Attitude Fault-tolerant Control Of Launch Vehicle Under Thrust Drop Fault

The launch vehicle is an indispensable vehicle for space activities such as manned spaceflight and commercial launch.The wide flight range of the launch vehicle,the drastically changing aerodynamic environment,and the impact of elastic vibration caused by the slender structure make the design of the control system more complicated.And the thrust failure of the launch vehicle may also reduce the tracking performance of the control system.In this paper,fault-tolerant control under the thrust fault of the launch vehicle is studied,with the aim of improving the tracking performance of the control system.The research content is as follows:First,a six-degree-of-freedom nonlinear mathematical model of the launch vehicle is derived,and the impact of the thrust drop failure on the control system is analyzed.Secondly,after a thrust drop fault occurs,a large disturbance torque will be generated.If the disturbance torque is not suppressed,it will cause a large attitude tracking error.In order to improve the attitude tracking performance after thrust decline fault,this paper designs a sliding mode observer to observe the disturbance,and feeds the observed interference back to the controller to suppress the disturbance.After the thrust drop fault occurs,the control force provided by the faulty engine will decrease,which reduces the total control force applied to the launch vehicle.In order to keep the total control force applied to the launch vehicle before and after the fault unchanged,this paper uses cascade pseudo-inverse control allocation.The cascade pseudo-inverse control allocation compensates the reduced control amount of the faulty engine by changing the control amount of the normal engine.Finally,simulations have verified that the fault-tolerant method can improve the attitude tracking performance of the launch vehicle control system after the thrust drop fault.The launch vehicle usually has a slender structure,which makes the launch vehicle suffer greater impact from the elastic vibration during flight.In order to suppress elastic vibration,the conventional method is to use a PD controller and a notch filter,which will affect the attitude tracking accuracy of the attitude control system.In order to improve the tracking accuracy of the attitude control system and achieve fault-tolerant control,this paper uses sliding mode fault-tolerant control based on the generalized proportional integral observer.The parameters of the observer are obtained through the optimization of the beetle whisker algorithm,and the parameters of the controller are determined through stability analysis.Finally,simulations verify that the fault-tolerant method can effectively ensure the stability and tracking of the launch vehicle attitude control system.Due to actuator installation deviation,external interference and other reasons,there is often uncertainty in the control distribution efficiency matrix.Convex optimal control allocation can be used to reduce the impact of this uncertainty.However,considering the needs of real-time control,the calculation of convex optimal control allocation takes a long time.In order to improve the real-time and robustness of control allocation,this paper proposes a neural network control allocation based on convex optimization.Firstly,the input and output data of convex optimal control allocation under different fault conditions are obtained,and these data will be used as training samples to train the neural network.Finally,a neural network control allocation based on convex optimization is formed.Simulation results show that the neural network control allocation method based on convex optimization can complete the control allocation task in real time when there is uncertainty in the control allocation efficiency matrix.