WINDBREAK EFFECTIVENESS FOR STORAGE-PILE FUGITIVE-DUST CONTROL

                     A Wind Tunnel Study
                              by

                      Barbara J. Billman

                             and

                         S.P.S. Arya
     Department of Marine, Earth and Atmospheric Sciences
               North Carolina State University
                   Raleigh, NC  27695-8208
            Cooperative Agreement Number CR-811973
                       Project Officer

                      William H. Snyder
             Meteorology and Assessment Division
           Atmospheric Sciences Research Laboratory
        Research Triangle Park, North Carolina  27711
           ATMOSPHERIC SCIENCES RESEARCH LABORATORY
              OFFICE OF RESEARCH AND DEVELOPMENT
             U.S. ENVIRONMENTAL PROTECTION AGENCY
        RESEARCH TRIANGLE PARK, NORTH CAROLINA  27711

-------
                          DISCLAIMER





     This report has been reviewed by the Atmospheric Sciences Research



Laboratory, U.S. Environmental  Protection Agency, and approved for



publication.  Approval  does not signify that the contents necessarily



reflect the views and policies  of the U.S. Environmental  Protection



Agency, nor does mention of trade names or commercial products constitute



endorsement or recommendation for use.

-------
                                ABSTRACT

     Results of wind-tunnel  experiments to  determine the optimal
size and location of porous  windbreaks for  controlling fugitive-dust
emissions from storage piles in  a  simulated neutral atmospheric boundary
layer are presented.  Straight sections of  windbreak material were
placed upwind of two non-erodible,  typically shaped piles and were also
placed on the top of one of  the  piles.  Wind speed, measured near the
pile surface at various locations  with heated thermistors, was isolated
here as the primary factor affecting  particle uptake.  Wind speed
distributions about the piles in the  absence of any windbreak and the
flow structure downwind of a three-dimensional porous windbreak are
presented.  Relative wind speed  reduction factors  are described and
efficiencies based on the relationship between wind speed and particle
uptake are given.  The largest and most solid windbreak caused the
greatest wind speed reduction, but similar  wind speed reductions were
obtained from several smaller windbreaks.   A 50% porous windbreak of
height equal to the pile height  and length  equal to the pile length at
the base, located one pile height  from the  base of both piles was found
to be quite effective in reducing  wind speeds over much of the pile.
Windbreaks placed on the top of  a  flat-topped pile caused large wind
speed reductions on the pile top,  but small, if any, reductions on the
windward pile face.  Windbreak effectiveness decreased as the angle
between the windbreak and the wind direction decreased.

-------
                                CONTENTS


DISCLAIMER 	   if

ABSTRACT	1	  i i i

FIGURES	    v

TABLES 	 vi i i

SYMBOLS 	   i x

ACKNOWLEDGEMENTS 	  xi i


1.   INTRODUCTION 	    1

2.   LITERATURE REVIEW 	    4
    2.1  Fugitive-Dust Emission Rates 	    4
    2.2  Flow About Windbreaks	    7
    2.3  Windbreaks for Storage-Pile Fugitive-Dust Control 	   10

3.   EXPERIMENTAL DESIGN AND INSTRUMENTATION 	   13
    3.1  Experimental Design 	   13
    3.2  Instrumentation 	   22

4.   FLOW ABOUT A POROUS WINDBREAK	   41

5.   FLOW ABOUT STORAGE PILES 	   51
    5.1  Conical Pile 	   51
    5.2  Oval, Flat-topped Pile	   53

6.   WINDBREAK EFFECTS ON FLOW ABOUT STORAGE PILES 	   58
    6.1  Windbreaks Upstream of the Pile	   58
         6.1.1  Relation to flow structure 	   58
         6.1.2  Results	   59
    6.2  Windbreaks on the Pile Top 	   75

7.   FURTHER ANALYSES AND DISCUSSION OF RESULTS 	   80
    7.1  Relation to Particle Uptake 	   80
    7.2  Comparison to Previous Studies 	   88

8.   CONCLUSIONS 	   90

9.   REFERENCES 	   93

-------
                                 FIGURES

Number                                                              Page

 2.1  Sketches of streamlines  about  windbreaks:  (a)  solid,
      (b) porous	   8

 3.1  Boundary-layer velocity  profile:  (a)  linear,
      (b) semi-logarithmic	  14

 3.2  Sketch of oval, flat-topped pile	  16

 3.3  Geometry of conical  pile and windbreak	  20

 3.4  Example hot-wire anemometer calibration  plots	  25

 3.5  Sketch of a pulsed-wire  anemometer  probe	,  27

 3.6  Typical signals observed from  various stages  of  the pulsed-wire
      anemometer	  29

 3.7  Example pulsed-wire  anemometer calibration plot	  31

 3.8  Resistance-temperature thermistor curves	  33

 3.9  Thermistor circuit	  34

 3.10 Dissipation factor vs. wind speed for Fenwal  GB31J1
      thermi stors	  34

 3.11 Top view of conical  pile.  Stars: Thermistor  positions on pile.
      Dots: Effective thermistor positions  due to pile rotation	  37

 3.12 Top view of oval, flat-topped  pile.   Dots: Thermistor positions.37

 3.13 Comparison of thermistor and single-wire wind speed
      measurements	  38

 4.1  "Streamlines" and normalized velocity vectors  downstream of a
      porous windbreak	42

 4.2  Relative wind speed  deficit downstream of a porous windbreak...  44

 4.3  Turbulence intensity downstream of  a  porous windbreak	46

 4.4  Normalized r.m.s. longitudinal velocity fluctuation downstream
      of a porous windbreak	  47

 4.5  Profiles of normalized wind speed downstream  of  a porous
      wi ndbreak	\	  48

-------
Number                                                              Page

 4.6  Profiles of normalized longitudinal  velocity  fluctuation
      downstream of a porous windbreak	  50

 5.1  u/Up about conical  pile for  no windbreak case...	  52

 5.2  u/Up about oval, flat-topped pile for no windbreak case	  54

 5.3  Corrected u/ur about  oval, flat-topped pile for  no windbreak
      case	  55

 5.4  Corrected u/ur about  oval, flat-topped pile at an angle to
      incident flow for no  windbreak case: (a) 20°,  (b) 40°	57

 6.1  Superposition of conical  pile positions and undisturbed
      "streamlines" downstream  of  a porous windbreak of height:
      (a) 0.5H, (b) l.OH, (c) 1.5H	  60

 6.2  Superposition of oval, flat-topped pile positions and
      undisturbed "streamlines" downstream of a porous windbreak
      height:  (a) 0.5H, (b)  l.OH,  (c)  1.5H	  61

 6.3  Wind speed reduction  factor  for  the  65% porous windbreak of
      height 0.5H and length l.OD  placed 1H from the conical pile
      base	  63

 6.4  Wind speed reduction  factor  for  the  65% porous windbreak of
      height 0.5H and length l.OB  placed 1H from the oval, flat-
      topped pile base	  63

 6.5  Wind speed reduction  factor  for  the  50% porous windbreak of
      height 1.5H placed  1H from the oval, flat-topped pile base with
      length 0.6B (solid  line)  and l.OB (dashed line)	  67

 6.6  Wind speed reduction  factor  for  the windbreak of height l.OH
      and length l.OD placed 1H from the conical pile  base with
      porosity 65% (solid line) and 50% (dashed line)	  69

 6.7  Wind speed reduction  factor  for  the windbreak of height l.OH
      and length 0.6B placed 1H from the oval, flat-topped pile base
      with porosity 65% (solid  line) and 50% (dashed line)	  69

 6.8  Wind speed reduction  factor  for  the  50%.porous windbreak of
      height 1.5H and length 1.5D  placed 1H (solid  line) and 3H
      (dashed  line) from  the conical pile base	  71

 6.9  Wind speed reduction  factor  for  the  50% porous windbreak of
      height 1.5H and length 0.6B  placed 1H (solid  line) and 3H
      (dashed  line) from  the oval, flat-topped pile base	  71
                                    VI

-------
Number                                                             Page

 6.10 Wind speed reduction  factor  for  the  50% porous windbreak of
      height l.OH and length  l.OD  placed 1H from the conical pile
      base oriented (a)  normal,  (b)  20°, and (c) 40° to the flow
      direction	 74

 6.11 Sketch of windbreak  positions  with respect to thermistor
      positions on the top  of the  oval, flat-topped pile	 76

 6.12 Wind speed reduction  factor  for  the  65% porous windbreak
      placed near the pile  leading edge: (a) height 0.27H and length
      0.5T, (b) height 0.14H  and length 0.5T, (c) height 0.27H and
      1 ength 0.16T	 77

 6.13 Wind speed reduction  factor  for  the  65% porous windbreak of
      height 0.27H and length 0.5T placed  near the pile centerline.
      Pile is: (a) normal,  (b) 40° to  the  incident flow	 78

 7.1  Efficiency (£3) vs.  height for windbreaks placed  3H from the
      pile base: (a) conical  pile, (b) oval, flat-topped pile	86
                                   vn

-------
                                  TABLES

Number                                                              Page

 3.1  Combinations  of windbreak  parameters used with the  conical
      pile	i	  21

 3.2  Combinations  of windbreak  parameters used with the  oval,
      f 1 at-topped pile	  23

 6.1  umax/ur for the various windbreaks placed upstream  of the
      conical  pile	  72

 6.2  umax/ur for the various windbreaks placed upstream  of the
      oval,  flat-topped  pile	  72

 7.1  Efficiency  (E^) for the various windbreaks placed upstream
      of  the conical  pile	  83

 7.2  Efficiency  (E^) for the various windbreaks placed upstream
      of  the oval,  flat-topped pile	  83

 7.3  Efficiency  (£3) for the various windbreaks placed upstream
      of  the conical  pile	  85

 7.4  Efficiency  (£3) for the various windbreaks placed upstream
      of  the oval,  flat-topped pile	  85
                                   V117

-------
                                SYMBOLS

a    constant

A    surface area

b    constant

B    oval, flat-topped pile base length [L]

c    constant

Cp   pressure drop coefficient

C    constant

d    distance between pulsed wire and sensor wire [L]

D    conical pile base diameter [L]

E    voltage [volt] or windbreak efficiency or effectiveness factor

EF   emission factor [M/L2T]

f    percentage of time that unobstructed wind speed exceeds 5.4 m/s
     at mean pile height

F    constant

h    windbreak height

H    pile height

i    electric current [amp]

k    von Karman's constant

K    thermistor dissipation factor [ML^/T^e]

L    windbreak length or a characteristic length [L]

n    exponent 1 or 3

p    nunber of days with > 0.25 mm precipitation per year

P    static pressure [H/LJ2] or distance between windbreak and pile
     upstream base [L]

PE   Thorntwaite's precipitation-evaporation index

Q    storage-pile eVosion rate [M/T]
                                   IX

-------
R    resistance [ohm] or wind speed reduction factor

s    pile surface area [L.2]

S    silt content
                                t
t    time or time-of-flight [T]

T    oval, flat-topped pile top length [L] or temperature [e]

u    wind speed [L/T]

u1   longitudinal  component of velocity fluctuation [L/T]

u    friction velocity [L/T]

U    mean wind speed or a characteristic wind speed [L/T]

V    voltage [volt]

w    oval, flat-topped pile base width [L]

w1   vertical component of velocity fluctuation [L/T]

x    Cartesian coordinate (streanwise) [L]

y    Cartesian coordinate (lateral) [L]

z    Cartesian coordinate (vertical) [L]

z0   surface roughness length [L]

a    exponent in King's law

2    thermistor constant [0]

Y    angle between direction normal to the pulsed-wire plane and the
     instantaneous velocity vector

5    boundary-layer height [L]

p    air density [M/L-3]

Pb   bulk density [M/L3]


Subscripts and special symbols

( )0   no windbreak case or free-stream value
                   \
( )}   value based on linear relation between wind speed and particle
       uptake

-------
(  )3    value based on  cubic relation between  wind speed and particle
        uptake

(  )a    ambient value

(  )i    specific pile location  value or index

(  )max  maximum value

(  )r    reference value

(  )R    reference value

(  )t    threshold value

(  )y    thermistor value

(  )     area-averaged value

-------
                            ACKNOWLEDGEMENTS





     The authors wish to acknowledge the advice and help given by



Dr. W.H. Snyder and his staff at. the EPA Fluid Modeling Facility.  The



cooperation and help from Messrs. W.B.  Kuykendal  and D.L. Harmon



(EPA-AEERL) is appreciated.   Mr.  W.R.  Pendergrass (Oak  Ridge Associated



Universities) provided the pulsed-wire anemometer.   Mr. R.  Lawrence



(KPN, International) provided the commercial  windbreak  material.
                                   xn

-------
                            1.  INTRODUCTION

     Total suspended particulate (TSP) levels in many regions of the
country do not meet the National Ambient Air Quality Standard; fugitive
dust from sources such as storage piles, materials transfer points,
unpaved roads, and agricultural  tilling contribute significantly to TSP
levels in some regions.  Wind erosion causes an estimated 30% of
storage-pile fugitive-dust emissions; load-in and load-out processes
(30%) and vehicular traffic (40%) cause the remainder (Cowherd et al.,
1974).
     Many states regulate fugitive-dust emissions by forbidding visible
fugitive dust beyond the property line surrounding the source; some
states have more strict storage-pile fugitive-dust regulations (e.g.
Bureau of National Affairs, 1982 and 1983).  Although limiting emissions
from a fugitive dust source may  not be required, a given industrial
facility may use the Environmental  Protection Agency (EPA) "bubble"
policy to their benefit by controlling fugitive dust rather than the
more costly process of upgrading the particulate controls on their
stacks to make them more efficient.
     Radioactive particulate is  also regulated.  EPA's "Environmental
Radiation Protection Standards for Nuclear Power Operations"  (40 CFR
190) provides limits for radiation doses received by the public.  The
Nuclear Regulatory Commission  (NRC) amended 10 CFR 20, requiring that
NRC licensees comply with the  EPA regulations.   Criterion 8 of that
amendment states that to control dusting from tailings,  dry tailings
shall be wetted or chemically  stabilized, unless the "tailings are
effectively sheltered from the wind, such as may be the  case where they

-------
are disposed of below grade and the tailings  surface is  not  exposed to
wind. . ."
     Storage piles, in addition to tailings,  are  found at mines,  mineral
processing plants, coal-fired electric generating plants, within  the
iron and steel  industry, etc. and contain  coal, coke, limestone,
aggregate, sand, etc.
     Early air pollution control  efforts emphasized  controlling
emissions from stacks rather than fugitive emission  sources  because the
greater bulk of pollutants came from stacks.   Now that efficient  control
methods for particulate matter from stacks are available, control
methods for fugitive dust sources are being tested.   The EPA Air  &
Energy Engineering Research Laboratory (formerly  Industrial
Environmental  Research Laboratory) requested  that the EPA Atmospheric
Sciences Research Laboratory (formerly Environmental  Sciences Research
Laboratory) conduct a wind tunnel study to assess windbreak
effectiveness for the control of fugitive  dust from  storage  piles.
This study was undertaken as part of a cooperative agreement between
the EPA and North Carolina State University.
     Since wind direction is variable, encircling a  storage-pile  with a
windbreak will  shield it from winds in all directions.   However,  this
may not be practical, particularly for an  active  pile, or economically
feasible due to size.  A straight windbreak placed to protect a pile
from the prevailing winds may be more practical.   A  windbreak may also
be placed on top of some piles to protect  certain areas, say where
material is removed from or added to the pile.  In the present study,
wind speed is isolated as the primary factor  affecting particle uptake,

-------
although moisture content, particle size,  and bulk  density affect



fugitive-dust emissions as well.   Wind speed was measured near the pile



surface with and without windbreaks of several  sizes  and porosities



located various distances upwind-or on the top of two typically shaped



storage piles.  No effort was made to simulate fugitive dust  emissions.



Effect of wind direction was also observed.   The wind speed patterns



were analyzed to determine the optimal windbreak porosity, height,



length, and location and to develop windbreak design  guidelines for



storage-pile fugitive-dust control.



     The following sections consist of a literature review, a



description of the experimental design, instrumentation, and  the flows



about a representative windbreak  and about the two  storage piles.   Wind



speed patterns for various windbreak cases are presented and  effects  of



the windbreak parameters are noted.  Finally, the wind speed



distributions are related to particle uptake.

-------
                         2.   LITERATURE  REVIEW





     Reviews of three topics  are presented  as background material  to



the present study.   The relationship  between wind  speed and  fugitive-



dust emission rate is examined first. Second,  the structure of  the



flow about a windbreak placed on a  horizontal surface  is examined.



Finally, a review of the literature dealing with windbreaks  for  storage-



pile fugitive-dust control  is presented.





2.1  FUGITIVE-DUST EMISSION RATES



     Mechanical  forces by such implements as bulldozers acting on  the



pile surface and by falling material  impacting  the surface,  freezing



and thawing, etc.,  create particles that may become  airborne.  Storage-



pile fugitive-dust emission rates depend upon the  stored material's



bulk density, moisture content and  particle size distribution, the



storage pile geometry, the wind velocity near the  pile surface and



other parameters.  However,  particle  uptake does not occur unless  the



wind speed is greater than a  given  value, the threshold velocity,  which



is dependent upon the type of stored  material,  its moisture  content and



particle size distribution.



     Several relationships between  wind  speed and  particle uptake  rate



are found in the literature.   Bagnold (1941)  suggested that  the  particle



uptake rate is proportional  to the  cube  of  the  wind speed.   Gillette



(1978a), in a wind tunnel test of the effects of sandblasting, wind



speed, soil crusting, and soil surface texture  on  wind erosion,  showed



that the soil particle flux is proportional to  the cube of the  friction



velocity (u ), where u  is determined from  the  mean velocity profile
           *       \  *


over a horizontal surface,

-------
                         u*     z
                    U=     In     >                               (2.1)
                         K      ZQ



where U is wind speed at height z, z0 is the surface  roughness  length,



and k is von Karman's constant (~0.4).   Blackwood  and Wachter (1978)


suggested that the storage pile emission rate,  Q  (mg/s),  may  be



expressed as



                    Q = (cu3 pb2 s°-345)/(PE)2,                   (2.2)



where u is wind speed (m/s), pb is bulk  density (g/cm3),  s  is pile



surface area (cm^), PE = Thorntwaite's precipitation-evaporation  index



(Thornthwaite, 1931), and c is a constant.   Axetell  (1978)  suggested  an


emission factor of 1.6u Ib/acre-hr, where u  is  in  m/s.  Finally,  wind



erosion emissions from active storage piles  may also  be estimated from



the EPA (1983) emission factor



                    EF = 1.9 [S/1.5]  [(365-p)/235] [f/15],         (2.3)



where EF is the TSP emission factor  (kg/day/hectare),  S is  silt content



(% of particles < 75 ym in diameter), p  is  number  of  days with  >  0.25 mm



precipitation per year, and f is the  percentage of time that  the



unobstructed wind speed exceeds 5.4 m/s  at the  mean pile  height.   This



equation is based on sampling emissions  from sand  and gravel  storage


piles and hence gives less accurate estimates when applied  to other


stored materials.  It implies that no emissions occur when  the



unobstructed wind speed is less than  5.4 m/s.


     Field tests with portable, open-floored wind  tunnels indicated that



threshold speeds, given in terms of threshold friction velocity (u )^,



are typically 0.2 to 2 m/s depending  upon the type of material



(Gillette, 1978b; Gillette et al., 1980; and Cowherd  et al.,  1979).   In
                   V

-------
other field tests, threshold speeds  of  about  10 m/s  at  a  height  of  15  cm


above a coal pile surface were estimated based  upon  the onset  of visible


particle uptake (Cowherd, 1982; Cuscino et  al., 1983).  Extrapolating


these speeds to a 10 m reference height from  the  velocity profile


(Equation 2.1) implies that very high mean  wind speeds  (e.g.  20 m/s)


are needed for erosion at the surface (z=0) to  commence.   Hence Cowherd


(1982) suggested that strong wind gusts, not  the  mean wind,  cause


erosion, although it was noted that  wind speeds 15 cm above  a  storage


pile surface may approach the 10 m reference  speed (Soo et al.,  1981).


     In the above relationships for  particle  uptake, emission  rate  is


independent of time.  However, unless an unlimited supply of erodible


particles is present, erosion will be time  dependent.   Erosion rate has


been observed to decrease with time  (e.g. Cowherd et al., 1979).


Cowherd (1982) suggested that erosion rate  is proportional to  the


amount of erodible material remaining and that  a  given  storage pile has


an "erosion potential" equal to the  total quantity of erodible material


present on the surface prior to erosion. The erosion potential  for


coal increased with wind speed.  An  emission  factor was then defined to


be dependent upon the frequency of disturbance  to the pile surface  and


the erosion potential at the expected wind  gust speed.


     A more comprehensive review of  storage pile wind erosion  emission


factors may be found in Currier and  Meal (1984).


     To summarize, particle uptake,  hence storage pile  emission rate,


clearly depends upon the ambient wind speed.   Previous  studies indicate


that uptake is highly sensitive to wind speed,  being related either
                   V

directly or to the cube of the speed, the threshold velocity,  or only

-------
occurring during high wind gusts.  If the wind speed over a given area
of a storage pile is reduced, then fugitive dust emissions from that
area will also be reduced.  The current study isolates the single
variable, wind speed, as the primary factor affecting emission rates
and attempts to evaluate the effectiveness of windbreaks in reducing
particle uptake.  A complete assessment, of course,  would require an
analysis of all factors, including material  type, particle size
distribution, moisture content, etc.

2.2  FLOW ABOUT WINDBREAKS
     Rows of trees or hedges have been used to protect homes and
agricultural fields from high winds for many years (e.g. Van Eimern,
et al., 1964).  Wind tunnel  and field experiments have shown that
windbreaks produce large areas of reduced wind speed in their lee.
Speeds near the surface are  typically reduced to half the upstream
value for a distance of 8 to 12 windbreak heights downstream of a
porous windbreak (e.g.  Raine and Stevenson,  1977).  Wind speeds are
also reduced upstream of a porous windbreak.  Perera (1981) observed
reductions of 50% one windbreak height upstream.
     Figure 2.1 shows typical streamline patterns for flow about solid
and porous windbreaks.   The  streamline patterns are  not drawn to scale;
Caborn (1957) and Naegeli (1953) found that  a "dense" barrier caused a
sheltered region up to distances of 10-15h downwind  and for a (50%)
porous barrier, 20-25h.  The general  flow features are of interest here.
Recirculation regions are evident both upwind and downwind of the solid
windbreak; they are regions  of low velocity  and high turbulence
intensity (Figure 2\l(a)).  Air incident on  the porous windbreak flows

-------
         (a)
         //////// //////////////////Y////7/77/
         (b)
Figure 2.1  Sketches of streamlines about windbreaks:  (a) solid,
(b)  porous.

-------
over and through the windbreak.   A region  of  reduced wind  speed  is
evident in Figure 2.1(b); the upwind recirculation  region  is  eliminated.
In wind tunnel experiments,  the  low wind speed  area moved  downstream  as
porosity (ratio of open to totaf cross-sectional  area)  increased  and, in
general, at a given downwind distance,  wind speed was less with  a lower
porosity windbreak (Raine and Stevenson,  1977).   With the  use of  pulsed-
wire anemometers which detect wind direction, other wind tunnel tests
showed that as porosity increased, the  downwind recirculation region
became smaller and moved downstream (Perera,  1981).
     Raine and Stevenson (1977)  made several  further observations.
Greater wind speed reductions and turbulence  intensity  (u'/U,  where u1
is the r.m.s. longitudinal  velocity fluctuation)  enhancements  were
observed with decreasing porosity.  But since the turbulent fluctuation
at a given location varied much  less with  porosity than did the mean
wind speed, the higher turbulence intensities were due to  the lower U.
The maximum in u1  was located just above and  extended downstream  from
the top of the windbreak.   In the region bounded  by the surface,  the
windbreak, and a line connecting the top of the windbreak  to  the  surface
just downstream of the velocity  minimum (x~8h,  where h  is  windbreak
height), the turbulent fluctuation u1 was  less  than that observed in
the absence of the windbreak, but above and downstream of  this region,
u' was greater, indicating the diffusion of turbulence  generated  in the
windbreak-induced  shear layer.   Greater magnitudes and larger  areas of
wind speed reduction occurred with smoother upstream terrain  and  lower
turbulence in the  approach  flow.
     Perera (1981)valso compared hot-wire  (HWA) and pulsed-wire (PWA)

-------
anemometer measurements downwind of a solid barrier.   Velocity measured



with the HWA was greater than that measured with  the  PWA for  heights



less than about 1.8h at a downwind distance of  5h,  but at higher



heights, the PUA measurements we're greater.  Longitudinal  turbulence



intensity from PWA measurements  were greater than the HWA measurements



at the downwind distance 5h for  heights  up  to 4h, where they  converged.



     Effect of thermal  stability was observed in  a  few field  and wind



tunnel studies.  Seginer (1975a) observed that  during unstable



conditions, wind speed at 2=0.25h beyond x=2.5h decreased with



increasing stability.  Between the porous windbreak and x=2.5h,



stability was not important.   Ogawa and  Diosey  (1980) observed that



the length of the recirculating  zone downstream of  a  solid fence was  a



maximum for near-neutral stability.  It  decreased rapidly as  stability



increased because the stable stratification inhibited changes in the



flow caused by the fence.  For unstable  stratification, the length of



the recirculating zone decreased slowly  with increasing instability.



Jacobs (1984) noted a faster recovery of wind speed with downwind



distance for unstable stratification.



     As expected, the location of the sheltered region shifts as the



wind direction varies from that  of the windbreak  normal (Gandemer,



1981; Seginer, 1975b; Mulhearn and Bradley, 1977).  Measurements along



the windbreak normal showed that the minimum wind speed increases  and



its location moves toward the windbreak  as  the  angle  between  the wind



direction and the windbreak normal increased (Seginer, 1975a,b).





2.3  WINDBREAKS FOR STORAGE-PILE FUGITIVE-DUST CONTROL
                   \


     Very little information on  the use  of  windbreaks as a storage-pile





                                   10

-------
fugitive-dust control  method is contained in the literature.   Low



efficiencies were reported by Bohn et al. (1978) and  Jutze  et al.



(1977), 30% and "very low," respectively; although  these  were only



estimates.  No guidelines on windbreak design and use are available.



Results of two laboratory investigations  have been  reported.   Results



of a water flume test indicated that the  windbreak  height should  be at



least 1.4H, where H is the pile height and that  the lower two-thirds of



the windbreak be 20% porous and the remainder, 50%  porous (Davies,



1980).  However, the author did not state whether more than one



windbreak height was used, nor which porosities  were  tested.   Details



of windbreak length with respect to the pile length and/or  width  and



windbreak position were not given, nor were the  details of  the



atmospheric boundary layer simulated.  Results from a wind  tunnel study



using a two-dimensional  windbreak and pile are reported by  Soo et al.



(1981) and Cai et al.  (1983).   Combinations of two  windbreak  heights



(0.25H and 0.5H), two porosities (solid and 33%  porous),  and five



positions (0, 1, 2, 3, and 4 pile heights from the  pile base) were



tested.  The wind tunnel speed and model  pile size  were chosen to match



the full-scale Reynolds  number (Re=UL/v,  where U and  L are  the



characteristic wind speed and length, and v is kinematic  viscosity),



based on a pile height of 3.05 m and an ambient  wind  speed  of 1.14  m/s.



A 33% porous windbreak of height 0.5H, placed the optimal distance  of



3H from the pile base reduced the wind speed by  50% just  above the



surface at the top of the windward face of the pile (Soo  et al.,  1981



and Cai et al., 1983).  The position of 2.5H was optimal  for a 33%



porous windbreak of height 0.25H (Soo et  al., 1981).   Lower wind  speeds
                                   11

-------
at the top of the windward face  of  the  pile were  observed with  solid



barriers of both heights at all  the positions  (Soo  et  al.,  1981).   In



their study the atmospheric boundary layer was  not  simulated, although



the flow was turbulent.          '
                                   12

-------
              3.  EXPERIMENTAL DESIGN AND INSTRUMENTATION




     Some details of the wind tunnel  boundary  layer,  the model  piles


and windbreaks and the experimental  plan  are given  in the following


subsections.                   '




3.1  EXPERIMENTAL DESIGN


     The experiment was conducted in  the  EPA Fluid  Modeling  Facility


(FMF) Meteorological Wind Tunnel  (MWT), a low-speed,  open-return  tunnel


having a test section 2.1 m high  x 3.7 m  wide  x  18.3  m long.   A complete


description of the wind tunnel  and operating characteristics  can  be


found in Snyder (1979).  Simulating the lower  part  of the atmospheric


boundary layer (ABL) and determining  model  size  and free-stream wind


speed are necessary to assure that wind tunnel  results will  be valid


for the full-scale case under investigation.   A  neutrally stratified


simulated ABL was generated using a  153 mm high  trip  fence placed near


the test section entrance.  Gravel roughness composed of pebbles  having


typical  diameters of 10 mm covered the tunnel  floor downstream of the


fence.  The tunnel ceiling height was adjusted to achieve a  zero


longitudinal pressure gradient  in the free stream.   The velocity  profile


in the surface layer is described by  Equation  2.1 and is shown in


Figure 3.1.  The tunnel free-stream speed was  4  m/s.   Roughness length


(z0) and friction velocity (u ) were  determined  from  the log-wind law.


The boundary layer was characterized  by a depth  of  approximately  1 m,  a


z0 of 0.12 mm, and a u  of 0.048U0, where U0 is  the free-stream speed.


Further details of the boundary layer are given  by  Castro and Snyder


(1982).
                   \

     Model size and free-stream wind  speed should be  determined from




                                   13

-------
o
o
o
o
o
                          (UJUl)  2
                                                                       o

                                                                       E
                                                                       en
                                                                       o
                                                                       01
                                                                       c
                                                                       O)
i-T-i 	 1 i 1 I I | 1 1 "I I 1 I 	 1 i ' i i i i i i i i i-r-t-r-r-r . r rf -, 	 r , . •. ,

1 '"\
'-  *~
>
^ _ ^ ^
• c «o
° T
• • cr •-
1 "i
^ -5 =
• * i 5
o
0 <*>
^. 4>
l/> — — f
en
u_
                            Ift
                                       cri     ru
                (LPUoisuaujip-uou)  9/2
                                  14

-------
matching the model and full-scale Reynolds numbers.   However,  with



typical scale reductions, the model  Re is typically  3 to 4 orders of



magnitude less than the full-scale Re.  Very high wind speeds  in the



tunnel are generally required to-match full-scale Re, but are  often



impractical.  Fortunately for wind tunnel modeling,  "geometrically



similar flows are similar at all  sufficiently high Reynolds numbers"



(Townsend, 1956).  That is, for Re greater than some minimum value, the



turbulent flow structure described in terms of characteristic  length



and velocity scales is independent of Re.  Since atmospheric flows are



almost always aerodynamically rough (for all  wind speeds), they  are



also Re-independent.  Hence wind  tunnel  velocities,  normalized by a



reference speed, are equivalent to normalized full-scale values.



     No two storage piles have the same  shape and size,  and active



piles have constantly changing dimensions.  For purposes of the  present



study, windbreak effects on two typical, but idealized,  pile geometries



are studied, with the results being generally applicable to many full-



scale piles.  Based on a survey of several coal  piles at electric



generating plants, typical  piles  are 11  m high, have slopes of 37°



and range from conical  to flat-topped in shape.   The two piles modeled



were an 11 m high conical pile with 37°  slopes (base diameter  29.2 m)



and an oval, flat-topped pile of  height  11 m, 37° slopes, and  base



dimensions 63 m by 78 m (Figure 3.2).  The larger pile is an elongated



frustum.  It is the lower part of a cone of base diameter 63.0 m and



height 23.7 m, which is cut-off at 11.0  m, with a 15.0 m extension



placed between the two halves of  the frustum.



     The model piles had to be small  enough to be within the surface
                                   15

-------
                             OJ
                             •c
                             OJ
                             c.
                             C-
                             c
                             4->
                              I
                             o

                             l«~
                             o

                             .c
                             o
                             4->
                             01
                             _i*
                             to


                             CM
                              •
                             CO

                             01
                             t_
                             3
                             CT>
16

-------
layer but large enough to facilitate measurements  and construction  of
windbreaks of height the same order as the pile height.   A full-scale
neutrally stratified ABL is typically 600 m high and  has  a surface
layer depth of 100 m (Counihan, -1975).  Since  typical  pile and  windbreak
heights are well  within the surface layer, matching the  ratio of  the
pile height to the boundary layer depth was not considered important.
Instead, Jensen's (1958) criterion of matching the ratio  of the model
height to the surface roughness length (z0) was used  to obtain  the
relevant model length scale.  Model  piles 110  mm high (1:100 scale)
resulted (Re=2.9 x 10 , based on UQ and the pile height).   The  model
surfaces were roughened to make them aerodynamically  rough with
roughness elements of size greater than approximately 400v/U0 (roughness
Reynolds number criterion) (Snyder,  1981).  The pile  could not  be
roughened with the same gravel  as that covering the floor of the  tunnel
because the 10 mm gravel was considered too large  for the pile  size.
Gravel having diameter less than 4 mm was used instead.   With a free-
stream speed of 4 m/s, the roughness Reynolds  number  criterion  discussed
above was met with the 4 mm gravel.   The terrain surrounding actual
piles is also usually rougher than the pile surface.
     Finally, the windbreak material  was chosen from  50%,  60%,  and  70%
porous synthetic material  that is commercially available  for use  in
windbreaks.  The openings of 25 mm x 12 mm and 25  mm  x 25 mm in the 60%
and 70% porous screens were considered too large for  the  overall
windbreak height (110mm) for direct use in the wind tunnel.   Hence, a
more uniform windbreak material  having the same aerodynamic drag  or
pressure-drop coefficient as the more porous commercial  screens was
                                   17

-------
desired.  Caput el  al.  (1973)  used the pressure-drop  coefficient  cp


to account for the porosity and size and shape  of  the windbreak


openings.  The pressure drop coefficient is  defined as


                    cp = AP/(0.5pU2),                              (3.1)


where AP is the static pressure drop across  the material,  p  is air


density, and U is reference wind speed, here, the  wind speed upstream


of the windbreak.


     The pressure drop coefficients of the three windbreak materials and


a 16 x 18 mesh nylon screen (65% porous) were determined  through  tests


in a smaller wind tunnel  (1 m  x 1 m x  4 m).  The material  was placed 1 m


from the entrance to the test  section  and fully covered the  cross


section.  The pressure drop was measured with pitot-static tubes  placed


upstream and downstream of the material, both connected to an MKS


Baratron (Type 170M-6B) capacitance manometer.   The output was digitized


at a rate of 200 Hz over the 60 s sampling time, then processed  on the


EPA FMF Digital Equipment Corporation  (DEC)  PDP-11/40 minicomputer.


The pressure drop coefficients for the 50%,  60%, and  70%  porous


materials and the mesh screen  were found to  be  5.5, 2.2,  1.4, and 1.8,


respectively, independent of wind speed for  U > 2  m/s. The  mesh  screen


was therefore used to represent the high porosity  material as its Cp


was midway between that of the 60% and 70% screens.   Note that cp for


the 50% porous windbreak material is nearly  three  times that for the


mesh screen.


     Given the boundary layer, the model piles, and windbreak materials,


the experimental procedure may be described.  It consisted of three

                    \
major parts:
                                   18

-------
          (1)  Measurement of surface wind speed patterns  on a conical
storage pile with and without a windbreak located upwind,
          (2)  Measurement of surface wind speed patterns  on an oval,
flat-topped pile with and without a windbreak  located upwind or on the
top of the pile, and
          (3)  Measurement of vertical  profiles  of velocity  downstream
of a porous windbreak over horizontal  terrain.
     The windbreak and conical  pile set-up is  shown in Figure 3.3.
Each windbreak had vertical  metal supports at  both ends, which were
tacked to the floor and to which guy wires were  attached.  The windbreak
material was folded underneath the gravel-covered plywood  sheets on the
tunnel floor.  Windbreaks of 50% and 65% porosity; heights 0.5H, l.OH,
and 1.5H, where H is the pile height; lengths  l.OD and 1.5D, where D is
pile base diameter; and positions 1H and 3H from the upstream pile base
were tested.  All combinations of the parameters result in six
windbreaks of each porosity  placed at two different distances from the
conical pile, i.e. a total of 24 cases (Table  3.1).  Upon  completion of
these tests, one windbreak was placed at angles  of 20° and 40° from the
position normal  to the incident flow.
     For the oval, flat-topped pile, one of the  windbreak  orientations
was similar to that shown in Figure 3.3; the longer axis of  the pile
being parallel to the windbreak.  The same windbreak porosities and
relative sizes and positions were used (the windbreak length is given
in terms of the pile base length (B))  except that not all  of the 24
cases were tested.  A few other relative sizes were tested:  heights
0.75H and 1.25H, and length  0.6B (length of flat top, see  Figure 3.2).
                                   19

-------
                                       QJ
                                       c
                                      •r-
                                       a

                                      T3
                                       C
                                       fO

                                       
-------
TABLE 3.1  COMBINATIONS OF WINDBREAK PARAMETERS  USED  WITH  THE CONICAL
                                  PILE
porosity (%)
50
50
50
50
50*
50
50
50
50
50
50
50
65
65
65
65
65
65
65
65
65
65
65
65
height (h/H)
0.5
0.5
0.5
0.5
1.0
1.0
1.0
1.0
1.5
1.5
1.5
1.5
0.5
0.5
0.5
0.5
1.0
1.0
1.0
1.0
1.5
1.5
1.5
1.5
length (L/D)
1.0
1.0
1.5
1.5
1.0
1.0
1.5
1.5
1.0
1.0
1.5
1.5
1.0
1.0
1.5
1.5
1.0
1.0
1.5
1.5
1.0
1.0
1.5
1.5
position (P/H)
1
3
1
3
1
3
1
3
1
3
1
3
1
3
1
3
1
3
1
3
1
3
1
3
        *This windbreak  was  also  placed  20°  and  40°  from  the
         position normal  to  the incident flow.
                                   21

-------
The other windbreak orientation  was  to  place  a windbreak on the pile
top to reduce wind speeds in active  regions of the  pile.  The windbreaks
were placed either close to the  centerline or at the upstream edge of
the pile top parallel  to the pile's  longer axis.  Two  heights, 0.14H
                                i
and 0.27H, and two lengths, 0.16T  and 0.5T, where T is the pile top
length, were tested.  For windbreaks in both  positions the pile was
rotated 20° and 40° to simulate  other wind directions.  The
specifications for all the windbreaks used with the oval, flat-topped
pile are given in Table 3.2.
     The final phase of the project  was to measure  velocity downstream
of the less porous windbreak oriented normal  to the air flow to
determine whether reverse flow was present and to relate the flow behind
the windbreak to that over a pile  placed in the lee of the windbreak.
The windbreak chosen was 50% porous, 112 mm high and 1180 mm long
(aspect ratio 10.5).  Vertical profiles of velocity were measured with
hot-wire and pulsed-wire anemometers along its centerline at downstream
distances of 1, 5, 8, and 12 windbreak  heights.

3.2 INSTRUMENTATION
     Velocity profiles were measured at selected locations over the pile
surface and flat tunnel floor with hot-wire and pulsed-wire anemometers
and wind speeds were measured at a number of  points over and close to
the pile surface with fixed thermistors. Hot-wire  and pulsed-wire
probes were attached to the MWT  instrument carraige which can position
the probe to within ±1 mm and may  be operated by remote control outside
the tunnel.  The Cartesian coordinate system  is oriented such that the
x, y, and z axes are longitudinal  (positive downstream), lateral  and
                                   22

-------
TABLE 3.2  COMBINATIONS OF WINDBREAK PARAMETERS USED WITH THE OVAL,
                            FLAT-TOPPED PILE
        porosity (%)  height (h/H)  length (L/B)   position (P/HJ

             50            0.5  '         0.6          1
             50            0.5           0.6          3
             50            0.5           1.0          1
             50            0.5           1.0          3
             50            0.5           1.5          1
             50            0.75          1.0          1
             50            0.75          1.0          3
             50            1.0           0.6          1
             50            1.0           0.6          3
             50            1.0           1.0          1
             50            1.0           1.0          3
             50            1.0           1.5          1
             50            1.25          1.0          1
             50            1.25          1.0          3
             50            1.5           0.6          1
             50            1.5           0.6          3
             50            1.5           1.0          1
             50            1.5           1.0          3
             50            1.5           1.5          1

             65            0.5           0.6          1
             65            0.5           0.6          3
             65            0.5           1.0          1
             65            1.0           0.6          1
             65            1.0           0.6          3
             65            1.0           1.0          1
             65           -1.0           1.0          3
             65            1.5           0.6          1
             65            1.5           0.6          3
             65            1.5           1.0          1
length (L/T) location
65*
65*
65
65
65
0.14
0.14
0.27
0.27
0.27
0.5
0.5
0.16
0.5
0.5
upstream edge
center! ine
upstream edge
upstream edge
center! ine
        *These windbreaks  were also placed  20°  and  40°  from the
         position  normal to the incident  flow.
                                   23

-------
vertical, respectively.   The corresponding  velocity  components  are  u,
v, and w.  Theory and operation  of the three  anemometers will be  briefly
discussed.
     Hot-wire anemometers are commonly used in  turbulent, wind  tunnel
flows.  A TSI, Inc. constant temperature hot-wire  anemometer  (model
1053B) with a TSI boundary-layer cross-wire probe  (model 1243)  was  used
to obtain vertical profiles in the undisturbed  (no pile, no windbreak)
boundary layer and downstream of a windbreak.   Mean  velocity, angle of
flow, r.m.s. turbulence velocities and turbulence  intensities in  two
directions (longitudinal  and vertical) and  Reynolds  stress were
obtained.  Yaw response corrections developed by Lawson and Britter
(1983) were applied to the turbulence intensity and  stress measurements.
The anemometer has been extensively used by other  investigators at  the
EPA FMF (e.g., Castro and Snyder, 1982; Pendergrass  and Arya, 1984).
The probe was calibrated in the free stream against  a  standard  pitot-
static tube each day.  The hot-wire output  was  fit to  a King's  law  form
by the hot-wire calibration routine CALLPA:
                £2 = aUa + b,                                     (3.2)
where E is output voltage, U is wind speed, and a, b and « are
constants.  CALLPA gives the best-fit a, b, and a.  An example  of the
computer output is shown in Figure 3.4.  The  hot-wire  output  was
digitized to 12 bit precision and processed at  a  rate  of 500  Hz on  the
PDP-11/40 with the program HOT.   90 s samples provided reasonably
repeatable results.
     The pulsed-wire anemometer is used to  measure mean and  fluctuating'
velocities in regions where turbulence intensity  is  very high or  flow
                                   24

-------
   ,_"     I    I    I  I



     £!  O«nc!piff»M §
         — L'I t\i —. -«f\t *
     .-*.  OO —— OO LJ
     •^   ...... Jg-
   ...fj  oooooo o
   ft.     .  '    '      rj
   t-                 CO

   =^_  (Db'.cri —UT— d
   0-
   o
   _1
   _



 J  I
   coZ
 ^^x

 CCli.
 X_l
 USX
                                                                                                                         o

                                                                                                                         Q.
                                                                                                                            (O
                                                                                                                            C-
                                                                                                                            ai
                                                                                                                           4-»
                                                                                                                            o
                                                                                                                            E
                                                                                                                            OJ
                                                                                                                            c
     _

  (•'• '  ut O i
    '-  -
  O OOOOOO o


•*                 "
^~ J0b~m—.LT—> d

•^cr »«u*>cD^.(ndi ^
                   r>
                   



 ^                 5
 '— l-OOujOO=»0> >.

  a -woo-eo:, Q
H°   wtc-o —10 —
till   ui=»r-uii-
So.     'tl  '  ' ''
       — — — — -.—

                                                                                                                           4J
                                                                                                                            O
                                                                                                                            a»

                                                                                                                            Q.
                                                                                                                            E
                                                                                                                           **
                                                                                                                             •

                                                                                                                           CO
                                                                                                                            a>
                                                                                                                            cr>
                                                            25

-------
reversal occurs.   Measurements  with  the  pulsed-wire anemometer were
taken downstream of a windbreak since  flow visualization indicated
intermittent flow reversal.
     A pulsed-wire probe  consists  of a pulsed wire normal to and between
two parallel sensor wires (Figure  3.5).  The pulsed wire is activated
by a voltage pulse, momentarily raising  the pulsed-wire temperature to
several  hundred degrees Celcius and  heating the air surrounding the
wire.  The heat,  acting as a  tracer, is  advected by the wind with the
velocity occurring at that instant.  Depending on the  instantaneous
flow direction, one of the two  sensor wires (operated  as resistance
thermometers) senses the tracer.   Thus,  the basic measurement is of the
time-of-flight of the heat tracer  from the pulsed wire to either of the
sensor wires.  Ideally, the time-of-flight t is
                    t = d/(Ucosy),                                (3.3)
where d is the distance between the  pulsed wire and sensor wire, U is
the magnitude of the velocity vector and Y is the angle between the
direction normal  to the probe plane  and  the instantaneous velocity
vector (Figure 3.5).  The probe plane  is the plane parallel to all
three wires.  In other words, given  the  separation distance between the
wires, the velocity component normal to  the probe plane is calculated
from the time-of-flight.   If  the velocity at the time  of pulsing is
close to zero or if the angle of flow  (Y) is greater than about 70°
(based on the sensor wire length and d), the heat tracer will not be
sensed at the sensor wires.   In such  cases, the velocity is recorded as
zero.  For operation in turbulent  flow the wire is pulsed many times
per second over a ^given time  period  to obtain repeatable mean and
                                   26

-------
                                           c
                                           i_
                                           CL
                                           o

                                           o
                                           O>
                                           c
                                            i
                                          •C
                                           QJ
                                           3
                                           CL
                                          00


                                          in
                                            •
                                          op

                                           OJ
                                           (_
                                           3
                                           O
27

-------
r.m.s. fluctuating velocity  components.
     The basic electronic components,  in  addition to  the  sensor wires,
are a differential amplifier,  a differentiator,  and two comparators.
The polarity of the output from the differential amplifier  indicates
which sensor wire received the tracer.  The  signal is then
differentiated to provide a  more clear indication of  the  time  the  heat
tracer was sensed since the  tracer diffuses  as  it is  advected.  One
comparator is set to be triggered by a positive sensor wire signal
(positive velocity) and the  other by a negative signal  (negative
velocity).  A 12 bit binary  counter begins when the pulse is fired and
stops when either of the comparators is triggered.  Hence the  time-of-
flight is measured and the flow direction is known.   The  other major
electronic component is a circuit to eliminate  the spike  on the sensor
wire circuit which occurs when the pulse  is  fired.  The sensor signals
at various stages of the electrical  operation may be  observed  on an
oscilloscope.  Typical signals are shown  in  Figure 3.6; the signal with
the spike, the signal  less spike, the differentiated  signal, and the
time-of-flight signal.  The  time-of-flight  signal indicates when the
counter stops.  Further details on the theory of the  pulsed-wire
anemometer are found in Bradbury and Castro  (1971).
     A PELA Flow Instruments,  Ltd. pulsed-wire  anemometer was  calibrated
in the tunnel free stream against a standard pitot-static tube.  The
probe plane was oriented perpendicular to the velocity component of
interest, U.  The FMF pulsed-wire calibration computer program PWCAL
fits the wind speed, U, and  time-of-flight,  t,  data to the  relation
                 '*  U = C/t  + F/t2                                (3.4)
                                   28

-------
 00
•O
 lU
 4J
 C
 (1)
 H
 0)
 -o
 x-x
 O
(0
c
oo
•rH
co
 o
 I
UJ
 O

 0)
 S
                                                                                 c_
                                                                                 Ol
O

                                                                                  in
                                                                                  O

                                                                                  in
                   c
                   CD
                   o
                  •r-
                   Q.

                  ^


                  VO
                    *
                  CO

                   O)
                   i_
                   3
                   Ol
                 29

-------
where C and F are constants.   An  example of  the  fit  Is  shown  in  Figure
3.7.  After one sensor wire was calibrated,  the  probe was  rotated  180°
to calibrate the other sensor wire.   Data were collected and  processed
by the program PWAHOT on the FMF' DEC PDP-11/44.   A sampling rate of
20 Hz over 3 minutes gave reasonably repeatable  results.
     The last instrument to be described is  the  thermistor which was
used to measure wind speeds near the pile surface at various  locations.
We have considered wind speed near the surface as the basic parameter
affecting particle uptake.   Unfortunately, it is difficult to measure
the wind speed near a roughened,  sloping surface, particularly when the
wind direction at the measurement point is unknown and  difficult to
determine.  Pitot-static tubes and hot-wire  anemometers require  certain
orientation with respect to the wind for accurate results.  Heated
thermistor beads were used  here for several  reasons.  First,  several
thermistors could be mounted on the pile surface, eliminating the
tedious task of moving a hot-wire from location  to location,  orienting
it properly and placing it  the same distance from the surface for  each
sample.  Once the thermistor was  fixed at the pile surface, errors
resulting from differences  of thermistor orientation and height, and
effects of local roughness  elements were minimized by comparing  wind
speed at a specific point on the pile for the cases  with and  without a
windbreak.
     The resistance-temperature relationship for thermistors  is
expressed as
                    R= R'exp{3[(T)-l - (I')-*]}.                 (3.5)
where R and R' are the resistances at temperatures T and T1,
                                   30

-------
                   TRIG LEVELS* 1. I.  Srf
                         U =  17£H.:)5/T
                                                             5.
l/T
0.00:53
0. 00 • 4?
0.10:33
0.00112
0.00095
0.00079
o.oocsa
0.00034

U IME"i
U.-fQ
3 . fi ?Ci
3 ^22
2! 850
2.310
1.830
1.27C
0.630
SSE
L 1C--
4 . HS3
H 3" ^
3i-i23
2.323
2.3:u
1.6i6
1.237
0.57J
= O.OC'499
L " - V ' 1' .
-C". 372
C.6G5
C.2S5

-C. C<2!
3. 53o
-2.553
6.309
S =0.9
     v.s
              .0002
                     .OOO't
                           .QOC5   .OD09   .00!

                            l/T C1/KICR25EC2
                                                              .0015
Figure 3.7   Example pulsed-wire  anemometer  calibration  plot.
                                31

-------
respectively, and 6 is a constant  dependent  on  the  thermistor material.
All  the thermistors (Fenwal  GB31J1)  were  calibrated in  a  hot  oil  bath-.
Calibration curves for two of  those  calibrated  and  the  Fenwal rated
curve, R(25°C) = 1000 ± 200 n, & = 3442 ± 90 K, are shown in  Figure  3.8.
Fenwal specified the thermistor time constant of 4  s in still air at
25°C.  The thermistor diameters were about 1 mm and their lengths were
about 1.5 mm.
     The electric circuit consisted  of  a  regulated  power  supply  of
constant voltage E and thermistor-resistor pairs in parallel  with the
power supply (Figure 3.9).  The voltage across  each series resistor  is
the output voltage (V-j).  Voltages V-j and E  were digitized at a  rate of
50 Hz and processed on the POP 11/40 minicomputer with  the computer
program THE.  A one minute sampling  tine  gave reasonably  repeatable
results.
     Thermistor anemometers operate  under the same  basic  principle as do
hot-wire anemometers, that is, the heat loss from the sensor  is  a
function of wind speed.  Rasmussen (1962) has shown that
                  i2R = K(TT - TJ,                                (3.6)
where i is current through the thermistor, R is thermistor resistance,
Tj and Ta are thermistor and ambient temperature, respectively,  and  K
is the dissipation factor.  K is a function  of  the  wind speed and the
properties of the fluid surrounding  the thermistor, and was determined
experimentally as follows.  The thermistor was  placed in  the  center  of
the smaller wind tunnel with the probe  support  body and the thermistor
oriented vertically (the same orientation with  respect  to the air flow
as for a thermistor1 mounted on the model  pile).  At each  wind speed, E
                                   32

-------
     10
a
z •
o
z
UJ
o
 I
z
o
 o
oc
     .1  -
    .01
                    Rated calibration
                    Thermistor  #1 calibration
                    Thermistor  #2 calibration
R(25<>C)   __§__
 1000  0   3442 K
  883  n   3509 K
 1193  0   3495 K
      .0025 .0026 .0027  .0023  .0029  .003 ,0:31 .0032 .0033 .:C3^  .0035 .CC3j

                        1/TEMPERfiTUPE  (K-'-l
  Figure 3.8   Resistance-temperature thermistor curves.
                                 33

-------
           _C
       L_   *"**
                  Figure 3.9  Thermistor circuit
      O
      o
            .1
                                    K(U) = 0.97 ua27
                                                   t __!_ . t. II
     1
U (m/s)
10
Figure 3.10  Dissipation factor vs.  wind speed for Fenwal GB31J1
thermistors.  Triangles: original data.  Squares and circles:
additional  data>
                              34

-------
and V-j were measured.  From those, i, R and Tj were calculated.  Ta was
measured with a YSI model 4320 temperature sensor.   Hence, K at each
wind speed was calculated.  By varying the wind speed from 0.3 m/s to
8 m/s, the functional relationship of K (mW/°C) to  u (m/s) was found to
be (Figure 3.10)
                    K = 0.97u0.27.                                (3.7)
Hence, thermistor output is related to wind speed.   It was originally
assumed that this relationship was valid for all  the thermistors used
in this project.  Later, two other thermistors were similarly
calibrated.  The curves of thermistor output (K)  vs. tunnel  wind speed
were similar to that of the original curve (refer to Figure 3.10); the
maximum differences in wind speed were ±10%.  The original calibration
tended to slightly overestimate wind speed as compared to the data from
the other thermistors.
     Tests of thermistor sensitivity to orientation were also conducted
in the smaller wind tunnel.  With the probe and support in the vertical
position, two tests were conducted.  First, the probe was rotated about
the vertical  axis, sampling the output for several  orientations at a
given wind speed.  The maximum difference in measured wind speed from
the tunnel speed was 5%.  Second, the wind speed  calibration was made
with the probe oriented vertically as described above, but rotated 90°,
to observe effects of the leads exposed between the thermistor and the
support.  Measured speeds were within the scatter of the calibrations
for the probe in the original  orientation.  Hence,  orientation about
the vertical  axis was not significant.  Yaw sensitivity was tested by
orienting the probe* and support in the horizontal,  normal to the air
                                   35

-------
flow, and rotating the support  in the xy-plane  to  angles up to  ±30°.
The maximum difference in measured wind speed from the tunnel speed was
8%.
     The thermistors were mounted normal  to  and about 2 to 3 mm  above
the pile surfaces; close enough to the surface  that the flow presumably
parallels the surface, and where wind speed  may be assumed to be
directly related to the surface shear stress.   Nine thermistors  were
mounted at different elevations on the conical  pile in the arrangement
shown in Figure 3.11.  81 thermistors were mounted on the oval,  flat-
topped pile in the arrangement  shown in Figure  3.12.
     Thermistor and hot-wire anemometer (TSI Model  1210, end-flow
single-wire) measurements were  compared to further substantiate
thermistor use as an anemometer and the use  of  Eq.  3.7.  For the pile
oriented in the position shown  in Figure 3.11 with air flow from the
left, wind speeds were measured by the middle thermistor on the  windward
side of the pile.  Wind speeds  were calculated  with the calibration
equation (Eq. 3.7) for several  wind tunnel tachometer settings.   The
pile was rotated approximately  45° and the hot-wire was placed at the
same relative position as the thermistor. Wind speeds were measured at
the same tachometer settings.  The results showed  good agreement (Figure
3.13).  However, slight differences in height above the surface,
individual roughness elements,  and thermistor and  hot-wire orientation
could cause larger differences  in the velocities measured by the two
probes.  The same may be expected for velocity  measurements using
different thermistor probes in  apparently similar  settings.
     As discussed above, the maximum error in wind speed due to
                                   36

-------
Figure 3.11  Top view of conical
pile.  Stars: Thermistor positions
on pile.  Dots: Effective thermistor
positions due to pile rotation.
Figure 3.12  Top view of oval,
flat-topped pile.  Dots: Thermistor
positions.
37

-------
     3.5 ••
±    2.5 -
LU
CC
 I
LU
O
~    1.5 +
tn
                     1     1.5     2     2.5     3     3.5

                        U,  THERMISTOR  W/S:
 Figure 3.13  Comparison of thermistor and single-wire wind speed
 measurements.
                                38

-------
thermistor orientation was ±8%.   Possible systematic  errors  resulting
from the application of the calibration curve  (Eq.  3.7)  to all  the
thermistors could have been eliminated if each  thermistor had  been
calibrated; however, due to time constraints, each  probe could not  be
calibrated.  Effects of individual  roughness elements and thermistor
height above the surface were other possible sources  of  systematic
error.  For the purposes of the  present study,  wind speed was  used  more
as a relative, than absolute, measure, i.e. wind speed with  and without
a windbreak or wind speed with one  windbreak and that with another  were
compared.  The error indicated above is for absolute  wind speed, but
relative wind speed was of more  importance here; its  error was expected
to be even less.
     In addition, thermistors cannot detect flow reversal, i.e. they
respond to wind speed and not wind  velocity.   If reverse flow  were
steady, the magnitude of the measured wind speed would be fairly
accurate.  On the lee side of the piles, with unsteady flow, the
measured average mean wind speed will  be higher than  the actual  (vector-
averaged) speed.
     Wind speed distributions were  measured on  the  piles in  the absence
of any windbreak at least once every three days and frequently once a
day to determine system repeatability.  The peak-to-peak difference at
a given position was ±10%, an r.m.s. error of ±2%,  indicating  quite
good repeatability.
     For the conical pile, each  run consisted of measuring the wind
speed with the nine thermistors, rotating the pile  30°,  measuring the
wind speed at thesexnine positions, etc., through 360°,  resulting in
                                   39

-------
108 data points per run (refer  to Figure  3.11).   For  the  oval,  flat-
topped pile, each run consisted of measuring  the  wind speed  at  the  81
positions shown in Figure 3.12.
                                   40

-------
                   4.  FLOW ABOUT A POROUS WINDBREAK


     Mean and fluctuating longitudinal  velocity components were measured

downstream of a representative windbreak (50% porous,  height  (h)=112 mm,
                                t
aspect ratio=10.5) with hot-wire (HWA)  and pulsed-wire (PWA)  anemometers

to quantitatively describe the flow structure in the sheltered area, to

determine if flow reversal occurred, and to further  explain results  of

the major part of the project.  Due to  the finite size of the windbreak,

the flow was expected to be three-dimensional.   Vertical  profiles  were

taken at distances of 1, 5, 8, and 12h  from the windbreak.   The results

from the two anemometers may also be compared.

     "Streamlines" for two-dimensional, steady  flow  were  obtained  from

the FMF computer program STRFUN which calculates the height at which a

given stream function value (fy) occurred (41 = /udz).   Since the flow

was likely three-dimensional, the calculated "streamlines" are not

necessarily the true streamlines;  knowledge of  the three  velocity

components are required to calculate the true streamlines.   However,

along the centerline not too far from and close to the windbreak,  the

flow is expected to be nearly two-dimensional.   "Streamlines" derived

from the pulsed-wire data are shown in  Figure 4.1.  The PWA data was

used because it was expected to be more accurate since the PWA senses

flow reversal.  Note the similarity to  Figure 2.1(b).   A  region of low

wind speed near the surface between 2h  and 8h from the windbreak is

evident, with the air flowing up and over this  region.  Flow

visualization indicated intermittent flow reversal  (unsteady  flow) near

the surface in this region.  Effects of the windbreak  are noticeable up
                   •\
to z=4h, with upward deflection apparent as far downstream as x=6h.


                                   41

-------
   in   in   in   in
                     in
        in
        11
t   1-
    t
        \ \\\   \ \
                           in
                           in
                           en
                           in
                            •

                           °
- -en





  in


  r-





  UD





  in

  =r





- -co





  LD
   •






- -o




  in
   •
  T-i
  I
                             JC


                             X
in in
       in rn in oj in
        •     •     •
       CO    C\l    *—«


            M/z
                     in o
                   •o
                   C
                   3
                   O
                                   O
                                   CL
o

E

O)

+J

c


-a

in
i_
o


o>
                                   o
                                   o
                                   -o
                                   OJ
                                   ISI
                                   •o
                                   c
                                   CO
                                   O)
                   H3
                   (U


                   4J

                   CO
                                   
-------
     For steady flows, velocity vectors are parallel  to streamlines.



Since the flow angle in the xz-plane was given by the hot-wire output,



the normalized velocity vectors (U/llR(h)) are also shown in Figure 4.1



(UR(h) is the wind speed at the Windbreak height, but in its absence).



The flow generally parallels the "streamlines" except in the region



below the windbreak height at x=lh, suggesting that the flow is three-



dimensional, i.e.  that some flow in the lateral  direction exists,  even



along the center!ine.  Vortices may be formed as the air flows round



the sides of the windbreak.  The hot-wire observed angles are not



nearly as large as those implied by the "streamlines" (derived from PWA



data) because of differences in wind speed profiles measured by the two



instruments.  The angles clearly show upward flow at x=lh and downward



flow at x=8h.



     Relative wind speed deficit may be defined  as [l)R(z)-U(z)3/UR(z)



where UR(Z) is the reference speed (measured with the HWA) at the



location of the windbreak but in its absence and U(z) is a speed



(measured with the PWA) at some distance downstream of the windbreak.



Again, PWA data are presented as they are expected to be more accurate.



Lines of constant  relative deficit are seen in Figure 4.2.  Downstream



distance and height are scaled by the windbreak  height h.  A large



region of wind speed reduction was observed downstream of the windbreak.



The height of the  region increased from z=1.4h at x=lh to z=3h at



x=12h.  Below z=lh, wind speeds were reduced at  least 50% from the



upstream value at  the same height.  The greatest deficits were observed



for heights less than z=0.5h between approximately 4 and 8h downstream.



In other words, the maximum deficit did not occur immediately downstream
                                   43

-------
                                                             
                                                             c

                                                             o
                                                             O)
                                                            •a
                                                             
                                                            0)
                                                            Qi
                                                            CM
                                                            
-------
of the windbreak, but occured farther downstream.   A strong gradient is



clearly evident just above and downstream of the windbreak.



     Lines of constant turbulence intensity are shown in Figure 4.3,



where turbulence intensity is the ratio of the r.m.s. fluctuating



longitudinal  velocity (u1, measured with the PWA)  at a given location



to the mean wind speed at that location.  The maximum was observed near



x=8h below z=0.5h.  Intensities of greater than 30%  were observed for



heights less than 1.5h between x=2 and 12h and at  x=lh just above the



windbreak height.  A pulsed-wire must be used here because with



intensities greater than 25-30% the errors with hot-wire data become



large.



     The fluctuating velocity component may also be described in



absolute terms.  Figure 4.4 again shows the distribution of u', but



normalized by UR(|I).  The presence of a maximum between z^l.ZBh and



1.5h suggests that greater turbulence is generated in the shear layer



separating from the top of the windbreak.   The turbulence diffuses with



downstream distance.  The normalized r.m.s. velocity fluctuations were



quite low (5-7%) just downstream of the windbreak.



     Figure 4.5 shows vertical  profiles of the longitudinal  mean



velocity component  (measured with both anemometers) normalized by UR(h)



as a function of downwind distance.  These are compared with the



reference wind speed profile.  Comparison  of the HWA and PUA results



show significant differences only in the region of high turbulence



intensity.  The HWA tended to overestimate the wind speed in the high



turbulence region.  No flow reversals in the mean  occurred, but



instantaneous reversals were indicated by  the PWA  (when the upwind
                                   45

-------
                                                                            O)
                                                                           TO
                                                                            c
                                                                            O
                                                                            c_
                                                                            o
                                                                            Q.

                                                                            <0
                                                                            o

                                                                            e

                                                                            OJ

                                                                           +J

                                                                            c

                                                                            o
                                                                            in
                                                                            c
                                                                            OJ
                                                                           Ol
                                                                           u
                                                                           Ol
                                                                           OJ
tn
                                                           o
                 C\J


                M/2
                         46

-------
in
in
  •

=r
in
  •
m
                                    ro
in
 •

nj
                                                               in
                                                      f\j
                                                                                 in
                                                                                 o


                                                                                iH
                                                                                               in
                                                                                                          O

                                                                                                          O
                                                                                                          
                                                                                                         -o
                                                                                                          3
                                                                                                            -o
                                                                                                             c
                                                                                                          CT>

                                                                                                          C

                                                                                                          O
                                                                                                             o
                                                                                                             a.
                                                                                                          E
                                                                                                           •
                                                                                                          t_


                                                                                                         •a
                                                                                                          OJ
                                                                                                          rsi
                                                                                                         ro
                                                                                                         E
                                                                                                    o



                                                                                                    03
                                                                                                      3
                                                                                                          t- 4J.

                                                                                                          3  O

                                                                                                          O> 3
                                                                                                         •t— r— »

                                                                                                         U- >»-
                                                   47

-------
       in
       co
        -S	3-
o
in

C\J      CM
                                  in
                                                tn
                            Q
                                         8
              _rn	  n
                    JUL
                                         G
                                             -
§§^
             in



             in





             CM



             LD
              •
           - -o




           --en
                                                    Ai
                                                   B-T
                                                   ^=^'v!_-
                                                          in
                                                          in
                                                                       O)
                                                                    
                                                Ol OJ
                                                «- E
                                                4-» O
                                                tn E
                                                C QJ
                                                S c
                                                O (O
                                                                    •o t.
                                                                    O" -r-
                                                                    o> 3:
                                                                    CL i
                                                                    «/) ~o
                                                                       Ol
                                                                    "O 
                                                M QJ
                                               •r™ ^^
                                               •— o
                                                •- OI
                                                o ^~
                                                                    QJ O
                                                                    r— U
                                                                    H— Q.
                                                                    l«-    .
                                                                    O O) "3
                                                                    (_ O 4->
                                                                    Q- C TJ
                                                                    in 01 t-
                                                                     .14-01
                                                 C QJ
                                                                    •I- •!- C
                                                                    U -- 1 (O
                          M/z
                                  48

-------
sensor wire sensed the tracer) and used to calculate the average over



the sample time.  The HWA time-averaged mean is not a vector average



since the flow direction is not sensed.  Hence, the PWA wind speeds



were smaller than the HHA indicated wind speeds.  Again, the profiles



show significant reductions in speeds near the  surface at x=5h and 8h.



The reductions from the reference values are quite clear.



     Figure 4.6 shows vertical profiles of the  r.m.s.  longitudinal



fluctuating velocity component normalized by U^(h).   Again,  results



from the HWA and PWA downstream of the windbreak and the HWA reference



profile are shown.  The PWA results were shown  in Figure 4.4 in a



different format.  The dashed lines separate the region in which the



r.m.s. fluctuations exceed 10% of the reference value from those in



which the fluctuations are less than or nearly  equal  to the  reference



value.  Increased turbulence is quite apparent  at x=lh just  above the



windbreak height.  The maximum fluctuation at each downwind  distance is



approximately twice the reference value at the  same height.   The



r.m.s. fluctuations in the lower region are approximately half the



reference r.m.s. fluctuations at x=lh.  Vertical  diffusion of the



increased turbulence in the shear layer is again apparent.  PWA values



tended to be higher than HWA values for heights greater than the



windbreak height and slightly less than HWA values for lower heights.



The greatest differences were in the region of  the highest r.m.s.



fluctuations.
                                   49

-------
in               in
  •                 *
rn       cr>      rvi       CM
in
s s ^'".B..
S

\
\
s X 8 e 8 S S a 5
\
0 ' ' ' /
efTl
H-J /
*•"* \ ''-* /
\ / ^
\ 7
\ 7
\ * /
CD
,&' g S Q?
o




-
31 LO  •— • LO O
m rv —
in s <*- 4->
3 I O
O •*-> O
in *- o %•£ .u
--H O J=. O
CX t-H r—
«O to -D 3
  i — 4->
. ._ -r- CU H-
0 *J E l/>
" — ' fQ O Qj CO
3 E -C  O) -M r—
0 C
~i as t- ai
-^ r— O t-
• cn <«- cu -r- 10
•r- 1 C O
mO ~TD "r^ *r--
O 
r^- c> • — o 3
> 3 — -U
-C cx cn 
•i— ••— ft)
^ C ^B O
O «- T-
'~ O> Q. 5
"O *~ CU
a.1 ••-(/) c:
Nl ««- -i-
•^- o ••
-fC i — t- CO   t- o o>
r— 3
O C fO >
t_ ••— 4J
O 	 1 ro O)
•o o
ID C
vo • c_ ai
1 O) OJ <*-
OJ t_ E O)
C- jQ O t-
3 -o e
cr> c a) a>
•f- -r- C J=
3
r~*
fO
c
a>
t.

-------
                      5.  FLOW ABOUT STORAGE  PILES





     The flow about the two storage piles will  be described.   It  is



important to locate the areas of high wind  speed since  that  is where



particle uptake is most likely to occur.   It  is also  important to locate



the areas of low wind speed; the areas where  wind erosion is  least



likely to occur.  The potential  fugitive-dust emissions from



uncontrolled (no windbreak protection) piles  are related to the wind



speed distributions shown here.






5.1  CONICAL PILE



     Figure 5.1 shows the top view of the pile  with contours  of



normalized wind speed, u/ur, where u is the wind speed  at the pile



surface measured with the thermistor and ur is  the wind speed at  the



equivalent full-scale height of  10 m in the absence of  the pile.   10 m



was chosen as the reference height rather than  11 m,  the pile height,



because 10 m is the standard height to which  "surface"  wind speed



measurements are usually referenced in the  meteorological literature.



Mote that the air flow in the figure is from  the left.   The areas of



maximum wind speed are near the  top of the  upwind face  but toward the



sides of the pile.  A high speed region (u/ur > 0.8)  is on the upstream



face, extending from near the crest down both sides.  The area of



minimum wind speed is in the lee near the top of the  pile with regions



of low wind speed extending down the pile on  both sides of the



centerline.  High speeds along the pile sides are expected because the



flow is accelerating round the pile.  The flow  separates on the lee



side, resulting in « region of low-speed recirculating  flow.
                                   51

-------
Figure 5.1  u/ur about conical  pile for no windbreak  case,
                           52

-------
5.2  OVAL, FLAT-TOPPED PILE
     Flows about the pile oriented with its  longer  axis  normal  to  the
air flow, then at 20° and 40° from the original  position are  described.
Figure 5.2 shows the top view of the pile with contours  of  normalized
wind speed.  This was the pile orientation during all  the windbreak
tests, except for the 20° and 40° pile orientations.   The overall
pattern, as indicated by the isotachs, is fairly symmetric  except  for
some asymmetry of flow in the lee.  Since the thermistors were  located
in a symmetrical pattern on the pile, a point-to-point comparison  of
measured wind speed about the axis parallel  to the  wind  direction  was
made.  Several significant differences were  observed.  Upon completion
of the runs with the windbreaks, wind speeds were measured  with the
pile oriented 180° to its original position. Again, the overall pattern
was similar to that shown in Figure 5.2,  but point-to-point comparisons
about the line of symmetry showed several  significant  differences.  The
differences likely resulted, in part, from the sources of error in the
thermistor measurements discussed in section 3.2.   Further, with nine
times as many thermistors on this pile, as compared to the  conical
pile, errors from applying the calibration equation (Eq.  3.7) to all
the thermistors could be greater than that discussed earlier.
Considering these uncertainties, the wind speed  pattern  was corrected
based upon the data from the 0° and 180°  pile orientations, assuming
flow symmetry around the pile surface.
     The corrected pattern is seen in Figure 5.3.   The highest  wind
speeds were observed on the windward face near the  top of the pile,
extending down the sides, similar to the  case with  the conical  pile
                                   53

-------
      flow
                                                          0.2
Figure 5.2  u/ur about oval,  flat-topped pile  for  no windbreak
case.
                                 54

-------
                                           .4
  _flow
direction
'bogt
                                    "at-topped p,,. for
                                                         no
                            55

-------
Again, the lowest speeds were observed in the  lee;  but  a  secondary

minimum also occurred on the top of the pile.

     Thermistor derived wind speed values for  the cases with  a windbreak
                                t
were also corrected.  The correction factor for each  thermistor  was  the

percentage difference between the value measured with the pile in the

0° position and the corrected value, both in the absence  of any

windbreak.  The data for all the windbreak cases and  for  all  the pile

orientations were corrected, using the same correction  factors which

ranged in magnitude from 0% to 50%, with 82% of the correction factors

being within ±20%.

     The normalized wind speed patterns for the pile  oriented at 20°

and 40° from its original position normal to the air  flow are shown  in

Figure 5.4.  The band of high speed near the top of the upstream face

was observed for both orientations.  The lobes of high  speed  extending

toward the base on the sides shifted with pile orientation.
                                   56

-------
                                                        o
                                                     CO
                                                     10 «»
                                                     10 H

                                                     0)
                                                     r~   *
                                                     1— O
                                                     0.0
                                                      i  o>
                                                     *J  VI
                                                     O
                                                     r—  OJ
                                                     to  t_
                                                     > -a
                                                     o -o

                                                     AJ •**
                                                     3  *
                                                     O
                                                     J=  O
                                                     •o  c
                                                     •o  o
                                                      Ci i—
                                                      OJ «->
                                                      (_ C
                                                      I- O>
                                                      O "O
                                                     in o
                                                         4->
                                                      v
                                                      L. cu
                                                      3 >—
                                                      CT S7>
                                                     v- C
                                                     U. 
-------
           6.  WINDBREAK EFFECTS  ON FLOW  ABOUT  STORAGE PILES



     The use of windbreaks for storage-pile  fugitive-dust control  is


based upon the existence of a  sheltered region  downstream of a


windbreak.  Effects of windbreak  height,  length,  porosity and position
                                 i

on wind speed about the piles  are discussed  for the  case of a windbreak


placed upstream of the piles and  on the top  of  the flat-topped  pile.


Relative wind speed reduction  and the  observed  maximum wind speed  are


used to assess the relative effectiveness of the  various windbreaks.



6.1  WINDBREAK UPSTREAM OF THE PILE


     Before the results are presented, general  windbreak effects are


hypothesized based upon flow structure and pile location in the


sheltered region.



6.1.1  Relation to flow structure


     Since height  and width of the sheltered region  are directly


related to windbreak height and length, windbreaks placed upstream of


the pile having dimensions less than the  pile height or length  are


expected to be less efficient  as  are smaller windbreaks placed  on  the


pile top.  Wind speed reductions  downstream  of  a  windbreak were greatest


near the surface and decreased with height (Figure 4.2).  Although some


reductions were observed at heights greater  than  the windbreak  height,


little, if any, reductions were observed  at  z=2h. Therefore, windbreaks


of height <0.5H cannot be expected to  have much effect on reducing the


high wind speeds occurring near the tops  of  the piles; the effect  may


be greater when the pile is farther from  the windbreak since the height


of the sheltered region increases with downwind distance.  Similarly,
                    \

windbreaks of length less than the pile base length  are expected to be



                                   58

-------
less effective since the pile is not "fully within"  the sheltered

region.

     Figures 6.1 and 6.2 show a superposition  of  the conical  and

oval, flat-topped piles' cross-sections at  the 1H and 3H positions  for

the three windbreak heights onto the "streamlines" downstream of a

windbreak of length 10.5 times its height  (Figure 4.1).   Streamlines  in

the pile's presence are not the same as those  shown; the location of

the low wind speed region with respect  to the  pile is of interest here.

The upstream face of both piles is in the same relative  position, but

the difference in pile size is quite apparent. For  the  0.5H  high

windbreak, the upstream face is in a lower  wind speed area when the

pile is in the 1H rather than the 3H position. For  the  l.OH  and 1.5H

high windbreaks, however, the upstream  pile face  is  in a lower wind

speed region when the pile is in the 3H position. The pile may be  in

an even lower wind speed region at 6H for the  1.5H high  windbreak.


6.1.2  Results

     Windbreak effects on surface wind  speed patterns are discussed in

this subsection.  Windbreaks were placed upstream of both piles, normal

to the wind direction, and at an angle  to the  wind direction  with the

conical pile.  The cases with a windbreak normal  to  the  air flow are

discussed first.  Figures showing normalized wind speed  for many of the

windbreak cases are not included because the change  in wind speed due

to a windbreak is of more interest.   With a windbreak, the wind speed

at a given location on the pile surface is  some fraction of that in the

unprotected case.  The relative amount  by which the  wind speed is
                   \
reduced is called the wind speed reduction  factor R^  and, in  percent,


                                   59

-------
                                                  e
                                                  
                                                 "O •*-*•
                                                  QJ  O
                                                  to  C

                                                 -o  ^
                                                  c

                                                     J3
                                                  c   .
                                                  o o
                                                 O
                                                 D.
                                                 r—  CT>
                                                 •t— -i—
                                                 Q. O>
                                                    jcr

                                                 IQ if—
                                                 O  O


                                                 O    O
                                                 O  t-
                                                 cx o
                                                 t-  o.

                                                 f^  1^
                                                 3
                                                 «/> <*-
                                                 VO  OJ
                                                     t_
                                                  OJ  4->
                                                 •^  o
                                                 u_ -a
                      M/Z
60

-------
                                                  . tf>
               . . o
                                               /
             y
                                                    .e

                                                     x
              . . CD
              . . to
                                                                 o
                                                              
                                                               -O
                                                               I  C
                                                               *• o

                                                              (O O
                                                              > CL
                                                              O
                                                              O M-
                                                                 O
                                                              C
                                                              o e
                                                             •r- 00
                                                             4J QJ
                                                             ••- t-
                                                              t/) -U
                                                              O 10
                                                              CL C
                                                              t- S
                                                              O) o
                                                              Q.-0
                                                              3
                                                             l^ =
                                                                 
                                                                 OJ
                                                             4->
                                                             •i— CO
M/z
M/z
           61

-------
is defined as
            Ri = K,i - Ui)/u0,i  x 100'                           (6-l)
where u-j and u0>-j  are wind speeds  at the  i-th  location  on  the  pile  for
the cases with and without the windbreak,  respectively.  R-j  is  zero
when the windbreak causes no change in the wind  speed,  and 100% when
the wind speed is  reduced to zero.   R-j of  25%  means  that the wind  speed
resulting from the windbreak's presence is 25% less  than the wind  speed
in the unprotected case.  It is important  to remember that the  combined
effects of windbreak height, length, porosity, and position, as well  as
turbulence in the  approach flow, and pile  shape  and  surface roughness
determine the wind speed distribution over the pile  surface.  Some of
these factors will be considered here.
     An optimum windbreak size (height and length) exists  since a  very
small windbreak is expected to be  ineffective  and  a  very large  windbreak
may be effective in reducing wind  speeds  but may turn out  to be too
expensive.  Contours of constant wind speed reduction factor R}
resulting from the 65% porous screen of height 0.5H, located 1H from
the pile base and  length l.OD (Figure 6.3) and l.OB  (Figure 6.4) are
shown.  Referring  also to Figures  5.1 and  5.2, in  the areas of  maximum
wind speed of the  no windbreak cases, these windbreaks  reduced  the wind
speed by approximately 20%.  For both piles,  relatively large  reductions
were observed over approximately the lower three-fourths of the upstream
face, with the greatest reductions apparent near the piles' centerlines.
Regions of wind speed increase (negative  reduction factor) were observed
on the lee side of the conical pile and on the top,  upstream half  of
the larger pile.  The increase in  speed observed on  the top of the
                                   62

-------
   direction"
                                                  .40
                                                   40
 Figure 6.3  Wind speed reduction  factor for the  65% porous windbreak
 of height 0.5H and length l.OD placed 1H from the  conical  pile base.
     flow
  direction^
Figure 6.4  Wind speed  reduction  factor  for  the 65% porous windbreak
of height 0.5H and length l.OB placed  1H from  the oval, flat-topped
pile base.
                                 63

-------
larger pile may be due  to  a  combination of air deflection about the
windbreak and the pile.  In  terms  of  fugitive dust control, these
increases are probably  not significant for the conical pile because the
wind speeds will  be relatively  Tow in this region unless the  reference
speed (ur) is very high,  but the increase could be significant for the
larger pile, again depending on ur.   This general pattern of  high wind
speed reduction in the  lower portion  of the upstream face and increase
in the conical  pile upper lee region  and the top upstream half of the
ovals flat-topped pile  was typical  of all the windbreaks of height
0.5H, but not of the higher  windbreaks.
     Slight asymmetry is  apparent  in  the lee of the larger pile in
Figure 6.4; this was also observed in the presence of other windbreaks.
Since the wind speeds are low in the  lee, even a small change in wind
speed appears as a large  relative  change in the reduction factor.
     For higher windbreaks (heights l.OH and 1.5H) upstream of the
conical pile, the area  of greatest wind speed reduction was the upper
part of the windward side, with typical reductions of at least 50%.
Areas of negative reduction  factors  (up to --15%) were observed with
windbreaks of height l.OH located  3H  from the conical pile; these
areas were significantly  smaller than those observed with lower
windbreaks (Figure 6.3).   Wind  speeds were reduced everywhere for the
other higher windbreaks with reduction factors ranging from 15% to 60%
on the lee side.  The reduction pattern for windbreaks of the same
porosity, length, and position  did not differ significantly when the
windbreak height was increased  from l.OH to 1.5H, except for  the least
porous windbreak located  farther  (3H) from the pile.   In that case, the
                                   64

-------
highest windbreak caused the greatest reductions  over much  of the pile.
The reductions in the lee near the pile top also  were higher for the
1.5H high windbreak.
     For higher windbreaks (heights l.OH and 1.5H)  upstream of the
oval, flat-topped pile, wind speeds were reduced  everywhere.  Increasing
the height from l.OH to 1.5H gave greater reductions  on  the pile top,
but the general pattern of wind speed reduction depended upon the
windbreak location.   In general9 a windbreak of height l.OH located  1H
upwind of the pile base gave similar wind speed reductions  (R-j) on the
top part of the upstream face and on the pile top.   Increasing the
height to 1.5H caused greater reductions only on  the  pile top.   A
windbreak of height  l.OH in the 3H position caused  higher reductions on
the upstream face than on the pile top.   Increasing the  height increased
the reductions on the top to approximately the same values  as those  on
the upstream face.  Relative reductions  of 30% to 60% occurred over
most of the top part of the upstream sloping face when windbreaks of
heights l.OH and 1.5H were close to the  pile (the 1H  position); for
windbreaks of both heights at the 3H position, the  corresponding range
of reduction factors was 45% to 75%.  The increase  in wind  speed
reduction on the top of the pile with windbreak height increasing from
l.OH to 1.5H was greater for windbreaks  located 3H  from  the pile base
(e.g., for 50% porous windbreaks at the  closer position, the reduction
increased from approximately 50% to 70%, but at the farther position,
it increased from approximately 30% to 60%).
     To summarize, windbreak height is clearly important, as
hypothesized in section 6.1.1.   The 0.5H high windbreak  is  not as
                                   65

-------
effective as higher windbreaks  (l.OH  and  1.5H), because the  forner
reduces wind speeds over a smaller portion  of the pile.   Higher
windbreaks reduce wind speeds over a  larger part of the pile's surface
and cause greater reductions  in the regions of high wind  speed observed
in the no windbreak case.   In general, with increasing windbreak  height
the wind speed reductions  increase also and cover a greater  portion  of
the pile, but with a conical  pile, the windbreak as high  as  the pile is
nearly as effective as a higher one.
     Increasing length is  expected to effect greater  reductions at the
pile sides.  This was observed  with the 50% porous windbreaks, although
the difference was slight  as  the length was increased from l.OD to 1.5D
and from l.OB to 1.5B with the  conical and  larger piles,  respectively.
Windbreaks of length less  than  the pile base were also placed upstream
of the oval, flat-topped pile.   The length  was equal  to the  maximum
length of the pile at the  top.   Very  high speeds were observed on the
upstream sloping face near the  pile sides with these  windbreaks.
Longer windbreaks located  upwind of the pile base caused  much higher
reductions in wind speeds  along the sloping sides and slightly higher
reductions at the sides of the  flat pile  top.  The reduction factors
for the two 50% porous windbreaks of height 1.5H at the 1H position  are
shown in Figure 6.5.  The  effect of length  at the sides is quite
evident.  Length becomes even more important if the incident wind is
not normal to the windbreak (see later discussion in  this subsection).
     Porosity was another windbreak parameter under investigation.   For
a given windbreak height,  length, and position, greater wind speed
reductions occured for the less porous  (50%) windbreak, particularly on
                                   66

-------
  flow
                      .40
 20
 20
 40'
   40
                        20
                          0
                       40
                     60
SSSrc
              67

-------
the windward face and top  of  the  piles.   The  reduction factors for the
two different porosity windbreaks of  height l.OH and length l.OD located
1H from the conical  pile and  the  two  windbreaks of height l.OH and
length 0.6B located 1H from the oval, flat-topped pile are shown in
Figures 6.6 and 6.7, respectively.   In both cases the distributions are
quite similar in shape, with  larger  reductions in the 50% porous case.
     The last windbreak parameter studied was the distance between the
windbreak and pile.  As discussed earlier, the effect of windbreak
position appears to be related to windbreak height.  Windbreaks of
height 0.5H caused greater wind speed reduction in the lower part of
the windward face and towards the sides when  placed 1H, rather than 3H,
from the piles' bases, although slightly  higher reductions were observed
along the top of the slope of the oval, flat-topped pile when the 0.5H
high windbreak was at the  3H  position.  Windbreaks of height l.OH
placed 3H from either pile caused greater wind speed reduction on the
windward face, but for the conical pile,  less reduction in the lee.
Windbreaks of height 1.5H  placed  3H  from  either pile caused greater
wind speed reductions on the  windward face than if the windbreak were
at the 1H position.  However, for windbreaks  of height l.OH and 1.5H
with the oval, flat-topped pile,  the  reduction on the pile top was less
when the windbreak was in  the 3H  position.  For both piles, the
reductions near the center!ine  along  the  pile slope were low near the
pile base and high towards the top of the slope with windbreaks greater
than 0.5H high at the 1H position.  In the 3H position, slightly higher
reductions occurred near the  top  of  the piles on the windward face and
much higher reductions occurred  in the lower  portion such that the
                                   68 .

-------
             flOW

          direction
                              60
Figure 6.6  Wind speed reduction factor for the windbreak  of  height
l.OH and length l.OD placed 1H from the conical pile base  with  porosity
65% (solid line) and 50% (dashed line).
             flow
          direction"
  Figure  6.7  Wind  speed  reduction  factor for the windbreak of height
  l.OH  and  length 0.6B  placed 1H from the oval,  flat-topped pile base
  with  porosity 65%x(solid  line) and  50% (dashed line).
                                   69

-------
reductions tended to be uniform  along the  slope.  Figures 6.8 and 6.9



show the wind speed reduction  distributions  for the two windbreak



locations (1H and 3H) for the  1.5H,  high 50% porous windbreaks of length



1.5D placed upstream of the conical  pile,  and  length  0.6B placed



upstream of the larger pile.   Clearly,  larger  areas of high  reduction



occurred for the 3H location with  the conical  pile.   Higher  reductions



occurred over much of the upstream face of the larger pile whereas



lower reductions occurred on the top for the 3H case.  Hence, the



overall  effect would depend upon the relative  magnitudes of  the local



effects.



     Patterns of wind speed reduction factors  clearly show effects of



windbreak height, length, porosity and  position.  The maximum wind



speed, independent of position,  may  also be  used to assess relative



effectiveness of the various windbreaks since  it is related  to the



maximum particle emission rate.  Values of maximum wind speed normalized



by the wind speed at the equivalent  full-scale height of  10  m in the



absence of the pile (ur) for the windbreak cases are  given in Tables 6.1



and 6.2 for the conical and oval,  flat-topped  piles,  respectively.



Note that the ratio umax/ur is 1.16  and 1.12 for the  unprotected  (no



windbreak) cases.  In general, higher maximum  wind speeds were observed



for windbreaks upstream of the oval, flat-topped pile than for the



corresponding windbreak with the conical pile, although many trends



were the same for both piles.



     All the windbreaks reduced  maximum wind speed, which is desired



for fugitive dust control.  Windbreaks  of  height 0.5H caused much



higher umax/ur than, did the higher windbreaks. Differences  in umax/ur
                                   70

-------
                flow
              direction"
                                6Q
                              80
X60
                          20
                                                          60
                                              80
Figure  6.8  Wind  speed  reduction factor for the 50% porous windbreak of
height  1.5H and length  1.5D  placed  1H (solid line)  and 3H (dashed line)
from the conical  pile base.
               flow
             direction
                                •40
                                   20
                             6Q
                                      20
Figure 6.9  Wind speed reduction factor for the 50%  porous windbreak of
height 1.5H and length 0.6B placed 1H  (solid line) and  3H (dashed line)
from the oval, flatv-topped pile base.
                                   71

-------
TABLE 6.1  umax/ur FOR THE  VARIOUS  WINDBREAKS  PLACED UPSTREAM OF  THE
                              CONICAL  PILE
            65% porous windbreak
50% porous windbreak
position: 1H
length: l.OD
height
0.5H
l.OH
1.5H
0.91
0.55
0.56
1.5D
0.93
0.59
0.60
3H
l.OD
0.91
0.54
0.50
1.5D
0.94
0.56
0.52
1H
l.OD
0.90
0.31
0.39
1.5D
0.93
0.34
0.42
3H
l.OD
0.82
0.37
0.25
1.5D
0.86
0.27
0.17
TABLE 6.2  umax/ur FOR THE VARIOUS  WINDBREAKS  PLACED  UPSTREAM OF  THE
                         OVAL,  FLAT-TOPPED  PILE
position:
length: 0.
height
0.5 H
0.75H
1.0 H
1.25H
1.5 H
0.
0.
0.
65% porous
1H
68
98
85
78
l.OB
0.96
0.70
0.69
windbreak
3H
0.
0.
0.
0.
6B l.OB
97 	
84 0.68
79 	
50% porous windbreak
1H 3H
0.6B
0.95
0.73
0.63
l.OB
0.92
0.58
0.45
0.48
0.45
1.5B
0.90
0.52
0.49
0
0
0
0
.6B
.93
.78
.66
l.OB
0.87
0.66
0.56
0.36
0.28
                                   72

-------
between windbreaks of height l.OH and  1.5H,  for the  same  porosity,



length, and position, were not  nearly  as  great as  for  those  between



0.5H and l.OH.  In general, for a given windbreak  height, length, and



position, the 50% porous windbreak caused lower umax/ur than did the



65% porous windbreak.  Placing  a given windbreak a distance  3H  from the



pile base was more effective than placing it  1H from the  conical pile



for nearly all the cases tested here.  For the larger  pile,  the 1H



position tended to be more effective except  for the  higher,  longer 50%



porous windbreaks and differences in umax/ur  between the  two positions



for the 65% porous windbreak were not  significant.   Windbreaks  shorter



than the length of the pile base were  clearly less effective.



     Given a windbreak height,  length, porosity and  location, the



direction of the incident wind  will also  affect the  wind  speed



reductions about a pile.  Wind  direction  effect was  studied  with the



50% porous windbreak of height  l.OH, length  l.OD,  placed  1H  from the



conical pile base.  Since the wind direction  is constant  in  a wind



tunnel, a windbreak oriented at an angle  to the tunnel center-line



simulates a full-scale case of  a fixed windbreak with  the air flow at



an angle to the windbreak different from  normal.   Figure  6.10 shows



reduction factors for wind directions  normal  to the  windbreak,  and at



20°, and 40° to the normal.  The windbreak positions are  also shown in



the figures.  The maximum u/ur  was 0.31,  0.69, and 1.12 for  the 0°,



20°, and 40° cases, respectively.  For the 20° case, a region of much



lower reductions was observed on the side of  the pile  opposite  the



windbreak, indicating that the  windbreak  length and  position are



important.  For the 40° case, the region  of  reductions greater  than 40%
                                   73

-------
   direction
      flow
    direction
      flow
    direction
                                                   60
                                                              (a)
                                                              (c)
Figure 6.10  Wind speed  reduction factor for the 50% porous windbreak of
height l.OH and and  length  l.OD placed 1H from the conical pile base
oriented (a) normal* (b)  20°,  and (c) 40° to the flow direction.
                                   74

-------
was quite small; indeed wind speed increases were observed.   Clearly,

windbreak effectiveness decreases with increasing angle of flow from

the normal.


6.2  WINDBREAKS ON THE PILE TOP

     In addition to placing windbreaks upstream  of both piles,  they  were

placed on the top of the oval, flat-topped pile  in the  positions shown

in Figure 6.11.  As stated earlier,  the 65% porous windbreaks were of

two heights and two lengths and were placed either near the  leading

edge or centerline of the pile.  For a given windbreak  in  both

positions, the pile was rotated 20°  and 40° from the  original position

to assess wind direction effects.

     Large reduction factors (up to  65%) on the  pile  top were observed

for all the cases.  The location and extent of the area with significant

wind speed reduction depended upon windbreak size, location, and angle

of the incident flow.  Higher, longer windbreaks caused larger  sheltered

regions as shown in Figure 6.12.  The width of the sheltered region

(parallel to the windbreak) is larger than the windbreak and areas of

wind speed increase (negative reductions)  were generally observed to

the sides of the windbreaks.  The greatest reductions were observed

between about 3h and 6h downstream of the  windbreak.  On the sloping

surfaces of the pile, reduction factors of 5 to  10% were observed

upstream of the three windbreaks placed near the leading edge and

negative reduction factors in the lee of the pile downstream of the

windbreaks.  A windbreak placed near the centerline caused wind speed

reductions upstream, as well as downstream, of the windbreak on the
                   V
pile top (Figure 6.13(a)).  The area of coverage and  magnitude  of the


                                   75

-------
Figure 6.11  Sketch of windbreak  positions  with  respect to thermistor
positions on the top of the oval,  flat-topped  pile.
                                 76

-------
A
   U
   0)
                                                                                               O)
                                                                                                  in
                                                                                               t.
                                                                                               
                                                                                                  cn
                                                                                                   0)
                                                                                                  -C
                                                                                               O
                                                                                               t_
                                                                                               O
                                                                                               Lft
                                                                                                   -
                                                                                               t-  en
                                                                                               O  c
                                                                                               O T3
                                                                                               •)->  C
                                                                                               O  nt
   f-.  •
 C Cvf I—
 o   . vo
•r- O •— •

 O 4-> O
 3 J=
-O  cr>-C
 O) -r- *J
 C_  O) 0>
   J= C
-o     o>
                                                                                               TO  O>
                                                                                                c  em
                                                                                               •r- TD f^
                                                                                               3  r— -^  ^
                                                                                               •r- «r- O
                                                                                               U. Q-- --
                                                 77

-------

                                                                                                              CsJ
                                                                                                                  O
                                                                                                               
                                                                                                               +-> •(->
                                                                                                                o c
                                                                                                                «o o>
                                                                                                               «*- o
                                                                                                               4J  CX
                                                                                                                o
                                                                                                                13  
                                                                                                               XJ  TJ
                                                                                                                OJ  1)
                                                                                                                O)  C
                                                                                                                CX
                                                                                                                U1 T3
                                                                                                                    O)
                                                                                                               •o  o
                                                                                                                c  
                                                          78

-------
reductions were greater for a higher windbreak.   Reduction  factors  for



the 0.14H high, 0.5T long windbreak placed  near  the  centerline with the



incident flow 40° to the windbreak normal clearly show  that the  area of



coverage shifted with the wind direction  (Figure 6.13(b}).  Smaller



reductions were observed on the slope downstream of  the windbreak.



Here the reduction factor was defined with  respect to the no  windbreak



case with the pile at 40° to the flow.



     To summarize, windbreaks on the pile top  reduce wind speeds on the



top both upstream and downstream of the windbreak.   Greater reductions



were observed a distance 3h to 6h downstream of  the  windbreak than



immediatedly in its lee.  Small  reductions  occurred  along the slope



upstream of the windbreaks placed near the  leading edge, but, in



general, the areas of high wind speed without  any windbreak were not



affected.  Hence windbreaks located on the  top of the pile  may not



provide much protection against high winds, but  may  be  used to locally



protect regions on the top of the pile.
                                   79

-------
             7.  FURTHER ANALYSES AND DISCUSSION OF  RESULTS





     In this section, windbreak efficiencies  based upon  the



relationships between particle uptake and  wind  speed are presented.
                                i


Then the results of both parts of the experiment (flows  about  a



windbreak and windbreak-protected piles) are  compared to results  of



previous studies.





7.1  RELATION TO PARTICLE UPTAKE



     Relative windbreak effectiveness may  be  assessed by several



methods.  Wind speed reduction patterns and the normalized maximum wind



speed have been described.  Further analyses  based on particle uptake



at wind speeds exceeding a threshold value will  be described.   Results



for windbreaks placed upwind of the pile will be discussed first.



     If emissions had been measured directly, the windbreak  efficiency  E



would have been defined as




                  E = 1 - (Q/Qo)                                  (7.1)



where Q and Q0 are the storage-pile fugitive-dust emission rates  with



and without the windbreak, respectively.   Since wind speeds  have  been



measured here, assumed relationships between  wind speed  and  emissions



are used to calculate efficiencies.  Depending  on the reference wind



speed, the wind speed on the lee side of the  piles and on part of the



top of the oval, flat-topped pile in the  absence of  any  windbreak may



be less than the threshold.  In general,  with certain windbreaks, the



speeds in these regions could become greater  than the threshold.



Furthermore, a windbreak may reduce wind speeds to values less than  the



threshold over partner all of the pile.  Hence  a better definition of



efficiency would include a threshol-d value.  However, threshold speeds  •



                                   80

-------
have been determined only for a few cases  as  discussed  in  section  2.1.
The relationship between threshold speed and  particle size and  moisture
content is not understood.  If a given  threshold  speed  were applied  to
the data here, a full-scale reference wind speed  would  have to  be
assumed.  To a first approximation, it  is  assumed here  that the
reference wind speed is sufficiently high  that  wind  speeds everywhere,
with and without a windbreak, exceed the threshold.  Percentage
efficiencies Ej and £3 are defined based upon linear and cubic  relations
between wind speed and particle uptake,  respectively.   Hence
                     = [1 - ICuiAO/lKjAi)]  x  100               (7.2)
                            i        i
                  E3 = [1 - RuO/lK i-)]  x  100             (7.3)
                            i         i
where the summation is over the entire  pile.   In  effect,  these
efficiencies are l-(un/u0n),  where n is  either  1  or  3  and  un  and  uQn  are
the area-averaged values over the pile surface  with  and  without a
windbreak, respectively.  Later in this  subsection,  as an  example,
threshold and reference wind  speeds are  assumed and  the  areal  extent  of
erosion is noted for the various windbreaks.
     Previous analyses by the present author  of the  data for  the  conical
pile utilized an overall or area-averaged  reduction  factor,
R = 1-u.j/u.j Q, and a windbreak  effectiveness  factor,  E  =  I-U^/UQ  ^6t
(Billman, 1984).   However,  if the objective was  to  define a  parameter
which is the fractional  amount  by which  fugitive emissions have been
reduced, then Ej  and £3 would be better  parameters  than R^ and  E.   In the
previous analyses, if the wind  speed at  some  points in  the no  windbreak
case were zero, theh R" or E would be infinite.   In  addition, the regions
                                   81

-------
of low wind speed in the absence  of  the windbreak make relatively large



contributions to R and E.   This  limitation  has been  removed by using the



terms E^ and £3 in the present analysis.



     EI for the windbreaks  placed upstream  of the conical and larger



piles with normal incident  flow  are  given in Tables  7.1 and 7.2,



respectively.  In general,  a windbreak is more effective  (higher E^)



when placed upstream of the conical  pile, as compared to  a windbreak



of the same relative size placed  upstream of the larger,  oval-shaped



pile.  Trends in Ej with changes  in  height, length,  location and



porosity.of the windbreak are similar for both piles.  In general,



increasing windbreak height is desirable.   Efficiencies range from



28-45% and 13-21% for 0.5H  high  windbreaks  placed upstream of the



conical and larger piles,  respectively, and for higher windbreaks,



44-77% and 27-62%, again for the  two piles.  The 1.5H height was more



effective than the l.OH height with  the oval, flat-topped pile,



reflecting increased wind speed  reductions  on the pile top for the



highest windbreak.  For most cases with the conical  pile, windbreaks of



heights l.OH and 1.5H were  equally effective, within experimental



scatter.  Efficiencies were higher with the less porous windbreak



material.  Except for the windbreaks of height one half the pile height



(0.5H), efficiency was lower when the windbreak was  not as long as the



pile base length.  Length was not significant for the 0.5H high



windbreaks because only the lower part of the pile sides  were



significantly affected with the  longer windbreak, not the higher part,



the location of high wind speeds in  the absence of any windbreak  (refer



to Figure 5.3).  However, with a higher windbreak, increased length
                                   82

-------
TABLE 7.1  EFFICIENCY (Ej) FOR THE VARIOUS  WINDBREAKS  PLACED UPSTREAM
                          OF THE CONICAL PILE

position
length:
height
0.5H
l.OH
1.5H
65%
J
l.OD

34
48
47
porous
1H
1.5D

32
45
44
TABLE 7.2 EFFICIENCY (


position
length:
height
0.5 H
n 7RH
1.0 H
1 ?RH
1.5 H

65%

0.6B

15

27

33
OF
porous
1H
l.OB

16

34

39
windbreak
3H
l.OD

28
53
55
50% porous
1H
1.5D l.OD 1.5D

30 46 45
52 66 67
54 64 65
El) FOR THE VARIOUS WINDBREAKS
THE OVAL
windbreak
3H
0.6B 1

13

28

38
, FLAT-TOPPED PILE
50% porous
1H
.OB 0.6B l.OB

18 20
	 	 41
37 34 53
	 	 56
44 58
windbreak
3H
l.OD 1.5D

36 36
65 71
71 77
PLACED UPSTREAM

windbreak
3H
1.5B 0.6B l.OB

21 15 17
	 	 34
51 31 49
	 	 57
59 43 62
                                   83

-------
caused wind speed reductions  in  the  high wind  speed area on the sides.



For both piles, the windbreak locations  of  1H  and  3H upwind of the pile



were better for the windbreaks of  height 0.5H  and  1.5H, respectively.



     Tables 7.3 and 7.4 list  £2,'the efficiency  based upon the u3



relation to dust uptake for the  various windbreak  cases for the conical



and larger piles, respectively.  Again,  the efficiencies tend to be



higher with the conical pile. Many  of the  same  qualitative results as



with EI are apparent.  The trends  are clearly  seen in a plot of



efficiency (£3) vs. windbreak height as  functions  of windbreak length



and porosity for both piles (Figure  7.1).   Windbreaks of the lowest



height were least effective;  windbreaks  higher than the pile were



significantly more effective  only  with the  oval, flat-topped pile.



Windbreaks shorter than the pile base length were  less effective than



those at least as long as the pile.   The less  porous material was



clearly more effective.  These trends were  observed for windbreaks at



either the 1H or 3H position.



     Although the highest efficiencies of 99%  and  96% correspond to the



50% porous material of height 1.5H,  length  1.5 times the base diameter



of the conical pile and equal to the base length of the oval, flat-



topped pile, respectively, located 3H from  the base of the piles, the



efficiencies of the more economical  windbreak  of the same porosity,



height equal to the pile height  and  length  equal to the pile base



length are only slightly lower (97%  and  89%, respectively).  Clearly,



the latter size would be preferable  on the  basis of cost effectiveness.



Any location between 1H and 3H from  the  base of  the pile could be



chosen depending onvthe convenience.
                                   84

-------
TABLE 7.3  EFFICIENCY (E3)  FOR THE VARIOUS  WINDBREAKS  PLACED UPSTREAM
                          OF THE CONICAL PILE
65%
position:
length: l.OD
height
0.5H
l.OH
1.5H
74
88
86
porous
1H
1.5D
72
85
82
windbreak
3H
l.OD
66
91
92

1.5D
67
90
91
50%
l.OD
82
97
95
porous
1H
1.5D
80
97
95
windbreak
3H
l.OD
76
97
98

1.5D
76
98
99
TABLE 7.4  EFFICIENCY (E3)  FOR THE VARIOUS  WINDBREAKS  PLACED UPSTREAM
                     OF THE OVAL,  FLAT-TOPPED  PILE
65% porous
position: 1H
length: 0.6B
height
0.5 H
0.75H
1.0 H
1.25H
1.5 H
42
63
67
l.OB
45
74
73
windbreak
3H
0.6B l.OB
39 	
66 79
75 	
50% porous windbreak
1H 3H
0.6B
48
73
80
l.OB
49
84
92
91
92
1.5B
51
89
92
0.6B
45
70
81
l.OB
50
74
89
95
96
                                   85

-------
                                                  o
                                                  CO

                                                  XJ   •
                                                  Ol  O)
                                                  O r—
                                                  
                                                  .*:  CL
                                                  
                                                  4->  O
                                                  O)
                                                  x:
                                                      O)
                                                  >> c
                                                  o  o
                                                  c  u
                                                  O)
                                                  LU  01
                                                      in
                                                      
-------
     As discussed earlier in this section,  windbreak  effectiveness may



also be assessed by assuming threshold (u^) and  full-scale  reference



(ur) speeds.  An example calculation was made  using the  u/ur  wind



tunnel data for windbreaks placed upwind, assuming a  reference wind



speed of 10 m/s and a threshold velocity of 2.8  m/s to determine the



percentage surface area over which no emissions  would occur (i.e. where



u < u^).  (This threshold speed was also used  in an example calculation



by Martin and Drehmel (1980), and may be applicable to some fine



material).  With no windbreak, 22% (25%) of the  conical  (oval) pile



surface area had wind speeds less than the  threshold.  With all the 65%



porous windbreaks and the 0.5H high 50% porous windbreaks,  40-65%



(24-49%) of the pile surface area would have no  dust  emissions.  The



area increased to 77-100% (40-99%) for the  other 50%  porous windbreaks.



Note that the percentage surface areas calculated here would  increase



with a decrease in reference speed and/or increase in threshold



velocity.



     For windbreaks placed on top of the larger  pile, efficiencies (Ej



and £3) were quite low, less than 12%.  As  discussed  in  section 6.2,



wind speeds on the pile top were reduced significantly,  but the



windbreaks had very little effect in the high  wind speed region on the



windward face.  This suggests that fugitive-dust emissions  on the top



of the pile may be controlled locally through  the use of a  windbreak.



Since the windbreaks were not very high, (0.14H  and 0.27H), in practice,



they could be portable, providing protection to  specific areas on the



top of the pile, depending upon the location of  activity and  wind



direction.         v
                                   87

-------
7.2  COMPARISON TO PREVIOUS  STUDIES
     Many results compare well  with  previous  studies  discussed  in
section 2, but several  windbreak  design  guidelines  are  introduced  here.
The flow structure about a porous windbreak will  be discussed first,
then that about windbreak-protected  piles.
     The flow structure downwind  of  a  porous  windbreak  was  qualitatively
similar to that observed by  Raine and  Stevenson  (1977).   The area  of
minimum wind speed and high  turbulence intensity  occurred not
immediately, but farther downwind of the windbreak  and  turbulence  was
generated in the shear layer at the  top  of the windbreak.  As observed
by Perera (1981), the hot-wire  anemometer (HWA) overpredicted wind
speeds in the highly turbulent  region.  Also, for heights greater  than
the windbreak height, the longitudinal turbulence intensity, measured
with the pulsed-wire anemometer (PWA), was greater  than that measured
with the HWA.  However, unlike  that  observed  by  Perera,  the intensity
was less when measured with  the PWA  for  heights  less  than In and the
mean wind speed measured by  the two  instruments was approximately  the
same for heights above about 2h.   Perera used a  solid barrier  (0%
porous); here it was 50% porous.
     Results with the windbreak-protected piles may be  compared with
two previous laboratory studies in which windbreaks were located upwind
of a pile.  Davies (1980) recommended  a  windbreak height of 1.4H,  which,
based on our results, should be very good.  We  did  not  examine  the case
of a windbreak of varying porosity as  Davies  did.  However, our results
cannot adequately be compared with those by Davies  since only  the  final
result was reported*, not the details of  the various cases tested.   The
                                   88

-------
present work extends that reported by  Soo et  al.  (1981)  and  Cai  et  al



(1983). . In both studies, optimal  windbreak  location  was found to be



related to the windbreak height,  and lower wind speeds were  observed



with less porous windbreaks.   In  the present  study, with three-



dimensional piles, effects of windbreak  length, additional windbreak



heights and porosities, and various incident  flow directions were



examined.



     Windbreak efficiencies presented  here are  generally much higher



than those estimated by Bohn et al. (1978) and  Jutze  et  al.  (1977),



indicating that windbreaks may be  a highly effective  fugitive-dust



control method.
                                   89

-------
                            8.   CONCLUSIONS

     This wind tunnel  study has  shown  that windbreaks normal to the wind
direction placed upwind of  a conical and a larger, oval, flat-topped
storage pile reduce wind speeds  hear the surface of the pile and hence
suggest reductions  in  fugitive-dust emissions.  Of the windbreaks
tested for each pile,  the largest  (height  1.5 times the pile height
(1.5H) and length equal  to  the larger  pile base length (l.OB) or 1.5
times the conical pile base diameter  (1.5D)) 50% porous windbreak
placed 3H from the  pile appears  to be  best in terms of greatest wind
speed reduction.  However,  all the 50* porous windbreaks at least as
high as the pile and as long as  the pile base had similar overall
effects.  Windbreaks of height and/or  length less than that of the pile
were clearly less effective.  Optimal  windbreak location appears to be
related to windbreak height, particularly  for the conical pile; the
higher the windbreak,  the farther  it  should be located upwind of the
pile.  However, locations farther  than 3H were not examined.  A farther
location could be more effective,  but  the  pile could then be beyond the
sheltered region, hence, a  less  effective windbreak position.
Windbreaks of height 1.5H caused greater wind speed reductions on the
top of the oval, flat-topped pile  than those of height equal to or less
than the pile height,  but the difference in the effectiveness of l.OH
and 1.5H high windbreaks with the  conical  pile was not significant.
     Windbreak length and position are even more important  in
determining effectiveness when the air flow is not normal to a
windbreak.  With a windbreak of  height and length equal to  the pile
dimensions, fairly nigh wind speed reductions resulted when the
                                   90

-------
windbreak was placed upwind normal  to the flow and also at an angle of
20° to the normal, but very little reduction occurred at an angle of
40°.
     Windbreaks (0.14H and 0.27H'high) placed on the top of the oval,
flat-topped pile caused large areas of high wind speed reductions on
the pile top both downwind and upwind of the windbreak, but very small
reductions to the high wind speeds on the windward face occurring in
the absence of any windbreak.  The area of greatest reduction was not
immediately downwind of the windbreak, but displaced farther downstream.
Changes in wind direction shifted the location of the sheltered region.
These results suggest that fugitive-dust emissions may be locally
controlled with windbreaks placed on  the top of a relatively level
storage pile.  In particular, portable windbreaks may be quite practical
since they could be positioned to protect active areas of the pile.
     Design guidelines developed by Soo et al. (1981) and Cai  et al.
(1983) have been extended since more  windbreak configurations were
examined and three-dimensional piles  were used.  Windbreak efficiencies
were generally much higher than expected (Bohn et al., 1978 and Jutze
et al., 1977).  With the design guidelines presented here, the use of
windbreaks for fugitive-dust control  appears promising.
     Wind speed was isolated here as  the major factor affecting storage-
pile fugitive-dust emissions, but storage-pile moisture content, type
of material stored and threshold speed also affect emissions.  A clearer
understanding of the relationship of  wind speed and threshold speed to
fugitive-dust emissions would allow for better analysis of the data
presented.  Additional  field measurements of fugitive dust from storage
                                   91

-------
piles with and without windbreaks  would be  helpful  for  comparison  to



the efficiencies and design guidelines  presented here.
                                   92

-------
                         9.  REFERENCES
Axetell, Jr., K., 1978:  Survey of fugitive dust from coal  mines.
    EPA-908/1-78-003, EPA Region VIII, Office of Energy Activities,
    Denver, CO 80295.
                                i
Bagnold, R.A., 1941:  The Physics  of Blown Sand and Desert  Dunes.
    Methuen, London, 265 pp.

Billman, B.J., 1984:  Windbreak effectiveness for the control  of
    fugitive-dust emissions from storage piles — A wind tunnel study.
    Presented at the Fifth Symposium on the Transfer and Utilization
    of Particulate Control Technology, Kansas City, MO, August 27-30,
    1984.  Proceedings to be published by the Environmental Protection
    Agency and Electric Power Research Institute.

Blackwood, T.R. and Wachter, R.A,  1978:  Source assessment: Coal storage
    piles.  EPA-600/2-78-004k, U.S. Environmental Protection Agency,
    Cincinnati", OH.

Bohn, R., T. Cuscino, and C. Cowherd, Jr., 1978:  Fugitive  emissions
    from integrated iron and steel  plants. EPA-600/2-78-050. U.S.
    Environmental  Protection Agency.

Bradbury, L.J.S. and I.P. Castro,  1971:  A pulsed-wire technique for
    velocity measurements in highly turbulent flows. J. Fluid Mech. 49:
    657.

The Bureau of National  Affairs, Inc., 1983:  Environment Reporter.
    336:0553-0554, Rule 302(f).

The Bureau of National  Affairs, Inc., 1982:  Environment Reporter.
    411:0516-0517, R336.1371, R336.1372.

Caborn, J.M., 1957:  Shelter-belts  and microclimate. Forestry Commun.
    Bull. 29:1.

Cai, S., Chen, F.F., and Soo, S.L., 1983:  Wind penetration into a
    porous storage pile and use of barriers. Environ. Sci.  Technol.
    H_: 298.

Caput, C., Belot, Y., Guyot, G., Sarnie, C., and Seguin, B., 1973:
    Transport of a diffusing material over a thin wind-break.  Atmos.
    Environ. ]_: 75.

Castro, I.P., and W.H. Snyder, 1982:  A wind tunnel study of dispersion
    from' sources downwind of three-dimensional hills. Atmos. Environ.
    Jj>: 1869.

Counihan, J, 1975: ^Adiabatic atmospheric boundary layers:  A review and
    analysis of data from the period 1880-1972. Atmos. Environ, j): 871.
                                   93

-------
Cowherd, Jr., C., K. Axetell,  Jr., C.M.  Guenther,  and G.  Jutze, 1974:
    Development of emission factors for  fugitive dust sources. EPA-450/
    3-74-037, U.S. Environmental  Protection Agency, Office of Air
    Quality Planning and Standards.

Cowherd, Jr., C., R. Bohn, and T.A. Cuscino, 1979:  Iron  and Steel
    plant open source fugitive emission  evaluation.  EPA-600/2-79-103.
    U.S. Environmental  Protection Agency, Research Triangle Park, NC.

Cowherd, Jr., C, 1982:   Emission  factors for wind  erosion of exposed
    aggregates at surface mines.   Proc.  75th APCA  Annual  Meeting. Paper
    82-15.5.

Currier, E.L. and B.D.  Neal, 1984:  Fugitive emissions from coal-fired
    power plants. EPRI  CS-3455.  Electric Power Research  Institute, Palo
    Alto, CA.

Cuscino, T., G.E. Muleski, and C. Cowherd, Jr., 1983:  Iron and steel
    plant open source fugitive emission  control evaluation.  EPA-600/
    2-83-110.  U.S. Environmental Protection Agency, Office of Research
    and Development, Research  Triangle Park, NC.

Davies, A.E., 1980:  A physical modelling approach to the solution  of
    fugitive emission problems. Proc.  73rd APCA Annual Meeting. Paper
    80-68.11.

U.S. Environmental Protection  Agency,  1983:  Supplement No. 14 for
    Compilation of Air Pollutant Emission Factors, 3rd ed., AP-42,
    Research Triangle Park, North Carolina.

Gandemer, J. 1981:  The aerodynamic characteristics of windbreaks,
    resulting in empirical design rules. J. of Wind Engineering and
    Industrial Aerodynamics. _7_: 15.

Gillette, D., 1978a:  A wind tunnel simulation of  the erosion of soil:
    Effect of soil texture, sandblasting, wind speed, and soil
    consolidation on dust production.  Atmos. Environ. 12: 1735.

Gillette, D., 1978b:  Tests with a portable wind tunnel for determining
    wind erosion threshold velocities. Atmos. Environ. 12: 2309.

Gillette, D.A., J. Adams, A. Endo, and D. Smith, 1980:  Threshold
    velocities for input of soil  particles into the air by desert soils.
    J. Geophys. Res. 85(C10):  5621, October 20, 1980.

Jacobs, A., 1984:  The  flow around a thin closed fence. Boundary-Layer
    Meteorology. 28: 317.

Jensen, M, 1958:  The model-law for phenomena in a natural wind.
    Ingenioren, Int. Ed.  Vol. 2, No.  4.
                                   94

-------
Jutze, G.A., J.M. Zoller, T.A. Janszen, R.S. Amick, C.E. Zimmer and
    R.W. Gerstle, 1977:  Technical guidance for control of industrial
    process fugitive particulate emissions. EPA-450/3-77-010, U.S.
    Environmental Protection Agency, Office of Air Quality Planning and
    Standards.

Lawson, Jr., R.E. and R.E. Britte'r, 1983:  A note on the measurement of
    transverse velocity fluctuations with heated cylindrical sensors at
    small mean velocities. J. Physics E: Sci. Instrum. 16: 563.

Martin, D.J. and D. Drehmel, 1980: Control methods for fugitive area
    sources. Proc. Fourth Symposium on Fugitive Emissions, Measurement
    and Control. EPA-600/9-80-041, U.S. Environmental  Protection Agency,
    Office of Research and Development, Research Triangle Park, NC,
    p. 402

Mulhearn, P.J. and E.F. Bradley, 1977:  Secondary flows in the lee of
    porous shelterbelts. Boundary Layer Meteorol. 12:  75.

Naegeli, W., 1953:  The braking effect of a forest on  wind.  Internat.
    Union Forestry Res. and Organisation, llth Congr.  Rome. Section 11,
    pp. 12-17. In: Windbreaks and Shelterbelts. WMO Technical Note
    No. 59. Van Eimern, et al., eds.

Ogawa, Y. and P.G. Oiosey, 1980:  Surface roughness and thermal
    stratification effects on the flow behind a two-dimensional fence.
    II.  A wind tunnel study and similarity considerations. Atmos.
    Environ. 14: 1309.

Pendergrass, W., and S.P.S. Arya, 1984:  Dispersion in neutral boundary
    layer over a step change in surface roughness--!.  Hean flow and
    turbulence structure. Atmos. Environ. 18: 1267.

Perera, M.D.A.E.S., 1981:  Shelter behind two-dimensional solid and
    porous fences. J. Industrial Aerodynamics. J3: 93.

Raine, O.K. and Stevenson, D.C., 1977:  Wind protection by model fences
    in a simulated atmospheric boundary layer. J. Industrial
    Aerodynamics. 2: 159.

Rasmussen, R.A., 1962:  Application of thermistors to  measurements in
    moving fluids. Rev. Sci. Instrum. 33: 38.

Seginer, I., 1975a:  Atmospheric stability effect on windbreak shelter
    and drag. Boundary Layer Meteorol. 8_: 383.

Seginer, I., 1975b:  Flow around a windbreak in oblique wind. Boundary
    Layer Meteorol. 9: 133.
                                   95

-------
Snyder, W.H., 1979:  The EPA meteorological  wind tunnel--It's design,
    construction and operating characteristics.  EPA-600/4-79-051. U.S.
    Environmental Protection Agency, Office  of Research  and Development
    Research Triangle Park, NC.

Snyder, W.H, 1981:  Guideline for fluid modeling of atmospheric
    diffusion.  EPA-600/8-81-009,'U.S.  Environmental  Protection Agency,
    Research Triangle Park, NC.

Soo, S.L., J.C. Perez, and S. Rezakhany, 1981:  Wind velocity
    distribution over storage piles and use  of barriers. Proc. Symposium
    on Iron and Steel Pollution Abatement Technology for 1980.
    EPA-600/9-81-017, U.S. Environmental Protection Agency, Office of
    Research and Development, Research Triangle Park, NC.

Thornthwaite, C.W., 1931:  The climates of North America according to a
    new classification. The Geographical Review. 21: 633.

Townsend, A.A., 1956: The Structure of Turbulent Shear Flow. Cambridge
    University Press, Cambridge, EngTand, 315 pp.

Van Eimern, J., Karschon, R., Razumova, L.A., and Robertson, G.W., 1964:
    Windbreaks and shelterbelts. WMO Technical Note No.  59.  188 pp.
                                   96

-------