Applying a momentum-based variable formulation in the SIMPLE algorithm to numerically solve thermo-buoyant turbulent flow in enclosures

Abstract

Natural or buoyant convection flow is an exemplary heat transfer phenomenon, with growing applications in various industries. This article develops a new algorithm, which models and solves the buoyancy-driven turbulent flows in enclosures more accurately than the past similar solvers. A careful literature review shows that the past existing approaches have mostly had serious limitations to apply their algorithms to buoyancy-driven flows with high temperature differences magnitude because of employing the classical Boussinesq approximation. As the novelty of this study, it benefits from a momentum-based variable approach in the context of the semi-implicit method for the pressure linked equations (SIMPLE) algorithm, which lets it accurately solve the strong compressible buoyant flows with high temperature differences. The algorithm is applied to both the Navier-Stokes and the accompanied turbulent flow governing equations using OpenFOAM 4.1 as the platform. To validate the developed algorithm, the current results are compared with experimental data in both square and tall cavities considering low (8.6 × 105), high (1.43 × 106), and very high (1.58 × 109) Rayleigh numbers. As the major contribution of this work, it improves the accuracy of the thermo-buoyant turbulent flow prediction at both low and high Rayleigh numbers. All test cases are carried out employing two different turbulence models of k-ω and k-ε. Furthermore, comparing the results of the present non-Boussinesq algorithm and those of the past developed methods with the experimental data, it is shown that the present algorithm provides a more accurate prediction for the temperature field, that is, <10% differences with the experimental data. Moreover, the present maximum velocity results surpass the solution of the past numerical methods and show <3% differences with the experimental data.