Methodology for the Evaluation of Natural Ventilation in ... - Cham

difficult to select the boundary conditions for modeling buildings, as was found in the analyseshere. With the experimental modeling, the ambient conditions were constantly measured toensure that the assumptions of consistency and steady state conditions could be used. If anynumber of variables changed, it affected the ambient temperature surrounding the model.Decreasing the number of inlet windows or outlets (windows or stacks) increased thetemperature inside the model and increased the heat loss to the test chamber. The geometry ofthe window and measurement location influenced the recorded inlet and outlet velocity.Water models aim to meet dynamic similarity requirements, but not necessarily thermal orgeometric requirements. Both heated and salt-water models demonstrate macroscopic flows forspaces, focusing on the restrictions between spaces and replicating the general phenomena fornatural ventilation, but not necessarily specifically for a building. They are constructed at smallscale to obviate the requirement for a large amount of space. However, they do require theequipment and apparatus to run the experiments. Water models do not have abundant amountsof detail, focusing rather on the connection of spaces and the restrictiveness between zones. Onelarge benefit of water bath models is the ability to incorporate flow visualization easily in theexperimental techniques. Dyed water, at the same temperature as the ambient, is easily injectedat the inlets and light shining through the model is easily projected on to a screen behind themodel. Water bath models have less than 1 percent of the heat input lost through the modelwalls (Gladstone and Woods 2001).With air models, the heat loss through the walls is significantly higher as described in chapterseven. The heat loss can be controlled to a certain extent, particularly with the exterior walls;however, heat transfer between floors is also a concern. The reduced-scale air model developedin this methodology allows for a certain amount of detail, which contributes to the success ofmodeling a prototype building. For example having the railings around the atrium in thecombined wind-buoyancy driven flow influenced the flow patterns in the atrium and in thenorthern half of the model. In general, with reduced-scale modeling, a small variation, as little as3 mm, can influence the resulting flow patterns. In the twelfth-scale air model, there is moredetail than most reduced-scale models. Therefore, there must be an increased attention to detailwith respect to major obstacles that might influence the flow. Airflow visualization is moredifficult with the twelfth-scale air model than the water models, but can demonstrate the flowpatterns more distinctly than the water models, which show whether a space is well mixed or not,and not specific flow patterns within an enclosed space.Numerical solutions are complicated in the modeling technique and require an experienced userto obtain meaningful results. If the boundary conditions are not known, describing the model inthe domain becomes much more difficult. With CFD simulations, complex mathematical modelsare solved iteratively until the solution converges. The selection of turbulence models willinfluence the results and the computing time required for the solution to converge. There islimited interaction between the surfaces and the working fluid, requiring assumptions ofadiabatic walls and no radiation between surfaces. However, CFD does provide detailedtemperature distributions, air velocities and flow patterns when modeled correctly. Thesesimulations can help explain experimental results both at the reduced and full-scale.164