基于神经网络滑模观测器的飞控系统故障诊断  

Research on fault diagnosis algorithm of flight control system based on neural network sliding mode observer

作　　者：权璐 姜斌[1] 杨蒲[1] QUAN Lu;JIANG Bin;YANG Pu(College of Automation Engineering. Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China)

机构地区：[1]南京航空航天大学自动化学院,南京210016

出　　处：《扬州大学学报（自然科学版）》2019年第2期51-55,共5页Journal of Yangzhou University：Natural Science Edition

基　　金：国家自然科学基金资助项目(61773201)

摘　　要：为了提高控制系统的安全性和可靠性,针对含有扰动的飞控系统执行器微小故障诊断问题,基于神经网络滑模观测器,提出了一种对系统执行器故障进行检测及估计的办法,并证明了该神经网络滑模观测器的稳定性.通过坐标变换将线性系统解耦成两个子系统:一个子系统不含扰动,对其设计神经网络观测器,实现对微小故障的检测;另一个子系统受到扰动和微小故障的双重影响,对其设计滑模观测器,消除未知扰动的影响,保证系统的强鲁棒性.通过飞控系统故障诊断试验平台验证了该方法的有效性.In order to improve the safety and reliability of the control system, a approach of fault detection and estimation based on neural network sliding mode observer for flight control system is proposed for the problem of actuator fault of the flight control system with disturbances, and the stability of the designed neural network sliding mode observer is proved. The linear system is decoupled into two subsystems by coordinate transformation. A subsystem does not contain disturbances and a neural network observer is designed to detect the incipient fault, another subsystem is affected by both the disturbances and incipient fault, and a sliding mode observer is designed to eliminate the influence of the unknown disturbances and guarantee the robustness of the system. Finally, the proposed actuator fault diagnosis method is tested in a flight control system fault diagnosis test platform, and simulation results demonstrate the effectiveness of the method.

关 键 词：微小故障检测 神经网络滑模观测器 飞控系统 

分 类 号：TP273.3[自动化与计算机技术—检测技术与自动化装置]