Adaptive Algorithm for Estimation and Control of Fedbatch Bioprocesses
To control a fedbatch bioprocess, a promising approach is the use of non-linear adaptive controller in order to cope with the non-linear time-varying nature of the dynamics [Bastin, Dochain, 1984]. This approach provides better guaranties of stability than the more classical one, such a PI control. Yet, these simple control algorithms designed with the physical behaviour of the process in mind leads to interesting results [Cooney, Duen-gang 1983]. The approach we propose combines both methods. It has a very straightforward physical interpretation while taking care of the time-varying and unknown parameters with an stable scheme of estimation for them. Another problem, that we come up against, is the lack of sensors for measuring the various concentrations in the reactor, specifically S, the substrate concentration we want to control. To address this problem we propose an indirect control of S via µ, the specific growth rate, which can be estimated on-line more easily than S. These algorithms are tested by simulation.
KeywordsSubstrate Concentration Specific Growth Rate Initial Error Physiological Model Ethanol Formation
Unable to display preview. Download preview PDF.
- Rajab A., (1986) Modélisation et conduite automatique de la fermentation de Saccharomyces Cerevisiae, Thèse d’état, INPL, France.Google Scholar