Assessment of thermal conductivity enhancement of nano-antifreeze containing single-walled carbon nanotubes: Optimal artificial neural network and curve-fitting
Document Type
Article
Publication Date
1-1-2019
Abstract
The neural network is a technique to reduce cost and time that can be a good alternative to practical testing. This technique, which has become more important with the advancement of computer science, can also be used to predict the properties of nanofluids. To prove this claim, in this research, an optimal artificial neural network (ANN) was designed to evaluation the thermal conductivity enhancement of the SWCNTs/EG-water nanofluid using experimental data. For this goal, reported experimental enhancement for various concentrations and temperatures were employed. 35 measured data obtained from experiments have been applied to utilize ANN modeling. 80% data were chosen for network training and the remaining data were adopted for network testing. Based on the minimum mean square error (MSE), ANN model with two hidden layers and 4 neurons in each layer was selected. In addition, a new correlation was presented for predicting the thermal conductivity enhancement. Comparative results showed ANN model can forecast the thermal conductivity enhancement of nanofluids appropriately. © 2019 Elsevier B.V.
Keywords
ANN, Thermal conductivity enhancement, Nano-antifreeze, SWCNTs, Experimental data, Curve fitting
Divisions
fac_eng
Publication Title
Physica A: Statistical Mechanics and its Applications
Volume
521
Publisher
Elsevier