Conference Proceeding

Streptomycin fermentation process modeling with principal componentanalysis and fuzzy model

Nat. Key Lab. of Ind. Control Technol., Zhejiang Univ., Hangzhou;
Pump Industry Analyst 02/2000; DOI:10.1109/WCICA.2000.862969 ISBN: 0-7803-5995-X In proceeding of: Intelligent Control and Automation, 2000. Proceedings of the 3rd World Congress on, Volume: 3
Source: IEEE Xplore

ABSTRACT Analysis, modeling and control for a fed-batch fermentation process still remain challenging issues. Based on principal component analysis (PCA) and fuzzy modelling a simple and efficient approach to monitoring fed-batch streptomycin fermentation is presented. The data obtained from an industrial streptomycin fermentation process is first analyzed with PCA so that the large multivariate data with highly correlated and noisy measurements can be compressed into a lower dimension space which contains most of the variance of the original matrix. Moreover, a fuzzy model is used to construct a product (antibiotic) concentration estimator of the streptomycin fermentation process, prior knowledge and expertise are important in fed-batch fermentation processes. The results of the fuzzy model compared with a linear multivariate regression model indicate the potential of the fuzzy model as a state estimator of all such industrial fed-batch processes

0 0
  • [show abstract] [hide abstract]
    ABSTRACT: The difficulty in using artificial neural networks (ANNs) to represent nonstationary processes such as biotechnical production processes has been coped with by process partitioning. By logically cutting down the whole process into segments corresponding to different process phases, or special process situations, and representing them by a corresponding modular set of neural networks it was possible to obtain well performing real-time state estimations and process predictions. At any given moment, more than one artificial neural net is employed to describe the process. A small fuzzy expert system, by which additional heuristic know-how about the process is exploited, is used to dispatch control to the individual ANNs appropriately, and to process the outputs of the individual nets. Knowledge conserved in the nets is continuously updated by means of a real-time learning procedure as they are activated. As compared with the process estimation and prediction by our previously developed distributed model which is based on the extended Kalman filter approach, the fuzzy-aided artificial neural network system turned out to be at least as accurate, and considerably faster to develop.
    Journal of Biotechnology. 01/1993;
  • [show abstract] [hide abstract]
    ABSTRACT: Despite extensive research efforts in biosensor technology, effective supervision and control of bioprocess systems still remains a formidable task. However, the inference of important biological process variables from existing on-line measurements provides one means of easing the burden. Inferencing techniques are primarily based upon sound process models, which for non-linear bioprocess systems can be problematic to derive. The ability of artificial neural networks to learn essential process non-linearities from plant data may provide a means by which to assist in estimator development. This paper investigates this prospect by considering the construction of artificial network-based biomass and penicillin estimators for on-line application to an industrial fermentation. Results from on- line industrial site tests are presented.
    Computers & Chemical Engineering. 01/1992;