Actuator Fault Diagnosis Of A Class Of Cascade Systems With Adaptive Observer Method

Nowadays,the scale and complexity of control systems are increasing dramatically.Consequentially,the problem of control system security is becoming increasingly prominent.The safety and reliability of control systems can be ensured by effective fault diagnosis methods.Therefore,it is of great significance to deeply study the fault diagnosis technology.Because the real systems are all nonlinear systems,and the system with high complexity has inevitably existed modeling errors,and there are various kinds of interference in actual operation,the development of fault diagnosis technology still has many deficiencies.The cascade system is a kind of special nonlinear system.It is composed of two cascaded subsystems.The control amount only acts on the drive subsystem.Therefore,it is of great significance to study actuator faults of such systems.Considering that the current research results mostly focus on the additive faults of actuators,the actuator gain fault which is universal has been neglected for a long time.In this paper,the fault diagnosis method based on analytical model is used.By designing the state observer,the actuator gain fault of the cascade system is systematically studied.A set of systematic and comprehensive fault diagnosis theoretical framework in cascaded system,including the ideal state,time-delay,model errors and external disturbance,is established.The main content of the paper is as follows:In this dissertation,firstly,aiming at a class of cascade system actuator gain fault model under ideal condition,a fault detection observer is designed to detect faults in time.Afterwards,an adaptive observer-based fast actuator gain fault estimation algorithm is designed to make online fault estimation.For the conclusions we have obtained,we have given strict proofs and conducted numerical simulations.The simulation results verify the validity of the method.Then for the actuator gain fault model of a class of cascaded systems with time-delay,the fault detection observer is designed,and the existence conditions of the detection observer are given to detect the fault occurrence in time.Then,the faults were estimated on-line through the designed adaptive diagnostic observer and diagnostic algorithm.We also provided strict mathematical proofs and numerical simulations for the conclusions.The simulation results verify the effectiveness of the method.Finally,on the basis of the contents of the first two chapters,the above conclusions are generalized to a more practical cascaded system actuator gain failure model,taking into account external disturbances and model errors.The robust adaptive diagnostic observer is designed for the case with on delay and time delay,and the fault estimation is carried out on line.The system state error and the fault estimation error index are attenuation to residual set.The conclusions have also been rigorously proved by mathematics.The simulation results also prove the validity of the method.