|
In a project sponsored by Axxion Corp., a major manufacturer
of personal computer enclosures and chassis, researchers at the
University of Texas used Coolit to analyze the impact of various
ventilation configurations on airflow within a desktop computer
chassis. By creating the simulations, the researchers were able
to alter the configuration without having to physically modify
any units.
The default chassis layout has three vents; two in the front
and one in the rear (see Figure 1). Because the rear vent is
closest to the processor heat sink and fan, it is reasonable to
expect that blocking this vent would result in reduced heat
transfer from the processor and possible overheating. Using
Coolit, however, it was found that this is not necessarily true
(see Figure 2).
As can be seen in the figures, blocking the rear vent
increased flow velocities through the front vents (greater
spacing between arrows on streamlines). This results in better
ventilation of the chassis as a whole and, therefore, reduced
temperatures. In addition, the back vent flow was colliding
with the front vent flow, creating a stagnation region near the
processor. Blocking the back vent eliminated this stagnation
region, further improving convection around the processor. Such
counterintuitive insights are made possible by CFD simulations - no
amount of prototype testing would obtain such detailed and
quantitative data.
To verify modeling accuracy, Coolit flow predictions were
compared to hot-wire anemometry measurements. Table 1 shows a
comparison between predicted and actual maximum vertical
velocity in the default chassis configuration.
|
Location
|
Measured Velocity
(m/s)
|
Coolit Predicted Velocity (m/s)
|
Percent Difference
(%)
|
|
right side of heat sink
|
4.93
|
5.13
|
4.1
|
|
right side of power supply
|
4.01
|
3.71
|
-7.5
|
|
right side of power supply fan
|
2.96
|
2.85
|
-3.7
|
|
upper vacant vent
|
1.64
|
1.61
|
-1.8
|
|
