Abstract

Zonal models are intermediate between CFD and single-zone models. They can generate results faster than CFD and more accurately than single-zone models. They provide global information on the distribution of airflow and temperature in a room and they have been used to predict moisture distribution, thermal comfort, contaminant distribution and personal exposure. The zonal models predict the airflow pattern reasonably well for natural convection since airflow is only due to buoyancy. For forced convection, they provide good prediction only in the specific zone, which indicates that the power-law model (PLM), used in the standard zone, is the main cause of the discrepancies of the zonal approach. In this study, three methodologies were investigated for improving the prediction capability of the PLM. The first was investigation of the significance of the value of the flow coefficient ( K ). In this methodology, one K value was used for each cell in the standard zone. Values other than the commonly used value of K = 0.83 were considered. The result revealed that for forced convection, the value of K has no impact on the prediction of the PLM. The second method employed CFD generated data to estimate K value for each cell. Two models were developed using this approach: PLM with variable K (PLMK) and PLM with variable K where K is a function of y/H (PLMK func ). The PLMK was found to provide better predictions than the PLMK func and the PLM. In the third methodology, the direct and indirect combinations of the PLM and the SDM were explored. The direct combination has provided different zonal models. The indirect combination has resulted in the modified power-law model (MPLM). The comparison of the predictions of the zonal models, for isothermal and non-isothermal room, with experimental data and CFD shows that the MPLM is superior to the other zonal models. Additionally, zonal approach was applied for the prediction of the temporal and spatial temperature distribution in a mechanically ventilated DSF and good agreement with experimental data was obtained. Since the zonal approach provides global information faster than CFD, potential applications of the zonal approach include the prediction of temperature distribution and contaminant dispersion in whole building. This is very important for protecting building occupants from fire accidents, chemical/biological attacks by terrorists and from natural occurring diseases such as SARS and avian flu