Page IV - 80 
Ventilation Design Handbook on Animal Research Facilities Using Static Microisolators 
4.2. 1 .2.2 Comparison of CFD Results vs. Experimental Data 
Presented below, in tables 4.2.02 to 4.2.12, are a series of comparisons between the experimental 
data sets and equivalent CFD simulation cage ACH for the chosen optimal values for the cage 
side crack loss coefficients. Note that, because of time constraints, only a representative sample 
could be considered from the wide range of experimental data available. The focus of the sample 
was to pick orientations that the cage was more likely to experience in the animal facility room 
environment, in particular, parallel cage orientation, and appropriate air velocities, in particular, 
40 fpm (0.2m/s) and below. 
A complete listing of the experimental cage ACH data is given in appendix I: section 2.2. 
The comparisons show good agreement between the experimental dari and CFD simulation 
results for the range of experimental series considered. In the majority of cases considered, the 
difference between the experimental and CFD results is under 20 percent. This error can be 
considered reasonable for this set of validation and calibration exercises. In particular, the CFD 
results show that the calculated value for the cage ACH is sensitive to the exact location of the 
sampling tubes, and the sampling tube holes themselves: this is because of the stratification of 
the CO 2 or SF 6 concentrations in the cages (see figure 4.59). Relatively small variations from the 
quoted location of the experimental sampling tubes would translate to errors in the CFD 
calculation. Further, as table 4.2.02 demonstrates, some level of error should be accepted in the 
experimental procedure. 
Table 4.2.02. Comparison of CFD results against Series Set Base: Parallel Orientation Results. 
Tunnel Velocity 
(FPM) 
Ventilation Rate (CFM) 
1 1/min CO 2 
Parallel - Heater On 
CFD 
Series Set Base 
Series Set Base 
(Repeat) 
20 
0.15 
0.19 
— 
30 
0.17 
0.21 
0.18* 
The percent difference between the two separate experimental readings is 14.3 percent. 
