Research On Transient Faults Detection For Networked Control System Based On Models

With modern industry constantly moving towards the direction of digital, networked,intelligent and comprehensive integration, networked control system (NCS) has beenwidely used especially in the aerospace, automotive, chemical plants and other complexengineering fields. Meanwhile, as the growth of system scale and complexity, therequirements of reliability and security of network control system have been very urgent.Transient faults are the dominated kind of threats to operation safety in networked controlsystems. And they have the natures of random occurrence and unpredictable, shortduration and easy to propagate in the system, it is difficult to detect and locate them in atraditional fault detection way. This paper presents a new method based on multiplemodels for transient fault detection in networked control system.In this paper, firstly features of transient faults in networked control system aresummarized, and then on the basis of investigating the models of networked controlsystem, a theory foundation of implementing transient fault detection is given, in whichmultiple models are used to represent deep knowledge of system. Secondly, the specificimplementation of these multiple models based on GTST and DFM are shown, and then atransient fault detection framework is proposed. Finally, a case study based on a flow ratenetworked control system is shown to verify the proposed multiple models based faultdetection method.