Computational analysis of magnetohydrodynamic natural convection in a square cavity with a thin fin

A numerical study of laminar natural convection in a square cavity with a thin fin that is under the influence of a uniform magnetic field is presented. The side walls of the cavity are kept at different temperatures and the horizontal walls are thermally insulated. An Adaptive Network-based Fuzzy Inference System (ANFIS) approach and an Artificial Neural Network (ANN) approach are developed, trained and validated using the results of Computational Fluid Dynamics (CFD) analysis. The effects of pertinent parameters on fluid flow and heat transfer characteristics are studied. Among these parameters are the Rayleigh number (103≤Ra≤106), the Hartmann number (0≤Ha≤100), the position of the thin fin (0.1≤Yp≤0.9) and the length of the thin fin (0≤Lp≤0.8). The results show that ANFIS and ANN can successfully predict the fluid flow and heat transfer behaviour within the cavity in less time without compromising accuracy. In most cases, ANFIS can predict the results more accurately than ANN. [Display omitted] •Magnetohydrodynamic laminar natural convection in a square cavity with a fin is examined.•Three numerical approaches of ANFIS, ANN and CFD are used in the investigation.•ANFIS and ANN accurately predict the cavity’s thermal performance in less time.•Magnetic field affects natural convection especially at higher Rayleigh numbers.•Fin’s length and position significantly affect the heat transfer rate of the cavity.