On-line state recognition in a yeast fed-batch culture using error vectors

• Published in: Biotechnology and bioengineering Vol. 47; no. 2; p. 165
• Main Authors: Shimizu, H, Miura, K, Shioya, S, Suga, K
• Format: Journal Article
• Language: English
• Published: United States 20.07.1995
• ISSN: 0006-3592
• DOI: 10.1002/bit.260470207

The physiological states with respect to cell growth and ethanol production in a yeast fed-batch culture expressed in linguistic form could be recognized on-line by fuzzy inferencing based on error vectors. The error vector was newly defined here in a macroscopic elemental balance equation. The physiological states for cell growth and ethanol production were characterized by error vectors using many experimental data from fed-batch cultures. Fuzzy membership functions were constructed from the frequency distributions of the error vectors and state recognition was performed by fuzzy inferencing. In particular, an unusual physiological state for a yeast cultivation, in which aerobic ethanol production was accompanied by very low cell growth, could be recognized accurately. According to the results of the state recognition, an energy parameter, the P/O ratio in the metabolic reaction model was adaptively estimated, and the cell growth was successfully evaluated with the estimated P/O. (c) 1995 John Wiley & Sons, Inc.