Forecasting of Convective Heat Transfer Coefficient in Turbulent Flow of Different Nanofluids in Circular Tubes, Using Artificial Neural Network

Modeling of turbulent convective heat transfer of nanofluids in circular tubes with constant temperature and constant heat flux boundary condition have been performed using artificial neural network. 610 sets of data have been gathered using previous investigations and have been used to train neural network (ANN). The investigated nanoparticles are: TiO2, Graphene, SiC, CuO, SiO2, Fe3O4, and Cu. The base fluid for all these nanofluids is water. The neural network used has 6 inputs, which includes: nanoparticle density, nanoparticle size, nanoparticle volume fraction, flow Re number, type of boundary condition (constant heat flux or constant temperature) and the amount of heat flux or temperature related to these boundary conditions. Also, the output of neural network is Nusselt number. Comparing our results with previous investigation, showed that the proposed ANN topology are in good agreement. In this study, the proposed topology of R2=0.9998 have been choosen between 400 examined ones.