Research On Fault-tolerant Control Algorithm For Hypersonic Vehicle

Hypersonic vehicle has broad development prospects and high research significance.As an important component of the hypersonic vehicle,the actuator will impact the stability of the system if facing failure.This paper studies the hypersonic vehicle while facing the situation that actuator gain loss fault and stuck fault.It includes an adaptive fault-tolerant controller,an adaptive neural network fault-tolerant controller and a fault-observer-based fault-tolerant controller.These controllers ensure the stability of the system while tracking the commands.Firstly,the background and significance of the research are expounded.According to the research progress at home and abroad,the concept and significance of the hypersonic vehicle fault,the research progress in fault-tolerant control and the main technical difficulties are presented.Through precise feedback linearization process,the model of hypersonic vehicle with actuator fault(gain loss fault and stuck fault)could be built based on the longitudinal dynamics model,which is separated from the hypersonic vehicle dynamics model.Secondly,considering the hypersonic vehicle actuator gain loss fault,an adaptive fault-tolerant controller based on the established longitudinal cruise segment model is proposed.The sliding-mode is designed based on the integral sliding surface and the double power reaching law.The adaptive update laws are designed to estimate the actuator gain loss factor.And the chattering is considered in the design,which effectively suppresses the chattering effect existing in sliding mode control.Numerical simulation results demonstrate the validity and robustness of the proposed method in the case of no fault,constant fault and time-varying fault.Thirdly,this paper proposes an improved adaptive neural network fault-tolerant control method based the adaptive fault-tolerant controller.Combined with the adaptive algorithm and the improved Radial Basis Function Neural Network method,this controller can make the hypersonic vehicle track the desired commands with simple algorithm structure,and does not need to estimate the value of the actuator fault or calculate a mass of real-time data.The simulation results show that the improved adaptive neural network fault-tolerant controller can handle faults faster and better than the adaptive fault-tolerant controller.Finally,considering the actuator gain loss fault and stuck fault,a fault-observer-based fault-tolerant controller is proposed in this paper.Under the combined effects of parameter perturbation of the aircraft system model,external disturbances and limited control input,an improved fast fault observer is designed to estimate the integrated faults in real-time.The fault-tolerant controller based on the fast fault observer can quickly solve the issue caused by actuator gain loss fault and stuck fault with the fault unknown.