航天推进技术研究院主办
[1]刘巾杰,许琪琪,扶山川,等.小样本下煤油加注温升模型参数估计方法[J].火箭推进,2019,45(06):90-94.
LIU Jinjie,XU Qiqi,FU Shanchuan,et al.Estimation method for coefficient of kerosene filling temperature rise model under small sample[J].Journal of Rocket Propulsion,2019,45(06):90-94.
点击复制
LIU Jinjie,XU Qiqi,FU Shanchuan,et al.Estimation method for coefficient of kerosene filling temperature rise model under small sample[J].Journal of Rocket Propulsion,2019,45(06):90-94.
小样本下煤油加注温升模型参数估计方法
《火箭推进》[ISSN:1672-9374/CN:CN 61-1436/V]
卷:
45
期数:
2019年06期
页码:
90-94
栏目:
测控与试验
出版日期:
2019-12-25
- Title:
- Estimation method for coefficient of kerosene filling temperature rise model under small sample
- 文章编号:
- 1672-9374(2019)06-0090-05
- 分类号:
- V511
- 文献标志码:
- A
- 摘要:
- 为消除火箭煤油加注实测温度与现行加注温升模型计算理论温度之间的偏差,准确估计煤油加注温升实现精确加注、精准控制,在靶场实测数据仅为小样本数据的现实条件下,采用贝叶斯回归方法,将经验参数作为先验信息给出待估参数的先验分布,由实测数据计算贝叶斯风险,通过最小化风险函数实现对煤油加注温升模型的参数优化估计。数值实验结果表明,与给定的经验参数和经典最小二乘估计参数相比,贝叶斯估计参数估算的加注结束温度与实测温度误差显著减小,且随着可用的数据样本容量相应增大,贝叶斯回归分析的参数估计精度呈现收敛趋势。贝叶斯方法进行煤油加注温升模型参数估计有效提高了加注温升模型精度,为提高推进剂加注精确度提供了方法基础。
- Abstract:
- In order to eliminate the deviation between the actual rocket’s kerosene filling temperature and the theoretical result calculated by current filling temperature rise model, and estimate kerosene temperature for accurate injection and precise control, Bayesian regression method was proposed to estimate fluctuation coefficient of the temperature rise model based on a very few sample data at launching site.It takes the empirical parameters as the prior information to give a distribution of parameters to be estimated, and uses real data to calculate Bayesian risk.The parameters of the filling temperature rise model were optimized by minimizing the risk function.Numerical results show that, compared with the given experience parameters and least square results, the kerosene temperature calculated by Bayes regression is much closer to the actual result.With the increase of available data sample capacity, the parameter estimation accuracy of Bayesian regression tends to converge.Bayesian regression can improve the precision of filling temperature rise model efficiently, and provides a foundation to improve the accuracy of propellant filling.
参考文献/References:
[1] 杨寒, 吴志坚.变频泵在地面煤油加注系统中的应用[J].火箭推进, 2015, 41(5):111-114.YANG H, WU Z J.Application of variable frequency pump in kerosene filling system on ground[J].Journal of Rocket Propulsion, 2015,41(5):111-114.
[2] 苏永芝, 李俊.常规加注虚拟仿真系统应用分析[J].北京信息科技大学学报(自然科学版), 2017, 32(3):75-79.
[3] 马原, 王磊, 孙培杰, 等.微重力环境低温流体无排气加注过程数值研究[J].低温工程, 2017(4):7-13, 20.
[4] 王正明, 易东云.测量数据建模与参数估计[M].长沙:国防科技大学出版社, 1996.
[5] 何晓群, 刘义卿.应用回归分析[M].北京:中国人民大学出版社, 2001.
[6] 朱蕾, 杨毅.改进贝叶斯算法在配网故障诊断中的研究[J].三峡大学学报(自然科学版), 2018, 40(2):72-77.
[7] 谢仕炜, 胡志坚, 吴方劼, 等.基于递推最小二乘法的多端口外网静态等值参数辨识方法[J].电力系统保护与控制, 2018, 46(3):26-34.
[8] 曹宁, 王晓锋, 徐亚栋, 等.基于逐步回归法的车载炮射击精度影响因素分析[J].南京理工大学学报, 2013, 37(4):557-560.
[9] 徐群.非线性回归分析的方法研究[D].合肥:合肥工业大学, 2009.
[10] PENG X Y, YAN Z Z.Bayesian estimation for generalized exponential distribution based on progressive type-I interval censoring[J].Acta Mathematicae Applicatae Sinica, English Series, 2013, 29(2):391-402.
[11] LIN Y J, LIO Y L.Bayesian inference under progressive type-I interval censoring[J].Journal of Applied Statistics, 2012, 39(8):1811-1824.
[12] 韦来生, 张伟平.贝叶斯分析[M].合肥:中国科学技术大学出版社, 2013.
[13] 徐芹.正态线性回归模型下Bayes估计的优良性[J].重庆工商大学学报(自然科学版), 2008, 25(4):347-349, 402.
[14] 田博, 陈舜杰, 周雯.基于最小风险贝叶斯决策理论的在线评价排名方法研究[J].上海管理科学, 2018, 40(2):91-95.
[15] 赵颖.国外新型运载火箭研制计划与发展现状[J].国际太空, 2002(10):22-27.
备注/Memo
收稿日期:2018-07-23; 修回日期:2018-12-16作者简介:刘巾杰(1989—),女,硕士,工程师,研究领域为液体火箭靶场数据处理
更新日期/Last Update:
2019-12-20