Research On State Estimation And Fault Diagnosis Of Nonlinear System Based On Extended Set-Membership Filter

With the increase of scale and complexity of industrial control systems,the demands for system state estimation and fault diagnosis are also growing.In order to achieve accurate estimation of the state of nonlinear engineering systems and improve the accuracy and stability of system fault diagnosis,relevant theoretical research has been carried out in this paper,which have theoretical and practical value for promoting state estimation and fault diagnosis research in the engineering field.The main work of this paper is as follows:The paper firstly introduces the relevant theoretical basis of nonlinear system state estimation and fault diagnosis technology.Then,based on the set-membership estimation theory,Improved Extended Set-Membership Filter(IESMF)algorithm is proposed to solve the problem of state estimation for nonlinear systems with unknown but bounded noise and disturbance in practical engineering.After linearizing the nonlinear system model,the first and second order error terms in Taylor expansion are analyzed to reduce the linearization error and improve the accuracy of state estimation.At the same time,it introduces quadrature rectangle decomposition and optimizes algorithm parameters to improve the stability of algorithm state estimation and the efficiency of filter calculation.Finally,simulation verification is carried out in a nonlinear power signal system.The IESMF algorithm shows high accuracy and robustness,which demonstrates the effectiveness and feasibility of the algorithm.Then,aiming at the problem of fault diagnosis of nonlinear system in actual engineering,the IESMF algorithm is applied to the fault diagnosis method based on filter to realize the fault detection and fault estimation of the system.The residual error is generated by the set-membership filter and the residual error evaluation function is calculated.Set an appropriate threshold,and compare whether the residual error evaluation function is greater than the threshold to detect the fault in the system.At the same time,an augmented observer is designed to estimate the fault amplitude and reconstruct the fault signal.In the simulation comparison experiments for pulse and periodic faults,the method can detect the fault occurrence and estimate the fault amplitude accurately,which proves the feasibility of the algorithm in fault diagnosis applications.Finally,in order to overcome the shortcomings of IESMF algorithm in fault diagnosis and further improve the accuracy and stability of fault diagnosis.Based on the original algorithm,Interactive Multiple Model algorithm,adaptive algorithm and Particle Swarm Optimization algorithm are introduced respectively.The three optimized algorithms can adjust the filtering effect adaptively to improve the estimation accuracy and stability of the algorithm.In the simulation verification of fault diagnosis,all three algorithms can detect system faults in time and accurately and realize accurate and stable estimation of fault amplitude under the premise of ensuring low missed detection and low false detection,which proves the effectiveness of algorithm optimization.