This paper addresses the problem of long sampling times arisen in state estimation of stochastic chemical systems. The discussed state estimation task implies that available measurements of some parameters (depending on the utilized technology) are used for calculation of remaining (not measurable) variables by means of Kalman filtering method. The most traditional way of state estimation is grounded in the well-known extended Kalman filter (EKF). However, it is found out that the traditional EKF may fail when the sampling (waiting) time is long enough. Thus, we show on an example of stochastic Batch reactor that modern state estimation methods, such as the continuous-discrete unscented and cubature Kalman filters and the accurate continuous-discrete extended Kalman filter, resolve successfully the mentioned inconsistency and work well for state estimation in chemical systems with infrequent measurements. These nonlinear filters evidently outperform the traditional EKF method and, hence, suit better for state estimation in chemistry research and industrial implementation.

CEMAT - Center for Computational and Stochastic Mathematics