摘要
在国内首次利用成熟的低成本火箭弹平台,开展超声速(马赫数>3)飞行试验的嵌入式大气数据传感系统技术研究。针对嵌入式大气数据传感系统的求解算法,测量系统和误差影响等关键技术问题,建立基于神经网络技术的求解算法和设计飞行试验方案,并完成飞行试验和数据分析研究。研究结果表明基于神经网络技术的求解算法具有较好的鲁棒性和较高的求解精度。测量结果与雷达测量结果基本吻合,验证了算法设计;测量结果相对于雷达测量结果,静压平均相对误差约为5.2%,最大相对误差18.8%;马赫数平均相对误差4.2%,最大相对误差14.9%。攻角和侧滑角的测量结果与理论弹道结果变化趋势接近。研究结果可为相关飞行试验技术研究提供参考。
The flush air data sensing（ FADS） system technology was researched by using mature and low cost supersonic（ Ma 〉 3） rocket projectile platform flight test for the first time in domestic. According to the key technology problems of algorithm for FADS solving and measurement system and the error influence,that FADS solving algorithm was built based on neural network technology and flight test scheme was designed,and the flight test and data analysis were done. Research results showed that the FADS algorithm based on neural network technique had good reliability and precision; The measurement results were basically consistent with the radar measurements results,which verified the algorithm design. The FADS measurement results relative to the radar measurement results,average relative error of static pressure was 5. 2% and the maximum relative error was 18. 8%; the average relative error of Ma number was4. 2% and the maximum relative error was 14. 9%; FADS measurement results of attack angle and side slip angle were close to the results theoretic trajectory. The research results could provide reference for the relative study of flight test technology.
出处
《宇航学报》
EI
CAS
CSCD
北大核心
2015年第10期1195-1202,共8页
Journal of Astronautics
基金
国家自然科学基金(11372040)
关键词
飞行试验
火箭弹
嵌入式大气数据传感系统
神经网络
计算流体力学
Flight test
Rocket projectile
Flush air data sensing system
Neural networks
Computational fluid dynamics
