EPA-600/4-76-047
September 1976
Environmental Monitoring Series
Laberatitrf
Office of and DeveBopmenf
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL MONITORING series.
This series describes research conducted to develop new or improved methods
and instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations. It also includes
studies to determine the ambient concentrations of pollutants in the environment
and/or the variance of pollutants as a function of time or meteorological factors.
This document is available to the public through the National Technical Informa-
tion Service. Springfield, Virginia 22161.
-------
EPA-600/4-76-047
September 1976
STACK PLACEMENT IN THE LEE OF A MOUNTAIN RIDGE
A Wind Tunnel Study
Alan H. Huber
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
Research Triangle Park, N.C. 27711
William H. Snyder
and
Roger S. Thompson
Meteorology and Assessment Division
Environmental Sciences Research Laboratory
Research Triangle Park, N.C. 27711
Robert E. Lawson, Jr.
Northrop Services, Inc.
Research Triangle Park, N.C. 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, N.C. 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica
tion. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
-------
ABSTRACT
In a wind tunnel study, the highly turbulent region in the lee of a
two-dimensional mountain ridge was found to consist of a large semipermanent
eddy that was caused by the separation of the main flow at the apex of the
ridge. General circulation of this eddy was in the main flow direction
alone, the upper edge, opposite the main flow direction along the ground
surface, and up the slope along the leeward ridge surface. Smoke visualiza-
tion arid hot film anemometry measurements showed that the cavity size and
shape were minimally affected by the thickness and turbulence intensity of
the approach boundary layer flow. In addition, the cavity size and shape
were not found to be affected by the detailed shape of the ridge, but were
strongly dependent upon the upwind terrain and the gross topographic features
(angles) of the downslope. The largest cavity was found to extend to two
ridge heights in the vertical and to ten ridge heights downwind.
A model stack was positioned to emit an air-methane mixture into the
cavity in the lee of the ridge. Longitudinal, lateral, and vertical concen-
tration profiles were taken. A tall stack placed to emit the test mixture
into the upper portion of the cavity produced higher ground level concentra-
tions near the downwind end of the cavity than did a shorter stack. The
maximum measured concentrations, however, were found to occur near the bases
of the shorter stacks.
The cavity region leeward of the model ridge was found to be highly
turbulent with significant plume downwash. For similar real-atmosphere
situations, it would be good engineering practice to avoid placement of any
significant source within the expected cavity region. Application of the
"2 1/2 times rule" (with respect to the ridge height) for stack construction
would be sufficient to avoid the highly turbulent region of the cavity
proper.
m
-------
CONTENTS
Abstract iii
Figures vi
Symbols ix
1. Introduction 1
2. Conclusions and Recommendations 3
3. Review of Literature 4
4. Design of the Experiment 9
5. Experimental Details and Equipment 11
6. Results 17
6.1 Phase I: Envelope and Cavity Visualization Results for
the 30 cm Gaussian Ridge and 30 cm Triangular Ridge 17
6.2 Phase I: Mean Velocity and Turbulence Intensity
Measurements for the 30 cm Gaussian Ridge and 30 cm
Triangular Ridge 17
6.3 Phase II: Concentration Measurements for a Stack
Placed in the Lee of the 30 cm Gaussian Ridge 21
6.4 Phase III: Cavity and Envelope Measurements in the
Lee of the 15 cm Gaussian Ridge 35
7. Discussion and Summary 39
References 42
-------
FIGURES
Number Page
1 Diagrammatic sketch of envelope and cavity regions bellind
a two-dimensional ridge ...... 2
2 Cavities behind obstacles (boundary layer thicknesses and
cavity dimensions are given as multiples of the obstacle
or step height) 7
3 Diagrammatic sketch of wind tunnel with vortex generators
and roughness slats in place 12
4 Mean velocity and turbulence intensity profiles 5.5 m
downstream from vortex generator trailing edge 12
5 End views of two-dimensional ridges .... 13
6 Phase III ridge configurations 16
7 Envelope size in lee of 30.5-cm Gaussian ridge 18
8 Cavity size in lee of 30.5-cm Gaussian ridge 18
9 Envelope size in lee of 30.5-cm triangular ridge .... 19
10 Cavity size in lee of 30.5-cm triangular ridge 19
11 Smoke envelope in lee of (A) 30-cm Gaussian ridge and (B) 30-
cm triangular ridge (The strut supporting the smoke tube is
laterally displaced from the smoke filled region.) 20
12 Mean velocity and turbulence intensity profiles for 30.5-cm
Gaussian ridge (separation not forced), U^ = 3.05 m/sec 22
13 Mean velocity and turbulence intensity profiles for 30.5-cm
Gaussian ridge (separation not forced), U^ = 9.15 m/sec 22
14 Mean velocity and turbulence intensity profiles for the 30.5-
cm Gaussian ridge, U^ = 3.05 m/sec 23
15 Mean velocity and turbulence intensity profiles for the 30.5-
cm Gaussian ridge, U =9.15 m/sec 23
-------
Ib Mean velocity and turbulence intensity profiles for the 30.5-
cm triangular ridge, U^ = 3.05 m/sec 24
17 Hean velocity and turbulence intensity profiles for the 30.5-
cni triangular ridge, U^ = 9.15 m/sec 24
Id Ground level concentration measurements with sampling probe
fixed near oase (x/H - 3 from ridge center, z/H = 0) of
30.5-crn Gaussian ridge (Stack was placed at four downwind
locations and its heigut was varied 25
19 Longitudinal ground level concentration profiles downwind
from stack placed at base (x/H = 2.7 from center) of 30.5-
cm Gaussian ridge 26
20 Vertical concentration profiles from stack placed at base
(x/H = 2.7 from center) of 30.5-cm Gaussian ridge,
HS/H - 0.5 28
21 Smoke visualization with 0.5 H stack placed at base of 30-cm
Gaussian ridge (x/H = 2.7 from ridge center) 29
22 Lateral concentration profiles from stack placed at base
(x/H = 2.7 from center) of 30.5-cm Gaussian ridge,
x/H - 1.25, HS/H = 0.5 30
23 Lateral concentration profile from stack placed at base
(x/H = 2.7 from center) of 30.5-cm Gaussian ridge,
x/H = 2.25, H$/H = 0.5 30
24 Lateral concentration profiles from stack placed at base
(x/H = 2.7 from center) of 30.5-crn Gaussian ridge,
x/h = 4.75, H$/h = 0.5 31
25 Elevated longitudinal concentration profile downwind from
stack placed at base (x/H = 2.7 from center) of 30.5-crn
Gaussian ridge, H /H = 0.5 31
26 Vertical concentration profiles from stack placed at base
(x/H = 2.7 from center) of 30.5-cm Gaussian ridge
HS/H = 1.5 32
27 Smoke visualization with 1.5 H stack placed at base of 30-cm
Gaussian ridge (x/H =2.7 from ridge center) 33
28 Lateral concentration profile from stack placed at base
(x/H = 2.7 from center) of 30.5-cm Gaussian ridge,
x/H = 1.25, Hs/H =1.5 34
29 Lateral ground level concentration profiles from stack
placed at base (x/H = 2.7 from center) of 30.5-cm
Gaussian ridge, z/H = 0, HS/H =1.5 34
vii
-------
30 Envelope and cavity size in lee of 15.2-cm Gaussian ridge
(60-cin simulated boundary layer) 35
31 Mean velocity profiles above center of 15.2-cm Gaussian
ridge . , 36
32 Envelope and cavity in lee of 15.2-cm Gaussian ridge
(15-cm natural boundary layer) 37
33 Envelope and cavity size in lee of 15.2-cm Gaussian
ridge (15-cm natural boundary layer with raised plane
approach) 37
34 Envelope and cavity size in lee of 15.2-cm Gaussian
ridge (60-cm simulated boundary layer with raised
plane approach) 38
35 Envelope and cavity size in lee of 15° sloping plane
(60-cm simulated boundary layer with raised plane
approach) ..... 38
VI
-------
SYMBOLS
L characteristic width of ridge
H height of obstruction (ridge, etc.)
H height of stack
D ' inside diameter of stack
<5 boundary layer thickness
U mean wind speed (a function of elevation)
Us wind speed at top of stack
U^ free-stream wind speed
/=T
/u1 root-mean-square of longitudinal velocity fluctuations
W stack effluent speed
v kinematic viscosity of air
Re Reynolds number
Q stack emission rate
C stack gas concentration
x concentration (a function of location)
C concentration expressed as the percent of stack gas concentration
M used as subscript to denote measurements under a model scale
situation
F used as subscript to denote measurements under full scale field
situation
R radius of curvature
x longitudinal distance (origin as specified)
y lateral distance from center of tunnel
z vertical distance from floor of tunnel
IX
-------
SECTION 1
INTRODUCTION
Aerodynamic effects induced by local terrain features can have a major
influence upon the dispersion of locally emitted effluents. Even when the
uest demonstrated control technology is applied, pollutant concentrations are
usually at levels far in excess of ambient air quality standards as tiie plume
leaves the stack. In addition, the plume may frequently be entrained in the
turbulent ecidies created by air flow over local terrain features and be
urought to the ground uefore the pollutant concentrations are reduced to
levels below ambient air quality standards. Thus, it is important that those
persons responsible for stack design and location have an adequate under-
standing of such adverse aerodynamic effects.
For neutral (neutrally stable) atmospheric flow, aerodynamic effects
evolve from interacting frictional forces and pressure gradients induced by
local surface roughness and terrain features. Adverse effects exist when
surface friction and pressure gradients combine to retard the surface layer
flow enough to produce separation of the boundary layer. Separation in a
neutral flow generally occurs near tiie apex of the terrain feature -- result-
ing in a stagnation region on the leeward side, often referred to as a "cav-
ity". At the point of separation, the main stream of flow is vertically
raised, causing a stagnation region (cavity) to develop below. Mean veloci-
ties in this cavity are reduced and the flow is highly turbulent. The flow
reattaches itself somewhere downstream of tiie obstacle. In the region of
reattachment, a portion of the flow is deflected upstream, forming a zone of
recirculation. The dividing "streamline" that separates the recirculating
flow from the main stream encloses tiie cavity, as shown in Figure 1. The
wake, defined as that region of the flow field tiiat is disturbed by the
obstacle, can extend very far downwind. (The "envelope" is the upper bound-
ary of the wake.) Far enough downwind, of course, the flow readjusts itself
to a uoundary layer appropriate to local surface roughness.
For a moderately stable atmospheric flow, the separation point generally
shifts to a position on the leeward side, which results in a smaller cavity.
For highly stable flows, separation may disappear entirely, or other phenom-
ena, such as lee waves and rotors, may be found. These phenomena also have
the potential to affect plume dispersion. Unstable flows, on the other hand,
can either enhance or delay separation.
Terrain features that most adversely affect the flow are two-dimensional
in nature. Lateral air motion around a hill results in a smaller cavity size
than would be observed for a two-dimensional ridge. A study of neutral flow
with separation occurring at the apex of a two-dimensional mountain ridge
would, therefore, best demonstrate the extent to which stagnation regions can
1
-------
influence trie dispersion of locally emitted effluents.
The objective of this study, conducted at the U.S. Environmental Protec-
tion Myency's Fluid Modeling Facility, was to simulate neutral atmospheric
flow over a two-dimensional mountain ridge and to examine the effects of the
ridge on stack plume dispersal and resultant ground level pollutant concen-
trations. Tne ultimate goal was to develop guidance for determining "good
engineering practice" when selecting the site for a stack.
Tnere were three major phases to this study. In the first phase, veloci-
ty and cavity size measurements were made in order to demonstrate that the
oasic flow structure is independent of the detailed shape of the ridge (Gaus-
sian or triangular). In the second phase, the concentration field resulting
from a source placed within the cavity was delineated. In the third phase,
the effect of variations in the approach flow on the cavity size and shape
was examined. The results presented in this report should enable the reader
to better understand the atmospheric flow structure so that good engineering
practice can be used in locating stacks or determining necessary stack
neights, and that "order of magnitude" ground level concentrations can be
predicted in the event the stack emits into the cavity region.
Figure 1. Diagrammatic sketch of envelope and cavity regions behind a two-dimensional ridge.
-------
SECTION 2
CONCLUSION AND RECOMMENDATIONS
The cavity region leeward of the model ridge was found to be highly tur-
bulent with significant plume downwash. For similar real-atmosphere situa-
tions, it would be good engineering practice to avoid placing any significant
source within the expected cavity region. Application of the "2 1/2 times
rule" with respect to the ridge height would be sufficient to avoid the
highly turbulent region of the cavity proper. Although the maximum depth of
the cavity was found to be 2 H, some margin of safety is well advised because
strong downdrafting occurs in the upper regions of the cavity. The maximum
horizontal extent of the cavity was found to be 10 H. Because the turbulence
intensity and thus the plume dispersion are increased in the envelope above
the cavity, part of a plume emitted above a cavity can, in only a short
distance, spread downward and thus become entrained within the cavity.
The concentration measurements from the stack placed in the cavity
region were helpful in providing a description of the effects that the ridge
had on stack plume dispersal. For the short stacks (0.5 H and 1 H stack
heigrits), the highest ground level concentrations were found at a position
very near the stack. For the 1.5 H stack, the highest ground level concen-
trations occurred significantly further downwind of the stack. From these
results, it is evident that instead of the strong immediate downwash, which
occurred for the shorter stack, the taller stack emissions are caught in the
outer recirculatiori region within the cavity. Maximum ground level concen-
trations, which did occur within the cavity region, were found to be of the
order 0.05 - 0.1 percent of the stack effluent concentration. These concen-
trations are undoubtedly significantly higher than would occur in the absence
of the mountain lee effects examined in this study. If further concentration
mappings were made, having a greater variety of stack locations within the
cavity and a greater variety of stack emission characteristics, a general
guideline for estimating ground level concentrations could be developed.
This additional delineation would be quite useful for those circumstances
that may arise when building a stack 2 1/2 times the ridge height would be
impractical.
Downwind of the cavity region the turbulence intensity was still signi-
ficantly greater than that found in the undisturbed flow. Further studies of
the behavior of plumes from sources placed downwind of the cavity region are
needed to characterize dispersion there.
-------
SECTION 3
LITERATURE REVIEW
The general engineering "rule of thumb" for avoiding plume downwash in
the lee of a ridge is to keep the height of the stack 2 1/2 times the height
of the ridge. According to Sutton (1960), the rule was probably derived by
Sir David Brunt from a study on the height of atmospheric disturbances over
a long ridge in connection with a British airship disaster investigation.
The downwind range of the strong downwash region (cavity) depends on the
general form of the obstruction being considered. Flow around a cube, for
example, was found to form a cavity extending 2 barrier heights downwind
(Evans, 1957). When the width perpendicular to the wind direction was in-
creased to 8 heights, the cavity length increased to 5 heights downwind, but
the vertical extent of the cavity region increased only slightly. The greatest
extent of the cavity region may, therefore, be expected to occur in the lee
of the two-dimensional ridge.
The ASME Guide (1968) reported that the plume from a stack placed in the
cavity leeward of a valley ridge became thoroughly diffused before passing
downwind to the wake region where the flow was in the direction of the upper
wind. The plume was distributed vertically throughout much of the valley
depth. Briggs (1969) reported that a terrain feature can affect a plume in
much the same manner as a building, which causes high concentrations of the
effluent to be brought to the ground. A brief discussion of effects induced
by terrain was presented by Strom (1968). The turbulent region on the leeward
side was described as a region in which the plume can be brought quickly to
the ground, diffused over a wide area, or, in extreme cases, transported up
the hill because of a large standing eddy. This turbulent region was reported
to extend to elevations well above the hill top, depending on the abruptness
of the windward side.
A detailed discussion of separation and wake formation was presented in
Halitsky (1968). Most of the experimental results reported, however, related
to three dimensional obstructions. Interestingly, he raised the point that:
"no wind-tunnel measurements of concentrations in plumes that have descended
into the wake appear in the literature, owing partly to the limited number
of facilities available for this type of test and partly to the policy of
industrial sponsors not to release such information." The interest in appli-
cation of fluid models to the study of air pollution has since increased, and
therefore the number of facilities and reports published in the literature
has been increasing.
According to Scorer (1955), the separation point is stationary when there
is a salient edge at the top of a hill or ridge, and he cites numerous, but
-------
observed.
f\ \n 4- f « !~ r~ *-
I (j I OCA 31 ICC L.
a wake zone in
ft _. .il
T I UVV
zinc sulfide
during which the
period of exper-
most common
wi lii the airstream
the lee of the ridge.
The
limited, field studies relating to the zone of recirculation and instances of
intense mixing and general downdrafting in the leeward regions of ridges.
Details are insufficient to draw firm conclusions relating to formation of
separated flows.
The general features of flow over a ridge were treated theoretically and
experimentally by Buettner (1964). A ridge station was constructed on the
lee side of the Ipsut Pass area of Mount Rainier National Park in Washington
as part of a study of the effect of terrain obstacles on the fallout of par-
ticulate matter through the atmosphere. Tracer particles of
were released and collected. Data were collected for 5 days
airflow approach was perpendicular to the ridge. During the
imental set-up, only light-moderate winds were
wind field occurrence was reported as a "
separating from the ridge top and forming
For this flow, the wind field was constant above and zero below a plane rep-
resenting the wake zone. Only a small amount of particulate penetrated down
through the horizontal vortex sheet. A contaminant released in the calm zone
was reported to meander in an unpredictable manner, previously, a lee eddy
with the main airstream moving first horizontally away from the ridge, then
down, and then up again close to the valley bottom was visually observed. At
this site, such a flow pattern was believed to exist only for strong winds.
Laminar flow complicated by the1?1 v.'ir.dc v^c reported to occu," wi.cn stable
settled conditions prevail and the gradient wind at ridge level was less than
6 knots (3.1 meters per second).
Some characteristics of wind field modification by natural obstructions
were reported by Gloyne (1965). An eddy flow 2 barrier heights in vertical
extent and 10 to 15 barrier heights in horizontal extent to the leeward side
of a "near solid" barrier was diagrammed. At ground level, the region of
disturbed flow extended to about 30 barrier heights. Downwind of a steeply
sloped, wooded hill with a wind blowing at right angles to its length, the
disturbed flow was repcrti-! tc alow cAlcnd UUWMNMIU iu douuc ju times its
height. Additional discussions relevant to wind modifications were also pre-
sented, and the point was made that each case must be assessed separately.
Slope angle and thermal stability and wind speed were influential factors in
determining the extent of terrain-induced disturbance
Pooler and Niemeyer (1970) presented, as part of a study evaluating
dispersion from tall stacks, several situations in which unexpectedly high
ground level concentrations could be associated with mountain lee effects.
On days with neutral flow, the plume fi OHI a suacK lucatea u riage heights
downwind from a 450-m ridge was carried down to ground level within a very
short distance. This phenomenon could well be a result of the strong down-
wash that occurs near the leeward edge of a standing eddy.
Corby (1954) reviewed the work of J. Forchtgott, who gathered about
35 different sets of observations involving five different mountain ridges
located in Bohemia. Mountain airflow was classified into four main types:
(1.) undisturbed streaming, (2.) standing eddy streaming, (3.) wave stream-
ing, and (4.) rotor streaming. The case of standing eddy streaming
-------
corresponded to the situation of boundary layer separation at the ridge apex
with cavity formation in the lee. This type of flow was reported to have been
observed frequently. Forchtgott implied that this situation was predominant
under moderate wind speed and wind shear conditions., Even for the cases with
smooth waves above, some form of turbulent wake was found in the lee of the
ridge.
A World Meteorological Organization technical note (1967) arrived at a
similar conclusion. Over rugged terrain, whether the flow aloft was smooth
or otherwise, it usually rested on a turbulent wake. Although little
descriptive detail of such regions was presented in the report, many photo-
graphs showed the wave structures above the wakes, as revealed by cloud
formations.
The results of a number of balloon releases made in two valleys in
Vermont were reported by Davidson (1963). Balloon releases were made at
several positions along the sides of ridges that had approximately 20° slopes.
Balloon paths were determined using theodolites. The limited results could
not be used to confirm a point of separation or the extent of a leeward cavity
region. The extreme turbulence generated in the lee of the ridges, however,
appeared to be dissipated at most elevations at a distance of 4 to 6 heights
downwind.
Halitsky, et al. (1966) reported comparisons between their wind tunnel
model results and Davidson's (1963) field observations in the lee of Green
Peak, Vermont. Best agreement resulted for the higher model wind speeds sug-
gesting that tests of this type be run with a minimum ridge height Reynolds
number of 1 x 105. The field observations of a cavity and wake flow generally
fitted the model test results.
An extensive literature review relating to both field and fluid modeling
studies and a discussion as to how mountainous terrain can alter atmospheric
airflow were offered by Orgill, et al. (1971). The authors reported that,
for neutral airflow over a mountain, a large semipermanent eddy occurs on the
lee side. An area in the central Rocky Mountains of Colorado was chosen for
a field and laboratory study of transport and dispersion over irregular ter-
rain. Two different atmospheric conditions were simulated: the thermal sta-
bility used in the wind tunnel model was near-neutral in the lower levels and
stable in the upper levels for one case and totally neutral throughout for
the other case. Field data yielded information on the mean velocity and dis-
persion characteristics over the local terrain. Totally neutral atmospheric
stability conditions were observed on only one day. No specific information
as to where and when boundary layer separation occurs or the size or shape
of the cavity region in the lee of ridges was reported in either the field
or laboratory study results. The purpose of the report was to generalize on
flow patterns in complex terrain on a much larger scaleo
Some discussion of fluid modeling studies on the flow behind regular-
shaped obstacles is relevant. Bradshaw and Hong (1972) compared the cavity
sizes found by several investigators. This comparison plus some recent ad-
ditions are shown in Figure 2. In the figure, the cavity size and boundary
layer depth are given in terms of the obstacle height, H.
6
-------
J
1.9-5.6
A.
?LARGE
1.7
-17-
-270-
PETRYK AND BRUNDRETT, 1967
-12.5-
IMAGABHUSHANAIAH, 1962
C.
3.3
-290-
D.
Re =10.000
1.7
I I-
:A:«, ET AL, 1974
3.5-8
PLATE AND LIN, 1964
MUELLER AND ROBERTSON, 1963
0.7
\- -- 6.8 -4
TANI, ETAL., 1961
BRADSHAW AND WONG, 1972
Figure 2. Cavities behind obstacles (boundary layer thickness and cavity dimensions are given as mul-
tiples of the obstacle or step height).
-------
Diffusion into the boundary layer downstream from a two-dimensional
sinusoidal ridge was examined by Plate and Sheih (1965). A continuous point
source of helium was placed at the crest of the model. The diffusion charac-
teristics were found not to differ from the case of a smooth flat plate when
the boundary layer did not separate from the model. Both plume height and
width were found to increase significantly in the immediate downwind neighbor-
hood of the model when separation did occur. Further downwind, however, the
rate of growth naturally decreased and became similar to that in an undisturb-
ed boundary layer. The size of the cavity in the lee of this model was not
shown.
Diffusion from a ground line source upwind from a fence was examined by
Plate (1967). Downstream from the reattachment point, the concentration pro-
file was found to follow a similarity law whose shape corresponded to that in
an undisturbed boundary layer. No measurements of concentrations inside the
cavity region were included.
The review of published field studies presented here strongly supports
the assertion that, on the leeward side of a mountain ridge, a recirculating
flow region with strong downwash and dispersion characteristics can exist.
However, information that could define the point of separation and the size
and extent of the cavity was not found. The point of separation appears to
be very much a function of mean flow speed and direction, atmospheric stabil-
ity, both the downslope and upslope angle of the ridge sides, and the location
of the ridge with respect to surrounding terrain.
For a particular situation, the greatest size of the cavity will take
place when separation occurs at the ridge apex. Both field results and fluid
modeling results confirm the natural expectation that the more obtrusive the
ridge, the larger the cavity region. Obstructions with salient features
should exhibit definite separation at their edges under all atmospheric con-
ditions. The size of the cavity region is greatest for isolated ridges with
steep sloping sides. Stable atmospheric conditions act to restrict the size
and extent of the cavity region.
-------
SECTION 4
DESIGN OF EXPERIMENT
To ensure that the flow in the model accurately simulates that in the
atmosphere, It is necessary to meet certain similarity criteria. Various
nondimensional parameters that characterize the flow in the atmosphere --
including the Rossby number, the RpynnlH niimhpri th.p Froude number, the
Richardson number, etc. -- must be matched in the model. Because this study
is concerned only with neutral atmospheric flows involving nonbuoyant
effluents and relatively small scales, the Richardson, Froude and Rossby
numbers may be ignored (Snyder, 1972). The remaining parameters of signifi-
cance are as follows:
H ' H ' H ' H ' U ' U
.
The first three of these parameters (length ratios) are easily matched
by constructing a scale model , but because no particular field situation was
modeled, idealized ridge shapes and representative values were chosen. The
next three parameters characterize the boundary layer approaching the ridge.
Two different boundary layers were used. One was a thick, simulated atmos-
pheric boundary layer; the other was the natural (thin) boundary layer
developed over the smooth wind tunnel floor.
The effluent speed to wind speed ratio was maintained at 3:2 in all
tests. This value is the minimum necessary to avoid downwash in the imme-
diate lee of the stack itself (Sherlock and Stalker, 1940). Under these
conditions (minimum effluent speed to wind speed ratio and zero buoyancy),
the effective stack height is ssc^ntl^lly t.u.c cac zz the physical stack
height. These conditions were chosen in order to be able to equate the
effective stack height in an undisturbed flow to the physical stack height.
Plume rise or downwash from the model stack placed in the lee of the ridge
are therefore associated with induced disturbances.
The last two parameters are the ridge Reynolds number (Re^ = U^H/v) and
the effluent Reynolds number (Re^ = WsD/v). For exact similarity, the model
ridge Reynolds number must equal the actual ridge Reynolds number. This is
not possible for the model scale u5cc! ar.cl, foi LunaLc"!^, is not necessary
(Snyder, 1972) because at sufficiently large Reynolds numbers, the pattern
of flow becomes independent of Reynolds number. For a mountain ridge having
a salient feature, near the top of the ridge, the boundary layer may be ex-
pected to separate near its apex (Scorer, 1968, p. 108). For a ridge not
having a salient feature, the separation point may fluctuate in position,
-------
and may even occur on the downwind side of tlie ridge, which, would result in
a smaller cavity. In the field, the ridge Reynolds number based on a rfdge
height greater than 75 m and free-stream wind speed greater than 3.0 m/sec
is greater than 1 x 107. For these very large Reynolds numbers, separation
is certain to occur at or slightly windward of the apex, even for a ridge
with a smooth rounded top [Scorer, 1968, p. 113). The Reynolds number for
the model mountain ridges lies between 1 x TO*4 and 1 x 1CP, which is much
smaller than the full scale Reynolds number. At these smaller Reynolds num-
bers, the location of the separation point is affected by the Reynolds
number. If the point of separation on the model occurs at its apex, however,
similarity of the two flow patterns should result. Hence, a small trip (1.25
cm-high slat, see Figure 11 A] was placed at the apex of the rounded ridge
in order to ensure separation at that point.
The plume behavior is independent of the effluent Reynolds number pro-
vided the flow is fully turbulent at the stack exit. The value of 2000 is
well-established for the maintenance of turbulent flow in a pipe. Because
the effluent flow appeared to be laminar as it was ejected from the stack,
internally serrated washers were placed inside the stack to ensure fully
turbulent flow at the exit (Lin et al., 1974).
Concentrations measured in the model can also be related to those that
would be measured in the field. For most situations, the stack emissions
may be assumed to behave as point sources. The stack gas concentration, GS
Cgram per cubic meter), is related to the stack emission rate, Q (gram per
second), by
Q = C
D2)W
S.
CD
The field concentration, xp> ^s linearly related to the model concentration.
XM; to the emission ratio, QM/QM ; and to the dilution ratio, U^H^ / UpHjs
These basic relations result in the expression
Xp - X
M (QF/QM)
CHM/HF)'
or
Cx/Cs)F = [(x/Cs)M]
[(D/H)pV(D/H)*]
(2)
With strict matching (identical effluent speed to wind speed ratios and stack
diameter to ridge height ratios), the model concentration ratios are equal to
the field concentration ratios, (x/Cs)^ - (x/C$)p- Relaxation of the
requirements for strict matching of the stack diameter to ridge height ratios
and effluent to wind speed ratios permits a wider applicability of the con-
centration measurements, provided the plume paths are similar.
10
-------
ECTION 5
EXPERIMENTAL DETAILS AND EQUIPMENT
The experiments were carried out in the meteorological wind tunnel of
" "" Facility. The wind tunnel has a test section measuring
EPA's Fluid Modeling
3.7 m x 2.1 m x 18.3 m, and the flow speed within the wind tunnel can be
controlled between 0
can be obtained from Snyder, et al., (1976).
Jeldils of the wind tunnel
In performing fluid modeling studies, it is important to simulate the
atmospheric boundary layer. The natural wind tunnel boundary layer that
develops over the smooth luime'i floor is generally too tnin and possesses
insufficient turbulence intensity for accurate simulation. A thick, turbu-
lent boundary layer was simulated by the method of Counihan 0969), who used
a castellated barrier placed at the test section entrance slightly upstream
from semiell iptical, wedge :u.apcd vertex geni.'al^, o. Counihan used three-
dimensional blocks downwind of the generators to maintain the boundary layer.
In the present study, two-dimensional roughness slats were found to be suf-
ficient to maintain the boundary layer in equilibrium. These experiments
used a 15-cm-high barrier, 60-cm-high generators, and 7 m of 0.635 cm x 1.43
cm slats placed across the floor perpendicular to the wind direction with
centers placed 15.24 cm apart (Figure 3). The ridge was placed just down-
stream of the last roughness slat. A 60-cm-thick boundary layer with an
approximate one-seventh power law mean velocity profile was measured in a
separate study 5.5 m downstream from the vortex generator trailing edge
CFigure 4). The one-sevcnlii iJioiilt; ii> generally accepted to represent
neutral atmospheric flow over relatively smooth terrain (open country, grass-
land -- Davenport, 1963). The vortex generator-roughness element combination
described above also produces a suitable turbulence intensity profile repre-
sentative of atmospheric finw n\/P>- ypia-Hwoiy cm^n-m terrain CFigure 4).
Three model ridges were constructed and are shown in Figure 5. One
ridge was triangular in shape with a 30.5-cm-high apex and sloping sides of
30°; the other two had sides with idealized shapes. The latter two ridges
were symmetrical about a uenuerl ine, eacn side of which could be divided into
three sections. The center section had a constant slope of 30°, and the
upper and lower sections were convex and concave outward, respectively. This
model shape appears to be very close to a Gaussian probability distribution
and is referred to thro"^1""'1" <~h"rc ^ep^t ?.i the Gaussian model ridge in
order to distinquish it from the triangular model ridge. The two similar
Gaussian ridges have apexes of 30.5 cm and 15.2 cm in height, respectively.
The three models will be referred to in the remainder of this report as the
30-cm triangular ridge, the 30-cm Gaussian ridge, and the 15-cm Gaussian
ridge.
11
-------
CASTELLATED
BARRIER
VORTEX GENERATORS
CEILING ADJUSTED
ROUGHNESS SLATS
Figure 3- Diagrammatic sketch of wind tunnel with vortex generators
and roughness slats in place.
1.50
1.25
1.00
1.50
6 = 60 cm
U00=7.5m/sec
1/7th POWER LAW
0.2 0.4 0.6 0.8 1.0 0
Figure 4. Mean velocity and turbulence intensity profiles 5.5 m downstream from vortex generator
trailing edge.
-------
A. GAUSSIAN RIDGE, H = 30.5 cm
B. TRIANGULAR RIDGE. H = 30.5 cm
C. GAUSSIAN RIDGE, H <= 15.2 cm
(SHAPE SIMILAR TO A)
Figure 5. End views of two-dimensional ridges.
In Phase I of this study, the primary goal was to examine whether the
flow structure is independent of the detailed shape of the ridge. Smoke
tests were made to defin^ t^° shanac nf fhe cavities and envelopes, and hot
film measurements were made to compare the flow fields above and downwind of
the 30-cm Gaussian ridge and the 30-cm triangular ridge. The approach flow
for all tests was that of the 60-cm simulated boundary layer.
For the 30-cm Gaussian ridge, smoke visualization indicated the cavity
region to be dependent on the free-stream velocity. A 1.25-cm-high slat was
placed along the apex of the ridge in order to trip the flow and induce sepa-
ration at that point. The triangular ridge exhibited flow separation at its
apex without the addition of a trip.
Further experimentation with the tripped, 30-un Gaussian-shaped ridge
and the triangular ridge indicated that the cavity region on the leeward
side was independent of free-stream velocity [Section 6.1). This was deter-
mined through smoke visualization. At various downstream positions, a smoke
source was raised in the vertical to a point where no smoke could be seen
recirculating back to the base of the hill. These vertical positions were
used to define the boundary shape of the leeward cavity. The "envelope" was
determined from observations of the maximum vertical spread of smoke emitted
within the cavity rsgior..
For mean velocity and turbulence intensity profiles, a hot film sensor
and anemometer were used. A Thermo-Systems Inc. Model 1054 A anemometer was
used in conjuction with the comoanv's Model 1210-20 hot film probe. The
resulting measurements for both the 30-cm tripped Gaussian ridge and the
triangular ridge were quite similar [Section 6.2). Because of the similar-
ity, only the tripped Gaussian ridge was used in further analyses [Phases II
and III).
In Phase II, the goal was to map the concentration field resulting from
13
-------
a stack emitting into the cavity region. The approach flow was the 60-cm
simulated boundary layer. A stack was placed in the center of the tunnel at
the leeward base edge of the tripped, 30-cm Russian ridge (2.7 H downwind
of the ridge center). The stack inside diameter (ID) was 0.89 cm, and the
outside diameter (OD) at the stack top was 1.1 cm. If the model ridge were
scaled to a 150-m mountain ridge, the stack would represent a 4.5-m ID stack
with a 1-m wall thickness. Although not crucial, these scales correspond
reasonably well to representative field situations. Longitudinal (x, downwind
from center of ridge), lateral (y, from center of tunnel), and vertical (z,
from floor of tunnel) concentration profiles were taken for stack heights of
0.5 H and 1.5 H and a free-stream velocity of 3.05 m/sec.
An air-methane mixture was ejected from the stack as a tracer gas. This
effluent simulated a neutrally buoyant plume because the amount of methane
in the gas mixture was only 1 percent and the stack gas temperature was
equal to the ambient air temperature. The stack exit velocity was controlled
to permit an effluent to wind speed ratio of 3:2. This ratio has general
acceptance as a "rule of thumb" for avoiding stack downwash. With the stack
placed in the free-stream flow, a minimal plume rise without any downwash
was achieved. The above stack conditions were selected so that all plume
rise and downwash could be associated with effects from the mountain ridge.
The stack exit velocity, W , was 4.57 m/sec, which made the stack Reynolds
number, WSD/V, equal to 2730. To ensure a turbulent flow from the stack,
two serrated washers were placed inside the stack at distances of 10 and 12
stack diameters from the orifice.
Concentration profiles were obtained by collecting and passing appropri-
ate air-methane samples through a Beckman Model 400 Hydrocarbon Analyzer,
which is a flame ionization detector. Its response time of 0.5 sec was too
long to examine any dispersion microstructure. However, time averages can
he related to steady-state averages occurring in similar full scale situat-
ions, A 2.5-minute averaging time was found to yield stable values of con-
centration. The sampling rate of the 0.16-cm OD probe was 200 cubic centi-
meters per minute. The diameter of the sample stream tube thus lies between
1 and 2 millimeters (mm), which represents the spatial resolution. For
ground level concentration measurements, the probe was kept slightly above
the surface (5 millimeters) to avoid any unnatural surface effects.
Also in Phase II, concentration measurements were made to verify the
visualization results of Phase I. The sampling probe was fixed to the wind
tunnel surface at a point three ridge heights downstream from the ridge
center (base). Pollutants emitted anywhere within the cavity region pro-
duced measurable concentrations at the ridge base as a result of intense
mixing and recirculation. The reason for locating the sampler near the
mountain ridge base was to be certain that measured concentrations could be
associated only with recirculation within the cavity. At several downwind
locations, the stack was fixed and extended vertically to the height at
which no measureable concentrations could be detected at the sampling point.
The size of the cavity region determined in this way was found to be in good
agreement (Section 6.3) with the results of the smoke visualization portion
of the study (Phase I).
14
-------
The objective of Phase III v;as to determine the effects that different
approach flows might have upon the size and shape of both the smoke envelope
and the cavity. Five different cases were examined (Figure 6). The 15-cm
Gaussian ridge was used in four cases and a 15° leeward sloping plane was
used in a fifth case. Because the ridge height was one-half that of the
previously tested ridges, the free-stream velocity for this part of the study
was doubled to 6.1 m/sec in order to maintain the same model Reynolds number.
The main reason for the use of a smaller mountain ridge was to more
easily accomodate a raised plain stretching from the apex of the ridge up-
stream to the test section entrance. With the raised plane in place, the
smoke envelope and cavity were measured for both a 15-cm natural boundary
layer resulting solely from flow over the smooth surface of the wind tunnel
(Case A) and for the 60-cm simulated boundary layer (Case B). Measurements
were also made for the 60-cm simulated boundary layer flow over a 15° sloping
plane placed downstream from the ridge apex (Case C).
For the remaining cases, the upstream plane was removed. The smoke
envelope and cavity were measuiou Tu, uuui uie 15-cm natural boundary layer
(Case D) and the 60-cm simulated boundary layer (Case E). Situations char-
acteristic of real atmospheric flow over mountain ridges will most likely
have approach flows that lie somewhere between a thick turbulent boundary
layer approaching over a raised ;:l?.r:2 :.r.d a t"ir, boundary layer approaching
over smooth terrain. The resulting smoke visualized cavity and envelope size
measurements are presented and their significance is discussed in Section 6.4.
The wind tunnel ceiling height was adjusted to compensate for blockage
effects of the models and to obtain a nonaccelerating free-stream flow above
the mountain ridge. The blockage effects of the 30-cm models were more dif-
ficult to minimize than those of the 15-cm model. For the 30-cm ridges, the
ceiling adjustments that were made resulted in velocity deviations of less
than 5 percent. For the 15-cm ridge, deviations were within 2 percent. A
0.65-cm pitot-static probe was uocJ Lu make ~iuny i cudinal velocity measure-
ments along the center!ine of the wind tunnel0 The vertical placement of
the probe was well within the free-stream flow away from the floor and ceil-
ing boundary layers.
15
-------
CASE A
CASE B
CASE C
CASE D
CASEE
Figure 6. Phase III ridge configurations.
16
-------
SECTION 6
RESULTS
This section presents those experimental results relevant to describing
the flow structure leeward of a mountain ridge. The cavity and envelope
sizes and shapes and the mean velocity and turbulence intensity profiles for
the Gaussian and triangular ridges at two different flow speeds are compared
in Sections 6.1 and 6.2. Section 6.3 presents the concentration data
resulting from stacks emitting into the cavity of the Gaussian ridge. The
effect of changing the approach flow conditions on cavity and envelope size
is presented in Section 6.4.
6,1 PHASE I: ENVELOPE AND CAVITY VISUALIZATION RESULTS FOR 30-cm
GAUSSIAN RIDGE AND 30-cm TRIANGULAR RIDGE
In order to have flow separation occurring at the apex of the 30-cm
Gaussian ridge, a trip as discussed in the previous section was required.
Figures 7 and 8 present the envelope and cavity smoke visualization measure-
ments that were made to demonstrate the effect of forcing separation of the
flow at the apex of the ridge and the effect of change in free-stream mean
velocity. With flow separation at the apex, the cavity depth and length
were two times larger than for the similar cases without the trip, and were
independent of the mean flow speed. The change in the free-stream mean
velocity showed little visible effects. The flow visualization measurements
behind the 30-cm triangular ridge [Figures 9 and 10) result in similar size
and shape of both the cavity and envelope as those for the tripped 30-cm
Gaussian ridge. The differences in boundary shape near the leeward side of
the comparative cavity regions are primarily caused by the increased diffi-
culty in acquiring the visualization measurements. At that position, down-
wash was less distinct and the smoke was more diffuse, making visualization
more difficult. The photographs of Figure 11 show the same envelope shape
for the two ridges.
The envelope height measurements in these and other situations to be
presented later were only taken to a limited extent downwind. The envelope
will continue to grow with increasing distance downwind: however, the magni-
tude of the flow disturbances associated with the envelope region naturally
decreases.
6.2 PHASE I: MEAN VELOCITY AND TURBULENCE INTENSITY MEASUREMENTS
FOR 30-cm GAUSSIAN RIDGE AND 30-cm TRIANGULAR RIDGE
The mean velocity and turbulence intensity profiles are presented in
Figures 12 through 17 for both upstream and downstream positions from the
30-cm Gaussian ridge and the 30-cm triangular ridge. An identical approach
17
-------
3.0 r-
2.5
2.0
1.5
1.0
0.5
0.0
O ENVELOPE; Uoo = 7.62 m/sec (TRIP)
D ENVELOPE; Uoo 3.05 m/sec (TRIP)
A ENVELOPE; Uoo = 7.62 m/sec (SEPARATION NOT FORCED, NO TRIP)
H = 30.5 cm
5
x/H
Figure 7. Envelope size in lee of 30.5-cm Gaussian ridge.
3.0
10
2.5
2.0
1.5
1.0
O CAVITY; Uoo 7.62 m/sec (TRIP)
Q CAVITY; Uoo 3.05 m/sec (TRIP)
A CAVITY; Uoo 7.62 m/sec (SEPARATION NOT FORCED NO TRIP)
H 30.5cm
0.5
Figure 8. Cavity size in lee of 30.5-cm Gaussian ridge.
18
-------
3.0
2.5
2.0
X
"^ 1.5
1.0
0.5
ENVELOPE; U
ENVELOPE; U
H = 30.5 cm
= 7.62 m/sec
= 3.05 m/sec
234567
x/H
Figure 9. Envelope size in lee of 30.5-cm triangular ridge.
2 345678
Figure 10. Cavity size in lee of 30.5-cm triangular ridge.
10
CAVITY; U- = 7.62 m/sec
CAVITY; U» = 3.05 m/sec
H = 30.5 cm
19
-------
Figure 11. Smoke envelope in lee of (A) 30-cm Gaussian ridge and (B) 30-cm
triangular ridge. (The strut supporting the smoke tube is laterally displaced
from the smoke-filled region.)
20
-------
flow, that is, the 60-cm simulated boundary layer, was used in all cases.
The mean velocity, U, is defined as the 1-minute average flow speed in the
longitudinal direction, x, at a given position--x, y, z. The turbulence
intensity is defined as the standard deviation of the velocity fluctuation
in the longitudinal direction, normalized with the mean velocity at the same
position. Measurements in regions having mean flow reversals are not quan-
titatively valid because the hot film cannot distinguish flow direction.
Smoke visualization quite clearly revealed upstream flow near the ground
level. The data presented, however, should permit valid qualitative compar-
isons.
The degree of disruption of the approach flow is small in extent, for
tne untripped ridge (Figures 12 and 13). The trip, however, has a major
influence on the flow structure (compare Figures 12 and 13 with Figures 14
and 15). The profiles around the 30-crn triangular ridge (Figures 16 and 17)
compare quite well with those of the tripped 30-cm Gaussian ridge (Figures
14 and 15). Note that, for all cases, the profiles are independent of the
free-stream mean velocity. The mean velocity profiles at the upstream base
location for all cases are similar to a one-third power law profile. The
change from the one-seventh power law profile of the approach flow (Figure
4) is caused by upstream blocking induced by the model ridge.
Included in Figure 14 are data obtained far downstream from the cavity
region. Note the three distinct parts of the mean velocity profile. The
bottom part represents the newly developing boundary layer, which has a one-
seventh power law profile, as expected. The middle part represents the
region of the readjusting boundary layer; the top part represents that of
the undisturbed boundary layer.
6.3 PHASE II: CONCENTRATION MEASUREMENTS FOR STACK PLACED IN LEE OF
30-cm GAUSSIAN RIDGE
Concentration measurements are presented in order to assist in describ-
ing plume behavior within the cavity region. To present the data in a form
for easy comparison, the measured concentrations have been nondimensionalized
with the stack gas concentration, C$. Thus, in the figures, the value C,
percent, is the measured concentration expressed as the percentage of stack
gas concentration. These data can be related to field situations as discus-
sed in Section 4.
The concentration values in Figure 18 were detected by a ground level
probe located near the base of the 30-cm Gaussian ridge. The stack was
positioned at four different downwind locations and raised in increments in
the vertical while the probe sampled at a fixed position. The vertical
position at which no ground concentrations were detected defined the upper
limit of the cavity. The cavity size thus determined is similar to that
found from smoke visualization (Section 6.1). Also note the similarity in
the concentration plots and the fact that significant ground level concen-
trations are found upwind of the stack.
Figure 19 gives the resulting ground level concentrations downwind from
a stack placed at the leeward ridge base for three stack heights. The
21
-------
-2.7
0 (CENTER OF RIDGE)
2.7
5
30.5 cm
0.0
0.5
0.0
0.8 1.0 1.2 0
TURBULENCE INTENSITY
15 30 45 60
"I/ u IU, percent
Figure 12. Mean velocity and turbulence intensity profiles for 30.5-cm Gaussian ridge (separation
not forced), Uoo = 3.05 m/sec.
5.0
4 5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
O x/H -2.7
D x/H 0 (CENTER OF RIDGE)
a x/H 2.7
ft x/H 5
H 30.5cm
MEAN VELOCITY
0.2 0.4 0.6 0.8 1.0 1.2 0
15 30 45
u'^ /U, percent
Figure 13. Mean velocity and turbulence intensity profiles for 30.5-cm Gaussian ridge (separation
not forced), Uoo= 9.15 m/sec.
22
-------
5.0
4.5
4.0
3.5
3.0
Ox/H -2.7
Dx/H 0 (CENTER OF RIDGE)
A x/H 2.7
x/H 5
x/H 10
"» x/H 20
h 30.5cm
MEAN VELOCITY
5.0
4.5
4.0
3.5
3.0
TURBULENCE INTENSITY
0.8
1.0 0
15 30 45
-]/ n'2 / u percent
60
Figure 14. Mean velocity and turbulence intensity profiles for the 30.5-cm Gaussian ridge,
Uoo = 3.05 m/sec.
5.0
4.5
4.0
3.0
S 25
O x/H -2.7
Dx/H 0 (CENTER OF RIDGE)
a x/H 2.7
x/H 5
H 30.5cm
2.0
Figure 15. Mean velocity and turbulence intensity profiles for the 30.5-cm Gaussian ridge,
Uoo = 9-15 m/sec.
23
-------
5.0
4.5
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
OX/H
" 0 x/H
a x/H
" x/H
H
-2.0
0 (CENTER OF RIDGE)
2.0
5
30.5cm
15 =_30 45
u'^ /U, percent
60
Figure 16. Mean velocity and turbulence intensity profiles for the 30.5-cm triangular ridge,
Uoo=3.05 m/sec.
Ox/H 0 (CENTER OF RIDGE)
A x/H 2.0
x/H 5
H 30.5cm
15 30 45
u'2/ U, percent
60
i velocity and turbulence intensity profiles for the 30.5-cm triangular ridge,
. = 9.15 m/sec.
24
-------
ro
en
H = 30.5 cm
c, % c,%
0 0.' 12 0 0.1 0.2
0.1 0.2
6 8
Figure 18. Ground level concentrations measurements with sampling probe fixed near base (x/H = 3 from ridge cen-
ter, z/H = 0) of 30.5 cm Gaussian ridge. (Stack was placed at four downwind locations and its height was varied).
10
-------
0.15
0.125
0.07
0.05
0.025
Figure 19. Longitudinal ground level concentration profiles downwind from stack
placed at base (x/H = 2.7 from center) of 30.5-cm Gaussian ridge.
26
-------
points of maximum concentration are found very close to the stack base,
perhaps even upwind of the stack base. The peak concentration decreases
and is shifted downwind as the stack height increases. This is typically
found to occur over smooth terrain. Beyond two ridge heights downwind from
the stack, however, the concentrations are highest for the elevated stack.
Figure 20 presents the vertical concentration profiles in the lee of
the 30-cm Gaussian ridge. The stack was fixed at the ridge base with HS/H
equal to 0.5. Because of the nearly stagnant mean flow in the lee of the
ridge and the general upward flow in the cavity region along the ridge sur-
face, a substantial plume rise occurs as indicated by the vertical concen-
tration profile. The concentrations on the leeward side of the cavity are
more uniform than those farther upstream, but little spreading occurs into
the region above the cavity (z/H = 2, as determined in Figure 18). The
photographs of Figure 21 show two realizations of the plume from the short
stack. These photographs show the substantial plume rise and very wide
dispersion in both the upwind and downwind directions and the large unstead-
iness of the flow. The lateral concentration profiles (Figure 22 through
24) show that the lateral plume width changes only slightly in the downwind
direction. Figure 25 vividly demonstrated how rapidly the pollutants are
diluted along the downwind direction near the elevation of the maximum
concentration.
Additional concentration measurements for the same stack, elevated to a
height 1.5 times the ridge height, were also made. Figure 26 presents a few
vertical concentration profiles for the elevated stack and shows the plume
rise to be essentially zero. Zero plume rise occurs because the elevated
stack while in the cavity is above the region of mean flow stagnation. The
elevation of the point of maximum concentration decreases with downwind
distance, showing evidence of the recirculation within the cavity. The
highly uniform concentrations below z/H = 1 are also evidence of the recir-
culation. The photographs of Figure 27 show three realizations of the plume
from the tall stack. The small plume rise, the large unsteadiness of the
flow, and the wide vertical dispersion are evident. Even for this elevated
stack, however, little dispersion into the region above the cavity occurs.
The lateral ground level concentration profiles (Figure 28 and 29) show
essentially identical spreads with only minor differences in their values
near the center (y/H = 0). From these results, it is evident that instead
of the strong immediate downwash that occurred for the shorter stack, the
taller stack emissions are caught in the outer recirculation region within
the cavity. The direction of the recirculation is down the leeward side of
the cavity and upstream along the ground.
27
-------
x/H = 4.75 FROM STACK
x/H = 2.25 FROM STACK
- x/H = 1.25 FROM STACK
H = 30.5 cm
Figure 20. Vertical concentration profiles for stack placed at base (x/H = 2.7 from center) of
30.5 cm Gaussian ridge, HS/H = 0.5.
28
-------
Figure 21. Smoke visualization with 0.5-H stack placed at base of 30-cm
Gaussian ridge (x/H = 2.7 from ridge center).
29
-------
0.06
0.04
0.02
0.0
-4
Figure 22. Lateral concentration profiles from stack placed at base (x/H = 2.7 from center)
of 30.5-cm Gaussian ridge, x/H = 1.25, HS/H = 0.5.
0.12
0.08
S 0.06
0.04
0.02
Figure 23. Lateral concentration profile from stack placed at base (x/H = 2.7 from center) of
30.5-cm Gaussian ridge, x/H = 2.25, HS/H = 0.5.
30
-------
0.06
0.04
0.02
Figure 24. Lateral concentration profiles from stack placed at base (x/H = 2.7 from center) of
30.5-cm Gaussian ridge, x/H = 4.V5, Hs/n = 0.5.
31
2/H =1
H =30.5 cm
oL
4
x/H
Figure 25. Elevated longitudinal concentration profile downwind from stack placed at
base (x/H = 2.7 from center) of 30.5-cm Gaussian ridge, HS/H = 0.5.
31
-------
x/H =4.75 FROM STACK
x/H =2.25 FROM STACK
A x/H 1.25 FROM STACK
H = 30.5 cm
\
LAC!
0.4
C, percent
Figure 26. Vertical concentration profiles for stack placed at base (x/H = 2.7 from center)
of 30.5 cm Gaussian ridge, Hg/H = 1.5.
32
-------
Figure 27. Smoke visualization with 1.5 H stack placed at base of 30-cm
Gaussian ridge (x/H = 2.7 from ridge center).
33
-------
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
O z/H 1.5
H = 30.5cm
1.5
0.5
0
y/H
0.5
1.5
Figure 28. Lateral concentration profile from stack placed at base (x/H = 2.7 from
center) of 30.5-cm Gaussian ridge, x/H = 1.25, HS/H = 1.5.
0.06
0.04
0.02
Ox/H = 1.25 FROM STACK
Dx/H = 2.25 FROM STACK
A x/H 4.75 FROM STACK
H 30.5cm
D
-3
2
y/H
Figure 29. Lateral ground level concentration profiles from stack placed at base (x/H = 2.7
from center) of 30.5-cm Gaussian ridge, z/H = 0, HS/H = 1.5.
34
-------
6.4 PHASE III: CAVITY AND ENVELOPE MEASUREMENTS IN LEE OF
15-cm GAUSSIAN RIDGE
The goal was to determine what effect the approach boundary layer
conditions have on the size and shape of the leeward cavity region Figures
30 through 35 contain the results of the third phase of this study] In
addition to the cavity and envelope measurements, some mean velocity profiles
taken with a pi tot-static tube above the ridge center were included.
Figure 30 shows the resulting cavity size and envelope for the 60-cm
simulated boundary layer approaching the 15-cm Gaussian ridge (Case E),
which is similar to that used in Phase I. The boundary layer thickness to
ridge height ratio was four (double that for the 30-cm ridge); however,
little difference in the size or shape of the cavity or envelope region
resulted (compare with Figure 7 and 8).
With the vortex generators and roughness removed (Case D), the
natural boundary layer over the smooth tunnel floor was found to be
depth, so that the boundary layer to ridge height ratio was about 1
mean velocity profile over the ridge center (Figure 31, Case D) was
jet and also to have a slightly larger shear in comparison with the
simulated boundary layer flow over the 30-cm Gaussian ridge (Figure
turbulence levels for the natural boundary layer may be expected to
less than those of the simulated atmospheric boundary layer. These
ences could explain the slightly larger extent of the cavity region
Figures 30 and 8 to Figure 32). The depth of the
those of Case E (Figure 30) and the similar case
resulting
15 cm in
The
found to
60-cm
14). The
be much
differ-
(compare
envelope agrees well with
in Phase I (Figure 7).
3.0
2.0
1.5
CASE E
OENVELOPE
D CAVITY
H 15.2cm
Figure 30. Envelope and cavity size in lee of 15.2-cm Gaussian ridge (60-cm simulated boundary
layer).
35
-------
5.0
4.5
4.0
3.5
3.0
2.0
1.5
0.5
O CASE A
D CASE B
A CASE D
IU = 6.1 m/sec
H 15.2cm
0.1
0.2 0.3
0.4 0.5
0.6
0.7 0.8
0.9
1.0
1.1
1.2
Figure 31 . Mean velocity profiles above center of 1 5.2-cm Gaussian ridge.
Changes in the size of both the cavity and envelope were found when the
mean flow approached the ridge along a raised plane elevated to the height
of ridge apex. Figure 31 presents the mean velocity profile above the center
of the ridge for the 15-ctn natural boundary layer (Case A), which is shown
to be similar to that of Case D. Both the wake and envelope size (Figure
33), however, are significantly reduced in comparison with the previous cases
(D and E) for an isolated ridge. Very similar results were found for the
situation with the 60-cm simulated boundary layer (Case B, Figure 34). The
effect of changing the boundary layer conditions appears to be minor.
The upwind terrain has been shown to have a direct effect on the size
of the cavity and envelope. In addition, angle of downs! ope from the ridge
apex can also have an effect on the size of the cavity region. With small
slope angles, the adverse pressure gradient in the lee of the ridge may be
expected to be lower, wnich would result in a smaller stagnation region.
The envelope and cavity size in the lee of a 15° slope (Case C) is shown in
Figure 35. The approach flow conditions were identical to Case B (Figure
31). As expected, the cavity size is reduced significantly.
36
-------
CASE D
OENVELOPE
D CAVITY
H 15.2cm
67 8 9 10 11 12
0.0
0 1
Figure 32. Envelope and cavity size in lee of 15.2-cm Gaussian ridge (15-cm natural boundary
layer).
CASE A
OENVELOPE
D CAVITY
0.0
Figure 33. Envelope and cavity size in lee of 15.2-cm Gaussian ridge (15-cm natural boundary
layer with raised plane approach).
37
-------
CASE B
O ENVELOPE
D CAVITY
H 15.2cm
10
Figure 34. Envelope and cavity size in lee of 15.2-cm Gaussian ridge (60-cm simulated boundary
layer with raised plane approach).
Figure 35. Envelope and cavity size in lee of 1 5° sloping plane (60-cm simulated boundary
layer with raised plane approach).
38
-------
SECTION 7
DISCUSSION AND SUMMARY
On the lee side of a mountain ridge, under neutral atmospheric condi-
tions, there is a highly turbulent region in which significant plume down-
wash and mixing occurs. This region, which is caused by the main flow
separating from the leeward surface of the ridge, is characterized by a
general air recirculation and is referred to as the cavity. Although other
types of phenomena are expected to occur in nonneutral flows, a firm under-
standing of the more simplistic neutral flow is a necessary first step.
Neutrally stratified and moderate to high wind conditions offer the greatest
potential for the natural occurrence of a separated flow region similar to
that examined in this study.
The literature search reported in Section 3 revealed little detailed
information that would relate the size or extent of the cavity to either the
downslope shape or the mean flow characteristics. The field studies do,
however, strongly support the assertion that a recirculating flow region with
strong downwash and dispersion characteristics often exists on the leeward
side of a ridge, A ridge with a salient peak or a steep sloping side can be
expected to exhibit separation at its apex. Under this circumstance, the
cavity will be its largest, both in depth and horizontal extent.
For the model Gaussian ridge, the point of separation was found to occur
considerably downstream from its apex, and flow visualization showed the
point of separation to fluctuate widely in position from one instant of time
to the next. In order to demonstrate the maximum possible extent of the
cavity, the main flow was forced to separate at its apex. Under neutral
stability and high wind speed conditions, the model results can be expected
to represent actual results behind similarly shaped ridges that exhibit
mean flow separation at their apexes.
For the 30-cm Gaussian ridge with separation occurring at its apex, the
maximum depth and horizontal extent of the cavity region were found to be
2 H and 10 H, respectively. Similar results were found for the 30-cm tri-
angular ridge. The mean velocity and turbulence intensity profiles were
also found to be similar for the two ridges. Thus, the flow patterns in the
lees of ridges that exhibited separation at their apex were found not to be
sensitive to the detailed shape of the downslope,
Cavity sizes in the lee of the tripped 15-cm Gaussian ridge were meas-
ured with four different approach flows. With the 60-cm simulated boundary
layer approach, which was similar to the situation examined for the 30-cm
ridge, the maximum depth and horizontal extent were found to be 2 H and 10 H,
39
-------
respectively. The cavity sizes and shapes were found to be only slightly
affected by the thickness and intensity of the approach Boundary layer, but
to be fairly strongly dependent upon the upwind terrain and the gross fea-
tures (angle) of the downslope.
An important objective of this study was to demonstrate the behavior of
a plume emitted from a stack placed within the cavity region. The 30-cm
Gaussian ridge with separation occurring at its apex and a stack emitting an
air-methane mixture as a tracer gas were used. The size of the cavity deter-
mined from the tracer gas measurements showed a maximum depth and horizontal
extent of 2 H and 8.5 H, respectively. The horizontal extent was only
slightly smaller than that found by smoke visualization for the same
situation.
Ground level concentrations were sampled through a probe fixed near the
base of the ridge (3 ti) for several stack heights with the stack placed at
2.7 li, 4 H, 6 H, and 8 H downwind. At all stack positions, the measured
ground level concentration increased with the lowering of the stack height.
The fact that significant ground level concentrations are found at a distance
of 5 H upwind of the stack confirms the existence of a strong recirculating
flow within the cavity region.
In order to quantitatively demonstrate the types of plume behavior that
exist for stack emissions within the cavity, longitudinal, lateral, and
vertical concentration profiles were taken throughout the cavity region for
three stack heights with the stack fixed at the base of the ridge. A stack
placed at this location results in measured concentrations (order of magni-
tude) that might occur if the stack were placed elsewhere downwind in the
cavity. Flow visualization and the concentration measurements aloft support
this assertion. For a stack positioned further upwind and along the ridge
slope, the upslope flow would dominate. This may provide characteristically
different concentration profiles.
With the stack positioned at the downwind base of the ridge, longitu-
dinal ground level concentration profiles were measured for three stack
heights. For the 0.5 H and the 1 H stack, the highest concentrations were
found at a position very near the stack. This is an indication of strong
immediate downwash. For the 1.5 H stack, the highest concentrations occur-
red significantly further downwind of the stack. Concentrations measured
near the mean flow reattachment point were highest for the 1.5 H stack. This
occurs because emissions near the upper boundary of the cavity region are
caught in the general recirculation, which downdrafts the plume towards the
reattachment point. Emissions from a lower stack are more rapidly downwashed
and dispersed. Thus, this is one case in which a taller stack does not
result in lower concentrations. Application of the "2 1/2 times rule" with
respect to the terrain feature would, however, result in significantly
reduced concentrations downwind because this stack would be well above the
cavity region.
Near the downwind end of the cavity (that is, near the reattachment
point), the flow is difficult to characterize, Part of the main flow recir-
culates within the cavity; the rest continues downwind forming a newly de-
40
-------
veloping boundary layer. The turbulence Intensity downwind of the reattach-
ment point fs still significantly greater than that found in the undisturbed
flow and tt decreases with downwind distance. This region can best be char-
acterized by the increased vertical and lateral spreading of a plume over
that occurring for a flow without the mountain ridge disruption, The above
assertions are generalizations drawn from a limited amount of smoke visual-
ization and mean velocity and turbulence intensity data taken downwind of the
reattachment point. Further studies of the behavior of plumes from stacks
placed downwind of the reattachment point are needed to further characterize
dispersion there.
41
-------
REFERENCES
American Society of Mechanical Engineers, 1973: Recommended Guide for the
Prediction of the Dispersion of Airborne Effluent?!Smith, M. (ed.).New"
York, ASME. 85 p.
Bradshaw, P. and F. Y. F. Wong, 1972: The Reattachment and Relaxation of
Turbulent Shear Layer. J. Fluid Mech.. 52_, 113-135. Part 1.
Briggs, G. A., 1969: Plume Rise. U.S. Atomic Energy Commission, Washington,
D.C. Critical Review Series No. TID-25075. 81 p.
Buettner, K. J. K., 1964: Orographic Deformation of Wind Flow. University
of Washington, Seattle, Washington. Prepared for U.S. Army Electronics
Research and Development Laboratory, Fort Honmouth, New Jersey, under Pro-
ject No. 1AO-11001-B-021-01, Contract No. DA 36-039-SC-89118. 70 p.
Corby, G. A., 1954: Airflow over Mountains: A Review of Current Literature.
Quart. J. Roy. Met. Soc., 80_, 491.
Counihan, J., 1969: An Improved Method of Simulating an Atmospheric Boundary
Layer in a Wind Tunnel. Atm. Environ., 3_, 197-214.
Counihan, J., J. C. R. Hunt, and P. S. Jackson, 1974: Wakes Behind Two-
Dimensional Surface Obstacles in Turbulent Boundary Layer. J. Fluid Mech.,
64, 529-563. Part 3.
Davenport, A. G., 1963: The Relationship of Wind Structure to Wind Loading.
In: Proceedings of Conference on Wind Effects on Buildings and Structures.
(National Physical Laboratory, Teddington, Middlesex, England, June, 1963)
HMSO, London, 1965. p. 54-102.
Davidson, B., 1963: Some Turbulence and Wind Variability Observations in
the Lee of Mountain Ridges. J. Appl. Meteorology, 2_(4), 463-472.
Evans, B. H., 1957: Natural Air Flow Around Buildings. Texas Engineering
Experiment Station, College Station, Texas. Research Report No. 59.
Gloyne, R. W., 1965: Some Characteristics of The Natural Wind and Their
Modification by Natural and Artificial Obstructions. Scientific Horticul-
ture, XVII, 7-19.
Halitsky, J., 1968: Gas Diffusion Near Buildings. In: Meteorology and
atomic Energy- 1968. Slade, D.H. (ed.). U.S. Atomic Energy Commission.
Report No. TID-24190. p. 221-255.
42
-------
Halitsky, J., G. A. Magony, and P. Halpern, 1966: Turbulence Due to Topo-
graphical Effects. New York University, New York, Geophysical Laboratory
Report No. TR-66-5. 75 p.
Lin, J. T.} H. T. Liu, Y. H. Pao, D. K. Lilly, H. Israeli, and S. A. Orsag,
1974: Laboratory and Numerical Simulation of Plume Dispersion in Stably
Stratified Flow over Complex Terrain. Flow Research, Inc., Kent, Washington.
Prepared for U.S. Environmental Protection Agency, Research Triangle Park,
N.C. under Contract No. 68-02-0800. Publication No. EPA-650/4-74-044. 70 p.
Mueller, T. J. and J. M. Robertson, 1963: Modern Developments in Theor.
Appl. i-iech., 1_, 326.
Nagabhushanaiah, H. S., 1962: Separation Flow Downstream of a Plate Set
Normal to a Plane Boundary. Ph.D. Dissertation. Colorado State University,
Fort Collins, Colo.
Orgill, M. M., J. E. Cermak, and L. 0. Grant, 1971: Laboratory Simulation
and Field Estimates of Atmospheric Transport - Dispersion Over Mountainous '
Terrain. Colorado State University, Fort Collins, Colo. Technical Report
No. CER70-71MMO-JEC-LOG40.
Petryk, S. and E. Brundrett, 1967: Department of Mechanical Engineering,
University of Waterloo. Researcii Report No. 4.
Plate, E. J., 1967: Diffusion from a Ground Level Line Source into the
Disturbed Boundary Layer far Downstream from a Fence. Int. J. Heat and
Mass Transfer. 10_, 181-194.
Plate, E. J. and C. M. Liu, 1964: The Velocity Field Downstream from a
Two-dimensional Model Hill. Part 1. Colorado State University, Fort
Collins, Colo. Technical Report No. CER65EJP-CWL41.
Plate, E. J. and C. M. Sheih, 1965: Diffusion from a Continuous Point
Source into the Boundary Layer downstream from a Model Hill. Colorado State
University, Fort Collins, Colo. Technical Report No. CER65EJP-CMS60.
Pooler, F., Jr. and L. E. Niemeyer, 1970: Dispersion From Tall Stacks: An
Evaluation. (Presented at 2nd International Clean Air Congress, Washington,
D.C. December 6-11, 1970, Paper No. ME-14D.) 31 p.
Scorer, R. S., 1955: Theory of Airflow over Mountains: IV-Separation of
Flow from the Mountain Surfaces, Quart. J. Roy. Met. Soc., 81, 340-350.
Scorer, R. S., 1963: Air Pollution. Oxford Pergamon Press. 151 p.
Sherlock, R. H. and E. A. Stalker, 1940: The Control of Gases in the Wake
of Smoke Stacks. Mech. Engr., 62, 455-458.
43
-------
Snyder, W. H., 1972: Similarity Criteria for the Application of Fluid
Models to the Study of Air Pollution Meteorology, Boundary-Layer Meteorology,
1(2), 113-134.
Snyder, W. H., R. S. Thompson, and R. E. Lawson, Jr., 1976: The EPA Meteoro-
logical Wind Tunnel: Design, Construction and Operating Details. In prepara-
tion for publication in the EPA Environmental Monitoring Series.
Strom, G. H., 1968: Atmospheric Dispersion of Stack Effluents. In: Air
Pollution. Vol. I. Stern, A. C. (ed.). New York, Academic Press. Chap-
ter 8.
Sutton, 0. C.s 1960: Discussion before Institute of Fuels, May 23, 1960.
J. institute Fuel (London), 33, 495.
Tani, I., M. luchi, and H. Komoda, 1961: Aeronautical Research Institute,
Tokyo, Japan. Report No. 364.
Tillman, W., 1945: Volkenrode translation MAP - VG 34 - 45T. British
Ministry of Aircraft Production, London, England.
World Meteorological Organization, 1967: The Airflow over Mountains. !
Geneva, Switzerland. Report No. 98. 43 p.
44
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/4-76-047
*. TITLE AND SUBTITLE
STACK PLACEMENT IN THE LEE OF A MOUNTAIN RIDGE
A Wind Tunnel Study
5. REPORT DATE
September 1976
6. PERFORMING ORGANIZATION CODE
3. RECIPIENT'S ACCESSION-NO.
7. AUTHOR(S)
Alan H. Huber, William H. Snyder*,
Roger S. Thompson, and Robert E. Lawson,
Jr.
8. PERFORMING ORGANIZATION REPORT NO.
Fluid Modeling Report No. 2
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
10. PROGRAM ELEMENT NO.
1AA603
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
In-house 1/75 - 6/75
14. SPONSORING AGENCY CODE
£
EPA-ORD
15. SUPPLEMENTARY NOTES
On assignment from the National Oceanic and Atmospheric Administration.
16. ABSTRACT
An investigation of the highly turbulent region in the lee of a two-dimen-
sional mountain ridge was carried out in the Meteorological Wind Tunnel of the U.S.
EPA's Fluid Modeling Facility. This highly turbulent region was found to consist of a
large semi-permanent eddy. Smoke visualization and hot film anemometry measurements
showed that the cavity size and shape are minimally affected by the thickness and tur-
bulence intensity of the approach boundary layer flow. In addition, the cavity size
and shape were not found to be affected by the detailed shape of the ridge, but were
strongly dependent upon the upwind terrain and the gross features [angles) of the
downs!ope.
A stack was positioned to emit an air-methane mixture into the cavity in the lee
of the ridge. Longitudinal, lateral and vertical concentration profiles were taken.
A tall stack placed to emit into the upper portion of the cavity resulted in higher
ground level concentrations near the downwind end of the cavity than did a shorter
stack. However, the maximum concentrations measured were found to occur near the
bases of the shorter stacks.
The cavity region leeward of the model ridge was found to be highly turbulent
with significant plume downwash. For similar real-atmosphere situations, it would be
good engineering practice to avoid placement of any significant source within the ex-
pected cavity region. Application of the "2 1/2 times rule" for stack construction
would be sufficient to avoid the highly turbulent region of the cavity proper.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDEDTERMS
COSATI Field/Croup
Air pollution
Wind tunnel
Boundary layer
Atmospheric diffusion
Mountain ridges
Chimneys
Downwash
13B
14B
20D
04A
08F
13M
20D
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
55
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
45
------- |