Font Size: a A A

Research Of Fault Diagnosis Method Based On DEDS

Posted on:2006-11-01Degree:MasterType:Thesis
Country:ChinaCandidate:H B XueFull Text:PDF
GTID:2178360182476673Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Fault diagnosis technology is developing with the increasing demand forsupervisory control system. Many different methods and techniques have beenproposed for diagnosis, we broadly provide a systemic and comparative study ofvarious diagnostic methods and briefly introduce some representative methods. TheDEDS approach to fault diagnosis is one of the newly developed methods. TheDiagnoser Theory is based on Finite State Machines (FSMs) or Automata for DEDSs.The system to be diagnosed is described as a formal language that modeled byautomata. A diagnoser built from the system model performs diagnosis. As for thelogical diagnoser, the system is said to be logical diagnosable if one can detect theoccurrence of a failure with a finite delay.The stochastic automata model inherits the structure of the logical automatamodel, in which a probabilistic structure is appended to the model to estimate thelikelihood of specific events occurring. And A-diagnosability and AA-diagnosabilityare introduced on the basis of the logical diagnosability. The stochastic diagnoserconstructed form the stochastic automata satisfies the Markov property. Not only thestochastic diagnoser can be used to analyze the A-and AA-diagnosability offline andto fault diagnose online, but also it can calculate the probability of each component ofthe state estimate. The processes of diagnosablity analysis offline and fault diagnosisonline are illustrated by examples from the door interlock device of the elevatorsystem.The failure mechanism of the elevator door system is researched thoroughly andthe stochastic automata approach is applied to the fault diagnosis of this system. Themodels of the elevator door interlock device and the elevator door system are builtand the stochastic diagnosers are constructed. We also improve the diagnosabilty ofthe elevator door system by adding sensor information and modifying the controlprotocol of the system.In order to determine the feasibility of the proposed fault diagnosis methodology,a test bed is required. The objective systems of the test bed are elevator systems andthe test bed consists of 11 virtual elevator models, 4 physical elevator models and 1supervisory control PC. The design of the two models is introduced and the testmethods are discussed. The other functions of the test bed are also introduced.
Keywords/Search Tags:Fault Diagnosis, DEDS, Stochastic Automata, Diagnoser, Elevator
PDF Full Text Request
Related items