Prediction of thermal conductivity of carbon nanotube-EG nanofluid using experimental data by ANN
Journal article
Authors/Editors
Strategic Research Themes
No matching items found.
Publication Details
Author list: Esfe M.H., Wongwises S., Rejvani M.
Publisher: Bentham Science Publishers
Publication year: 2017
Journal: Current Nanoscience (1573-4137)
Volume number: 13
Issue number: 3
Start page: 324
End page: 329
Number of pages: 6
ISSN: 1573-4137
eISSN: 1875-6786
Languages: English-Great Britain (EN-GB)
View in Web of Science | View on publisher site | View citing articles in Web of Science
Abstract
Background: The artificial neural network has been employed to predict the thermal conductivity of the carbon nanotube–ethylene glycol (CNT-EG) nanofluid based on experimental data. The main aim of this study is to find the best training algorithm for modeling the thermal conductivity of nanofluids. Methods: Different activating functions and two training algorithms have been tested to train the neurons. The architecture of this modeling is the same and consists of one hidden layer with two neurons. The input parameters of the network include 20 data of temperatures (15–55°C) and volume concentrations (2.2–7.8 vol.%), and the output of the network is the thermal conductivity coefficient. Results: The results indicate that the trainbr algorithm with the Elliotsig activating function responses have a higher regression coefficient and a lower mean square error. The results show also that an artificial neural network can estimate the experimental results with high precision in a wide range of temperatures and concentrations of carbon nanotubes. Conclusion: The comparative graph with experimental data and artificial neural network modeling results in terms of temperature for different volume fractions revealed that the neural network can estimate the experimental results with high precision at a wide range of temperatures and concentrations of CNTs. Also, the results indicated that the neural network was not a proper tool for outside of the available data and should be used in the same range in which it was trained. © 2017 Bentham Science Publishers.
Keywords
CNTs-EG nanofluid, Train algorithm
Contact Information