Artificial brain structure-based modeling to predict the photo-thermal conversion performance of graphene nanoplatelets nanofluid using experimental data

Radiation is one of the means of thermal energy and heat transfer. Therefore, investigating and determining materials’ abilities to absorb energy and heat can be beneficial in design process of different thermal systems. In this study, thermal radiation characteristics (e.g., absorption rate and transmittance coefficient) of graphene nanofluids in different wavelengths are predicted using modeling approach based on feedforward neural network with Levenberg–Marquardt backpropagation algorithm as a type of artificial neural network (ANN). For model development, data from experimental measurements of thermal conductivity of nanofluid with graphene nanoplatelet/water deionized in the laboratory were used. Different wavelengths ranging from 200 to 2500 nm and various mass fractions of 0.025, 0.050, 0.075 and 0.100 were considered in developing the model. Overall, 924 data samples were used of which 648 were for ANN training and the rest were for calibration and validation. To ensure high accuracy level of modeling and prediction, statistical indexes such as root-mean-square error, mean absolute percentage error and coefficient of determination ( R 2 ), were employed, which indicate values of 0.041, 0.059 and 0.998, respectively. Modeling results and associated statistical indexes based on experimental data demonstrate high accuracy and validity of the ANN modeling compared to other prediction methods.