Nonliear Fault Tolerant Control For Near Space Vehicle

Near space vehicle (NSV) plays an important role on national military security and theexploitation of space, and it have attracted more and more attention from many countries in the world.The NSV shows some unique characteristics during its operation, such as, multi-flighing states,multi-tasking modes, large flight envelop and so on. Moreover, the especial and complex flightenvironment makes it a challenging problem to guarantee its security and reliability. Thus, the faulttolerant control design for NSV flight control system is an innovative and challenging topic. In thisdissertation, it carries on some investigations in NSV dynamical modeling and faulty modelinganalysis, and it mainly focuses on the fault diagnosis and fault tolerant control design for NSV controlsystem based on advanced nonlinear control.First, the flight motion nonlinear model of NSV with six degrees of freedom and twelve statevariables is established and analyzed. The coefficients of aerodynamic force and moment arepresented as nonlinear fuctions of Mach number, angle of attack, and control surface deflections, andthe moment of inertia and center of gravity are functions of the vehicle's weight. The propulsiondevice uses the combination of athodyd and rocket engine. The state responses of open-loop dynamicsindicate that the presented model can describe the characteristics of NSV. Thus, the model can be usedas the theoretical analysis and simulation platform.Then, all the possible fault types of flight control system, including actuator fault, sensor faultand system fault, are presented and analyzed, and the characteristics of faulty models are provided inthe form of parameterized model. Otherwise, the control allocation problem of NSV dynamics isanalyzed, and actuator fault modes are also introduced. It makes a contribution to designing faultdiagnosis scheme and fault tolerant control.Next, a T-S fuzzy model for the nonlinear dynamics of NSV is established based on T-S fuzzymodeling technique, by selecting proper fuzzy rules and operation points. Then, a fault tolerantcontrol (FTC) strategy is proposed for NSV with actuator fault using sliding mode observer technique,and a set of sliding mode observers are designed to generate residuals which contain the faultinformation. Then, a novel fault diagnostic algorithm is given to estimate the actuator fault occurred.Utilizing the obtained on-line fault estimation information, a fault accommodation scheme isdeveloped to compensate for the effects of actuator fault. Based on Lyapunov stability theory, asufficient condition to guarantee the stability of closed-loop system is derived in terms of linearmatrix inequalities (LMIs), which can be easily solved by Matlab LMI toolbox. Finally, simulationresults are presented to demonstrate the efficiency of the proposed approach.Later, a fault diagnosis and active FTC approach is investigated based on adaptive methodology to deal with actuator fault problem of NSV. An adaptive fault diagnosis observer (AFDO) is firstdesigned, the residual is defined as the norm of the output error between AFDO and actual system todetect the actuator fault, and an adaptive fault estimation algorithm is proposed to estimate the fault. Afeasible adaptive algorithm is explored and it is shown that observer parameters can be obtained bysolving LMIs. An active FTC is designed by utilizing the diagnostic fault information to compensatefor the loss of actuator effectiveness. Finally, simulation results on the longitudinal model of NSV arepresented to demonstrate that the proposed approach not only can detect and estimate the faultefficiently, but also accommodate and tolerate the fault.Further, a FTC scheme based on backstepping and neural network (NN) methodology isproposed for NSV attitude dynamical model. The linearly parameterized radial basis function (RBF)NNs are employed to approximate unknown system faults and disturbance respectively, and thenetwork weights are adapted using adaptive on-line parameter-learning algorithms. Then an adaptivebackstepping based FTC is designed to compensate for the effect of system faults. The asymptoticalstability of the closed-loop system and uniform boundedness of the state tracking errors are provedaccording to Lyapunov theory. Finally, the designed strategy is applied to NSV attitude dynamics, andsimulation results are presented to demonstrate that the proposed approach can compensate for theeffects of both fault and uncertainties. The implantation of NN expands the applications ofbackstepping method.Finally, a robust tracking FTC scheme based on adaptive sliding mode technique is proposed. Amodified sliding model tracking control approach, using a continuous strategy of integral slidingmode control (SMC), is designed to improve the position tracking precision. This scheme can solvethe chattering problem without loss of robustness and control accuracy. The SMC is robust tounmeasurable uncertainties and disturbances, and the actuator fault is compensated for by an integralaction term. The stability of the closed-loop system is proved according to Lyapunov theory. To verifythe effectiveness and application value of the proposed control scheme, a numerical simulation isperformed on NSV attitude system in re-entry mode. Simulation results demonstrate that the systemhas good tracking performance in spite of the actuator fault and system disturbances.