论文信息 - AN APPROACH TO THE ISOLATION OF SENSOR AND ACTUATOR FAULTS BASED ON SUBSPACE IDENTIFICATION �

AN APPROACH TO THE ISOLATION OF SENSOR AND ACTUATOR FAULTS BASED ON SUBSPACE IDENTIFICATION �

This paperaddressestheproblemof sensorandactuatorfault isolationin MIMO discrete-timelinear time-invariant systems.Theconsideredclassof faultsarerepresentedby scalingfactorsfor both thecontrolsignalsandthemeasured (output)signals,andconstantoffset terms.Thestartingpoint is theidentificationof a state-spacemodel,thatrepresents thedeterministicpartof the faulty systemup to a similarity transformation,from corruptedby noiseinput outputmeasurements. Basedontheobtainedmodelwederiveaprocedurefor theisolationof thefaultsin thesystem.It first solvesa suitablydefinednon-linearoptimizationproblemover thescalingfactors.Necessaryconditionfor theexistenceof unique solutionto theoptimizationproblemis given,andthentheuniqueglobaloptimizeris derived in analyticalform. Given the optimal scalingfactorswe next find the similarity transformationthat transformsthe identifiedmodelto the initial statebasis,which is a necessarystepfor thedeterminationof theoffset factors.Theapproachis verified in a simulation studywith a linearizedmodelof the lateraldirectionaldynamicsof anF-14aircraft, in which simultaneousfaultsin all sensorsandactuatorsweresimulated. keywords: fault isolation,subspaceidentification,non-linearleastsquares

S. K. Kanev

[1] Michèle Basseville. On fault detectability and isolability , 1999 .

[2] Michel Verhaegen,et al. Identification of the deterministic part of MIMO state space models given in innovations form from input-output data , 1994, Autom..

[3] Peter S. Maybeck. Multiple model adaptive algorithms for detecting and compensating sensor and actuator/surface failures in aircraft flight control systems , 1999 .