Research Of Fault Accommodation Method For Nonlinear System Based On Extended State Observer And Inverse System

With the rapid development of contemporary technology,various of control systems have become more and more complex and large in practical engineering,and have stronger nonlinearity.Once the system failure,minor failure of the system canaffectperformance and serious failure of the system canmake the whole system collapse,which can lead to catastrophicaccidents.Therefore,the study of the security and reliabilityfor the nonlinear system has become particularly eager and important.In view of this,consideringbetween the non-linear,the actuator failureand the external disturbances,and other factors of the actual system,this thesis proposes three methods for a class of nonlinear systems with actuator settings and time-varying faults,that is extended state observer(ESO),new extended state filter based on extended state observer(NESF-ESO)and new extended state filter(ESF).Then,the problem of the fault accommodation is research with the Inverse system theory.Finally,the corresponding simulation experiment study are given using the classical mathematical model Van Der Pol oscillator of mechanics,physics and electronics.specific content includes:1)Research of fault accommodation for continuous system based on time-varying gain ESO and inverse system method.For the purpose of effective diagnosis of a class of nonlinear system with the time-varying failure of the actuator,a method of the fault accommodation was proposed based on time-varying gain extended state observer(ESO)for the inverse system.Firstly,the traditional constant gain ESO was improved by using a time-varying gain design method.Secondly,the nonlinear system was linearized to the corresponding linear system based on the compensation and regulation of the fault estimation value for the original system and the introduction of the inverse system.And then,the robust guaranteed cost control was designed.Finally,a nonlinear system simulation example of Van Der Pol oscillator shows the effectiveness of the proposed approach when occur constant alues and time-varying failure2)Research of fault accommodation for discrete system based on NESF-ESO and inverse system method.Aiming at the problem of the linearization error and filtering divergence for traditional fault diagnosis methodofnonlinear system,First,a new extended state including the nonlinear dynamic and fault of the system is constructed,and the sufficient conditions of the augmented systemare presented.second,the nonlinear system was linearized to the corresponding linear system based on the compensation and regulation of the fault estimation value for the original system and the introduction of the inverse system.And then,the robust guaranteed cost control was designed.Finally,a simulation example shows theeffectiveness and superiority of the proposed approach.3)Research of fault accommodation for discrete system based on new ESF and inverse system method.In view of the traditional filter algorithm ESF effect problem commonly,studied a new type of ESF for discrete nonlinear system with signals noise fault filtering and accommodation method.First,a new type of ESF,recursive learning methods to improve the state of noise system and the accuracy of fault estimation;Secondly,after using new ESF filtering fault estimation as a failure to adjust item,combined with the discrete inverse system control algorithm,the original nonlinear system is linearized;Then consider the inverse system control method is not form a closed loop and inverse modeling errors,we design a robust guaranteed cost controller under the discrete time;Finally,still use classic nonlinear mathematical model of the Van Der Pol oscillator to show the effectiveness.