Page IV ■ 98 
Ventilation Design Handbook on Animal Research Facilities Using Static Microisolators 
significant impact on the airflow and temperature distributions. This is particularly important 
when using a diffuser of the radial type since they have inherently low momentum, so the jet is 
easily affected by small, induced air currents. 
The problems are clearly identified by the fact that measured values of air speed are well in 
excess of the theoretical limit of 0.1 1 m/s (22 fpm) well away from the diffuser jet (see appendix 
I: section 4.1.2). The CFD results only show these larger air speeds where they would be 
expected - close to the supply air diffuser and the room exhaust. The advantage of the 
computational fluid dynamics approach is that, by definition, the predicted values for the 
parameters in the analysis have to be consistent since the approach is based on the fundamental 
laws for the conservation of mass, momentum, and energy. The experimental approach does not 
benefit from any such luxury since the measurements are each independently recorded by 
parameter and by location. Air velocities of the low speeds seen in a room (in this case typically 
below 0. 10 m/s (20fpm)), for example, are inherently difficult to measure for several reasons: 
• The low velocities are very susceptible to fluctuations as well as variations in the boundary 
conditions over the period of the experiment. Extending the sampling period for individual 
points to obtain a more representative average can be counterproductive since it increases the 
total period of the experiment and makes it more difficult to maintain the boundary 
conditions for the experiment. 
• The constant temperature thermistor based sensing head (used in hot film anemometers) 
becomes inaccurate at low speeds because the heating of the head causes local air velocities 
of a similar order. 
• The presence of the measurement equipment, and any traversing equipment (see figures 4.32 
and 4.33) can influence the local velocities. 
The scalar quantities such as temperature and concentration are easier to measure, and as this 
section (and the following section, section 4.2. 3. 2. 2 show) shows good correlation can be seen 
between the measured and predicted data. 
Given that the predictions of temperature and concentration are good then, by the laws of 
conservation the air velocities must also be well predicted, with the discrepancies between 
measured and predicted data being accounted for by difficulty in measurement outlined above. 
It should be also be remembered that the CFD model of the radial diffuser was found to 
accurately predict the manufacturer’s throw data (see section 4.2.2), and so a good degree of 
confidence should be place in the CFD representation of the diffuser. 
As far as the temperature measurements are concerned, the experimental data plots, for example, 
figure 4.72, highlight the difficulty in maintaining isothermal conditions within test facilities. In 
this example, the differing boundary wall temperatures have a marked impact on the internal 
room temperatures. 
