Neural Network Modelling of Fermentation Taking into Account Culture Memory
In the present paper the problem of chemostat modelling using the neural networks is considered. Two kinds of neural network models taking into account the culture memory are proposed. The first one is a feedforward neural network with time delayed feedback connections from and to the output neurons. The second one is a feedforward neural network with time delayed feedback connections from the input neurons describing only the specific growth rate. The obtained neural network model is adopted in the classical nonlinear model. Simulation investigations and comparison of the both models are carried out. The interpretation of the parameters according to their biological means is made.
KeywordsSpecific Growth Rate Neural Network Model Time Instant Output Neuron Feedforward Neural Network
Unable to display preview. Download preview PDF.
- 1.Wang, N.S., Stephanopoulos, G.: Biotechnology and Bioengineering Symp., 14, 635 (1984)Google Scholar
- 2.Thibault J., Grandjean, B.: Proc. of the IFAC Int. Symp. Advanced Control of Chemical Processes ADCHEM’91, Toulouse, France, 295 (1991)Google Scholar
- 3.Thibault, J., Cheruy, A.: IMACS Annals on Computing and Applied Mathematics Proc. MIM- S2’90, Brusels (1990)Google Scholar
- 4.Thibault, J., Van Breusegem, V., Cheruy, A.: B & B, 36, 1041 (1990)Google Scholar
- 5.Thibault, J., Van Breusegem, V.: Bioautomation, 12, 79 (1992)Google Scholar
- 7.O’Neil, D.G., Liberatos, G.: B & B, 36, 437 (1990)Google Scholar
- 8.Grueger, W., Grueger, A., Brock, T.D. (eds): Biotechnology: A Textbook of Industrial Microbiology. Suderland, MA, USA: Sinauer Associates, Inc 1984.Google Scholar
- 9.Petrova, M., Koprinkova, P., Patarinska T.: Biotechnology and Biotechnological Equipment 1 88, (1994)Google Scholar