Statistical analysis of enriched water heat transfer with various sizes of MgO nanoparticles using artificial neural networks modeling

In the present research, experimental data of the heat transfer coefficient for MgO aqueous nanofluids are modeled by the MLP artificial neural networks (ANN), for 4 types of nanoparticles with diameters of 20, 40, 50 and 60 nm, at 4 solid volume fractions of 0.5%, 1%, 1.5% and 2%, and at 11 Reynolds numbers from 3000 to 25,000. Modeling and predicting of the data like the empirical data, is proved the well capability of the ANN in modeling the data related to nanofluids’ heat transfer coefficient. Another interesting point is that, the heat transfer coefficient increases by decline of the nanoparticles’ diameter, and there is a direct relationship between the rise of solid volume fraction and the heat transfer coefficient. The increment rate of heat transfer coefficient, remained unchanged by increasing the Reynolds number, and increased with the rise of solid volume fraction. Present investigation showed that, ANN is able to save all the rules hidden in these changes with a high accuracy and take the advantage of them to predict the other data. In addition, a correlation in terms of variables affecting the heat transfer coefficient is obtained and presented in the article. [Display omitted] •Modelling and predicting the various size of MgO-water nanofluid data is well proved.•Heat transfer coefficient increases by reducing of the nanoparticles’ diameter.•A direct relationship exist between the volume fraction and heat transfer coefficient.