Simulation of buoyancy-driven natural ventilation of buildings—Impact of computational domain

Abstract Two computational domains have been used for simulation of buoyancy-driven natural ventilation in vertical cavities for different total heat fluxes and wall heat distributions. Results were compared between cavities with horizontal and vertical inlets. The predicted ventilation rate and heat transfer coefficient have been found to depend on the domain size and inlet position as well as the cavity size and heat distribution ratio. The difference in the predicted ventilation rate or heat transfer coefficient using two domains is generally larger for wider cavities with asymmetrical heating and is also larger for ventilation cavities with a horizontal inlet than those with a vertical inlet. The difference in the heat transfer coefficient is generally less than that in the ventilation rate. In addition, a ventilation cavity with symmetrical heating has a higher ventilation rate but generally lower heat transfer coefficient than does an asymmetrically heated cavity. A computational domain larger than the physical size should be used for accurate prediction of the flow rate and heat transfer in ventilation cavities or naturally ventilated buildings with large openings, particularly with multiple inlets and outlets. This is demonstrated with two examples for natural ventilation of buildings.