PARAMETER-ESTIMATION TOWARD FAULT-DIAGNOSIS IN NONLINEAR-SYSTEMS USING A MARKOV MODEL OF SYSTEM DYNAMICS

A model-based parameter estimation method for nonlinear systems that d oes not require the linearization of the system equations and that can account for uncertainties in the monitored data as well as the parame ters (e.g., random variations) is described. The method is particularl y suitable for fault diagnosis because of its capability to assign pro babilities of occurrence to user-specified parameter magnitude interva ls that may be associated with system faults. The method regards syste m evolution in time as transitions between these intervals as well as user-specified magnitude intervals of the dynamic variables. These tra nsition rates are obtained on-line from the system model and the monit ored dynamic variable data and constitute a Markov chain in discrete t ime. The method then compares predicted and observed data at a given t ime step to narrow the estimated parameter range in the next time step . Implementations using a second-order van der Pol oscillator and a th ird-order system describing temporal xenon oscillations in a hypotheti cal reactor indicate that the method is computationally efficient and can be used for multiparameter estimation with incomplete information on the system state.