摘要
以50°弯角多圆弧单列叶栅和串列叶栅为研究对象,应用数值模拟手段对所设计的两类叶栅进行数值分析,获取了其不同来流马赫数条件下的性能和流动细节。应用NSGAⅡ遗传算法结合BP神经网络技术,对串列叶栅的五个关键几何参数进行优化设计,验证了优化方法的可行性。研究结果表明:在全攻角范围内,串列叶栅的静压比都高于单列叶栅,负攻角范围内,串列叶栅损失低于单列叶栅;经过优化,串列叶栅在大负攻角下的性能略有降低,同时改善了正攻角性能,在4°攻角、0.8马赫数时静压比提升4.3%,总压损失系数降低42%;优化后串列叶栅在全工况范围内性能都要优于单列叶栅,并且串列叶栅最大压比点和最小损失点攻角均向右漂移2°。
Flow characteristics and detailed flow structure were obtained by numerical simulations for both single and tandem cascades of 50° blade turning,where the multi-circular-arc airfoil was used. Based on the ge?netic algorithm coupled with neural network, five key geometric parameters of tandem cascade were optimized successfully. The result indicates that compared with the single cascade,the static pressure ratio of tandem cas?cade was higher at all incidence conditions,and the total pressure loss coefficient was smaller within the range of negative incidence. After optimization, the performance of tandem row was deteriorated at large negative inci?dence and improved at positive incidence. The static pressure ratio increased by 4.3%and total pressure loss coef?ficient decreased by 42%,at incidence angle of 4° and inlet Mach number of 0.8. Better performance of tandem cascade than single one was achieved at all incidence conditions, and the optimum incidences for the highest static pressure ratio and the smallest total pressure loss coefficient shifted to the right by 2°.
出处
《推进技术》
EI
CAS
CSCD
北大核心
2014年第11期1469-1474,共6页
Journal of Propulsion Technology
基金
中央高校基本科研业务费专项资金(3102014KYJD004)
关键词
串列叶栅
叶型优化
神经网络
遗传算法
Tandem cascade
Blade optimization
Neural network
Genetic Algorithm
