutral network,a multi-segment optimization design method was proposed for the typical blade profile of an ultra-supercritical steam turbine.Optimization under multiple operating conditions was performed for the original and optimized blade profiles.The result showed that the concaving design for the suction side of blade profile reduced the energy loss under high Mach number conditions.Synthetically with the subsonic design technique for the blade leading edge and pressure s...
出处
《推进技术》
EI
CAS
CSCD
北大核心
2009年第3期314-317,共4页
Journal of Propulsion Technology
关键词
超声速流
叶片
气动特性
优化设计
Supersonic flow
Blade
Aerodynamic characteristic
Optimization design
