极大似然法在飞机起飞性能参数辨识中的应用  

Application of maximum likelihood estimation to the parameter identification of airplane take-off performance

作　　者：于雪梅[1] 程伟[1] 谷伟岩 

机构地区：[1]北京航空航天大学航空科学与工程学院,北京100191 [2]中航通用飞机有限责任公司试飞交付中心,广东珠海519015

出　　处：《飞行力学》2012年第3期201-205,共5页Flight Dynamics

摘　　要：以Y12飞机起飞性能飞行试验实测数据为例,研究了利用极大似然法进行飞机起飞性能参数辨识的问题。首先通过理论分析建立了Y12飞机双发起飞性能数学模型;然后从简化灵敏度导数计算角度提出了以飞机在起飞滑跑过程中速度增量为观测量的观测方程,分析给出了待辨参数和利用极大似然法进行参数识别的方法;最后针对极大似然法的几个要素,如迭代初值、代价函数、迭代次数、样本长度、试验数据段的选取等对辨识结果的影响进行了讨论和分析。结果表明,利用极大似然法可以快速、准确地辨识出起飞性能所需的参数,并具有较高的精度,在今后的性能飞行试验数据分析中应注重推广和应用。Take take-off performance flight test data of airplane Y12 as example,we use maximum likelihood method to investigate the take-off performance parameter identification.First,the mathematical model of all-engine take-off performance on airplane Y12 was established via theoretical analysis.Then,the observed equation with speed increment as a observed quantity was put forward to simplify the calculation of sensitivity derivatives,and the identification parameters and method of using the maximum likelihood method were also given.Finally,the influence of several elements about maximum likelihood estimation,such as initial iteration value,cost function,iteration number,sample length,and the selection of the test data segment,etc.to identification results was also discussed and analyzed.The results show that maximum likelihood method can identify take-off performance parameters quickly and accurately,and results have a higher precision,so the method should be generalized and applied to other kind of flight test data and of other airplanes.

关 键 词：极大似然法 飞行试验 参数辨识 观测方程 

分 类 号：V212.1[航空宇航科学与技术—航空宇航推进理论与工程]