Measurement of the thermal conductivity of MWCNT-CuO/water hybrid nanofluid using artificial neural networks (ANNs)

In this paper, artificial neural networks (ANNs) are developed to predict the thermal conductivity (k (nf)) of multi-walled carbon nanotubes (MWCNTs)-CuO/water nanofluid. After generating experimental data points, an algorithm is proposed to find the optimum ANN regarding the best performance. Three different states including ANN, experimental, and fitting method have been evaluated, and their errors in k(nf) prediction have been investigated. Regarding the obtained results, it can be seen that the best and worst neuron numbers are 8 and 31, respectively. Then, using curve fitting method, the behavior of nanofluid is predicted by a surface equation with third order. Finally, the ANN results and fitting results have been compared. Finally, it is found that the ability of the ANN to predict the k(nf) is greater. It was also found that the ANN has better performance and correlation and thus less error in the predicted data. On the other hand, comparing methods in predicting the k nf is an important issue. The use of ANNs in predicting the k(nf) as a new approach can lead to a great contribution in determining the most desirable performance and achieving the best and most accurate state. In addition, mean squared error (MSE) has obtained 2.4451e-05 for fitting method. According to the experimental data, it can be seen that in phi = 0.6% and T = 50 degrees C, an increase of more than 30.38% has occurred in the k(nf) compared to the ambient temperature.

Automated summary

In this paper, artificial neural networks were developed to predict the thermal conductivity of multi-walled carbon nanotubes-CuO/water nanofluid, with the best ANN model and curve fitting method used to predict the behavior of the nanofluid. Results showed that the ANN outperformed in predicting k(nf) with better performance and correlation.