基于神经网络的自修复飞行控制系统故障认定

Fault recognition of self-repairing flight control system using neural networks

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 。

The resource of maintenance diagnostic information in the indirect self repairing flight control system is from the fault detection module, which disadvantage is to correspond the real time of controller and the detection precision is difficult. And there is no fault detection module in the direction self repairing flight control system, but it needs the module of fault recognition of control system A method of training neural networks to recognize fault using augment information from the direction self repairing flight control system is presented. A sub neural network is designed to recognition the fault again and the accuracy of recognition is improved. Results of simulation test have demonstrated that the method in this article works well. The Self Repairing Flight Simulation Platform verifies this flight simulation system. It is a new scheme for getting the flight maintenance diagnostic information and alerting pilots.