Model Based Fault Diagnosis And It’s Application In Satellite Gesture System

Nowadays,with the rapid development of science and technology,the demand for safety and reliability of control systems is getting higher and higher.From a positive perspective,he has brought us enormous economic benefits and comfortable living environment.On the contrary,any failure in automation will bring disastrous accidents which can not be underestimated.Therefore,the importance of fault diagnosis becomes increasingly apparent.Fault diagnosis is an indispensable step to check the state of the system and improve the stability of the system.In recent years,under the background of increasing demand for the stability of the system,it has been pushed to a certain research height by experts.The first step of fault diagnosis is the fault detection technology,whose role is to make a certain judgment of whether the system fails,to avoid the fault caused greater harm to the system.The second step is the fault estimation technology,which is based on the first step.It is used to estimate the fault boundary value online and plays an important role in fault diagnosis.However,the fault estimation is more difficult,and it is more challenging than fault detection.In recent years,the research results of fault diagnosis technology emerge in endlessly,and the diagnosis technology is more abundant,but there are still some shortcomings,which limit the scope of application of various technologies.It can be seen that the problem of fault diagnosis is still a subject to be further studied.In this paper,the observer based fault diagnosis for aerospace systems is studied.The main work is as follows:Firstly,although the traditional adaptive design method is simple,it is difficult to satisfy the dynamic performance of complex systems because of equality constraints.Therefore,the traditional adaptive fault diagnosis method is improved.The improved algorithm significantly improves the speed and accuracy of fault estimation.The proof process of traditional adaptive fault diagnosis is improved.Equality constraints are eliminated,diagnosis rules are relaxed and application scope is expanded by using inequality transformation.Secondly,a multi-observer fault isolation method is proposed,and a two-level fault diagnosis scheme is designed.After fault detection,a specific fault isolation scheme is implemented.This scheme can locate the fault location after fault diagnosis.Thirdly,according to the characteristics of disturbance uncertainty of actual system,the observer and neural network method are combined to realize the detection of small faults and enhance the ability of system fault detection.Finally,it is applied to Aeronautical System for simulation,and the effectiveeness of the method is verified.