Monitoring Biological Modes in a Bioreactor Process by Computer Simulation.
Abstract
This paper deals with the general framework of fermentation system modeling and monitoring, focusing on the fermentation of Escherichia coli. Our main objective is to develop an algorithm for the online detection of acetate production during the culture of recombinant proteins. The analysis the fermentation process shows that it behaves like a hybrid dynamic system with commutation (since it can be represented by 5 nonlinear models). We present a strategy of fault detection based on residual generation for detecting the different actual biological modes. The residual generation is based on nonlinear analytical redundancy relations. The simulation results show that the several modes that are occulted during the bacteria cultivation can be detected by residuals using a nonlinear dynamic model and a reduced instrumentation.