Model-Based Estimation of Reaction Rates in Stirred Tank Bioreactors

Abstract

In this paper an adaptive model-based algorithm is proposed for the on-line estimation of reaction rates in stirred tank bioreactors. The main design condition imposes that the observation errors reflecting the mismatch between the estimated parameters and the 'true' values follow second-order dynamics of convergence. The gain matrices are shown to be functions of the state and of user-defined damping coefficients and natural periods of oscillation for second-order trajectories. The application of the algorithm is illustrated with a simple case-study involving the estimation of the specific reaction rate for a single substrate, single product scheme.

Adaptive linearizing control of bioreactors

“…The adaptive feature comes from the on-line estimation of the required process time varying parameters. The adaptive algorithm proposed enforces a desired and pre-set second order convergence dynamics as originally introduced by Oliveira et al (1). Formulating the estimator on this basis leaves the user with the choice of two simple and intuitive tuning parameters with physical meaning -a damping coefficient and a natural period of oscillation.…”