Multiple-model control strategy for a fed-batch high cell-density culture processing

The fermentation of microorganisms is an important process in the biotechnology industry. To enhance the volumetric productivity, high cell concentration and cell productivity are required. Exponential feeding profiles are routinely applied to control growth rate and limit the production of inhibiting metabolite by-products. However, during aerobic high cell-density culture (HCDC) processes, both the rate of oxygen mass transfer and dissolved oxygen concentration decrease with increasing biomass. Due to the interactive influence of both oxygen and carbon on the growth rate of microorganisms, therefore, a good operating methodology in HCDC processes must maintain the relational concentrations of both substrates. The primary object in this work is to construct a simple and robust control strategy for this specific purpose. The overall control structure includes an optimal feedforword controller and a multiloop feedback controller. In each feedback loop, multiple-models are used such that the proposed control structure is simple to build. The feedback controllers, tuned by internal model control (IMC) principle, are scheduled by multiple-model in each loop. The control structure is shown by simulation to be robust with respect to mismatches in the model parameters and also in the model's form. A worst case analysis on 20% mismatch in model's five parameters showed a productivity loss of only 0.093%. Even though there exists oxygen mass transfer limitation, this control structure can still maintain good performance. Furthermore, an override control strategy can be added in the overall control structure to handle the oxygen-limiting problem. (c) 2005 Elsevier Ltd. All rights reserved.