Volume II - Section I - Introduction 
Pa 2 e II - 5 
photoperiod on the mice. The supply air was assumed to have a zero concentration of CO 2 . This 
source allowed the additional concentration of CO 2 in the air to be calculated in the simulation. It 
also allowed the concentration of NH 3 , among other things, to be calculated by scaling even 
though it has a different molecular weight from both air and CO 2 . This was possible as the 
magnitude of the source was very small and the resulting concentrations were so low as to have a 
negligible effect on the density of the mixture of air, CO 2 , and NH 3 . In effect, the CO 2 and NH 3 
are intimately mixed with, and flow with, the air. 
Experimental data later showed the generation rate of CO 2 was actually higher than the source 
used in the CFD simulations at 0.90 g/hr/lOOg mouse body weight. This means the concentration 
of CO 2 in the room and cages was derived from the simulated concentration multiplied by a 
scaling factor (0.90/0.76). The concentrations of NH3 in both the cages and the room were also 
derived by scaling the concentration with a factor specified in the post-processing of the data. 
This factor was assumed to vary according to the number of days that passed since the bedding 
in the cage was changed, along with the average relative humidity in the cages. See volume I, 
section 4.2. 1.2 for the experimental determination of the factors. 
Background levels of CO 2 and NH 3 were assumed to be zero. This means that all values quoted 
in the CFD section of the report are relative to the background level. If an absolute value for CO 2 
is required, an additional amount in the range of SOOppm to 700ppm should be added for most 
locations. 
The remaining cage boundary conditions are associated with the transfer mechanisms for 
air/gases to enter/leave the cage. The cracks at the side of the cage were modeled as 6.35e-3m 
(0.25”) high planar resistances, with the loss coefficient for these resistances having been 
determined from the cage wind tunnel CFD simulations (see volume I section 4.2. 1.2). The top 
of the cage, which was filtered, was defined as a combination of a planar resistance and a planar 
source. The determination of the loss coefficient for the resistance, and the coefficient for the 
source has been outlined (see volume I, section 4.2. 1.1). 
The spacing of the cages on the shelves was dependent on whether the racks were single density 
(7 cages per shelf), or double density (14 cages per shelf). In the single density cases, the cages 
were centrally located in the short dimension, and equally spaced in the long dimension. The 
spacing was 4.88e-2m (1.92”) from comer of cage to comer of adjacent cage. In the double 
density racks, the cages were equally spaced in both the long and short dimensions. The spacing 
was 2.20e-2m (0.87”) and 4.88e-2m (1.92”), respectively. 
Change Station Model 
Two alternative change stations were considered in this project. Both stations were constmcted 
primarily from rectangular blocks and triangular prisms. The internal stmcture and flow field 
were of no concern in this project. It was only the effect of the station on the room airflow that is 
of importance. 
The first design is shown in figure 1.03. The station had overall dimensions of 1.32m (52”) x 
0.86m (34”) x 1.83m (72”). This design was effectively passive in terms of direct flow field 
interaction. In particular, the station internally recirculated a flow of 350 cfm (1.65e-l m^ /s). 
