基于非结构动网格的导弹动导数智能补偿辨识算法

臧剑文; 刘君; 苏红星; 刘凯

doi:10.13224/j.cnki.jasp.20240330

基于非结构动网格的导弹动导数智能补偿辨识算法

doi: 10.13224/j.cnki.jasp.20240330

大连理工大学力学与航空航天学院，辽宁大连 116024

基金项目: 国家自然科学基金（U2141229）

详细信息

作者简介:
臧剑文（1998-），男，博士生，主要从事飞行器气动参数辨识及数值模拟研究。E-mail：zangJW@mail.dlut.edu.cn

中图分类号: V211.7
计量
- 文章访问数: 596
- HTML浏览量: 333
- PDF量: 47
- 被引次数: 0
出版历程
- 收稿日期: 2024-05-22
- 网络出版日期: 2025-01-08

Intelligent compensation identification algorithm for missile dynamic derivatives based on unstructured dynamic mesh technology

School of Mechanics and Aerospace Engineering，Dalian University of Technology，Dalian Liaoning 116024，China

摘要

摘要:
提出了一种基于非结构动网格方法的导弹动导数智能补偿辨识算法，采用非结构动网格方法与强迫谐和振荡等方法相互组合，快速获取导弹动导数的离线数据，根据离线数据与试验值的偏差构建深度神经网络，在线智能补偿导弹气动导数。该算法仅依靠离线数据及深度神经网络，对标准模型Finner导弹的纵向气动导数进行智能补偿辨识，实现了对静稳定性导数和组合动导数的精准预测，且补偿后的动导数残差降低75%。该算法计算精度高，可推广至横向和航向的动导数智能补偿辨识。
- 非结构动网格 /
- 静稳定性导数 /
- 组合动导数 /
- Finner导弹 /
- 智能补偿辨识
Abstract:
An intelligent compensation identification algorithm for missile dynamic derivatives based on the unstructured dynamic mesh method was proposed. By combining the unstructured dynamic mesh method with forced harmonic oscillation and other methods, this algorithm could quickly obtain offline data of missile dynamic derivatives. A deep neural network was constructed based on the deviation between the offline data and experimental values to perform online intelligent compensation for missile aerodynamic derivatives. Relying solely on offline data and a deep neural network, this algorithm can intelligently compensate and identify the longitudinal aerodynamic derivatives of the standard Finner missile model, achieving accurate predictions of static stability derivatives and combined dynamic derivatives, with the residuals of the compensated dynamic derivatives reduced by 75%. With high computational accuracy, it can be extended to intelligent compensation identification of lateral and directional dynamic derivatives.
- unstructured dynamic mesh /
- static stability derivatives /
- combined dynamic derivatives /
- finner missile /
- intelligent compensation identification

HTML

图 1 非结构动网格分区示意图

Figure 1. Sketch of unstructured dynamic zoned grid

下载: 全尺寸图片幻灯片

图 2 网格点间弹簧示意图

Figure 2. Schematic diagram of springs between grid points

下载: 全尺寸图片幻灯片

图 3 迟滞环示意图

Figure 3. Sketch of the hysteresis loop

下载: 全尺寸图片幻灯片

图 4 动导数智能补偿辨识策略图

Figure 4. Strategy diagram for intelligent compensation identification of dynamic derivatives

下载: 全尺寸图片幻灯片

图 5 Finner导弹模型示意图

Figure 5. Schematic diagram of the Finner missile model

下载: 全尺寸图片幻灯片

图 6 Finner模型三维分区网格示意图

Figure 6. Schematic diagram of the three-dimensional partitioned grid of the Finner model

下载: 全尺寸图片幻灯片

图 7 α=0°非结构分区网格

Figure 7. Unstructured grid partitioning at α=0°

下载: 全尺寸图片幻灯片

图 8 压力分布云图（Ma=2，α=1°）

Figure 8. Pressure distribution contour （Ma=2，α=1°）

下载: 全尺寸图片幻灯片

图 9 速度分布云图（Ma=3，α=1°）

Figure 9. Velocity distribution contour （Ma=3，α=1°）

下载: 全尺寸图片幻灯片

图 10 迎角与俯仰力矩系数变化曲线

Figure 10. Angle of attack and pitching moment coefficient variation curve

下载: 全尺寸图片幻灯片

图 11 不同来流马赫数下俯仰力矩系数与迎角的迟滞环（κ=0.005）

Figure 11. Pitch moment coefficient hysteresis loops at different Mach numbers and angles of attacks （κ=0.005）

下载: 全尺寸图片幻灯片

图 12 静稳定导数辨识结果

Figure 12. Identification results of static stability derivatives

下载: 全尺寸图片幻灯片

图 13 组合动导数辨识结果

Figure 13. Identification results of combined dynamic derivatives

下载: 全尺寸图片幻灯片

图 14 气动导数预测补偿值与期望补偿值对比和预测补偿值与期望补偿值的残差

Figure 14. Comparison between predicted compensation values and expected compensation values of aerodynamic derivatives and residuals between predicted compensation values and expected compensation values

下载: 全尺寸图片幻灯片

图 15 气动导数预测补偿值与自由飞试验值对比和智能补偿方法与传统计算方法的残差对比

Figure 15. Comparison between predicted compensation values of aerodynamic derivatives and free flight test values and comparison of residuals between intelligent compensation and traditional calculation

下载: 全尺寸图片幻灯片

表 1 非定常模拟计算工况

Table 1. Unsteady simulation calculation conditions

参数	数值
参考密度/（kg/m³）	1.182
参考温度/K	293.15
参考气体比热比	1.4
角速度/（rad/s）	314
谐振振幅/rad	0.0174
减缩频率	0.005

下载: 导出CSV

表 2 谐和振荡法计算动导数结果

Table 2. Results of calculating aerodynamic derivatives using the harmonic oscillation method

Ma	$C_{\mathrm{m}}^\alpha $	$C_{\mathrm{m}}^{\dot \alpha } + C_{\mathrm{m}}^{\omega _{\textit{z}}}$
1.58	−41.08	−523.52
2	−22.78	−420.70
2.5	−16.77	−348.54
3	−13.32	−299.71
3.5	−11.02	−256.93
4	−9.27	−242.42
4.5	−7.91	−257.15

下载: 导出CSV

表 3 迟滞环法计算动导数结果

Table 3. Results of calculating aerodynamic derivatives using the hysteresis loop

Ma	$C_{\mathrm{m}}^\alpha $	$C_{\mathrm{m}}^{\dot \alpha } + C_{\mathrm{m}}^{\omega _{\textit{z}}}$
1.58	−35.61	−527.93
2	−23.59	−427.60
2.5	−17.29	−354.72
3	−13.70	−300.81
3.5	−11.31	−259.75
4	−9.51	−230.03
4.5	−8.09	−203.31

下载: 导出CSV

表 4 递推最小二乘法计算动导数结果

Table 4. Results of calculating aerodynamic derivatives using the recursive least squares

Ma	$C_{\mathrm{m}}^\alpha $	$C_{\mathrm{m}}^{\dot \alpha } + C_{\mathrm{m}}^{\omega _{\textit{z}}}$
1.58	−35.64	−527.85
2	−23.60	−429.52
2.5	−17.83	−355.02
3	−13.72	−301.81
3.5	−11.30	−263.15
4	−9.50	−235.34
4.5	−8.09	−211.17

下载: 导出CSV

表 5 神经网络智能补偿动导数的计算误差

Table 5. Neural network intelligent compensation for aerodynamic derivative calculation errors

误差类型	数值
平均绝对误差	0.75783
均方误差	1.11820
均方根误差	1.05750

下载: 导出CSV

参考文献(27)

[1]	曹爽, 叶正寅, 李恒. 高超声速飞行器非线性纵向运动稳定性研究[J]. 飞行力学, 2021, 39(1): 12-17. CAO Shuang, YE Zhengyin, LI Heng. Research on nonlinear longitudinal motion stability of hypersonic vehicle[J]. Flight Dynamics, 2021, 39(1): 12-17. (in Chinese CAO Shuang, YE Zhengyin, LI Heng. Research on nonlinear longitudinal motion stability of hypersonic vehicle[J]. Flight Dynamics, 2021, 39(1): 12-17. (in Chinese)
[2]	李正洲, 高昌, 肖天航, 等. 超/高超声速飞行器动态稳定性导数极快速预测方法[J]. 航空学报, 2020, 41(4): 123545. LI Zhengzhou, GAO Chang, XIAO Tianhang, et al. Extremely efficient prediction technique of dynamic derivatives for super/hypersonic flight vehicles[J]. Acta Aeronautica et Astronautica Sinica, 2020, 41(4): 123545. (in Chinese LI Zhengzhou, GAO Chang, XIAO Tianhang, et al. Extremely efficient prediction technique of dynamic derivatives for super/hypersonic flight vehicles[J]. Acta Aeronautica et Astronautica Sinica, 2020, 41(4): 123545. (in Chinese)
[3]	HUI W H, TOBAK M. Unsteady Newton-Busemann flow theory Ⅱ bodies of revolution[J]. AIAA Journal, 1981, 19(10): 1272-1273. doi: 10.2514/3.7860
[4]	SEIFF A. Secondary flow fields embedded in hypersonic shock layers[R]. NASA TN D-1304, 1962.
[5]	ERICSSON L E. Unsteady aerodynamics of an ablating flared body of revolution including effect of entropy gradient[J]. AIAA Journal, 1968, 6(12): 2395-2401. doi: 10.2514/3.5000
[6]	ERICSSON L E. Unsteady embedded Newtonian flow as basis for nose bluntness effect on aerodynamics of hypersonic slender bodies[J]. Astronautica Acta, 1973, 18(3): 309-330.
[7]	卢学成, 叶正寅, 张伟伟. 超声速、高超声速飞行器动导数的高效计算方法[J]. 航空计算技术, 2008, 38(3): 28-31. LU Xuecheng, YE Zhengyin, ZHANG Weiwei. A high efficient method for computing dynamic derivatives of supersonic/hypersonic aircraft[J]. Aeronautical Computing Technique, 2008, 38(3): 28-31. (in Chinese LU Xuecheng, YE Zhengyin, ZHANG Weiwei. A high efficient method for computing dynamic derivatives of supersonic/hypersonic aircraft[J]. Aeronautical Computing Technique, 2008, 38(3): 28-31. (in Chinese)
[8]	史爱明, 杨永年, 叶正寅. 结合CFD技术的跨声速动导数计算方法研究[J]. 西北工业大学学报, 2008, 26(1): 11-14. SHI Aiming, YANG Yongnian, YE Zhengyin. A more accurate method for calculating transonic dynamic derivatives (TDDs) using present state-of-the-art CFD[J]. Journal of Northwestern Polytechnical University, 2008, 26(1): 11-14. (in Chinese SHI Aiming, YANG Yongnian, YE Zhengyin. A more accurate method for calculating transonic dynamic derivatives (TDDs) using present state-of-the-art CFD[J]. Journal of Northwestern Polytechnical University, 2008, 26(1): 11-14. (in Chinese)
[9]	米百刚. 基于CFD的动导数计算及非线性气动力建模技术[D]. 西安: 西北工业大学, 2018. MI Baigang. Dynamic derivative calculation and nonlinear aerodynamic modeling technology based on CFD[D]. Xi’an: Northwestern Polytechnical University, 2018. (in Chinese MI Baigang. Dynamic derivative calculation and nonlinear aerodynamic modeling technology based on CFD[D]. Xi’an: Northwestern Polytechnical University, 2018. (in Chinese)
[10]	米百刚, 詹浩, 王斑. 基于刚性动网格技术的动导数数值模拟[J]. 航空动力学报, 2014, 29(11): 2659-2664. MI Baigang, ZHAN Hao, WANG Ban. Numerical simulation of dynamic derivatives based on rigid moving mesh technique[J]. Journal of Aerospace Power, 2014, 29(11): 2659-2664. (in Chinese MI Baigang, ZHAN Hao, WANG Ban. Numerical simulation of dynamic derivatives based on rigid moving mesh technique[J]. Journal of Aerospace Power, 2014, 29(11): 2659-2664. (in Chinese)
[11]	陶洋, 范召林, 赵忠良. 基于CFD的带控制舵导弹的动导数计算[J]. 航空动力学报, 2010, 25(1): 102-106. TAO Yang, FAN Zhaolin, ZHAO Zhongliang. Predictions of dynamic damping coefficients of basic finner based on CFD[J]. Journal of Aerospace Power, 2010, 25(1): 102-106. (in Chinese TAO Yang, FAN Zhaolin, ZHAO Zhongliang. Predictions of dynamic damping coefficients of basic finner based on CFD[J]. Journal of Aerospace Power, 2010, 25(1): 102-106. (in Chinese)
[12]	常思源, 田中伟, 李广利, 等. 基于气动导数的高压捕获翼飞行器纵向稳定性数值研究[J]. 中国科学: 技术科学, 2024, 54(2): 275-288. CHANG Siyuan, TIAN Zhongwei, LI Guangli, et al. Numerical study on longitudinal stability for HCW aircraft based on aerodynamic derivatives[J]. Scientia Sinica (Technologica), 2024, 54(2): 275-288. (in Chinese doi: 10.1360/SST-2022-0309 CHANG Siyuan, TIAN Zhongwei, LI Guangli, et al. Numerical study on longitudinal stability for HCW aircraft based on aerodynamic derivatives[J]. Scientia Sinica (Technologica), 2024, 54(2): 275-288. (in Chinese) doi: 10.1360/SST-2022-0309
[13]	BHAGWANDIN V, SAHU J. Numerical prediction of pitch damping stability derivatives for finned projectiles[R]. AIAA 2011-3028, 2011.
[14]	SILTON S I. Numerical experiments on finned bodies[R]. AIAA 2014-3021, 2014.
[15]	BUNESCU I, HOTHAZIE M V, PRICOP M V, et al. Numerical investigation of basic finner model in roll motion as complement to the experimental work[R]. AIAA 2024-2348, 2024.
[16]	袁先旭. 非定常流动数值模拟及飞行器动态特性分析研究[D]. 四川绵阳: 中国空气动力研究与发展中心, 2002. YUAN Xianxu. Numerical simulation of non-constant flow and analysis of vehicle dynamic characteristics[D]. Mianyang Sichuan: China Aerodynamics Research and Development Center, 2002. YUAN Xianxu. Numerical simulation of non-constant flow and analysis of vehicle dynamic characteristics[D]. Mianyang Sichuan: China Aerodynamics Research and Development Center, 2002.
[17]	袁先旭, 张涵信, 谢昱飞. 基于CFD方法的俯仰静、动导数数值计算[J]. 空气动力学学报, 2005, 23(4): 458-463. YUAN Xianxu, ZHANG Hanxin, XIE Yufei. The pitching static/dynamic derivatives computation based on CFD methods[J]. Acta Aerodynamica Sinica, 2005, 23(4): 458-463. (in Chinese doi: 10.3969/j.issn.0258-1825.2005.04.012 YUAN Xianxu, ZHANG Hanxin, XIE Yufei. The pitching static/dynamic derivatives computation based on CFD methods[J]. Acta Aerodynamica Sinica, 2005, 23(4): 458-463. (in Chinese) doi: 10.3969/j.issn.0258-1825.2005.04.012
[18]	杨士斌, 郭东, 钱宇, 等. 一种高效的组合动导数分量数值模拟算法[J]. 航空动力学报, 2018, 33(8): 1974-1980. YANG Shibin, GUO Dong, QIAN Yu, et al. An effective method for numerical simulation of individual components of pitching combined dynamic derivative[J]. Journal of Aerospace Power, 2018, 33(8): 1974-1980. (in Chinese YANG Shibin, GUO Dong, QIAN Yu, et al. An effective method for numerical simulation of individual components of pitching combined dynamic derivative[J]. Journal of Aerospace Power, 2018, 33(8): 1974-1980. (in Chinese)
[19]	刘绪, 刘伟, 柴振霞, 等. 飞行器动态稳定性参数计算方法研究进展[J]. 航空学报, 2016, 37(8): 2348-2369. LIU Xu, LIU Wei, CHAI Zhenxia, et al. Research progress of numerical method of dynamic stability derivatives of aircraft[J]. Acta Aeronautica et Astronautica Sinica, 2016, 37(8): 2348-2369. (in Chinese LIU Xu, LIU Wei, CHAI Zhenxia, et al. Research progress of numerical method of dynamic stability derivatives of aircraft[J]. Acta Aeronautica et Astronautica Sinica, 2016, 37(8): 2348-2369. (in Chinese)
[20]	MCGOWAN G, KURZEN M, NANCE R, et al. Computational investigation of pitch damping on missile geometries at high angles of attack[R]. AIAA 2012-2903, 2912.
[21]	HASSAN D, SICOT F. A time-domain harmonic balance method for dynamic derivatives predictions[R]. AIAA 2011-1242, 2011.
[22]	MOHAMED K, NADARAJAH S, PARASCHIVOIU M. Detached-eddy simulation of a wing tip vortex at dynamic stall conditions[J]. Journal of Aircraft, 2009, 46(4): 1302-1313. doi: 10.2514/1.40685
[23]	RAVEH D E. Identification of computational-fluid-dynamics based unsteady aerodynamic models for aeroelastic analysis[J]. Journal of Aircraft, 2004, 41(3): 620-632. doi: 10.2514/1.3149
[24]	VLADISLAV K, NODERER K D. Modeling of aircraft unsteady aerodynamic characteristics: Part 1 postulated models[R]. NASA-TM-109120, 1994.
[25]	GUSTAFSON W. The Newtonian diffuse method for computing aerodynamic forces[R]. Washington DC: Lockheed Aircraft Corporation Report No. IMSD 5132, 1958.
[26]	WEINACHT P, STUREK W B. Computation of the roll characteristics of a finned projectile[J]. Journal of Spacecraft and Rockets, 1996, 33(6): 769-775. doi: 10.2514/3.26836
[27]	王钦超, 李世超, 高宏力, 等. 高超声速风洞短时气动力智能辨识算法研究[J]. 力学学报, 2022, 54(3): 688-696. WANG Qinchao, LI Shichao, GAO Hongli, et al. Research on intelligent identification algorithms for short-term aerodynamics of hypersonic wind tunnels[J]. Chinese Journal of Theoretical and Applied Mechanics, 2022, 54(3): 688-696. (in Chinese WANG Qinchao, LI Shichao, GAO Hongli, et al. Research on intelligent identification algorithms for short-term aerodynamics of hypersonic wind tunnels[J]. Chinese Journal of Theoretical and Applied Mechanics, 2022, 54(3): 688-696. (in Chinese)