Prediction of RO membrane performances by use of artificial neural network and using the parameters of a complex mathematical model

Prediction of RO membrane performances by use of artificial neural network and using the... Some mathematical models have properly predicted the RO membrane performances. The equations of these models which were usually complex and time consuming were solved algebraically and numerically. The modified surface force-pore flow model is one of the best models which has predicted the RO membrane performances, for example separation factor (f), pure solvent flux (N P) and total flux (N T), better than the others. In this study, these performances were computed by use of an artificial neural networks technique by applying the parameters of this model and the physical properties of the membrane. A back-propagation feed-forward network with three layers including 9 neurons in the first layer and one neuron in the output layer was used. Minimum error was found with 20 neurons in the second layer, by trial and error. Some experimental data were used for simulating the network. The network was trained in two subsequent steps including feed-forward and error back-propagation. The datasets were randomly divided into three parts: 70 % of them were applied for training, 15 % were used for validating, and the remaining 15 % were applied for testing. The predicted values of the network were compared with experimental data existing for RO membrane performances (f, N P, and N T). A mean square error less than 0.0007 was achieved and a correlation coefficient with more than 0.99 was derived for the test datasets. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Research on Chemical Intermediates Springer Journals

Prediction of RO membrane performances by use of artificial neural network and using the parameters of a complex mathematical model

Loading next page...
 
/lp/springer_journal/prediction-of-ro-membrane-performances-by-use-of-artificial-neural-tJmFEwYGpd
Publisher
Springer Netherlands
Copyright
Copyright © 2012 by Springer Science+Business Media Dordrecht
Subject
Chemistry; Catalysis; Physical Chemistry; Inorganic Chemistry
ISSN
0922-6168
eISSN
1568-5675
D.O.I.
10.1007/s11164-012-0835-z
Publisher site
See Article on Publisher Site

Abstract

Some mathematical models have properly predicted the RO membrane performances. The equations of these models which were usually complex and time consuming were solved algebraically and numerically. The modified surface force-pore flow model is one of the best models which has predicted the RO membrane performances, for example separation factor (f), pure solvent flux (N P) and total flux (N T), better than the others. In this study, these performances were computed by use of an artificial neural networks technique by applying the parameters of this model and the physical properties of the membrane. A back-propagation feed-forward network with three layers including 9 neurons in the first layer and one neuron in the output layer was used. Minimum error was found with 20 neurons in the second layer, by trial and error. Some experimental data were used for simulating the network. The network was trained in two subsequent steps including feed-forward and error back-propagation. The datasets were randomly divided into three parts: 70 % of them were applied for training, 15 % were used for validating, and the remaining 15 % were applied for testing. The predicted values of the network were compared with experimental data existing for RO membrane performances (f, N P, and N T). A mean square error less than 0.0007 was achieved and a correlation coefficient with more than 0.99 was derived for the test datasets.

Journal

Research on Chemical IntermediatesSpringer Journals

Published: Oct 16, 2012

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve Freelancer

DeepDyve Pro

Price
FREE
$49/month

$360/year
Save searches from Google Scholar, PubMed
Create lists to organize your research
Export lists, citations
Read DeepDyve articles
Abstract access only
Unlimited access to over
18 million full-text articles
Print
20 pages/month
PDF Discount
20% off