Sec.9.

ESP
for special ct
mo$tmmo

ElectrostaticPrecipitators 9.2.I Ploil

Thit
boilers" cefi
g.I..INTRODUCTION alyic crdn
(furmad
.. An electrostatic pr.9 ipitator (ESP) is a pa4igle control device thal uses elec,trical forces hr
to move the particles out of the flowing gas stream and onto collector electrodes.They range.in voltege rfo
size from those installed to'clean the flue gasesfrom the largest powerplants to'those used as eachfurp
small household air cleaners.ESPs account for about 95Voof all utility'particulate controls in the Th
U.S. (Offen and Altman :1991). The particles are given an electrical charle by forcing'them to quite trll Tl
pass through a corona, a region in which gaseousions flow. The electrical field that forces the needfuqq
charged particles to the walls comes from discharge electrodesmaintained at high voltage in the secrftm'd
center of the flow lane. usdltdi
Once the pafiicles are on the collecting electrodes,they must be removed from the sur- n
face without reentraining them into the gas stream. This is usually accomplished by knocking bgmrn
them loose from the plates; allowing the collected layer of particles to slide down into a hopper crctbl
from, which they are evacuated. Some ESPs remove the particles by intermittent . or continuous dccubr
washing with water. Precipitators are unique among particulate matter control devices in that the agCthrl
forces of collection act only on the panicles and not on the entire gaseousstream..This phenome- trfiflh
non typically results in a high collection efficiency (above 99.57o)with a very low gas pressure zmfiQ
drop. k
cles (Lrl
enqlbr
9.2 TYPES
OF ESPS sm&dL
Ar
l*ting d
gas. Ib c
Your objectivesin studying this sectionare to Evenh{Y
levelsdl
1. IdentifythemostcommonESPconfigurations: (1) plate-wire,(2) flat &
plate,(3) tubular,(4) wet, and(5) two-stage. ping th{
2. Describethe operatingcharacteristics andadvantages of each coal ffyd
configuration. later serti:
3. Understand the importanceof reentrainment,gassneakage, anddust recaptured
resistivity. providcdt
of thetutd
 371
Sec.9.2 TyPesof ESPs

have been developed

ESPs are configured in several ways. Some of these configurations
economic reasons' A description of the
for special control actiois, while others have evolved for
most common configurations follows'

PreciPilolors
9.2.1Plote-Wire
Thisconfigurationisusedinawidevarietyofindustrialapplicationsincludingcoal-fired
recovery boilers, petroleum refining cat-
boilers, cement kilns, solid waste incinerators,paper mill
coke oven batteries, and glass furnaces
alytic cracking units, sinter plants, open hearth furnaces,
(Turner et al. 1988 a).
sheet metal' Weighted high-
re_i In a plate-wire ESP, gas flows between parallel plates of
I ::- voltage wire electrodes hang i"t*""n the plates as shown schematically in Figure 9'1' Within
| .\ it flows through the unit.
each flow path, the gas must*passeach wire sequentially as
&e Theplate'wireEsPallowsmanyflowlanestooperateinparallel,andeachlanecanbe
handling large volumes of gas' The
I ri,-r quite tall. Therefore, this type of precipitator is well suited for
causesthe plates to be divided into
te need for rapping the platesio disiodge the collected material
be rapped independently' The power supplies are
the sections,often three or four in series,which can
higher operating voltages'
usually sectionalizedin the samemannerto obtain
voltage (220440 V) to pulsat-
;ur- The power suppliesfor the ESP convert theindustrial AC
The voltage applied to the electrodes
ing ing DC voltage in the range of 20 to 100 kv as needed'
an action known as a "corona'" The
Fter causesthe gas between thJplates to break down electrically,
negative corona supports a higher volt-
ous electrodesusually are given a negative polarity becausea
generated in the corona follow elec-
the age than a positive before sparking occurs. The ions
"o.*onu each wire establishes a charging
ne- tric field lines from the wires to the collecting plates. Therefore,
ilre zone through which the particles must pass'
the ions' Small aerosolparti-
Particles pusring through the charging zone absorb some of
their total charge becomes large
cles (less than l-pm diameter) can absorb tens of ions before
diameter) can absorb tens of thou-
enough to repel further ions. Large particles (more than 10-pm
on the large particles'
sandsofions. The electrical forces are therefore much sffonger
closer and closer to the col-
As the particles pass each successivewire, they are driven
to keep them uniformly mixed with the
lecting walls. The turbulence in the gas, however, tends
electrical and dispersive forces'
gas. The collection process is a cJmpetition between the
so that the turbulence drops to low
Eventually the particles approach close enough to the walls
levels and the particles are collected'
efficiency of the ESP' The rap-
Reentrainmentof the particles during rapping reducesthe
of the particles (typically l2Vo for
ping that dislodges the accumulated dust also pto3""tt some
particles are then processedagain by
coal fly-ash) back into the gas stream. these ieentrained
of the ESP have no chance to be
later sections,but the particles reentrained in the last section
into the space
recaptured and so the unit. Part of the gas flows around the charging zone
"r"up"align the electrodes.this is called "sneakage" and amountsto 5Voto IOVo
provided to support ani
flow to mix with the main gas stream
of the total flow. Antisneak-agebaffles force the sneakage

.dftnq*ilHrnF||ilr"
 372 Chap.9 ElectrostaticPrecipitators Sec.9

causesan
High-voltagewires high reen
Collectorplates ticles.
I
(1-2 trm
velocity a

9.2.3Tu

tube. Tod
Gas flow Figure 9.L Schematic monly ap
diagram of a plate- There are
wire ESP es are mu

9.2.4U

for collection in later sections.But, again, the sneakageflow around the last section has no oppor-
tunity to be collected. The q'ater
Another major factor that affects the performance of ESPs is the resistivity of the collect- a sump fc
ed material. Becausethe particles form a continuous layer on the ESP plates, all the ion current corona-
must pass through the layer to reach the ground plates. This createsan electric field in the layer dled mur
which can become large enough to cause local electrical breakdown called "back corona.'oBack
corona is prevalent when the resistivity of the layer is high, usually above 2 x 1011ohm-cm. It
9.2.5Ir
reducesthe collection ability of the unit becausesevereback corona causesdifficulty in charging
,
the particles. Conversely, at resistivities below 108ohm-cm, the particles hold on to the plates so
elecrical
loosely that reentrainmentbecomes much worse. The resistivity is strongly affected by tempera-
collectior
ture, moisture, gas composition, particle composition, and surfacecharacteristics.
device r*i
this confi
9.2.2FlcrlPlotePrecipilolors high-resis

A significant number of smaller ESPs (50-100 m3/s) use flat plates instead of wires for smaller d
the high-voltage electrodes.The flat plates increasethe averageelectric field and the surface area filter- ion
available for the collection of particles. Flat plates can not generate corona by themselves, so d&arro r
corona generating electrodes are placed ahead of and sometimes after the discharge plate zone. tigle-'a.
These elecffodesmay be sharp-pointedneedlesattachedto the edgesof the plates or independent
corona wires. Unlike plate-wire or tubular ESPs,this design operatesequally well with either pos-
itive or negative polarity. A positive polarity reducesozone generation.
A flat plate ESP operateswith little or no corona current flowing through the collected
dust, except directly under the corona needles or wires. This makes the unit less susceptible to
back corona than conventional precipitators. However, the lack of current in the collected layer

*#i-
0rs Sec.9.2 Typesof ESPs 373

causesan electrical force that tends to dislodge the layer from the collecting electrode leading to
high reentrainmentlossesduring rapping. The dislodging electrical force is stronger for large par-
ticles.
Flat plate ESPs have wide application for high-resistivity particles with small MMDs
(1-2 pm). Fly-ash has been successfully collected with this type of precipitator, but low-flow
velocity appearsto be critical for avoiding high rapping losses.

9.2.3TubulorPrecipitolors
The original ESPs were tubular, with the high-voltage wire running along the axis of the
tube. Today they comprise only a small portion of the precipitator population and are most com-
hc monly applied where the particulate is either wet or sticky-for example, in sulfuric acid plants.
There are no sneakagepaths and, becausethey are usually cleaned with water, reentrainmentloss-
es are much lower.

9.2.4WeI Precipilotors

Any of the precipitator configurations describedearlier may be operatedwith wet walls.

The water flow may be applied intermittently or continuously to wash the collected particles into
a sump for disposal. This type of ESPs has no problems with rapping reentrainmentor with back
corona. The wash increasesthe complexity of the device, and the collected slurry must be han-
dled more carefully than a dry product adding to the expenseof final disposal.

9.2.5Two-Stoge
Precipifotors
This configuration separatesparticle charging and collecting functions to optimize the
electrical conditions for each. Charging requires a high current density and electric fieid, whereas
collection requires high electrical fields but much less current. The two-stage ESp is a series
device with the dischargeelectrode, or ionizer, preceding the collector electrodes.Advantages for
this configuration include more time for particle charging, less propensity for back corona with
high-resistivity ash, and economical construction for small sizes.
This type of precipitator is usually applied for gas flow rates of 25 m3/s or less. The
lftr
smaller devices are often sold as preengineeredpackage systemsconsisting of a mechanical pre-
EJ
filter, ionizer, collecting-plate cell, after-filter, and power pack. Recent work suggeststhai the
Li"
addition of a precharger to a conventional precipitator is an economical way to it from
IE
sirgle- to two-stage operation when handling high-resistivity ash (Offen and Altman "*u"rt
k: 1991).
F!-

Etail
C ILr
l\er

,*,M*,g,.
374 Chap.9 Eleotrostatic
Precipitators Sec .9.

9.3 ETECTROSTATIC
PRECIPITATION
THEORY Exampl

Dur
Your objectivesin studying this sectionare to 1,fi
Thc
1. Develop the grade-efficiency equation for an ESP. (a)
2. Estimate the average operating electric field strength for an ESP. " (b)
3. Estimate the total charge on a particle by diffusion and field charging.
4. Estimate grade efficiencies of an ESP with Feldman's model. Solutior

(a)
The theory of ESP operation requires many scientific disciplines to describe it thorough-
tor
ly. The ESP is basically an electrical machine. The principal actions are the charging of particles
and forcing them to the collector plates. The transport of the particles is affected by the level of 0.2
turbulence in the gas. The particle properties also have a major effect on the operation of the ESP.
o)
9.3.1Porticlecollection
The electric field in the collecting zone producesa force on a particle proportional to the
magnitude of the field and to the particle charge:

F"= QpE"
(e.1) l
usually v
where
terminal'
F"=force due to the electric field, newtons (N)
form dist
ep = chargeon the particle, coulombs (C) laminar t
E" = strength of the electric field in the collecting zone, volVmeter (V/m) to that ol
The motion of the particles under the influence of the electric field is opposedby the vis- equation
cous drag force of the gas. When the drag force exactly balances the electrostatic force, the parti-
cle attains its terminal velocity, 21.Assuming that the particle obeys Stokes's law,

QpErC" l
ut=- ''
3np.D, (e.2)
where

where
C, = Cunningham correction factor
D, - particle diameter

F + J '
 Sec.9.3 Electrostatic
Precipitation
Theory 37S'.

Example9.1 Terminal velocity of charged Particle In.Dlectric Field

Dustparticlesof 1.O-pmdiameterwith an electricchargeof 3 x l0-16C anda densityof

'
1,000kg/m3comeunder:the: influenceof an electricfield with a strengthof 100,0@V/m.
Theplilticlesaresuspended in'air at 298'K and l'atm.
''
(a) Estimatetheterminalvelocityof thepaiticles. ,
* (b) Calculatethe ratio of the electrostaticforce to the force of gravity on the particle
,
Solution
- -.----- --

(a) For 1.0-pmdiameterparticlesin air at 298K and i atm,the'Cunninghamcorrectionfac-

tor is C" = 1.168.FromEq.(9.2),ut= (3 x l0-re)(10s)(1.168)/l(3nXl.S4
x l0-s)(tGr)l=
0.202mls

(b) The ratio ofelectrostaticto gravity force actingon the particle is.: .,

-Fo
6qoE"
= -=
(o)(:,,ro 16)(rbs)
=5,847
Fs npogDf n(roooxs.s)(to-tt;

, , Equation(9.2) givesthe particlevelocitywith respectto still air. In the ESP,theflow is

usually vely tur,bulent,with instantaneousgas velocities of the samemagnitudeas the particles'
terminalvelocify,but in randomdirections.Theeddyingmotionowingto turbulencecausesa uni-
form distribution of particlesthroughmost of the flow passage.The capturezoneis limited to the
laminar boundarylayer in the viiinity of the collecting elecuode.The physicalsituationis similar
to that of turbulent-flow settling chambers(seeProblem 7.13). The resulting grade efficiency
equationis

'n(Do)=
/\Fl
1-exPlell
: ' (e.3)
t'AJ
: :: :'
where
A" = total collecting electrode area
Q=gas volumeffic flow rate
,976 Chap.9 EleetrostaticPrecipitators Sec.9

Example9.2 CollectionEfliciency of lirbular ESP 9.3.2El

Assur4qthat thq conditionsof Exampleb.1 conespondto a tubularESPwith a diameterof 1
thecollectingelectrodeof 2.9 m anda lengthof 5.0 m. If the gasflow rateis 2.0 m3ls,esti- which th
matethe collection efficiency for 1.0-pmdiameterparticles.Corroboratethe assumptionsof which rou
turbulentflow andStokes'slaw applicability. one setof
1
Solution negativec
ducedby r
Figure 9.2 showsschematicallythe tubularprecipitatorarrangement.Assumingturbulent morefree
flow, Eq. (9.3)givesthe gradeefficiency.Thecollectionareafor a tubularprecipitatoris experim
the areaof the curvedsurface,A,=np2= n(2.9)(5)= 45.55m2.Therefore,rl(1 pm) = | -
exp [-(45.55)(o.2o2y(2.0)]= o.99 (99vo).

where
T
D = 2 . 9m J

4
f
.F
(
Figure 9.2
1
Tubular ESPof grating lh
Exarnple9.2 field is in
voltage u
complerq1
is givenbl
To corroboratethe assumptionof turbulentflow, calculatethe Reynoldsnumberfor the gas
flow throughthe circular conduit,Re - pvDlp, wherey is the averagegasvelocity. For the
givenconditions,y = (4)(2)ltn(2.g)21
= 0.3m/s.TheReynoldsnumberis Re= (1.185) where
(0.303)(2.9y(1.84x l0-5) = 56,600.Therefore,the gasflow is fully turbulent. d
d
To corroboratethe applicability of Stokes'slawocalculatethe terminalReynoldsnumberfor x
theparricles,
Reo=pupollt= (1.185x0.202)(1}6y(1.84 x 10-5)= 0.013.Therefore,
Stokes'slaw applies.
Exanpf

Apl
diil

fi:4,*-
 precipitation
Sec.9.3 Electrostatic Theory 377

9.3.2Eleclricotoperotingpoint
The electrical operating point of an ESP section is the value of voltage
and current at
which the section operates.The best collection occurs when the highest
electric field is present,
which roughly correspondsto the highest voltage on the electrodes.The
term section represents
one set of plates and electrodeshaving a common power source.
The lowest acceptablevoltage is the voltage required for the formation
of a corona. The
negative corona is formed when an occasional free electron near the
high-voltage electrode, pro-
duced by a cosmic ray, gains enough energy from the electric field to
ionize the gas and produce
more free electrons' The electric field for which this processis self-sustained
has been determined
experimentally. For round wires, the field at the surfaceof the wire is given
by White (1963) as

E o = 3 x 1 0 6 / [ s . g+. 0 . 0 3 " { i e J r , ]
(e.4)
where
E0 = corona onset field at the wire surface,V/m
s.g.= specific gravity ofthe gas,relative to air at 293 K and 1 atm
rw = radius of the wire, m
f = roughnessfactor
For a clean smooth wire,/- 1.0; for practical applications,/= 0.6 is
a reasonablevalue
(Flagan and Seinfeld 1988).
The voltage that must be applied to the wire to obtain this value of field
is found by inte-
grating the electric field from the wire to the collecting electrode.
In cylindrical geometry, the
field is inversely proportional to the radial distance. This leads to
a logarithmic dependenceof
voltage on electrode dimensions. In the plate-wire geometry, the field
spatial variation is more
complex, but the voltage still .exhibits the logarithmic dependence.The
corona onset voltage, Vp,
is given by

Yo=Ar*nldl
(e.5)
where
d = ontercylinderradiusin a tubularESp
d = (4/n)Wfor plate-wireESp
I4l= wire-plateseparation

Example9.3 Corona OnsetVoltagein plate-Wire ESp

A plate-wireESPhandlesair at 400 K and 110kPa.Theplatespacingis 300mm,

andthe
diameterof the dischargewiresis 4 mm. Estimatethecoronaonsetvoltaee.

'lm;nEi|l-
378 Chap.9 ElectrostaticPrecipitators Sec.9.

Solution
, ,, Equation (9.4) gives the corona onset field at,the wiie surfaCe.The specific gravity of the
'gasis
s.g.= PT,"y'P,"fT=(1rc)Q93)lt@00x101.3)l
=0.795.Therefore,
E0)'(319o;(0.6)
" where
[0.795+0.03(0.79510.002)0.5]=2.51x106V/m.Calculated=4Wln=4(150)t/n=I9l Es
mm. Equation(9.5)givesthecoronaonsetvoltage,Vs= (2.5I x 106X0.002)
ln (I9ll2) = T
22,900Y. P
TI
,Theelectric field is strongestal,ongthe line from wire to plate and is approximatedvery
well, exceptnearthe wire, by
Example
E -V
Dma)c - -
w (e.6)
Estin
Ther
where V = applied voltage. The electric field is not uniform along the direction of gas flow.
-
Turner et al. (1988 a) suggestthat the averagefield for an ESP section is given.by
Solutim
F -E-*
K.
(e.7) Equr
B€cr
where K is a constant which depends on the ESP configuration and the presence ofback corona. quu
Table 9.1 gives values of K for some coffrmon situations.
:,, . a
,, , .,:
trG
: ., I

Toble9.I Rofioof Moximumto AverogeEectricFieldsin ESPs 9338

Confizuration Back corona o
Plate-wire Noa t.75 imtotli
Plate-wire Severe 2t0 image&
Flat plate Noa 1.26 esad
one, loc*
Flat plate Severe 1,80
chrge is si
aResistivityof.thecollecteddustlessthan2 x 10tt ohm-cm. on apfltid
TurneretaI.(1988a).
Source: It
Difhrsirnr
the ims I
When the electric field throughout the gap between the wire and the collecting electrode they temi
becomes sftong enough, a spafk occurs. The voltage cannot be increased beyond this point with- chargingir
out severe sparking'occurring. A reasonable estimate of the sparking field strength, E , is given by chargiagd
(Turner et al. 1988a):
 4-,'"'?.*
'S
Sec.9.3 Electrostatic
Precipitation "
Theory 379

t
E, = 6.3r.tOt{zz:p)o
\rl (e.8)

where
E" = sparkingfield strength,V/m
T = absolutetemperature, K
P - gaspressure,atm
TheESP generallyoperatesnearthis point.E_*is equalto or lessthanE".

ry
Example 9.4AverageElectric Field in a plate-Wire ESp

Estimate the averageelectric field and the operating voltage for the ESp of Example
9.3.
The resistivity of the collected dust is 1011ohm-cm.

Solution

Equation(9.8)givesE, = 630,000l(273/400)GI0/101.3)10.s = 496,000y/m. = E*.

Becausetheresistivityof thecollecteddustis lessthan 2 x l}tt ohm-cm,thereis no back
corona.FromTable9.1,K = 1.75,therefore,Eay= 496,000/it75 = 2g3,400V/m.
Theoper_
atingvoltageis V = WE*o,=(0.l5)(496,000)= j 4.4 kV .

9.3.3PorticleChorginE
Charging of particles takes place when ions bombard the surface of a particle.
Once an
ion is close to a particle, it is tightly bound becauseof the image charge
within the particle.The
image charge is a representationof the charge distortion that occurs
when a real charge approach-
es a conducting surface.The distortion is equivalent to a charge of opposite
magnituJe to ihe reat
one' located as far below the surface as the real charge is above it.^itre
motion of the fictitious
charge is similar to the motion of an image in a mirror, hence the name.
As more ions accumulate
on a particle, the total charge tends to prevent further ionic bombardment.
There are two principal charging mechanisms: diffusion charging
and field charging.
Diffirsion charging results from the thermal kinetic energy of the ions
overcoming the repulsion of
the ions already on the particle. Field charging results when ions
follow electric field lines until
they terminate on a particle. In general, both mechanismsoperate
for all sizes of particles. Field
charging is the dominant mechanism for particles greater ihan about pm,
2 whereas diffusion
charging dominatesfor particles smaller than about 0.5 p*.

.r-..*..{d. {.4ilcrr-..

,il$a,
380 Chap.9 Electrostatic'Precipitators $ec.9,3

Diffusion charging produces a logarithmically increasing level of charge on particles

given by (White 1963)

/ \ 2nukTD^
-" -ulnll+t/r1)
qjt)=
e (e.e)
L)
o
where .l 90
-
ea particle chargeby diffusion
a
€o = freespacepermittivity(8.845x 10-12C2Al-m2)
k = Boltzmann'sconstant(1.38x 10 23J4()
e = electroncharge(1.67x 1&ts C)
/ = exposuretime
ra = characteristictime for diffusion charginggiven by
-u.a__ t n l S m k- n T
^
ezNDo
Fi(

N = numberof ionsper unit volume

r??= massof an ion
For typicaloperatingconditions,N =2x 101sions/m3,m= 5.3 x 10-26kg (Crawford
1976,Licht 1980). where
Diffusion chargng neverreachesa limit, but it becomesvery slow after threecharacter-
. Qs'
istic times.For fixed exposuretimes,the chargeon the particle is proportionalto its diameter.
T'
Thesanrrr

.-
Calculatethe chargeby diffusionon a 1.0-pmdiameterparticleasa functionof time.The
gasis air at 400 K and 1 atm.The ion densityis 2 x 101sions/m3,andthemassof a typical where
ion is 5.3x lF26 kg. tr=
:t I
E=
Solution . : TI
whyfifld
stanti$giE
Calculatethe characteristic
time for diffusioncharging,17= (8.85x l}l2) I(8)(5.3x
10-26)(1.3816-z:;1n)(400;1o.s711t.Utx 10-re)2(2
x 10r5X10-6)l = 1.36x 10-5s. FromEq.
"

of the chargeaequiredby the particleby.diffusion. : .i

whereB=ir
lyrz\l(2x
rs Sec',9.3,,.El'ectriostatic
Precipitation
Theory' 381

c.)
o
9A

Figure9.3Chargin$by diftu-
sion of a 1.0-pmdiarneter
p a r t i c l e a t 4 0 0 ' K: '.r': ' : '
0.01 0.1 1 10
Exposure
time(ms)

g.
Field chargingatsoexhibits a characteristictime-dependence,
given by

/ -\ --t t
Q
47\t)= (9:10)
t+ry

r- 4r = saturationcharge(chargeat infinite time)

T/ - field chargingtime constant
The saturationchargeis given by

n"=(#) neaEDf (e.11)

': ,
r=dielectricionstantoftheparticle r.:,;
E = externalelectricfield appied to theparticle
The saturationchargeis proportion"l ,: ,1" squareofthe particle
size, which explains
why field chargingis the dominantmechanismfor larger particles.
ile field chargingtime con-
stantis given by

ry= 199-
" NeB (e.r2)

lbf,=in mobility' usuallyof the orderof 10+ m2lV-s.For typicalcondirions, (4)(g.g5

17= x
l|}rzX(2 x loq(r'6r * 1gt11lr)l = 0-0011 s. For practical purposes,the saturationcharse
382 C h a p . 9 Elect'rostatic
Precipitato
rs Sec.9

may be takenas attainedat t = 100If. = 0.1 s. Becausethe residencetime of the particlesin the
precipitatorgenerallyexceedsa few seconds,it is usually assumedthat the particlesattainthe sat-
urationchargesoonafter enteringthe device. , I
Strictly speaking,both diffrrsion and field chargingmechanismsoperateat all times on
all the particles,andneithermechanismis sufficient to explain the chargesmeasuredon the parti
cles. It has beenfound empiricallythat a very good approximationto the measuredchargeis
givenby the sumofthe chargespredictedby Eqs.(9.9)and(9.10)independently ofone another.

qp(t)=q[t)+qAt)
(e.13)

I account l
Example 9.6 TiotalParticle Charge I effect of

For the following setof conditions,calculatethe ratio 41q, asa function of particle sizefor
/ = 0.1,1.0,and10.0s.

r=5.0 E=3x10sV/m Z=300K

B =2.2x 10{ mzls-V N =2x 1015
ions/m3
m = 5 . 3x 1 F 2 6 k g whereI
(9.3)tor
Solution

The field chargingtime constantis from Eq. (9.I2), tf = 0.000503s. Thus,assumethat for r
> 0.05s the field chargingmechanism is saturated.Equation(9.13)becomes
where
Qp= r.787x to-so] + 1.438x t0-t2Drm(t+ t.lsx toloroo)

The first term in this equationis the saturationfield charge.Therefore,

Q p_ ,
- -
I . !,%g n(r+t.tsxrorooo)
Qs
,E",d
whereDo is in metersand r in seconds.The following tablepresentsthe ratio elq, asa becm
functionof particlesizeandexposuretime.
 Theory' 383
Sec,9.3 Eleclrosta''tic:Frecipitatbn

qy'qs
l:=,0.1s / = 1 . 0s r=10.0s
36;15 54.59 73.TI
0,0'1"
636 8.2L. .10.06
0;10
1.91 .:.: ,2.09
1.00 ,1.72
1.09 1 . 11 _1 . 1 3
10.0

'
Feldman(1975) .., ^rnnnser{
proposedanotherrepresentation
reoresentation-of the'total chargeon a particle to
of the'tott
to Cochet(1961)the combined
accountfor the behaviorJf tf'" 'otmicron parl"gs' According
effect of field chargrnganddiffusion chargingleadsto:

=l('.
qp(l . - L-
1,eoED:
f;l ,'4--]
-A)(".')l t+71 (e.14)
L \
incorporatedEqs' (9'1a) and(9'2) into
wherex,= mean-freepathof the gas.For t >> xf,Feldman
(9.3)to obtain
f , -
u:"
rn, ,_l
(e.1s)
' \ r ' = 1- expl-^'?
n(a) \c,ro)l
L 3PQ 'i

where

*L\'. , 4*,-t) I
'-l\^-+/
F=f(l (e,16)
('.^lt'..rl]
precipitator,
In a single-stage theelectricfield-in1le vlcrmtybf thecollectingelectrode'
E'"' Equation(9'15)
E., and* J!"tric field, E, are^**"0 "quuri" trtgit*uee field'
"n"ti[
becomes
'
tL(c,FDpl
n(pp)=1-expl-*run
(e.17)
I
T
go* "'fi*o
 384 Chap.9 Eleotrostatic
Precipitators S€e.E

Example 9.7 Grade-Efficiency Curve for Single.StageESP

pfl
The averageelectricfield in a single-stage,plate-wireESPiq 300,000V/m. The gasflowing bd
throughit is air at298K and I afrn.The dielectricconstantof the particlesis 5.0. The spe- d
cific collectionarea(SCA= AotQ)is 78.8m2l(ml/s).Showthatthegradeefficiencyfor par,
ticulateremovalgoesthrougha minimum for particlediametersbetween0.1 and 1.0pm.

Solution

Equation(9.17)gives '::.1' E
'
: . . . I I - , €o
x 10s}zg p 7t3( 1.84x 1F)l }
rt(Dp) = 1 - exp {-(78.8X8.85x 10-12X3
- I - exp 1.137x IN C,FDrl.
{-
"p
E
For air at298Kand1 atm,]v=0.0667pm.For r = 5.0,EQ,(9.16)becomes

r =(t +0.1334\2+ -----8-

\\ Dr; L [ t +' l ' 0 . 1 3 3 4 \
,; )
9J0
where Do' is in microns. The following table summarizes,thecalculationsto estimatethe
ESP'sgradeefficiencyfor variousparticlesizes. II

T
0.01 22.68 206.0 1.000 L
0.05 5.06 13.77 0.981 ,"",.
0.1 2.91 5.94 0.860 rtl

0.2 1.89 3.46 0.774

0.4 r.42 2.64 0.819 Glfi
0.6 I.28 2.43 0.880 GH
0.8 r.2l 2.34 0.923 Eb
1.0 G-rtl
I.l7 2.30 0.953
ad
2.0 1.08 2.21 0.996 *,1

i
i
i
I

lI
Sec.9.4 OverallEtficiency

Figure 9.4 is a graphicalrepresentationof theseresults.It showsa "window" for

particle sizesbetween0.1 and 1.0 pm where the penetrationincreasessignificantly. This
behaviorwas first documentedexperimentallyby Abbott and Drehmel(1976) and hasbeen
observedby manyotherssince.

I 0.9
C)
r)
b 0.85
C)
cl
6 0.8
Figure 9.4 Grade effi-
ciency versus particle
sizefor an ESP

9.4 OVERALL
EFFICIENCY

Yourobjectivesin studyingthissectionareto
1.Estimatetheoverallefficiencyof anESPwith Deutsch-Anderson
equation.
quadrature
2. EstimatetheoverallefficiencythroughGaussian formulas.

Equation (9.3) as it stands,without refinements,is known as the Deutsch-Anderson

equationand hasbeen usedas the basisfor much work on precipitators.Although it is a grade-
efficiency equation,it has been the practiceto use it for overall efficiency calculations.This is
doneby replacingal with a so-calledeffectitremigration velociry,wr. This quantity is takento rep-
resentthe collection behaviorof an entire dust of a certainkind and undera certain set of operat-
ing conditions.It is back calculatedfrom experimentaldataand so is really a performancepara-
meter.Equation(9.3) becomes:
 R
i*:'
-i,-'

386 Ghap.,9,,,:Eleo,trostatie'Precjpitator€ S ;'$#

'r .,:..SCA=4cr.
(e.18)
:.. . . , ' : : , ' , ' .Q
, ,.'.t.,'

Turner et al. (1988a) calc.ulated migrationvelocitiesfor threemain precipitatortypes:

plate-wire,flat.plate, and wet wall ESPsof the pfale-wire Epe.,Ttiil following'threetables; keyed
to overall designefficiency, summarizethe migration velocities under varioris.conditions.In
T
Table9.2,the migrationvelocitiesfor plate-wireESPsaregivenfor.conditionsof no backcorona
andseverebackcorona.In Table9.3, they aregivenfor a wet wall precipitatorassumingno back
coronaand no r4ppingreentrainmdnt.In Table 9.4,the flat plate ESP migrationV'eloOities are s
given only for no back coronaconditionsbecauseflat plate ESPsappearto be lessafftcted by
high resistivify duststhan the plate-wiretype.
"It is generallyexpectedfrbm experiencethat the migration velocity will decreasewith (
increasingefficiency.However,in Tables9.2 through9.4the migrationvelocitiesshow some
fluctuations.'itis is becausethe numberof sectionsmustilcreaseasthe efficiencyincreases, and
ihe Changin€i's€ctionalization df:",,-, the overall migration velg-city..This effect is particularly
I
noticeable,for example,in Table9.4 for glassplants.
I

Exarnple 9.8 Design of Flat Plate ESP Basedon Migration Velocity

Calculatethe total collectorplate areafor a flat plate ESPtoliontrljl-"fly-trsfremissimst oml

j
a coal-fuedboilerburningbituminouscoal.Theflue gassfreamis 24 m3lsat 436K.
Analysis'ofthoash,- shows,a,resistivity
of 12 x.l0l"l"ohm"cm,. ifheov,erallef,ficiency a At4iX
reryired.ls99.9%. bAt5q
c At34
r:
Solution Source
ftomJ/
Calculatethe SeA from Eq, (9,18).The fly.astundgr.lirn v,eloclry,is0.160.m/s(seeTable
9.4).Then,SGA'--*llr{1'- 0;999/0.16='43..2m2(mals)'Thetstal.collecforplateareais
thenAc = (43.2)(24)= 1,037m). F-rrl

Er
A more generd:nrcthodtorc,aleil$ate overall efficienciesis"baee-d..
on thg grade-efficiency cq
equation for the'ESPFnd the ,size'distribrrtionfunction of lhe aerosolpopulation e*rteringtho m
device.If the aerosolis log-normally distributed,the calculationof overall efficiency can be done
numericallythroughGaussianquadratureformulasas shownin Chapter7.
s"tq
A
4
11

t
i
I
ll
F
)
i
 387
Sec.9.4 OverallEfficiencY

Toble9.2Plote-wireEsPMigrotionvelocities(m/s)
Efficiency
(vo)
Particlesource 95 99 99.s 99.9
BituminouscoalflY-asha
(no BC) 0.126 0.101 0.093 0.082
(BC) 0.031 0.025 0.024 0.021
Othercoala
(noBC) 0.097 0.079 0.079 0.072
(BC) 0.029 0.022 0.02r 0.019
Cementkilnb
(noBC) 0.015 0.015 0.018 0.018
(BC) 0.006 0.006 0.005 0.005
Iron/steelsinterPlantdust
with mechanicalcollectof
(noBC) 0.068 0.062 0.066 0.063
(BC) 0.022 0.018 0.018 0.017
IncineratorflY-ashc
(noBC) 0.153 0.114 0.106 0.094

aAt 420K. BC = backcorona.

b At 590K.
. Ar 395K.
with permission
Source:From Turner,J.H., et ar.JApcA, 3g:45g(1988). Reprinted
fromJAPCA.

Example 9.9Designof ESP Basedon Grade-EfficiencyEquation

the grade efficiency
Estimate the overall efficiency of the ESP of Example 9.8 integrating
the device are log-normally distributed with MMD = 16
equation.The particles entering
mm and og =3.0. The dielectric constant of the particles is 5'0'

Solution
field strength' E'*= E'=
At a remperatureof 436 K, Eq. (9.8) gives the maximum electric
Table 9'1 givesK = l'26'
433,000 v/m. For a flat-plate configuration with no back corona,
SCA = 43'2 s/m' The sub
Then, Eo'.= 433,O001I'26= 344,000 V/m' From Example 9'8'

,{#iil''
 388 Chap.9 EleetrostaticPrecipitators Sec.9.

Toble9.3wet woll Plole-wireESpMigrotionvelocities(m/s) prog

Emp
Efficiency (%) ic pr
Particle sourcea 9s 99 99.5 gg.g equ
Bituminous coal fly-ash glve
0.314 0.330 0.338 0.249
Other coal 0.400 0.427 0.441 0.3t4
Cement kiln 0.064 0.0s6 0.050 0.057
Iron/steel sinter plant dust
with mechanical collector 0.140 0.131 0.133 0 . 11 6
aAll sourcesat 370 K, no back corona.
Source:From Turner,J. H. et al. JApcA, 3g:45g(19gg Reprintedwith permission
).
fromJAPCA.

Toble9.4FlotplofeESpmigrotionvelocities(m/s).
Efficiency (7o)
Particle source 95 99 99.s 99.9
Bituminous coal fly-ashu 0.132 0.151 0.186 0.160
Other coalu 0.155 0.t12 0.151 0.135
Cement kilnb 0.024 0.023 0.032 0.031
Glass plantc 0.018 0.019 0.026 0.026
kon/steel sinter plant dust
with mechanical collectora 0.t34 0.121 0 . 1 31 0.124
Incinerator fly-asho 9.5CC
0.252 0.169 0.2r1 0.183
a At 420 K, no back corona.
b At 590 K, no back corona. 1'(
c At 530 K, no back corona.
l.
d At 395 K, no back corona.
Source:From Turner,J. H. et ar.JApcA. 3g:45g(19gg). Reprintedwith permission 7.
ftomJAPCA.
rs Sec,9.4 OverallEfficiencY 389

I- program.FIl.NCTION HE'TA(D)calculatesthe gradoefficiencyfrom Eqs-,(9'16)and(9-.17).

bt"piti"uf correlationqto estimatethe viscosity an! the mean,freepathfor air at atrnospher-
ic piqssureasa-functionof temperaturearepart of the sub_prog-rqn1 ThSerade-effici:n:y :
eqgation is inte$ated numerically with 4.16-pointGau$s-Hermite quadrature formula to
givelanoverallefficiencyof 99,67p.,

FUNCTTONHETA(D) :. :| '
REAL LAMBDA, DIEL, EL, SCA, VIS, EO, T, D
PARAMETER (E0 = 8,85E:12)
" DATA T, DrEL,EL, scA /436.0,5.,3.4485,4321,
b 61ryrp1=r.*tz.*r,aveoA/D)*(1.257+.4+EXP(-.SS*p/t-AlWSOe))
erAicpi=tr.*2.*LaMenA/D)*r'2a2.*(@IEL- 1.)/@IEL+2.))/
*: I (1.+2.*LAMBDAIb)
n vIS = 1.72E-5* ("il273.0)** A.71
LAMBDA = 6.71E-l 1*T*'t1.21
IF (0.55*D/LAMBDA.LT. 80,) TFIE,N
CUNING=CLJN(D)
ELSE
CIINING= 1. + 2.*LAMBDA/D
ENDIF
F
ARGUM=SCA*80*EL*'<2*CUNING*AFAC(D)*D(3.*VIS)
.
IF (ARGIJM.GT. 8O.O) THEN
PEN=0.0
ELSE
PEN=E)(P(-ARG[M)
ENDIF
HETA=1.-PEN
END

TOTHEMODEL
9.5 CORRECTIONS

Your objeptivesin'studyrng thiq sectidn are to

1. Identify thosephenomenaaffectingthe performanceof the ESPnot
incorporatedin the grade-efficiencymodel.
2. Define the quality factor of the flow distribution andincludeit in the
efficiencycalculations.
3. Define the lossfactor owing to reenffainmenJ andgassneakage'and
includeit in the efficiency calculations.
390 Chap.9 ElestrostaticPrecipitators Sec.9.

The experimental values of the effective migration velocity do not behave in all respects
as predicted by the the.oreticalmodel. There are various aspectsof the performance of an ESP not
incorporated in the model. These are sometimesreferred to as non-Deutschianphenomena.Some
of the most significant arenon-uniforrn gas velocity distribution; gas sneakage; and.rapping reen- this manr
trainment. Possible corrections to the model for these items follow. equivalen
to the "a1
9.5.1Non-Uniform
GosVelocityDistribution Equation

The theoretical model developed assumesthat the fluid velocity parallel to the collector
surface is everywhere the same and equal to the volumetric flow rate divided by the total cross-
sectional area. This condition is almost impossible to achieve in practice. Uniform, low-turbu-
lence gas flow is very important for optimum precipitator performance.The detrimental effect of
non-uniform gas flow is twofold. First, owing to the exponential nature of the collection mecha-
nism uneven treatment of the gas lowers collection efficiency in the high-velocity zones to an Solving F
extent not compensatedfor in the low-velocity zones. Second, high-velocity regions near collec-
tion plates can sweepparticles back into the main gas stream.
Although it is known that a poor gas velocity distribution results in reduced collection
efficiency, it is difficult to formulate a mathematical description for gas flow quality. The follow-
ing development, presentedby McDonald and Dean (1982), demonstratesthe general considera-
tions to be made in accounting for the effects of a non uniform gas velocity distribution on collec-
tion efficiency. Eq. (9. 18) can be writren as

1
= expl;:_)= *n (-
Ptid (e.1e) yield valu
# When car
where achievedt
Pttd = ideal overall penetration
Ar = inlet crosssectionalarea
va = ayat&Eagasinlet velocity Exampl
k = A,w/At
(a)
The precipitator can be divided into a number of imaginary channels corresponding to
pitot tube traverse points. The penetration for all the channels can be summed and averagedto The
obtain the mean penetration with an actual velocity distribution instead of an assumeduniform Cal(
velocity. (b)'
The
NNl.

h=-: \ ryPr1'd=
.,1- \ rir-i,, Solutiol
t\vaj=l lvro
(e.20)
j=t

(arI
where the r
 ,Corrections
Sec.,9.5, to the Model 391

N = number of points for velocity traverse

vy= point values of gas velocity
For any practical velocity distribution and efficiency, the mean penetration calculated in
this manner will be higher than that calculated based on an average uniform velocity- This is
equivalent to a reduced "apparent" migration velocity. The ratio of the original migration velocity
to the "appareni" migration velocity is called the qualityfactor Q, of the velocity distribution.
Equation (9,20) can, then; be w,rittenas:

/,\ _k
pt= exp
l- {l= + vje vj
Nvo (9.2r)
\ 0u,/ j =t

Solving W. Q.2l) for the quality factor,

'
d=- ---ft --
t" (pr) (e.22)

2,,"t)
".*[rt
The use of Eq. (9.22) with measured velocity traverses in a working precipitator may
yield values of $ as high as 2 to 3 in a poorly regulated flow pattern (McDonald and Dean 1982).
When care is taken, through scale model studies, values of Q as low as l.l or 1.2 should be
achieved (Licht 1980).

Example 9.10 Flow Quality Factor Calculations

(a) The ideal overall penetration for an ESP is l/p 255uming a uniform velocity OistriUuiion.
2, v2 = vo, and vZ= 3vol2.
Calculate the flow quality factor and the actual overall penetration.
(b) The ideal overall penetration for an ESP is 17oassuming a uniform velocity distribution.
The flow quality factor is 2.0. Calculate the actual overall penetration.

Solution

(a) FromEq. (9.19),Hvo= - (n Pti\ = - ln (0.01)= 4.605.Thefollowing tablesummarizes

the calculationsfor the quality factor andactualpenetration:

i, l,*-
392 Chap.9 EleetrostaticPrecip|tators Sec.9.

T
agesetsal
T
1.0 material*
0.01 0.01
section.Tl
1.5 0.0447 0.06705
I = 0.07710

FromEq.(9.2L),Pt =ElN = 0.M7ll3 = 0.0257.Equation(9.22)yields

0 = ln (0.01)/ln(0.0257)
= 1.258. whereRR

(b) From Eq.(9.22),Pt = 1p6a)r/O= (0.01)0.s

= 0.10.

wherelf
Commenls
Theseexamplesemphasize theimportanceof a uniformvelocitydisribution in an ESP's
performance.Partb showsa lO-fold increasein penetrationfor a flow quality factor of 2.0! I
However,Hein (1989)developeda computermodelthat showsthepossibilityof improving =
andLF
the performanceof an ESPby using a controllednon-uniformgasdistibution at the inlet
mumlntr
and outlet facesof the precipitatorwhenthereare significantrappingreentrainmentlosses.

9.5.2Sneokogeond Roppingneentroinmenl

Sneakageand rapping reentraihmentlossesare best consideredon the basisof the sec-

tions within an ESP.Considerflrst the effect of sneakage.'Assuming
that the gasis well mixed E---f
betweensections,the penehationfor eachsectioncanbe expressedas:
Er
b
P/"= 5"+ (t -s,y)rr.(E) (e.23) lr
L
where
P/r= ssslion'soverallpenetration sffi
S7y= fraction of the gasbypassingthe section(sneakage)
Q' - gasvolumeflow in the collection zone= O(1 - Srs) El
Pt" (Q') = overall penetrationin the collection zone 5t
 to the Model
Sec.9.5 Correetions 393

The penetration for the entire ESP is the product of the section penetrations. The sneak-
ase sets a lower limit on the collection efficiency through each section.
The collected dust accumulates on the plates until they are rapped, when most of the
material falls into the dust hopper. A fraction of it is reentrained by the gas flow and leaves the
section. The average penetration for a sectionoincluding sneakageand reentrainment is

p/,= 5" * (t - s7v)rr"(o - rr"(E)]

) + nn((r- sru)fr (e.%)

whereRR = fraction reentrained.Equation(9.24) canbewritten as

Pt,= 7P 1(t -tr)rt"(O) (e.2s)

whereLF (lossfactor) is definedas

L F=SN+nn- ( su) kn) (e.26)

Fly-ashprecipitatorsanalyzedin this way have averagevaluesof,S; = 0.07, RR = 0.12'

andLF = 0.182(Turneret al. 1988a). To achievea given overallpenetration,choosethe mini-
mum numberof sections,.{", suchthat:

*1"4
r"' ' lnlLF)
(9.27)

Example 9.11Sneakageand Reeatrainment Losss

Estimatethe minimum numberof sectionsfor a fly-ash ES? if the overall efficiency must
be99.9Vo.Assume thatthe lossfactorowingto sneakageandreentrainment is 0'182.
Assumingt}ratall the sectionshavethe same estimate
overallpenetration, the penetrationin
the collection zoneoPt", of eachsection. I '.

Solution

Equation(9.27)givesN" > ln (1 - 0.999)/ln(0.182)= 4.06"Therefore,the ESPshouldhave

5 sections.Assumingthat all the sectionshavethe sameoverall penefiation,
394 Chap:9 .,Eleetrostatic,Precipitators Sec.9.

'l
/, ,,,
pt\r,,ril _f-:_
for a
^ .F
pt"=-_: (9.2g) For i
-*__
1_LF

in F.q.(9.28),p/c = [(0.00I)0.2-0.1,s2]t(t-0.tg2) = 0.0g46.

Substituting

Example9.L2corrections,to PenetrationPredictedby the ESp Model

The collectionareaof the ESPof Example9.9 is equally divided into six sectionsin the
directionof the gasflow. Valuesfor sneakage andrappingreentrainment arc lTVo and.i,2Vo,
respectively;theflow qualityfactor'is't.Z-nitirnatethe actualoverallcollectionefficiencv
for theESP.

Solution

lhe penetrationpredictedby ,Feldman'smodel,Ptid, mustbe correctedfor non-uniformgas

flow distribution,numberof sections,andreducedgasflow-owing to sneakage-through
thecollectionzone.Thecorrectedvaluecorresponds to pt", andisgivenby

r - s")l
,, =(rr'i)t'lon"( (9.29)
"
Substituting .2x6x0.e)1
in Eq. (9.29),Prc= (0.004)vt(r = 0.426.Thelossfactoris IF = 0.10+
-
0.12 (0.1X0.12) -- 0.208.FromEq.(9.25),Pt,=9.545,thenpt= (0.545)o=0.026,mdnar
=97.4V0. , ,

Example9.l3 DesignofnsP. Including Gorrectionsto Model

::r, i .:':: .:j,.tt- ,.
.'..- .;- ;,-: :ji. , , , ..1 ..:. .i
.1.. ,: r,;:,,.:l:. -i: I ,

': '. iConsidertheeonditionsof Exar.nptres

9.9 and:9.12.Redesign,the ESPto,maintain,an
actual
overallcollectionefficiency
sf 99.6Vo. :: ,,:i.',,' '^ i: :

Solution

Equation (9.27) givesthe optimurn nurnber of sectionsthe ESP should have: N" > ln
(0.004/ln (0.208) = 3.52. Therefore,N, = 4 sections.Equation (9.28) gives rhe actual over-
all penefation in the collection zone of each of the ESP's sections:Pr" = [10.664;o.zs
-
- =
0.2081(1 0.208) 0.055. Eq. (g.2g),appliedto a single section, givesPxid=PrcO(l-sff)= !4

(0.055;t'os0= 0'0468. The computer program ESP calculatesthe collection areaper section
rS to the Model
Sec.9.5 Corrections 395

for a given value of Pfd, as predicted by Feldman's model. The answer is 3/4.3 m2lsection.
For four sections.the total collection arcais L,297.2 nP.

PROGRAMESP
c
C CAITULATES THE COLLECTION AREA REQT]IREDFOR A GTVENIDEAL
C PENETRATION,PENID, AS PREDICTEDBY FELDMAN'S MODEL
c
REAL PENID, MMD, NMD, SICG, Q, LAMBDA, VIS, Xl, X2' SL
REALRTBIS, XACC, EL, CT]N,AFAC, PEN, DIEL
INTEGERN
LOGICAL SUCCES
COMMON/BLOCKI/ PENID,MMD, SIGG,Q, LAMBDA, VIS' EL, DIEL' EO'
EXTERNALFLINCP
CUN(D)=l.+(2.*LAMBDA/D)*( 1.257+.4*DG(-.55*D/LAMBDA))
AFAC(D)=(1.+2.*LAMBDA/D)**2+2.*((DIEL-l.) KDIEL+Z'))l
* (I.a/.*LAMBDAID)
*uprD)
X I --3. *VIS'<Q*LOG(PENIDy(E0:i'EL**2*cUN(IvIIvrD)*AFAq@MD)
X2=SIGG*Xl
CALLZBRAC(FIJNCP,Xl, X2, SUCCES)
IF (SUCCES)THEN
XACC=ABS((xI -x2y10000.)
SIFRTBIS(FUNCP,Xl, X2, XACC)
', '
PRINT *, ' COLLECTION AREA PERSECTION SL, SQ METERS'
GOTO 10
ENDIF
'
PRINT *, 'FAILIJRE IN BRACKETING SL
10 sroP
l
END
c
I
BLOCKDATA
REAL PENID, MMD, SIGG, Q, LAMBDA, VIS, EL, DIEL, EO
COMMON/BLOCKI/PENID, MMD, SIGG,Q,LAMBDA' VIS' EL' DIEL' EO
DATA PENID/0.0468/,MMD/16.0/, SIGG/3.0/,Q/24.00/,LAMBDA/
* 0.1050/,vlsl2.4E-5l,EU3.44E5l,DEU5./,E0/8.8 . '5 E - 1 8 /
END
c-
orNcrro*FUNcP(sL)
X, A
REAL PENID,MMD, SIGG,Q, LAMBDA, VIS, EL, DIEL' EO,.PEN;
REALM\4D, EP.S,PENC,F{.INCBSL ..
INTEGERN
COMMON /BLOCKI/ PENID, MMD, SIGG, Q, LAMBDA, VIS, EL, DIEL, EO
PARAMETER(N=10,EPS=1.0E-06)
DIMENSION X(I'T),A(N)
t96 Chap;,9, Electrostatic
Precipitator:s Sec.9

'
CIJN(D)=l.+{2.*LAMBDAID)*(l.257+.4*EX(..55tD_&AMBDA)) 9.6.tSi
AFAC(D)=( l.+2. *LAMBDA/D) * t2t?. * (DmL- L) t(DEL+2.)) |
* (1.+2.*LAMBDATD)
CALL HERMIT(N,X, A, EPS) ,,,, ,,:a
NMD=E)G(LOG(MMD)-3*(LOG(SIGG))**2)
svstemto
,SUM,A=0;0 may'rc.i
DO I0I=1,N :: r :, melal of.r
' kept'taut
D=DG(SQRT(2.){'LOG(SIGG)*X(D+LOC(NND)
IF (0.55*D/LAMBDA .LT. 80..) THEN . Recent,d
CUNINHUN(D), . i'
, ,1
ELSE . ,i : ther,'!:aS
CTJNING=1. + 2.*LAMBDA[ dry.'pr-tir
]
ENDIF becorlu*
ARGLJM=SL*E0*EL{.*2*CUNtr{G*AFAC(D)*D/(3. tq*y15'
plate.$ir
rF (ARGTJM.cT. 80.0) THEN
PEN=0.0 :
' '.,','f
inthe,lag
EI.SE r' , r'. .. . ' .r..'.
PEN=EIP1-4116g14) '',, trodes.lA
ENDIF standard
SLMA=S{JMA+A(I)*PEN*E)(P(3.*SQRT(2.)*LOG(SIGG)*X(D)
- ' ': ' I
- i' -r-r '' : ' no regard
lo CONTINUE
PENGSUMA(SQRT(3.1416)*EXP(4.5*(LOG(SrGG;**2;;' trainrnen
FUNCEPENC-PENID ' r,:. . affanged
END ,l 1

9.6 PRACTICAL
DESIGNCONSIDERATIONS
where
I
I
Your,objegtiv,esin,shrdying this seclion are to Sitrcc-

1. Estimatethedimensioras:of anESP,given'thecollectionarea..,,
2. Estimatethe totalpower.consumption:of an ESP. , ,:!
3. Understandthe importanceof flue gasconditioningto maintain
properresistivityof thecollecteddust. qrkcni

The completedesignof an ESPincludessizinganddeterminingthe configurationof the

collectingaf diylryrse,glectrglel, calculating*r?p,oqerc-o.4sumption,
an{ ggecifyingrapping,
dustremoval,andflue gasconditioningsystems.
Sec',9.6' PracticalDesignConsideiations 397

9.6.1Sizingthe eleclrodes

., The dischargeelectrodesystemis desig.nedin conjunctionwith the collection electrode

systemto maximize the electric current and field stren$tr. The shapeof the dischargeelectrode
may be in the form of cylindrical or squarewires, barbedwire, or stampedfrom formed strips of
metalof variousshapes.Most ESPsin the U.S. usecylindricalwires of about2.5 mm diameter
kept taut by weights,whereasEuropeandesignsfavor rigid, mast-typesupportsfor the wires.
Recentdesigns!n bothcontinentsuserigid dischargeelectrodes (CooperandAlley 1986).
The numberand sizeof the collectionplatesdependon the total collection areaspecified,
the gasflow velocity,the platespacing,andthe aspectiatio (length-height). An ESPcollectinga
dry particulatematerial runs a risk of nonrapping,continuousreentrainmentif the gas velocity
becomes, ttremaximumacceptable
too large.For fly.ash applications, velocityis about1,5m/s for
plate-wireESPsandaboutI m/sfor flat-plateESPs.
Most U.S. utility precipitatorshave 230-mmplate spacing,althoughsystemspurchased
in:thelast few yearsgenerallyhave300-mmspacingwhich allowsthe useof rigid dischargeelec-
trodes.A few U.S. utilities are currentlyinstallingESPswith the new Europeanplate spacing
standardof 400 mm (Offen andAltrran 1991). :
The DeutschequationrelatesESP performanceto specific collecting surfaceareawith
no regardto the length-heightaspectratio. However,becauseof gassneakage andrappingreen-
trainment losses,the actual precipitator's performance is.dependent on how the surfaceis
arranged. For collectionefficienciesgraterthan997o,theaspectratio shouldbe keptabove1.0.
The numberof ductsfor gasflow, Nd, is given by

o
N4= ---:- (e.30)
2WvoH

where
W = wire-plate spacing
H =plate height
Since each plate has two collecting surfaces, the total collection area is given by

. ,a

' :r: .I ::
A"=2N;RH'. i .: ,. (e.31)

wtere R is the aspectratio.CombiningEqs.(9.30)and(9.31),theplateheightis

(SCR)uow
s = (e.32)
.R
7

398 Ghap.,9 ElectrostaticPrecipitators Sec.9.

Example9.14 Dimensionsof Plate-Wire ESP The follow

Estiinatethe dimensionsof a plate-wireESPtreating 333 m3lsof gaswith a total i2latearea

: -. .::
, o! !+,OO_O m] for 99!o collection of fly-ash. Platespacingmustbe 300 mm to allow the use
of rigid dischargeelectrodes.
where 142
Solution
.]
Choosetypicalvaluesof gasvelocityandaspectratio:yo=,1.5rnls,R = 1.0.CalculateSCA Exampl
, -- A/Q - I4,W01333= 42.0 s/m.For a wire-platespacingof 150mm, Eq. (9.32)gives11=
' Esti
9.45m. Sincethe aspectratio is unity, Z = H = 9.45m: Equation.(9.30)
givesNa=78..3;
chooseNa=79ducts.Thewidthof theESPis (79X0.30)=23.7 m. fan r
Becausethe numberof ductswasroundedto the next integer,the actualcollection areais
,
.:1. i Solutiol
on collectionefficiency(14,110m2versusl4,0qg m2),and
, , higherthanthat specified.based
thegasvelocityis lower thanthevalueinitially chosen(1.49m/sversus1..5rls). Both fac-
Ass
, tglr. improvethe performanceof the ESP.Figure 9.5 is a sebematicdiugtumof theresulting
design. (e.
' 63

9.6.3H

I
1,-,^ (of thfl:
alwayscr
marir
be rrd
faces"ft
1

ofHq
Figure 9.5 Dimensions of
the ESPof Example 9.14 ofcd
Fotu
comh
affi
9.6.2Powerconsumplion .:. fad
in tug
The two main sourcesof powerconsumptionof an ESParefrom coronapower andgas
rcqnfu t
pressuredrop. Total pressuredrop for a precipitator and associaledductwork is usually of the
order of 1.0kPa (Turner et al. 1988a). Coronapower is a strongfunction of overall penetration.
llqr rr
opcrft
I
t

{
1
s Sec.9.6 PracticalDesignConsiderations 399
a

The following correlation is basedon actual operational data presentedby White (1984):

w.= * l'LZl
' -ohts.s
\ (e.33)
Ptl

where 142"is the corona power in watts.

Example9.15Power Consumptionof ESP

Estimate the total power consumption of the ESP of Example 9.I4. Assume that the motor-
fan efficiency is60Vo.

|5
Solution

= 555kW. Equation
dropof 1.0kPa.Thefanpoweris (333X1.0y(0.6)
Assumea pressure
(9.33)givesthecoronapower, kW.Thetotalpoweris
w" =333(t15.8+1.1710.01)=l'1.5
632.5kW.

9.5.3FlueGos ConditioningSystems t

The major factors affecting fly-ash resistivity are temperatureand chemical composition
(of the fly-ash and of the combustion gases).For a given chemical composition, fly-ash resistivity
always a distinct maximum at temperaturesabout 395 to 450 K' The temperatureof the
"^hiUitt cannot
maximum resistivity is unfortunate for power boiler operators.ESP operating temperatures
without risking condensation of sulfuric acid on some of the colder sur-
be much below 395 K
at temperatures much higher than 450 K results in unnecessary loss
faces. Conversely, operating
of heat out the stack, reducing the thermal efficiency of the plant (Cooper and Alley 1986)'
Resistivity decreaseswith increasedfuel sulfur content. becauseof increasedadsorption
(>2Vo sulfut)
of conductive gases,such as SO:, by the fly-ash. The combustion of high-sulfur coal
produces sufficient SO3 to condition the fly-ash to a low resistivity, as shown on Figure 9.6. The
small
combustion of low-sulfur coals (<1% sulfur) with highly alkaline ash produces only
in very high ash resistivities. In the intermediate case, the existing data
amounts of SO:, resulting
for medium-sulfur coal exhibit considerable scatter in the relationship between the SO2 and SO3
1990
in the gas entering the ESP (Harrison et al. 1988). The Clean Air Act Amendments of
2000'
require electric utilities to reduce SO2 emissions by approximately 10 million ton/yr by
\lan1' eristing plants are switching to low-sulfur coal, with negative impacts on the electrical
operarionof ESPs.

ilmlF

I"
-
400 Chap.9 Precipitators
Electrostatic Sec.9

(
l012
agents us
gaseous€
Low-sulfur coal available
10rr

b 9.7 CC
High-sulfur coal
1010

Figure 9.6 The effect of Yo

10e
temperature and fuel sul-
1.
420 M0 460 520 fur content on fly ash
,:
resistivity
(K)
Temperature

Other SO2 "dry" reduction strategies,such as spray dryers, furnace sorbent injection, and costsfor
conversion to fluidized-bed combustion, have a significant impact of ESPs located downstream.
These techniquesincreasethe particulate loading and changethe resistivity of the collected mater- 9.7.1I(
ial, generally making it more difficult to remove. Durham et al. (1990) reported resistivity values
of sorbent-fly-ash mixtures of the order of 107to 10eohm-cm at spray dryer conditions. Although
the injected material had a high resistivity at 420 K, the surface conditioning provided by the iary equ
increasedmoisture and cooling was sufficient to reduce it dramatically. At such low resistivities, sitefacil
the electrostatic force holding the particles onto the collector plates is reduced and the particles
are easily reentrained.They found that reentrainmenthad to account for 6OVoof the observedpen- ( 1 9 8 8b r
etration per section to correctly predict ESP performance at spray dryer conditions (see Problem ing area
e.19).
The cohesive characteristicsof the sorbent-fly-ash mixture improve by using additives
such as ammonia. The interaction of ammonia and SO3results in the formation of ammonium
bisulfate and sulfate, which condenseon the surface of the particles. These surface deposits are where
viscous and cohesive,which reduce rapping and nonrapping reentrainment.
The atmospheric fluidized-bed combusror (AFBC) is an emerging technology for the
control of SO* and NO* emissions from coal-fired power plants. Crushed coal is fed into a bed of
inert ash mixed with limestone or dolomite. The bed is fluidized by injecting air through the bot-
tom at a controlled rate. The coal burns within the bed and the SO* formed reacts with the lime-
Toble
stone to form a dry calcium sulfate which is removed with the fly-ash in a downstreamparticulate
collection device. The system can remove up to 95Voof the SO2 and up to 80Voof the NO* emis- Plate ar
sions (U.S. Congress,1991).However, Altman (1988) reportedsevereESP operationalproblems 930to -l
at a full-scale coal-fired power plant converted to AFBC operation due to the high resistivity of 1,6m
the collected dust. Flue gas conditioning to reduce the particulate resistivity was been considered. Source
JAPCA
Sec.9.7 eostingConsiderations 401

ESPs' Conditioning
Gas conditioning equipment is frequently used to upgrade existing
A typical dose rate of the
agents used include sor, grsoo, sodium compounds, and ammonia.
systemscommercially
gaseousagentsis 10 to 30 ppm by volume. There are several vendors and
gives good results with reasonably small expense'
available. Gas conditioning usually

CONSIDERATIONS
9.7 COSTING

Your objectivesin studying this sectionare to

1. EstimatetheTCI for varioustypesof ESPs.
precipitation.
2. EstimatetheTAC for electrostatic

This sectionpresentsmethodsto estimatethe total capitalinvestment and total annual

costsfor anESPsystem.This materialis takenprimarilyfrom Turneret al' (1988b)'

9.7.rTCI
The TCI includes costs for the ESP structure, the internals, rappers,power supply, auxil-
and off-
iary equipment, and the usual direct and indirect installation costs.Land, working capital,
site facilities are not normally required.
et al'
ESp equipment costs are almost always correlated with the collecting area. Turner
(19gS b) obtained cost quotes from precipitator vendors and regressedthem against their collect-
ing areas.These costs,updatedto June 1990, are given by

EC= aA! o34\

Ac = collectingarea,mz
c, b = regressionparameters
Table 9.5 lists the values of a and b for different rangesof collecting area.

Toble9.5 RegressionPqrometersfot ESPEquipmenlCosf Equoliono

Plate area (m2) e b
930to 4,600 ,551
1.600to 93,000 715 0.8431
from
). Reprintedwith permission
Source:FromTurner,J. H. et al.JAPCA,38:715(1988
J.|PCA. ^Updatedto June 1990.
 Chap.9 Precipitators
Electrostatic Sec.9.
402

These costs apply to all ESP configurations, except the two-stage. They include the fol-
lowing: ESP casing, pyramidal hoppers,rigid electrodesand internal collecting plates, trans- Toble9
former-rectifier (TR) sets and microprocessor controls, rappers, inlet and outlet nozzles and dif-
fuser plates, weather enclosure and stair access, structural supports, and insulation (7'6 cm of Costs
fiberglass encasedin a metal skin). 5E
Two-stage ESPs are usually limited to small sizes and are sold as modular units. To be ES
consistent with industry practice, the equipment costs (in $ of June 1990) are given as a function Inr
of flow rate through the unit (Turner et al. 1988b), Ta
Frt
EC = 21,200+ 41,500lnQ 1.0<Q,d.O.o Pu
(e.35)
Ins
Totr
where Q is the actual volume flow rate through the unit. This cost is for modular units fully Indi
assembledmechanically and electrically, and mounted on a steel structural skid. TC]
The purchasedequipment cost, B, is the sum of the equipment cost, auxiliary equipment,
instruments and controls, taxes, and freight. TCI is estimated from a series of factors applied to Source: I
the purchasedequipment cost to obtain direct and indirect costs for installation. Table 9.6 summa-
rizes the most important factors for average installation conditions. For two-stage ESPs, pur-
chased as packaged systems,installation costs are greatly reduced. Turner et al. (1988b) suggest ash
that, in this case,TCI = I.258. Esti
oP'e
$0.
9.7.?TAC retu
equ
Direct annual costs include operating and supervisory labor, maintenance (labor and
materials),utilities, dust disposal, and waste water treatment for wet ESPs. Typical operating Solutio
labor requirements are 2 h/shift. Supervisory labor is taken as l5%a of operating labor'
Maintenance labor is estimated as 660 h/yr; annual maintenancematerials are approximately IVo The
of the purchasedequipment cost. I <r
Ifthe collected dust cannot be recycled or sold, it must be disposedof. Disposal costs for -++-a
nonhazardouswastes are typically $3O/metricton. Disposal of hazardouswastes may cost 10 (50
times as much. If the dust collected can be reused or sold, a recovery credit should be taken.
For capital recovery calculations, ESP systemsare assumedto have averagelifetimes of
-{>\
20 yr. Overhead is calculated as 60Voof total labor plus maintenancematerials. Property taxes,
fan
insurance,and administrative chargesare estimatedas 47o of theTCI.
[{:']
,l-ri

ttf :
Example9.16Capital Investmentand Annual Cost of ESP ann

Assume a plate-wire ESP is required for the control of fly-ash emissionsfrom a coal-fired
boiler burning bituminous coal. The flue gas streamis 26.3 m3/s at 435 K and has an inlet
 Sec.9.7 CostingConsiderations 403

Toble9.6CopitolCostFoclorsfor ESps
Costs Factor
Direct costs
ESP and auxiliary equipment costs A
Instrumentsand controls 0.10A
Taxes 0.034
Freight 0.05A
Purchasedequipment cost B = 1.184
Installation direct costs 0.678
Total direct costs 1.678+SPu+Bldgb
Indirect costs 0.578
TCI 2.248+SP+Bldg
Source:Vatavuk and Neveril (1980). usite Preparation. bBuildings.

ash loading of 0.01 kg/mz. The design SCA for 99.9Voovenll efficiency is 62.4 slm.
Estimate the TCI and TAC for this application. Assume that the unit operatesfor 8,640 h/yr.
Operating labor rate is $15/h, maintenancelabor rate is $20lh. The cost of electricity is
$0.061kw-h, dust final disposal costs are $30/metric ton. The minimum attractive rate of tne
return is l2Volyr. Assume that no buildings or special site preparation is needed.Auxiliary
equipment needed-ductwork, fan, motor, and stack-cost $65.000.

Solution

The total collection areais (62.4)(23.6)= 1,473 np. From Eq. (9.34) and Table9.5, EC =
4,551 (1473)0'6276 - $443,100.The sum of the ESP and auxiliary equipmentcosts
is A =
443,100+ 65,000= $508,100.From Table 9.6, the purchasedequipmentcost is B = 1.18
(508,100)= $600,000.TCI=2.24 ($600,000)= $1,344,000($ of June 1990).

Assuming a total pressuredrop of 1.0 kPa and a mechanical efficiency of 60 percent for the
fan, the total fan and corona power requirementsare given by t(1.0)(23.6)10.61+23.6
[0.1158 + (0.00117/0.001)]= 70 kW. The total massof dust collectedyearly is (26.3X0.01)
(0.999X3,600X8,640y1,000 ='l ,330 metric ton. For a useful life of 20 yr and a yearly rate
of return of l2Vc, the capital recovery factor is 0.1339/yr. The following table presents
annualcostsfor this project.

it

lli
I
I

I
-
I
 404 Precipitators
C h a p . 9 Elbctrostatie Referen

Item Annualcost(,$lyr)
Direct annual costs
: =
(3X360X15)
Operatinglabor: 76,200',,
Supervisory: =
(0.15X16,200) .2,400
..
, 8,20A
ancelabor:(660)(20)='
=
(0.01X600,000)
materials:
Maintenance 6,000
c
Electricity:(70x8,640x0.06)= : 40.000 .,8
.
Wastedisposal:(7,330)(30)= 2r9.900
Total direct annualcosts 2y7,700
lndirect annualcosts I
, ., Overhead:(0.6X37,800)= .:::,.::. : :: ..,.
22,68A.,..,.. A
=
(0.04X1,344,000)
Tax,insurance,administration:' 53,760
I
=
cost:(0.1339X1,334,000)
-Capitalreco-very 178.600 A
Total !r{ir..eclannualcosts 255,040,
TAC :1 ,':', , , $552,740 T
I

I
J

9.8 CONCTUSION : :::'

i :.::
I
The selectionof particulaieconffol equipmentin the nearfuture will dependlargelyon
regulatorytrendsand the technicaland commercialsuccessof ongoingparticulatecontrol
i
researchand developmentefforts.Well-designedmodernESPscan meetemissiontargetslower
- thanthe currentNSPS.Requirements for greaterparticulatereductionsor stringentemissionlim- .L
I
its basedon respirableparticulatesmay'limit the useof conventionalESPs.Severaloptionsfor
significantlyenhancingprecipitatorscollectionefficiencypresentlyunderactiveconsideration are
tunedspecificallyfor fine particlecontrol;andthe additionof a very {
wet ESPs;pulseenergization
compactpulsejet baghouse(to be describedin detailin Chapter10)as a polishingstepfollowing
* ttt'
Errs have been,thgworkhorseof p#culate control,equ.ip fo, ulmoutu ry.
"ettt
and manufacturers
However,their future is uncertainand will dependon the ability of researchers
to retain the reliability and simplicity of currentdesignswhile improving their toleranceto varia-
tionsin particulatecharacteristics.
lors References 405

: REFERENCES

Eng.Prog.r72:47(1976).
Abbott,J. H., audDrehmel,D. C, Che.m.

AltmanR. F.,|APCA,38:1455(1988).

Coclret,R. Colloq. Intern. CentreNatl -Rech..Sci.[Paris],l02z33I ( 1961).

Cooper,C. D., andAlley, F. C. Air PollutionControl:A DesignAppronch,PWS

Boston,MA (1986).
Engineering,

Crawford,M. Air PollutionControlTheory,McGraw-Hill,New York (1976).

Durham,M.D., Rugg,D. E., Rhudy,R. G., andPuschaver,

E. J. "/.AirWasteManage.
(1990).
Assoc.,4O:112

Feldman,P. L. PaperNo. 75-02.3,68thAnnualMeeting,Air PollutionControl

Association,Boston,MA (1975).

Flagan,R. C., andSeinfeld,J.H. Fundamentals

of Air PollutionEngineering,Prentice
Hall,Englewood Cliffs,NJ (1988).

Harrison,W. A., Nicholson,J. K., DuBard,J. L., Carlton,J. D., andSparks,L. E.

JAPCA,38:209 (1988).

(1989).
Hein,A. G. JAPCA,39;:766

/on Licht, W. Air Pollution Control Engineering:Basic Calculationsfor Particulate

rrol Collection,MarcelDekker,New York (1980).
fer :
im- McDonald,J. R., andDean,A. H. ElectrostaticPrecipitator Manual,NoyesData
for Corporation,ParkRidge,NJ (1982).
tre
Ery (1991).
Offen,G, R., andAltmanR. F. J,Air WasteManage.Assoc.,.4l;222
lmc
I
Turner,J. H., Lawless,P. A., Yamamoto,T., Coy, D. W., Greiner,G. P.,McKenna,J.
Fv. D., andVatavuk,W. M. JAPCA,38:458 (1988a).
Icrs
iia- Turner,J. H., Lawless,P. A., Yamamoto,T., Coy, D.W., Greiner,G. P.,McKenna,J.
I D., andVatavuk,W. M. JAPCA,38:715 (1988b).
I
r$06 Chap.9 EleotrostaticPrecipitators Problen

U.s. congress,office of rechnologyAssessrnent,

EnergyTechnologychoices:shaping 92r.Ih
Our Future,OTA-E-493,Washington,DC (1991).
As
Vatavuk,W. M., andNeveril,R.B. ChemicalEngineering,pp. ls!-Ilz(November 3, 0.25m3A
1980). tic field
theESP.
White,H. J. 'Control of Particulates
by Electrostatic
Precipitation,"Chap.12in
Handbookof Air PollutionTechnolclgy, S. CalvertandH. M. Englund(Eds.),Wiley,
NewYork(1984).

White,H.J.Industrial ElectrostaticPrecipitaion Addison-Wesley,

Reading,MA 93b. Etil
(1963).
Cor
dropm
the plm

9.{r.Ca
PROBTEMS Ap
andfud

The problemsat the end of eachchapter have been,groupedinto

four classes(designatedby a superscriptafter the problem num-
: 9.S.Ar
ber)
F{i
Class a: Illustrates direct numerical application of the formulas in the 9.4. Tbr
text.
Class.b: Requires elementary analysis,of physical situations, based on
the subjectmaterial in the chapter.
9JlAr
Class c: Requires somewhatmore mature analysis.
C1assd: Requirescomputer solution. nq

9.14.Terminal velocity of a charged particle in an electric field :

RepeatExample9.1 for 2.Qpm diameterparticles.Assumethattheparticlecharge 9.7r- Flt
is proportionalto its surfacearea.Estimatethe telminal velocity of'the particles.
Dcl
Answer:0.375m/s
Problems 497

9.2r.Dimensionsof a.tubular ESP .

A small tubularESPis to:produce:acollectionefficieney of 99Vowhenhandling
025 m3/sof air at 773 K and I atmcontaining2.5-trtmparticles. The strengthof the elec-
tric field is 200,000V/m andtheparticlechargeis 10-15C. Determinethe dimensionsof
theESP.Thevelocityof the gasesthroughtheprecipitatormustnot exceed0.318m/s.
Answer:L=1.28m

9.3b.ESP overall efficiency to satisfy NSPS

Considerthecoal-firedpowerplant of Example5.1. Assumingthat25Voof the ash

dropsout of the furnaceasslag,calculatethe efficiencyof anESPto removefly-ashif
the plant is to meetthe 1980federalNSPSfor particulates., : ,, :
'
Answer:991.6|%

9.4u.Corona onsetvalue in plate-wire ESP

A plate-wireESPhandlesair at 700 K and 101.3kPa.The platespacingis 460 mm
andthe diameterof the dischargewiies:is5 mm. Estimatethe coronaonsetvoltage.
, r:.

9.5a.Averag€:electricfield in a plate-wire ESP

Estimatethe averageelectricfield andthe operatingvoltagefor the ESPof Problem
9.4.Theresistivityof the collecteddustis 1012ohm-cm. : '
Answer:I19 kV/m

9.@.Averageelectric field in a flat plate ESP :

:.'..,
RepeatProblem9.5butforaflat-plateESP'confiluration.: j
_:.

9.7r-Particle saturation charge

Detenninethe time constantand saturationchargefor freld chargqng
of 5.0-pmpar-
408 Chap.9 Eleotrostatic:Precipitators Proble

l0l6'ions/m3and:the
ticles having a dielectricconstantof 4.0 if the ion conCentration,is 9.13b.Ef
field strengthis 150kV/m. Assumethattheion mobility is2.2x lF m2lV-s. ,, se'l
. .: I ', :r ,. :l{nswer:0.21fC Deutscb
emission
,- ' , -' : l
: i:;

9.88.Total particle charge according to Feldman

Estimatethe total chmgeon theparticlesof Problem9.7 according'toEq. (9.14)for where
a chargingtime of 0.1 s. Assumethatthe gasmean-freepathis 0.1 pm.
Scl
Answer:
0.215
fC
Q=
9.9b.Effect of gasflow rate on ESP grade efficiency KW
by 20 percent,esti-
ConsidertheESPof Example9.7.If the gasflow rateincreases ,S=
matethe percent
increasein penetrationfor 1.O-pmparticles. ,An
Answer:66,5Vo
,(a),
tur-,q
witham
9.10b.Effectof average,electricfield strength on ESP performance estim&t
Considprthe ESPof Example9.7.If theaverage by
electricfield strengthdecreases
20Vo,estimatethe-percentincreasein penetrationfor tr.0-pmparticles. (b)
Answer:200Vo xatio,iti

9.11.s.Effective migration velocity

Estimatethe effectivemigrationvelocity for the fly-ash of Example9.9. 9.1'fr.(l
h
donetrc
9.12^.ESPfor the control otp-missions,fronr a cementkiln , ; ,
.:
The g4ses.froma cementkiln flow at a volumetricrate of 380 m3/minat 590 K and
1 atrn.A plate-wireESPwillremove 99.94oof theparticulatemattercarriedby the
gases.Assumingno back corona,estimatethe total collection areabasedon the effectiye
' ::: 'r ' ' ': '
migrationvelocitiesof Table9.2.
Answer:2,430m2
)rs Problems

9.13b.Effect of surfur contentof the fuer on

ESp per{ormance
selzlerandwatson(JApcA,24:115, 1974)proposed
thefollowing empirical
.{ Deutsch-typeequationf9r the overallpenetrationof an ESp for
the controlof fly-ash
emissionsfrom a pulverizedcoal_firedfurnace:

ptr4= sapi- r.+ 0u

o.o+so(scel
{rlr) l-s-10'"1
where
t \gl \eal l

rc SCA = specific collection area.s/m

- gasvolumetric
Q flow rate, m3ls
1(W= power input to the dischargeelectrode,
kW
.l = sulfur content of the fuel, weight percent
AIl = ash content of the fuel, weight percent
(a) A precipitator with an overall efficiency
^ of 97.7vooperareson a coal with a sul-
fur-to-ash ratio of L.5 : r2.3. To help meet
so2 emissions,r*o*ar, a coal is substituted
with a sulfur-to-ashratio of 0.8 : l2.5.If all
the other operating variables are the same,
estimatethe new collection efficiency.

(b) what percentchangein coronapower

is necessary if, for thenew sulfur-to-ash
ratio,it is desiredto restorethecolrectionefficiency
to it, originutvalue?
Answer: 26.7Voincrease

9.14a.Coronapower for a fly-ash ESp

Problem9'13 givesselzlerandwatson'sempirical
equationfor the overallcollec-
tion efficiencyof an Esp.A precipitatortreats
r,322m3/sof flue gasto removefly_ash
particlesfrom a pulverizedcoal-fredpowerplant.
The designoverallremovalefficiency
is 99.5Vo. Thetotal collectionareais 7g,320m2;the coal
contains l.\Vo S andI2Vo ashr.
Estimatethecoronapowerfor theseconditions.

Answer:650kW
410 Chap.9 ElectrostaticPrecipitators Problen

9.15a.Effect of back coroRaon plate-wireESPperformance where

Considerthe conditionsdescribedin Example9.9.Assumingthat aplate-wirepre-
cipitator is usedwith the sameSCA, integratethe grade-efficiencyequationto calculate
theresultingoverallpenefration
(a) With no backcorona
(a) I
Answer: l.9Vo
flow qual
b) With severeback corona tion.
Answer:6.5Vo

o)l
9.1"6d.Multiple cyclone precleaner for,an ESP I : I :: velocity d
507o,esti
The gas flow rate through the ESP of Example 9.9 increasesto 30 m3/sbecauseof
reentraim
changesip the capacity of the boiler. Instead of increasing,thecollection area,the ESP is
fitted with a multiple cyclone precleaner with an overall removal efficiency of 80%. The
aerosolpopulation leaving the cyclones is log-normally distributed with MMD = 7.0 pm
and o, = 2.5. Estimate the overall efficiency of the combined systemat the new gas flow
9.19c'd.E
rate.
hy
Answer:99.6Vo existingc
the exisil
'
9.L7t. Effect of non-uniform gas velocity orr ESP performance andtheu
SprayDrl
Example 9.9 calculatesthe overall efficiency oi an gSp asr,iming u ooiform gas
velocity distribution. If the distribution is characterizedby a quality factor of 1.2, esti- -
mate the actual efficiency assumingno gas sneakageor reentrainmentlosses. Paramrfi
:i, : r Answer:99To
ffiE
Teryerfr
Plateacel
9.18b.Flow quality factor and statistical measures of velocity nonuniformity Nrdcrd
Pt#ili
McDonald and Dean (1982) presentedthe following empirical relationship based Aner4ed
on:a pilot plant study between Q,the normalized standarddeviation of the gas velocity InletrilEd
:'
distribution(of), an4the idealpenetrationpredicted:' :' MMD.;
oa
Inhbr
0 = 1 + 0.766(1 - rtid)o;'trc + 0.0755oylnPtid ffictfri
 Problems 411

where

(a) Estimate oJfor the flow distribution conditions of Example 9.10. Estimate the
flow quality factor predicted under those conditions by McDonald and Dean's correla-
tion.
Answer:Q = 1.224
(b) The ideal overall collection efficiency of an ESP is 99.97oassuminga uniform
velocity distribution. If the normalized standarddeviation of the actual distribution is
50Vo,estimatethe resulting overall efficiency assumingno lossesby gas sneakageor
reentrainment.
Answer:99Vo

9.19c'4.ESP performance in SOz dry scrubbing applications

Dry scrubbing processesoffer potential cost-effectiveretrofit FGD technologiesfor
existing coal-fired electric generatingstationsfaced with acid rain legislation. However,
the existing particulate control equipment must be capableof collecting both the flyash
and the unreactedinjected sorbent.The following data were obtained at the TVA 10 MW
Spray Dryer/ESP Pilot Plant (Durham et al. 1990):

Parameter BaselineConditions SprayDryer Conditions

Flow rate, m3/s 16.35 15.5
Temperature, K 430 336
Plate areaper section, m2 3r4.4 314.4
Number of sections 4 4
Plate spacing,mm 254 254
Average electric field, kV/m 300 3',70
Inlet size distribution
MMD, pm 8.0 10.0
08 2.5 2.3
Inlet dust concentration,g/m3 2.634 2r.r3r
Outlet dust concentration,g/m3 0.00453 0.0355

.qrtm'

t,diiili*li.:rq,;'
412 Ghap.9,ElectrostaticPrecipitators Proble

Gassneakage andtheflow qualityfactorwereestimatedas0.05,and 1.1respec-. : mm. The

tively for both setsof operatingconditions.The dielectricconstantof theparticlesfor cm. Frm
both conditionsis about8.0. mentabo
(a) Estimatethereentrainment ing:
fraction,RR,at baselineconditions.
(
Answer:
o'09 (
,. : , ,,, :r .. : ,
(b) EstimateRR at spraydryerconditions (
Answer:0:lS8 Ctn
. ' (c),Noticethat at spraydryerconditionsthe outletdustconcentration increases sig- merit As
nificantlyascomparedto thebaselineconditions.Designa,newESPthat,at spraydwer andmain
conditions,achievesthe outletconcentrationthatcharacterizedbaselineconditions. devioes c
Specifythenumberof sectionsandthecollectionareaper sectior,, , . ,
, . ,. . , , , .

9,20b,,Diqgnsionsofan ESP
:l
specifythe dimensionsof an ESPtorprocess 50,000m3/minof,a flue gaswith
9?:5%.o1e_ratt.narticulate
remorralefficiency.The effectivgmigrationvelocityof thepar-
ticlesis 0.1 m/s.Assumethat theplatespacingis 400 mm; the aspectratio is 1.0;andthe
gasvelocityis'not to exceed1.0m/s.Specifythetotal iollebtion,area,
the numberof
channels,the dimensionsof the EsR andthe coronapowerrequfiements.' , ,,'r' '
a ;,, i ,,, .: ..,
i. , . . t: ; .:,it.. .,.t
:

Answer:A, = 44'300mz

9.21". Total capital investmentforan ESP

' ' r Estimatrim fCt andthe annualcapitalrecoverycostfor the ESPof,P*obletri9.20,'
"
Assume thatthecostof auxiliaryequipmentis'$215,O00.
No specialsitepr.eparationor
buildingsareneeded.Assumea usefullife of 20 yr.qndno salvagevalue.Theminimum
attractiverateofreturn is I\Volyr.
l: ._ :: ,, ,

Aruswer:CRC= $2,586,400/yr

9.22d.Plate-wireESPfor a Portland:cement,kiln 1,...

In Problem8.L7, amultiplecyclonesystemwasdesignedasa precleanerfor the
conffol of particulateemissionsfrom a Portlandcemert kiln. For eachof the alternatives 1

considered in that problem,designa plate-wireESPsuchthatthecombinedsystemis i

t
t
aapable.ofsatisfyingtheparticulateNSPSfor the source.Thq,p-late.sp3cin€ mustbe 400
Problems 413

mm. The dielectricconstantof the particlesis 6.14;their resistivityis below 10ll ohm-
cm. From scalemodels,gassneakage is expectedto be about0.10;rappingreenfrain-
mentabout0.10;andthe flow quality factor l.2.For eachalternative,specifythe follow-
ing:
(a)Numberof sections
(b)Totalcollection
area
(c)TCI andTACof thecombined system
multiplecyclone-ESP

Choosebetweenthethreealternativessuggested basedon the EUAC;me.asure of

merit.Assumethatthe ESPhasa usefullife of 20 yr with no salvagevalue.Operating
andmaintenancelabor ratesare$15/hand $20lh,respectively.The dustcollectedin both
devicescanbe recycledto theprocessat no additionalcost.