Evaluation of the effects of the presence of ZnO- TiO2 (50%-50%) on the thermal conductivity of Ethylene Glycol base fluid and its estimation using Artificial Neural Network for industrial and commercial applications

In this study, the thermal conductivity (knf) of ZnO-TiO2 (50 %-50 %)/ Ethylene Glycol hybrid nanofluid using Artificial Neural Networks (ANNs) was predicted. The nanofluid was pre-pared at different volume fractions (u) of nanoparticles (u = 0.001 to 0.035) and temperatures (T = 25 to 50 degrees C). In this study, an algorithm is presented to find the best neuron number in the hidden layer. Also, a surface fitting method has been applied to predict the knf of nanofluid. Finally, the correlation coefficients, performances, and Maximum Absolute Error (MAE) for both methods have been presented and compared. It could be understood that the ANN method had a better abil-ity in predicting the knf of nanofluid compared to the fitting method. This method not only showed better performance but also reached a better MAE and correlation coefficient.(c) 2023 The Author(s). Published by Elsevier B.V. on behalf of King Saud University. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Automated summary

The study predicted the thermal conductivity (knf) of ZnO-TiO2 (50% - 50%)/ Ethylene Glycol hybrid nanofluid using Artificial Neural Networks (ANNs). The nanofluid was prepared at different volume fractions (u) of nanoparticles (u = 0.001 to 0.035) and temperatures (T = 25 to 50 degrees C). The study introduced an algorithm to determine the optimal number of neurons in the hidden layer and used a surface fitting method for knf prediction. The results showed that ANN method outperformed the fitting method in terms of knf prediction, with better MAE and correlation coefficient.