Application of Digital Filtering Methods to State and Parameter Estimation in Process Plant

Abstract

The application of on-line state variable and parameter estimation for chemical processes, with particular reference to a pilot plant scale double effect evaporator, has been investigated. The investigation has shown the requirement of some adaptive modification to the recursive Kalman Filter to be necessary. This adaptive modification prevents the occurrence of bias and divergence of the estimates of the state of the system, when investigating poorly understood chemical processes. A mathematical model of the double effect evaporator is derived along with heat transfer correlations that are suitable for implementation within the Kalman Filter. Finally a computer package is developed, which incorporates an adaptive filter, for on-line estimation of the state and parameters of a chemical process. The computer package is shown to operate successfully in real time on the double effect evaporator.