Moving-horizon state estimation applied to an industrial polymerization process

In this paper, a simplified version of a model available from the literature is used to illustrate some of the issues encountered when applying a moving-horizon state estimator (MHSE) to industrial polymerization processes. Results of an online implementation will be presented. This online implementation of moving-horizon state estimation is used to investigate the effects of weight selection, initial conditions and disturbance model structure on estimator performance. Orthogonal collocation is used to discretize the equations allowing simultaneous solution of the estimation problem using commercially available numerically efficient optimization software. These results demonstrate that this class of nonlinear horizon-based estimators can be implemented on industrially relevant problems with modest computing hardware.