Research On Aircraft Self-repairing Controlwith Multiple Faults Based On Adaptive Control

Flight control system is an important part of aircraft design. In the process of flight, aircraft is sensitive to external disturbance and other uncertain factors, such as faults. This paper proposes adaptive fault-tolerant control method based on adaptive control. Three different planes are discussed with model uncertainties, external disturbances and actuator faults. The main research content is as follows:We develop an adaptive fault-tolerant control for fixed-wing flight vehicles with multiple faults. Longitudinal dynamics of a hypersonic air vehicle with uncertain parameters and actuator faults are considered. Nonlinear dynamic inverse is chosen as the basic controller and the robust controller is designed based on SVM and adaptive control theory. Considering both the slow variation faults and fast abrupt faults, we design adaptive laws separately to reconstruct and update the controllers. System traceability and stability are strictly proved. Robustness of the proposed system makes it suitable for fault-tolerant design of the advanced aircraft.Then, we amend a simple linear model of a Quanser helicopter. Complex coupling of attitude angles are considered. By mechanism analysis method, nonlinear differential equations are accurately described of pitch, roll and yaw axes. These equations can be concluded as a 6-order nonlinear state space equation. Linearization work is done according to the new model. By MRAC(model reference adaptive control), modeling and other system uncertainties are compensated effectively and anti-interference ability of the system are enhanced.Based on the nonlinear equations of a quadrotor, position tracking with unknown parameters and fault-tolerant control scheme is proposed in this paper. First, multi-loop structure is designed with multiple time-scale analysis. Then, a sliding mode fault observer is added into each loop to accommodate the effects of all the uncertainties. This controller shows great applicability because of the simple structure. The only information the controller needs to know is the limits of uncertainties. Since the above controller can not guaranteed robustness when the system model is unknown, an online adaptive estimation method is proposed to learn the parameter information.Finally, the simulation is presented with several situations to verify the effectiveness and feasibility of the proposed scheme.