摘要
为改善横向波纹隔热屏综合冷却效果,采用三维数值模拟和基于支持向量机的代理优化模型,在给定的单位面积冷却空气流量下对发散冷却结构参数进行了优化研究。设计变量选取为气膜孔直径、气膜孔排布的展向间距和流向间距,以面积平均综合冷却效率作为目标函数,通过遗传算法搜索获得了设计变量区间内的优化设计点。在隔热屏单位表面积冷却空气流量Gf = 2.647 kg/(
加力燃烧室隔热屏广泛采用致密孔的发散冷却结构。大量研究表明,发散冷却能够在壁面形成全覆盖的气膜,具有较为理想的冷却效
针对波纹隔热屏,Shinbo
从以上研究工作中可以发现,发散冷却结构的开孔率及孔排布对于波纹隔热屏的冷却效率影响很大,然而,由于波纹隔热屏的结构影响因素多、变化规律复杂,在工程设计中必需通过多参数影响的耦合分析才能遴选出合理的冷却结构参数匹配,以实现其综合冷却效果的优化。对此,本文采用基于CFD的数值优化设计方法,针对特定的波纹结构和接近真实的加力燃烧室气热条件下,对横向波纹隔热屏发散冷却结构的主要结构参数,如气膜孔直径(d)、展向间距(P)和流向间距(S)进行优化设计分析,以获得高的隔热屏壁面综合冷却效率。
如

图1 计算域示意图
Fig.1 Schematic sketch of computational domain
计算域边界条件设置如下:主次流进口均为质量流量进口,主流进口质量流量为 0.185 9 kg/s,温度为2 250 K,总压为2.562 MPa;次流进口温度为900 K,质量流量按照隔热屏单位表面积冷却空气流量Gf = 2.647 kg/(
选取气膜孔直径d、展向间距P和流向间距S为设计变量,设计变量的选取参考某航空发动机的横向波纹隔热屏几何尺寸,设计区间如
对于发散冷却,以追求冷却壁面的综合冷却效率为目标,构建优化模型
(1) |
约束条件为
, , |
式中:F为适应度函数;ηav,A为隔热屏发散冷却段面积平均综合冷却效率。
局部综合冷却效率η定义为
(2) |
式中:Tc和T∞分别表示次流和主流的入口温度;Tw为壁面温度。
采用支持向量机(Support vector machine,SVM)构建代理模型,结合遗传算法(Genetic algorithm,GA)对设计变量进行全局寻优。该优化基本流程如下:
(1) 确定优化问题与目标函数,采用拉丁方设计法构建正交实验
(2) 采用ANSYS Fluent计算软件进行流场数值模拟,获得综合冷却效率随设计变量变化的数据样本,建立支持向量机优化代理模型;
(3) 结合遗传算法对设计变量进行全局寻优,以壁面综合冷却效率为目标函数获得最优值下的横向波纹隔热屏发散冷却结构。
采用ICEM软件对计算域进行网格划分。为保证网格质量,网格划分采用结构化网格。在气膜孔、壁面处采用附面层局部加密,保证

图2 计算网格
Fig.2 Computational grids
基于ANSYS Fluent软件进行求解,湍流模型选择参考已有的研
湍流模型验证按照文献[

图3 湍流模型验证
Fig.3 Validation of numerical results of turbulence model
SVM是一种小样本的机器学习算
基于对数据样本G ={(Xi, yi), i=1,2,…,Ns}(Xi为输入向量,yi为输出值,Ns为样本容量)的机器学习,SVM通过
(3) |
式中:k为核函数;参数a和参数b可通过求解式(4)和
s.t. (4)
式中c为惩罚因子。
(5) |
核函数k选用径向基函数
(6) |
式中δ为核函数参数。
SVM的预测性能取决于惩罚因子和核参数。本文通过试错法来确定这两个参数,获得的最佳核参数和惩罚因子分别为0.8和2.5。

图4 SVM预测值与CFD计算值
Fig.4 SVM estimated values vs. CFD computed values
利用基于支持向量机构建的代理模型,采用遗传算法搜索最优设计点。遗传算法是一种模拟生物进化机制、全局概率化寻找最优解的算法。其主要特点是直接对结构对象进行操作,具有较好的全局寻优能
通过优化过程,获得了在给定隔热屏单位表面积冷却空气流量下的优化结构参数(
为了说明优化参数的作用和影响机制,从
诸多研究表明,发散冷却结构的气膜孔开孔率是影响冷却效率的一种重要无因次结构参数。开孔率越大意味着气膜孔在壁面上的分布越加致密,有利于相邻气膜射流之间的融合,形成更好的表面气膜层覆盖。如




图5 壁面温度分布云图
Fig.5 Temperature distributions on wall surface
3种参考冷却结构中,参考结构3的开孔率仅为1.2%,由于其相邻气膜孔之间的展向和流向节距比(P/d和S/d)大,相邻气膜射流在喷吹后的聚合发展较为缓慢,尤其是在发散段的前区,气膜在壁面上的覆盖效果差(

图6 典型截面无量纲温度分布云图
Fig.6 Dimensionless temperature distributions on several sections

图7 典型截面速度矢量图
Fig.7 Cross sectional velocity vector

图8 沿流向局部综合冷却效率分布
Fig.8 Local overall cooling effectiveness along streamwise direction

图9 沿展向局部综合冷却效率分布
Fig.9 Local overall cooling effectiveness along spanwise direction
由此可见,横向波纹隔热屏发散冷却结构参数影响具有高度的耦合,如何在上述结构参数寻求合理的匹配,涉及多参数的优化设计。
针对一种横向波纹隔热屏发散冷却结构,在给定的单位面积冷却空气流量下,以发散冷却结构气膜孔直径d、展向间距P和流向间距S为设计变量,以横向隔热屏综合冷却效率为优化目标,基于支持向量机模型结合遗传算法在设计区间内寻优,获得了横向波纹隔热屏发散冷却结构的优化设计参数。在隔热屏单位表面积冷却空气流量Gf = 2.647 kg/(
参考文献
CERRI G, GIOVANNELLI A, BATTISTI L, et al. Advances in effusive cooling techniques of gas turbines[J]. Applied Thermal Engineering, 2007, 27: 692-698. [百度学术]
KREWINKEL R. A review of gas turbine effusion cooling studies[J]. International Journal of Heat and Mass Transfer, 2013, 66: 706-722. [百度学术]
LIN Y Z, SONG B, LI B, et al. Investigation of film cooling effectiveness of full-coverage inclined multihole walls with different hole arrangements:ASME GT2003-38881[R]. [百度学术]
S.l.]: ASME, 2003. [百度学术]
LIGRANI P, GOODRO M, FOX M, et al. Full-coverage film cooling: Film effectiveness and heat transfer coefficients for dense and sparse hole arrays at different blowing ratios[J]. ASME Journal of Turbomachinery, 2012, 134: 061039-1-13. [百度学术]
谢婕, 张靖周. 带偏转角气膜出流发散壁冷却特性的数值分析[J]. 南京航空航天大学学报,2013, 45(2): 157-161. [百度学术]
XIE Jie, ZHANG Jingzhou. Numerical simulation on cooling characteristics of effusion wall with deflection film outflow[J]. Journal of Nanjing University of Aeronautics and Astronautics, 2013, 45(2): 157-161. [百度学术]
杨卫华, 卢聪明, 郑建文. 气膜-发散冷却结构冷却效果的实验研究[J]. 南京航空航天大学学报,2014, 46(4): 517-523. [百度学术]
YANG Weihua, LU Congming, ZHENG Jianwen. Experimental investigation on cooling effectiveness of film-effusion cooling[J]. Journal of Nanjing University of Aeronautics and Astronautics, 2014, 46(4): 517-523. [百度学术]
YANG Zhimin, ZHANG Jingzhou. Experiment on adiabatic film cooling effectiveness in front zone of effusion cooling configuration[J]. Transaction of Nanjing University of Aeronautics and Astronautics, 2014, 31(3): 338-344. [百度学术]
刘友宏, 李英, 杨旭. 冲击/发散冷却层板隔热屏冷却性能及对比[J]. 航空动力学报, 2014, 29(6): 1272-1278. [百度学术]
LIU Youhong, LI Ying, YANG Xu. Cooling performance and comparison of impingement/effusion cooling lamilloy used as heat shield[J]. Journal of Aerospace Power, 2014, 29(6): 1272-1278. [百度学术]
SHINBO K, KOIDE Y, KASHIWAGI T, et al. Research of heat transfer of a liner for an afterburner: AIAA Paper 97-3005[R]. [百度学术]
S.l.]: AIAA,1997. [百度学术]
CHAMPION J L, DESHAIES B, CURTELIN R, et al. Aerodynamical structure of the wall flow over a wavy surface partially cooled by air injection through multi-perforations: AIAA Paper 99-1016[R]. [百度学术]
S.l.]: AIAA, 1999. [百度学术]
FUNAZAKI K, IGARASHI T, KOIDE Y, et al. Studies on cooling air ejected over a corrugated wall: Its aerodynamic behavior and film effectiveness: ASME 2001-GT-143 [R]. [百度学术]
S.l.]: ASME, 2001. [百度学术]
唐婵, 常海萍, 毛军逵. 离散孔纵向波纹隔热屏气膜冷却特性研究[J]. 工程热物理学报, 2007, 28(3): 487-489. [百度学术]
TANG Chan, CHANG Haiping, MAO Junkui. Numerical simulation of discrete holes on the longitudinal ripple wavy liner[J]. Journal of Engineering Thermophysics, 2007, 28(3): 487-489. [百度学术]
唐婵, 常海萍. 发散孔纵向波纹隔热屏气膜冷却特性[J]. 航空动力学报, 2009, 22(1): 37-41. [百度学术]
TANG Chan, CHANG Haiping. Numerical simulation of effusion holes on the longitudinal ripple heat shield[J]. Journal of Aerospace Power, 2009, 22(1): 37-41. [百度学术]
常国强, 常海萍, 常飞, 等. 多孔纵向波纹表面气膜冷却效率实验研究[J]. 航空动力学报, 2009, 24(3): 513-518. [百度学术]
CHANG Guoqiang, CHANG Haiping, CHANG Fei, et al. Experimental investigation on film cooling effectiveness of multi-hole at longitudinal wavy surface[J]. Journal of Aerospace Power, 2009, 24(3): 513-518. [百度学术]
SINGH K, PREMACHANDRAN B, RAVI M R. Numerical investigation of film cooling on a 2D corrugated surface[J]. Numerical Heat Transfer, Part A, 2016, 70(11): 1253-1270. [百度学术]
SINGH K, PREMACHANDRAN B, RAVI M R. Experimental and numerical studies on film cooling of a corrugated surface[J]. Applied Thermal Engineering, 2016, 108: 312-329. [百度学术]
REN Haoliang, LIU Youhong. Experimental investigation of fluid flow and heat transfer characteristics of a longitudinal corrugated liner for a combustion chamber[J]. Applied Thermal Engineering, 2016, 108: 1066-1075. [百度学术]
REN Haoliang, LIU Youhong, DU Liwei. An experimental study of flow and heat transfer performance of a longitudinal corrugated liner for a combustion chamber[J]. Applied Thermal Engineering, 2017, 127: 1305-1316. [百度学术]
渠立红, 张靖周, 谭晓茗. 发散孔横向波纹隔热屏气膜冷却特性研究[J]. 工程热物理学报, 2016, 37(7): 1532-1537. [百度学术]
QU Lihong, ZHANG Jingzhou, TAN Xiaoming. Numerical investigation of film cooling characteristics for effusion holes on a transverse ripple heat shield[J]. Journal of Engineering Thermophysics, 2016, 37(7): 1532-1537. [百度学术]
QU L H, ZHANG J Z, TAN X M, et al. Numerical investigation on adiabatic film cooling effectiveness and heat transfer coefficient for effusion cooling over a transverse corrugated surface[J]. Chinese Journal of Aeronautics, 2017, 30(2): 677-684. [百度学术]
渠立红, 张靖周, 谭晓茗. 发散孔结构参数对横向波纹表面气膜绝热冷却效率的影响[J]. 航空动力学报, 2018, 33(3): 590-596. [百度学术]
QU Lihong, ZHANG Jingzhou, TAN Xiaoming. Effects of structural parameters of effusion holes on adiabatic film cooling effectiveness over transverse corrugated surface[J]. Journal of Aerospace Power, 2018, 33(3): 590-596. [百度学术]
曾文明, 谭晓茗, 张靖周, 等. 高冷气温度下横向波纹隔热屏气膜冷却特性研究[J]. 推进技术, 2019, 40: 180305. [百度学术]
ZENG Wenming, TAN Xiaoming, ZHANG Jingzhou, et al. Numerical investigation of film cooling characteristics of a transverse ripple heat shield with high temperature coolant[J]. Journal of Propulsion and Power, 2019, 40: 180305. [百度学术]
张玲, 王冲. 不同横斜槽结构对气膜冷却效果影响的正交模拟[J]. 推进技术, 2016, 37(5): 922-929. [百度学术]
ZHANG Ling, WANG Chong. Orthogonal simulation of effects of different transverse declining slot structures on film cooling[J]. Journal of Propulsion and Power, 2016, 37(5): 922-929. [百度学术]
HARRISON K, BOGARD D. Comparison of RANS turbulence models for prediction of film cooling performance: ASME GT2008-50366[R]. [百度学术]
S.l.]: ASME, 2008. [百度学术]
SILIETI M, KASSAB A J, DIVO E. Film cooling effectiveness: Comparison of adiabatic and conjugate heat transfer CFD models[J]. International Journal of Thermal Science, 2009, 48: 2237-2248. [百度学术]
渠立红, 张靖周, 谭晓茗. 狭缝喷注-发散冷却的综合冷却效果数值研究[J]. 推进技术, 2018, 39(4): 849-856. [百度学术]
QU Lihong, ZHANG Jingzhou, TAN Xiaoming. Numerical investigation on overall cooling effectiveness for a combined scheme of slot injection and effusion cooling[J]. Journal of Propulsion and Power, 2018, 39(4): 849-856. [百度学术]
GUSTAFSSON K M B, JOHANSSON T G. An experimental study of surface temperature distribution on effusion-cooled plates[J]. Journal of Engineering for Gas Turbines and Power, 2001, 123(2): 308-316. [百度学术]
CORTES C, VAPNIK V. Support-vector networks[J]. Machine Learning, 1995, 20(3): 273-297. [百度学术]
WANG C H, ZHANG J Z, ZHOU J H. Optimization of a fan-shaped hole to improve film cooling performance by RBF neural network and genetic algorithm[J]. Aerospace Science and Technology, 2016, 58: 18-25. [百度学术]