Predictive modeling and loose-loop control of perfusion bioreactors

A 10 L bioreactor was inoculated with 2.5/spl times/10/sup 5/ cells/mi and grown as a batch until the modeling predicted that the glucose concentration would be less than the set point concentration of 10 mM. The computer-controlled feeding to the reactor was started approximately 6 h after the sampling took place. The glucose uptake rate (GUR) modeling and prediction horizon estimators were obtained from equations which are presented. The predictive nonlinear modeling and daily samples were able to maintain the process at the glucose set point with a standard deviation of 0.35 mM. This compares favorably with a semi-online process sampling and analysis system with a linear controller and hourly sampling with a standard deviation of approximately 0.66 to 0.72 mM (K.B. Konstantinov, 1996). The controller response to deviations from the set point is nonlinear. If the concentrations are close to the set point value, then little adjustment is made. However, with larger deviations, the more active the controller becomes in terms of adjusting the flow rate. The modeling of perfusion bioprocesses was mediated in a multiple model adaptive framework, with a predictive controller used to maintain process control. Nonlinear modeling in a predictive control framework has been effective to control substrate concentrations with daily sampling. Data filtering protocols were included as process noise was shown to have significant effects.