Gaussian Dispersion Model

Z-axis through stack

Heplmeceraerline

Y-axis is transverse
to wind

X-axis in
a- direction of wind
t e e d ak

Courdiss 9 ad **alame fox ite

Ciwka unz ide
ruPind

Concentration
Air pollution law in most industrial countries
bascd on concentration of contaminants
NAAQS in US
-

Need method to predict concentrations at any

given location
-

Any given set of pollutant

Meteorological conditions
At any location
For any time period
Bul even best currently available concentralion
models are far from ideal
 Concentration

Commonly express concentration as ppm or

Hg/m3
Parts per mil ion (ppm)= 1 volume of
-

I ppm volume gaseous pollutant_

10 volumes (pollutant + air)
Hg/m' micrograms/cubic meter
 Factors that determine Dispersion

Physical nature of effluents

Chemical nature of effluents
Meteorology
Location of the stack
Nature of terrain downwind from the stack
 Stack Effluents
Gas and particulate matter
Particles< 20 um behave same as gas
-
Low setting velocity
Particle> 20 um have significant settling velocity
treated in
Only gases and Particles < 20 um are

dispersion models
Others are treated as
particulate matter

Assumes eflluents leave the stack with sullicient

momentum and buoyancy
- Hot gases continue to rise
enim i d

Assumptions
Efluents leave the stack with sufficient
momenturî and buoyancy
Hot gases continue to rise
Plume is deflected along its axis in
proportion to the average wind speed (u)
 Gaussian or Normal Distribution

Gaussian distribution model

Dispersion in y andz directions uses a
double gaussian distribution --

plumes
Dispersion in (x, y, z) is three-dimensional
Used to model instantaneous pulf of
emissions
dd

Gaussian or Normal Distribution

Pollution dispersion follows a distribution

function
Theoretical form: gaussian distribution
function
 Gaussian or Normal Distribution

20 451

xmean of the distribution

a standard deviation
Gaussian distribution used to model probabilities, in this
context fomula used to predict steady state concentration
a a point down stream
 Gaussian or Normal Distribution
What are some properties of the nomal
distribution?

Point o
inflection

FIGURE 41 The Gaussian or normal distnbution tunctwon for ditlerent values of

x, ard a

f(x) becomes concentration, maximum at center of plume 10

Gaussian or Normal distribution
68% of the areafall within I standard deviation of
the mean (u Vd).
95% of area fall within 1.96 standard deviation of
the mean (u t 1.96 o).
99.7% of the area fall within 3 standard deviations
of the mean ( 3 o)
 Gaussian dispersion model
Dispersion in y and z directions are
modeled as Gaussian
.Becomes double Gaussian model
Why doesn't it follow a Gaussian
distribution in the x direction?
Direction of wind
 Gaussian Dispersion Model
For localized point sources -stacks
General appearance
Plume exits at height, h,
Rises an additional distance. Ah
-

buoyancy of bot gases

called plune rise
reaches distance where buoyancy and pward momentum cease
Exit velocity. V,
Plume appears as a point source emitted at height H
h,+Ah
Emission rate Q (g's)
Assume wind blows in x direction at speed u
-

n is independent of tinme. elevation. or location (not really true)

Gaussian Dispersion Model

fauvalet o

al a f i t a k regtH
FIGURE 42 Adieson m o i wih vrtiai r
 Gaussian Dispersion Model
Stack gas transported downstrgam
Dispersion in vertical direction governed by
atmospheric stability
Dispersion in horizontal plane governed by
molccular and eddy diffusion
X-axis oriented to wind direction
Z-axis oriented vertically upwards
y-direction oriented transverse to the wind
Concentrations are symmetne about y-axis
axis
and z-

15
itpdi

As distance
increase so does
dispersion

Souros al
coordinates
(0. 0.

Figure 20.3
Behavior of the downwinc, elevated transverse concentraton profilies as
a function of distance downward.

17
a scE. Coopet aud Alen. J002
Figure 204
Coordinate system showing Gaussian distrbtions the harizontal
and venical
Anpet rom Tume 90

g 4 Cocpa ud Aley. 00
 Point Source at Elevation H

Assumes no interference or limitation to

dispersion in any direction

Cu, )2uo,0, (4-6)

20,

a d 2, are locatiou of centerline of plune

Yo taken as base of the stack

is H
Q- emission strength of sorce (1nss tine) gs
u- average wind speed thru the plhune m's
C concetration-gin' (Notice this is not ppm)
a, and o, ae horizoutal and vertical standard deviations in meters
19
 Wind Velocity Profile

Wind speed varies by height

International
standard height for wind-speed
measuremeats is 10m
Dispersion of pollutant is a funetion of wind
speed at the height where pollution is
emitted
But difficult to develop relationship
between height and wind speed
mddgdl

Point Source at Elevation H without Reflection

Cu. . 2uo, xp -. xp (46

3 ten
gives coceutration on the ykuterline of the plyie
gves concenutration as yod uove in the sideyays direction (t y
duection). direction doesu't umatter becauy(t y gives a positive
value
gives concentration as you nove in the'vertical directiou (t z
direction). direction doesu't uatter because (t (z- H)F gives ia
positive value
Concentrations are symmetric aboul y-nxis and z-axis
Same concentration at (z-H)= 10 m a s (z-H)=10 m
Close to ground symunetry is distubed
tpd

Gaussian Plume Example

A factory emits 20 g/s of SO, at height H
(includes plume nse)
Wind speed 3 m/s (u)
-

A t a distance of 1 km downstream, o, and a, are

30 m and 20 m (given. otherwise we would have
to look up)
What are the SO, concentrations at the centerline
of the plume and at a point 60 meters to the side
and 20 meters below the centerline
 Gaussian Plume Example

Cu..) 2uo,d, 4 8)

Q-20g's of SO Z-H 0 So second half

u-3 ' s (u) of equation
a, and o, are 30 m and 20 m goes to 0
y 0 andz -
H
So reduces to
Cx0.0) 20 gs .0017gm'= 1770n
2113 30*20)

At centerline y and Z are 0

 Gaussian Plume Example

Cu.y.) 2uo,0, (4-8)

O exp-1/2C) +((2-HE]
2lu a, , la,
20gs exp-12 [(-60mE+-20m =
1330N20) (20-m)
(0.00177 g'u) (exp)-0.000145im'or 14S.231g/
At 20 and 60 meters
 Pasquill-Gifford Curves
Concentrations correspond to sampling times of
approx. 10 minutes
Regulatory models assume that the concentratiens
predicted represent 1-hour averages
Solid cures represent rural values
Dashed lines represent urban values
Estimated concentrations represent only the lowest
several hundred meters of the atmosphere
 Pasquill-Gifford Curves
o , less certain than o,
- Especially for x>1 km
For neutral to moderately unstable
atmosphericeconditions and distances out to
a few kilometers, concentrations should be
within a factor of 2 or 3 of actual values
Tables 3-1: Key to stability classes