<> "The repository administrator has not yet configured an RDF license."^^ . <> . . . "Neural and hybrid neural modelling and control of fed-batch fermentation for streptokinase: Comparative evaluation under nonideal conditions."^^ . "Fermentations involving competition between two or\r\nmore kinds of cells under nonideal conditions show\r\ncomplex profi les that are sensitive to the extra-cellular\r\nenvironment. These fermentations therefore require\r\naccurate and rapid on-line data acquisition and control.\r\nHowever, both on-line measurements and modelling\r\nare diffi cult and expensive for large bioreactors, thus\r\nlimiting the usefulness of model-based control. While\r\nneural networks offer an alternative, they require\r\nextensive training and can be diffi cult to optimize\r\nfor large arrays. Hybrid networks combining\r\na few neural networks with some mathematical\r\nequations offer a good compromise. The possibility of\r\nusing a hybrid model for simulation-cum-control has\r\nbeen examined here for the fed-batch production of\r\nstreptokinase. Under noideal conditions, hybrid neural\r\nmodels outperformed both mathematical models\r\nand arrays of neural networks, thus suggesting their\r\nviability for large-scale fermentation monitoring and\r\ncontrol.\r\nLes fermentations provoquant une compétition entre\r\ndeux ou plusieurs sortes de cellules dans des conditions\r\nnon idéales montrent des profi ls complexes qui\r\nsont sensibles à l’environnement extra-cellulaires. Ces\r\nfermentations nécessitent donc une acquisition et un\r\ncontrôle en continu des données qui soient précis et\r\nrapides. Toutefois, les mesures et la modélisation en\r\ncontinu sont diffi ciles et coûteuses pour les grands\r\nbioréacteurs, ce qui limite l’utilité du contrôle basé sur\r\ndes modèles. Les réseaux neuronaux sont une autre\r\npossibilité, mais ceux-ci nécessitent un entraînement\r\npoussé et peuvent être diffi ciles à optimiser pour de\r\ngrands dispositifs. Les réseaux hybrides combinant\r\nréseaux neuronaux et équations mathématiques\r\noffrent un bon compromis. La possibilité d’utiliser\r\nun modèle hybride pour la simulation et le contrôle\r\na été examinée dans ce travail pour la production\r\nà alimentation discontinue de streptokinase. Dans\r\ndes conditions non idéales, les modèles neuronaux\r\nhybrides offrent une meilleure performance que les\r\nmodèles mathématiques ou les dispositifs de réseaux\r\nneuronaux, et il pourrait donc s’avérer viable pour la\r\nsurveillance et le contrôle de fermentation à grande\r\néchelle"^^ . "2004" . . "82" . "3" . . "wiley"^^ . . . "Canadian Journal of Chemical Engineering"^^ . . . . . . . . "P R"^^ . "Patnaik"^^ . "P R Patnaik"^^ . . . . . . "Neural and hybrid neural modelling and control of fed-batch fermentation for streptokinase: Comparative evaluation under nonideal conditions. (PDF)"^^ . . . . . "Neural and hybrid neural modelling and control of fed-batch fermentation for streptokinase: Comparative evaluation under nonideal conditions. (Indexer Terms)"^^ . . . . . . "Neural and hybrid neural modelling and control of fed-batch fermentation for streptokinase: Comparative evaluation under nonideal conditions. (Image (PNG))"^^ . . . . . . "Neural and hybrid neural modelling and control of fed-batch fermentation for streptokinase: Comparative evaluation under nonideal conditions. (Image (PNG))"^^ . . . . . . "Neural and hybrid neural modelling and control of fed-batch fermentation for streptokinase: Comparative evaluation under nonideal conditions. (Image (PNG))"^^ . . . . . . "Neural and hybrid neural modelling and control of fed-batch fermentation for streptokinase: Comparative evaluation under nonideal conditions. (Image (PNG))"^^ . . . . . "HTML Summary of #954 \n\nNeural and hybrid neural modelling and control of fed-batch fermentation for streptokinase: Comparative evaluation under nonideal conditions.\n\n" . "text/html" . . . "TP Chemical technology"@en . .