--
Fluid Modeling Facility
U S ENVIRONMENTAL PROTECTION AUtNCY
RESEARCH TRIANGLE PARK. NORTH CAROLINA 27 7
-------
Smoke visualization study shows rooftop exhaust being
trapped in airflow wake behind cubical model.
Plume from short stack collects in airflow wake behind
model.
Plume from tall stack having low exit I'elocity produces
downwash behindstack.
-------
6OOR770O4
o
Introduction
Plume from tall stack having sufficient exit velocity remains
aloft to disperse smoke.
Pollutants from a variety of sources are constantly
emitted into the atmosphere where meteorological
forces transport, diffuse, or otherwise affect their
concentrations. Because the I'.S. Environmental Pro-
lection Vgencv is charged bv Congress with establishing
and enforcing air pollution control standards to protect
the public health and welfare, it conducts research pro-
grams to describe and predict the el lects ol atmospheric
phenomena on emissions.
Measurement of pollutant concentrations at a
specific site is relative!) simple—instruments merely
col led samples at \arious locations. Predict ion of pollu-
tant levels, however, requires knowledge ol the charac-
teristics and emission levels ol the pollutant and the
atmospheric characteristics that influence pollutant
dispersal.
An effective method of characterizing atmospheric
diffusion inv olv es placing a carefully constructed model
of a pollutant source, such as an industrial plant, in a
chamber where, using wind or water, an accurate repre-
sentation of the atmosphere can be reproduced. Exami-
nation ol the ellects o! these artificial atmospheres on
model pollutant emissions provides researchers with a
greater understanding of the interaction of meteoro-
logical factors and air pollution.
To carry out this Iv pe of research, EPA's Meteorology
and \ssessment Div ision has established a Fluid Model-
ing Facility, w hie h features sev eral u ind tunnels, a water
channel-lowing lank, and support facilities.
-------
Wind in the atmosphere is a highly complex, con-
stantly changing phenomenon that—along with stack
height, surface terrain, and other factors—affects the
diffusion of pollutants in the atmospheric boundary
layer. This layer is that region of the atmosphere close
to the surface of the earth (600 meters) in which meteor-
ological factors and surface topography influence the
flow. For the casual observer, an easy method of visually
relating pollutant emissions to meteorological effects is
through the examination of the exhaust plume from an
industrial smokestack. Harmful pollutants, which may
be present in the plume, can be diluted to safe levels by
mixing with the air as it moves away from the stack or
can rapidly drop to the ground and adversely affect the
health of residents near the source.
Both wind speed and direction change with time—
from one instant to the next, from one hour to the next,
and from one day to the next. Meteorologists have found
that wind effects can be separated roughly into two
scales of motion: large-scale motions that last an hour
or more (weather), and small-scale motions that last less
than an hour (turbulence).
These two types of motion can often be seen by study-
ing a smokestack plume for a period of time. At any
given instant, the centerline of the plume will normally
form a reasonably straight line after its initial bending
over at the stack top and in the absence of obstruction
to its motion. This straight centerline indicates a reason-
ably constant wind direction. Spreading of the plume—
caused by smaller scales of motion called gusts, eddies,
or, more generally, turbulence—occurs as the plume
travels downwind. It is these smaller scales of motion,
with mean wind speed and direction remaining constant
for approximately one hour, that can be simulated in a
wind tunnel or a water channel.
The rate of dilution in the plume, however, can vary
drastically from one day to the next even if wind speed
and direction remain constant. This variability is re-
lated to atmospheric stability coupled with solar heating
METHODS FOR PREDICTING POLLUTANT FLOW
AROUND STRUCTURES*
MATHEMATICAL
MODELS
FLUID
MODELS
FIELD
PROGRAMS
ACCURACY
GOOD
BETTER
BEST
RESOLUTION
TIME
COST
COARSE
(20 m)
SHORT
(2 TO 4 weeks)
MODERATE
($25,000)
INTERMEDIATE
(3m)
SHORT
(2 TO 4 weeks)
LOW
($10,000)
FINE
(1 m)
LONG
(I year)
HIGH
($100,000)
'NUMBERS REPRESENT ONL Y ROUGH ESTIMA TES (COMPARA-
TIVE ORDERS OF MAGNITUDE) FOR SPECIAL CIRCUMSTAN-
CES; THEY MA Y NOT BE USED FOR INTERPOLATION, EXTRA-
POLATION, OR IN ANY OTHER WAY AS A BASIS FOR
ESTIMATION.
-------
r-S-k ,
Modeling Atmospheric Diffusion
or radiative cooling of the ground and results in the fan-
ning, coning, fumigating, lofting, or looping of the
plume. Because local topography also influences turbu-
lence, plume behavior for a given specific stability, wind
speed, and wind direction may still be drastically dif-
ferent if a tall building or hill is located in the vicinity
of the stack.
EPA's primary interest in the modeling of atmos-
pheric diffusion processes is an outgrowth of the estab-
lishment of National Ambient Air Quality Standards,
which are the maximum levels of a given pollutant that
are permitted in the ambient air. Control techniques,
which are applied to sources of pollution, and control
strategies, which determine the necessity for controlling
respective sources, are used to assure that pollutants
emitted into the atmosphere do not exceed these maxi-
mum levels. The dispersal of pollutants to and at ground
level, however, depends on atmospheric diffusion and
transport. To assure that ground-level concentrations
are kept within the standards, three methods are avail-
able to predict the likelihood of exceeding an air quality
standard at a particular location:
• Mathematical models can be used to evaluate alter-
native control strategies, but they require gross
simplifications. These models are not exact because
the fundamental fluid dynamics processes involved
in the dispersal are not sufficiently understood and
because computer memories are still far too small to
keep track of the detailed eddy motions that occur in
the atmosphere. Moreover, present mathematical
models are not yet adequate for calculating concen-
trations of contaminants when the plume is strongly
affected by obstructions.
• Field programs apparently provide the most reliable
results but are very expensive and time consuming.
Because meteorological conditions are not control-
lable, study periods in excess of a year must be spent
in the field to obtain a proper range of conditions,
and even then specific sets of conditions may not
occur. Furthermore, it is impossible to investigate
the impact of potential changes in alternative control
strategies for a source by means of field programs.
• Fluid models appear to work best where mathemati-
cal models fail, that is, where obstructions such as
buildings and hills block wind flow. Fluid models
also show great promise for simulating surface-
induced airflows such as heat island circulation and
mountain valley winds. Atmospheric conditions may
be programmed into a fluid model so that years of
field time are reduced to a few weeks. Fluid model
studies can reduce the resources required for field
studies and facilitate the development of better
mathematical models.
A complete research program includes comparison
and feedback among the three methods in order to gain
a deeper understanding of the processes associated with
atmospheric transport and diffusion of pollutants.
-------
-------
Components of Facility
The Fluid Modeling Facility, one of only a few atmos-
pheric dispersion modeling facilities in the world, is
available to all EPA organizations, to other Federal
agencies, and to state air pollution control agencies.
A water channel-towing tank, a meteorological wind
tunnel, and an instrument calibration wind tunnel
represent the basic components of the facility. In addi-
tion, the facility includes a model shop, an electronics
shop, a darkroom, and a chemical laboratory. A mini-
computer, including an analog to digital converter,
magnetic disk and tape drives, electrostatic printer-
plotter, and a CRT display unit, is available for real-time
data acquisition and analysis. Flow rates and
concentrations are measured by various electronic,
chemical, and mechanical equipment. The staff includes
professionals trained in environmental fluid dynamics,
model makers, computer programmers, and laboratory
and electronic technicians.
-------
38 m-
SOUND-ATTENUATING
x-ENCLOSURE
CEILING
[CONTRACTION/" HOIST
VARIABLE
-HEIGHT CEILING
FAN. ,-DIFFUSER EXHAUST
FLOWSTRAIGHTENER
AND SCREENS
TEST
SECTION
MNLET NNLET
SILENCER TRANSITION
-------
SPECIFICATIONS
DIMENSIONS
OVERALL LENGTH
TEST SECTION
LENGTH
TEST SECTION
WIDTH
TEST SECTION
HEIGHT
CONTRACTION
RATIO
TOTAL POWER
TYPE OF POWER
SPEED CONTROL
METEOROLOGICAL
WIND
TUNNEL
38 m(125 ft)
18.3 m (60 ft)
3.7 m (12ft)
2.1 m (7 ft)
2.8:1
75 kW (100 hp)
1.8-m (72-in.)
AXIAL FAN
AC MOTOR
WITH EDDY
CURRENT COUPLER
APTI
WIND
TUNNEL
11 m (35 ft)
3 m (10 ft)
1 m (3 ft)
1 m (3 ft)
4.5:1
20 kW (25 hp)
1.2-m (48-in.)
AXIAL FAN
DC MOTOR
WITH SCR
CONTROL
I Meteorological Wind Tunnel
A meteorological wind tunnel differs in two basic
respects from an aeronautical wind tunnel. First be-
cause the top of the atmospheric boundary layer is
usually much higher than the buildings immersed in it,
the simulated boundary layer in the meteorological
wind tunnel must be quite deep in order for the model
buildings to be of reasonable size. In tests in an aero-
nautical wind tunnel, on the other hand, great care is
taken to minimize the depth of the boundary layer.
Second, high wind speeds are generated in aeronautical
wind tunnels to compensate for the reduced size of the
models. In meteorological wind tunnels, however, wind
speeds generally are reduced so that buoyancy effects,
which are very important in atmospheric flows, can be
reproduced.
In operation, models are placed on a turntable that
can be rotated to simulate different wind directions,
and smoke is released from model stacks for flow visu-
alization studies. Air is drawn into the tunnel through
a flow-straightening honeycomb, and "vorticity
generators" trip the flow at the entrance to the test
section to create a thick boundary layer, which simulates
that of the atmosphere. If quantitative concentration
measurements are required in the study, hydrocarbon
gas is used as a tracer in the stack gas, and samples are
taken at various locations in the test chamber. The air
is exhausted back into the room.
The ceiling of the test section is adjustable to com-
pensate for blockage effects of the model. Acoustic
silencers minimize the noise from the fan. An instru-
ment carriage provides for three-dimensional position-
ing of measuring probes anywhere in the test section by
remote control and with readout to within ± 1 milli-
SPEED
0.5 TO 10 m/sec
(1.5 TO 30 ft/sec)
0.3 TO 21 m/sec
(1 TO 70 ft/sec)
meter.
A smaller wind tunnel, which belongs to the Air Pol-
lution Training Institute (APTI), is available for the
calibration and response testing of wind-measuring
instruments and smaller scale studies.
-------
SPECIFICA TIONS
DIMENSIONS
Ot'ERALL LENGTH 35 m (114 ft)
TEST SECTION LENGTH 25 m (8.3 ft)
TEST SECTION WIDTH 2.4 m (8 ft)
TEST SECTION HEIGHT 1.2 m (4 ft)
WATER CHANNEL DRUE
POWER
TYPE OF DRI1 E
75 kW (100 hp)
1.5 m (60-in.) AXIAL IMPELLER
SPEED CONTROL AC MOTOR WITH EDDY
CURRENT COUPLER
SPEED RANGE 0.1 to 1 m sec (0.3 to 3 ft sec)
TOWING CARRIAGE
POWER
TYPE OF DRIl E
3.7 kW (5 hp)
CABLE
SPEED CONTROL AC MOTOR WITH EDDY
CURRENT COUPLER
SPEED RANGE 1 to 50 cm/sec (0.03 to 1.6 ft/sec)
STRATIFICATION CAPABILITY
ARBITRARY STABLE DENSITY PROFILE
SHAPES WITH SPECIFIC GRAIITY FROM 1.0
TO 1.2 BY USING SALT WATER
-------
,-ft-u/^l
Water Channel-Towing Tank
FRESH TANK
WATER TRUCK
DRAIN DRAIN
The water channel-towing tank was installed in the
Fluid Modeling Facility to make possible the study of
dispersion under stably stratified atmospheric condi-
tions. The dual-purpose unit is of closed-circuit
design, with a pump in the return leg on the bottom and
the test section (free surface) on the top. The test section
is constructed of acrylic plastic in an aluminum
framework.
In the water channel mode of operation, the pump
recirculates water through the test section, and the
facility is used in a manner similar to that of the wind
tunnel. Models are fastened to the floor of the test
section; dyes are used for flow visualization studies and
for quantitative concentration determinations. The
channel is supported on jacks that can be adjusted
to tilt the entire unit to compensate for the pressure
drop through the test section.
In the towing tank mode of operation, the ends of the
test section are blocked with gates, and the test section
is filled layer by layer with salt water, each layer of dif-
ferent density. Atmospheric density gradients are
modeled by the density gradients of the salt water.
Models are attached to a turntable that is suspended
from a towing carriage into the water, and towed the
length of the test section, making possible the study of
flow and dispersion around buildings and complex
terrain under stably stratified atmospheric conditions.
A filling system comprised of a brinemaker, five large
tanks, and numerous pumps and valves provides the
capability of filling the test section with a desired stably
stratified salt-water mixture in approximately four
hours. Any type of stable-stratification from elevated-
or ground-based inversions to neutral conditions may be
simulated in the towing tank mode of operation.
-------
0
DC
CD
OC
CONCENTRATION
o
GC
CD
O
GC
CONCENTRATION-
-------
Modeling in Action
An old "rule of thumb" says that a stack placed next
to a building must be at least 2'/2 times the height of the
building to avoid down wash of the plume in the wake of
the building. Downwash would result in high concentra-
tions of pollutants at ground level. A wind tunnel study
showed this to be a good rule for a conventionally
shaped building. For a tall, thin building, however, the
rule was demonstrated to be unnecessarily conservative
and, therefore, wasteful. The photographs and related
concentration profiles show plume behavior from
model buildings and exhaust stacks in the wind tunnel.
Comparisons of the illustrations on the facing page
show that a thin building has essentially no effect on
plume behavior when the stack is l'/2 times the height
of the building. The illustrations on this page, however,
show that downwash occurs behind a wide building
when the stack height is only 1'2 times the building
height.
This study, then, benefitted the consumer by demon-
strating that the construction of costly tall stacks is
not always necessary.
CONCENTRATION-
-------
oottop emissions in wind tunne
jle stratification in towing tank
leulral stratification in towing tank
Highway vehicle study in wind tunnel
Building downwash in wind tunnel
-------
Recent Publications
Thompson, R.S. and D.J. Lombardi. 1976. Dispersion of Roof-Top
Emissions from Isolated Buildings-A Wind Tunnel Study. U.S.
Environmental Protection Agency, Research Triangle Park, N.C.
(in review).
Snyder, W.H., R.S. Thompson, and R.E. Lawson, Jr. 1976. The EPA
Meteorological Wind Tunnel: Design, Construction, and
Operating Details. U.S. Environmental Protection Agency,
Research Triangle Park, N.C. (in preparation).
Huber, A.H. and W.H. Snyder. 1976. Building Wake Effects on Short
Stack Effluents. Presented at Third Symposium on Atmospheric
Turbulence, Diffusion, and Air Quality, Raleigh, N.C. October.
Snyder, W.H. and R.E. Lawson, Jr. 1976. Determination of a Necessary
Height for a Stack Close to a Building-A Wind Tunnel Study.
Atmos. Env. (in press).
Thompson, R.S. and W.H. Snyder. 1976. EPA Fluid Modeling Facility.
Presented at EPA Conference on Modeling and Simulation,
Cincinnati, Ohio. April.
Huber, A.H., W.H. Snyder, R.S. Thompson, and R.E. Lawson, Jr. 1976.
Plume Behavior in the Lee of a Mountain Ridge-A Wind Tunnel
Study. Presented at EPA Conference on Modeling and Simulation,
Cincinnati, Ohio. April.
Snyder, W.H. 1974. Fluid Modeling Program of the Meteorology
Laboratory, U.S. Environmental Protection Agency. In: Air
Pollution: Proceed ings of the Fifth Meeting of the Expert Panel on
Air Pollution Modeling. NATO Committee on the Challenges of
Modern Society, p. 31-1 to 31-47.
Snyder, W.H. 1972. Similarity Criteria for the Application of Fluid
Models to the Study of Air Pollution Meteorology. Boundary Layer
Meteor. 3(2): 113-134.
-------
ENVIRONMENTAL
RESEARCH
CENTER
For further information on the Fluid Modeling Pro-
gram, contact:
Chief, Fluid Modeling Section
Atmospheric Modeling and Assessment
Branch
Meteorology and Assessment Division
Environmental Sciences Research Laboratory
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
------- |