Fault Diagnosis And Isolation Based On The Adaptive Observer In Complex Networks

With the rapid development of science and technology,people and things are more and more closely related.Many systems in real life can be described as complex networks,such as power systems,transportation systems,biological systems and so on.In the event of a failure in these big systems,it would have a huge impact on our lives,even devastating effects.Therefore,the researches of fault diagnosis and fault isolation in complex networks have become an important subject in the field of complex network research in recent years.This paper is mainly based on the state observer to propose the methods of the fault diagnosis and fault isolation in complex networks.The thresholds of fault diagnosis and isolation are related to the uncertainties of the network and are adaptive.This paper has far-reaching significances on the research of the fault diagnosis and fault isolation in complex networks.The main contributions are listed as follows:Firstly,this paper studies a class of complex network with uncertainties,taking the design of the fault diagnosis and isolation observer into considerations.Initially,a fault diagnosis observer is designed and a threshold function is obtained for fault diagnosis.In order to counteract the effect of uncertainties on the threshold,a method of constructing threshold function dependent on uncertainties is proposed.Then,the fault isolation observers and the adaptive threshold function corresponding to it are designed in the purpose of the fault isolation.Finally,the accuracy of the method is verified by simulation.Secondly,this paper also studies the fault diagnosis and fault isolation in a class of complex network system with uncertainties,taking the connection fault and sensor fault into considerations.First of all,the fault diagnosis is studied in the whole complex network system,and the process of the fault diagnosis is mainly consisted of the design of the fault diagnosis observer and the adaptive threshold function corresponding to the fault diagnosis observer.Then,the process of the fault isolation is mainly composed of the isolation of the connection fault and the sensor fault in the complex network.Finally,the accuracy of the method is verified by simulation.Thirdly,this paper also studies the multiple failures in a class of complex network system with uncertainties,taking various faults into considerations,such as the controller failure,connection fault and so on.Based on the knowledge of the design of the observer,this paper firstly does some researches about the fault diagnosis in the complex network,and then the fault of the complex network can be detected by the comparison between the state error and the adaptive function.When a fault occurs in the complex network,the fault isolation observers are activated to isolate the type of fault.Finally,the accuracy of the method is verified by simulation.