间接自修复飞行控制系统的维护诊断信息来源是故障检测 ,它的缺点是控制的实时性与检测精度难以协调 .直接自修复飞行控制系统取消了故障检测模块 ,但是需要进行控制系统的故障认定 .本文利用直接自修复飞行控制系统的补偿信息训练神经网络来进行故障认定和二次故障认定 ,使故障认定算法处于告警实时性级别 ,从而提高认定精度 .故障飞机自修复飞行仿真结果表明 ,文中所采用的神经网络故障认定方法是行之有效的 .此飞行仿真系统通过了自修复飞行仿真平台的验证 ,为飞行维护诊断信息获得和驾驶员告警提供了一种新思路 The self-repairing flight control system maintenance diagnosis information source is the fault detection, but its shortcoming is that the real-time control and the detection accuracy are difficult to coordinate.The direct self-repairing flight control system cancels the fault detection module, but needs to carry on the control system breakdown confirmation. In this paper, the compensation information of direct self-repairing flight control system is used to train the neural network to identify the fault and identify the secondary faults, so that the fault recognition algorithm is in the real-time level of the alarm, so as to improve the accuracy of the verification.The fault self-healing flight simulation results show that, The neural network fault identification method is effective.This flight simulation system has passed the verification of the self-repairing flight simulation platform, which provides a new idea for the flight maintenance diagnostic information acquisition and driver alarm