Sliding Mode Observer-based Fault Diagnosis For Flight Control Systems

As the scale and complexity of the control system continue expanding,the requirements of system reliability and safety are also increasing.The aerodynamic characteristics is complex and flight conditions are special in the aeronautical aircraft,which also suffers from external disturbances,actuator faults and sensor faults.Therefore,it is a practical work to study the fault diagnosis of flight control systems.Considering the nonlinear factors such as modeling uncertainty and external disturbance,the types of faults are analyzed in the flight control systems of unmanned aerial vehicle(UAV).Based on sliding mode observer theory,the fault diagnosis algorithm of the flight control system is designed.Main research work is presented as follows:Firstly,the research background and significance of the paper,the concept and classification of the fault are introduced.The recent development of the observer-based fault diagnosis technology is presented.Then,a sliding mode observer-based fault diagnosis method is proposed to detect the actuator faults for a class of linear systems with disturbance and model uncertainty.The linear matrix inequalities(LMI)are used to solve the design problem of the sliding mode observer-based control.The stability of the designed sliding mode observer can be proved by Lyapunov stability theory.Secondly,a method based on neural network sliding mode observer for detecting and estimating the system actuator fault is proposed to solve the incipient fault diagnosis problem of the flight control system,and the stability of the neural network-based sliding mode observer is proved.The linear system is decoupled into two subsystems by coordinate transformation,with one dedicated to design the neural network observer for detecting the incipient fault,which does not contain disturbances,and another to design the sliding mode observer to attenuate the unknown disturbances and achieve the robustness of the system,which is affected by both the disturbances and incipient fault.Thirdly,a robust sliding mode observer-based fault diagnosis method is proposed to estimate sensor faults and detect the actuator faults for a class of uncertain nonlinear systems.The original system with sensor faults,actuator faults,and unknown inputs is transformed into an augmented singular system by taking the actuator fault vector as a part of an extended state vector,which has only actuator faults and unknown inputs.For the constructed singular system,a robust sliding mode observer is used to simultaneously estimate the states of the original system and sensor faults.By solving an optimization problem,the observer gain matrices are calculated according to the linear matrix inequalities.Meanwhile,the actuator fault detector is operating when an actuator fault occurs.Finally,according to the research content of the subject,in order to provide a simulation for the fault diagnosis study of the aeronautical aircraft flight control systems,the hardware and software system of the semi-physical experimental platform named Aircraft Fault Diagnosis Experimental Platform is constructed.The sliding mode observer-based fault diagnosis algorithm proposed in this paper is applied to the platform as an example.The experimental results not only demonstrate the rationality of the platform design and construction,but also verify that the designed sliding mode observer can achieve good performance of fault diagnosis,,maintain effective tracking,and recover the system performance.