Automatic Generation of Process Models for Fed-Batch Fermentations Based on the Detection of Biological Phenomena Dissertationsschrift
-
- Englisch ausgewählt
19,90 €
inkl. gesetzl. MwSt.,
Lieferung nach Hause
Beschreibung
Produktdetails
Einband
Taschenbuch
Altersempfehlung
18 - 99 Jahr(e)
Erscheinungsdatum
02.02.2015
Verlag
EpubliSeitenzahl
184
Maße (L/B/H)
21/14,8/1 cm
Gewicht
244 g
Auflage
1
Sprache
Englisch
ISBN
978-3-7375-2957-0
Using dynamic models to describe biotechnological processes leads to a better understanding of the complex process dynamics and helps to find optimal conditions that improve the process significantly. However, the development of adequate mathematical models is generally difficult, tedious, time-consuming, and requires extensive prior experimentation.
This work presents an algorithm that automatically proposes process models from an automated processing and analysis of data from (fed-)batch experiments. For this purpose, the algorithm first uses different data smoothing and interpolation techniques to account for a typically noisy and poorly time-resolved data set. Then, the measurements are numerically compensated for the influence of feeding and sampling. To reveal the qualitative behavior of the measurements, a method is used that divides the compensated curves into several episodes in a probabilistic framework.
Based on this qualitative description, crucial information about the underlying reaction network can be obtained. For this, different biological phenomena describing the relation between several measured variables are defined. Rules to (dis-)prove these phenomena are applied to the data. The detected biological phenomena and the used measured variables then lead to an automated proposal of several model structures with different degrees of complexity.
Furthermore, a procedure to detect model deficiencies is drafted. The phenomena detection is applied to simulations of a model and compared to the measurement-inherent phenomena. This approach is initially tested on simple case studies.
The presented algorithm is applied to fed-batch experiments of three different organisms. Structured models are generated, identified, and validated. The results show that the models still need to be improved, but, in many cases, are able to describe the dynamics satisfactorily. Thus, the presented approach helps to speed up the modeling process significantly.
This work presents an algorithm that automatically proposes process models from an automated processing and analysis of data from (fed-)batch experiments. For this purpose, the algorithm first uses different data smoothing and interpolation techniques to account for a typically noisy and poorly time-resolved data set. Then, the measurements are numerically compensated for the influence of feeding and sampling. To reveal the qualitative behavior of the measurements, a method is used that divides the compensated curves into several episodes in a probabilistic framework.
Based on this qualitative description, crucial information about the underlying reaction network can be obtained. For this, different biological phenomena describing the relation between several measured variables are defined. Rules to (dis-)prove these phenomena are applied to the data. The detected biological phenomena and the used measured variables then lead to an automated proposal of several model structures with different degrees of complexity.
Furthermore, a procedure to detect model deficiencies is drafted. The phenomena detection is applied to simulations of a model and compared to the measurement-inherent phenomena. This approach is initially tested on simple case studies.
The presented algorithm is applied to fed-batch experiments of three different organisms. Structured models are generated, identified, and validated. The results show that the models still need to be improved, but, in many cases, are able to describe the dynamics satisfactorily. Thus, the presented approach helps to speed up the modeling process significantly.
Kundinnen und Kunden meinen
Verfassen Sie die erste Bewertung zu diesem Artikel
Helfen Sie anderen Kund*innen durch Ihre Meinung
Kurze Frage zu unserer Seite
Vielen Dank für dein Feedback
Wir nutzen dein Feedback, um unsere Produktseiten zu verbessern. Bitte habe Verständnis, dass wir dir keine Rückmeldung geben können. Falls du Kontakt mit uns aufnehmen möchtest, kannst du dich aber gerne an unseren Kund*innenservice wenden.
zum Kundenservice