Study on performance optimization of single expansion ramp nozzle based on depth neural network
-
摘要:
针对高超声速飞行器机体/推进系统一体化布局下超燃冲压发动机单边膨胀喷管（SERN）的推力最大化、力矩匹配和几何可约束设计要求，提出了一种基于深度神经网络（DNN）的单边膨胀喷管性能优化方法。基于单边膨胀喷管数值仿真数据集，建立基于深度神经网络的喷管壁面压力分布预测模型，对喷管性能影响参数进行了灵敏度分析，并结合优化算法对其性能进行优化。研究结果表明：基于Unet-L3卷积神经网络构建的单边膨胀喷管沿程壁面压力分布预测模型具有较高的精度；基于喷管壁面压力分布预测模型和差分进化算法的单目标优化算法无法同时对单边膨胀喷管的推力系数和推力矢量角进行优化；而结合喷管壁面压力分布预测模型和混合优化算法对单边膨胀喷管推力系数和推力矢量角进行多目标优化，可在推力系数减小
0.0116 （相对降低1.17%）的情况下使得推力矢量角从1.54°降低至0.39°（相对降低74.65%），能在满足喷管推力性能的要求下实现飞行器后端横向载荷的降低，有利于宽速域飞行器的操稳和配平。Abstract:In response to the requirements of thrust maximization, torque matching and geometric constraint for single expansion ramp nozzle (SERN) of scramjet due to the integration of aircraft/engines, a novel method based on depth neural network (DNN) for SERN performance optimization was proposed. Based on the data set from numerical simulation, the predicting model for calculating the SERN’s wall pressure distribution was established by DNN, which can be applied to optimize the SERN performance combined with the optimization algorithm, and the sensitivity analysis of nozzle performance on geometric parameters was carried out. The results showed that: the prediction model based on Unet-L3 convolutional neural network for predicting the wall pressure distribution of SERN had rather high accuracy. The single-objective optimization algorithm based on the DNN prediction model and the differential evolution algorithm can rarely optimize the thrust coefficient and thrust vector angle simultaneously. The multi-objective optimization for thrust coefficient and thrust vector angle of SERN can be achieved by using the DNN prediction model and hybrid optimization algorithm. By multi-objective optimization, the reduction of thrust coefficient by
0.0116 (relatively decreased by 1.17%) can decrease the thrust vector angle from 1.54° to 0.39° (relatively reduced by 74.65%), and help to reduce the trimming resistance of hypersonic aircraft and the difficulty of flight control. -
图 8 测试集壁面压力分布预测结果
Figure 8. Prediction results of wall pressure distribution in test set
图 10 推力系数单目标优化曲线
Figure 10. Curve of single objective optimization for thrust coefficient
图 11 推力矢量角单目标优化曲线
Figure 11. Curve of single objective optimization for thrust vector angle
表 1 单边膨胀喷管数据库取值范围
Table 1. Sampling range of SERN database
变量 取值范围 α/（°） 0～10 β/（°） 5～15 γ/（°） 0～10 y2/H 1.5～2.5 
下载: 导出CSV
表 2 基准单边膨胀喷管几何特征参数及性能
Table 2. Geometric characteristic parameters and performance of reference SERN
α/（°） β/（°） γ/（°） y2/mm Cfx |θ|/（°） 5 10 5 2H 0.9878 1.517
下载: 导出CSV
表 3 单边膨胀喷管推力系数优化结果
Table 3. Optimization result of SERN thrust coefficient
α/（°） β/（°） γ/（°） y2/mm Cfx |θ|/（°） 9.997 10.421 0.0447 2.4999 H0.9957 1.5433
下载: 导出CSV
表 4 单边膨胀喷管推力矢量角优化结果
Table 4. Optimization result of SERN thrust vector angle
α/（°） β/（°） γ/（°） y2/mm Cfx |θ|/10−5（°） 7.255 6.3956 6.5019 1.9331 H0.9782 6.72
下载: 导出CSV
表 5 单边膨胀喷管多目标优化结果
Table 5. Multi-objective optimization result of SERN
α/（°） β/（°） γ/（°） y2/mm Cfx |θ|/（°） 9.825 6.306 0.4241 2.1036 H0.9841 0.3912
下载: 导出CSV
表 6 不同优化方案下单边膨胀喷管性能对比
Table 6. Comparison of SERN performance under different optimization schemes
优化方案 优化结果 Cfx |θ|/（°） 推力系数单目标优化 0.9957 1.5433 推力矢量角单目标优化 0.9782 6.72×10−5 多目标优化 0.9841 0.3912
下载: 导出CSV
-
[1] KAZMAR R. Airbreathing hypersonic propulsion at Pratt & Whitney: overview[R]. AIAA-2005-3256, 2005. [2] SZIROCZAK D, SMITH H. A review of design issues specific to hypersonic flight vehicles[J]. Progress in Aerospace Sciences, 2016, 84: 1-28. doi: 10.1016/j.paerosci.2016.04.001 [3] EDWARDS C, SMALL W, WEIDNER J, et al. Studies of scramjet/airframe integration techniques for hypersonic aircraft[R]. AIAA-1975-58, 1975. [4] LEDERER R, KRUEGER W. Nozzle development as a key element for hypersonics[R]. AIAA-1993-5058, 1993. [5] 张艳慧. 非对称大膨胀比喷管设计及性能分析[D]. 南京: 南京航空航天大学, 2006. ZHANG Yanhui. Single expansion ramp nozzle design and performance analysis[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2006. (in ChineseZHANG Yanhui. Single expansion ramp nozzle design and performance analysis[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2006. (in Chinese) [6] 全志斌, 徐惊雷, 莫建伟. 单边膨胀喷管膨胀型面的非线性缩短设计[J]. 推进技术, 2012, 33(6): 951-955. QUAN Zhibin, XU Jinglei, MO Jianwei. Design of nonlinearly compressed SERN profile[J]. Journal of Propulsion Technology, 2012, 33(6): 951-955. (in ChineseQUAN Zhibin, XU Jinglei, MO Jianwei. Design of nonlinearly compressed SERN profile[J]. Journal of Propulsion Technology, 2012, 33(6): 951-955. (in Chinese) [7] JU Shengjun, YAN Chao, WANG Xiaoyong, et al. Optimization design of energy deposition on single expansion ramp nozzle[J]. Acta Astronautica, 2017, 140: 351-361. doi: 10.1016/j.actaastro.2017.09.004 [8] OGAWA H, BOYCE R R. Nozzle design optimization for axisymmetric scramjets by using surrogate-assisted evolutionary algorithms[J]. Journal of Propulsion and Power, 2012, 28(6): 1324-1338. doi: 10.2514/1.B34482 [9] MO Jianwei, XU Jinglei, QUAN Zhibin, et al. Design and cold flow test of a scramjet nozzle with nonuniform inflow[J]. Acta Astronautica, 2015, 108: 92-105. doi: 10.1016/j.actaastro.2014.12.005 [10] YU Kaikai, XU Jinglei, LV Zheng, et al. Inverse design methodology on a single expansion ramp nozzle for scramjets[J]. Aerospace Science and Technology, 2019, 92: 9-19. doi: 10.1016/j.ast.2019.05.054 [11] 陈以勒. 超燃冲压发动机喷管智能设计与性能研究[D]. 南京: 南京航空航天大学, 2021. CHEN Yile. Intelligent design and performance study of scramjet nozzle[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2021. (in ChineseCHEN Yile. Intelligent design and performance study of scramjet nozzle[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2021. (in Chinese) [12] MIYANAWALA T P, JAIMAN R K. An efficient deep learning technique for the Navier-Stokes equations: application to unsteady wake flow dynamics[EB/OL]. (2018-08-15)[2023-08-20]. https://arxiv.org/abs/1710.09099v3. [13] 金鑫, 殷建业, 王健志. 基于深度学习的飞行载荷测试与反演方法研究[J]. 航空工程进展, 2020, 11(6): 887-893. JIN Xin, YIN Jianye, WANG Jianzhi. Research on deep-learning-based flight load test and estimation method[J]. Advances in Aeronautical Science and Engineering, 2020, 11(6): 887-893. (in ChineseJIN Xin, YIN Jianye, WANG Jianzhi. Research on deep-learning-based flight load test and estimation method[J]. Advances in Aeronautical Science and Engineering, 2020, 11(6): 887-893. (in Chinese) [14] SEKAR V, JIANG Qinghua, SHU Chang, et al. Fast flow field prediction over airfoils using deep learning approach[J]. Physics of Fluids, 2019, 31(5): 057103. doi: 10.1063/1.5094943 [15] 马博文, 巫骁雄, 于洋. 基于机器学习方法的压气机落后角与总压损失预测代理模型[J]. 航空动力学报, 2023, 38(7): 1675-1690. MA Bowen, WU Xiaoxiong, YU Yang. Surrogate model for deviation angle and total pressure loss prediction of compressor based on machine learning methods[J]. Journal of Aerospace Power, 2023, 38(7): 1675-1690. (in ChineseMA Bowen, WU Xiaoxiong, YU Yang. Surrogate model for deviation angle and total pressure loss prediction of compressor based on machine learning methods[J]. Journal of Aerospace Power, 2023, 38(7): 1675-1690. (in Chinese) [16] ZHU Linyang, ZHANG Weiwei, SUN Xuxiang, et al. Turbulence closure for high Reynolds number airfoil flows by deep neural networks[J]. Aerospace Science and Technology, 2021, 110: 106452. doi: 10.1016/j.ast.2020.106452 [17] 朱剑琴, 李地科, 陶智, 等. 基于约束神经网络的气膜冷效分布预测方法[J]. 航空动力学报, 2023, 38(7): 1537-1545. ZHU Jianqin, LI Dike, TAO Zhi, et al. Predicting method of film cooling effectiveness distribution based on constrained neural network[J]. Journal of Aerospace Power, 2023, 38(7): 1537-1545. (in ChineseZHU Jianqin, LI Dike, TAO Zhi, et al. Predicting method of film cooling effectiveness distribution based on constrained neural network[J]. Journal of Aerospace Power, 2023, 38(7): 1537-1545. (in Chinese) [18] YU Jian, HESTHAVEN J S. Flowfield reconstruction method using artificial neural network[J]. AIAA Journal, 2018, 57(2): 482-498. [19] YU Kaikai, CHEN Chong, CHEN Yile. Inverse design of nozzle using convolutional neural network[J]. Journal of Spacecraft and Rockets, 2022, 59(4): 1161-1170. doi: 10.2514/1.A35243 [20] 王骥飞. 高超声速飞行器气动外形一体化设计方法研究[D]. 西安: 西北工业大学, 2018. WANG Jifei. Research on integration design methodology of aerodynamic shape for hypersonic aircrafts[D]. Xi’an: Northwestern Polytechnical University, 2018. (in ChineseWANG Jifei. Research on integration design methodology of aerodynamic shape for hypersonic aircrafts[D]. Xi’an: Northwestern Polytechnical University, 2018. (in Chinese) [21] SPAID F, KEENER E. Hypersonic nozzle/afterbody CFD code validation: Ⅰ experimental measurements[R]. AIAA-1993-0607, 1993. -
点击查看大图
计量
- 文章访问数: 355
- HTML浏览量: 245
- PDF量: 43
- 被引次数: 0