ANN and RSM‐Based Comparison on a Natural Convective Octagonal Cavity Containing Ternary Hybrid Nanofluid Including External Magnetic Field

The primary aim of this work is to explore how an external magnetic field affects the natural convective fluid and thermal transmission in an octagonal cavity filled with a Cu‐CuO‐Al 2 O 3 ‐H 2 O ternary hybrid nanofluid because of the numerous practical uses of natural convection. The cavity's center is occupied by a circular uniform heat source (T h ), and the vertical walls act as a heat sink (T c ). Thermal insulation is installed in the remaining walls. The finite element method (FEM) is used to numerically simulate the governing equations. Artificial neural network (ANN) is utilized for data prediction, and another statistical technique called response surface methodology (RSM) is also applied. The results are expressed in terms of average Nusselt number (Nu av ), isotherms, streamlines, and velocity profiles for different values of Rayleigh number ( Ra ), Hartmann number ( Ha ), and nanoparticle volume fraction ( ϕ ). RSM is also used to make a proper correlation between Ra , Ha , and ϕ with the response Nu av . The findings of examining the rate of sensitivity show that although Ha shows the opposite tendency, Ra increases thermal performance. Moreover, adding solid particles of Cu, CuO, and Al 2 O 3 to H 2 O rather than the base fluid increases the rate of heat transfer by up to 10.73%. A strong basis for thermal system optimization using a ternary hybrid nanofluid is provided by the combination of numerical, statistical, and ANN methods for the investigation and guess of heat passage methods.