摘要
随着现代航空发动机性能的持续提高，气冷涡轮的设计变得越来越重要。对于无冷气条件下的优化设计，很多人都做了研究。但对于有冷气条件下的优化设计，目前还未见到这方面的文献。本文联合采用人工神经网络和遗传算法对某气冷涡轮级的叶型与级间匹配进行了三维优化设计。流场计算采用全三维粘性流N-S方程求解。优化过程采用在无冷气条件下，先在粗网格上进行计算，再在细网格上进行校核的方法来减少优化时间，提高优化效率。结果表明，在无冷气的条件下，静叶和动叶的损失都有所下降，性能提高，涡轮级效率提高1％，对此结果在不同冷气量下校核。级效率提高0．80％～0．92％，提高趋势与无冷气时基本相同，说明该方法可以用于气冷涡轮的优化设计。
With the continuous improvement in performance of modern aeroengines,the design of air-cooled turbines has become ever more important.Many people have conducted a study of optimized design of modern aeroengines without the use of cooling air.However,with respect to the optimized design with cooling air,to date,no literature has yet been available.By a combined use of an artificial neural network and a genetic algorithm,a three-dimensional optimized design was conducted of blade profiles and inter-stage matching of an air-cooled turbine stage.By solving a full three-dimensional viscid flow N-S equation,a flow field calculation was performed.The optimization process involves the adoption of a method,under which the calculation was first conducted on a coarse grid without the use of cooling air,followed by a check calculation on a fine grid so as to shorten optimization time and enhance optimization efficiency.The calculation result shows that without cooling air the losses in both stator and rotor blades will somewhat be reduced and the performance enhanced with an increase in the turbine stage efficiency by 1%.With different cooling-air flow rates,the check calculation result shows that the stage efficiency will rise by 0.80% to 0.92%.The tendency featuring an increase in efficiency is basically identical to the case when no cooling air is used.This indicates that the method under discussion can be used for the optimized design of air-cooled turbines.
出处
《热能动力工程》
EI
CAS
CSCD
北大核心
2006年第5期450-455,共6页
Journal of Engineering for Thermal Energy and Power
关键词
气冷涡轮
三维优化
遗传算法
人工神经网络
air-cooled turbine,three-dimensional optimization,genetic algorithm,artificial neural network
