For accurate prediction of flow, scalar transport and wall heat and mass transfer in complex building space we propose a time-dependent RANS (T-RANS) approach which resolves in time and space the large-scale convective motion and associated deterministic eddy structure. The residual (“subscale”) turbulence is modeled by a single-point closure. The method can be regarded as Very Large Eddy Simulations (VLES) since the deterministic and modeled contribution to the turbulence moments are of the same order of magnitude. The modeled part becomes dominant in the near-wall regions where there are no large eddies and the proper choice of the subscale model is especially important for predicting wall friction and heat transfer. We use an ensemble-averaged 〈k〉 - 〈ε〉 - 〈θ2〉 algebraic stress/flux/concentration closure as the subscale model which can provide information about the stress and heat/species flux anisotropies. The method is especially advantageous for predicting flows driven or affected by thermal buoyancy, for which the conventional eddy-viscosity/diffusivity RANS models and gradient transport hypotheses are known to fail even in simple generic configurations. The approach was validated in a series of buoyancy-driven flows for which experimental, DNS and LES data are available. Examples of full-scale application include computational simulations of real occupied and furnished residential or office space in which the furniture elements and persons are treated as passive blocking elements. The simulation showed that the T-RANS approach can be used as a reliable tool for a variety of applications such as optimization of heating and ventilation system, building space insulation, indoor quality, safety measures related to smoke and fire spreading, as well as for accurate wall heat and mass transfer predictions.

This content is only available via PDF.
You do not currently have access to this content.