Appendix I 
Page A I- 129 
If the secondary velocity fluctuation were filtered out, the velocities in figures 4.1.01 and 4.1.02 
can be reconstructed as in figure 4.1.03, which only contains the primary velocity fluctuation. 
The turbulence intensities yielded from figures 4.1.01 and 4.1.03 will be approximately the same. 
From figure 4.1.03, the primary velocity fluctuations have a frequency of 3.7 Hz for U d = 350 
fpm and 3.5 Hz for U d = 150 fpm. Therefore, the velocity fluctuation frequency (f v ) for U d = 350 
fpm can be considered as 4 Hz (>3.7). 
It has been proven that the energy density in a turbulent flow attenuates quickly at high 
frequencies (typically higher than 10 Hz). The relationship between the energy density and the 
frequency of the air velocity fluctuation is approximately in the order of f v ' 3 (Hinze, 1975). 
Examples of energy density spectra of air velocities are shown in figures 4.1.04 and 4.1.05 
(Zhang, 1991). In the case of figure 4.1.01, in which the sampling frequency was 250 Hz, the 
secondary velocity fluctuation had a frequency of approximately 60 Hz. Compared with the 
primary velocity fluctuations at approximately 4 Hz, the energy density (Ed) in the secondary 
velocity fluctuation, a negligible fluctuation, was: 
E d 1 f = 60 
E d I f = 4 
1 % 
Other error sources such as anemometer time response and instrumentation accuracy are usually 
well over 1 percent of error range and should be greater concerns than the secondary high 
frequency velocity fluctuations. 
Since the air jet in this study has a discharge velocity of 282 fpm (from Eqn. 4.2), which is less 
than 350 fpm and has a much smaller jet momentum than that in figure 4.1.01, the velocity 
fluctuation frequency should be smaller than 4 Hz. Therefore, a sampling frequency of 
fs > 2 f v = 8 (Hz) (4.6) 
should be sufficiently high for this application. Thus, an anemometer with a time response of less 
than 0.125s is required, i.e., the anemometer should be capable of sampling 8 times per second. 
The BERL thermister type anemometer is capable of measuring velocity fluctuations at 20 Hz 
sampling frequency. This thermister omni-directional anemometer was compared with two 
commercially available anemometers: a hot wire anemometer (TSI, model IFA 100) and an 
omni-directional anemometer (TSI, model 8470-13E-V). 
Comparison results are shown in figures 4.1.06 and 4.1.07. The thermister and the TSI hot wire 
anemometer yielded approximately the same results (figures 4.1.07 to 4.1.09), but the TSI omni- 
directional anemometer could only respond at very low frequencies. From figure 4.1.07, the 
thermister anemometer responds to the velocity within 0.5 percent of error what the TSI hot wire 
anemometer does when the air velocity is higher than 20 fpm. In figures 4.1.08 and 4.1.09, the 
thermister anemometer gives a very similar frequency response to that of the TSI hot wire 
anemometer. These data show that the thermister omni-directional anemometer is suitable for 
this application. 
