Research Of Robust Fault Diagnosis And Active Fault Tolerant Control Technology On Dynamic System

Fault diagnosis and fault tolerant control technology play a key role on the reliability of the dynamic system.With the development of the control theory and the applications of the multi-disciplinary,the engineers advance the high performance index,high reliability and high safe standard for the equipments in the system design.The fault diagnosis and fault tolerant based on analytical model methods have great efficiency but without the other hardware redundancy.The accuracy of the dynamic system described decides effectiveness of fault diagnosis and fault tolerant algorithms based on analytical model.Because the systems are nonlinear intrinsically,the model can describe the system thoroughly.Therefore the complex systems also raise the rates of fault occurrence as well.Operating the equipments or reducing the system performance index properly can avoid the system break down,especially for the safe critical systems.Therefore,fault diagnosis and fault tolerant control theory have the important meaning for the theory research and applications.This paper researches the robust fault diagnosis and active fault tolerant control technology for uncertainty linear and nonlinear systems respectively.The main works are as follows:TheH_-/H_?fault diagnosis observer for the linear system with model uncertainties and unknown input disturbances that its disturbances distribution matrix is known is proposed.The algorithm improves the robust against the disturbances and the sensitivity for faults.The design of the fault diagnosis observer is converted into the linear matrix inequality constraints formulations by the Schur complement lemma.In order to decrease the alarm rates in the fault diagnosis,the adaptive threshold design method is proposed.According to the convex optimization problems formulations,it is easy to achieve the fault diagnosis observer.Besides,the proposed algorithms guarantee the optimization trade-off between the robust against the unknown inputs disturbances and the sensitivity for faults.The robust fault diagnosis algorithm is proposed for the linear system with the model uncertainties and unknown input disturbances that its distribution matrix is unknown.The model estimation prediction unknown input observer is designed,and it improves the deficiency of the traditional unknown input observer on assumption that the disturbances distribution matrix is known.The object function about the disturbances distribution matrix is constructed by the convex quadratic programming optimal principle.An optimal estimation algorithm to the unknown input disturbances distribution matrix which meet the objective function minimum and a method of designing the fault diagnosis system gain matrix which makes the variance of the state estimation error minimum are proposed.The design of generalized fault diagnosis observer(GFDO)and active fault tolerant control system(AFTCS)scheme are proposed for linear system with external disturbances and actuator fault.Unlike the traditional studies assume that external disturbances are norm bounded,the novel external disturbances constraint condition that satisfies the system stability is derived based on Gronwall Lemma.And then,the GFDO which can estimate system states and fault simultaneously is designed by two performances index defined.Besides,the AFTCS is designed in that obtaining desired performances under the fault case.When the fault is diagnosed by GFDO,the normal controller switches to AFTCS.A robust fault diagnosis scheme for Lipschitz nonlinear system is designed and a novel algorithm for a robust fault diagnosis observer is proposed in this paper.The robust performance index is defined to ensure the robustness of the observer designed.The norm of most unknown input disturbances are assumed bounded at present.However,some systems are proved unstable under traditional assumptions.In the proposed algorithm,the external disturbances constraint condition that satisfies the system stability is derived based on Gronwall Lemma.The design procedure of the observer proposed is implemented by pole assignment.Adaptive threshold is generated using the designed observer.The active adaptive composite fault tolerant control algorithm is proposed for the Lipschitz nonlinear system with the model uncertainties and actuator fault.The algorithm proposed includes the external disturbances estimation observer and generalized robust fault diagnosis observer design.Compared with the traditional methods that assume the disturbances are known or apply the constant gain matrix fault tolerant controller,the paper has less constraint conditions for the system considered and has more extent applications.Besides,the active adaptive composite fault tolerant controller not only can assure the faulty system function well but also can assure the ideal state response results for the faulty system.