Modelling of the p,p'-dinitrodibenzyl electroreduction by using an artificial neural network

Hdl Handle:
http://hdl.handle.net/10149/98933
Title:
Modelling of the p,p'-dinitrodibenzyl electroreduction by using an artificial neural network
Authors:
Olea, M. (Maria)
Affiliation:
Babes-Bolyai University of Cluj-Napoca. Department of Chemical Engineering. Romania.
Citation:
Olea, M. (2007) 'Modelling of the p,p'-dinitrodibenzyl electroreduction by using an artificial neural network', Match: communications in mathematical and computer chemistry, 57 (3), pp.735-748.
Publisher:
University of Kragujevac, Faculty of Science
Journal:
Match: communications in mathematical and computer chemistry
Issue Date:
2007
URI:
http://hdl.handle.net/10149/98933
Additional Links:
http://www.pmf.kg.ac.rs/match/content57n3.htm
Abstract:
An artificial neural network-based model was developed to predict the performances of the p,p -dinitrodibenzyl (DNDB) electroreduction under different operating conditions. Six feed-forward networks with a three-layer structure were trained and then optimized through the validation process. The performance of the neural network was examined with respect to the learning coefficient, number of hidden nodes and number of iterations required to reduce the total error to an acceptable level. The best predictions were obtained with a neural network having an 8-12-2 architecture and "tansig" "tansig" as threshold functions. The 8 nodes of the input layer correspond to two input variables, current density and quantity of electricity passed, and six state variable, temperature, initial concentration of DNDB in the electrolyte, mass transfer coefficient, electrode surface area, number of electrons involved, and volume of the catholyte. The 2 nodes of the output layer correspond to current efficiency and yield of the main reaction product, the p,p'-Diaminodibenzyl (DADB).
Type:
Article
Language:
en
Keywords:
artificial neural network; dinitrodibenzyl electroreduction
ISSN:
0340-6253
Rights:
Articles published in "MATCH Communications in Mathematical and in Computer Chemistry" are not protected by any kind of copyright. Thus these may be freely reproduced and "repositoried". What we only expect is that in all such cases it be indicated where the paper has been published. [Email from Editor of MATCH]
Citation Count:
0 [Web of Science and Scopus, 17/05/2010]

Full metadata record

DC FieldValue Language
dc.contributor.authorOlea, M. (Maria)en
dc.date.accessioned2010-05-17T07:56:09Z-
dc.date.available2010-05-17T07:56:09Z-
dc.date.issued2007-
dc.identifier.citationMatch: communications in mathematical and computer chemistry; 57(3):735-748en
dc.identifier.issn0340-6253-
dc.identifier.urihttp://hdl.handle.net/10149/98933-
dc.description.abstractAn artificial neural network-based model was developed to predict the performances of the p,p -dinitrodibenzyl (DNDB) electroreduction under different operating conditions. Six feed-forward networks with a three-layer structure were trained and then optimized through the validation process. The performance of the neural network was examined with respect to the learning coefficient, number of hidden nodes and number of iterations required to reduce the total error to an acceptable level. The best predictions were obtained with a neural network having an 8-12-2 architecture and "tansig" "tansig" as threshold functions. The 8 nodes of the input layer correspond to two input variables, current density and quantity of electricity passed, and six state variable, temperature, initial concentration of DNDB in the electrolyte, mass transfer coefficient, electrode surface area, number of electrons involved, and volume of the catholyte. The 2 nodes of the output layer correspond to current efficiency and yield of the main reaction product, the p,p'-Diaminodibenzyl (DADB).en
dc.language.isoenen
dc.publisherUniversity of Kragujevac, Faculty of Scienceen
dc.relation.urlhttp://www.pmf.kg.ac.rs/match/content57n3.htmen
dc.rightsArticles published in "MATCH Communications in Mathematical and in Computer Chemistry" are not protected by any kind of copyright. Thus these may be freely reproduced and "repositoried". What we only expect is that in all such cases it be indicated where the paper has been published. [Email from Editor of MATCH]en
dc.subjectartificial neural networken
dc.subjectdinitrodibenzyl electroreductionen
dc.titleModelling of the p,p'-dinitrodibenzyl electroreduction by using an artificial neural networken
dc.typeArticleen
dc.contributor.departmentBabes-Bolyai University of Cluj-Napoca. Department of Chemical Engineering. Romania.en
dc.identifier.journalMatch: communications in mathematical and computer chemistryen
ref.citationcount0 [Web of Science and Scopus, 17/05/2010]en
or.citation.harvardOlea, M. (2007) 'Modelling of the p,p'-dinitrodibenzyl electroreduction by using an artificial neural network', Match: communications in mathematical and computer chemistry, 57 (3), pp.735-748.-
All Items in TeesRep are protected by copyright, with all rights reserved, unless otherwise indicated.