Thermal conductivity modeling of graphene nanoplatelets/deionized water nanofluid by MLP neural network and theoretical modeling using experimental results

The purpose of this study is to predict the thermal conductivity of graphene nanofluid by using multilayer perceptron (MLP) artificial neural network (ANN). Experimental measurement results of thermal conductivity of graphene nanoplatelets/deionized water nanofluid in 25 °C to 50 °C temperature and in weight percentages of 0.00025, 0.0005, 0.001, and 0.005 have been used in order to modeling by artificial neural network. Furthermore, in order to evaluate accuracy of the model in predicting the coefficient of nanofluid thermal conductivity, indexes of root mean square error (RMSE), coefficient of determination (R2), and mean absolute percentage error (MAPE) have been used which are equal to 0.04 W/mk, 99% and 0.26% respectively. In this study, considering all common methods for theoretical modeling of nanofluid thermal conductivity, Nan's theoretical method has been applied for assessing the importance of modeling and predicting the results using ANN. According to our research, the results of indexes and predictions show high accuracy and certainty of ANN modeling in comparison with experimental results and theoretical models.