简介:Thispaperpresentsanovelapproachtodetectanddiagnosefaultsinthedynanmicpartofachisofstochasticsys-tems.theSuchagroupofsystemsaresubjectedtoasetofcrispinputsbuttheoutputsconsideredarethemeasurableprobabilitydensityfunctions(PDFs)ofthesystemoutput,ratherthanthiesystemoutputalone.Anewapproximationmodelisdevelopedfortheoutputprobabilitydensityfunctionssothatthedynamicpartofthesystemisdecoupledfrontheoutputprobabilitydensityfunctions.Anonlinearadaptiveobserverisconstructedtodetectanddiagnosethefaultinthedynamicpartofthesystem.Convergencyanalysisisperfomedfortheerrordynamicsraisedfromthefaultdetectionanddiagnosisphaseandanapplicabilitystudyonthedetectionanddiagnosisoftheunexpectedchangesinthe2Dgrmmagedistributionsinapaperformingprocessisincluded.