Diagnosis of a Continous Stirred Tank Reactor Using Kalman Filter

This paper is dedicated to complex system diagnosis. Therefore, a Kalman filter has been proposed to cope with external disturbances and unpredictable faults that are associated with chemical processes. The Continuous Stirred Tank Reactor (CSTR) model with parametric uncertainties is linearized around a chosen steady state and discretized to apply the diagnosis approach. A Kalman filter is employed to generate conversion estimate for the CSTR using only temperature measurements and precedent estimates. Numerical simulation Results of a no isothermal CSTR with coolant jacket dynamics are presented. Robustness with respect to interpolation errors, disturbances and model parametric uncertainties is achieved by using the obtained Kalman filter.