Research On Fast Diagnosis And Fault-tolerant Control Of Small Fault Based On Sliding Mode Theory

With the rapid development of modern industrial control system,the integration scale of the system continues to expand,and the complexity of the system tends to increase geometrically.Partial failure of the system often pose a great impact on the whole system.Therefore,fault diagnosis,fault tolerant control and other redundant functions are designed to improve the stability and safety of the control system,so that the system can avoid failure as much as possible or minimize the loss when fault occurs.It is of great significance to study the early stage of fault,to detect and isolate fault timely,and to control fault tolerance when necessary.This paper mainly focused on the study of rapid diagnosis and identification of faults in a small period and fault-tolerant control optimization.The main work is as follows:(1)Concerning the fast fault estimation problem with both noise disturbance and small amplitude faults,this thesis puts forward a new fast estimation method of comprehensive observer.Firstly,by constructing suitable transformation matrix,the original system is decoupled into two subsystems.For subsystems with noise disturbances and small faults,a synovial observer is designed to suppress the noise;By combining with the subsystem which contains small faults,this thesis puts forward a kind fast estimation algorithm of small faults,which obviously improves the tiny fault estimation performance.By using Lyapunov stability theory,to prove the stability of the composite observer proposed in this thesis is proved.Finally,a simulation test based on quadrotor UAV model is carried out to verify the effectiveness of the method.(2)For the problem of steady and fast estimation of small fault system in case of noise interference of large amplitude,this thesis puts forward a combined with high-order sliding mode robust estimation method of filtering.On the basis of the above increase the influence of noise disturbance,for subsystem contains only small unknown fault,combining the Runge-Kutta method to design a new adaptive observer,better improve the estimation precision;Then a robust estimation method is designed based on the high order sliding mode filter.Finally,by comparing the results of several model simulations and several algorithms,it is proved that the new algorithm can improve the performance of fault estimation under the condition of strong noise.(3)On the basis of the rapid immunity fault diagnosis research,the method of gap measurement is introduced to define a new method of fault hierarchies,with which the degree of fault severity can be classed intuitively and conveniently,the purpose of targeted fault-tolerant control of the fault system can be reached by combining with the fault-tolerant control method,and the system robustness can be improved more effectively.Then,two kinds of adaptive optimal fault-tolerant controllers are designed,and the effect of faulttolerant control is compared and analyzed.