舵面故障会影响飞机的操纵性,为实现飞机的控制重构需在线诊断出舵面的故障信息。针对多操纵面飞翼构型飞机,提出了一种在线诊断其舵面故障的方法。飞翼构型飞机舵面故障在线诊断需辨识的操纵导数个数较多,直接使用最小二乘法难以精确辨识得到各操纵导数。基于限定记忆最小二乘法,采用分组辨识法对飞机的操纵导数进行辨识,减少了每次辨识操纵导数的个数,使得各操纵导数可精确辨识得到。针对舵面卡死故障的特点,首次提出了假定卡死故障情形法,使得被卡死舵面及其卡死偏角均可得到精确诊断,进而可诊断出各舵面各种类型故障的信息。

It is essential for flight control reconfiguration to obtain on-line information of the control surface failures as they will influence the controllability of an aircraft. This article provides a method to diagnose on-line the fault of multi-control effector flying wing control surfaces. The number of the control derivatives needed to be estimated in on-line fault diagnosis of such a flying wing is more than that of a conventional aircraft. Consequently, it is difficult to estimate the control derivatives of the flying wing accurately by the least square method. Based on the limited memory least square method, the control derivatives of the flying wing are identified in groups. Therefore, the number of the control derivatives needed to be estimated in each group is reduced and thus the derivatives can be estimated precisely. A method of assuming a deadlock case is presented for the characteristic of surface deadlock control, by which the control surface of the deadlock and its deflection can be identified precisely. Thus the fault information of control surfaces of different types is acquired.