Predicting the effects of magnesium oxide nanoparticles and temperature on the thermal conductivity of water using artificial neural network and experimental data

Highlights: • Proposing a correlation for estimating thermal conductivity of MgO-water nanofluid. • Artificial neural networks with various numbers of neurons have been assessed. • Comparing output of ANN with the results of the proposed empirical correlation. • ANN modeling was more accurate than curve-fitting method. The current paper first presents an empirical correlation based on experimental results for estimating thermal conductivity enhancement of MgO-water nanofluid using curve fitting method. Then, artificial neural networks (ANNs) with various numbers of neurons have been assessed by considering temperature and MgO volume fraction as the inputs variables and thermal conductivity enhancement as the output variable to select the most appropriate and optimized network. Results indicated that the network with 7 neurons had minimum error. Eventually, the output of artificial neural network was compared with the results of the proposed empirical correlation and those of the experiments. Comparisons revealed that ANN modeling was more accurate than curve-fitting method in the predicting the thermal conductivity enhancement of the nanofluid.