Toward a Systematic Method for the Generalization of Fermentation Data
A common problem encountered in many research and development laboratories is the management and better utilization of large volumes of fermentation data collected in the course of basic fermentation research and product development programs. Besides rudimentary evaluation of fermentation performance, it would be highly desirable if the same data could be further utilized in order to extract and retain important features which are generic to the biological system and of essential value for fermentor control and optimization. In this context, the problem is defined as one of identifying a set of key culture variables and biological parameters which best represent the state of the organism as well as the fermentor. The approach discussed in this paper builds upon prior research on data reconciliation and fermentor identification and monitoring of intracellular metabolism, and extends the methods of generalization and pattern recognition to the problem of key feature identification and correlation in fermentation processes. Such a method, holds the promise of significantly enhancing the value of primary fermentation data as well as contributing to the development of supervisory control systems for fermentation processes.
KeywordsFermentation Performance Fermentation Feature Intracellular Flux Data Reconciliation Fermentation Data
Unable to display preview. Download preview PDF.
- 1.J.A. Romagnoli and Geo. Stephanopoulos, Chem. Eng. Sci., 36, 1849 (1981).Google Scholar
- 4.J.J. Vallino and G. Stephanopoulos, Annals N.Y. Acad. Sci., (in press) (1987).Google Scholar
- 5.J.J. Vallino and G. Stephanopoulos in Frontiers in Bioprocessing, S. Sikdar, P. Todd and M. Bier (eds.), (in press) (1988).Google Scholar
- 6.T.M. Mitchell, "Generalization as Search", p. 517 in Readings in Artificial Intelligence, B.L. Webber and N.J. Nilsson, (eds.), Morgan-Kaufmann (publ.), Los Altos, CA (1981).Google Scholar