Free Convection of Hybrid Nanofluids in a C-Shaped Chamber under Variable Heat Flux and Magnetic Field: Simulation, Sensitivity Analysis, and Artificial Neural Networks

In the present investigation, the free convection energy transport was studied in a C-shaped tilted chamber with the inclination angle α that was filled with the MWCNT (MultiWall Carbon Nanotubes)-Fe3O4-H2O hybrid nanofluid and it is affected by the magnetic field and thermal flux. The control equations were numerically resolved by the finite element method (FEM). Then, using the artificial neural network (ANN) combined with the particles swarm optimization algorithm (PSO), the Nusselt number was predicted, followed by investigating the effect of parameters including the Rayleigh number (Ra), the Hartmann number (Ha), the nanoparticles concentration (φ), the inclination angle of the chamber (α), and the aspect ratio (AR) on the heat transfer rate. The results showed the high accuracy of the ANN optimized by the PSO algorithm in the prediction of the Nusselt number such that the mean squared error in the ANN model is 0.35, while in the ANN model, it was optimized using the PSO algorithm (ANN-PSO) is 0.22, suggesting the higher accuracy of the latter. It was also found that, among the studied parameters with an effect on the heat transfer rate, the Rayleigh number and aspect ratio have the greatest impact on the thermal transmission intensification. The obtained data also showed that a growth of the Hartmann number illustrates a reduction of the Nusselt number for high Rayleigh numbers and the heat transfer rate is almost constant for low Rayleigh number values.