Research Of Fault Diagnosis Based On Unknown Input Observer

Nowadays,we can see the fast development of production technology and the scientific theory.With the fast improvement,the scale of modern control systems is becoming larger as well as more complicated.Therefore,the systems have a increasing possibility to arise fault.We must design and build a monitoring system in order to improve the reliability and security of the system.The monitoring system should be able to analyze the reason and characteristics of the fault as well as take some effective methods to solute the fault.At last,the monitoring system have a capability to prevent the fault from happening.This research is based on observers for fault diagnosis of the unknown input observer.The main research work of my study is as follow:At first,the dissertation introduces the theory of fault diagnosis object that are based on mathematical models.The introduction includes the common system model,the fault diagnosis model and the testability of using this method to deal with the fault.The dissertation discusses the measures that based on the unknown input fault diagnosis and have a simulation for the linear system.The diagnosis measure which is based on the mathematical model depends on the accuracy of the model.Therefore,when we use this measure to diagnose fault of a uncertain system,it is necessary to ensure the uncertain factors have the lowest affect to the result of the diagnosis and the efficiency of the diagnosis.This is the most important factor of the model-based fault diagnosis.For the system which has model uncertainties,the dissertation discuss the theory based on H_? filter and design the fault detection filter based on it.This measure transform the problem of how to design the fault detection filter into how to get the matrix inequality constraints by H_? norm theory.It makes the nonlinear problem to a linear problem.And an iterative algorithm is used to transform the filter design problem into the optimization problem with linear matrix inequalities constraints.With the problem of setting up a threshold value,this dissertation uses a auto-adaptive threshold method and then it studies the robust fault detection for the uncertain system.A method of robust fault estimation based on adaptive theory and sliding mode theory is proposed for the uncertain system which has unknown input signal.With the fault reconstruction,the slide mode observer is used.In order to compensate the effect of the uncertain input to the error system,the adaptive law is used.The sliding mode observer theory is also used for the estimation of the system fault.Then considering about the H_? filter,proper H_? filter index is designed to diagnose the fault.At last,a conclusion of the research work of this dissertation is given and a discussion of the further research in this field is raised.