Page FV - 78 
Ventilation Design Handbook on Animal Research Facilities Using Static Microisolators 
Placed on the shelf was the CFD model representation of the cage. A typical representation of 
the cage, without instrumentation, is shown in figure 4.55. The dimensions of the cage were set 
as 0.27 m x 0.16m x 0.21m (10.7” x 6.38” x 8.39”): these dimensions retained the same volume 
as in the physical case. The sides of the cage were modeled as thin plates, with the thickness and 
conductivity of the plates set to those of the cage polycarbonate. The bottle was represented 
using a combination of rectangular prisms and cuboid blocks. The volume of the bottle was 
retained, as was the location of the bottle in the cage. A food supply was modeled using two 
triangular prisms. The bedding was represented as a dimension 0.27m x 0.16m x 1.3e-2m (10.7” 
x 6.38” x 0.5”) rectangular block. The alternative representations of the mice heater were 
modeled using rectangular blocks. In particular, the original heater was modeled as a 1.59e-2m x 
0.32e-2m x 0.32e-2m (5/8” x 1/8” x 1/8”) block, with the heat flux set to 2.3W, while the 
huddled mice were modeled as a block of dimension 0.1 lm x 8.6e-lm x 2.2e-2m (4 1/4” x 3 3/8” 
x 7/8”), with the surface temperature set to 26.7°C (80.0°F). 
There are two transfer mechanisms for the air and tracer gas to enter/ leave the cage, namely the 
top of the cage, that includes filter media, and the side cracks of the cage. 
The top of the cage has two constituent parts that had to be represented using CFD boundary 
conditions: the filter media; and the top of the cage itself, that consists of regular arrays of holes 
in the polycarbonate material. The filter material was identified as Reemay #2024, 12 mils, 2.1 
oz/ yd . Using manufacturer’s data, a pressure drop vs. wind tunnel speed graph could be plotted 
in figure 4.56. The profile was then approximated to a polynomial expression that could be 
converted to CFD boundary conditions. In particular, the polynomial expression can be 
expressed as: 
DP = 70.277 v 2 + 307.37 v (4.22) 
As the average velocities through the filter are relatively small (of the order of 0.17 cfm (8e-4 
m/s)), the linear contribution dominated the pressure drop. Using boundary conditions defined in 
section 5. 1.6.2, the first term of the right hand side was represented using a planar resistance, 
while the second term was represented using a planar source of momentum. The loss coefficients 
were set appropriately for each boundary condition to replicate the polynomial expression. As 
the flow through the filter media is laminar, the turbulent viscosity at the plane of the media was 
reduced to very low levels. To achieve this, the value of k (the turbulent kinetic energy) was set 
at le-5 at the planar source, while e (the rate of dissipation of k), was set to le5. 
The cage top material itself was represented through the calculation of the free area ratio of the 
top surface, and the addition of the loss coefficient to the planar resistance term. The free area 
ratio was calculated to be 0.35, that gives a loss coefficient of 12.35 (Idelchik (1989)). 
The settings for the side crack boundary conditions were the most proolematical to specify 
because of physical uncertainties. In particular, the top lid of the cage does not fit well on the 
lower section of the case because the meshing is often deformed. The first step was to define 
