Multi-grid prediction of conjugate heat transfer and air flow in buildings

The heat conduction through the walls changes the heat load and its distribution in a room, and thus affects the air flow pattern in a buoyancy-controlled ventilated room. This paper presents a methodology of how conjugate heat transfer and air flow in a room can be handled in an efficient computational fluid dynamics (CFD) code. The wall and indoor air regions are simulated simultaneously. The standard k-e model is used for modelling the turbulence in rooms. The transport equations take the same form in both fluid (air) and solid (wall) regions. The boundary conditions for the momentum and energy equations are specified at the outer surface of the walls. A multi-grid solver is applied to the problem of conjugate heat transfer and turbulent indoor air flow. The local grid-refinement technique is introduced to add local grid points in the regions where variable gradients are large. The multi-grid solver shows a much faster convergence rate than the single grid solver for the non-isothermal cases in this study. The developed approach is applied to a modelled 3D room and a room with displacement ventilation. This approach together with the previously developed general thermal boundary conditions, allows the study of interactions between indoor and outdoor environments.