Khanzadi, Kobra (2014) Modeling and optimization of enzymatic synthesis of caffeic acid phenethyl ester using artificial neural network and genetic algorithm. Masters thesis, University of Zabol.
|
Text
Modeling and optimization of enzymatic synthesis of caffeic acid phenethyl ester.pdf Download (360kB) | Preview |
Abstract
In this study, the reaction of caffeic acid and 2-phenyl ethanol in the presence of immobilized lipase from Candida Antarctica ( Novozym 435) was modeled and optimized in isooctane system using artificial neural network and genetic algorithm methods in order to obtain caffeic acid phenethyl ester. For this purpose, a 5- level -4 variable central composite rotatable design was used for modeling the enzymatic reaction using artificial neural network. Independent variables were temperature, time, molar ratio of substrates and enzyme amount; while the molar conversion of caffeic acid to ester was considered as a dependent variable. The Levenberg-Marquardt algorithm was used as learning algorithm of artificial neural network. Therefore, first, the modeling was carried out by artificial neural network and using Levenberg-Marquardt algorithm. The best model includes a network of four inputs, 10 neurons in hidden layer and one output (4-10-1). After modeling by artificial neural network, genetic algorithm was used for optimization of the model. The optimized conditions were: time: 60 hrs, temperature: 62˚C, molar ratio of substrates: 0321 (2-phenyl ethanol: caffeic acid) and enzyme amount: 322 PLU. Under these conditions, the actual and predicted of molar conversion of caffeic acid to ester were 21012 and 100054, respectively.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Modeling; optimization; enzymatic synthesis; caffeic acid phenethyl ester; artificial neural network; genetic algorithm |
| Subjects: | Q Science > QD Chemistry |
| Depositing User: | admin admin1 admin2 |
| Date Deposited: | 24 Oct 2016 08:52 |
| Last Modified: | 24 Oct 2016 08:52 |
| URI: | http://eprints.uoz.ac.ir/id/eprint/884 |
Actions (login required)
 |
View Item |
