United States EPA-600/9-85-Oil
Environment^ Protection
A9encv April 1985
>EPA Research and
Development
THIRD U.S./U.S.S.R.
SYMPOSIUM ON
PARTICULATE CONTROL
Prepared for
U. S. /U. S. S.R. Bilateral Agreement
Prepared by
Air and Energy Engineering Research
Laboratory
Research Triangle Park NC 27711
EPA LIBRARY SERVICES RTP NC
EPA-600/9-85-011
TECHNICAL DOCUMENT COLLECTION
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development. U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grquping, was consciously
planned to foster technology transfer and a maximum Interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the MISCELLANEOUS REPORTS series. This
series is reserved for reports whose content does not fit into one of the other specific
series. Conference proceedings, annual reports, and bibliographies are examples
of miscellaneous reports.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
EPA-600/9-85-011
April 1985
THIRD IT.,5./ILSJ3.R. SYMPOSIUM
ON PARTICIPATE CONTROL
EPA Project Officer: Jaroslaw Pekar
Air and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina 27711
Prepared for:
U. S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
ABSTRACT
The Third U.S./U.S.S.R. Symposium on Particulate Control was sponsored by the Particulate
Control Technology Project Group of the U.S./U.S.S.R. Stationary Source Air Pollution
Control Technology Working Group under the U.S./U.S.S.R. Environmental Agreement. Papers
were presented by Soviet specialists from scientific research and design institutes
and by representatives of the United States Environmental Protection Agency and
representatives of the private sector. The topics included: prediction of back
corona formation, fly ash. resistivity prediction, improved ESP mathematical model,
calculation of effects of back corona in wire-duct electrostatic precipitators, study of
chemical conditioning of flue gas before an ESP on a 500 MW power unit, sodium
conditioning test with EPA mobile ESP, fly ash dielectric properties and critical current
density, improved design for general industrial application of precipitating electrodes
for ESP's, performance analysis of a hot-side ESP, dynamics and strength of corona
electrodes in industrial ESP's, progress on electrostatic precipitators for use at high
temperatures and high pressures, cleaning of hot aspirated air from sinter machines and
clinker coolers, ESP model for the TI-59 calculator, problems of fly ash removal at
electric power stations, particle collection by granular bed filters and dry scrubbers,
filter materials for cleaning gases and possible areas of their application, high
temperature ceramic filters, methods and instruments for measuring size of finely
dispersed particles less than 0.3 micron for process gases, particle measurement in
the U.S.A., current trends in aerosol dispersion analysis, instruments for automatic
monitoring of particle-size distributions, methods and instruments for measuring the dew
point of process emissions, and design of an improved impactor.
ii
-------
Table of Contents
Page
Abstract ii
Foreword vi
Forcasting the Occurrence of Back Corona
in Electrostatic Precipitators, (USSR) 1
I.K. Reshidov
V.M. Tkachenko
A.Ya. Rachek
A.T. Lyapin
V.A. Rodionov
A Technique for Predicting Fly Ash Resistivity, (USA) 13
Roy E. Bickelhaupt
Leslie E. Sparks
An Advanced Mathematical Model for Electrostatic
Precipitator Calculating 23
(USSR)
I.V. Ermilov
T.I. Dmitrieva
Yu.M. Morozov
Calculations of Effects of Back Corona in
Wire-Duct Electrostatic Precipitators, (USA) 41
Phil A. Lawless
Leslie E. Sparks
Study of Chemical Conditioning of Stack Gases before
Electrostatic Precipitation at 500 MW Power Station (USSR) .... 69
I.A. Kizim
B.V. Zolotaryov
A.A. Troitsky
M.A. Golosov
Sodium Conditioning Test with EPA Mobile ESP, (USA) 80
Steven P. Schliesser
Improving Electrical Feeding System for
Electrostatic Precipitators, (USSR) 114
V.I. Sikorskiy
V.V. Kutlyashov
Fly Ash Dielectric Properties and Critical Current Density, (USA) . . . 125
John P. Gooch
Jack R. McDonald
Leslie E. Sparks
111
-------
Table of Contents (cont'd)
Page
Research, Development and Introduction of Advanced
Precipitating Electrode Designs of Industrial Purpose
Electrostatic Precipitators (USSR) ................ 140
A.I. Zavyalov
V.B. Mescheryakov
Performance of a Hot-Side Electrostatic
Precipitator, (USA) ....................... 145
G.H. Marchant
John P. Gooch
Leslie E. Sparks
Study of the Dynamics and Strength of Coronal
Electrodes in Electrostatic Precipitators, (USSR) ........ 155
E.N. Kurbatskiy
V.B. Meshcheryakov
E.N. Rudometov
Progress on Electrostatic Precipitators for
Use at High Temperature and High Pressure, (USA)
George Rinard
Donald Rugg
Robert Gyepes
James Armstrong
Dennis Drehmel
Purification of Hot Aspirated Air from Sintering
Machines and Clinker Coolers, (USSR) ............... 172
V.M. Tkachenko
V.A. Rodionov
A.D. Mai gin
A Model of Electrostatic Precipitation for
TI-59 Calculator, (USA) ..................... 178
Leslie E. Sparks
Urgent Problems of Ash Trapping at Thermal
Power Stations, (USSR) ...................... 186
L.I. Kropp
G.S. Chekanov
Particle Collection by Granular Bed Filters and
Dry Scrubbers, (USA) ....................... 19g
Dennis Drehmel
R. Parker
S. Calvert
-------
Table of Contents (cont'd)
Page
Filtering Material for Gas Cleaning and
Possible Areas for Its Use, (USSR) 212
I.K. Goryachev
Yu.V. Abrosimov
High Temperature Ceramic Filters, (USA) 218
Dennis Drehmel
R. Parker
S. Calvert
Methods and Equipment for Measuring of Finely
Dispersed Particle-Size Less Than 0.3 Micron 228
for Industrial Gases, (USSR)
N.G. Bulgakova
D.L. Zelikson
Particle Measurement in the USA, (USA) 239
W.B. Smith
K.M. Gushing
D.B. Harris
Present-Day Tendencies of Aerosol Disperson Analysis, (USSR) . . . 268
D.L. Zelikson
N.G. Bulgakova
Instruments for Automatically Monitoring Particle-
Size Distribution, (USA) 275
W.M. Farthing
W.B. Smith
W.B. Kuykendal
Methods and Devices for Measuring Industrial
Gases' Dew Point (USSR) 303
A.I. Manyurov
N.G. Bulgakova
E.I. Akopov
Design of an Improved Impactor, (USA)
W.B. Smith
K.M. Cushing
D.B. Harris
-------
FOREWORD
The primary objective of the Third U.S.-U.S.S.R. Symposium on Participate Control, held
September 10-12, 1979, in Suzdal, U.S.S.R., was to discuss work done by American and
Soviet Specialists on particulate control and measurement technology.
The two symposium co-chairmen were: (U.S. side) A. B. Craig, Deputy Director of
the Industrial Environmental Research Laboratory * U.S. Environmental Protection Agency,
North Carolina; and (Soviet side) I. K. Reshidov, Director of the Scientific Research
Institute of Industrial and Sanitary Gas Cleaning (NIIOGAZ), Moscow, U.S.S.R.
The following organizations participated:
U.S.A. - Industrial Environmental Research Laboratory,*U.S. Environmental
Protection Agency, Research Triangle Park, N.C. 27711
U.S.S.R. - Scientific Research Institute of Industrial and Sanitary Gas Cleaning
(NIIOGAZ), Moscow
- Semibratovo Branch of NIIOGAZ
- Zapozozhye Branch of NIIOGAZ
- Special Dasign Office for Gas Cleaning and Dust Eliminating Equipment
(SKBGPO), Semibratovo
- State Institute for Gas Cleaning Equipment Desing (GIPROGAZOOCHISTKA),
Moscow
- All-Union Heat Engineering Institute imeni Dzerzhinskiy (VTI), Moscow
- State Research and Design Institute for Environmental Protection in the
Building Materials Industry (NIPIOSTROM), Novorossiysk
- Moscow Energy Institute (MEI), Moscow
- Moscow Transport Engineering Institute (MIIT), Moscow
- All-Union Electrotechnical Institute imeni Krzhyzhanovskiy (VEI),
Moscow
(*) Now, Air and Energy Engineering Research Laboratory.
vi
-------
I.K. Reshidov, V.M. Tkachenko, A.Y. Rachek, A.G. Lyapin,
E.N. Andrusenko, G.I. Vyalkova, V.A. Rodionov
FORECASTING THE OCCURENCE OF BACK CORONA IN ELECTROSTATIC
PRECIPITATORS
For efficient removal of high-resistance dusts from gases in electro-
static precipitators, it is important to know under what gas-cleaning
parameters that back corona may arise in the equipment.
Predicting 'back corona by the magnitude of resistivity in the
dust layer is not always justifiable, since the resistivity magnitude
depends not only on properties of the dust and gas, but also on dust
layer forming methods and dust resistence measurements. Furthermore,
back corona is caused not only by dust resistivity, but also by cur-
rent density in the precipitator. Therefore, it is most advisable to
directly record back corona and establish the parameters under which it
may occur.
•
A method for detecting back corona and for determining the criti-
X*
cal parameters of its occurence has been developed by NIIOGAZ, This
method is based on various effects of the dust layer in the precipitating electrodes
upon ;the voltampere characteristics of the corona in relation to the
dust's resistance.
It is known that, in the case of non-high-resistance dust and absence
d>f back corona, the voltampere characteristics of dust-lac.en electrodes
have less slope than clean precipitating electrodes. This is caused by a
a drop in voltage in the dust layer and by blocking of coronal discharge
t.
by volumetric charging of dust.
In the case of high-resistance dust, various effects of the dust
layer in the precipitating electrodes on the corona voltampere charac-
teristics may occur. Figvl shows the voltampere coronal characteristics in a "conductor-
-------
plane" electrode system in the presence of a high-resistance
fly-ash layer from ekibastuz coal on the plane's surface as well as
during its absence. Discharge distances are.the same in all cases.and
comprised 125mm. For needle coronal electrodes (fig.la), the current
magnitude of the corona characteristic at the beginning section in the
presence of a layer is somewhat less than corresponding current values
in the absence of a dust layer. This is caused by blocking action of the
layer's potential, with further voltage increase, the slope
of the characteristic increases and intersects the characteristic in
clean electrodes. This is related to the manifestation of back corona. As
seen from fig.lb, £t is obvious that by using bar wire electrode, back
corona manifests itself practically simultaneously with the base corona,
despite the fact that under these circumstances the average value of the
specific current of the corona per unit of length is less than that for
a needle-type electrode. This occurence is caused by considerable change
in the bar wire corona current with respect to tijne and the maxural current density
value for the bar wire electrode is significantly higher, than for a
needle electrode (1).
The dependence of corona current differences for clean and dusty-laden
electrodes on voltage (fig.2) is plotted on the basis of the considered
voltampere characteristics. In the case of needle electrodes, it has variable
character both according to its magnitude, and to its sign. Corona
current difference for bar wire electrodes increases with the rise of
voltaoe. but has a negative sign. These represent typical_relationships
for high-resistance dust.
73ie correct approach for determining voltage in the dust layer on
jbrecioitatina electrodes befpre the onset of back corona according to
corona voltampere characteristics for clean and dust-laden electrodes (2)
has been established by probe measurement. It is shown that back
-------
corona formation corresponds to the maximum of the dependence of
Voltage difference, for dust-laden and clean electrodes on corona current.
This is confirmed by the occurence of glow on the precipitating electrode
and also because the probe measurement begins to trap opposite-sign
ions.
Consequently, the maximum of corona current difference dependence
for.clean and dust-laden electrodes on voltaqe corresponds to the start
of back corona occurence. It is obvious from fig.2 that the largest va-
lue of corona current difference for needle electrodes is reached at
the critical voltage ^critical= 30 kil°v°lts. At this voltage, corona
current in the discharge intervals with dust-laden electrodes (fig.la) is equal to
the critical value J 'tical~ ^06 milliamperes/m.
•
A knowledge of critical values for corona current density /jv-i+i-ai
in the precipitating electrode and voltage in a dust layer U
of determined thickness 0 allows the resistivity critical value of the
dust layer to be found:
Critical = Uc -critical
--.j^..' -,i • —«.
critical .Jcritical
Use of the stated method gives the opportunity to determine not
only critical electrical parameters for back corona occurence, but also
critical values such as, for example, temperature and gas humidity. In
order to do this, it is necessary to change the parameter of interest until that time
when the back corona occurs or disappears.in this manner it is possible to
ft*//A .-»••»•»*•
establish an area of implementation of back corona during the tsapping of
high-resistance dusts in the precipitator.
The developed method may be realized in principle in a multi-field
precipitator/ if the last field is going to have clean electrodes.
A special complex of equipment was created at NIIOGAZ, which allows
the determination of critical parameters of back corona occurence in labo-
ratory as well as in industrial conditions by means of direct dust and gar
3
-------
sampling from the gas-duct in a model precipitator. Electric heaters
for gas and a steam generator installed at the inlet of the model,
which has both clean and dust-laden electrodes make it possible to vary
the temperature and moisture content of the studied gas.
Aspirated dust, sampled from the hopper of a DVPNI-4x20 precipi-
tator in a center plant at a metalurgical works, was studied in the lab-
oratory uder .various air temperature with the aid of this equipment complex. Aspirated
air, entering the unit at a temperature of about 110°C from various sec-
tions of the sintering machine, is cleaned in this precipitator.
It has beeen established as a result of research that the proba-
bility of a back corona occurence during trapping of the indicated
dust will appear at air temperatures of more than 70°C, moisture con-
tent of 9-10g/m (of moist gas), and during an exceeding of average
current density of 1.3-2.5 milliamperes/m . At lower current densities,
the back corona does not occur during an increase of air temperature to
150'C.
At NIIOGAZ's Zaporozhe branch, another method has been developed
which also makes it possible to determine several quantitative characteristics of
back corona. The impulse characteristics of discharge from the surface
of thivfe±gh-resistance dust layer serve as the basis of this method.
A differential sensor, which gives basic corona signals and constant
interference source signals and separates only the back corona ioni-
zation process impulses, is proposed for the measurement of back corona
impulses (3,4). The initial back corona voltage can be determined bv
the manifestation of the first pulses while back corona intensity can
be determined bv the repetition freauencv of these pulses.
•£*•'•' t,
By placing the differential sensor together with the qesoRjLzinq
electrode directly in the gas-duct or in a separate chamber, it is possi-
ble to eredict the conditions of back corona occurence prior to con-
struction of industrial precipitators.
4
-------
A general view of the unit with coronizina intervals is shown
in fig.3. The unit's discharge interval consists of a coronizinq elec-
trode (1) and a back corona sensor (2). Part of the gas from the
duct (9-) enters the chamber (3) through the connecting (5). In
the corona discharge field, a dust layer is formed on the working sur-
face of the sensor (2). The amount of gas and its temperature is
controlled in the unit. According to sensor signals, the begin-
ning back corona occurence conditions are determined (current density,
averaged electrical field intensity) and its intensity.
A back corona measuring device was used fbr study of back corona occurence
conditions in a gas-dust stream from the casting-bed of a high-capacity
blast furnace in a metallurgical works, from electric-arc ovens for
synthetic corundum smelting in an abrasive mill and from electric-arc
ovens for silicon smelting in aluminum plants. In all cases, the gas-
dust stream passed through the coronizing interval. The test results have
shown that the sensor works satisfactorily.
The measurement results taken are introduced in fig.4-6. At a
humidity more than lOg/nm , the back corona is not present until breakdown
voltages in the gage on dust from blast-furnace cast-houses (fig. 4)
In this case, there are no back corona impulses. On high-resistance
synthetic corundum dust, the back corona occurs at a voltage of 9.3
kv in the coronizing electrode and with a 9.7 microampere current
through the discharge interval (fig.5). Under the increase of voltage
back corona impulse frequency grows and reaches 10kHz.
The measuring device allows monitoring of corona parameters durinq
introduction into the dust-gas stream of moisture,necessary for its
removal. The temperature of gases from from silicon smelting furnaces
decreases in relation to the amount of directlv introduced moisture,
and the intensity of back corona drops simultaneously. At 200°C, the
5
-------
fig.l. -Volt ampere -Character I^ta-ds of the corona for a "wire-plane"
electrode system:
1. during the presence of a dust layer on the plane surface.
2. for clean electrodes
a. neddle electrode
b. bar wire
-------
-002
fig.2. Relationship of corona current variability, for
clean and dust-laden electrodes:
1. needle electrode
2. bar wire.
-------
4
00
Fig.3. Bac* corona measurement device, 1. coronal electroae, 2. differential sensor,
3. wooing chamber, ,.5. „. duct outlet and inlet, 6. porthoie, 7,„.connecting pipe,
fronor.cess. 10.-collector, Hr ejector connecting pipe, 12,13.14. baffles, 15.thern,o-
meter connecting pipe, 16, porthole, ". quartz window.
-------
8 12 16 20 2/kw
fig.4. Corona voltampere characteristics in the
measurement device for aspirated air from a blast-furnace cast-house.
-------
8.0
2.0
/
0 16 ^3248
fig.5. The dependence of impulse repetition frequency of back corona on:
a. discharqe voltaae: b. discharge current.
Gas-dust flow feom electric-arc synthetic corundum smelting furnaces.
-------
kHz
1250
ODD
750
500
250
0
200 t'C
fig.6. Effect of temperature of a gas-dust stream
from silicon smelting furnaces on the frequency of
back corona impulse repetition.
11
-------
back corona repetition frequency comprised 1500kHz. At temperatures
less than 124°C, discharges from the high-resistance layer are not
observed (fig.6}.
Thus, the developed methods and equipment permitpredicting the
electrical and technological parameters under which back corona may
occur in the electrostatic precipitator.
.BIBLIOGRAPHY
1. Reshidov, I.K., "Back Corona Occurences in Electrostatic Precipitators",
collection of reports of an inter-regional seminar on gas-cleaning,
Yaroslavl, 1972.
2. Levitov, V.I., Reshidov, I.K., "On the.Use of an Automated Method of
Probe Characteristics in Several Electrostatic Gas-Cleaning Problems",
in the collection Strong Electrice' Fields in Technological Processes,
Moscow, "Energy", pp.145-158, 1969.
3. Rachek, A.Y.,"A Sensor for Detection of Back Corona in Electrostatic
Precipitators", in "Industrial and Sanitary Gas-Cleaning", no.4, p. 54,
1976.
4. Rachek, A.Y., Reshidov, I.K.,"Compensations of Current Shift
During Impulse Electrical Discharge" in "Electronic Materials Treatment",
no.2, p.36-39, 1977.
12
-------
A TECHNIQUE FOR PREDICTING FLY ASH RESISTIVITY
Roy E. Bickelhaupt
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35205
Leslie E. Sparks
Industrial Environmental Research Laboratory
Environmental Protection Agency
Research Triangle Park, N.C. 27711
(This document has been reviewed and approved for
publication by the U. S. EPA.)
13
-------
INTRODUCTION
A research objective that has been pursued for several years is to predict fly
ash resistivity as a function of temperature based'on coal ash chemical analysis and
a flue gas composition stoichiometrically calculated from the as-received, ultimate
coal analysis. This goal has been difficult to attain because one must be reasonably
certain that the laboratory resistivity data utilized in developing a predictive tech-
nique agrees with the resistivity values that the precipitator will experience. This
is a difficult problem that is magnified when the sulfur oxides are introduced into
the experimental environment.
EXPERIMENTAL SCOPE AND APPROACH
A large number of commercially produced fly ashes was chemically and physically
characterized for this work. ' By having a large data base, many ranks of coal were
represented, and one can assume that the effect of factors such as specific surface
and ash layer porosity were minimized.
Resistivity as a function of temperature for a given set of electrical and en-
vironmental conditions was determined using all the available fly ashes. From these
data, one can relate fly ash 'resistivity to fly ash chemical composition for one set
of experimental conditions and selected temperatures. From the original group, 16
ashes were selected to investigate the effect of the variation in environmental water
concentration and applied voltage gradient on resistivity. Six of these ashes were
further utilized in experiments to determine the effect of sulfur trioxide on resis-
tivity. From these experiments, expressions were developed for predicting resistivity
for a given temperature, water concentration, sulfur trioxide concentration and ap-
plied electrical stress.
EXPERIMENTAL PROCEDURES
The equipment and experimental technique have been previously illustrated and
discussed1 in detail and will be'only briefly mentioned here. When the effect of
sulfur trioxide was not under consideration, ASME, PTC-28 test cells2 were used in
14
-------
a stainless steel environmental chamber. Resistivity was determined under an applied
voltage gradient of 2 kV/cm from 460°C to 85°C in an environment of nitrogen, 5% oxy-
gen, 13% carbon dioxide, 500 ppm sulfur dioxide and 9% water by volume. In other
similar experiments the water concentrations were 5% and 14%. When the effect of
applied voltage gradient was of interest, the test temperature was maintained at
162°C.
A special test cell and environmental chamber were developed with which the ef-
fect of sulfur trioxide on resistivity was evaluated. The tests were conducted iso-
thermally at several temperatures using sulfur trioxide concentrations of 2, 5 and
10 ppm in an environment of air and water vapor. A radial-flow, concentric-electrode
test cell allowed the determination of resistivity for an ash layer one mm thick.
This thin layer of ash can be equilibrated with the environment containing sulfuric
>
acid vapor in a reasonable length of time.
EXPERIMENTAL RESULTS
Effect of Ash Composition
Figure 1 shows the relationship between the measured resistivity and the combined
atomic concentrations of lithium and sodium for 33 fly ashes. .These data were taken
from resistivity versus reciprocal absolute temperature plots for the individual ashes
f
at 1000/T(°K) = 2.4 (144°C, 291°F). Prevailing test conditions included the previously
described simulated flue gas containing 9% water, applied electrical stress of 2 kV/cm,
and no sulfur trioxide added.
The expressions defining the curve produced by linear regression analysis are
shown. One can either calculate the resistivity for the specific set of experimental
conditions prevailing using these equations or read the resistivity value from the
figure. A coefficient of correlation of -0.97 was determined. This coefficient de-
fines the degree of fit between the data.and the linear regression curve, and a value
of 1.00 would define perfect correlation between the two factors. Since it has been
15
-------
shown3 that lithium and sodium are principal charge carriers when experimental environ-
ments excluding sulfuric acid vapor are used, the relationship shown in Figure 1 was
expected.
Effect of Water Concentration
Figure 2 illustrates the effect of changes in water concentration for one of
the ashes evaluated. The water concentrations used were selected based on the data
sheets completed by the ash-supplying power stations which showed a range of 6 to
13 volume percent. These curves were generated using the previously defined base-
line test conditions with of course the exception of water concentration and are
similar to those found by other investigators." The attenuation of resistivity due
to increased water concentration is observable at about 230°C and becomes very signi-
ficant at the lower temperatures.
One way of displaying these data in a form suitable for use in the prediction
of resistivity is shown in Figure 3. In this interpretation, the resistivity data
of the previous figure have been plotted as a function of water concentration for
several isotherms. Expressions developed from data such as these can be used to
correct the resistivity value' predicted for a given set of baseline conditions to
a value for some other set of conditions. For example, the average slope of the
resistivity-water concentration curve at a temperature of 144°C was -0.085. This
is based on the data accumulated from the selected 16 ashes used to evaluate the
effect of water concentration. A simple algebraic expression can be used to convert
the resistivity value for 9% water shown in Figure 1 to the value for some other water
concentration.
log p = log p +
-------
logp : logarithm of resistivity for a specific lithium plus sodium concentration,
c, and a water concentration of 9 volume percent. Value obtained from
Figurte 1.
W : volume percent water concentration to which the resistivity is to be
corrected.
W : water concentration used in establishing Figure 1, nine volume percent.
c
S : A log p/A% HzO; -0.085 for 144°C and water concentrations between 5% and
w
15%.
Effect of Applied Electrical Stress
Increasing the electrical stress on an ash layer causes a decrease in the mea-
sured resistivity. A variation in the magnitude of the effect of applied voltage
gradient on resistivity has been observed during in situ testing as well as in labora-
tory tests by other investigators. 5 ASME PTC-28 suggests determining resistivity
just prior to dielectric breakdown. A research program involving dozens of ashes
and a multiplicity of test conditions cannot afford to do this. Therefore tests are
conducted on a few ashes to establish a relationship between resistivity and applied
voltage gradient. Data for other conditions and ashes are then calculated from this
relationship using an expression such as:
log p = log p + (E - E )S (2)
^* Mc,w,e * Kc,w e c e
logp : logarithm of resistivity for a specific lithium plus sodium concentra-
Cf w, e
tion, c, a water concentration W , and an applied voltage gradient E .
Yf 6
logp : previously defined.
CfW
E : applied voltage gradient to which log p is to be corrected.
6 G f W
E : applied voltage gradient used in establishing Figure 1, 2 kV/cm.
C
S : A log p/AE; -0.030 for 144°C and an applied voltage gradient range of
2 to 10 kV/cm.
17
-------
Effect of Sulfur Trioxide
The effect of sulfur trioxide on resistivity in an environment of water and air
was examined using a limited number of ashes and tests. The procedures and results
for individual tests are detailed in reference 1. Figure 4 shows the results for
six tests conducted on one ash to demonstrate the combined effect of sulfur traoxide
concentration and temperature on resistivity.
All data thus far obtained af\, 147°C using 2 kV/cm voltage gradient and a base-
line environment of air containing 9 volume percent water are shown in Figure 5.
Six ashes were used in conjunction with sulfur trioxide concentrations of nominally
2, 5 and 10 ppm. At this time, it is believed that the data base is too small to
make other than guarded statements regarding the quantitative effect of sulfur trioxide
on resistivity. It is obvious that the effect can be dramatic in that the presence
of 10 ppm of sulfuric acid can reduce the resistivity two or more orders of magnitude.
The initial evidence suggests that the presence of sulfuric acid in the environ-
ment provides an alternate conduction mechanism. Therefore, other than the fact that
specific ashes have different affinities for sulfuric acid vapor, there would seem
to be no relationship between the acid and the ash composition with respect to conduc-
tion. Presently it is suggested that the effect of sulfuric acid can be combined
with the other factors that influence resistivity by considering them as two conduc-
tion mechanisms and determining a resultant resistivity from the equation for parallel
resistances.
p x p
*s c,w,e
Pr = r^P (3)
Ks c,w,e
Pr: resultant resistivity combining the effects of composition, water concentra-
tion, applied voltage gradient, and sulfuric acid concentration.
Pg: resistivity resulting from the effect of environmental sulfuric acid con-
centration taken from Figure- 5.
18
-------
p : previously defined, equation (2).
c,w,e
ILLUSTRATION OF RESISTIVITY PREDICTION
The information required to predict resistivity in the described manner is the
as-received, ultimate coal analysis and the chemical composition of the coal ash.
A stoichiometric calculation of the combustion products is made using 30% excess air
to determine the concentration of sulfur trioxide and water. The sulfur trioxide
value is taken as 0.3% of the calculated sulfur dioxide value. The coal ash is pre-
pared by first igniting the coal at 750°C in air, passing the ash through a 100-mesh
screen, and then igniting the ash a second time at 1050°C ± 10°C in air for a period
of 16 hours. Good agreement in chemical analysis has been obtained between coal ashes
produced in this manner and their respective fly ashes. The usual chemical analysis
of the coal ash in weight percent expressed as oxides is performed. The analysis
is converted from weight percent as oxides to molecular percent as oxides. The atomic
percentage of the lithium and sodium is taken as 66.6% of the molecular percentage
of the oxides.
The sum of the atomic percentages of lithium and sodium is used to determine
the resistivity value, p , graphically from Figure 1 or by computation. Using the
concentration of water determined from the combustion products calculation and equa-
tion (1), the predicted resistivity in terms of ash composition and water concentra-
tion, p , is determined. For the 5 mm ash layer thickness used, it was found that
C t W
dielectric breakdown generally occurred at applied voltage gradients of 8 to 12 kV/cm.
Therefore, it was arbitrarily decided to use 10 kV/cm as the electrical stress at
which the resistivity is predicted. Using equation (2) and E = 10 kV/cm, the pre-
dicted resistivity is put in terms of ash composition, water concentration and di-
electric breakdown stress, p . This value then is the predicted resistivity ex-
c,w,e
elusive of the effect of sulfuric acid.
19
-------
Eastern ashes are generally defined as those containing low concentrations of
alkaline earth elements k
-------
1013
U
E
o
•*
CM
II
X
O
1012
10"
GO
UJ
ec
Q
ui
CC
UJ
5
1010
109
1 [
In v * • + b In x
y • e«-xl>
INTERCEPT- 1-25.435
SLOPE - b - -2.129
COEFFICIENT - R - 0.97
OF
CORRELATION
• EASTERN ASH
O WESTERN ASH
1000/T(°K> 2.8
°C 84
OF 183
2.4 2.0 1.6 1.2
144 227 352 560
291 440 666 1040
TEMPERATURE
Figure 2. Typical resistivity-temperature data
showing the influence of environmental
water concentration.
o
5
1012
IQll
w
w
UJ
cc
109
J I
0.1 1.0
ATOMIC PERCENT • LITHIUM + SODIUM
10.0
Figure 1.
Resistivity as a function of combined lithium and
sodium concentrations for a specific set of test
conditions.
0 5 10 15
v/o H2O
Figure 3. Resistivity as a function of environmental
water concentration for various test
temperatures.
-------
OAIR+ 9v/o WATER ...^vinc
D AIR + 9v/o WATER +*5 ppm SULFUR TRIOX DE
A AIR + 9v/o WATER +»9 ppm SULFUR TRIOXIDE
ID"
10"
9
o
Figure 4.
1000/T(°K) 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 1.4
°C 40 60 84 112 144 182 227 283 352 441
°F 103 141 183 233 291 359 441 541 666 826
TEMPERATURE
Typical resistivity versus reciprocal absolute temperature
data showing the effect of injected sulfur trioxide.
OPEN - EASTERN
CLOSED - WESTERN
10
PPM SULFUR TRIOXIDE
Figure 5. Resistivity as a function of environmental sulfur trioxide
concentration for six fly ashes.
22
-------
I.V. XRMILOV, T.I. DMITRIEVA, Y.M. MOROZOV
AN ADVANCED MATHEMATICAL MODEL FOR ELECTROSTATIC PRECIPITATOR
CALCULATING
At the present time, a mathematical model has been developed
n the USSR for calculating the efficiency of dry, plate-type electrostatic pre-
ipitators trapping non-high-resistance dusts under known initial data
.bout the electrical conditions, gas-dust stream parameters, and the
quipment's geometrical characteristics (1,2,3). The model was deve-
.oped on the basis of research on basic dust precipitation processes in
:he ESP: kinetic charge and particle drift velocity (4,5), distri-
sution of dust concentration profiles (6) , and dust entrainment during
electrode rapping (7). Latest achievements in the area of calculating
fields with a volume charge (8), were used in the model. The cor-
rectness of this usage is asserted by field intensity measurements
taken by the probe characteristic method (9). Calculated values for
efficiency satisfactorily agree with experimental figures taken in
industrial equipment (3).
However, calculating efficiencies according to this model is
is somewhat time-consuming and in large-scale computations, use of
a computer is required. This makes analysis of several factors
of dust-trapping and -selection of the optimal variant of the gas-
cleaning set-up difficult. In connection with this, it is necessary
to create a simplified, engineering method of calculation that
has acceptable accuracy and exactitude in its computations. Moreover,
in developement, computation of re-entrainment is needed, which was
considered in preceding works as being dependent only on rapping
conditions of the electrodes (1,2). Developement of methodoloqies for
dieting the level of working voltages and corona ignition voltages
in industrial precipitators is also necessary. This data is used in
23
-------
the model during calculation of field intensity in the unit and in-
•fluence the accuracy of determining particle drift speed.
Solving these problems permits creating a finished mathemati-
cal model for calculating precipitators, where only the eauipment's geome-
trical characteristics and the gas-dust stream's parameters are used
as starting data. Prognosing the gas-dust stream's parameters in re-
lation to production technology is a problem of more removed perspec-
tive.
Therefore, the problem of developement of of an improved mathe-
matecal model of calculating precipitators occurs, in which the problems
listed above are reflected. Such a model for calculating precipitators of
thermal electric power stations is introduced in this work. One may
hope that the approach used will turn out to be acceptable even for
calculating precipitators of other industries.
Tn no. 10 of the bibliography, a simplified methodology for calculating
precipitator efficiency has been introduced, that was developed with
use of the moment method.. However, calculations according to this methodology
do not always satisfactorily agree with experimental data. This is
connected with the fact that in the simplified Tnethodology, active factors
are considered less than they are in the methodology from no.l fiseebibl.).
It is therefore advisable to develop a. simplified calculation methodo-
logy on the basis of the methodology from no.l (see bibl.). For this
it is necessary to to analize the significance of numerous factors en-
tering into the calculation, in order to choose the main ones.
Calculation of the factors' significances were conducted to fit to
the two most widely used types of precipitators: 1) those with coronal
electrode'.frame design (UG) ; and 2) those with free-hanging coronal
electrodes (UGT). In UG-type units, the relative area of the unit's active
zone, SA (eoual to the ratio of coronal electrode height to
24
-------
the height of the precipitating electrodes) is 0.91 and in UGT-type units,
Table 1 introduces results of significance calculations considered in
the exact methodology of factors: non-uniformity of dust concentration profiles,
particle charges at the coronal electrodes, particle passage through "semi-active"
precipitator zones, non-linearity of resistance of the particle movement medium,
kinetics of charge and particle entrainment during electrode rapping. The
calculations were carried out to fit the electrostatic precipitators at the
Ladyzhin State Regional Electric Power Station and the Slantsevskit cement
plant.
Table 1
Factors
s:
y- 1
/- >L<
0 QT T 0?_T H£
non-unifprmity of concentration ¥•£ A*?t»»e*UD
i uuii + yj.aij..i.y uj. concentration T A T oc
profiles-,- I.O 1.25
"Particle1 charge at the 0,91
coronal electrodes I.O 1.6
Particle charge 0.91 0.995
kinetics I.O 0.87
Kon-linearity of resistance 0,91 0.98-0.99
.of the medium I.O 0.97
"Particle passage 0,91 0.2
""through "semi-active" zones T^Q 1,0
"Dust entrainment during 0.91 O.T-0.2
rapping ' 1,0 0,01-0.05
7'/l
The values, -— , introduced in table 1, represent the ratio of the
values for dust re-entrainment, calculated with and without taking into
consideration one of the factors. As seen from the table, the most
25
-------
significant factors are the particle passage through "semi-active" zones and
entrainment during electrode rapping. It is necessary to consider particle
charge at the coronal electrodes and concentration non-uniformity profiles
only for electrodes with S* i£1.0. This is connected with the fact that during
Sa V 0.91, dust entrainment through the precipitator is basically determined
by the "semi-active zone, in which the influence of charge at the coronal
electrodes and of the non-uniformity of profile concentration is relatively low.
The remaining factors may be disregarded during development of the simplified
method.
During these simplifications, exact formulas from 1 (in bibliography)
may be taken thus:
Liku.
where t\JJo ~ particle drift speed at the precipitating electrode
'* L
T. overall relative length of precipitator fields
He
<~ho - spacing between precipitating electrodes
U, - average gas velocity
coefficient of non-uniformity of the field's gas
velocities ( £ c = /V* K3
coefficient, considering field weakening in the
"semi-active" zone (7^*0,5)
26
-------
Drift velocity of particles of diameter d are calculated according to the
formula:
(2)
where - the dielectric constant
• field intensity at the preci pita tor electrode
- gas pressure in mm hG
U - dynamic viscosity of the gas
£c and % - Cunningham correction, computed according to the formulas
/ *• I .
*$5/5- 0,56 ^ Iff t ^5 V~p~ /' ^^7« "' C'>8 '
27
(4)
(SI
-------
where / - distance between coronal elements
U - precipitator's working voltage
Uo - corona ignition voltage
^9 - effective radius of coronal electrode
Calculation of the overall degree of gas-cleaning during the absence entrainment
of precipitated dust is done according to the expression;
(7)
For density of particle distribution by size j (d), the log-normal law
was taken:
•!- m
where o - average standard deviation
dgQ - average median diameter
Trapping efficiency with regard to entrainment during electrode rapping
is found according to the expression:
where K - generalized coefficient of re-entrainment.
28
-------
The expression (1) may take the form:
d. •£-
'so ( 12
The parameter^ is analagous according to its structure by the index
of the exponent to Deutsch's formula and it may be called the parameter of
precipitation of poli-dispersed dust.
The parameter ft considers the effect on dust precipitation of diffusion
particle charging and Cunningham's correction. This is equal to the ratio of
the components of drift speed, conditioned by these factors, to the basic
components of drift speed, caused by impact charging of particles when
d - d5Q.
In relation to the formulas (7-10) introduced above, the degree of
gas-cleaning depends only on three dimension!ess parameters -j3 tft and 0
Result analysis on a computer of the numerical calculation of the degree
of gas-cleaning according to formulas 7-10 has shown that the degree of gas-
cleaning may be approximated by the following formula:
a / I")
w'
29
-------
The value of the parameter A, dependent on 0 and /t , is determined from
the graphs in fig. 1 and 2. The relationship^0.42 in introduced in the same
graph for convenience in calculation.
In calculating the degree of gas cleaning, it is possible to use the
following approximation formulas for parameter A:
ft 4 -.S-±-2i2§j___-iii__ for - 0.91
(-9.23,4 )+0.r 0<£* 0.5
ft * 2*§_iJk96 __ - itn k_ for
-I) O.Wiep(-i?;s*.>,a.T~ 0^^0.05
£or
'
The generalized re-entrainment coefficient K is calculated according to
the formulas:
A = 1-0.275 K
-------
where
— .= nj
//* //
H c -V7- '- relative height of precipitating electrodes
no
« u
u. s 77*" - relative gas velocity
- relative dust mass on electrodes.
The basic parameters are equal to the following measures: Ho = 8 m.,
Uo = 1 m/sec., Mo - 1 kg/m.
The values of the coefficients KJ, K2, K3 are introduced in table 2.
Formulas 17 and 18 were obtained by means of generalization by method of
regression analysis of experimental data according to re-entrainment coefficients,
taken from industrial equipment in the USSR.
Table 2
HeJige. of parame-
ter measurement t
H t U. . /IT
0 - 0,1
0,1*0.5
0,5*1*0
1,0*1.5
1,5*2.0
V/ (//'*>
0 - 0.3
0,3-0.7
0,7-1.0
1,0-1,2
1,2-1.1
We**)
0 - 0.1
0,1-0,8
0,8-1.0
1,0-1.15
1.15-1,3
>IV ./
I - 0.0
0.8-0.1
0.1-0,2
0,2-0.1
0.1-0.01
Comparison of the calculations according to the simplified (11-13) and
exact methodologies has shown that the maximum error of calculations at Sa = 0.91
31
-------
does not exceed +_ by effective drift velocity and about 50% by re-entrainment.
At Sa = 1.0, the maximum errors may be higher and may measure in separate
instances +. 15% from the effective drift velocity.
Table 3 introduces comparison results of calculation and experimental
values for cleanly efficiency in industrial precipitators in the USSR and USA.
The calculation was done by exact and simplified methodologies. The experimental
values, necessary for calculation, and the cleaning efficiency of American
precipitators were taken from no. 12 (bibliography), as well as during joint
research at the Allen Steam Plant in the USA by a group of Soviet and American
specialists (bibl. no. 13).
Comparison of calculated values according to the exact and simplified methodo-
logies and experimental values for efficiency attests to their satisfactory agreement.
The average error in calculating entrainment according to the exact methodology
doe? not exceed 10% and according to the simplified methodology, no more than 20%
of the experimental particle entrainment in industrial precipitators.
It is worth mentioning that the agreement occurs both in experiments when
the electrode rappers are shut off (k=l), and during rapping. This attests to
the fact that, in the model, basic physical processes, occuring in industrial
precipitators, both during electrode rapping and without electrode rapping, are
apparently reflected correctly. The circumstance is especially reassuring that
the agreement between the calculated and experimental values occurs in industrial
precipitators in the USA, which are essentially different from Soviet precipitators
according to their design (Sa=l) and data according to which the advanced model
was not used during the development.
Experimental data were generalized by the regression analysis method, in
order to predict the working voltages and coronal ignition voltages of industrial
electrostatic precipitators where these voltages are necessary for calculating
32
-------
drift velocity. These data were obtained for 10 industrial electrostatic
precipitators of thermal electric power stations using the principal grades
of coal utilized in the USSR.
A type of the regression equation was selected on the basis of the
experimental relationship obtained during research on laboratory and pilot
plant electrostatic precipitators. This is relative in part to the dependence
of the working voltage on the distance between electrodes and on dust
resistivity, etc. Six influencing factors were originally considered in the
regression equation. However, after testing factor significance, it turned out
to be possible to ignore the influence of incomplete combustion, sulfur content,
and also the relative gas density in the equation determining working voltage,
whose final form is:
2/l/(
where: F - area of the precipitating field of the precipitator, fed by one unit,
m2, F«f= 1000m2,
8 - gas moisture content, %, B( - 6%,
P - dust resistivity, Ohmmeters,Dtf = 5-10 Ohmmeters,
h0 - distance between coronal and precipitating electrodes, m;
hS = 0.1375 m.
These same computations were carried out for determining the voltage of corona
discharge ignition. However, in this case, the relative gas density turned
out to be a significant factor. The resulting equation for calculating
ignition voltage has the following form:
33
-------
Where: o - relative gas density,
-------
BIBLIOGRAPHY
1. Ermilov, I.V., "Calculating the Degree of Gas Cleaning in Plate-type
Electrostatic Precipitators", in "Industrial and Sanitary Gas Cleaning",
No. 1, 1977.
2. Ermilov, I.V., "Calculating the Efficiency of Thermal Power Station
Electrostatic Precipitators", in "Reports - Symposium on Controlling Particle
Emissions Into the Atmosphere and Other Related Problems, 26-28 May, 1976,
Kishinev".
3. Ermilov, I.V., Dmitrieva, T.I., "Mathematical Modelling of Electrostatic
Precipitators and Its Use For Industrial Units in the USSR", in "Reports of the
Second Soviet-American Symposium on Technology of Solid Particle Removal, 26-29
September, 1977", vol. 1, Environmental Protection Agency, Research Triangle
Par1, USA.
4. Ermilov, I.V., "Calculating the Degree of Gas Cleaning in Electrostatic
Precipitators", in "Electricity", no. 3, 1976.
5. Ermilov, I.V., "Experimental Research in Microparticle Charge In the
Corona Discharge Field", in "Electricity", no. 2, 1974.
6. Ermilov, I.V., "The Distribution of Dust Concentration In the Precipitator
Corona Discharge Field", in "Electricity", no. 7, 1974.
7. Sanaev, Y.I., Ermilov, I.V., "Experimental Research of the Processes of
Dust Entrainment From the Precipitating Electrodes in Electrostatic Precipitators"
in "Industrial and Sanitary Gas-Cleaning", no. 1, 1976.
8. Vasyaev, V.I., Vereshchagin, I.P., "Toward Calculating Characteristics
of Unipolar Corona Discharge In An Electrode System, 'Wire Rows Between Plates'",
in Electricity", no. 5, 1972.
35
-------
9. Vereshchagin, I.P. et al, Bases of Dispersion-Systems Electrical and
Gas Dynamics, Moscow, Energy, 1974.
10. Mirzabekyan, G.Z., Udalova, V.I., "Calculating Precipitator Work Efficiency
During the Dispersion Phase's High Concentrations", in "Electricity", 1976, no. 6.
11. Idelchik, I.E., The Aerodynamics of Industrial Equipment, Moscow, Energy,
1974.
12. Particulate Collection Efficiency: Measurements on Three Electrostatic
Precipitators. Environmental Protection Technology Series, EPA-600/2-75-056,
October 1975.
13. MacDonald, J.R., Mathematical Modelling of Electrostatic Precipitators
in the USA, Reports From the Second Soviet-American Symposium on the Technology
of Solid Particle Removal, 26-29 September, 1977, vol. 1, Environmental Protection
Agency, Research Triangle Park, USA.
36
-------
60
fig.l. Relationship of the coefficient A to b when Sa = 1,
37
-------
10
o 20 ^o eo jo joo
*
fig.2. The relationship of the coefficient A to b when Sa = 0.91,
= l-exp(-KA c°-
38
-------
Table 3
<£>
NO.
1
2.
3.
4.
5.
6.
7.
8.
9.
10.
II.
pltator
«*.«!»
uu-4-S) '
m-S-2-55
t>|*Tlm-ntjl
ua-f-*j
PCD-J-50
UCZH-265
Expcrlnu-ntal
P"
35.5
42.4
3«%2
40.0
52.0
45.8
36.3
3Z.9
32.5
33.4
29.6
I
19.2
21.7
20.0
22.0
28.6
22.4
«
^3.0
12.0
12.0
16.0
i
12
10
6.6
5.0
10,0
7.56
16.0
10.0
10.97
12.2
10.97
!
0.1375
0,1375
0.1625
0.1325
0.1375
0.1625
8.D75
0.1375
O.II43
O.II43
O.II43
1 4- i s: !
J 1 «J4 !
! i '
0.18 0.915
0.18 0.89
0.18 0.86
0.18 0.9
0.18 0.89
0.19 0.88
0.18 0.915
0,18 0.915
0.175 1.0
0.175 1.0
0.229 1.0
k
i.i
i.*
1.55
1,05
1.4
I.*
I.I
I.I
1.15
1,15
1.15
j^«<|,
I.M5 386.5
Z.I 415
l.«f 422
1.5 403
l.fll 418
I. Ill, 427
1.3* 400
1.11 393.5
|.« 427
l.tt 575
l.ifc "602 ":
! wa* Masurad In laboratory conditions
at the ga«'a working tcraperatura.
3) Data from noa. 9-11 waa taken from procipltatori at
thb Gorga», San Juan, and Allen Thermal Electric Statl6na in th«
USA.
-------
Table 4
Typo of preclpl-
tator
ECtrJ-177_
Fxporirental
UG2-4-SJ
UGZ=4-53
PCOS-3-50
Fxpcrlrental
Experimental
_p. PCDS-4-42
0
DGPS-2-42
PCOS-3-SO -
Experimental
f.-f
3046
20
1123
Tnr>
1255
1247
1247
934
1380
1255
20
i S
i
0.743
0.724
0,701
D.693
0,842
0,706
0,697
0.685
0.622
0.686
0.693
j-4; \
j
-------
CALCULATIONS OF EFFECTS. OF BACK CORONA
IN WIRE-DUCT ELECTROSTATIC PRECIPITATORS
by
Phil A. Lawless
Research Triangle Institute
Research Triangle Park, N.C. 27709
and
Leslie E. Sparks
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
(This document has been reviewed and approved for
publication by the U. S. EPA.)
41
-------
ABSTRACT
A model is used to calculate the effects of back corona on the
voltage-current characteristic, current density distribution at the
collector plate, and particle collection for five simple mechanisms
describing the generation of back corona. Each mechanism is capable
of simulating the V-I curves characteristic of precipitators operating
in back corona, but one in particular is identified as a potential
back corona mechanism on the basis of generating hysteresis in the
V-I curve, in accord with limited experimental evidence.
42
-------
ACKNOWLEDGMENT
This work was supported by the Industrial Environmental Research
Laboratory (Research Triangle Park, N.C.), Environmental Protection
Agency under Grant No. R80-58-9701. 0. H. Turner is grant adminis-
trator.
43
-------
CALCULATIONS OF EFFECTS OF BACK CORONA
IN WIRE-DUCT ELECTROSTATIC PRECIPITATORS
Back corona in an electrostatic precipitator is a condition in which
electrical breakdown in a high resistivity dust layer creates positive
ions which enter the normally negative ion cloud in the duct of the pre-
cipitator. 1>2 The formation of the positive ion current leads to a
characteristic increase in the slope of the voltage-current characteris-
tic. ' This increase is due to the increased numbers of positive and
negative ions and can be so large as to give near-vertical or even
/
negative slopes to the V-I curves. This voltage-current characteristic
is the most common indication of the presence of back corona because
direct detection of free ions ;0f both polarities is very difficult. In
this paper, we relate mechanisms for back corona formation to the cal-
culated V-I curves produced and examine the implications for particle
charging and collection.
The approach used in the calculation is a numerical solution of
Maxwell's equations for the electric field. The overall approach is
similar to that of McDonald, et al.. and a detailed description is to
be published.
The equations describing the distribution of electric fields and
flow of current are Maxwell's equations and the current continuity equa-
tion. In this formalism, it is more convenient to work with flux and
line integrals of the electric field and the current density. These
are, for the static case:
44
-------
/E • ds = /-£- dv (1)
V °
E • dA = 0 (2)
L
y j • ds = 0
(3)
In these equations, s is a surface or flux integral evaluated over
the surface of a grid/cell, which encloses the volume V. The line
integral is the term that couples adjacent grid cells and leads to a
solution for the entire region.
The relationship between the current density and electric field is
given by
j = pbt , (4)
where p is the ion density and b is the ion mobility. For the present,
we assume that positive and negative ions have the same mobility and that
any particulate space charge has a zero mobility. It is further assumed
that ion recombination or neutralization effects are negligible, so that
separate current continuity equations can be written for both types of
i ons.
The equations are solved by an iteration procedure involving an
underrelaxation of error terms. The starting values for the electric
field are derived from Peek's semi-empirical relation for the electric
45
-------
field at the surface of the corona wire at the onset of corona, and the
Q
electrostatic potential distribution of Cooperman for the wire-duct
geometry. The potential expression is differentiated either numerically
or analytically to give initial values to all the components of E and E
A y
as a function of position. For a given negative ion current at the plate,
an appropriate space charge density is calculated at the corona wire and
propagated to the plate. If the local current density at the plate
exceeds a threshold value, a certain density of positive ions is liberated
and propagated back to the corona wire. Equation 1 is then solved to
obtain new values for E ; Equation 2 is used to calculate an error term
J\
for E ; and a portion^of the error is then used to obtain a new value for
E . When the error terms for all E are adequately reduced, the equations
j y
have been solved, and the corona wire voltage can be obtained by integrat-
ing the fields from the plate to the wire.
The electric field model predicts standard clean-plate V-I curves
quite well and gives excellent agreement with experimental values of the
space potentials measured by Penney and Matick.9"
-------
BACK CORONA MECHANISMS
Although recent experimental work has begun to delineate the nature
of back corona mechanisms, ' revealing streamer and glow modes of
propagation, the mechanisms of generating positive ions in the particu-
late layer are still uncertain. Electrical breakdown of the particles
1 2
or the interstitial gases in the particle layer are postulated, ' but
detailed descriptions are lacking. The approach taken in the present
work is to assume several simple model mechanisms for the generation of
positive ions and to explore the consequences of each. All of these
mechanisms are described in terms of a critical current density, since
experimental evidence indicates the sudden onset of back corona as the
i
current density is increased. All the proposed mechanisms are assumed
to be local, with back corona onset in a given location having no direct
effect on adjacent locations. Finally, all the mechanisms assume a con-
tinuous generation of ions as opposed to the pulsing effects observed
experimentally. The five mechanisms chosen are described below.
I. Positive ions are generated in proportion to the amount
that the total current density exceeds the critical
current density. This can be written:
'e(JY-Jr)/Exb, j >j
x c x x c (5)
0 , j jx JG >
where e is an adjustable dimensionless parameter and jc
is the critical current density. Since the generation of
positive ions occurs in the dust layer near the plate,
only the x-component of field is significant. For values
of 6 near 1, this mechanism tends to keep the electric
field at the layer constant by reducing the net space
charge as the current density increases.
-------
II. Positive ions are generated in proportion to the critical
current density once the total current density exceeds
the critical level:
(YJC/
>+ = lo
jx
-------
V. The positive ions are generated in proportion to the nega-
tive ion current density squared:
(9)
where j_ is the x-component of the negative ion current
density~and < is an adjustable dimensionless parameter.
This mechanism describes a situation in which the positive
ion current does not directly contribute to the heating
effect, as would be the case if generation occurred at the
surface of the layer.
In all the mechanisms except II, the adjustable parameter affects
the ratio of the positive ion current to the negative ion current, and,
/
in general, values can be found for which the ratio is nearly 1 over an
extended range of currents. Mechanism II is different in that the
adjustable parameter affects only the size of a step increase in posi-
tive ion density, and the positive and negative current densities will
be equal at a single current only.
The effects of the resistivity of the dust layer are not considered
here. The dust is assumed to have zero thickness, and effects of posi-
tive ion generation on the dust surface or within the dust layer are not
considered.
EFFECTS OF BIPOLAR CHARGING
Where ions of opposite polarity impinge upon a particle of radius a,
a charging equation of the following form results:
|| = 3rt2[j.(l-q/qs)2 - j+(l+q/qs)2] , (10)
49
-------
where the sign convention chosen assumes that qg is negative and is
given by
qs = 12Trae0E , (11)
where E is the electric field magnitude far from the disturbing effects
of the particle. The charge q stops changing at an equilibrium value
given by
1/2
(12)
The factor multiplying q we call Q for convenience and is less
than 1 (even negative) for all non-zero values of j+. Recalling that
qs does not depend on current density, the factor Q represents the
degradation in particle charge that can be expected under bipolar charg-
2
ing conditions, and the quantity E Q is proportional to the particle
drift velocity, under bipolar charging conditions.
After choosing a fairly standard precipitator configuration and
operating conditions (wire radius = 1.34 mm, wire-wire spacing =
114.3 mm, plate-plate spacing = 228.6 mm, T = 615 K), each back corona
mechanism was used to generate V-I curves. The critical current density
-5 2
was set at 4 x 10 A/m , and the particulate space charge, q, was
-5 3
chosen to be either zero or -1 x 10 C/m .
The average current density was initially set to zero and then
incremented upward in steps to a maximum value, calculating the corona
wire voltage at each step. The electrical conditions used as the
50
-------
starting values for a given step were the final solution values of the
proceeding step, to mimic a real measurement of a voltage-current curve.
After the maximum was reached, the current density was decremented to
observe hysteresis effects, but many mechanisms showed no hysteresis
effects at all. Some calculations could not be carried to completion
because of numeric instabilities which were encountered, generally in
cases with extreme back corona. Particular attention was paid to
achieving a near-vertical V-I characteristic and to checking for hystere-
sis effects, since these are observed characteristics in back corona
situations.
Figures 1 through 5 present the voltage-current density curves
calculated for the five mechanisms. In most instances, near-vertical
curves could be obtained, and in some cases hysteresis was observed.
Note that hysteresis was expected in mechanisms II and III (and was
calculated), but it also appeared in mechanism IV in an unusual fashion.
The particulate space charge tends to suppress the spread of back
corona laterally between the wires. In Figures 2 and 3, the steps in
the curve are due to the cell structure of the calculation, with each
step corresponding to the formation of positive ions in one or two cells
as the local current density exceeds the threshold. With q = 0, the
transitions occur in all the cells over a narrow range of total current
densities.
CONSEQUENCES OF BACK CORONA
1 p
In the Deutsch-Andersen theory of precipitator collection, the
drift velocity of the charged particles appears as a multiplier in the
argument of an exponential term in an expression for the collection
51
-------
efficiency and is, therefore, especially significant. Recalling that the
2
drift velocity is roughly proportional to E Q, we have plotted this quan-
tity for the five mechanisms in Figures 6 through 10. The terms were
evaluated at the place, because collection occurs at the plate, and a
simple linear average over all the cells was used.
The figures illustrate that the presence of back corona has a strong
effect by whatever mechanism it occurs and whether or not a particulate
space charge is present. Two mechanisms are worthy of note: in mecha-
2
nism II on the descending legs of the hysteresis curves, the E Q product
average actually becomes negative; and in mechanism III, for y = 0.5,
2
the E Q product approaches zero rapidly as the current density increases.
i
Both of these conditions are disastrous for particle collection. The
other mechanisms, particularly in the presence of space charge, exhibit
2
local E Q products which are zero or negative but for which the overall
average is positive.
In terms of operating strategy in a precipitator suffering back
2
corona, the E Q product is generally a maximum at current densities which
avoid initiating back corona, i.e., low current densities.
2
The E Q curves also point out the enhanced drift velocity in the
presence of a particulate space charge. Even though back corona degrades
2
the E Q produi
higher level.
2
the E Q product, adding space charge can boost it to a significantly
COMPARISON WITH EXPERIMENT
An actual V-I curve (Figure 11) of a precipitator in back corona
was obtained on an X-Y recorder, measuring the corona wire voltage and
collector plate current as the voltage was slowly increased to a
52
-------
maximum and then decreased to zero. The participate space charge was
zero (no participate feed).
The ascending leg of the curve exhibits both increases and decreases
in the voltage as the current rises. More importantly, the X-Y recorder
reveals small variations in voltages that would likely be missed in a
point-to-point type of measurement. The descending leg of the curve is
more regular but still has the small step-like character that the ascend-
ing leg has.
Of the five mechanisms, only mechanism II has the potential of such
a pronounced hysteresis, with step-like increases and decreases in voltage.
Mechanism II is incomplete, however, in that it postulates only a single
/
threshold. In order to produce a characteristic similar to Figure 11,
multiple closely spaced thresholds would be required.
A composite voltage-current characteristic consisting of the sum of
five voltage-current curves from Figure 4, which were offset from each
other by 100 volts, is shown in Figure 12. These offset voltages corre-
spond to diameter variations of less than one percent, which are within
2
the realm of feasibility for dirty corona wires. The E Q curves for the
composite mechanism would be similar to those of Figure 7, with the more
rapid variations smoothed out by an averaging process.
Other mechanisms can produce V-I characteristics similar to the
ascending leg of Figure 11 with proper choice of the threshold current
density and the adjustable parameters, particularly mechanism IV. The
difficulty is in giving physical meaning to the parameters, which in
these calculations are arbitrary. In mechanism II, the adjustable
parameter controls only the size of the increase in current and can
potentially be related to the properties of the dust layer. It is the
53
-------
superposition of many small steps, near the same threshold, which can
produce the steep increase in current; and it is the sum of the increases
which leads to the hysteresis of the descending leg.
CONCLUSIONS
In spite of the simplicity of the proposed back corona mechanisms,
a reasonable approximation to the V-I characteristic of a precipitator
operating in back corona has been obtained. Although refinements in
experiment will lead to more complex models, the general effect of back
corona on degrading collection efficiency in these mechanisms suggests
that operation at reduced current densities will be beneficial, although
i
not extremely so, and that an enhanced particulate space charge in the
precipitator would be advantageous. Operation of a precipitator on the
descending leg of a voltage hysteresis curve should be strictly avoided.
The electric field model has proven to be a versatile calculational tool
for diagnosing the relatively inaccessible conditions in the electro-
static precipitator and will continue to be of value as the phenomenon
of back carona is investigated further.
54
-------
REFERENCES
1. S. Oglesby, Jr., and 6. B. Nichols. Electrostatic Precipitation
(Marcel Dekker, New York, 1978).
2. H. J. White. Industrial Electrostatic Precipitation (Addison-Wesley,
Reading, Mass., 1963).
3. H. W. Spencer, III. "Electrostatic Precipitators: Relationship
between Resistivity, Particle Size, and Sparkover," EPA-600/2-76-
144 (1976), U.S. Environmental Protection Agency.
4. H. J. White. J. Air Poll. Cont. Ass. 24, 314 (1974).
5. J. R. McDonald, W. B. Smith, H. W. Spencer, III, and L. E. Sparks,
J. Appl. Phys. 48, 2231 (1977).
6. P. A. Lawless and L. E. Sparks, to be published.
7. F. W. Peek, Jr. Dielectric Phenomena in High Voltage Engineering
(McGraw-Hill, New York, 1929)
8. P. Copperman. Unpublished. See Reference 5.
9. G. W. Penney and R. E. Matick. Trans. Am. Inst. Electr. Engr. 1 79,
91 (1960).
10. S. Masuda. Inst. Phys. Conf.' Ser. No. 27, Ch. 3 (1975).
11. S. Masuda and A. Miguno. Proc. 4th Int. Clean Air Conference,
Paper No. V-47 (Tokyo, May 1977).
55
-------
FIGURE CAPTIONS
1. Voltage-current density curves for Mechanism I.
2. Voltage-current density curves for Mechanism II. Hysteresis is
indicated by the dashed curves.
3. Voltage-current density curves for Mechanism III. Hysteresis is
indicated by the dashed curves.
4. Voltage-current density curves for Mechanism IV. Hysteresis is
indicated by the dashed curves.
5. Voltage-current desntiy curves for Mechanism V.
2
6. E Q as a function of current density for Mechanism I (6=0.5).
-—(B-0).
2
7- E Q as a function of current density for Mechanism II (y=1.0).
(Y=0); •-•-•(turrent decreasing).
2
8. E Q as a function of current density for Mechanism III (6=0.5, I;
6=0.4, •; 6=0.5, A). (6=0); -(current decreasing).
2
9. E Q as a function of current density for Mechanism IV (£=0.020).
(e=0); •-•-•(current decreasing).
2
10. E Q as a function of current density for Mechanism V (<=0.10).
—-U-0).
11. Voltage-current curve of a precipitator section in back corona.
12. Voltage-current density curves with hysteresis for a composite
Mechanism II.
56
-------
Mechanism I
6
*o
v.
<
"3 3
i5 c
^ 0)
Q
5 2
a
U
q«0 C/ms
0.5
q--lO"'C/m*
0.6
0.5
0.4
16
18
20
22
Voltage tKV)
Figure 1
24
26
-------
q «0 C/m3
I I
Mechanism H
q»-Kf* C/m3
o
x
S 3
00
(A
e
0)
3
O
y =0,0
16
18
20
22
24
26
Voltage (kV)
Figure 2
-------
q « 0 C/ms
I I T
Mechanism HI
q»-lo"5 C/m*
o
x
•S 3
Ui C
vo 4>
O
•a
O
0.4
0.5
0.3
16
18
20
22
24
Voltage (kV)
Figure 3
26
-------
g
X
«« .,
e 4
• 3
«>
O
s
Mechanism IZ
q *0 C/m3
0.023
q.-IO"5 C/m3
.0.023
0.020
0.018
16
18
20
22
Voltage (kV)
Figure 4
24
26
-------
Mechanism 3t
6
O
x
(0
c
Q>
o
q « 0 C/ms
.0.10
0.05
q--IO"5 C/m3
,0.10
0.05
I I
16
18
20
22
Voltage (kV)
Figure 5
24
26
-------
Mechanism I
1000
I 2 3
Current Density (A/m*) jc I04
Figure 6
62
-------
Mechanism E
1000
1234
Current Density (A/m2) x !04
Figure 7
63
-------
Mechanism HI
1000
Current Density (A/ma)x!04
Figure 8
64
-------
Mechanism JZ
1000
100
•>
a:
|0
Current Density (A/m*)xl04
Figure 9
65
-------
Mechanism 3C
1000
100
o
a:
10
Current Density (A/m*) xlO4
Figure 10
66
-------
6.O
5.0
4.0
£. 3.0
2.0
1.0
0.0
10
20 30
Voltage (kV)
40
Figure 11
67
-------
o
X
6 4
oo
to
c
0)
Q
flJ O
t_ c
O
Mechanism 31 Composite
q « 0 C/ms
-KfsC/ml
/I
I
/
I
16
18
20
24
26
Voltage (kV)
Figure 12
-------
I.A. Kizim, B.V. Zolotarev, A.A. Troitskii, M.A. Golosov
REPORT: STUDY OP CHEMICAL CONDITIONING OF STACK GASES
BEFORE ELECTROSTATIC PRECIPITATION AT A 500-MEGAWATT POWER GENERATOR
In accordance with basic directions for developement drawn UD
for the national economy of the USSR for 1976-1980 and long term
prognoses for the coming decade, the nation's energy output will be
increased chiefly through -construction of large-scale thermal elec-
tric power stations and and central heating and power plants working
mainly on low-grade hard fuel, primarily low-sulfur (SP<1%) and low
moisture coals from the Ekibastuz and Kuznets basins.
Combustion products of these coals includes many aluminum sili-
cates (up to 93-94%) while the share of alkaline metals (sodium, li-
thium, potassium) is insignificantly small. Therefore with exhaust
gas temperatures of 130-150°C, this ash has a high resistivity, com-
prising 10^^-10^-2 ohm-meters.
As practice shows, deep cleaning of combustion product fly-ash
of given coals in electrostatic precipitators reaching 99-99.5% is
difficult, in as much as the high resistivity in these units causes
intense back corona, sharply lowering ash-trapping efficiency.
Efficient ways for solving the problems of trapping high-resis-
tivity fly-ash are:
- precipitator installation in a gas temperature zone of 350-
400CC, due to which the volume resistivity of the ash deter-
mining its resistivity sharply decreases;
- lowering exhaust gas temperature according to comparison
with projected values when the ash resistivity is determined
by increased surface conductiveness;
- addition to the stack aases of various chemical conditioning
' 69
-------
substances and water steam, providing the required value
of surface conductivity for the ash layer.
Each of the listed methods has its own advantages anc insuffi-
ciencies. For deciding the advisability of using one of these me-
thods, it is necessary to conduct investigation into working condi-
tions at full-scale electrostatic precipitators.
Chemical conditioning is one of the promising directions for
increasing the degree of gas-cleaning during trapping of high-
resistance fly-ash.
Conditioning gas by humidifying is usable in those situations
where the exhaust gases have a high temperature and lowered moisture
content. Under these conditions, it has been established that the introduction of
moisture has a favorable effect on increasing electrical strength and lo-
wering gas viscosity. Water adsorption on ash particle surfaces
grows proportionately to the content of moisture in the gases and
inversely proportional to temperature.
With humidifying stack gases, besides lowering the ash resi-
stance , gas cooling occurs, which in its turn, lowers the resisti-
vity. However, introducing humidity conditioning to thermal power
stations gives rise to great difficulties in as much as for these
purposes, considerable input of water and steam is required. Other
than this, in power station conditions, it is impossible during
sufficiently low exhaust gas temperatures to produce full water
evaporation because it leads to condensation in gas conduits, pre-
cipitators and exhaust fans and their surfaces become sticky, which
guarantees corrosion.
803 and ammonia are other conditioning reaCcnT«f for stack gas
conditioning.
70
-------
The relationship of the degree of gas cleaning to sulfur content
is known. If follows from this relationship that, upon lowering sul-
fur content, efficiency of ash trapping is sharply lowered. Therefore
during lower-sulfur coal fly-ash trapping, SCU is added to stack
gases to compensate for the lack of it and to raise ash-trapping
efficiency. Sulfur trioxide, reacting with the ash, forms sulfates
capable of raising surface conductivity of the ash particles.
Introducing ammonia to the stack gases gives a positive effect
with low sulfur content, just as with high sulfur content. At the
same time, ammonia may be introduced to the gas tract both after
the air pre-heater and before it.
Sulfur dioxide is three times more expensive than ammonia under
the same quantities of reagent and approximately the same efficiency.
Therefore, ammonia was used in ekibastuz coal fly-ash trapping in
industrial experimenting with chemical conditioning at the Troitskaya
State Regional Power Station.
Industrial tests were conducted in 1977 on the 500 mW capacity
power unit no. 8 with a flow-through boiler with a steam-production of
1650 tons/hr.
Ekibastuz coal, type SS, was the fuel used, and has the following
characteristics.
p
ash content - A = 36.5S
moisture - WP = 5.7%
content of combustibles in exhaust s| - 1.5%
heat of fuel combustion QJJ = 17.6 *ilobules Heat
71
-------
Four regenerative air pre-heaters were installed at the boiler's
outlet into the exhaust tract. Fly-ash trapping is done by two four-
2
field precipitators with active sections of 265m each. The precipitators
are placed after the air-preheaters.
The precipitator's active zone consists of a system of C-shaped
precipitating electrodes and needle-strip coronal electrodes. Rapping
is done by hammers.
The precipitator fields are supplied with rectified high-voltage
by ATF-1600-type step-up rectifiers. Each field consists of two
electrically isolated parts (semi-fields). The high-voltage current
supply circuit is the "unit-semifield".
A 14% ammonia water solution, which was fed into the gases after
the air pre-heater, was used during testing.
The lay-out for ammonia conditioning consisted of an intermediate
space of 5m volume, equipped with a level indicator,, two pumps mounted
parallel to each other on one fra-e, distribution collectors for acqueous
ammonia solution and compressed air and shut-off armature.
Ammonia enters the intermediate space from the reagent stock station.
From the intermediate space, the ammonia is fed into the distribution
collector, which is a pipe with branches having calibrated apertures.
The compressed air system is analogously arranged. Compressed air
atomizes the ammonia water solution.
During the test period, the boiler agregate worked with nominal
parameters, namely:
unit's electrical load - 511 megawatts
superheated steam pressure - 23.53 - 23.73 megapaschals
boiler's steam yield - 1655-1665 tons/hr
72
-------
superheated steam temperature - 540-543°C
The precipitator test results are introduced in table 1.
Table 1
•
•
Precipitator parameters. '
•
•
Ammonia content (parts per million)
Gas temperature at inlet, °C
Gas temperature at outlet, °C
Gas rarification, kilopaschals
m3
Gas volume at inlet, thousands j~
without ! with
•
conditioning conditioning
•
•
-
165
158
3.3
1520
3D
164
157
3.3
1535 .
40
JG4
157
3.3
1545
Gas dustiness, g/mj
at inlet
at outlet
Gas speed into precipitator, m/s
Effective particle drift velocity, cm/s
Degree of gas cleaning,%
Average precipitator
52.8
0.99
1. 71
4.93
97.5
30.9
0.95
54,7
0.3D
1.74
6.85
99.35
36.6
0.37
54.7
0.20
1.75
f-3
S9.56
36.9
0.33
Average precipitator corona current, amps 0,95
Industrial precipitator test results have shown that with a
velocity of 1.74m/s, the introduction of 30 parts per million of ammonia
resulted in an increase of cleaning of from 97.5 to 99.35* and in a
decrease of exhaust dust by 3.2 times.
73
-------
The addition of 40 parts/million of ammonia, under other equal
conditions makes it possible to increase ash precipitation efficiency
up to 99.56% and decrease outlet dust content from 0.99g/m to
0.20g/m (i.e. by 5 times).
The increase in the degree of gas cleaning was due to lowering
the intensity of back corona. This is supported by increasing the
working voltage an average of 10-25% and lowering the magnitude of
corona current along the precipitator fields by 1.7-4.6 times.
Voltampere characteristics were taken both with stack-gas conditioning
and without it. With the introduction of ammonia conditioning, voltampere
characteristics shift to the right, during which corona current is lessened
and the loop, characterizing back corona intensity, narrows.
Measurement results of ekibastuz coal-ash resistivity give indirect
confirmation of the lessening of back corona intensity during ammonia
introduction. The resistivity measurement was taken both with and without
•
ammonia stack-gas conditioning. According to measurement data, the
resistivity of fly-ash at the precipitator inlet, _1.2-10 ohmmeters,
taken during gas conditioning by a 14% ammonia water solution, was lowered
e
by a factor of 3 and comprised £-10 ohmmeters.
The fly-ash resistivity at the precipitator inlet was determined
by use of a unit for measuring dust resistance (DRM).
The DRM unit's active principle lies in measuring dust resistance
of a layer of a given thickness, formed in the corona discharge field.
Constancy of the layer's thickness is assured by dust precipitation on the
comb-type precipitating electrode. Two such electrodes, arranged strictly
on one plane, form the comb gap.
-------
0 tn
er "uiati 1.1113
were mdae or cue i eia i
ekibastuz and kuznets TR-type coal -ash resistivity to temperature
in a range up to 350°C, when resistivity is determined by volume
conductivity (fig. 1.). The ash's chemical composition strongly
effects its resitivity magnitude (table 2). Silicon and aluminum
dioxide with a few admixtures of iron, calcium, etc. Oxides are the
basic components of fly-ash. Ekibastuz coal fly-ash contains 64% Si02
and 22% AlgOg, which determines its infusibility and abrasiveness.
Table 2 •
components
ft 3E2 4, & Q /fr 0 3%
calci-
nation loss
content,
% of mass
I 64.4 22 4.6 J,5. 0.4- ' 0.6' I.I
U.w
Ash dispersion composition (table 3) is determined in industrial
conditions by the rotational separation method. Ash sampling was done
at *he precipitator inlet without ammonia conditioning, and at the outlet,
both with and without the conditioning.
Tobfe 3
r**T"tic2^ <3 i r*i Htur
d, m fflj/'-f TjiniP
content of parti-
cles of diameter
less than d,
% of mass
terfainal velocity
in air, orv/second
2,5
94
O.C4 '
4,0
•
'84
O.C9
6,3
82
0.24
JO
70
C.C2
16
55
1. 53
25 40
37 20
4.0 9.0
d^ A— 21 micro]
,
standard de-
viation,
b = 3.2
It has been established experimentally, that for ash samples
containing less than 10% FegO-j, the magnitude of resistivity at a
temperature of 356°C depends on the ^0 content (fig. 2).
"TS
-------
Ash resistivity, measured in industrial conditions, practically
coincides with the value, determined by chemical composition with help
of the given relationship. The average resistivity value at 350°C,
determined by the relationship in fig. 2, comprises 7-10 ohmmeters
(0.35; Na20) for ekibastuz coal and 2.3-107 ohmmeters (0.65% Na20) for
kuznets coal. Therefore, as a result of studies, the relationship of
resistivity to temperature was derived. It has been shown that with a
gas temperature increase to 350°C, the resistivity decreases by three
orders of magnitude, and its value in these conditions may be approximately
determined according to its content of alkaline. oxide metals.
The effect of small quantities of ammonia introduced into stack gases
is connected with particle surface conductivity in the ash precipitating
layer. With the addition of ammonia, its reaction with water steam and
SO,, contained in the stack gases, occurs.
NH3+ S03 + H20 -» (NH42 S04
Ammonia sul fates formed during reaction increases the surface
conductivity of fly-ash particles.
One of the basic characteristics of preci pita tors is particle
drift velocity. As a result of conditioning, the effective drift
velocity grew 1.44 times and measured 7.34cm/sec. under reagent consumption
at 40 parts per million. In this case, the degree of gas cleaning
measured 99.5%. Increasing the precipitator size 1.5 times was required
to get such a degree of efficiency without gas conditioning.
76
-------
Thus, industrial condition studies of ammonia stack gas conditioning-
during trapping of high-resistance ekibastuz coal-ash has shown that this
method is an efficient means for increasing the degree of cleaning.
77
-------
v, ohm-
iheters
fO
8
c
\
\
w
1CO
200
SCO
C'C
Fig.l« DfrpcAct&»\c-Y of kuznets and ekibastuz coal ash resisti-
vity crx temperature.
•—• kuznets coal
ekibastuz coal
78
-------
.meters
Fig. 2.
under 35Q°C (L-0 %
(Fe-O, content< 10%)
1. Ekibastuz coal
2. Kuznets coal
of ash resistivity
Na20 content
79
-------
SODIUM CONDITIONING TEST
WITH EPA MOBILE ESP
Steven P» Schllesser
Acurex Corporation
(Although the information in this document has been funded
by the U.S. EPA, it does not necessarily reflect the views
of the Agency, and no official endorsement should be inferred.
The paper is as delivered at the meeting; Tables 2 and 4B are
missing.)
, 1.0 SUMMARY AND CONCLUSIONS
The objective of this pilot program was to determine the condition-
ing effects of adding anhydrous sodium carbonate into a "cold-side"
slipstream with respect to the collection performance of an electrostatic
precipitator (ESP). A power plant combusting low sulfur,.low sodium'
western coal generated the high resistivity ash (2.5 x 10 OHM-CM
@270*F) for conditioning evaluation.
A performance evaluation was conducted on a pilot scale precipitator
which treated the base and sodium-conditioned flyash. The program,
conducted over several weeks, consisted of twenty days of operating and
testing. For each ash species, the pilot precipitator treated 28.3
m /min (1,000 acfm) of flue gas at an average of 110*C, maintaining a
specific collection area equal to 57 m /nr/sec. (290 ft /Kefm).
In situ resistivity measurements, precipitator operating conditions,
and particulate concentration and sire distribution measurements con-
stitute the data assembled for the comparative demonstration.
The following results reflect the effects of conditioning the
base ash with a 1.0-1.51 concentration of sodium carbonate as Na.Ot
1L A sixfold reduction in specific resistivity,
suppressing.the base ash from 2.1 x 10 OHM-CM
to 3.7 x 1011 OHM-CM.
80
-------
2. Relative Improvements in achievable current
densities ranging from a factor of two to six.
Under steady state conditions, maximum current
densities were 1-8 nA/cm for the base ash,
•nd 10-24 nA/cm for the conditioned ash.
. 3*. Enhancement of particulate collection levels,
resulting in an improvement in collection
efficiency from 98.19 to 99.46 percent,
associated with a reduction in outlet loading
from 140 to 40 mg/DSm.
4. Improvement in fractional efficiency character-
istics, particularly in the fine particle range,
as demonstrated by the following results:
FRACTIONAL EFFICIENCY?
0.3u 0.5u l.Ou 3.0u
Base Ash 69.4 81.2 93.9 98.9
Conditioned Ash 91.fr 93.7 97.4 99.7
Oh this pilot scale basis the difference
in collection efficiency between the base
ash and the sodium-conditioned ash was
significant enough to move performance
from noncompliance (0.122 lb/10 Btu) to
compliance (0.0361 lb/10° Btu).
Assimilation of the test results prompt the following conclusionst
1. Conditioning by the admixture of conductive
material to reduce specific resistivity of
the process effluent, in quantities small
enough to be potentially economical, has
been demonstrated.
2. A material for such conditioning, which has
the potential to be economical, commercially
available, and environmentally acceptable has
been identified.
3. This material can be utilized fpr flyash treatment
without special equipment or extensive mainte-
nance requirements by simply Injecting it into
• pilot scale process stream.
81
-------
4. The effects of using sodium carbonate as a
conditioning agent coincide with alternative
•ethodst
a. An Increase In the operating
corona points for each preci-
pltator field.
b. An Increase in collection perfor-
•ance, primarily In the fine particle
alee range.
Mote:
Federal new source performance standards for coal-fired steam
electric plants of,250 x 10 Btu or greater heat Input for particulate
natter -0.1 lb/10 Btu. 2 hour average.
2.0 INTRODUCTION
The program objective was to determine the effects on preclpltator
operation and collection performance of adding sodium carbonate into
the exhaust of a boiler burning low sulfur low sodium western coal.
The specific effects of interest were:
1. particulate resistivity level
2. precipitator operating conditions
3. particulate removal characteristics
The feasibility of utilizing conditioning agents at the test
site (Montana Power Station, Colstrip, Mont.) is documented by
another power facility (Montana Power/Corrette Station, Billings,
Mont.) using coal from the same strip mine. Injection of proprietary
conditioning material into the flue gas before it entered the preci-
pitator dramatically affected operating corona points and Improved
particulate collection performance so that compliance limits were met.
In situ resistivity measurements at 135*C at.the test site determined
the flyash resistivity to.be in the lower 10 OHM-CM range, outside
the preferred range of 10 to 10 OHM-CM. Compositional analyses
of the coal and flyash provide the basis of this excessive resistivity.
I.e., low sulfur (1.01 percent) and low sodium (0.31 percent) content.
\
The experimental conditioning agent was anhydrous sodium carbonate
(Na2 CO.), a dry, powdery material with size distribution comparable to
flyash {mean site * 15 y). Charging was achieved by metering the
•aterlal with a calibrated screwfeeder, and entraining it with a cc
82
-------
pressed air supply through an injector. Identical test methodology
employing a pilot-scale preclpitator vas implemented for a two phase
prograa:
1. Operation with sodium carbonate addition
and a A - 5 percent rate relative to flyash
concentration.
2. Operation without conditioning additives.
f
The pilot scale ESP is one of three mobile field units owned by
the Utilities and Industrial Power Division, Industrial Environmental
Research Laboratory-U. S. Environmental Protection Agency (EPA),
Research Triangle Park, North Carolina. A pilot scale scrubber and
baghouse complement a fleet of conventional collection systems designed
to offert
Comparative evaluation of removal and economic
performance factors among the three conventional
control systems on either traditional sources
or process streams, or inadequately documented
particulate-laden process streams.
A parametric evaluation characterizing a given
control device/emission source in search for
optimum design criteria.
Feasibility/demonstration programs on novel or
•odlfied processes.
Specific problem-solving programs in control
technology.
3.0 DESCRIPTION OF FACILITIES AND INSTALLATION
Power Plant
The test program was conducted at the Montana Power Company
Station at Colstrip, Montana. The recently constructed facility
represents current design and operating methodology for coal-fired
electric-power-generating stations. Modular venturi scrubbers
provide emission control for two 350 megawatt (MW) boilers.
More pertinent information is available via a performance evaluation
program conducted by Southern Research Institute.
Design operating data from the test facility is listed in Table 1,
along vlth power production data for several days during the test
prograa. Table 2 presents coal and flyash analyses, as well as gas
composition data.
83
-------
Table 1. COLSTRIP POWER PLANT DESIGN AND OPERATING DATA
Unit 11 Operating Data As Designed
358
330
282
212
1) Coal Feed Rate, tons/hr. (appr.)
220
185
165
120
2) Steam Production. Ib/hr. (design) 2.464.261 2.228.286 1,871.761 1.403.821
3) Output Steam Pressure, PSI (design)
2,520
2.400
2.400 2.400
CD
4) Output Steam Temperture, *F (design) 1.000
1.000
1.000 1.000
3) Generation Rate. MW
358
330
282
212
6) Gas Volume. Kcfm
1.208
1.114
952
716
7) Temperature at Cold-side
Air Prcheater, *F
300 to 250'F (290*F Normal)
-------
PILOT ESP FACILITY
General Description
The mobile ESP facility consists of two separate units Mounted on
freight trailers 40 feet in length. Figure 1 is a floor plan of the
two units. The first unit is the process trailer which houses a five-
section ESP and all auxiliary equipment. This equipment consists of:
• Heat-traced, insulated 10" duct.
— Flow rectification equipment containing
vaned turning elbows and dlffuser sections.
- An I. D. fan with a cooling system.
- A screw conveyor and gate valve for the
removal of the collected material.
- .Electromagnetic vibrators for rapping the
collection plates.
/
- Five transformer-rectifiers, with rated
capacity of 50 KV and 10 MA. DC.
The system is designed to operate up to 1,000*F over a flow range of
28-35 m /min (1,000-3,000 acfm). Figure 2 shows an isometric of the
high voltage field design.
The second unit is a control/laboratory trailer containing all
process controls, monitors and recorders plus provisions for an
analytical laboratory and spare equipment storage.
Pilot ESP design specifications are compared with the specifi-
cations typical of full scale precipitators in Table 3. Comparison
of each design parameter between the pilot and full scale units
•hows thatt
a. Fixed design parameters on the pilot unit lie
midway in the normal range for full scale installa-
tions.
b. Operational parameters contain the flexibility
to cover the spectrum typical of commercial
units.
c. The single exception to conformity with
commercial installations is the plate area,
the inherent concession of pilot-scale
methodology.
85
-------
CO
STORAGE^
/WORK TOP
\
T 1 %
T5T1
CONTROL ARFJl _^/ CONTROL PANEL
WORK TOP
LABORATORY
AREA
' \ 7
ORf \ ITRANSFOI
TRANSFORMER MOTOR CONTROL
CENTER
UMBILICAL __
CONNECTIONS
TRANSFORMERS
RS
t
I.D. FAN
'• v ' i u. • • ,u i i iu
l__J I 1 C I I 1 I 1
•cr
f-4
\
HIDE INLET
ANGLE SAMPLING
VANED SECTION
DIFFUSER
\ \
HTSH VOLTAGE FIELDS (5)
OUTLET
SAMPLING
SECTION
Figure 1. Plan view of mobile ESP facility
-------
TROLLEY
•VIBRATOR
THERMAL
INSULATION
BAFFLE
SUPPORT
RACKET AND
INSULATORS
PLATE
COLLECTION
ELECTRODE
HIGH
VOLTAGE
CORONA
FRAME
WIRE
DISCHARGE
ELECTRODE
SPACER
RODS
Figure 2. Removable high voltage field
87
-------
Table 3. FULL SCALE AND PILOT ESP DESIGN SPECIFICATIONS
Full Scale"
Discharge Wire Diameter, en 0.25
Wire-Wire Spacing, cm 10-30
Wire-Plate Spacing, cm 10-15
Specific-Collection Area, 20-150
(in /mVsec)
Collection Area/T-R Set, m2 50-750
Aspect Ratio 0.5-1.5
Specific Corona Power 100-1,000
(vatts/m /sec)
Current Density, nA/cm 5-70
Gas Velocity, m/sec 1-5
No. of Fields in Series 2-8
Pilot ESP
0.25
18
13
20-100
9
0.67
100-1.000
5-100
0.5-2
2-5
88
-------
SITE INSTALLATION
A slipstream was withdrawn at a location downstream from the
combustion-air preheater and upstream from the pollution control
equipment. A 10" pipe located across the midpoint of the process
stream served as the means of sample extraction. Sixty feet of
heat-traced, insulated 10" pipe transported the slipstream to the
pilot precipitator. Ambient air quality standards dictated the
need for returning the particulate-abated stream to the access
locations for subsequent sulfur dioxide (SO*) removal. An access
door (2V x 4*) was fitted to the stack to accommodate the extraction
and return connections, along with a 4" coupling for in situ resis-
tivity measurements. Figure 3 illustrates the positioning of the
access, ducting, process unit, and sampling and Injection locations.
4.0 PROGRAM METHODOLOGY
PILOT PRECIPITATOR OPERATION
To ensure that the test program was conducted on a sound basis
with specific guidelines, protocol for installation, preparation,
operation and data acquisition was established and is discussed below.
Installation
Considerations for locating the slipstream access point required
that:
1. The conditions of the process stream be
typical for those of full-scale control
unit installation.
2. A representative stream could be withdrawn
from the process effluent, i.e., flow dis-
turbances are relatively absent.
3. The location and routing be appropriate for
the transport and maintenance of a representa-
tive stream.
Operation
The normal operating schedule allowed for one- week test increments,
starting at noon Monday and continuing until Friday evening. Operating
on a continuous rather than daily start-up/shutdown basis enhanced
data credibility and reliability for the following reasonst
1. Steady state conditions would be maintained
rather than approached.
89
-------
PRIMARY STREAM
LOCATION LEGEND
(1) RESISTIVITY MEASUREMENT,
(2) SODIUM INJECTION
(3) INLET PARTICIPATE SAMPLING
(4) RESISTIVITY MEASUREMENT, PILOT STREAM
(5) OUTLET PARTICULATE SAMPLING
PRIMARY
STREAM
PILOT ESP
Figure 3. Pilot ESP site installation
-------
2. Extended operating time would permit clearer
Illustration of time-related characteristics,
e.g., flyash accumulation on corona vires and
collection platea.
3. Corrosive conditions due to condensation
accompanying startup and shutdown periods
would be minimized.
4. A uniform daily testing program could be
•ore reliably established and maintained.
Because the reliability of the high voltage cables was limited,
approximately half of the test was conducted on an Interrupted rather
than continuous schedule. Prior to weekly start-up, the precipltator's
internal parts were cleaned in order to restore initial conditions.
This process consisted of removing flyash from discharge wires and
collection plates. Removal of the precipitated flyash between the
base ash and sodium test Increments was necessary to isolate the
causal relationships of the flyash species. Since it has been shown
that, due to reentralnment, rapping factors such as frequency and
intensity have significant effects on effluent loading, the rapping
cycles were terminated during the partlculate measurement periods.
Data Acquisition
The following operating data were recorded on a semi-hourly basis:
• inlet gas temperature
• outlet gas temperature
• Individual voltage levels
• individual current levels
Additionally, MV-IH data (corona discharge curves) were generated
before and after the daily sampling periods, and following the weekly
Internal cleaning. Sampling activities were conducted during normal
working hours, with operating data acquisition and analytical activities
being performed during the remainder of the day.
PARTICULATE MEASUREMENTS
Partlculate measurements were taken using a sample train, as
depicted schematically in Figure 4. Stainless steel filter holders
were used with 47mm glassRfiber filters to provide mass concentration
data. Seven stage Brinks and U of W Mark III Impactors containing
greased substrates were employed to measure the inlet and outlet size
distributions, respectively. A common grease mixture (20 percent
Apiezon "H" with benzene) was applied to the metallic substrates used
with each Impactor. Quality assurance measures and procedures for
conditioning, greasing, and weighing substrates were conducted in
91
-------
DRVING
COLUMN
SAMPLE
(•AS
VACUUH
PUMP
METERING ORIFICE
CONDENSER
CONTROL
VALVE
HEAT
EXCHANGER
COIL
DRV
GAS
METER/ THERMO-
COUPLE
BY-PASS
VALVE
MARNEHELIC
GAUGE
Figure 4. Sample train schematic
-------
accordance with Reference 6. Sample ports, available at virtual plug-
flow locations, consisted of 1" stainless steel couplings welded to
the duct, 10.75" I. 0.
Extractive probes (1/4" stainless tubing, 12" long) were mounted
in the ports and fitted with interchangeable nozzles to.allow isokinetic
sample removal. The probes were positioned at average velocity locations*
Particulate concentrations and size distribution determinations were con-
ducted at inlet and outlet locations. Two determinations of each
type at both locations constituted a test unit consisting of eight
individual measurements. Substrate and filter weight gains were
processed nightly, providing qualitative feedback on precipitator
performance and data quality.
RESISTIVITY MEASUREMENTS
A point-to-plane probe was used to measure in situ resistivity.
Maintenance and procedural operations were conducted in accordance
with the recommended practice for resistivity measurements. The
primary location for these determinations was a reduced velocity area
(3-5 m/sec.) immediately upstream from the first high voltage field.
Longer sampling periods (1-2 hours) were required for the relaxed flow
region (1-2 hours) than for the primary process Cs-1 hour). Quality
assurance tests were conducted in situ at the slipstream access,
yielding consistently comparable resistivity levels.
i
Three values of resistivity are reported for each test unit,
and are labeled:
* ^A ^park Method •* Breakdown Strength)
• P. (Spark Method at Electric Field
Strength - 10 kv/on)
• Pc (Re: V-I Method).
Each resistivity value is calculated as the ratio of the electric
field to current density. A complete discussion of the equipment,
procedure, and significance of resistivity measurement is contained
in Reference 7.
5.0 DISCUSSION OF PERFORMANCE RESULTS
The data base presented in the following sections consists of
five measurement categories:
• mass balance of sodium material
• specific resistivity
i
• operating measurements
93
-------
• particulate Bass measurements
• particulate size distribution measurements
Unavoidable fluctuations in boiler load, gas temperature, and
precipitator operation occurred throughout the test program, but
these fluctuations did not, in the final analysis, obscure the
results of the demonstration test program. The data clearly support
the general conclusions that, for aodium carbonate addition:
•,
• specific resistivity of the particulate was
reduced,
• precipitator operating levels were Improved, and
• particulate collection efficiency (total and frac-
tional) was enhanced.
CONDITIONING BT SODIUM CARBONATE ADDITION
Resistivity conditioning was achieved by metering and pneumatically
injecting anhydrous sociiura carbonate into the slipstream immediately
adjacent to the primary stream (Figure 3). Mixing was achieved by
injecting the material into the center of the slipstream, and trans-
porting the mixture through 60 feet of duct, four 90* elbows, and
the flow rectification sections at the trailer entrance. A previous
In-house program, In which sodium carbonate was injected 35 feet
upstream from the flow rectification section, demonstrated, by analyses
of samples taken from each field, that equal and consistent distribu*'
tion could be achieved. Injection was accomplished through use of a
screwfeeder, an air Injector, and an air compressor. Figure 5
illustrates the configuration of the injection assembly. The material
was metered through the calibrated screwfeeder into the injector,
where compressed and induced air entrained and delivered it through
a gently bending probe Into the slipstream.
The amount of sodium carbonate to be injected was calculated
as sodium oxide, with conversion being made on the basis of molecular
weight ratio:
Mole. Wgt, Na-0 - 62
= . 0.585
Mole. Wgt, Na20 CO -106
Thus, under most conditions, 11.5 gm/min of sodium carbonate
was added to represent 6.7 gm/min as sodium oxide, reflecting a
5.0 percent injection rate relative to a mean flyash rate of 135 gm/min.
Analysis of the collected ash from the high voltage fields
(Table 4) shows an average 1.16 percent differential collection rate
94
-------
in sodtun oxide content between the base and conditional ash. These
independent determinations reveal an unaccountable portion for a mass
balance of sodium material. The discrepancy between the injection and
collection rates suggests two possible occurrences:
1. Deposition of the sodium material between
injection and electrostatic collection, and
2. Penetration of the sodium through the ESP.
The latter contingency cannot be well supported, since the unaccountable
portion of sodium carbonate is ten times the concentration of the
penetrating ash. The author suggests that the unaccountable sodium
material was deposited during transport to the ESP through a non-optimum
injection system.
The beneficial irony of sodium conditioning is that one gram of
sodium oxide caused an equal amount of flyash to be collected. Numeri-
cal analysis correlating the precipitated amounts of sodium and the
causal increase in flyash shows a 1.16/1.0 relationship.
The term "conditioning" in the context of controlling particle
resistivity usually implies the addition of moisture or chemicals to
the carrier gas. In a broader sense, conditioning includes effectively
controlling particle resistivity by any appropriate means, such as
temperature or composition control.
Conditioning by sodium carbonate injection, unlike moisture and
chemical conditioning, does not change the conductivity of the original
flyash particles. The conditioning effect is achieved by simply
injecting a conductive material into the stream, which, when collected
on the plates with the original high resistivity material, provides
ionic carriers or additional electrical paths for the corona current.
The insulating effect of the high resistivity flyash is reduced, allowing
the preclpitator to treat the stream at a lower resistivity level.
The purpose of these .tests was to determine the improvement in
precipltator performance associated with sodium carbonate addition.
To date, this approach to conditioning has several limitations:
• Some applications require large quantities
of conductive material.
• Some economically viable conductive materials
are unavailable*
• The material is difficult to adequately distri-
bute into the process stream.
• Uncertainties result from selective precipi-
tation of the conductive material3
95
-------
INDUCED AIR
1. VIBRASCREW FEEDER SCR-20. 3/8 SCREW. 0-0.1 CFH
t FABRICATED CONNECTION BETWEEN FEEDER AND INJECTOR
3. JET INJECTOR. McMASTER-CARH, P.O. BOX 4355. CHICAGO 60680. CAT. 4977K-11
4. FABRICATED PIPE
1 COVER PLATE APPRO* V DIAMETER. WELDED TO BENT PIPE
I COMPRESSED AIR SUPPLY LINE. APPROX. 70 PSIQ
Figure 5. Depiction of sodium Injection equipment
96
-------
CD
Table 4A FLYASH CONSTITUENCY DATA
NO Sodium Injection
ti2o
H«20
*2°
HgO
C«0
fe203
A1203
sioz
T102
P2°5
so3
tot
6/24
Cell
1
0.04
0.60
0.8
7.2
18.5
4.0
23.8
41.5
1.0 x
0.4
1.1
0.5
6/24
Cell
2
0.05
0.58
0.7
7.5
18.3
3.5
23.8
41.1
- 1.0
0.4
-
0.5
6/24
Cell
3
0.05
0.56
1.0
7.7
19.5
3.8
-
-
0.9
0.3
-
0.7
6/30
Cell
1
0.04
0.56
0.7
5.8
17.0
3.7
21.4
44.5
0.9
0.3
1.3
0.5
6/30
Cell
2
0.04
0.74
0.7
6.4
17.7
3.5
23.3
43.2
1.0
0.3
-
0.6
6/30
Cell
5
0.04
0.75
0.9
6.7 ...
17.8
3.2
24.2
41.9
0.9
0.4
-
0.6
8/23
Cell
3
0.04
0.61
0.8
6.5
19.6
3.9
23.7
41.8
0.7
0.5
1.2
0.3
8/23
Cell
4
0.04
O.61
0.8
6.8
19.3
3.2
24.4
40.7
0.7
0.9
1.4
0.2
8/23
Cell
0.04
0.71
0.8
6.5
18.6
3.4
25.0
40.2
0.6
0.6
1.7
0.6
8/24
Cell
3
0.04
0.59
0.7
6.4
19.3
3.3
23.3
42.1
0.6
0.5
1.2
0.5
8/24
Cell
4
0.04
0.59
0.7
6.8
20.5
3.8
23.6
39.3
0.7
0.6
1.3
0.4
8/24
Cell
0.04
0.61
0.8
6.8
19.7
3.3
-
-
-
0.6
-
-
-------
RESISTIVITY DATA AND RESULTS
In order to quantify the resistivity suppression of sodium con-
ditioning, resistivity measurements were taken on (1) the primary
stream near the pilot stream access location, (2) the slipstream
containing the base flyash, and (3) the slipstream containing the base
ash plus sodium carbonate.
All determinations were made with an in situ probe designed by
Southern Research Institute . The design of the probe and the procedure
provide the following data as presented in Table 5:
1. PA - Resistivity value taken at the breakdown
strength of the sample.
2. P. - Resistivity value for electric field
strength - 10 KV/CM.
3. PC - Resistivity value determined by the comparison
of volt-amp relationships before and after
precipitation of the sample.
4. EA - Electric field strength at breakdown.
P. is the preferred resistivity value, because voltage is applied
directly to the electrostatically precipitated sample, and determination
is made at the breakdown strength of the sampled ash layer.
The other two resistivity values, PB and PC, are less significant to
precipitator applications, but are offered to support the comparison
between the two ash species. The last column in Table 5, EA, represents
the breakdown strengths of the Individual samples. Compilation of over
35 tests resulted in the following log-averaged resistivity values:
12
«. 2.5 x 10 OHM-CM (on the primary stream
at 130-135»C).
b. 2.1 x 1012 OHM-CM (on the slipstream at
107-115*C with the base ash).
c. 3.7 x 10U OHM-CM (on the slipstream at
107-115'C with the base ash and 1.2 percent
sodium oxide/baae ash addition).
Comparison of these values indicates that:
• The characteristic resistivity of the gas
and partlculate stream were preserved during
transport to the pilot unit.
98
-------
Table 5. RESISTIVITY DATA
Run Mo.
6/22/7?!*
6/23/77
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
PA
A
2.9 £12
2.1 £12
. . .
5.9 Ell
1.1 £12
1.6 Ell
4.1 Ell
5.0 Ell
3.2 Ell
1.0 E12
6.0 £12
2.1 £12
...
...
2.0 Eft
...
8.0 Ell
4.1 £12
2.5 £12
9
B
2.1 £12
2.7 £12
...
5.2 Ell
2.2 E12
...
6.8 Ell
1.0 Ell
5.6 Ell
1.7 £12
1.0 £12
5.0 £12
...
...
3.6 Ell
...
1.3 £12
6.5 £12
6.0 £12
P
C
4.5 Ell
3.8 Ell
3.6 £11
2.7 Ell
2.1 £12
4.0 Ell
7.3 Ell
7.6 Ell
7.3 Ell
2.2 E12
2.5 £12
2.2 £12
1.5 £12
6.2 Ell
1.5 E12
2.2 £12
2.5 £12
2.6 £12
2.8 £12
41,
10,
. _
19
35
4
32.
26.
10,
20
25
£A
A
22. 24
17. 12. 17
.
27
26. 26
25. 27
8. 50
. .
„ .
55.
. .
29.
39.
32.
.
_
33. 73
.
50, 40
26
37
2.17 Ell • Log Mean Resistivity for Conditioned Ash 0 107-115*0
3.70 £12 - Log Mean Resistivity for Bash Ash t 107-115*0
PA - Resistivity Determined by Spark Method 0 Breakdown Field (OHM-CM)
P_ « Resistivity Determined by Spark Method 8 Electric Field
Strength - 10 KV/CM
PC » Resistivity Determined by V-I Method
EA • Electric Field Strength § Breakdown (KV/CM)
* Data Obtained at Pilot ESP Inlet
** Data Obtained on Primary Stream at the Pilot Stream Access Location
99
-------
• A sixfold reduction in specific resistivity
is attributable to sodium injection.
Research by Bickelhaupt has shown that, for surface resistivity
of flyash, the alkali metal constituents, particulary sodium,, serve
as the principal charge carriers In the conduction process. This
observation led to two previous field tests demonstrating that
•odium conditioning did reduce specific resistivity, and that the
reduction was predictable by the relationship of:
VP2 • 2
Pj • measured resistivity in OHM-CM for W2
P2 - predicted resistivity in OHM-CM for W2
V. « measured weight percent Na20 in ash
W2 • added weight percent Ha20 in ash.
Demonstration of the sensitive and predictable relationship
between sodium content and resistivity is shown from the test data:
For Wx - 0.63, W2 - 1.79, P - 2.1 x 1012:
P2 • 2.6 z 1011 OHM-CM
The predicted value correlates with the measured resistivity value
of 3.7 x 10" OHM-CM.
OPERATIONAL DATA
Operating corona points and the associated V-I data over the
allowable range were taken regularly during the program. The
operating corona data reflect the electrical conditions during the
sampling period for a given test. It should be noted that the control
of the transformer-rectifier sets was accomplished manually, as the
pilot facility was not equipped with automatic controllers customarily
included in full-sired preclpltators. It should also be pointed out
that the operating corona points were set 4 KV lower than the maximum
allowed. This was done to prevent frequent high voltage cable break-
downs responsible for delays incurred during the initial testing
phase.
Table 6 and Figure 6 depict actual and comparative current
densities for the base ash and ash-with-sodium conditioning tests.
The above table and figure provide a histogram of the current
densities related to on-line time following Internal purging. The
degradation rates are comparable for each species, but the base ash
curves show more severe current reductions as they approach steady
100
-------
Table 6 FRECIPITATOR ELECTRICAL OPERATING DATA -
SECONDARY VOLTAGE (KV), CURRENT DENSITY (nA/ca )
Base Flyaeh
Cell
1 39 .11
2 39.21
3
4
5 ,41,28
Tine*
(day.) 1
Teat
Mo. 1
40,4.2 40.2.3
40.3.1 40.1.1
39.9.0 39.6.3
1% 2
2 3
38,18
39,24
39.22
*
11
39.10
37,16
36.12
*
12
40.10
40.21
40.25
1
13
40.27
38.21
38.35
*
17
38 .10
38.11
36.26
1
18
36 .6.0
40.7.0
32,13
2
19
36.4.0
38.3.0
32 .8.0
3
20
Base Plyash * Sodiua
Cell
1
2
3
4
5
•in*
46,36 46,13
39.27 42.16
38 31 38 .30
» h -i
39.8.0
36 .9.0
35.14
39,28
35,41
1%
42.7.8
41,19
37^6
37,26
37,23
2
44 ,9.5
40,18
40.20
39,24
2%
42.10
41.13
40.21
3
42 ,9.0
41,10
36.5.0
31!
38.23
40^4
38^8
. %
40,14
40,23
40,50
1
41-19
39-20
2
Teat
No. 4 5 6 7 8 9 10 14 15 16
Elaaped tiae operating since internal purge.
-------
50
40
30
20
10
0
O: BASE FLYASH
O* FLYASH I SODIUM
INLET
FIELD
±30
CO
§20
I10
n 0
50
10
30
20
10
0
OUTLET
FIELD
.1
1 2
ON-LINE TINE (DAYS)
Figure 6. Current density vs. on-line time
102
-------
state conditions. The inlet field histograms reflect s 2-3 factor
improvement in achievable current densities for the sodium conditioning
test series* having achieved a 10 nA/cm level compared to a 4 nA/cm
level for the base ash series. A consistently comparable relationship
exists for the outlet fields, with the sodium series achieving
20 nA/cm , relative to 7 nA/cm . The second or middle field histogram
indicates improvement for the sodium series, achieving 20 nA/cm
compared to 4 nA/cm for the base ash.
Reasonable correlation exists between the independent resistivity
measurements and the achievable precipltator current densities approach-
ing steady state conditions. The relationship between resistivity (P)
and current density (C. D.) is defined by P-E/C. D., where E is the
dielectric (breakdown) strength of the collected dust layer. Assigning
the limiting breakdown strength of 10 KV/cm for each ash species, one
can calculate the achievable current density from.the log-averaged
resistivity values. For the base ash (2-2.1 x 10 OHM-CM), the
calculated.current density is 4.8 nA/cm • For sodium-conditioned ash
(3.7 x 1011 OHM-CM the calculated value is 27 nA/cm . These theoreti-
cally calculated values correspond with the.empirical values averaged
from the operational data (5.0 and 15 nA/cm , respectively).
Comparison of the achievable voltage levels shown in Figure 7
for the two ashes are not as dramatic as the current density comparison
but higher voltages were consistently obtained for the sodium treated
ash.
The relationship of the voltage applied to each precipitation field
and the resulting current (corona discharge curves) may be analyzed to
gain Insight into precipitator operation and performance. Figure 8
portrays the corona discharge curves for the base and conditioned ashes
in graphical fora. These curves show the characteristic profile of
the volt-amp relationships from each field for steady state conditions.
Knowledge of the meaning of the V-I curves is a fundamental tool
necessary for proper operation of a precipitator. These data can
characterize the following parameters/conditions:
• corona current leakage
• corona initiation voltage
.• effective corona wire size
• effective wire-to-plate spacing
• alignment effects
• specific resistivity level
• specific collection area
103
-------
50
40
30
~ 50
§2 40
o 30
50
40
30
« — — —OSBASE FLY*SH
Q ei v*e" ft SODIUM
INLET
UO Q "Q— •
1 1 1 1
1 23
-
- Q— — n— — • — ^S— — — — — ..
°""
1 1 1 1
123
"" OUTLET
FIELD
Q
B^-i__^ ^_
J .. 1 ., ? , 1
2 3
ON-LINE TIME (DAYS)
Figure 7. Operating Voltage vs. on-line time
104
-------
30
I 20
!
10
_ : CURVE WITH BASE ASH
_ ; CURVE WITH SODIUM
SPARKOVER
OUTLET
FIELD
SECOND
FIELD
10
20 ?0 40
APPLIED VOLTAGE (KV)
Ftgure 8. Corona discharge curves
105
-------
• breakdown voltage
• achievable current density
• concentration of fine particulate
• relative position of each precipitation field
(inlet, second ..., outlet)
• variations in emission source process
Analysis and interpretation of these curves provide the basis
for the following generalities referenced in other sections of this
report:
1. A reduction in specific resistivity between
the base ash and sodium-plus-ash test series.
2. An Improvement in achievable corona point for
the sodium series.
3. An improvement in expected collection performance
for the sodium series.
4. The degradation in performance related to elapsed
time from internal clean-up.
5. A small increase in effective vire sice due to
ash build-up.
PARTICULATE COLLECTION RESULTS
Particulate Concentration Results
• Table 7 shows the averages of the inlet and outlet concentrations
| per teat used to calculate the average collection efficiency values in
! Table 9 and Figure 9. The inlet data exhibit considerable scatter
'; resulting from the inconsistent probe catches apparent in the raw data.
i Due to the .reduced flyash concentration and probe catch material, the
1 outlet data offer reasonable agreement and consistency. •
i ,
: The apparent scatter In the particulate data depicts an ordered
trend when the elasped operating time from the precipltator "cleaning"
i is taken into account. As previously noted, the corona wires and
I collection plates were restored to a clean state between conditioning
I and noncondltloning test series. The degradation in performance
j evident particularly for the base ash was caused by the accumulation
, i of high resistivity ash on the collection plate and the discharge !
t ,.,, • electrodes. The marginal differences in performance for the tests |
..(' .after recent cleaning do not reflect the performance levels achievable ;
I v .;; Y I •_..; ]._. .• ;
I'A'.'. MIH'.i-.i it
106
-------
for steady state operation. According to a best fit approach from
Figure 14, sodium conditioning vas responsible for reducing outlet
loadings from 140 to 40 mg/Ds • (0.122 Ib to 0.0361 lb/10° Btu),
resulting in an Improvement In collection efficiency from 98.19 to
99.46 percent*
Table 7
Average of Paniculate Data
(ag/SDm-*)
Test
1
2
3
4
5
6
7
B
9
10
11
12
13
14
15
16
17
18
19
20
Inlet
Outlet
182
133
394
29.1
45.6
40.1
46.3
47.6
38.2
36.2
33.2
30,
35,
17,
22.9
34.1
50.4
44.6
117
137
Meant 7,557
\
it;:
t..
4 . _.
P/V.I
107
-------
Table 8. SUMMARY OF PARTICULATE DATA AND EFFICIENCY RESULTS
Data
Inlet concentration
(Mg/DffiT)
Outlet concentration
(Mg/DNtr)
Inlet mean particle
•ire (y)
Outlet Mean particle
•ice (y)
Efficiency
Total mass (Z vt.)
Fractional f:
3.0y
1.0 y
0.5 y
0.3 p
Base Flyash
7,550
140
15
2.6
98.2
98.8
94
81
70
Base Ash with Sodium
7,550
40
15
2.1
99.5
99.7
98
94
92
108
-------
4x10"!
3
SxlO2
•
4
w
i.
2x10
r2
2X101
I
I
2 3
ON-LINE TIME (OATS)
Figure 9. Prectpitator collection performance
109
-------
Table 9 PARTICLE SIZE DATA AND
FRACTIONAL EFFICIENCY RESULTS1
Test Mo. Mass Mean Diameter
Fractional Efficiency
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Base ash
Sodiusi me
Inlet
8.6
9.7
26.0
13.5
13.0
8.8
7.4
7.8
7.4
9.9
11.5
21.0
8.0
23.3
17.0
16.0
12.0
22.5
27.0
29.0
•eant
ant
Outlet
1.6
1.7
3.5
1.6
1.6
1.2
1.4
1.7
3.0
2.1
1.9
1.2
1.4
2.3
4.8
2.1
1.4
5.5
4.7
2.6
2.1
29.3
89.7
87.9
85.7
93.3
93.2
94.9
79.6
93.3
37.1
61.1
92.5
96.7
91.5
98.3
37.5
85.0
90.7
83.2
69.4
91.8
42.3
90.9
90.5
88.3
94.0
94.6
96.4
83.7
95.6
71.2
72.0
95.9
98.5
93.4
98.4
78.7
88.6
93.5
88.0
61.2
93.7
90.7
94.9
87.2
93.5
97.4
97.7
98.7
94.9
97.8
97.7
97.0
98.7
99.1
98.1
99.1
93.2
97.0
93.1
90.0
93.9
97.4
89.7
99.2
96.0
99.7
99.5
99.5
99.8
99.8
99.6
99.6
99.4
99.9
99.9
99.9
99.9
99.7
99.7
99.5
96.0
98.8
99.7
Data obtained from daily averages.
PARTICLE SIZE AND FRACTIONAL EFFICIENCY RESULTS
Brink and the University of Washington (U of V) imp actors were
employed for obtaining size measurements on the inlet and outlet
streams* respectively. Quality assurance measures and proceudres
for conditioning and weighing substrates were performed in consistence
with EPA/IERL specifications. Due to small weight losses on the
greased substrates for the D of tf, trial impactor flowrate was reduced
for all subsequent tests from 189cc/S to 142cc/S. Evidence of sub-
strate scouring was consequently eliminated with flowrate reduction.
Table 9 presents the particle size data and fractional efficiency
results for all test runs. Mass mean diameter on the inlet ranged from
8-29 , with an average N.M.D. of 15 • The outlet stream contained
M.M.D's from 1.2—5.5 « with 2.1 and 2.6 being the average mean size
for the sodium conditioning and base ash effluents, respectively.
110
-------
Figure 10 demonstrates the Improved fractional collection performance
associated with sodium conditioning, particularly in the subnlcron
range.
REFERENCES
1. Private Communication with Bob Olmstead, Plant Superintendent,
Montana Power Corrette Station, Billings, Montana.
2. McCain, J.D., CEA Variable Throat Venturi Scrubber Evaluation.
EPA-600/7-78-094, U.S. Environmental Protection Agency, Research
Triangle Park, N.C., June, 1978.
3. White, H.J., INDUSTRIAL ELECTROSTATIC PRECIPITATION, Reading,
Mass., Addison-Wesley, 1963, p. 359.
4. "MOBILE ESP OPERATING MANUAL," Environmental Sciences Group,
Naval Surface Weapons Center, Dahlgren, VA., October, 1976.
5. Spencer, R.W., A'Study of Rapping Reentrainment in a Nearly
Full-Scale Pilot Electrostatic Precipltator. EPA-600/2-76-140,
U.S. Environmental Protection Agency, Research Triangle Park, NC,
1976.
6. Harris, D.B., Procedures For Cascade Impactor Calibration 6
Operation in Process Streams, EPA-600/2-76-144, U.S. Environmental
Protection Agency, Research Triangle Park, HC.
7. Nichols, G.N., Test Methods 6 Apparatus For Conducting Resistivity
Measurements, U.S. Environmental Protection Agency, Contract No.
68-02-1083. Final Report No. 3121-III., Sept.,1977.
8. White, R.J., Op. Cit., p. 312.
9. Bickerhaupt. R.E., Surface Resistivity & The Chemical Composition
of Flyash, Proceedings From Symposium On Electrostatic Precipitators
For The Control of Fine Particles, For National Environmental
Research Center, PB 240 440, Jan., 1975, p. 246.
10. Bickelhaupt, R.E., SODIUM CONDITIONING TO REDUCE FLY ASH RESISTIVITY,
EPA-650/2-74-092, U.S. Environmental Protection Agency, Research
Triangle Park, NC.
11. Banks, S.M., McDonald, J.R., Sparks, L.E., VOLTAGE CURRENT DATA
FROM ELECTROSTATIC PRECIPITATORS UNDER NORMAL AND ABNORMAL CONDI-
TIONS,
Proceedingst PARTICULATE COLLECTION PROBLEMS USING EPA'S IN
THE METALLURGICAC INDUSTRY, EPA-600/2-77-208, U.S. Environmental
Protection Agency, Research Triangle: Park, NC, October, 1977.
Ill
-------
ACKNOWLEDGEMENTS
This program was sponsored by EPA with participation by Montana
Power Company, Southern Research Institute, Monsanto Research Corpora-
tion, and Acurex/Aerotherm. The program was initiated under EPA Con-
tract No. 68-02-1816 with Monsanto Research Corporation, and
completed under Contract No. 68-02-2646 with Aerotherm.
The author expresses sincere appreciation to the following
individuals for their involvement with and contribution toward this
program:
Ivan Bonnette, Ray Hoffman, and Bruce Knusten of Montana
Power Company, Colstrip, Montana;
Dale Harmon and Les Sparks of EPA, Research Triangle
Park, North Carolina;
Grady Nichols and Jerry Sutton of Southern Research
Institute, Birmingham, Alabama;
Don Zanders, Billy Bowles, Tony Hojtowicz, and
Mark Wherry of Monsanto Research Corporation, Dayton,
Ohio;
Hal Buck, Mike Griffin, Randy Page, and Clyde Stanley
.of Acurex/Aerotherm, based at Research Triangle Park,
'North Carolina.
112
-------
5
4
3
2
I 1.0
^ 0.8
§
I 0.6
oc
3
0.4
0.2
O-.BASE FLYASH
QifLYASH S001UN
I
I
III I i II I I
20 '10 60 70 80 90 95 9899 99.6 99.9
FRACTIONAL EFFICIENCY, I HT.
Figure 10. Comparison qf mean fractional efficiency results
-------
V.I. SIKORSKI, V.V. KUTLYASHOV
IMPROVING ELECTRICAL FEEEDING SYSTEMS FOR ELECTROSTATIC
PRECIPITATORS (Thyristor Automatic Precipitator Voltage Regulator)
NIPIOTSTROM's Electrostatic Gas-Cleaning Laboratory has deve-
loped an automatic electrical-feeding parameter regulator for gas-
cleaning precipitators. It is intended for use in step-up transformer-
rectifier units with thyristor control of primary-power circuits.
The power circuit of such a unit is illustrated in fig. 1. The auto-
matic regulator (AR) carries out, by means of the thyristor control
unit (TCU), phase regulation of tension in the primary coil of the
high-voltage transformer (HVT), and consequently, in the precipitator
electrodes too.
During research on the converter-electrostatic precipitator
system as an object of phase regulation, basic automatic control
aims were formulated:
1} Continuous electrical power supply parameter support for
precipitators on a level assuring eificient gas-cleening;
2) Increasing corona discharge stability in the precipitator's
volt-ampere characteristic section corresponding to sparking and spark-
over;
3) Prevention of emergency conditions arising due to electri-
cal parameter instability of the regulated unit.
Electrostatic precipitators, examined as to their electrical
load qualities, possess many characteristic features.
First of all, they show instability of the voltampere charac-
teristic, bound with changes in many factors influencing the corona dis-
charge intensity in the environment of an industrial stack.
114
-------
Pig. 1. Functional scheme of automatic regulator ar.s
electrical scheme of power circuit of a thyristcr converter
unit.
precipitat
electrode
cs
PVC -
SDS -
AR - automatic regulator
FSF - feedback signal former
TVC - thyristor voltage control
TCU - Thyristor control unit
HVT - high-voltage transformer
TIRF - Thyristor impulse regulation former
ARC - angle regulation corrector SL
APC - amplitude-phase converter
MOPC ~ -maximum current power control
TCS? - transient condition signal former
PRC - program regulation control
CLSF - current limiter signal former
BRSF - basic regulator signal former
current sensor
• precipirator voltage
controller
spark discharge
selector
- sensory defense ur.
shorti.ic limiter
115
-------
The precipitator's voltampere characteristic instability
leads to frequent overloads in the power supply as a result of the
letter's power limitations. An automatic regulation system should,
therefore provide limits to the precipitator1s working current.
The second feature is frequent sparking and spark-over
-•in the precipitating space. They are, essentially, brief shorts
in the feeding source and are accompanied by significant current
impulse surges in the power circuit, and have thermal and dynamic
effect upon the circuit's elements. Along with this, spark discharges
are one of the basic conveyors of information about processes in the
precipitating space. Spark intensity characterizes the position at a
given moment of the working point in the voltampere characteristic
and its degree of closeness to spark-over conditions. Using spark in-
tensity, along with other electrical parameters, it is possible to
indii.ctly make a judgement as to the degree of back corona intensity,
dust content in the gas stream and the degree of dust contamination of
the ESP electrodes.
Spark intensity is characterized.by_the frequency of dis-
charge and by the energy given off by each spark.
Energy, given off by spark discharge, is proportional to
the electrode system's capacity and to the voltage at the moment of
discharge, squared.
The fullest characteristic of spark intensity is conversion
of average frequency into average spark energy- this is the average
capacity of the spark process. However, due to difficulty in meas-
uring this parameter, upon development of a regulator serving as
the spark process intensity characteristic, the relative duration
of transfer elements in the precipitator current was taken, indirect-
ly connected with spark capacity. To accomodate relative duration, it
is necessary to to understand the relationship of sum duration
116
-------
sf these processes having a space for integration period, to the
integration period in relation to the expression:
n
Pu - ftui » n'tu = fu.tu,
where: Pu - spark intensity,/*$/min;
T - integration period, min.
n - number of sparks during integration period;
tui- duration of ith spark, vs;
fu== - average spark frequency, I/ min;
tu= frtux _ average spark duration,ys
n
The third pecularity of precipitators is the considerable
electrical capacity of its electrode system.
The precipitator capacity, together with the inductive ele-
ments of the feeding unit's power circuit, form oscillation circuits.
During sparking and arc discharges, shock excitation of these circuits occurs, =.-.-
___ ~l
acconpanied by high and low,frequency oscillations. These oscillations
p *
cause impulse surges in the unit's circuit elements and give the es-
sential action to the spark formation process in the precipitating space.
Specifically, excitation of the low-frequency oscillation
circuit formed by the precipitator's capacity and the primary power
circuit's inductiveness, having resulted due to tension reduction
after spark discharge, is one of the causes of spark-over. Under these
circumstances, there occurs instead of separate sparks in the precipi-
tating area, a lengthy series of spark-overs in every half-period of
line-voltage.
It is necessary to relate to precipitator peculiarities,
examined in the light of electrical capacity, the increased proba-
bility of dead-end shorting and shorting through low resistance,
connected basically with breaks of coronal electrodes and the jamming
of dust in the precipitator hoppers.
117
-------
At low angles of thyristor conductivity, caused by the
preceding conditions and also due to actuation of the operating current limiting
system, the average current value of short circuits may be equal to
or less than the unit's nominal current, whereas its amplitude value
reaches levels representing a danger to power circuit elements. Semi-
conductor mechanisms, thyristors and high-voltage rectifiers, are
especially vulnerable in this case.
A specific feature of thyristor phase control of the
power transformer's primary circuit is the possible occurence of the half-wave
feeding of the primary coil. This is due to the circuits' discrete-
ness of transmission of positive and negative half-periods, relay
characteristics of the thyristors and the short duration of control-
ling impulses.
Half-wave conditions may subdivide into stable, caused
by irreversible damage to one of the circuit's branches, and quasi-
stable conditions, lasting a considerable period due to internal
positive feedback, present in the process because of, for example,
•the occasional not switching in of single thyristors.
Positive feedback is formed in this case as a
result of the presence in the circuit of non-linear induction
of the power transformer's primary coil. Following transmission, the
current half-cycle by-passes the non-remagnetized circuit. This induces
a sharp increase of its duration and amplitude. If such a half-cycle
increase prevents switching in of a thyristor in the successive half-wave,
a quasi-stable half-wave occurs that is, in contrast to the stable,
a reversible event.
The original omission of one of the thyristors is possible,
for example, in two instances:
1) On decrease by the regulator of the angle of inclusion/
06, of the thyristors to the measure of the least natural phase angle
118
-------
of the primary circuit, upon which cureent constriction angle a3
exceeds the angle of inclusion and the including impulse will be
given to the thyristor's control electrode with lack of anode
voltage?
2) On spark discharge in the precipitator, occuring in
work conditions when the angle of thyristor inclusion is less
than 90 degrees, the current has in this case a purely in-
ductive character.and the angle of constriction 03 approaches,
according to its value, 90 el.degrees.
Half-wave feeding conditions are, according to their
consequences, close to short circuits. Therefor an automatic
regulator should contain a device, preventing their development
during the occasional switch-off of thyristors.
On the basis of the preceding, the observation can be
made -hat stable and efficient work by unit systems and precipi-
tators may be assured under the condition that the automatic re-
gulator will carry out multipurpose control according to the
following directions:
1) That it contunuously mauntain in the precipitator a given
•relationship between spark intensity and electrode voltage, with the
purpose of optimizing gas cleaning;
2) That it extinguish spark-overs and prevent occurence of
secondary spark discharges, thus increasing the coronal discharges
stability.
3) Automatically limit in the precipitator the critical average
current value, in relation to the given unit;
4) Prevent occurence of half-wave conditions in primary power
circuit feeding;
5) During shorting in the precipitator, to disconnect the unit
after the necessary time lag;
119
-------
6) Limit the average short circuit current during the protective
time lag to a value, not exceeding 10% of the unit's nominal
current.
The automatic regulator, fulfilling these functions,
consists of the following functional circuits and units (fig.
1):
1) The basic control signal former circuit, forming,
along with the precipitator unit system, the basic automatic
regulating circuit; this contains the spark discharge selecter
(SDS), basic regulator signal former (BRSF) with the program
regulation control (PRO, the amplitude-phase converter (APC),
and the thyristor impulse regulation former (TIRF);
2) The transient condition signal former circuit,
consisting of the spark discharge selector (SDS)/ the transient
condition signal former (TCSF) and the remaining elements of the
basic circuit, except for the regulator signal former (BRSF);
3) A circuit, correcting the basic signal accord-
ing to the value of the working current, which consists of
a current limiting signal former (CLSF) and the maximum cur-
rent force control (MCFC);
4)A circuit for basic signal correction according
to the allowed value of the angle of inclusion for power thy-
ristors, consisting of a feedback signal former according
to inclusion angle former (RCSF) and an angle regulation cor-
rector unit (ARC);
5) A system for protection against precipitator
short circuits, including a sensory defense unit (SDU) and a
shorting current limiter (SL).
120
-------
Signals enter the regulator's input from the current
sensor (CS), from the precipitator voltage control (PVC) and
the thyristor voltage control (TVC).
The control angle of power thyristors, represented
by the phase angle of of transmission of the currents short-
term inclusion impulses, fed into the thyristor control junctions
of the thyristor control unit (TOJ), represent the regulator input signal.
The automatic regulator works in the manner described
below. After the unit is switched on, the basic signal former
circuit provides a smooth increase of the thyristors1 control
angle in relation to the regulator's basic booster character-
istic, thus the precipitator's voltage and current steadily
rise. With spark discharge in the precipitator, the spark
discharge selector divides transient elements in the current
and rectified voltage and forms right-angle impulses, whose
duration is equal to the duration of the spark current. The
succesion of these impulses is subjected to sticking non-
linear integration into the basic regulator signal former
unit. At the same time, each impulse induces intermittent
output-signal decrease in the regulator to the value, pro-
portional to the selctor's impulse duration, in relation to
the function:
where Aay - decrease in cogtrol angle of indi-
vidual sparks, el. degrees
Kn = 0.65; $3 = 18.6 - regulator constants
au - selector impulse duration, el.'ldearees,
121
-------
In intervals between spark discharges, the control
angle steadily grows* with proportionately decreasing speed.
Such a regulator output signal provides intermittent voltage
reduction for the precipitator during discharges and smooth
current build-up until the next spark-over. This improves
to a considerable degree the coronal discharge's stability.
The succession of drops and increases in control
signals forms a somewhat dynamic level of thyristor angle
control, determined by spark discharge intensity.
The dependence of averaged angle control on spark-
ing intensity represents the static characteristic of the ba-
sic signal regulator former circuit.
The approximate formula of the static characteristic
takes the form:
Pu - 3.33-103.k3.a'y,
"
where Pu=fu-tu - sparking intensity, ys/min.
a'y - rate of change of control angle, el.
degrees/us
Automatic control of feeding parameters, carried out
by the basic circuit, is directed at the system's retention on
the basic signal former's static characteristic which is in the
given case the regulator's programm.
The basic regulator signal former's circuit can be con-
sidered an astatic regulator of a set relation between sparking
intensity and the power thyristors1 control angle.
The quality of the control process can be characterized
by dynamic error, the static error being equal to zero.
122
-------
The automatic regulator has five fixed control programs
to which correspond five static characteristics, distinguished
by their slope toward the Pu axis. Program switching is produced
with the help of the program regulation control (PRO, which
changes the value to K .
P
With each spark-over the transient condition signal
former is triggered by the spark discharge selector's (SDS)
forward front. The action of the basic regulation program is
temporarily terminated and the transient condition program, is ac-
tivated, which determines thyristor control angle for the next
few half-periods of network voltage.
In accordance with this program, the control signal is
removed in entirety, at the moment of discharge, and is absent
during two half-cycles of network voltage. Then exponential
signal build-up follows with a time constant of 20 vs up to
its discharge at the level of the basic signal.
The described process provides quick-acting spark
extinguishing (spark current does not last more than 0.01
seconds) and entirely precludes a lengthy series of spark
discharges in the precipitator. Thus, coronal discharge sta-
bility is raised and erosion of precipitator electrodes is
lessened to a considerable degree.
The prevention of developing half-wave conditions is
done by correction of the value of the regulator-produced thy-
ristor control angle.
The basic condition for occurence of half-wave con-
ditions, as shown above, is the exceeding by the angle of
current constriction a3 of the angle of thyristor connection
oB, given by the regulator.
123
-------
The feedback signal former (FSF) fixes in each half-cycle
of circuit voltage the closing of current constriction accordina
to the moment of anode voltage restoration in the thyristor unit
and generates the signal, equal according to its value to the cur-
rent maximum allowable thyristor control angle (fig.l).
This signal is compared in the angle regulation
corrector (ARC) with the signal produced by the basic regulator
circuit. The least of the angles, compared in the ARC, enters
the thyristor impulse regulation former.
In this way, during occasional current constriction,
the thyristor inclusion angle increases according to the degree
of this constriction and development of half-wave conditions
does not occur.
Unit protection from possible current overloads is
provided by automatic limitation of the working current's
average value.
• *
A signal is continuously formed in the current limi-
ter signal former (CLSF, fig.l), proportionate to scale set by
the maximum current power control (CFC) to the average pre-
cipitator current. The value of this signal is compared
with some fixed level laid in the regulator. During a current
overload of the unit, the average current signal exceeds the
indicated level and at the signal regulator former (BSRF) input
a negative feedback siemal will appear. The basic regulating
circuit, actina to extinouish this signal, lowers the preci-
oitator's current to the qiven value.
In this manner, the current limiting circuit, along
with elements of the basic regulating circuit form a static sta-
bilizor for current of one-way action. Acceptable accuracy in
124
-------
supporting the set current value is provided by high coefficients of
amplification of the regulating tract and of the thvristor control
unit.
Seiect.ion of direct short circuit conditions in ±he
precipitator for purposes of their elimination, is devised on the
principle of current protection with voltometric blocking. In
the sensory defense unit (SDU), the device for defense against
short circuits is continuously compared with precipitator average
current values and voltage. If due to a lack of voltage the
current value exceeds the sensitivity threshhold, the device
switches on an element independent of time lag (not shown in
fig.l) and also switches on the short circuit current limiting
unit (SCL). The latter, acting through the transient condition
signal former, instantaneously limits the power thyristors con-
trol angle to a value'*: providing short circuit average current
(nearly 10% of the unit's nominal" current) defense sensitivity minimally
»
necessary. According to the time lag lapse, the unit cuts off.
The automatic regulator is a contactless device using
semiconductors. This established its high working reliability
and durability.
125
-------
FLY ASH DIELECTRIC PROPERTIES
AND CRITICAL CURRENT DENSITY
by
John P. Gooch
Jack R. McDonald
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35205
and
Leslie E. Sparks
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
(This document has been reviewed and approved for
publication by the U. S. EPA.)
126
-------
1. Introduction
A critical factor in the design of an electrostatic precipitatou for a given
application is the prediction of the operating values of applied voltage and current
density which can be achieved under the design conditions. For example, if the elec-
trical resistivity of the collected dust and the particulate space charge are not
limiting factors, it is expected that an average current density of ^ 50 to 60 nA/cm2
at average applied voltages of -\> 45 kV will be achieved with conventional parallel
plate electrode geometry. The practical significance of such operating parameters
is that collection efficiencies exceeding 99.5% for coal fly ash can be obtained with
a specific collecting area of 50 m2/(m3/sec). In contrast, if the current density
and voltage are limited by the dust properties to 5 nA/cm2 and 35 kV, specific col-
lecting areas exceeding 100 m2/(m3/sec) are required to achieve 99.5% collection
efficiency with all other conditions equal. Empirical methods are generally used
to estimate electrical operating points from properties of the collected dust layer.
The precision and applicability of these methods require improvement in view of the
high level of performance needed for most electrostatic precipitatot applications.
As a result of the importance of understanding the processes involved in pre-
dicting electrical operating conditions, the Environmental Protection Agency is spon-
soring research concerning the physical processes by which charge transport occurs
across the interelectrode space and through the dust layer on the collection elec-
trodes. The overall objective of this research effort is the development of a meth-
odology which will allow a more precise prediction of the optimum electrical operating
point for a specific application given the following: (a) the properties of the fuel
undergoing combustion (in the case of coal) or the properties of the particulate under-
going collection, (b) the temperature, pressure and composition of the gas in the
precipitator, (c) the electrode geometry, and (d) the type of electrical energization.
127
-------
2. Description of Field Measurements
A study of the parameters influencing operating values of secondary voltage and
current density is in progress at Southern Research Institute. The study has been
initiated with a measurement program conducted at several coal-fired power boilers.
Measurements which were performed included the following:
e Secondary voltage-current data were obtained from the precipitator power sup-
plies under normal operating conditions.
• In-situ determinations of ash resistivity and breakdown strength were conducted
with a point-plane probe.
• In-situ determinations of particle size distribution were performed with in-
ertial impactors at the precipitator inlets.
• Effective flue gas ion mobilities were determined with a wire-cylinder probe.
• Chemical analyses were performed on collected samples of coal, fly ashf and
flue gas.
The particle size distribution data and the effective ion mobilities are used
in the theoretical calculation of voltage-current relationships and electric field
profiles between electrodes with a mathematical procedure described by McDonald, et
al1 and discussed elsewhere.2 This paper is limited to an examination of the rela-
tionship between fly ash dielectric properties and apparent values of critical current
density observed for full-scale precipitator transformer rectifier sets.
Table 1 contains data obtained at precipitator installations collecting fly ash
with high to moderate electrical resistivity. Comparisons are made between apparent
and measured dust breakdown strengths. The "selected" operating points were subjec-
tively chosen based on an indication of incipient electrical breakdown in the dust
layer. The selection is based on either an apparent deviation from the expected form
of the voltage-current relationship in which
128
-------
j = k(v-v )n (1)
s
where: j = current density, amps/m
k and n = constants
V = corona starting potential, volts
s
V = potential, volts,
or an obvious onset of sparking. The actual observed operating points were not chosen
because of the complications introduced by the response of various automatic control
systems to sparking or back corona.
The "apparent" -breakdown strength of the dust layer is calculated by using the
relationship
EB~ javg p (2)
where: E• •» = apparent breakdown strength, volts/m
o
j = average current density, amps/m2, at selected operating point
p - measured resistivity, ohm-m
The measured breakdown strengths given in Table 1 were obtained from a point-plane
device In '.:ha Ilue gas at a voltage immediately prior to dielectric failure. The
estimated operating points were obtained from the measured values of in-situ dust
resistivity and the assumed breakdown strengths of 1 and 5 kv/cm suggested by Hall.3
Similar data are given in Table 2 for two precipitators collecting low resis-
tivity fly ash. Figures 1 and 2 display the voltage-current density relationships
for Plants A and F representing the high and low resistivity cases, respectively.
The secondary voltages were obtained with calibrated voltage divider assemblies which
were attached to the high voltage output of the transformer rectifier sets under study.
3. Discussion
It is generally recognized that factors other than dust dielectric properties
can influence the allowable electrical operating parameters in full scale precipitators.
Specifically, electrode design, electrode alignment, voltage waveforms and control
129
-------
system responr.e, collection area per transformer-rectifier sot, flue gas composition,
temperature and pressure, and electrode cleaning system effectiveness are considera-
tions which can significantly alter precipitator electrical conditions for a given
application. However, the comparisons of measured and apparent breakdown strengths
given in Tables 1 and 2 are of interest in that the deficiencies of the existing
empirical approach and the complexity of the task of predicting operating points are
illustrated.
Plants A and B are typical of large specific collecting area units which have
been constructed to collect high resistivity fly ash. The ratio of measured to ap-
parent E values are within-the range which can be expected due to dust layer non
B
uniformity, and the useful input power is clearly limited by the fly ash dielectric
properties. The increase in apparent E •* f rom inlet to outlet is hypothesized to
B
result at least in part from the variation of fly ash breakdown strength with layer
t-hickness. It has been observed in both field and laboratory environments that the
breakdown strength of fly ash decreases with increasing layer thickness. Therefore,
the thinner dust layers which usually are present in outlet fields may. be able to
tolerate higher average electric fields. Another possible contributing factor is
the variation of resistivity due to particle size distribution variations which occur
from inlet to outlet.
Plant C exhibits an unusual variation in inlet to outlet operating points, and
electrode cleaning system ineffectiveness is a probable cause. The ratio of actual
to apparent EQ decreases to less than unity for the outlet fields, which suggests
that either; (1) the point-plane probe collected a dust sample not representative
of the collected dust layer in fields 7 and 8, or (2) a back corona discharge exists
which is not detectable from the V-I curves.
The low values of current density achieved at Plant D are difficult to explain
in terms of the measured properties of the fly ash. However, the sparking threshold
is clearly influenced by dust or flue gas properties, as the data in Figure 3 illu-
strate. Thccc voltage-current curves were obtained with and without SO3 conditioning
130
-------
for the indicated TR Bet. Plant D has i\ relatively ineffective rapping yyntcm, and
it is probable that residual dust layers which were present on the collecting elec-
trodes possessed a higher resistivity than that measured with the point-plane probe
during the test series. If a residual layer with a resistivity of 2.7xl012 ohm-cm
were present, the apparent E* values would fall within the expected range of 5 to
13
10 kV/cm. These circumstances illustrate both the difficulty of estimating the limita-
tion imposed by high resistivity and the value of flue gas conditioning as a means
of alleviating the problem. Note that the assumption of 1 to 5 kV/cm as an apparent
breakdown strength interval failed to bracket the selected operating points for the
four plants described in Table 1.
Plants E and F represent precipitator installations which are not performance-
limited by dust resistivity. The low apparent values of E-suggest that the sparkover
voltages are determined by factors other than dielectric failure in collected dust
layers. Although the current densities achieved at Plant E were relatively low, the
collection efficiency was relatively good because of the high operating voltages.
For both plants E and P.- it is apparent that electric field intensity in the dust
•
deposits does not provide a basis for estimating operating points. For these cir-
cumstances, however, high levels of performance can be achieved without excessive
plate area requirements, and the accurate prediction of achievable current density
is less critical than for the plants listed in Table 1.
The variations in measured and apparent breakdown strength in Tables 1 and 2
illustrate the need for a thorough understanding of the electrical properties of col-
lected dust layers. McDonald and Mosely"* have proposed an approach for describing
the electrical characteristics of such layers based on coupling the current transport
equation and Poisson's Equation. In this method, the electrical breakdown of the
dust layer is attributed to enhanced values of the local electric field (defined as
the electric field at a point between particles) due to the combined effects of polari-
zation of the particles in the dielectric layer and the contribution to the electric
131
-------
field from the r.pace charge distribution in the aolid phase of the dust layer. The
average value of the electric field for which breakdown will occur can be computed
as a function of dielectric constant of the particles by specifying a breakdown
strength for the trapped gas. Table 3 shows the results of calculations for several
cases of interest. This approach implies that the operating current density limit
could be predicted from measurements of resistivity and dielectric constant in an
appropriate environment. Other factors may require consideration, however, such as
charge carrier density gradients which may develop in collected dust layers over
extended time periods.
In order to construct appropriate charge transport models for collected dust
layers, certain experimental studies need to be performed to determine appropriate
values of the parameters and boundary conditions that must be utilized in modeling
the electrical phenomena. These studies include measurements to determine the fol-
lowing :
• whether electrical breakdown is due to gas breakdown or particle breakdown
• el'ci.i.ii; potential distribution and concentration profiles of charge carrying
species in the dust layer
• values for transference numbers, activation energies, conductivities, diffusion
coefficients, and mobilities of the different charge carrying species
• values of the dielectric constants for different dusts in simulated flue gas
environments.
4. Summary
The actual and apparent breakdown strengths of electrostatically collected dust
layers exhibit large variations which are not adequately represented by existing con-
cepts. Physical models of the charge transport process across the interelectrode
132
-------
space and through the collected dust layers have the potential o£ providing a method-
ology for accurate prediction of optimum electrical operating conditions utilizing
dust and flue gas properties and the design parameters of the electrostatic precipitator,
Fundamental studies of dust layer properties are continuing with the objective of
developing the required predictive methods.
References
1. McDonald, J.R., W.B. Smith, H.W. Spencer, and L.E. Sparks. A Mathematical Model
for Calculating Electrical Conditions in >?ire-Duct Electrostatic Precipitation
Devices. J. Appl. Phys., 48(6):2231-2246 (1977).
2. McDonald, J.R. A Mathematical Model of Electrostatic Precipitation: Revision
1. Volume 1, Modeling and Programming. EPA-600/7-78-llla, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, 1978.
3. Hall, Herbert J., "Trends in Electrical Energization of Electrostatic Precipi-
tator. Proceedings of the Electrostatic Precipitator Symposium, Feb 23-25, 1971,
Birmingham, Ala.
4. McDonald, J.R., R.B. Mosley, and L.E. Sparks. An Approach for Describing Elec-
trical Characteristics of Precipitated Dust Layers. Submitted for publication
in JAPCA, March, 1979.
133
-------
20.4
17.0
13.6
M
Z
til
Q
I-
UJ
oc
oc
D
u
10.2
6.8
3.4
10
20
30
VOLTAGE, kV
40
50
CO
Figure 1. Voltage-current curves for plant A.
134
-------
47.G
40.8
34.0
CM
,0
27.2
V)
Z
IU
Q
cc 20.4
DC
o
13.6
6.8
0.0
10
20
30
VOLTAGE, kV
40
50
. 60
Figure 2. Voltage-Current Curves for Plant F.
135
-------
20.4
17.0
13.6
CM
E
,0
=
H
z 10.2
01
O
1-
2
uu
1
8 6.8
3.4
0
1
•
V^MB
— ^^^^
WITHOUT S03
^H^HV
I
1
.1
1
1
WITH SO3
1
1
1
I
f
1
1
r
i
\
1 __
i
/
i
1
1
I
t
I
1
10
20
30
40
50
60
VOLTAGE, kV
Figure 3. Voltage - current curves for Plant Dt Fourth Field
136
-------
Plan
A
Field ir.
Direction of
Cas Flow
Selected Operating
Point
kV nA/cm2
37.0
36.0
33.5
32.0
34.0
36.0
36-0
33.0
33.0
36.0
28.0
29.0
32.0
32.0
27.0
25.0
2.0
2.0
2.0
3.0
4.0
5.3
5.5
10.0
12.0
2.5
44.0
48.0
2.0
2.0
2.0
2.0
Select-.>• operating Points and Apparcn: ..kdown Strengths
High to Moderate Resistivity
Apparent Measured
Electrode
Geometry
30.48 cm plate spacing,
discharge electrodes
ere 4.26 mm square,
fluted & twisted held
in a rigid mast.
m
25 cm plate spacing,
discharge electrodes
are spiraled, 2.5 mm
in diameter & held in
a rigid frame.
22.86 cm plate spacing,
discharge electrodes
are round(2.68 mm
diameter) & weighted.
22.86 cm plate spacing,
discharge electrodes are
round(2.68 mm diameter)
and weighted.
Temperature
°C
138
135
144
164
Resistivity Area/TR set
ohm-era m2
3x10 l2 2675.5a
2675.5°
2675. 5a
2676.5a
2x10" 3518.6
3518.6
3518.6
3518.6
3518.6
SxlO11 1424.2
1424.2
1424.2
3xl011 961.5°
961. 5a
961. 5a
961. 5a
V
kV/cm
5.8
5.8
5.8
8.7
7.2
9.5
9.9
18.0
22.0
1.25
22.0
24.0
0.54
0.54
0.54
0.54
Eg Eg Measured
kV/cm tB Apparent
27.2 4.7
4.7
4.7
3.1
26.6 3.7
2.8
2.7
1.5
1.2
10.5 8.4
0.48
0.48
13.8 25.5
25.5
25.5
25.5
Estimated Operating
Point nA/cE*
IkV/cm 5kV/c3
0.33
0.50
2.0
3.7
1.7
2.5
10.0
18.5
UJ
a. collection area per bushing
-------
Table 2
itlng Prints and Apparent Breakdcwi
Low Resistivity
Strength
Plane
.- 1
r t
Ca
old in
rtc'ion of
fc r lo«'
1
2
3
1
2
3
4
Selected Operating Apparel
Point
kV
46.0
45.0
45.0
55.6
51.7
48.9
48.5
r.A/c=.:
8.0
8.0
24.5
20.0
27.3
36.0
46.0
Electrode
Gcoir.etry
27.9 cm plate spacing,
4.2 ran square twisted
Temperature
"C
153
Resistivity
ohm- cm
2x10"
discharge electrodes held
in a rigid mast.
25.4 cm plat- spacing,
discharge electrodes are
rigid tubular barbed
electrodes
160
2x10 10
Area/TR set
m2
2458. la
2458. la
2458. la
1475. 2a
1475.23
737. 6a
737. 6a
V
kV/cra
0.16
0.16
0.54
0.40
0.55
0.72
0.92
kV/cra
14.0
17.0
Eg Measured
Eg Apparent
88.0
88.0
26.0
42.0
31.0
24.0
18.0
Esticated Operating
Point nA/c-.2
IkV/ca 5XV/K3
50.0
50.0
250.0
250.0
a. collection area per bushing
00
00
-------
Table 3. Average electric field at breakdown for
various dielectric constants/ temperatures/ and pressures
U)
VO
E_= 30 kV/cm at
r>
293°K, 760 nun Hg
D = 20.8 kV/cm at
D
422°K, 760 mm Hg
ED = 11.4 kV/cm at
ts
616°K, 608 ITCT Hg
K
1
2
3
4
5
6
7
8
9
10
E /E EAVe f°r
loc Ave Breakdown (kV/cm)
1.50
3.38
5.25
7.13
9.00
10.88
12.75
14.63
16.50
18.38
20.00
8.88
5.71
4.20
3.33
2.76
2.35
2.05
1.82
1.63
E /E EAve f°r
loc' Ave Breakdown (kV/cm)
1.50
3.38
5.25
7.13
9.00
10.88
12.75
14.63
16.50
18.38
13.90
6.15
3.96
2.92
2.31
1.91
1.63
1.42
1.26
1.13
E /E EAVS f°r
loc Ave Breakdown (kV/cm)
1.50
3.38
5.25
7.13
9.00
10.88
12.75
14.63
16.50
18.38
7.60
3.37
2.17
1.60
1.27
1.05
0.89
0.78
0.69
0.62
-------
A.I. 7avialov, V.B. Meshcheriakov
RESEARCH, DEVELOPMENT AND INTRODUCTION OF ADVANCED
PRECIPITATING ELECTRODE DESIGNS OF INDUSTRIAL PURPOSE ELECTRO-
STATIC PRECIPITATORS
In connection with the growth in volumes of dusts removed
from stack gases with the aid of industrial ESP's, special atten-
tion is being paid to the problem of the reliability and durability
of their mechanical components. In the course of the last ten years,
NIIOGAZ's Semibratovsk branch and the Moscow Institute of Trans-
port Engineeers have conducted complex experimental-theoretical
research into the dynamics of precipitating electrodes of industrial
purpose ESP's.~Curing the course of research, a science-based ex-
perimental prccec*ure-.-£L) -wasdeveloped along with methodology for the
computational evaluation of precipitator electrode rapping effi-
ciency where the rapping was carried out by the impact method. (2) .
As a result of research, reasons for the unsatisfactory degree
of cleaning the electrodes from accumulated fly-ash were established,
as well as the causes for the breakdown of rapping band and rap- •
ping mechanisms. The basic research results are published in vari-
ous articles and have also been presented in symposiums (3,4).
Related to the number of basic factors causing unsatisfac-
tory electrode performance are rapping blow energy losses at the
junctures of the electrodes with the rapping band and the dampening
of the disturbance with great element length. Density of contact in
the areas of energy transmission plays a large role. Specially set
experiments have shown that the presence of a cap to the magnitude
of ^.Imp may lower the magnitude of the transmitted impulse bv two
140
-------
or' more times. With fly-ash falling into this cap, the lowering of
the magnitude of impulse becomes more substantial (5).
New design variants are being developed for improving the
work of precipitating electrodes. The main direction of this deve-
lopment is increase of element width and raising the contact reli-
ability of the elements with the rapping band, widely distributed
electrode designs may be operatively redesigned with the aim of in-
creasing their rapping efficiency and raisng their operating reli-
ability and endurance. Design variants developed in NIIOGAZ's Semi-
bra tovsk branch may be used for this.
0.35m~wide precipitating elements a're doubled and hung excen-
trically;; This measures increases the excentricity, lowers the num-
ber of places of energy transmission and increases the contact relia-
bility.
The precipitating elements in the electrode are combined into
two groups (6). The first groups elements are hung excentrically for
takinc the blow from the raoo: cr mechanism's hammer. Elements of the
\
second group are also hung excentrically but are drawn against the
rapping band for taking the rapping blow against the supports. Such
design (see fig.2) gives an opportunity for considerably raising the
average level of maximum accelerations on the electrode. The element1
solid contact with the rapping band is also provided by the precipi-
tating electrode with vertical alignment of the elements to the rap-
ping band. A precipitator element variant is proposed with vertical
element support against, the rapping band and with a mechanism for
element rapping (7) in order to eliminate differences in accelera-
tion at the first and last elements fron the place of ir—sact (•fue
to non-uniform pulse distribution). Such design provides uniform
distribution of the shock impulse according to the electrode's
141
-------
height and width and also a sufficient degree of cleaning with
a smaller number of impact loads. The low level of vibration of
neighboring unrappable elements during the drop from a specified
height of one (in this case, the fourth) element (see fig.3) and
the opportunity for regulated rapping speed make it possible to
decrease re-entrainment, but a decrease of the number of impacts
make it possible to increase electrode reliability.
All of the suggested design variations on precipitating elec-i-
trodes passed".a stage of stand and industrial testing and accounted
well for themselves. Thus, for example, testing of advanced designs
(fig.l and 2) at the Troitskaya State Regional Electric Power Plant
have shown that these designs possess great advantages in comparison
with earlier-used designs. .
At the Sredneuralsk copper-smelting works, nine-month testing
of the precipitating electrode with vertical element support and
element-by-elanent rapping system was conducted (7) . These tests showed
that a high degree of rapping efficiency (9Cr-95%) is obtained
through one-two "raise and drop" cycles.
Precipitating electrodes with two groups of elements (fig.2)
are recommended for introduction in serially produced and redesigned
precipitators of various types.
Precipitating electrodes with vertical element-try-element rapping
are advisable for use in precipitators removing strongly-aglutina-
ting dusts (cement dust, etc.) and for high-resistivity dusts.
142
-------
BIBLIOGRAPHY
1. Zavi-alov , A.I., Methods and Research in Dynamics of Precipitating
Electrodes, "Tsement", 1975, No.3.
2. Meshcheriakov, V.B., Numerical Evaluation of Precipitatinc Elec-
trode Rapping Efficiency in Industrial Precipitators, in the collec-
tion "Industrial and Sanitary Gas Cleaning", TsINTIKhlMNEFTEMASh,
1977, No.l.
3. Meshcheriakov,, v.B. et al, Results of Theoretical and Experimental
Studv of Dynamics of Precipitator Electrodes : Report Abstracts of the
Symposium "Electrical Gas Cleaning Methods", Moscow, 1973.
4. Meshcheriakov, V.B., Zavialov, A.I., Dynamics of Precipitator Elec-
t
trodes and Predicting Effectiveness of their Regeneration ; Report to
the Second Soviet-American Symposium on Particulate Control, Washing-
u-n, 1977.
5. Zavialov, A.I., Meshcheriakov, V.B., Research in the Dynamics of
Advanced Precipitator Electrod-..3, in the collection "Industrial and
Sanitary Gas Cleaning", TsINTIKhlMNEFTEMASh, 1977, No.6.
6. Zavialov*, A.I., Sobchuk, S.I., Smirnov, L.P., Nagornyi, V.V.,
The Precipitator Electrode, author's certificate, patent no. 472690,
bulletin "Discoveries, Inventions, Industrial Models, Trade Marks",
1975, No.21.
~7. Zavialov/ A.I., Nagornyi, V.V., Panasenko, V.I., Sobchuk, S.I.,
The Electrostatic Precipitator, author's certificate, patent no. 523712
letin "Discoveries, Inventions, Industrial Models, Trademarks",
1P75, No.29.
143
-------
-ILJJ
JtJJJl
JU s
f j g. .1.
Precipitating electrode
with paired elements.
Mff.
.fig. 2
Precipitating electrode
with two groups of ele-
to
30
to
O £ {f f~ number of
element.
-fig.3. '
Acceleration distribution in pre
cipitating electrode with- elemen
t>y-*?elefnent rappincr-
-------
PERFORMANCE OF A HOT-SIDE ELECTROSTATIC PRECIPITATOR
G.H. Marchant, Jr.
John P. Gooch
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35205
Leslie E. Sparks
Industrial Environmental Research Laboratory
> Environmental Protection Agency
Research Triangle Park, N.C. 27711
(This document has been reviewed aid approved for
publication "by the U. S. EPA.)
145
-------
INTRODUCTION
The performance of a hot-side electrostatic precipitator installed on
a large power plant boiler was investigated. The tests were made on Unit 3
of the Navajo Generating Station of the Salt River Project in Arizona. Over-
all and fractional collection efficiencies were measured across one chamber
of the unit, and fractional collection efficiency was measured for the entire
precipitator. Voltage-current characteristics of the power supplies were
measured. An engineering analysis for the installation included estimates
of the specific collecting areas required for improving performance of the
hot-side precipitator and for replacing it with a cold-side precipitator.
The electrostatic precipitator has two levels (Figure 1), each with eight
isolable chambers (Figure 2), and was designed to operate with a total volume
flow rate of 1860 m3/sec at 350°C with 99.5% collection efficiency. Each
chamber contains 6 electrical fields in the direction of gas flow and 35 gas
passages spaced 22.9 cm apart. A total of 48 transformer-rectifier sets power
parallel fields in parallel chambers. The collection electrodes are 1.83
m in depth and 9.14 m high. The discharge electrodes have diameters of 2.68
mm and the average spacing between each wire is 22.9 cm. Each precipitator
has a total collecting area of 112,372 m2, which results in a design specific
collection area of 60.4 m2/(m3/sec).
TEST RESULTS
The overall fly ash collection efficiency (by mass) of one isolated
chamber (Chamber 8) of the electrostatic precipitator was found to be 99.22%,
based on mass train measurements at the inlet and outlet of the chamber.
Test data are shown in Table 1. In these tests, average secondary voltage
was 22 kV, average secondary current density 40 nA/cm2, specific collection
146
-------
area 52.6 m2/(m3/sec), and fly ash resistivity (determined in situ) SxlO9
ohm-cm at 350°C. Plant operating problems prevented the measurement of col-
lection efficiency for the entire precipitator by a mass train. However,
a limited number of measurements were made by impactor traverses in the stack
and at the precipitator inlet, producing the results in Table 2. The stack
data indicated a mass emission rate of 31 ng/J, of which 9 ng/J consisted
of particles with diameters < 2 ym. The gas volume flow rate in the stack
was consistent with that at the outlet of Chamber 8. Oxygen contents, de-
termined simultaneously at the inlet and in the stack, indicated that air
in-leakage across the precipitator and air preheater was about 11%.
Figure 3 compares the particle-size distributions measured with impactors
at the inlet to the electrostatic precipitator and to Chamber 8. The bimodal
particle-size distributions are typical of those produced by pulverized coal-
fired boilers, with one mode a maximum at about 2 ym, and the other at a diam-
eter > 10 ym. The mass median diameter of the fly ash, based on impactor
data, is about 13 ym. If it is assumed that the difference in mass loadings
measured by the impactor and the mass train sampling system is due to inef-
ficient capture of particles > 20 ym by the impactor, the corrected mass
median diameter is 16 ym.
Concern about the effects of soot blowing on the particulate loading
of the flue gas entering a hot-side electrostatic precipitator led to com-
parative tests which showed that there were significant (at the 90% confi-
dence level) increases in loading at the inlet of Chamber 8 during soot blow-
ing, as measured with a mass train. However, no significant differences were
observed at the chamber outlet as the result of soot blowing. Figure 4 shows
147
-------
the particle size distributions obtained with impactors at the inlet to Cham-
ber 8 with and without soot blowing. Most of the increase in loading during
soot blowing was due to particles > 8 ym.
Figure 5 compares the differential particle-size distributions in the
stack and at the outlet of Chamber 8. The stack data exhibit substantially
higher loadings of particles with diameters from 0.8 to > 10 um. These dif-
ferences are also reflected in the fractional collection efficiency results
shown in Figure 6. The fractional collection efficiency curve showed a mini-
mum value of 92% at a particle diameter of about 0.50 um. The disagreement
between the impactor and ultrafine results in the 0.5 pm region is partly
due to experimental error.
Values of electrical resistivity of the fly ash measured at 350°C in
situ and in the laboratory are in reasonable agreement with those measured
for other fly ashes of similar chemical composition. The resistivities re-
mained relatively constant during the test program.
Coal and ash analyses indicated constant quality of the coal supply.
Attempts to measure sulfur dioxide in the flue gas indicated that concentra-
tions were never above the detection limit of ^ 0.5 ppm at the inlet or either
of the two outlet sampling locations.
The voltage waveforms and secondary voltage-current relationships ob-
served during the tests appeared to show the effects of back corona. Figure 7
presents voltage-current curves obtained for Chambers 7 and 8 (secondary volt-
ages were determined from voltage divider readings). The operating voltages
were lower than anticipated, especially for the outlet fields. Several ob-
servations indicated that fly ash deposits on the collection electrodes af-
fected the functional relationship between applied voltage and corona current:
148
-------
The V-I curves do not respond to changes in discharge electrode diameter to
the extent theoretically predicted (Figure 8). The curves from the outlet
field show some hysteresis. The shapes of the curves are affected by elec-
trode cleaning (Figure 9). Voltage-current waveforms (Figure 10) suggest
a back corona discharge at high current levels.
Although the precipitator has not maintained design efficiency, it has
operated reliably. The most significant maintenance problems have been air
infiltration and ash build-up in hoppers.
ENGINEERING ANALYSIS
The estimated cost of the electrostatic precipitator installation is
given in Table 3. These costs were calculated from the 1973 and later con-
tract costs, plus a 20% distributable cost, plus 9% of the contract and dis-
tributable cost for engineering costs, and escalation of each cost element
to 1977 at 7.5%/year.
Table 4 shows the annual operating cost, including maintenance cost as-
signed to the electrostatic precipitator.
The performance tests indicated that the two principal causes of poor
performance of the precipitator were the relatively low operating voltage
and the relatively low value of specific collecting area. To indicate changes
in precipitator design that should overcome these deficiencies, values of
design parameters listed in Table 5 were calculated for an improved hot-side
unit. The basic geometrical configurations of the existing unit were arbitra-
rily retained. A mathematical model was used to extrapolate the apparent
collection efficiency of the existing unit (98.6% at a specific collecting
area of 53.15 n»2/(m3/sec)) to give a specific collecting area of 78.74 m2/(m3/sec)
149
-------
required for 99.5% collection efficiency. To reliably attain a 99.5% collec-
tion efficiency in the presence of the unfavorable electrical operating condi-
tions requires a design specific collecting area of 93.90 mz/(m3/sec), which
provides a safety margin of about 20%. The design parameters for the improved
precipitator are listed in Table 5. The new design has an increase in collec-
tion surface of 83% over the existing unit. The estimated capital cost for
this design is $60,440,000, or $75.5/kW, based on 800 MW generating capacity.
No retrofit charges are included in the estimate.
Table 5 also includes data on two cold-side designs for 99.5% minimum
collection efficiency, for which the estimated capital costs are $52.4/kW
and $65.1/kW, for collecting fly ash with electrical resistivities of 9 x 1010
and 7 x 10 ohm-cm, respectively.
150
-------
TABLE 1
AVERAGE IHLBT AND OUTLET PARAMETERS. ISOLATED CHAMBER I
Inl«t Outlet Collection Efficiency.
*C 3*1 330
Sl.S
Temperature
CM volue* riov Rat*, dm'/MC 44.4
NIK Loading, «/dn*
lavactor
H>» Train
Miaeber of Taata
nqwctoc
Naea Train
5.19
*.77
0.0314
0.0529
32
II
»9.26
»».22
TAILS 2
AVERAGE IMLR AMD- STAC! PARAMETERS
Inlet Stack
•ratura, *C 3
-------
oa ti t. ox
111
{{{(it
uo M h. e x
Figure 1. Ductwork and precipitator arrangement
for Navajo Station, Unit 3.
Figure 2. Precipitator chamber arrangement.
1C*;
it*.
103:;
101 ;
• CMAHHII t IHLJT
O MAIM INLCT
irfM— nl | I
ID"1 1CP 101 1C?
PARTICLE DIAACTER
Flgur* 3. Av«r«g« inlet differential >i
-------
PErCTRATICN-EFFICIEJOr
•HE 1*477
10^
10°-
UCT*
p
o CHUMEK IOUTUT. 7/1»2»TT
icr1 10° lo1 10s
PARTICLE DIAMETER (MICROMETERS)
Figure 5. Outlet differential size distribution.
99.33-
39.35:
39.3-
>. 33.8-
Q 39.5:
E -1
2 3B"
Q
K •
35-
90-
fesyeaj ,
BO-
K
. '
_ . * ULTtUFINE
"^ * • 1WACTOR [
A ULTNA>*INE
| CHAMMm, ..MPACTW i
•% **
' ' * -^ '
* / • 1 "-"
'. ' A * n, . * ^
1 •* *
• , i i
nfU
2^s2{R» • -
"
r* 10'1 10° lo1 K
•O.Ol
•O.O5
•O.I
•o.e g
0.5 ^
1 I
.5 s
Q
Q
•5
•10
an
/°
Figure 6. Fractional efficiency for Chaaber 8
and total ESP.
S£J
S9.I
EU
eu
«u
4M
JU
IM
W
mo
1&0
tfto
u
o
«^
—
o
t
O |
1
O A
O A
A
O
•
O a
O A
0 A
0 a
,c
ky
r x
^
dfi«^
00
oo
QD
00
OO
CD
CD
CD
0 A IS
' 1 • 1 -«£>-
O H FIELD
V a FIELD
O F FIELD
O E FIELD
A D FIELD
O C FIELD
• OFERATINO FONT
E
^r
'•^"H
v 0
v 0
O
• o
v O
i.i.
O C FIELD Ch 74
• O C FIELD CH 1-1 LARGE WIP.EI
METER READING!
FROM V21/77
Figure 7. Voltage-current curves for Chamber 7 and I,
July 12-13, 1977.
Figure B. Theoretical end experimental voltage-current
relationships for various wire diameters.
153
-------
1
5
»
fc
«
Q
C
u
C
3
Ml
76
70
65
60
65
60
46
40
36
30
26
20
16
10
5
I.I 1 1 1
A C FIELD. Ch 7-8 8/21/77 (SRI DATA)
- O C FIELD. Ch 6-6 MAY 1977 (NO CLEANING))
Q C FIELD. Ch 7-8 MAY 1977 NEW WIRES. >
CLEAN PLATES J W'P' DATA
(ALL METER READINGS - UNCOHRECTED) Q
-
O
A
0
"* ™
- • —
* o Q
— * _
A
_ «
A
a
- -
D 0
-
A o
~ 0
-
-Do -
A
0
1 1 1 1 1
18 20 22 24 26 28 30
*--
s..
••>
^*
•**.
»•—
***•
•N.
*--
*\
\
/•
/
•v
V
s~
/
V
*\
X
-
s
f
\ ,
\f
y
\
\
(
1
\
\i
f
/
/
\
\
f
1
;
/
-i
-2
-3
-4
0
-1
-2
-3
0
-1
-2
_1
-2 CORNA START
KNEE OF CURVE
MAXIMUM
-2 AUTOMATIC
OPERATING POINT
0 6 10 15 20 26 30 35 40 45
VOLTAGE. kV
Figure 9. Voltage-current relationships for C Fields,
Ch. 7 4 8 and Ch. 5 S 6.
Figure 10. Voltage waveforms for C Field,
Chambers 7 and 8.
-------
3.K. Kurbatskii, V.B. Keshcheriakov, E.K. I:, seto*,
ST'JTY OF THE DYNAMICS AND ST5TSGTH CF COSOKAL
SLSCTRODEJ1 IN ELECTROSTATIC :-ri:iPITATORS
Existing systems of coronal electrodes in industrial
electrostatic precipitators do not to the full degree meet the
racuirements presented them in their section where settling
fly-ash is removed, and in their reliability and durability.
This report elucidates study results from the work of
the two most widely distributed in practice coronal elec-
trodes: those with elements hung from support beams and those
with frame-construction electrodes. A complex experimental-theore-
tical study was conducted for the obtaining.reliable reults
j during investigation of the electrodes' dynamic work.
In the course of theoretical study, integral and differential
equations were compiled describing processes of formation of
contact forces and the transmission of disturbances in the
system's unitr (1). Approximate methods fer calculating maximal
accelerations and stresses in the units of coronal electrodes
under pulsed loads have been developed. In these calcu-
lations, all physical-gecmetric electrode characteristics and con-
ditions of the formation of inlet -ffect are considered. Means of
rigid and hanged hanging of the elements to the support beam were
separately ^considered.
j Contact force during transverse impact to the support beam is
determined by the following non-linear integral equation:
(I)
where: p(t) - sought contact force f - density of the material
t - time F. - transverse area c-! beam
section
• - initial impact velocity _ propagation of the dis-
m - mass of the striking body ^ waves in the besr.
KB - coefficient, related *•
physical-gec-ez_-ic cr
terisrics of r!-«> «--
155
-------
As a result of the impact to the beam, flexible elements hung to
•"zne bssr. are subject to a loss of stability in a snail portion cf
ier.rth -y the value { with a maximal deviation of Uf (these para-
meters depend upon intensity of the iarast ar.d upon the character
cf attachment of the el r.ments to the beam). The disturbance taJ-.ar. fai-
ths elersnt is propagated along its length and as a result all points of
the element are subject to the acceleration.
acceleration at the points of the coronal elements
i durinc rierid attachment to the support beam is determined by the
I ar^rcximate formula:
i
I a-
\ where: ?. , t, - maximal value of contact force and duration of contact
< c, c2 - velocity of propaqaticn of longitudinal and transverse
i waves in flexible coronal elements.
t
\ y,Y\. - coefficients of internal friction in support beam
? material and in the element, determined experimer tally
| z,x - distance from the point of the impact to the point cf
| attachment of the element and coordinates cf the
\ • point 'o* the elenent.
• Maximum stress in coronal ' elements is determined accordinej to
the formula:
\ where: Q - force of element stress, F2 - area of its transverse section.
; During developement of a method of experimental szudy of the
f dynamics and durability of coronizing electrodes with flexible elements,
; specific features of the studied system were considered. The i-ipulse
! character of charging, great element flexibility and small driver, mass
; impose essential limitations on the choice of mcritors and methods fcr
| measuring dynamic characteristics.
; Inertial-type monitors based on resistor, capacity, electrc-
• magnetic and piezoelectrical converters he -a received the grea-est
distribution for measurement of the parameters of movement (2) . 7cr
': measuring accelert-ion occuring in cor~ni:ir.g electrodas, pieicelss-
156
-------
trie accelerometer-sensors which possess wide rer.ses cf r.aasurer.snt
far amplitude and frequency and small dimensions tnd nasa were user.
i • The method of fastening the sensor and its t-ss hava an effect
| en the aess-jrecsant results when the measurement sensor is fixed directly to the sur-
; face of the object. In this case, the error, caused by the rains of attacrrar-, did not
exceed 10%. For determining the effects of the monitor's mass, special
: experiments were carried out and the problem of flexible element cs-
cillaticn with concentrated mass was theoretically solved. The calcu-
I Jated dependence of measured accelerations on the mess of the ser.-
| sor end experimental data for a frequency of 2.OS kilchertz is inrro-
f dueed in fig.l. The obtained relationship allows either selection cf sen-
[ scr mass providing desired accuracy of measurement, or calculating the
| true value of acceleration with respect to the figure for measured ac-rele-
i rations with a known sensor mass. Relative measurement error of accele-
; rations as taken by the described method did not exreed 19%.
: As a result of experimental investigation of dynastic characteri-
stics, the effect of a whole series of factors on the level and cha-
t
| racter of the distribution of maximum accelerations (those normal to
I coronal electrode surfaces) were established. These factors are: method of fiadner
r
| of tiie elements to the support beam; element length- and their stretchino
' strength; mass of the rapping mechanism's hantner and initial, velocity of the djnrrt.
\ Fig. 2 introduces basic results of dynamics studies for coronal
I electrodes with hanging elements. The results are relative to cases of
i rigid element mounting to the support beams.
I The contact force from a blow to the coronal electrode unit
/
i frame (see fig. 3) is determined using a non-linear integral equation
| similar to equation 1, with the only difference being the last term in the
\ right part of the ecuation before the integral - a factor of the following type:
i . /
where 7, - area of transverse section of frame rod taking the blow.
F, - *rea of transverse secti- n of frame rod, taking the longitudinal
blow.
fs a result of the iiroact to the frame assembly, a disturbance is propagated through
all of the rods. Wave reflection and refraction occurs in the franvr
assemblies, and as a result of this, the erslitade cf the disturbance steadily
157
-------
diminishes. Tig. 2 is a diagrammatic calculation of t. .istribution
cf rr.a:-:ir.un accelerations through the frame wich regard : the tra.-.s-
mirsion cf impact energy to all coronal eler-er.ts.
r. section of the distribution of the level of maximum accelera-
tions lengthwise along rod A-A, taking the longitudinal blow, is shown
in fig. 4. Experimental results are shown for comparison. Discrepancy
cf the calculations with the experimental data is explained by the
fact that in experiments, sufficiently close contact of the coronal
elements with the frame rod A-A was not provided, in support of this,
a. dotted line in fig. 4 shows the result of calculation with the com-
plete absence of coronal electrodes.
The experimental-theoretical study cf dynamic characteristics
undertaken allows calculation of coronil electrode dust-removal
efficiency still at the electrostatic precipitator's planning stage,
it allows choosing optimal variants of their design , and also evalua-
tion of the reliability and durability of their work.
On the basis of results from the research conducted in NIlOGAZ's
SerdJbratovsk branch, improved coronal electrode design variants have been developed.
At present, these are being studied as laboratory bench-scale models.
BIBLIOGRAPHY
1. Goldsmit, V., Shock; Theory and Physical Properties of Colliding
Bodies, Moscow, Stroizdat, 1965.
2. lorish, Y.I., Vibrometrics, Moscow, Mashinostroenie, 1?63.
158
-------
. - 5 10 ....
Fig.l. Relationship of measured accelerations to mass of
of the* sensor. .
35 m
—. _ calculation
0'_ experiment Vs
v _ experiment ,- '\V2
_ experimentV.
D D DQ
rrij
Fig.2._f.elatic-shi? of mean., level of maxi:r.un accelerations
159
-------
Fig.A. Section
Of distribution
•i 1~,
Of accelera- ' "
"lions along the
line A-A.
r~ e
v.,3
C.25
rig] 37 calcuTaEea''distribution of level of accelerations along
fraiae-desrgn'coronal electrode. - . .
calculation"(elements not considered)
160
-------
PROGRESS ON ELECTROSTATIC PRECIPITATORS FOR
USE AT HIGH TEMPERATURE AND HIGH PRESSURE
by
George Rinard, Donald Rugg
Robert Gyepes, James Armstrong
Denver Research Institute
and
Dennis Drehmel
Particulate Technology Branch
Industrial Environmental Reseach Laboratory
Environmental Protection Agency
Research Triangle Park, N. C. 27711
(This document has been reviewed and approved for publication by
the U.S. EPA.) '
INTRODUCTION
The electrostatic precipitator (ESP) is one candidate apparatus for
use as the tertiary collector in a pressurized fluidized-bed combustion
(PFBC), combined-cycle power plant. This application would require
operation of the ESP at temperatures on the order of 1000°C and pressures
of 1 MPa. ESP's in this range of temperature and pressure may also find
application in coal gasification and magnetohydrodynamic (MHO) power
generation. Whili some work has been done on ESP's operating under
these conditions, the feasibility of commercial application of ESP's for
use with PFBC is yet to be demonstrated.
Presently work* is being conducted to determine the feasibility of
ESP operation under these extreme conditions. The project is to design
and build a high temperature pressure vessel and test ESP operations
under flow conditions. Many of the technical problems to be overcome
relate to the materials that can be used for extended periods of time at
the high temperature involved. Under the present schedule, initial
testing will start about December 1979.
BACKGROUND
All work to date on ESP's at high temperature and pressure has
utilized cylindrical electrode geometry. Some of the earlier work on
*This work was supported under EPA Research Grant R805939(010) through EPA's
Industrial Environmental Research Laboratory, Research Triangle Park, N.C.
161
-------
on high temperature, high pressure (HTHP) ESP's was done at General Electric
In this work the negative corona characteristics of air and methyl
chloride were determined over a temperature range of 20 to 500°C and
pressures from 100 to 500 kPa. This work concluded that, for a range of
2
current densities of 68 to 680 nA/cm , the temperature and pressure had
no effect on the corona discharge other than on the density of the gas.
Thus, the characteristics were uniquely determined by the relative
density of the gas between the electrodes.
Several years later corona characteristics were measured at Princeton
for both positive and negative corona for air and nitrogen at temperatures
2
from 65 to 825°C and with pressures from 10 to 800 kPa. At this time
it was reported that positive corona characteristics were a function of
gas density only. However negative corona characteristics were found to
depend on gas temperature as well as gas density. In addition, pressure
dependent instabilities were observed at high temperature for both
polarities.
Shale's work at the Bureau of Mines in 1963 also showed that
negative corona depended on gas temperature as well as gas density.
This work was done in a 5 cm diameter cylindrical ESP for temperatures
of 315°C to 815°C and pressures of 100 to 640 kPA, using negative discharge
corona. The results indicate stable precipitator operation for the full
range of pressure and temperature up to 730°C. Above 730°C, operation
was limited to pressures above atmospheric, because sparking occurred
before corona could be generated at lower pressures. The length of the
corona wire was 107 cm1 the 5 cm diameter pipe section was 61 cm long.
2
The current densities were in the range of approximately 3-5 nA/cm .
This is considerably higher than normally encountered in low-temperature,
atmospheric-pressure ESP's. An effective ion mobility was calculated
from the experimental data: with all else constant, it was found to
decrease with gas density, to increase with both temperature and field
strength. The higher effective mobility, at higher temperatures and
field strengths, was attributed to larger components of electron current
since the mobility of ions changes very little with these parameters.
A
Cooperman predicts high thermal ionization rates at high temperatures.
Trace quantities of alkali metals could serously effect ESP operation at
temperatures above 800°C. However, Shale's3 conclusion concerning
162
-------
thermal ionization is that this should not be a problem over
the range of temperatures and pressures considered.
A year later Shale repeated his earlier experiments, using positive
discharge corona. The same procedure that was used for negative corona
was used, with the polarity reversed. This work showed essentially the
same corona-start voltage for positive or negative corona, except that,
at above 650°C, the negative corona-start voltage was considerably lower
than for positive corona. Sparkover voltage was considerably higher for
negative corona below 200°C, at which temperature it was equal to that
for positive corona. Above 200°C, sparkover voltage for negative corona
was lower than for positive corona, becoming about equal to the corona
onset voltage at 800°C. The positive corona sparkover voltage, however,
remianed high at higher temperatures. From this result, Shale predicted
that positive corona would be more efficient at high temperatures. He
calculated that at 650°C a positive-corona ESP would have to be four
times as large as a negative-corona ESP to achieve the same collection
efficiency.
In 1969 Shale reported on a multitube, high-temperature, high-
pressure ESP operating at 800°C and 640 kPa. The equipment utilized a
modified atmospheric gas/air combustor. The ESP, designed and constructed
under contract with Research Cottrell, was 1.5 m in diameter and 9.1 m
high. There were 16 tube-type collecting electrodes each 15.2 cm in
2 3
inside diameter and 1.8 m long. Its SCA was 19.7 m /(m /sec), with all
tubes parallel with the gas flow. The power supply was 45 kV unfiltered
2
dc (70 kV peak) at 250 mA. This supply would allow about 1 nA/cm ,
which, although lower than used in his earlier experiments, is high
compared to atmospheric ESP's.
Shale found that he obtained higher collection efficiencies using
negative corona. He obtained much higher currents with negative corona
and operation was spark-limited. On the other hand positive current
amplitudes were about 20 percent of those obtained with negative corona
even though the voltage was considerably higher. He did not get sparking
at the maximum positive voltages, which would indicate that the maximum
voltage was still too low. With the equipment available, he was not able
to operate in the high voltage positive corona range where he predicted
that collection efficiencies would be higher. The apparent reason for
163
-------
this is that he increased the size of the tubes from 5.1 to 15.2 cm but
did not increase the power supply voltage sufficiently to be able to
operate in the range where higher efficiency was expected for positive
corona. The average measured efficiencies were 75 to 77 percent for
positive corona. The average measured efficiencies were 75 to 77 percent
for positive corona and 91 to 96 percent for negative corona.
In 1971 a larger scale HTHP ESP was tested by Brown at Research Cottrell
the temperature was as high as 940°C at pressures up to 1.05 MPa. The
ESP was a 20.3 cm diameter, 4.6 m long pipe. The corona V-I characteristics
and collection efficiencies agreed well with those obtained by Shale.
Negative corona currents were much higher for a given voltage than
positive corona currents, and negative corona collection efficiencies
were much higher. Collection efficiencies for negative corona were as
high as 91 percent; the SCA was about 37.4 m2/(m3/sec^.
Recent work on HTHp ESP's was also done at Research Cottrell by
8 9
Feldman. This work presents clean plate V-I corona curves for both
polarities at temperatures up to 1093°C and pressures up to 3.5 MPa.
The collector eletrode was a 7.6 cm diameter tube. In this work negative
corona appeared to provide higher sparkover voltages than positive
corona even in the range of temperature and pressure used by Shale and
Brown.
A recent review of HTHP ESP work in the Soviet Union is given by
VaTdberg. This work covers the temperature range to 400°C and pressures
of 100 to 600 KPa. Cylindrical and point-plane ESP's were used. The
cylindrical ESP's used tubes 6.4, 9.9, and 14 cm in diameter and 49.8 cm
long. The corona electrode was a strip-needle configuration, unlike the
smooth round wire or twisted wires utilized by U. S. investigators.
Their conclusions indicate once again that negative corona provides more
efficient collection than positive corona. Another cylindrical ESP
(with a 15.2 cm diameter tube, 2.4 m long) was tested on a blast furnace
exhaust at 99.9 percent efficiency. The SCA was 19.7 m2/(m3/sec). This
ESP operated at a temperature of 250°C and a pressure of 300 kPa. The
corona wire was operated at a negative voltage of 85 to 90 KV.
An HTHP ESP experiment is being conducted in Essen, Germany.11 The
apparatus consists of a recirculating pressure chamber, with provision
for dust entrainment; it too is cylindrical.
164
-------
The importance of hot gas cleaning is summarized in a recent EPA
12
report. The importance of fluidized-bed technology and hot gas cleaning
13
to the power industry is summarized in the 1977 FBC workshop proceedings.
A theoretical review of important parameters relating to dust collection
at HTHP is given in another EPA report. This report emphasizes that
the behavior of these parameters in the range of temperature and pressure
are only partially understood.
A review of HTHP ESP work is given by Robinson who indicates that
as pressure is increased a critical pressure may be reached where the
sparkover voltage is lower than the corona onset voltage. The critical
pressure is a function of corona electrode size and relative gas density.
A range of operation suitable for PFBC ESP's is given.
The results of this past work indicate stable operating regions for
HTHP ESP's suitable for use with PFBC. There appears to be agreement
that ESP's can be operated effectively at HTHP. The high temperature
has a degrading effect on performance, but these effects are countered
by increased pressure. If the corona characteristics are very nearly
depending on density only, at the pressures and temperatures under
consideration, one could expect two to three times the operating voltage
that is normally found in a typical cold-side precipitator. There are
many questions. How will gas density affect its viscosity and, in turn,
particle migration velocity? What SCA's will be needed? How should the
sections be subdivided mechanically and electrically? What about ash
resistivity and consistency at high temperature? What about thermal
ionization, particularly those trace compounds with low iom'zation f
potentials that may occur in stack gas? The object of the present
project is to test HTHP ESP performance in the laboratory, using reentrained
fly ash, and to determine answers to these questions.
LABORATORY MODEL HTHP ESP
A laboratory model HTHP ESP system is presently being designed and
is shown pictorially in Figure 1. The air compressor will supply 27.2
kg-mol/hr at 1 MPa to the burner package. The burner is a specially
designed high pressure unit capable of burning either methanol or No. 2
fuel oil. The burner has a maximum heating capacity of 244 J/s and a
maximum turndown ratio of 4:1. Gases at temperatures up to 1000°C
165
-------
leaving the ESP will be cooled to 260°C before the outlet sampling port.
Cooling will involve an air cooled jacket and cooling fins.
The planned pressure vessel is shown in Figure 2. It has a cool
outside pressure shell. The hot gases enter near the bottom and exit
near the top of this shell. The inside of the shell is lined with
castable .refractory. The chamber will be capable of accommodating
collecting tubes up to 30 cm in diameter. The collecting tube will be
supported by three rapping rods which will extend through the top of the
pressure vessel.
The corona wire will be supported by an air cooled, one piece, high
density alumina feedthrough. Cooling will be provided by cooling fins
inside the pressure shell and a water jacket outside. Radiation shields
on the corona wire will help prevent radiant heat from reaching the high
voltage feedthrough. The maximum design temperature for the high voltage
feedthrough is 260°C.
The shell and flanges will be carbon steel. The shell will be
rolled plate stock, with butt-welded ASA flanges. Sufficient insulation
will be provided to maintain a maximum temperature of 110°C on all
carbon steel components and welds. The collector tube will be 18 BWG
(1.25 mm) 310 stainless steel or equivalent. The corona wire will be
Hastelloy X, and the radiation shields will be 310 stainless steel.
Other materials Considered were Iconel, tungsten, and Incoloy. Inconel
and Incoloy alloys do not have the creep strength required at 980°C and
tungsten oxidizes at temperatures above 425°C. The cooling section of
the outlet will be 316 stainless steel or Hastelloy X.
A cooling water system will cool the top and bottom of the precipitator
vessel to a temperature where the high voltage feedthrough can operate
at a maximum temperature of 260°C. The system will consist of a cooling
water jacket around the top of the vessel, a circulation pump, and air
cooled heat exchanger, and associated piping and controls.
Laboratory tests were conducted to simulate the air flow and electrical
conditions of the HTHP ESP. For air flow modeling, the following was
assumed: the main gas velocity through the ESP is 1 m/sec, the collector
tube diameter is 29.2 cm, and the gas pressure is 1 HPa. The Reynolds
number was calculated to be,
Re = 1.4 x 104
166
-------
To obtain the same Reynolds number of the half scale model at an
ambient temperature of 21°C at Denver's altitude of 1.6 km required an
average velocity of
Vale = T-81"/5"
Air velocity traverses were made at 13 points along the length of
the model including the hopper and high voltage feedthrough area. The
velocity distribution had a normalized standard deviation of less than
10 percent in the collector tube between 3 tube diameters from the inlet
to 1 tube diameter before the outlet.
Smoke tests in the model indicated that essentially stagnant conditions
existed in the hopper and high voltage feedthrough areas.
The results of the air flow modeling indicates that good quality
air flow should exist in the HTHP ESP.
The electrical modeling was done in a 23 cm diameter tube to determine
if a lower corona wire guide would be required and to examine potential
sparking problems at the inlet and outlet ports. The tests were performed
at ambient conditions and the results extrapolated to indicate what
could be expected at the relative gas densities of the HTHP ESP.
The results^of these tests indicate that a lower corona guide will
be required to prevent pendulum action of the corona wire. To ensure
good electrical insulation of the guide, the hopper area, where the
guide is located, will be water cooled.
The tests also show that sparking should not occur with clean
conditions below 200 kV. For a corona wire diameter of 0.635 cm, the
calculated negative corona onset voltage is 130 kV. It is expected to
2
be possible to obtain corona current densities as high as 1 a/cm .
While the electrical tests at ambient conditions should not be used
to predict actual performance of the HTHP ESP, they were done to help
indicate practical parameters for the working unit.
REFERENCES
1. Keller, L. R. and Fremont, H. A., "Negative Wire Corona at
High Temperature and Pressure," Journal of Applied Physics, Vol. 21, pp.
741-4, August 1950.
167
-------
2. Thomas, J. B. and Wong, E., "Experimental Study of dc Corona
at High Temperatures and Pressures," Journal of Applied Physics, Vol.
29, No. 8, pp. 1226-30, August 1958.
3. Shale, C. C., Bowie, W. S., and Holden, J. H., "Feasibility of
Electrical Precipitation at High Temperatures and Pressures," Bureau of
Mines Report of Investigations RI 6325, 1963.
4. Cooperman, P., Commun. Electron., 75, 792, 1964.
5. Shale, C. C., Bowie, W. S., Holden, J. H., and Strimbeck, G. R.,
"Characteristics of Positive Corona for Electrical Precipitation at High
Temperatures and Pressures," Bureau of Mines Report of Investigations
RI 6397, 1964.
6. Shale, C. C. and Fasching, G. E., "Operating Characteristics
of a High-Temperature Electrostatic Precipitator," Bureau of Mines
Report of Investigations RI 7276, 1969.
7. Brown, R. F. and Walker, A. B., "Feasibility Demonstration of
Electrostatic Precipitation at 1700°F, "Journal of the Air Pollution
Control Association, Vol. 21, No. 10, pp. 617-20, October 1971.
8. Feldman, P- L., "Development of a High Temperature Electrostatic
Precipitator," Progress Report, EPA Contract 68-02-2104, Cottrell
Environmental Sciences to EPA, Research Triangle Park, N. C., June 1975.
9. Feldman, P. L., "High Temperature, High Pressure Electrostatic
Precipitator," in EPA/DOE Symposium on High Temperature/High Pressure
Particulate Control, EPA-600/9-78-004 (NTIS No. CONF-770970, 3/78,
Washington, D. C., September 20-21, 1977.
10. Val'dberg, A. Y., Danilin, V. V., Lyapin, A. G., and Tkachenko, V. M.,
"Electrical Gas Cleaning at Higher Pressures," in Second US/USSR Symposium
on Particulate Control, EPA-600/7-78-037 (NTIS No. PB 279-625, 3/78,
Research Triangle Park, N. C., September 2629, 1977.
11. Weber, I.E., "Problems for Gas Purification Occurring in the
Use of New Technologies for Power Generation," EPA/DOE Symposium on High
Temperature/ High Pressure Particulate Control, EPA-600/9-78-004 (NTIS
No. CONF-770970, 3/78, Washington, D. C., September 20-21, 1977-
12. Parker, R., Calvert, S., and Drehmel, D. C., "High-Temperature
and High-Pressure Particulate Control Requirements, EPA-600/7-77-071
(NTIS No. PB 271-699) July 1977.
13. "Proceedings on the Fluidized Bed Combustion Technology
Exchange Workshop, Vol. 1 and 2," sponsored by ERDA and EPRI. CONF-770447-P-l,
April 13-15, 1977.
168
-------
14. Calvert, S., Parker, R., and Drehinel, D. C. "Effects of
Temperature and Pressure on Particle Collection Mechanisms: Theoretical
Review," EPA-600/7-77-002 (NTIS No. PB 264-203), January 1977.
15. "Air Pollution Control, Part 1", edited by Werner Strauss,
John Wiley, New York, pages 283 to 298.
169
-------
HIGH VOLTAGE
FEED THRU
RAPPING
RODS
BURNER
PACKAGE
FUEL TANK
SILENCER
AIR COOLING JACKET
FINNED COOLING PIPE
OUTLET
SAMPLING
FIGURE 1. HTHP-ESP SYSTEM
-------
-7- H.V. FEEDTHROUGH
WATER COOLING JACKET
COOUNG FINS
'RADIATION SHIELD
RAPPING RODS
COLLECTING PLATE
CORONA WIRE
INLET
OUTLET
PRESSURE SHELL
BLANKET INSULATION
CORONA WIRE GUIDE
WATER COOLING JACKET
FIGURE 2. HTHP-ESP VESSEL
171
-------
V.M. Tkachenko, V.A. Rodionov, A.D. Malgin
PURIFICATION OF HOT ASPIRATED AIR FROM SINTERING
MACHINES AND CLINKER COOLERS
Presently in the USSR, dry electrostatic precipitators
have been accepted as the basic unit in planning the purification
of aspirated air in iron industry sintering plants.
With an increase in aspirated air temperature, the re-
sistivity of dust caught in the precipitators rises and this may
cause back corona in the precipitator's electrodes. It is known
that the back corona considerably lowers the precipitators1 work
efficiency.
Therefore, one of the basic research problems of electro-
static precipitation of aspirated air from sintering machines is
in determining maximum temperature and other parameters of the
aspirated air, under which precipitators can efficiently work
without back corona.
Such research has been conducted by the State Scientific
Research Institute for Gas Purification in Industry and Sanitation
(NIIOGAZ) in the sintering plant of a metallurgical factory
where aspirated air from AKM-312 sintering machines was cleaned
on installment of 8 DVPNI-4x20 electrostatic precipitators. Before
passing the precipitators, the aspirated air at this site was
mixed in a common collector. This air was from various sources:
sites where charge was loaded into compacters, from hammer mills,
sifters, conveyer sinter overfills, covering of front and rear
portions of line coolers, and others. Cleansed aspirated air
enters an exhaust fan from each precipitator and is then expelled
172 *
-------
through a common flue into the atmosphere.
The electrostatic precipitator DVPNI-4x20 is a dry
gas, vertical, plate-type unit with coronal needle electrodes
and has the following technical characteristics :
1. Number of sections 4
2. Active sectional area of sections, m2 20
3. Number of precipitating electrodes in
one section 18
4. Precipitating electrode dimensions, m 7.9x4
Each precipitator section is fed by a separate unit, the
ATF-250, with nominal rectified current of 250mA and a maximal
voltage value of 80 kv.
During research on precipitators, sintering machine tech-
nological condition was characterized by the following indices:
CaO
sinter basicity (the relationship = ) ran 1.16-1.30, and
iron content in sinter :was 57.9-58.7%.
From the eight operating precipitators, one was chosen
for research at whose intake the temperature was maximal and
ranged from 107-115 °C. The air's moisture content had a value
of 7.5-13g/m^ with an mean value of ll.lg/m^ (of moist air).
Under these conditions, the cleansed air's dust resistivity
changed in the range of (1-8) 10^° ohm .cm. (i.e. on a critical
value level) . Average particle size at the intake, as determined
by liquid sedimentation, was 19 microns.
Under the indicated moisture-temperature conditions
of aspirated air, dust resistivity values, and the tested cur-
rent density (up to 0.23mA/m2), the precipitator operated with-
out back corona. Coinciding corona volt-ampere characteristics,
measured under voltage rise and drop attests to the lack of
173
-------
back corona. A sample of such characteristics is introduced
in table 1. It is known, that with the presence of back corona
in precipitators, the volt-ampere characteristic runs higher
at a drop in voltage than with a voltage increase.
With air entering the precipitator at an operational velocity cf
0.75m/s, the degree of cleaning's mean value comprised 99.38%
when the residual dust content was 13mg/m^(n), and with an in-
crease in the velocity up to 1.15 m/s, the degree of purifi-
cation was lowered to 96.8% and the mean value of residual
dust after electrostatic precipitation rose to 87mg/m3(n).
The indicated precipitator work efficiency was ob-
tained with feeding the units under conditions of periodical
spark-overs and optimized electrode rapping the interval be-
tween rprecipitator electrode rappings were 125 minutes, and its duration was for
five minutes, with 40 and 20 minutes correspondingly for coro-
nal electrodes.
Thus,- at temperatures of sintering machine aspi-
rated air reaching 110°C, and a moisture content of 7.5-13g/m3
(of moist air), the DVPNI-4x20 vertical precipitator works effi-
ciently without back corona under a current density of 0.23mA/m2,
which conforms well with results, predicted for back corona occurehce
with use of NIIOGAZ equipment (fig.l.).
Aspirated air from clinker coolers Contains a
large amount of coarse dust fractions, whose median particle di-
ameter is 30-40 microns with a dispersion of 4.5-5. The dust is
noted for its abrasiveness and high resistance which usually
ranges 109 to 10^° ohm-in-
Aspirated dust temperature depends on processing con-
ditions and may change from 120°C to 280°C.
174
-------
10
18
22 26 ' 30
38 42 &
Fig.l. Precipitator_ corona volt-ampere-characteristics for the
DVPNI-4x20: 1. at a rise in voltage 2. at a drop in voltage; air temp-
erature in precipitator, 107°C, 11.2g/m3 air-moisture.
175
-------
In using electrostatic precipitators for cleaning aspirated gases, trou-
bles occur, connected with the dust's high resistance. To efficiently purify the gas,
it is necessary to condition the dust-gas mixture entering the cleaning process. It
will, otherwise, complicate and raise the expense of cleaning to a considerable degree.
Using cyclone dust-removal equipment will not meet sanitary requirements
set for cleaning aspirated air. The abrasiveness of coarse dust particles rapidly erodes
cyclone elements.
Ihe high temperatures of aspirated gases impede the use of filter bags
equipped with inexpensive filtering materials made from synthetic fibers.
As test results from NIIOGAZ's Semibratovsk branch show, using gravel
filter-cyclones has been shown to be highly-promising for cleaning clinker cooler
gases. Ihe filter-cyclone's distinctive feature is its three-stage cleaning process.
A precipitation chamber, spread between the two rows of cyclone elements, is the
first step in the process. The third step is a filled layer made up of 2-4rrm gravel.
A significant characteristic of the unit is the double layer positioning
of the filled layer and its regeneration by agitation and reverse blowing with air.
The blown air which passes through the filter layer from below upwards attracts dust
particles and carries them to the cyclone, where the large particles of sinter material
settle out. The fine sinter particles together with the blown air stream enter other
sections which operate as filters and settle out in the gravel layer.
Agitation has the main significance of separating dust from the filled
layer granules. Moving relative to each other during granule agitation, the dusts,
which cannot be blown out by an air-jet stream, are conducive to separation.
The filter-cyclone FTGN-120, with a 120m filter surface, is
used to clean aspirated air from a "Volga 35" clinker cooler in one
cement plant. The unit is estimated to clean gas at a rate of 100-120
thousand m /hour. Experience with filter-cyclones in industrial condi-
tions has shown that dust concentration at the unit's outlet is
176
-------
40-50mg/m^, and the degree of purification about 99%. The
unit's hydraulic resistance is 2500-3000 Pascals. Studies
have shown that 80% of the dust settles in the cyclone and
dust chamber and only 20% enters the filled layer.
Uniform distribution of the granules in the layers
in one and the same section has the essential effect on the
operation of" the unit. In those trays where the gravel layer
is lower, the hydraulic resistance is less, the stream of
dust and gas passes through it with great speed, and the
degree of purification is reduced. In connection with a
build-up of a considerable dust layer, regeneration in
these trays passes less intensively.
Filter-cyclone operation in cement factories has
shown the unit's sufficient reliability.
177
-------
A MODEL OF ELECTROSTATIC PRECIPITATION
FOR TI-59 CALCULATOR
by
L. E. Sparks
Particulate Technology Branch
Utilities and Industrial Power Division
Industrial Environmental Research Laboratory
Research Triangle Park, N. C.
ABSTRACT
A mathematical model of electrostatic precipitation has been developed
for the Texas Instruments TI-59 calculator. The equations used in the
model are presented. The calculated results of the model are shown to
be in good agreement with the results of the EPA-Southern Research
Institute Computer Model for Electrostatic Precipitation.
(This document has been reviewed and approved
for publication by the U. S. EPA.)
178
-------
INTRODUCTION
Recent advances in calculator technology have made available calculators
with tremendous computing capability. This computing capability can be
used to model electrostatic precipitation without a large computer.
This paper describes such a model for the Texas Instruments TI-59 Calculator.
THE MODEL
The ESP Model for the TI-59 Calculator is based on the EPA-Southern
Research Institute (EPA-SoRI) ESP Computer Model Revision I by McDonald.
A brief discussion of the theory behind the model and the assumptions in
the model are given here. The reader interested in more detail should
consult the referenced reports.
Steps in Electrostatic Precipitation Process
The electrostatic precipitation process can be divided into 4
steps:
1. Corona generation and establishment of electric field.
2. Particle charging.
3. Particle collection.
4. Removal of the collected material.
A complete model of the process would handle all four steps. However,
Step 4 is too complex to be modeled at this time and is neglected by
both the EPA-SoRI Model and the TI-59 Model. Step 1 is modeled in
detail by the EPA-SoRI Model but is neglected in the TI-59 Model. The
TI-59 Calculator does not have sufficient power to handle the computation
used in the computer program. Work is underway to develop approximations
for use in the TI-59. When these approximations are developed they will
be incorporated into the model. Until then the user must supply as
input data the applied voltage and current density.
Particle Charging
Particle charging generally takes place by two mechanisms—field
charging and diffusion charging. ' For large particles field charging is
by far the dominant mechanism. Diffusion charging dominates for very
179
-------
small particles. Particles of major interest in air pollution (those with
o
0.1 ? d < 2 um) are charged by both mechanisms. Pontius et al have shown
that the following approximation .for the charge on a particle agrees with
experimental data and detailed theory fairly well.
:p = _nd C, (b d/aEAvNt [1 + 2 (£i)] + C?J In U^JV 1 > (D
2 ' b Nt + C1 K * d ^Ll4'
rrp "-'•= number of charges
Cl = 4Ve
C2 = k/e
EO = permiti vity of free space
e = charge on electron
k = Boltzman's constant
b = ion mobility
EAv = average electric field
N = free in density
t - residence time for charging
K = perticle dielectric constant
v = mean thermal speed of ions
d = particle diameter
The charge on a particle, q, in coulumbs is given by
q = np e
This equation is used in the TI-59 program and in the approximation procedure
in the EPA-SoRI Model.
The average electric field used in the calculation is given by
EAv - U/H (2)
where U = the applied voltage qv
H = wire to plate spacing, m
The free ion density, N, is given by
N - bEAv/j
where b = average ion mobility,
2
j = current density a/m
180
-------
Particle Collection
Particle collection in an ESP is given by the Deutsch Andersen equation
Pt(d) = exp [ - u(d) A/V] (3)
where ui(d) = electrical migration velocity of particles
with diameter d, m/s
2
A = collector plate area m
.3
V = volumetric flow rate of gas m/s
The ratio A/V is called the specific collector area (SCA).
The electrical migration velocity near the collector plate for small
particles is given by Stokes' law as
0)
(d) = q EpC'/3nyd (4)
where q = the particle charge, coul. (q = en)
Ep = the electric field at the plate, v/m
C1 = the Cunningham correction factor
.,*»*
A = 1.246 + 0.42 esp (-0.87 d/2A)
X = mean free path of gas
ip~7 (296T2/
X = mean free path at 23°C and 76 cm Hq pressure
o
P = barometric pressure, cmHg
T = temperature, °K
vi = viscosity of gas kg/m-s (10 poise = 1 Kg/m-s)
McDonald reported that for ESP collecting flyash
Ep • EAv/1.75 (5)
This estimate of E is used in the TI-59 Model.
181
-------
McDonald also reported that equation (4) underpredicts the migration
velocity for real ESP. He recommends that the migration velocity be corrected
by an emperical factor to improve agreement between predictions and theory.
The corrected migration velocity is given by
a(d) = u;(d):i X (1.7 - 0.45d) (5)
where ^(d), is the uncorrected migration velocity Equation (5) applies for
H
0.2786 +0.0755 aln [l/Pt(d)] (8)
a = the normalized standard deviation of the gas flow
(o = 0.25 is generally considered good)
Note that both B and F are particle size dependent.
182
-------
The overall penetration, Pt , is given by
PtQ = /°>t(d)f(d)dd (9).
and the corrected overall penetration, Ptn", is given by
PtQ ' = /"Pt(d)' f(d)dd (10)
where f(d) is the fraction of particles with diameters between d and d + dd.
TI-59 ESP Model
A mathematical model based on the above equations has been programmed for
the Texas Instruments TI-59 Calculator. In the TI-59 Model the ESP is
divided into user specified time increments. The division of the ESP into
time increments is necessary to account for the time dependent nature of the
particle charging and particle collection process. For each time increment
the calculator calculates particle charge, particle migration velocity, and
particle penetration through the increment. All these calculations are
performed for each particle diameter in the user specified particle size
distribution. The penetration of a given particle diameter through the
ESP is given by
n
Pt(d) =|| Pt(d)1 (11)
where Pt(d). = penetration of particles of diameter d through
A.U
the 1 increment.
An effective migration velocity for each diameter for the entire ESP is
calculated from
u(d) - - In Pt(d) (12)
e A/V
The penetration corrected for non-ideal factors Pt(d)1 is calculated
from
Pt(d)1 = exp - ("(d)e A/V) (13)
B x F
183
-------
These calculations are performed for each of the m particle diameter in
the user specified particle size distribution. The overall penetration
is given by
m
Ptn = - Pt(d). f. (14)
0 1 = 1
and the corrected penetration
i m
Pt_ - X Pt(d). f. (15)
0 1=1 n 1
The results of the TI-59 Model are in good agreement with results of the
computer Model as shown in Table I.
CONCLUSION
The TI-59 Model provides a useful tool for examining the performance of
an ESP. The answers are obtained quickly and without a large expensive
computer.
REFERENCES
1. McDonald, J., A Mathematical Model of Electrostatic Precipitation
(Revision 1) in two volumes. EPA-600/7-78-11 la and EPA-600/7-78-llb.
2. Pontius, D. H., L. G. Felix, J. R. McDonald and W. B. Smith,
Fine Particle Charging Development. EPA-600/2-77-173, August 1977.
3. Gooch, J. P., J. R. McDonald, and S. Oglesby, Jr., "A Mathematical
Model of Electrostatic Precipitation." EPA-650/2-75-037, 1975.
184
-------
Table I. Comparison of TI-59 and EPA-SoRI
Computer Model Results
Diameter
0.30
0.40
0.525
0.7
0.95
1.25
1.6
2.05
3.15
4.0
5.25
7.0
9.0
13
Pt(d) Computer
0.15
0.16
0.16
0.16
0.15
0.13
0.11
0.10
0.086
0.073
0.049
0.024
0.012
0.0037
Pt(d) TI-59
0.13
0.15
0.15
0.15
0.14
0.13
0.12
0.10
0.074
0.060
0.042
0.021
0.010
0.0036
Conditions:
U = 26.7 KV j = 13.3 na/cnT
SCA = 56.81 m2/m3/s
185
-------
Urgent Ash-Tra??ir.c Problems a_t Thermal Power Stations
L.I. Krop?, G.S. Cheko.r.cv'
Thermal power engineering's basic orientation to
the use of solid fuels, including its inexpensive types, in-
creases the urgency of protecting the atmosphere from stack
gas solid contaminants. The extraordinary diversity of domes-
tic fuels is responsible for the substantial variety in emis-
sion of harmful contaminants in combustion products. Table 1
includes data on several fuels and their usage in larce-
scale condensation--type thermal power stations.
Table 1
I
! gas discharge. Cirou
Fuel Jjiormal conditions)
Berezov lignite
KuzrfeEz'SS
Donetz low-grade coal
Ekibastuz coal
Estonian shale
Moscow lignite
Mazut^' ^J
4,35
4,1
4.0
4,05
5,2
4.8
4.0
! .fly-ash
29
82
97
253
320
242
0,4
In the hygienic evaluation of the harmfulness
of fly-ash from thermal power stations, at the present time,
it is accented that its charactersitics most nearly correspond
to non-toxic dusts raised, for example, from road surfaces
by wind or traffic. In reality, it is known however that there are
harmful substances in the composition of fly-ash, such as
186
-------
free calcium oxide and silicon oxide, incompletely-burnt fuel
particles and snail quantities of still nore harmful substances
such as lead, arsenic, mercury compounds, and fron some fuels,
radioactive elements (1). The content of these contaminants
found in ash is significantly higher than their averaoe con-
tent in the_earth's crust. Therefore, only some fuel ash can be
equated with non-toxic dusts. At the present time, detailed ana-
lyses of the various fuel ashes are still lacking and the full
measure of their toxicity has not been determined. In connection
with this, it has not been excluded that, subsequently, maximum
concentration standards (MFC) for ash in the atmosphere will be
established differentially with respect to peculiarities of
the various fuels1 compositions.
Zn addition to this, MFC values established by sani-
tary institutions only indicate that dust concentration in the
air by weight should not exce i the given value. With this,
dust dispersion composition is not specified. Meanwhile, it is
known that dust with a particle size over 10 microns doesn't pene-
trate human lungs but is arrested in the upper respiratory
track where it is comparatively easily expelled from the orga-
nism. On the other hand, dust particles that are around 1 micrcr.
in size do penetrate the lungs, remain there and cause many
illnesses. From that point of view, fly-ash released into
the atmosphere past thermal power station fly-ash control equipment is
i3angerous, since it has a nearly 30% composition of particles less than
5 microns.
It should also be noted that the atmosphere's ability to
cleansing itself of dust particles is ir. a very large decree
dependent on the dispersion composition of the dusts. Thus, particles
187
-------
of a size larger than 10 microns subsiderelatively quickly to
the earth's surface, 4-10micron particles rise another kin. with
the smoke current and are spread along thousands of kilometers of the earth's
surface. Particles less than4 micron?precipitate poorly in
rain, but slowly subside, reaching the surface from the heiqht
of a kilometer in the course of a year. Particles that are less
than a micron in size are involved in Brownian motion
and are spread like atmospheric gas molecules, settling to the
surface only under special conditions, namely, with moisture
condensation around them. Thus, highly dispersed fly-ash is
capable of accumulating- in the atmosphere, which can lead to
'undesirable consequences (2).
In this connection, gaseous stack gas contaminants,
such as sulfuric and nitrous oxides are less dangerous since
their presence in the atmosphere is limited to approximately
five days, during which they are removed from the air, dissolv-
ing into open water and precipitation, reacting with ammoimia
secreted by microorganisms, and absorbed by soil and vegetation.
It should also be considered that atmospheric doist
disperses solar radiation and reflects part of the energy
reaching earth back into space. Measurements show a signifi-
cant loss of solar radiation intensity in large cities and
industrial centers, especially the ultra-violet region of the
spectrum, which is the most active biologically. Light disper-
sion by one and the same amount by weight of dust is inverse-
ly proportional to particle size to the sixth degree. The
intensity of light dispersion depends not only on particle
concentration, but also on the thickness of the dusty air
layer.
188
-------
Taking all sanitary and ecological considerations into
account together, it is necessary for power generating units
of the new large-scale thermal power stations burning solid
fuel to provide the maximum technologically attainable degree
of ash removal from stack gases, independent of stack height.
For lower capacity thermal power stations, and also for those
burning fuels with lower ash content, the regulated degree of ash-cleaning
may be lowered somewhat. Indices based on such an approach are
presented in table 2 (3). In the following, these indices may
be more rigid, according to the degree of completion of the dust-
trapping equipment.
Table 2
.4Type— and capacity af- power—-
generating station burning
.srlid fuel. " -•
Neces&ary-- degree- -of - stack-
cleaning for fly-ash, --no
•Given ash con-
tent of fuel
%k-g71WO~kcal
I. SREPS*, 2400 MW output 98.5-99.0
TETS** 500 MW output
2. SREPS , from 1000 to 2400 MW 98.0
TETS, from 300 to 500 MW
_ - " QA 0
o. SREPS, less -than.- LQOO- .MW ^ -• ^
TETS, less than 300 MW
j f "s-5
j
'" 99.0
98.5
96.0
gas •- • ; \
lower • ;
j **> 5
i '
1 1
! ;
99.5
99.0
98,0- :
* SRFPS - State Regional Electric Power Station
** TETS - Central Heat and Electric Power Plant
For meeting the requirements introduced in table 2,
large-scale coal-burning power generators are eauipped with elec-
trostatic precipitators. For low -capacity power stations and boile:
generators, scrubbers and inertial action dry dust-catching
189
-------
equipment is used.
Technically feasible adaptation of parameters
of exit gases (temperature and sometimes moisture content)
is of major significance to conditions for efficient ash-
trapping by electrostatic precipitators, and it is especial-
ly urgent with large-scale generating units burning berezov,
ekibastuz, of kuznetz coal. In connection with this, it is
advisable in some instances to re-examine norm definitions
for £.Xi'"f gas temperatures and judge this index on the
basis of a summary minimum outlay of steam generators and
ash-trapping equipment.
A complex approach based on close interrelation-
ship of electrostatic precipitators and convective parts of
the steam generator may be, as experience shows, widely
used not only in designing new thermal power stations, but
in those .already operative, that are burning solid fuels.
Thus lowering discharged gases' temperatures by 15-20° C
within allowable limits before they reach the precipitators
may in many cases cut fly-ash emission by 30-40? with a si-
multaneous increase in ash-trapping reliability. An analogous
effect is also attained by introducing measures to lower
the excess coefficient of air in discharge gases to the
value Aa— 0.2-0.25. Such an approach, along with cutting
fly-ash emission, assures perceptible fuel and electrical
economy.
With many coal-burning thermal power stations
equipped during a period of."less demand for atmospheric purity,
modernizing active electrostatic precipitators to increase
their efficiency and working reliablity , the basic resource
190
-------
for cutting £ly -ash emission from energy production, is a com-
plex problem. Considering here industrial site limitations and
thermal power station working conditions, it is especially
important to find methods of intensifying processes in electrostatic
precipitators, to find layout methods and techno-econoir.ic
expediencies in a variety of cases where there is some increase
of hydraulic resistance in modernized ash-trapping installations,
in order to attain the necessary indices. Solving these problems de-
mands considerably widening the typical dimensions of precipita-
tors installed in power stations according to the height of pre-
cipitating electrodes, width of the units, etc.
One of the problems with electrostatically cleaning
stack gases is in providing stable and efficient precipitation
while burning a quantity of mazut at the same time hard fuel is
being burned. In connection with the insolvability of the above problem,
electrostatic precipitators are in many instances shut off while hard
fuel and mazut are simultaneously burned, which is impermissi-
ble. Meanwhile, in connection with present fuel-supply condi-
tions, simultaneous burning of both of these fuels will be
practiced at many power stations.
A considerable share of energy production in ther-
mal power plants will be based on using low-sulfur, low-mois-
ture coals, especially ekibastuz and kuznetz coal, whose com-
bustion products possess unfavorable electrophysical proper-
ties and therefore," solid contaminants are porrly removed by
electrostatic precipitators, due to the so-called back corona.
Relative to this, criteria can be used for predicting
ash behavior in precipitators in burning diverse fuels whose
numerical value is related to factors determining volumetric and
surface resistivity of the ash layer vprecipitators
191
-------
during cleaning of combustion products of a given fuel (4).
The proposed criterion has the form:
K = (Al 20°4S iOf>)*AP
(WP-f9HP)SP
Here all designations are commonly accepted. The criterion is
a dimensionless value. During its derivation, the following circumstan-
ces were considered. Aluminum and silicon oxides have the high-
est volumetric resistivity in ash composition. The higher the
content of these oxides in the ash, the more likely will the
ash layer in the precipitator have a high volumetric resisti-
vity. Therefor, the value representing the sum of these oxides
is placed at the root of the criterion.
It is known that calcium oxide also has a high resisti-
vity value, however, zxmchemical analysis of ash under calcium
oxide it is understood that not only free calcium oxide is con-
tained in the ash, but also all other calcium compounds, among
them sulfate having a significantly lower resistivity value
and always present in ash.Thus, calcium oxide is not included
among the oxides determining high volumetric resistivity of ash.
In order to reflect the effect of processes determin-
ing surface layer resistance, the value representing fuel ash
content relationship to sulfur content has been in the criterion.
with this, it is proposed that the amount of sulfur oxides formed in
the boiler is proportionate to fuel sulfur content, but dust
layer surface resistivity changes correspondingly to the amount
of ash per unit of forming sulfur trioxide.
In actuality, the amount of sulfur trioxide found
in stack gases entering electrostatic precipitators depends
on the content of alkaline compounds in the ash. However, considering
that the result of chemical reaction of sulfur trioxide
192
-------
alkaline compounds forms sulfates also having a measure of volu-
metric resistivity, adjustment for sulfur trioxide binding with
ash is not introduced in the criterion.
With the increase in fuel combustion products of the
content of water vapor formed from burning fuel hydrocarbons
and "from moisture eveaporation, the temperature rises, upon
which sulfuric acid steam condensation begins on ash parti-
cle surfaces and the probability of an ash layer with a high
volumetric resistivity at 130-150°C lessens. Therefore, the
value characterizing combustion product moisture, W°+9H^, is
brought into the criterion.
Given in table 3 are the value K's compilation results
for the majority of the USSR's energy fuels, which are arranged
in order of the value's decrease. The greatest values for K were
rdCt'vtdf j0
-------
scale of the given problem, it is advisable to widen research
and testing of electrostatic technical processes for overcoming
the harmful effects of back corona in precipitators in ~&fa>
burning of such fuels. In this regard, works on optimization
of pulsed supply presnt interest. Moscow Power Engineering In-
stitute has suggested a method of feeding precipitators by
square - wave alternating voltage (5) and also,
in the USA, there has been recent development of
systems for preparatory ionization of stack gases enter-
ing the precipitators. Advisable as well are non-electric
methods of dry ash-trapping, applicable according to efficiency,
but in a greater measure satisfying working conditions of ther-
mal power stations. These are more applicable than electrostatic
precipitators, that are, for example, less sensitive to increase
in stack gas temperatures encountered in steam generator operation.
To achieve all-out cutting out of fly-ash emission, it
will also be necessary in coming years to study -fKe. increasing
efficiency and operative reliability of a large stock of
dust control equipment used in operating electric power plants
and also at new low and medium capacity thermal-electric power
plants burning fuels with a low calcium oxide content in the ash.
Relatively inexpensive scrubbers with Venturi co-
agulators have been developed and are being used in many
thermal power stations eqipped with reverse scrubber systems
cleaning stack gases at 96% with a hydraulic resistance near
100mm Hg (6).
Completed research results show that the means
of optimizing Venturi tube scrubbing and estimating physi-
cal-chemical reaction properties of gas-forming and liquid
phases is the completely practicle construction of economi-
cally adaptable equipment assuring reliable gas scrubbing
194
-------
at 97.5-98.5% efficiency with burning of the majority of solid fuel
types under condition of reverse water supply.
Reliable performance of the scrubber system is the necess-
ary condition for highly effective work by scrubber units. The mechan-
ises intended for ash-removal from dust control equipment hoppers should also
exclude the suction of air into the equipment. This should also eliminate
dust spilling into the scrubber building due to single instances of
loading of considerable amounts of ash into the hoppers, such as, for ex-
ample, when the precipitator electrodes are rapped after lengthy
shut-offs of the rapping mechanism, and ensure sufficiently
rapid emptying of hoppers full of ash as need arises.
All of these demands are met by ash-removal systems,
adapted for use in high-capacity power stations equipped with elec-
•
trustatic precipitators. Beneath every precipitator bin in these
systems is placed a pneumatic-layer seal batcher or a pneumatic-
layer seal batcher switcher. Pneumatic layer seals are analogous
by active principle to hydroseals, in which aerated ash takes the
role of fluid (7).
Seals are combined with a system of jets by which ash
enters the mixing unit where it is mixed with water and later
pumped to an ash dump. If there is a need for use of dry ash, then
part or all of the ash gathered by the jets, is pumped pneuma-
tically to a special ash deposit, from which it is taken as needed.
Work experience with such ash removal systems asserts
their reliability and economy- Furthermore, pneumatic layer seal
batchers placed under precipitator bunkers allow the precipitating
electrodes to be rapped after maximally lengthy periods, based
only on the thickness of the ash layer on the electrodes and its
195
-------
effect on the precipitator's electrical condition. This practical-
ly implies that rapping intervals of the first-field electrodes
may be increased as much as 1-2 hours, the second field, by 4-8
hours, and those of other fields by as much as a day or more.
Such a method of electrode rapping, aside from
increasing dust-trapping efficiency, assures an increase of essential
work reliability by the mechanical part of precipitators,
as well as considerable cut in energy expenditure in rapper
and ash removal drive mechanisms.
The work of wet scrubbers is still dependent in a
large degree on ash removal systems. For preventing water-
table pollution, water ash removal systems are made reverse-
able. Due to frequent ash contacts with stack gases, the sali-
nity of water circulating in reverse water ash removal systems
increases to 10g/l and more, and the pH indicator reaches
11.5-12.0 and at higher ash calcium oxide content, pH reaches
12.6. Using such water for irrrigating scrubbers is connec-
tecPdifficulties coming from carbonate and calcium sulfate
deposits forming in pipelines, jets, nozzles, on the walls of
intake pipes and demisters.
On the basis of research and work experience with
scrubbers irrigated with back-water, it has been established
that, as a means of preventing carbonate deposits, it is neces-
sary to change irrigation water's pH to 10.5, and then sulfate
deposits will not form if the production of Ca2*and SO4 lon
concentration does not exceed a value, equal to 1500. The last
condition may be maintained if the reverse water ash removal
system's work is provided a set regime including a sufficiently
lengthy period for water to clarify at the ash dump and also
an. insignificant addition of industrial water to the system.
196
-------
With concern to lowering alkalinity, this may be done by adding
a corresponding amount of acid.
At the present time industrial testing is being carried
out on a method of neutralizing alkaline clarified water by use
of stack gases, which other than cutting expenditure on reagents,
2+
allows for lowering Ca ion concentration in the water, which is
important for preventing sulfate deposits.
Table 3
1 —
Fuel j
Deposit j
I !
Kuuchekin
Ekibastuz
Tpmusin_open-pit_
Nos. 7~8 Kuzbass"
Kemerov
rJVvrrmcirl r>p»n — pi -h
Nos. .3-4 Kuzbass.
-Kara.g.an.din.._
Promprodukt
Kedrov .open-pit
Kuzbass
Osinnikugol
Krasnogorsk open-pit
Kuzbass
Gusinozersk
Listviansk- open-pit
Kuzbass
Donetz A-Sh
Vakhrushi open-pit
Kuzbass
Novosergievsk open-pit
Kuzbass
Krasnoborodsk open-pit
"Kuzbass"
Kolmogorsk open-pit
Kuzbass
K
2
230
177
175
143
160
94.8
«h *
84
79
63.5
48.5
46.7
46.0
44.0
43.0
41.0
35.7
I :Fuel }
i 'Deposit j
! 3 !
Donetz ARSh
Raichikhin
Kharanor
Irsha-Borodinsk
Estonian shale
__I_ntin
Donetz G Promprodukt
-Badaevsk open-pit
Kuzbass
Cheliabinsk
Tserezovsk
Podmoskovnoe
Lvovsko-Volynsk
Donetz
Itatsk
Donetz D
Donetz G
Donetz T ~
Kizelovsk
Angrensk
197
1
V "*"" l
K !
A I
35.0
33.7
31.0
30.0 '
20.0 ;
22.0
21.0
20.0
17.4
16.0
15.4
15.0
14.3
12.7
12,57
12.4
TT . 5
• ^ f W
10.4
10.0
-------
BIBLIOGRAPHY
1. Zalogin, N.G., Kropp, L.I., "The Work of Electric Power Stations '
and Protection of the Environment", in "Thermal Power Engineering",
1975, no.4, p.5-7.
2. Zalogin, N.G., Kropp, L.I., "On the Necessary Degree of Clean-
ing Fly-Ash From Stack Gases At New Large-scale Thermal Power
Stations", in "Electric Power Stations", 1977, no.6, p.2-3.
3. Kropp, L.I./ "Ways of Abatting Harmful Emissions From Thermal
Power Stations", in "Thermal Power Engineering", 1978, no.11,
p.2-7. • .
4. Zalogin, N.G., Kropp, L.I., Shmigol', I.N., "Criteria for
Evaluating the Behavior of Fly-Ash in Electrostatic Precipitators",
in "Electric Power Stations", 1978, ho.6, p.15-16.
5. Mirzabekyan, G.Z., Rudenko, V.M., Shevalenko, I.S., "Study
of Electrostatic Precipitator Feed Conditions, Assuring the
Absence of Back Corona Discharge", in "Electron Treatment of
Materials", Academy of Sciences of the Moldavian SSR, 1978,
no.2, p.70-73.
6. Kropp, L.I., Akbrut, A.I., Ash-Control Equipment With
frenturi Tubes at Thermal Power Stations, Moscow, "Energy",
1977, 159pp.
7. Chekanov, G.S., "The Use of Pneumolayer Mechanisms for
Ash-Removal from Control Equipment Hoppers" in "Industrial
and Sanitary Gas-Cleaning", 1978, no.l, p.9
198
-------
PARTICLE COLLECTION BY GRANULAR BED FILTERS
AND DRY SCRUBBERS
by
D. C. Drehmel
Particulate Technology Branch
Industrial Environmental Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, N. C. 27711
and
R. Parker and S. Calvert
Air Pollution Technology, Inc.
4901 Morena Blvd.
San Diego, California 92117
(This document has been reviewed and approved for
publication by the U.S. EPA.)
To Be Presented at the
Third US/USSR Symposium on
Particulate Control
Moscow, USSR
September 1979
199
-------
Introduction
Granular bed filters and dry scrubbers may be defined as any collection
system comprised of stationary or moving discrete, relatively closely
packed granules as the collection medium. With respect to motion of the
granules, granular bed filters may be classified as moving or fixed bed
filters. Dry scrubbers may be very similar to moving bed filters except
that the gas stream may be accelerated before contacting the granules in
order to maximize collection from impact!on.
A moving bed filter is usually arranged in a cross-flow configuration.
The bed is a vertical layer of granular material held in place by louvered
walls. The gas passes horizontally through the granular layer while the
granules and collected dust move continuously downward and are removed
from the bottom. The dust is separated from the granules by mechanical
vibration. The cleaned granules are then returned to the overhead
hopper and the panel by a granule recirculation system.
CPC Dry Scrubbe*-
The Combustion Power Company's "dry scrubber" is an example of a
continuously moving bed filter. The granular bed material flows downward
between two concentric cylinders. The gas passes through the bed and is
filtered by the granules. The granules are recycled pneumatically and
the collected dust particles are disengaged from the granules and sent
to a conventional baghouse.
The performance of this device has been reported by Wade, et al.
(1978). They conducted extensive cold flow tests to investigate the
effects of bed depth, granule diameter, and other parameters on the
collection efficiency. Particulate loadings ranged from 0.46 to 4.6
g/Nm (0.2 to 2.0 gr/scf). The superficial gas velocity was varied from
200
-------
20 to 80 cm/s (40 to 160 ft/min). The medium flow rate was varied from
0.4 to 1.6 kg medium/kg air. Pressure drop ranged from 1.2 to 5.7 kPa
(5 to 23 in. W.C.).
Unfortunately, the fractional efficiency can be low for submicron
size particles. However, the CPC moving bed filter is capable of
particulate removal efficiencies in excess of 98% for particles larger
than 1 pmA in diameter. Submicron particles can be collected at an
efficiency in excess of 90% only in cases with high velocities, high
loadings, and low collection media recirculation rates. Beds with larger
thickness-to-granule-diameter ratios were most effective in the capture
and retention of particles in the 2 to 5 ymA diameter range. Also,
Intermittent media movement was shown to improve efficiency by a few
percent. However, the economics of this operational technique have not
been analyzed.
High temperature tests of the moving bed filter are planned. No
high temperature data are available at this time.
The major advantage of the moving bed filter design is that the bed
granules are removed and cleaned out of the primary gas stream. This
enables efficient cleaning and a relatively steady collection efficiency.
Also, it is not necessary to isolate filter units during cleaning so
that the total filter area open to gas flow is available for filtration
at any time.
The moving bed design also has some limiting operating characteristics.
The granule recirculation system adds significantly to the operating
cost. Particle reentrainment caused by the relative motion of the granules
limits the granule flow rate and affects the overall collection efficiency.
Erosion of the retaining grids, louvers, and transport system components
may be a problem, especially in high temperature and pressure systems.
The collected dust particles cannot form a filter cake so that the
201
-------
operating efficiency will be essentially that of a clean bed. Temperature
losses may be large and will be proportional to the heat capacity and
temperature of the recirculated granules.
It may be possible to resolve most, if not all, of these problems
through further development and testing. Performance data at high
temperatures and pressures will be important in identifying the most
serious operational problems.
APT Dry Scrubber
EPA is developing through its contractor, Air Pollution Technology,
Inc., a dry scrubber system which can be used for high temperature and
pressure gas cleaning. This system has been reported by Calvert et al.
(1977) and Patterson, et al. (1978).
This system is somewhat similar to a venturi scrubber system in
that it uses relatively large particles as collection centers for the
fine particles in the gas stream. The principal advantage with this
system is that it maximizes the collection efficiency of individual
collector particles and thereby reduces the number of collectors that
need to be cleaned and recycled.
The collector particles introduced to the gas stream collect fine
particles by mechanisms such as diffusion, inertia! impaction, interception,
and electrophoresis. The larger size of the collector particles allows
easy separation from the gas stream by methods such as cyclones and
gravitational settling.
The first step in the particle collection process involves introducing
the collectors to the gas stream. This can involve pneumatic or mechanical
injection. The second step involves contacting the collectors with the
gas in such a way as to encourage the movement of the fine particles
202
-------
toward the collectors. A venturi device can be used for the contactor,
in which case the system would be analogous to a venturi scrubber (with
solid collectors instead of liquid drops).
The next process step is to remove the collector particles after
they have captured the fine particles initially present in the gas.
This is accomplished by using the large size and mass of the collector
particles to separate them from the gas. A cyclone separator, gravity
settler, or virtual impactor could be used for this step. Two streams
leave the separator: 1) the cleaned gas stream; and 2) the flow of
collector particles to the next step. The final process involves either
discarding the collector particles or cleaning them for recycle and
disposing of the particulate matter collected from the gas stream.
The particle collection efficiency and pressure drop for an APT dry
scrubber with co-current flow can be predicted with the same relationships
that define co-current wet scrubber performance. The theoretical performance
of the APT dry scrubber has been determined based on the venturi scrubber
•
performance model of Yung, et al. (1977). The predicted efficiencies at
20°C and 820°C are similar but slightly lower at 820°C. Efficiencies
will be somewhat lower at high pressure for the same pressure drop.
Experimental work has been done to determine fine particle collection
efficiency in an APT dry scrubber in order to confirm the predictions
obtained from available mathematical models. A dibutylphthalate (DBP)
aerosol was used in collection efficiency experiments with 100 \an mean
diameter sand as collector particles* The DBP aerosol had a mass median
aerodynamic diameter of 1.3 ymA and a geometric standard deviation of
2.0.
The prediction compares well with the experimental curve. Higher
collection efficiencies can be achieved using denser collector particles.
203
-------
For this reason experiments were also conducted with 125 jum nickel
beads.
Particle penetration data for all runs with nickel and sand collectors
were calculated in terms of the 50% cut diameter as a function of gas
pressure drop. These data were found to fall on the line which represents
the best available relationship for industrial scale wet venturi scrubbers.
Therefore, the APT dry scrubber follows the same primary collection
efficiency/power relationship as wet venturi scrubbers.
The overall efficiency of the APT dry scrubber will depend on the
reentrainment characteristics of the specific system configuration in
addition to the primary collection efficiency. Particle and collector
properties, system geometry, flow rates, and other parameters will
influence reentrainment.
APT has built an atmospheric fluidized bed coal combustor which
will be used for testing a pilot plant dry scrubber at high temperature.
Electrostatic augmentation is also being investigated as a means of
increasing the collection efficiency independently of pressure drop, and
possibly improving the adhesion of fine particles to collector particles.
Fixed Bed Filters
Fixed bed filters operate in two modes: filtration and cleaning. During
filtration the bed is stationary. The gas passes through the bed and
collected particles are deposited within the bed and on the bed surface.
During cleaning the bed is isolated from the main flow and agitated
mechanically or pneumatically by a reverse flow of gas.
Currently two fixed bed devices are being developed: the Rexnord
gravel bed filter and the Ducon granular bed filter. The Rexnord
204
-------
filter uses a rake-shaped stirring device to agitate the bed during
cleaning. This loosens the filter cake which is then removed by a reverse
flow of clean air.
The Ducon granular bed filter cleans the bed by a reverse flow of
gas which fluidizes the bed and elutriates the fine collected particles.
The Ducon filter was tested on the effluent from a fluid bed catalytic
cracking unit regenerator at an oil refinery (Kalen and Zeng, 1973).
The gas was at 370°C to 480°C and 1 to 1.5 atm. The dust loading ranged
from 0.34 to 1.94 g/m (0.15 to 0.85 gr/acf). A collection efficiency
of 85-98% was obtained on dust with a mass median diameter of 35 urn and
a geometric standard deviation of about 4.
A high temperature and pressure design of the Ducon filter was
tested at the Exxon Miniplant (Hoke, et al., 1978). Initially severe
plugging of the bed retaining grids was encountered. This problem was
resolved by eliminating the grids and redesigning the bed housing to
provide sufficient freeboard above the bed to allow cleaning of the bed
without loss of bed material.
A number of operating problems were encountered during the Exxon
tests of the modified Ducon filter. The lowest demonstrated particulate
outlet concentration was 68.6 mg/Nm (0.03 gr/scf) which was considered
to be too high to protect a gas turbine and borderline for meeting
current emissions regulations. The use of smaller filter media could
improve efficiency. However, at times the filtration efficiency was
very poor and the outlet particulate'concentrations were as high as 700
to 1,200 mg/Nm3 (0.3 to 0.5 gr/scf).
It was also observed that the efficiency decreased with time in the
longer runs, dropping from 90% initially to about 50% later in the run.
205
-------
Loss of filter medium during blowback was another recurring problem.
Further attempts were made to use 50 mesh retaining screens, but they
failed because of plugging. Additional tests with 10 mesh screens also
resulted in. significant screen plugging.
A significant buildup of particles in the filter beds was also
observed amounting to about 30* of the weight of the filter medium. A
possible steady long term increase in filter pressure drop may result
because of this. However, no significant increase in filter pressure
drop was noted during any of the shakedown runs.
It was also observed that the particles were not only building up
in the beds, but were uniformly mixed with the filter medium. It is
possible that the buildup and mixing of particles in the bed could be
responsible for the increase in the particle concentration in the outlet
gas with time.
Another potential problem with the current design was its vulnerability
to upsets. When upsets occurred, such as bed plugging or loss of filter
medium, the operating problems caused by such upsets required shutdown
of the system. Another problem which may be unique to the Mini-plant
was the interaction of the granular bed filter with the rest of the FBC
system during the blowback cycle. An increase in system pressure was
noted during blowback, resulting in problems with the coal feed system
which is controlled by the differential pressure between the coal feed
vessel and combustor. This required modifications to the coal feed
•
control system to minimize the effects.
Granular bed filter performance data for all runs give a range of
efficiency from 31 to 97*. The efficiencies are based on an inlet
concentration of 2.3 g/Nm (1.0 gr/scf) which is the average for the
206
-------
emissions from the secondary cyclone. The fractional efficiency decreased
rapidly with decreasing particle size below 1 urn.
The granular bed filter test program was suspended in November 1977.
In all runs in which more than one outlet concentration was measured, it
was observed that the outlet concentration increased with time. Exxon
was not able to demonstrate that the current EPA emission standard (0.1
lb/10 Btu or 0.05 gr/scf) could be met for more than a few hours of
operation. In no run was the anticipated new standard (0.05 lb/10 Btu
or approximately 0.025 gr/scf) satisfied.
Discussion
The three devices discussed in this paper give a range of design
features available for collection of particulate on hard granules (see
Table 1). At one end of the spectrum are fixed granular bed filters
which rely on collection throughout the bed material until a layer or
cake is formed. The cake provides greater filtration efficiencies
especially for submicron particles. Optimizing the performance of the%
fixed granular bed filter requires a cleaning system which preserves
some of the cake while preventing unacceptably high pressure drops.
Moreover, it is essential that the cleaning systems not allow particulate
to work its way through the bed either by insufficient cleaning or by
motion of the granules.
At the other end of the spectrum is the dry scrubber which relies
only on impaction for collection and a cake is never formed. In this
case, the gas velocity is high to optimize impaction. Collection of
submicron particles may be augmented by charging particles and granules
with different signs.
207
-------
Table 1. Summary of Performance Information
CPC Dry
Scrubber
Device
APT Dry
Scrubber
Ducon
Filter
Superficial gas velocity, cm/s
Pressure drop, KPa
Bed depth, on
Granule diameter, cm
Efficiency at 1
Efficiency at 6 pro, %
20-80
1.2-5.7
20-40
0.08-0.2
78
93
3,000-6,000
2-7
NA
0.01
96
99
45
8(from predic
3.8
0.04
82
96
208
-------
In the middle Is the dry scrubber in which the granules move slowly
enough to form a bed which acts like a filter. This device works partially
by impaction and partially by filtration but cannot be optimized for
either since the gas velocities are low and the bed is removed before a
cake is formed. However, some designs using intermittently moving
granular beds do establish and preserve a cake for better filtration.
The information in Table 1 is a quick comparison of the points in
the spectrum discussed above. Pressure drops for all three devices tend
to be the same. The APT dry scrubber has a very high gas velocity and
no bed. The CPC dry scrubber has a low gas velocity and a thick bed.
The Ducon filter has a low gas velocity and a thin bed. The thick bed
is used with the CPC dry scrubber to ensure good filtration in the
absence of cake filtration or high gas velocity for impaction collection.
Based on these data, the APT dry scrubber gave the best performance.
Conclusions
A spectrum of particulate collecting devices has been discussed in
terms of three examples under development. Advantages of these differing
approaches are:
APT Dry Scrubber (dry scrubber)
High collection efficiency possible from impaction and electro-
static attraction.
Collecting granules are removed from the system for easy cleaning.
209
-------
Ducon Filter (fixed bed filter)
High collection efficiency possible from cake filtration.
Low attrition rates of bed material.
Minimum energy requirement because of low heat loss and no
bed recirculation power.
CPC Dry Scrubber (moving bed filter)
Collecting granules are removed from the system for easy
cleaning.
Electrostatic augmentation may provide high collection
efficiencies.
Choosing between these alternatives important considerations include
cost, the adhesiveness of the particulate, and the importance of energy
conservation. For example, if the particulate is sticky, a system which
removes granules for easy cleaning will be necessary. On the other
hand, if a fixed bed filter can be cleaned and removal of granules
implies a high energy loss, cost considerations may favor the fixed bed
filter.
210
-------
References
Calvert, S. et al., Paper #111-5, in EPA/DOE Symposium on High
Temperature, High Pressure Paniculate Control (September 1977), EPA-
,600/9-78-004 (NTIS CONF-770970), March 1978.
Hoke, R. C. et al., Miniplant Studies of Pressurized Fluidized-Bed
Coal Combustion: Third Annual Report, EPA 600/7-78-069 (NTIS PB 284534),
April 1978.
Kalen, B. and Zeng, F. A., CEP 69 #5, pg. 67, June 1973.
Patterson, R.G. et al., APT Dry Scrubber for Particulate Collection
at High Temperature and Pressure, Symposium on the Transfer and Utilization
of Particulate Control Technology, Denver, Colorado, July 1978.
Wade, G. et al., Granular Bed Filter Development Program, Report to
DOE, FE-2579-19, April 1978.
Yung et al., Venturi Scrubber Performance Model, EPA 600/2-77-172,
NTIS PB 271-515/9BE, August 1977.
211
-------
I.K. Goryachev, y.v. Abrosimov
FILTERING MATERIAL FOR GAS-CLEANING AND POSSIBLE AREAS
FOR ITS USE
The ever-greater distribution of dust removal of industrial
and exhaust gases from chemical, metallurgical, construction indus-
tries and others takes in use of bag houses supplied with high-ef-
fectiveness filtering materials.
'•From the assortment of filtering material serially produced
in the USSR, approval can be given to material possessing features
for withstanding the extended effects of high temperatures, corro-
sive action of dusts and gases with preservation of highly filtra-
tional, durable and regenerative features.
For cleaning aspirated air in the ventilation systems of
work areas,and air sucked out from transfer operations, transport and
packing of fine jarticulatesf. (inexpensive synthetic materials having
a high dust-trappina coefficient^ -under relatively low hydraulic
resistance are widely used. These materials efficiently trap such
high-dispersion dusts as industrial-grade carbon, zinc oxide, etc., as-
suring sanitary norms for dust emission from exhaust systems (1).
For cleaning industrial gases at temperatures reaching 130°
C, filtering fabrics and non-woven materials made from Lawson ftrane.
note: Soviet equivalent of Dacron) fibers are receiving ever wider use
Lawson fabrics possess high indices for dust-separation, strength
and^regenerative properties. The warp and woof of these fabrics are
made from staple fiber. Experimental work at NIOGAZ's Semibratbvsk
Branchy conducted- on an industrial filter FRKDI-1100 durina the
cleaning of qases from «lectr.tc-*rc steel-smelting furnaces has
shown that the staple Lawson fabric L-3 has a dust-removing effi-
212
-------
ciency of 99% and a high regeneration capacity (2).
Non-woven needled Lawson materials have wide use in scrim
filters with impulse blowers. They are relatively simple to manu-
facture, inexpensive and possess high filtering and strength cha-
racteristics. Thus, for example, in removing lead monoxide dust
with a 3 micron median particle diameter, FRKI-type bag houses,
furnished with needled non-woven material, had an operational
effectiveness for dust removal of 99.8% during the course of a
year and a half, and the dust content of the expelled gases dur-
ing this time did not exceed 4-6mg/m .
Materials made from poliacrylnitryl nitron fibers are used
in many industries. Nitron '." filter ~ bags have withstood tempera-
tures up to 130°C for extended periods. They possess a relatively
high chemical stability.and sufficient durability against wear
and bending. At non-ferrous metallurgy sites, nitron fabric's com-
prise nearly 40% of the total amount of all used filtering materials.
*
Using increased thermal-resistant synthetic fiber filtering ma-
terials (oxalon, sulphon) makes "possible an increase in filter •efficiency
by means qf_reducing_a£r"infl6w_for the cooling of the gases.- The
determine'd "heat resistance limit of such fabrics is 180-200°C.
In cleaning rotary kiln gases in zinc plants, oxalon fabric
provides dust-removal efficiency at 99.5% with the temperature of
the gas being cleaned at 200-210eC (3).
Glass filtering fabrics lare widely used for cleaning gases
with temperatures reaching 250DC in technical-grade carbon produc-
tion, cement plants and in ferrous and non-ferrous metallurgy.
Glass fabrics possess a complex of physico-technological and chemi-
cal properties not characteristic to any fabric of organic fibers.
The basic advantage of glass fabric over other filtering materials
213
-------
is its high thermal resistance and maintenance of its mechanical
strength." Glass fabric works efficiently for a period of 1.5-2
years in FR-5000 type filters with fabric regeneration by back-
blowino in technical-gra^e carbon plants. Glass fabric possesses
however a relatively weak resistance to frequent bending and abra-
sion. Accordingly, glass fabric is not used in scrip-design filters.
In the USSR a special anti-static filtering material has
been developed for bag-houses working in dangerously-explosive
surroundings for removal of dusts that are capable of building up
static electricity.
Testing this antistatic material in industrial working con-
ditions in a chemical-metallurgic plant for a long period showed
its high dust-removal efficiency and reliability in providing
secure filter work.
Asbestos fabrics are sometimes used with filtering fabric-
corrosive hiah-temperature gases. Asbestos fabrics are sufficiently
resistant to acids and alkali and withstand high temperature.
However, due to their low strength, they are used only in low pres-
sure drop conditions.
Different fiber materials are added to asbestos fabrics to
raise their strength. Endurance indices are raised in this case,
but the resistance to increased temperatures is lowered somewhat.
Fabric and non- woven material prepared from needling
methods is used as filtering partitions for cleaning hot gases with
a temperature higher than 220°C.
Thus, for example, serially-manufactured metallic fabric
sieves (4) are widely used for removal of chemical regeants and
particularly pure chemical substances (potassium carbonate, nickle,
cadmium, and the oxides of zirconium, silver, vanadium, molibdenum.
214
-------
tunqsten, etc. ) at exhaust eras temperatures up to 500*C. Basicallv.
fabric nets made from KhlSNlOT-grade steel wire and NMZhMts-2.5-1.5
grade nickle-copper alloy (monel) are being used. KhlSNlOT-grade
austenite chrone-nickle steel is characterized by higher heat resis-
tance, full non-magnetization and good weldability.
For fibers of 0.12 to 0.18mm diameter, intermittent resis-
tance is 5£ = 95mg/mm2, and relative elongation is o « 20%.
NMZhMts28-2.5-1.5-grade nickle-copper alloy (monel) is dis-
tinguished by its high corrosion resistance to acids, alkalis and
qas media.
This alloy is distinguished by its hig'h mechanical properties,
considerable heat resistance, and weak magnetic properties.
It is necessary to note that fabric fiber screen made by
interweaving wire warp with v oof •without mesh formation in the beginning
of the filtration process is inferior in dust-trapping efficiency to
natural-fiber fabrics.
However, metallic woven net is distinguished by high heat-
resistance and corrosion resitance, and provides the very same de-
gree of trapping after an elementary dust layer has formed on its
surface as natural fiber fabrics.
On the basis of experience in industrial operation of filters
furnished with metallic net fabric for trapping potassium superoxide,
and the oxides of nolibdenum and tungsten, a type-dimensional series of filters
for cleaning gases up to a temperature of 500°C was developed.
These filters regenerate themselves by section by compressed
air blasts (from 1.5 to 2 kg/cm2).
non-woven filtering materials lacking a scrim base are current-
ly being developed and tested in the USSR. These materials are, by
tlheir strength characteristics, slightly inferior to needled
215
-------
materials with scrim bases but in their filtering characteristics,.
are not inferior to the best foreign models. The cost of producing
needled filtering materials lacking scrims is several times lower than
scrim material.
New technology introducing non-woven filter material lack-
ing scrims simultaneously allowed solving the problem of producing
high-temperature filtering materials from oxalon fibers with thermal
resistance up to 230°C and production of material from metallic fi-
bers with thermal resistance up to 50'0°C. By its mechanical properties,
felting without scrim from oxalon fibers is not inferior to analogous
materials on a scrim.
Residual dustiness in cement and quartz dusts (dgQ-Smicrons)
3 2
at a specific gas load of 4m /m min. and an intake dust content of
8p/nm3 for felting without scrim did not exceed 7.5mg/nm3. Pelting
regeneration efficiently vorked with compressed air back blowing
(P=1.2kg/cm2).
Needled felting was manufactured in the USSR -from metallic fiber with the
aim of jraking filtering material with a heat-resistance up to 500°C.
Microfiber made from lKhl8N9T-grade steel was used as the raw material.
Comparative stand testing of a sample of metallic fiber and
of Sl26-.12Khl8NlOT State Standard 3187-76 woven wire filter screen
in removing quartz and cement dust (<350=8 microns) has shown that use
of non-woven material instead of nets allows an increase of gas filter
performance by 1.5-2.0 times.
216
-------
1. Ooryachev, I.K., Kazakov, A.K., "Experience in Using Small Dimension
Cellular-Design fabric Filters in Exhaust Lines", in the book Industrial
Gas Cleaning and Aerodynamics of Dust-Trapping Equipment , Yaroslavl, 1975.
pp.118-119.
2. Pozin, L.M., Istomin, V.S., Malgin, A.D., "Cleaning Arc Steel-Smel-
tinq Furnace Gases by FRKDN-llOO-Type "Baq Houses" in" Industrial and Sanitary
Gas Cleaning", 1977, no.5, p.2.
3. Gordon, G.M., Kaplan, V.G., Luk, V.I., "Study in Filterinq Fabric
Made from Highly Thermal-Resistant Synthetic Oxalon Fiber", in "Indus-
trial and Sanitary Gas-Cleaning",1978, no.4, pp.4-5.
4. Abrosimov, Y.V., NIIOGAZ's Scrim Glass Fabric Filters, Mashinostroenie,
1972.
Translation by Carl Gruber, IERL, RTF, N.C., July 1979.
217
-------
HIGH TEMPERATURE CERAMIC FILTERS
by
D. C. Drehmel
Parti oilate Technology Branch
Industrial Environmental Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, N. C. 27711
and
R. Parker and S. Calvert, APT, Inc.
4901 Morena Blvd.
San Diego, California 92117
(This document has been reviewed and approved
for publication by the U. S. EPA.)
To Be Presented at the
Third US/USSR Symposium on
Particulate Control
Moscow, USSR
September, 1979
218
-------
INTRODUCTION
Conventional fabric filters are limited to operating temperatures
below 250°C. The maximum temperature varies with the specific fabric
and is the temperature at which accelerated fabric deterioration or
abrasion occurs.
Glass fiber bags are the most common type used for higher temperature
applications and are limited to about 300°C. The glass fibers are
coated with a silicone, silicone-graphite, or equivalent finish in order
to provide lubrication between the fibers. Unfinished glass fibers
experience extreme abrasion and unsatisfactory bag life. The temperature
limit of glass fiber fabrics is directly related to the temperature
limit of the finish.
In fact, it is generally true that in practice the effect of temperature
and pressure on filtration mechanisms was not a determining factor in
the application of high efficiency filters. The main problems are the
physical and chemical effects of a high temperature environment on the
filter materials. These effects may appear as reduced mechanical strength
and resilience or loss of adhesion, leading to mechanical leakage and
decrease in efficiency.
PREVIOUS WORK
Filtration media for extreme temperatures and pressures have been
investigated by many authors (see Table 1). Silverman and Davidson
(1956) suggested the use of ceramic fibers sandwiched between layers of
woven metallic or ceramic fabrics for filtration at temperatures to
1,100°C and higher. Billings et al. (1955) and Silverman (1962) discussed
the use of metallic fiber "slag/wool" filters for high temperature
219
-------
Table 1. High Temperature Filtration
Investigators
Silver-man (1956)
Davidson
Billings (1955)
Silverman (1962)
First (1956)
Kane (1960)
Drehmel (1977)
Shackleton
Drehmel (1977)
Temperature
Range, C°
to 1100
300-650
760
815
360-815
Media
woven metallic
or ceramic
"slag/wool"
ceramic fiber
ceramic felt
ceramic membrane
Efficiency
Range, %
NA
75-98
82-91
99
81-99
Ci1i berti
220
-------
filtration of open-hearth furnace fume. They found that efficiencies
ranged from 75 to 98% for the slag-wool filter at temperatures from 300
to 650°C and dust loadings from 0.1 to 1.1 g/Nm3. A continuous slag-
wool filter was designed and tested at 300-400°C. The efficiency ranged
from 10 to 80% for Fe203 particles with a mass median diameter of 0.65 urn.
First et al. (1956) and Kane et al. (1960) reported on the use of
ceramic fiber filters capable of withstanding temperatures up to 1,100°C.
First et al. (1956) measured collection efficiencies at 21°C and 760°C.
The mass median particle diameter was 8.5 pm with a particle density of
6.4 g/cm .
A composite filter comprised of fibers from 4 to 20 ym in diameter
gave over 99% collection for all temperatures. The fractional efficiency
for the composite tests is greater than 95% for particles larger than 1 um.
THEORETICAL PREDICTIONS
Filtration theory has been reviewed by many authors including
Davies (1973) and Rich (1966). Using the theory presented by Davies to
predict the collection efficiency of a clean fiber filter at high temperature
and high pressure, the following parameters were assumed:
2
Filter weight = 0.026 g/cm
Fiber density =2.53 g/cm
Fiber diameter = 5.0ym
Superficial velocity = 12 cm/s
3
Particle density =2.5 g/cm
Air properties at:
Temperature = 20 and 1,100°C
Pressure =1 and 15 atm
These parameters simulate a typical aluminum-silicate ceramic
paper.
221
-------
At high temperature and low pressure the Brownian diffusion regime
becomes very significant and the collection efficiency of particles smaller
than approximately 0.5 ym increases dramatically. At high temperature and
high pressure this effect is less apparent.
In the inertial impaction regime, high temperature and high pressure
reduce the collection efficiency from that obtained at standard conditions.
Even at high temperature and pressure, however, the predicted collection
efficiency is effectively 100% for particles larger than 2 ym. This is
consistent with the data reported earlier from the work by First et al. (1956).
RECENT DEVELOPMENT WORK ON BAG FILTRATION
Recent development work on high temperature and pressure ceramic fiber
filters involves three groups of materials:
1. Woven structures - Cloth woven from long-filament yarns
of ceramic fibers.
2. Papers - Ceramic structures produced from short lengths
of fibers, generally held together with binders.
3. Felts - Structures produced to form mats of relatively
long fibers. These materials are known as blankets in
the insulation industry. They tend to be less tightly
packed than conventional felt materials.
Ceramic fiber filters have two major drawbacks at high temperature and
pressure. First, they must be very durable. Conventional filter bags are
expected to last at least 1 year (see Billings and Wilder, 1970). Bag life
at the Nucla Power Plant was estimated to be 5 ±1.3 years. For typical
cleaning pulse frequencies, a bag may have to withstand a few million pulses
in its lifetime. For this reason, blanket or felted ceramic fiber materials
are expected to be the most promising. They combine good filtration properties
with relatively high strength.
222
-------
The second drawback is the size of typical fabric filter installations.
Most conventional fabric filters operate at superficial velocities in
the range of 1 to 3 cm/s. Slightly higher velocities (up to 10 or 20
cm/s) are possible with some felted fabrics although bag life will be
shortened.
In comparison, granular bed filters (GBF) operate at superficial
velocities from 40 to 80 cm/s. For a given gas flow rate, fabric filters
will require from 4 to 20 times the surface area required by granular
bed filters. This is especially important at high pressure where the
cost of the pressure vessel can be a significant fraction of the capital
cost. For greater efficiency, both baghouses and GBF's should be
designed to maximize the surface to volume ratio.
These problems are being considered. Preliminary experience at
high temperature and pressure indicates that at least three configurations
show promise, having survived 50,000 cleaning pulses at 815°C and 9 atm.
Test conditions were:
Temperature - 815°C
Pressure - 930 kPa (9 atm)
Air-to-cloth-ratio - 5 to 1 (2.54 cm/s)
Cleaning pulse pressure - 1,100 kPa
Cleaning pulse interval - 10 s
Cleaning pulse duration - 100 m/s
Dust - redispersed.
The three filter media configurations tested were:
1. Saffil* alumina mat contained between an inside and
an outside layer of 304 stainless steel knit wire
screen.
*ICI United States, Inc., Concord Pike and New Murphy Road, Wilmington, DE 19803
223
-------
2. Woven Fiberfrax* cloth with Nichrome** wire scrim insert.
3. Fiberfrax blanket contained between an inside
and an outside cylinder of 304 SS square mesh
screen similar to common window screen. The
ceramic fiber blanket was held in position
between the screens with 302 SS wire sewn between
the screens.
Pressure drop during the tests was controlled by the rapid cleaning
pulses and in general remained less than about 5 kPa (20 in. W.C.).
Formation and removal of the filter cake for these configurations and
test conditions presented no problems.
The average outlet loading during the Fiberfrax blanket test was
o
0.0055 g/Nm (0.0024 gr/scf). The fly ash dust dispersion apparatus
used was suitable for filter loading tests but may not be representative
of the dust characteristics and size distribution to be encountered in a
real-life application.
•
To obtain test data in a real pressurized, fluidized-bed combustion
application, a ceramic baghouse was tested at the Exxon Miniplant. The
test program has shown that the bags will survive pressure drops as high
as 10 kPa and can be kept clean at air-to-cloth ratios as high as 10 to
1 (5 cm/sec).
WORK ON MEMBRANE FILTRATION
Several available ceramic materials in many configurations have
been evaluated as possible high-temperature filters. One of the most
promising materials tested was the ceramic cross flow monolith, Thermacomb.***
*Carborundum Co., P. 0. Box 808, Niagara Falls, NY 14302
**Driver-Harris Co., Harrison, NJ 07029
***3M, 3M Center, St. Paul, MN 55101
224
-------
This material is composed of alternate layers of corrugations separated
by thin filtering barriers. This type of configuration affords a large
amount of filter surface in a very small volume.
The Thermacomb cross flow structure is made up of several layers in
the following pattern: a thin (0.25 - 1.5 mm) porous cordierite sheet,
a layer of cordierite corrugations similar in appearance to those used
in cardboard, another flat sheet of cordierite followed by another layer
of corrugations oriented 90% from the corrugations below. Now available
forms of the material have 1.97, 3.15, or 4.72 corrugations per cm. A
similar material is manufactured by W. R. Grace & Company.* This
material has perpendicular dividers with rectangular holes, rather than
the triangular holes of Thermacomb. The Grace material tested had
approximately 8.5 holes per cm and an equal number of layers per cm.
These materials have many properties that make them attractive as
filters. Among these are (1) working temperature to 1,200°C, (2) very
good mechanical strength despite thin separators, (3) excellent resistance
to thermal shock, (4) excellent resistance to corrosive atmospheres and,
(5) very high surface area to volume ratios.
Efficiency of filtration was determined with a limestone test dust.
The mass median diameter was typically 1.4ym and the geometric standard
deviation was 3.0.
Results of the high temperature Thermacomb tests show that the
overall collection efficiency averaged 96.4%. No problems were encountered
in cleaning the filter media by reverse pulses of compressed air. It
was possible to clean the filter and return to a stable pressure drop
even with relatively heavy dust loadings.
*W. R. Grace and Co., 1114 Avenue of the Americas, New York, NY 10036
225
-------
Similar results were obtained in a limited number of tests on the
W. R. Grace material.
CONCLUSIONS
A number of investigators have shown that high collection efficiencies
are possible with ceramic filtration at high temperature (greater than
800°C) and/or high pressure (greater than 900 kPa). The recent work has
tested both ceramic bags and rigid ceramic filters. Advantages of these
differing approaches are:
Ceramic Bag Rigid Ceramic Filter
High collection efficiency High collection efficiency
Easy to clean Compact
Resists failure because of Resists failure because of
thermal shock high pressure drop
The endurance of both media is unknown. If the rigid ceramic filter can
be used without clogging or thermal shock, it could be maintained in
service for many years. However, ceramic bags are expected to have a
limited life because of the less durable nature of the bag structure.
Tests to date indicate that ceramic bags will easily survive up to 50,000
cleaning pulses or the equivalent of 1 year's light service.
REFERENCES
Billings, C.E. and Wilder, J. E., Handbook of Fabric Filter Technology,
Volume I, EPA No. APTD 0690, NTIS No. PB 200-648, December 1970.
Billings, C. E., et al., JAPCA 5 #3 p. 159, November 1955.
226
-------
Davies, C. N., Air Filtration. Academic Press, New York, 1973.
Drehmel, D. C. and Ciliberti, D. F., Paper #77-32.4, APCA Annual
Meeting, Toronto, June 1977.
Drehmel, D. C. and Shackleton, M.S., Paper #17, Third Symposium on
Fabric Filters for Particulate Control, Tucson, December 1977.
First, et al., I&EC Chem 48 #4. p. 696, 1956.
Kane, L. J. et al., Bureau of Mines Report 5672, 1960.
Pich, J., In Aerosol Science, C. N. Davies, ed., Academic Press,
New York 1966.
Silverman, L. and Davidson, R. A., U. S. Patent # 2,758,671, August
1956.
Silverman, L., U. S. Patent #3, 063,216, November 1962.
227
-------
N.G. Bulgakova, D.L. Zelikson
METHODS AND EQUIPMENT FOR MEASURING OF FINELY DISPERSED
PARTICLE SIZE LESS THAN ^MICRONS FOR INDUSTRIAL GASES
The task of measuring size of finely-dispersed particles less than
ftjmicrons is at the present time very urgent. This is in connection
with the fact that such particles are usually poorly removed by dust-
cleaning equipment and are emitted into the atmosphere along with exhaust
gases. For example, in gases from electric power stations that burn
coal, particle content after precipitators reaches 20% for particles
less than a micron in diameter and as much as 5-102 of those less than
0.3 microns, while in ferrol alloy production, content of such particles
is 70% after cleaning by wet scrubbers. As recent research has shown,
it is namely these particle-fractions that are a serious source of air
pollution in connection with their peculiar toxicity. Therefore from
the point of view of increasing the effeciency finely-dispersed particle
removal, abating their emission into the atmosphere, and selection and
evaluation of the work of dust-control equipment, the development of
methods and equipment for measuring sub-micron particles provides further
technical progress in the area of protection and renewal of the environment.
In as much as the indicated sub-micron diameter range of particles
is close to the operative limits of pitical, acoustical, aeromechanical,
etc. measuring equipment, choosing the method of measurement presents
known difficulties.
Out of the methods known for sizing sub-micron particles, the
following should be noted.
228
-------
Electron microscope method. Based on the visual study of individual '
particles. The measurement process is time-consuming, and requires use
of high-cost and operationally complex equipment. The method does not
ensure automation of the measurement process.
Electrical methods. Based on registering the electrical charge of
the particles and their separation in electrical, magnetic or electromagnetic
fields. The methods are sufficiently simple and allow quickly obtaining
results and automation of the measurement process. The basic shortcoming
of these methods lies in the necessity of preliminarily preparing the
gas sample in the majority of industrial discharges, i.e. to cool and
lffa
rar4fy the gas. The equipment for sample preparation is often more
complex than the measuring unit.
This method has been fulfilled in the USA in the particle electrical
mobility analyzer unit.
Diffusion methods. Based on particle separation under the influence
of diffusion forces with aerosol'Movement in a laminar flow. This
method is implemented in a unit-incorporating diffusion batteries of
various design with condensation nuclei counters. This method may be
used during variation in particle size distribution in air aerosol flow,
but requires preliminary sample preparation.
Inertia! methods. Based on particle separation from aerosol stream
by the Tatter's flow around the settling surfaces. This method is the
simplest and most operative. It allows chemical sample analysis when
needed. Direct removal of the sample in the gas conduit excludes the
need of preliminary sample preparation.
229
-------
The method is accomplished by a cascade-impactor, which is widely
used for dispersion analysis of particles of 0.3-30 microns.
Theoretical and experimental research (1) has shown the possibility
of using cascade impactors for analysis of sub-micron particles in a
size range of 0.01-0.3 microns. For this, it is necessary in the unit's
last stages to form a sufficiently low absolute pressure of 10-lOOmm Hg
(10-10 pascals) to assure sub-micronic particle settling through an
adjusted Cunningham coefficient (1).
Cascade impactor calculation is based on known theoretical premises
(2).
The main calculation formula links diameter dgQ of the particles
(with 50% settling efficiency in each stage of the impactor) with nozzle
diameter D; gas flow rate and viscosityl/Tk; particle density^, and
with Cunningham's correction alsoQf)«
• J
The latter is described by the following expression:
where: '] is the average length of the free run of gas molecules, which
is determined by gas pressure.
230
-------
It follows from considering these expressions, that by reducing the '
range of analyzed particles to the sub-micron level is accomplished by
varying the three parameters, l/^, J> and Cft) .
The increase of gas velocity ( ) is limited to the speed of sound.
Nozzle diameter reduction is limited by the technological capability of
producing aperture smaller than O.lmm in diameter. Aside from this, the
theory of inertia! particle settling derives from the viscosity condition
of the gas flow in the nozzle, but with a decrease in the nozzle diameter
and increase in the Cunningham correction, i.e. reduction in pressure,
the length, of free run becomes commensurable with the diameter of the
flow and the flow regime crosses into the molecular.
Increasing the Cunningham correction is possible increasing the
average molecule run length, i.e. decreasing absolute pressure. With
7.6mm Hg pressure, the correction C(p) comprises more than 2000 for
particles less than 0.01 microns.
In this manner the possibility in principle for the use of cascade
impactors for analyzing sub-micronic particles is obtained. Several
such impactors are known, such as, for example, that developed in the
USA 0).
At NIIOGAZ a eight-stage sub-micron cascade impactor model (fig.l)
has been produced with the following characteristics:
- least nozzle diameter - 0.2mm.
- greatest flow velocity in nozzles - 200m/s
- lowest pressure in final stages - 20mm Hg
- gas flow - 5 lit./min.
- unit dimensions - 0 65mm; 150mm length
Parameters of the sub-micron cascade impactor are given in table 1.
231
-------
Table 1
°*j
staaes j
]
i
I
2
3
4
5
6
7
8
nozzle - ji
aperture :
diameters:
in • •«
Iran's ;
10
0,8
0,8 '
0,8
0.2
0.2
0,2
0.2
lumber
of.
ipertures
I
4
4
4
18
15
23
44
j
gas exhaust j
velocity, J
m / sec. j
6,4
40
40
40
150
200
200
200
static pressure
in the stages
A D mm Hg
760
750
740
730
620
420
220
20
t
: value of
! Q 50>mic'
1 •
i
5
0,950
0,950
0,950
0,250
0,125
0.050
0,010
The table shows values for particle diameter with regard to the
Cunningham correction.
By design, the impactor is made in the form of diaphragm-stages
with 60mm diameter by 7mm thick and are connected by four pins. The
nozzles are made in the form of af.rture in a stainless steel plate of a
thickness of 0.3mm and a diameter of 8mm. The plates are hermetically
glued-in to the diaphragm. Filtering material or lubricant serves as
the precipitating surface in the first four stages, and in the rest of
the stages, pieces of optical glass K-8 (0 10x5), are placed at a distance
of 1mm from the nozzles. An exterior view of a separate diaghragm-stage
is presented in fig. 2. The glass precipitation surface may be covered
by a special lubricant or may be used for measuring settlement by microscopic
methods.
The cascade impactor was tested on a laboratory stand, the scheme
of which is shown in fig. 3a and 3. The stand consists of an aerosol
generator 1,2, drying system 3,4, and measuring and regulating units 5-
232
-------
9. Air was bubbled through a solution of methylene blue with subsequent
drying of the aerosol in the air stream served as the aerosol generator.
The compressed air is.fed into the atomizer (1), at the end of which
there is a thin-walled head-piece with 0.5mm apertures. The aerosol
formed, bubbling through the methylene blue solution, enters the drying
column (3,4), filled with silica jel impregnated with calcium chloride,
where it is dryed and leaves the measuring system. Under these conditions,
the aerosol contains solid methylene blue particles.
An AFA filter was used for measuring the discharged aerosol's
concentration. The particle dispersion composition was controlled with
the aid of a diffusion battery (7). A submicron impactor (6) measured
particle size. Control and measurement of the air flow rate was done by
the valves C9} and rheometers (8).
Aerosol with an average methylene blue particle size of 0.3-0.8
microns was produced on the stand. Table 2 shows results of one of the
analyses conducted by the eight---age cascade impactor. The number of
particles settled on the diaphagm-stages was determined by weight. The
flow rate of air through the impactor was 5 lit./min.
233
-------
Fig.l. General view of a submicron cascade impacter
234
-------
Fig.2. Diaphragm-stages Cmatchbox shows scales)
~ig.3a. General view of the laboratory stand.
235
-------
Fig.3. Schematic of laboratory stand.
-------
Table. 2.
1 .
NO. <
of !'
stages {
•
I
2
3
4
5
6
7
8
filter
amount of
gained, mg.
-
0.25
0,15
Or25
0,40
0.15
0,10
0,10
-
1,40
i i
*, % share '
; !
1 !
1 1
! !
-
17,86
10. 71
17.86
28,58
10,71
7,14
7.14
-
j total content %
I
1
1
_
17*86
28,57
46,43
75,01
85.72
92,86
100
I ,
! Ct 50' micl
*
1
i
5
0,95
0.95
0,95
0,25
0,125
0,05
04OI
According to this table's data, the distribution of methylene blue
particles by size is plotted in a probability-logarithmic coordinate
grid (fig. 4). As can be seen from the figure, the average particle
size 0.71 microns.
In analyzing the results taken here, it can be concluded that
cascade impactors may be used to analyze submicron particles down to
0.01 microns.
237
-------
sc-
CO
70
tt>
50-
20
'/0
d
microns
0.6 1.0
Fig.4. Methylene ^lue dispersion composition, as
taken by the submicron impactor.
238
-------
PARTICLE MEASUREMENT IN THE U.S.A.
W. B. Smith
K. M. Gushing
Southern Research'Institute
2000 Ninth Avenue South
Birmingham, Alabama 35205
D. B. Harris
Industrial Environmental Research Laboratory
Environmental Protection Agency
Research Triangle Park, N.C. 27711
(This document has been reviewed and approved for
publication by the TJ. S. EPA.)
239
-------
PARTICLE MEASUREMENT IN THE USA
INTRODUCTION
Measurements are made to characterize the emissions from sources in order
to obtain the data needed to assess the impact on health and the environment,
to develop and enforce emission regulations, and to develop control strategies.
For airborne particulate emissions, the required data include the size distribu-
tion, concentration, and composition of the particles.
Methods are described in this paper for measuring the particulate emis-
sions from stationary sources. Measurements made inside ducts or stacks are
referred to as instack or in situ measurements. The fraction of the particles
emitted from a process that escape directly into the atmosphere (unducted) are
called "fugitive" emissions and must also be measured. If measurements are
made in the atmosphere around an industrialized area where more than one source
contributes to the atmospheric burden, the measurements are termed "ambient"
measurements. Methods used in the USA to measure particle size and mass in
situ, in fugitive emissions, and in the ambient are listed in Table 1.
Manual methods for sampling in situ, fugitive and ambient aerosols are
discussed in this paper. A second paper includes a discussion of automatic
systems for particle sizing.
SAMPLING IN PROCESS STREAMS - IN SITU SAMPLING
MASS CONCENTRATION AND MASS EMISSION RATE
Gravimetric methods using filtration for obtaining samples are used for
determinations of mass concentration and mass emission rates. Samples are
obtained by withdrawing measured gas volumes from the process stream for equal
240
-------
TABLE 1. METHODS AVAILABLE FOR MEASURING PARTICULATE SIZE AND MASS
MASS
SIZE
AMBIENT
Filters
Hi-Vol
Others
Piezoelectric
Nephelometer
8-gauge
FUGITIVE
Filters
Hi-Vol
Others
Piezoelectric
Nephelometer
B-gauge
Lidar
STACK SAMPLING
Filters
Charge transfer
8-gauge
Peizoelectric
Optical
Transmissometers
Lidar
Impactors
Hi-Vol
Mega Vol
Others
Dichotomous Sampler
Diffusion Batteries
Electrical Mobility Analyzers
Optical Counter
Piezoelectric
Impactor
Centrifuge
Cyclones
Centrifuge
Filters
Hi-Vol
Mega Vol
Others
Dichotomous Sampler
Diffusion Batteries
Electrical Mobility Analyzers
Optical Counter
Piezoelectric
Impactor
Centrifuge
Cyclones
Fugitive Assessment
Sampling Train (FAST)
Centrifuge
Elutriators
Impactors
Cyclones
Optical
Electrical Mobility Analyzers
Diffusion Batteries
Surface lonization
241
-------
time intervals at a series of points across the duct. These traverse points
are located at the centers of zones which divide the duct into a number of
zones of equal area. The number of zones into which the duct is divided de-
pends on the distances to upstream and downstream flow disturbances and the
duct dimensions. All samples are taken under isokinetic conditions.
One of two general types of sampling systems is used for these determina-
tions depending on the test conditions and the purposes for which the data are
being acquired. The simpler system, illustrated in Figure 1, utilizes a filter
holder and attached sampling nozzle which are introduced directly into the gas
stream. The filtration medium may be a flat glass fiber filter, a glass fiber
thimble or an alundum (ceramic) thimble. The choice depends on gas tempera-
ture, the concentration of the particulate matter, and the intended use of the
data. A pitot tube and thermocouple located adjacent to the sampling nozzle
make it possible to continuously monitor the gas velocity and permit correction
of the sample flowrate as required to maintain isokinetic sampling conditons.
On those occasions when the velocity field within the duct is stable over long
periods of time separate traverses may be made for determination of velocity
and dust sampling, in which case the pitot tube on the sampling probe may be
omitted.
Testing for compliance with air pollution regulations is generally done
with the system shown in Figure 2. This system, known as the U.S. Environmental
Protection Agency's Method 5 sampling train, uses a nozzle and heated probe to
withdraw the sample from the duct to a filter located within a temperature
controlled oven.
The filter may be followed by either a condenser and moisture absorbent
or by a series of impingers and an absorbent. A leakless pump is used to pro-
vide the necessary suction downstream of which are located a dry gas meter and
an orifice-type flow meter. The latter is used for monitoring and setting the
sampling rate during a test. This system provides for simultaneous determina-
tions of volumetric gas flows within the duct being tested, mass emission rate,
mass concentration, and moisture content. The use of appropriate reagents in
242
-------
NJ
•P-
GLASS FIBER THIMBLE FILTER
HOLDER AND PROBE(HEATED)
SAMPLING
NOZZLE
V^
U^'
LJ
"^-^ ) ) 1
CONDENSER
DRYER
CHECK
WAI WC
-f>~-
DRY TEST METER
VACUUM GAGE
AIR-TIGHT PUMP
Figure 1. American Society For Testing and Materials' Particulate
Sampling Train.
-------
IMPINGER TRAIN OPTIONAL:
MAY BE REPLACED BY AN
EQUIVALENT CONDENSER
HEATED PROBE
AREA FILTER HOLDER
THERMOMETER
CHECK
VALVE
MANOMETER DRY TEST METER AIR TIGHT PUMP
Figure 2. U.S. Environmental Protection Agency's Method 5 Particulate
Sampling Train.
244
-------
the impinger portion of the train permits simultaneous determination of one
or more vapor phase components of the flue gas as well (e.g., SO3, SOa). The
probe connecting the nozzle and filter housing must be glass lined or, if longer
than 2.5m,.may be stainless steel or Incalloy. With the exception of tests of
fossil fuel utility boilers the temperatures may be raised to 165°C to prevent
or minimize the addition of sulfuric acid condensate to the particulate catch.
Sampling rates range from 20 &pm to 200 &pm.
PARTICLE SIZE DISTRIBUTIONS
The methods most widely used in the U.S.A. for the determination of par-
ticle size distributions of industrial flue gas particulate matter are based
on inertial separation. The separators include cascade impactors, cyclone
separators, and other forms of air elutriators, including the Bahco "micro-
particle classifier." Because of the lack of resolution in the fine particle
end of the size spectrum and the problems inherent in the redispersion of
collected aerosols, the use of laboratory devices such as the Bahco are gener-
ally being discarded in favor of other devices which perform the size separa-
tion during the sampling operation.
The cascade impactor is the device most commonly used at this time for
determining particle size distributions. In the devices, an aerosol sample
is withdrawn isokinetically from the gas stream and passed sequentially
through a series of impingement stages. Each succeeding stage operates at
higher jet velocities than the previous stage thereby removing successively
smaller particles at each stage. Determinations of the amount of particulate
matter collected on each stage is then made either by gravimetric or chemical
analysis. The particles size range spanned by an impactor is usually from
about 0.50 ym to 10 jam. The number of stages ranges from about 6 to 13 and
flowrates range from about 0.6 Aprn to 30 £pm. Prototypes of impactors pro-
viding size distribution information to diameters as small as 0.02 ym have
been constructed and successfully tested at industrial sites. Figures 3 and
4 show two of the more popular conventional (0.5 ym to 10 ym) impactors used
in the U.S.A.
245
-------
NOZZLE
T
t
PRECOLLECTION
CYCLONE
JET STAGE
(7 TOTAL)
COLLECTION
PLATE
SPRING
0700-14.1
Figure 3. Modified Brink Model BMS-11 Cascade Impactor.
246
-------
JET STAGE O-RING
COLLECTION PLATE
INLET
7 \
FILTER HOLDER
COLLECTION
PLATE (7 TOTAL)
JET STAGE
(7 TOTAL)
0700-14.2
Figure 4. University of Washington Mark III Source Test Cascade
Impactor.
247
-------
The cyclone separator is a second type of inertial separator which is
in fairly wide use at the present time. Small cyclones have been demonstrated
to provide rather sharp particle size selection characteristics as illustrated
in Figure 5. This shows calibration data obtained using monodisperse aerosols
for a "cascade cyclone" designed and constructed by Southern Research Institute
for the U.S. Environmental Protection Agency. The system operates at a nominal
flowrate of 28 Apm and provides particle size information from about 0.3 ym to
7.0 ym. Figure 6 illustrates the cyclone system. Cyclonic type collectors are
particularly useful when sampling gas streams having high dust concentrations,
which tend to lead to rapid overloading of impactors, and when large samples
are needed for chemical or toxicological studies. They may also be a valuable
tool for determining size distributions in high temperature process streams
(600°C to 1100°C) such as those associated with coal gasifiers and fluidized
bed combustors for combined cycle electrical generating stations. Cascade im-
pactors may not be suitable for high temperature studies because of a lack of
a suitable substrate on which to collect the particles. Small cyclonic sepa-
rators are also frequently used as precollectors upstream of cascade impactors
to eliminate rapid overloading of the first impaction stage. Figure 3 shows
one application of a cyclone precollector.
A three-stage "cascade cyclone" operating at a flowrate of 140 s£pm is in
routine use today as a part of the Source Assessment Sampling System (SASS).
The SASS, illustrated in Figure 7, is used for providing semiquantitative
information on the form, nature, and composition of pollutants emitted from
industrial processes for the purpose of assessing their potential impact on
the environment. The system provides size segregated samples of particulate
matter in four size ranges (greater than 10 ym, 3 ym to 10 ym, 1 ym to 3 ym,
and smaller than 1 ym), usually in sufficient quantities for chemical analysis
and for toxicological studies to determine their potential impact on human
health and terrestrial and aquatic ecosystems. In addition to the particulate
samples, the system also provides samples of organic and inorganic vapors for
similar analyses.
Optical and electron microscopy are also used occasionally for determining
particle size distributions, but difficulties in obtaining unagglomerated
248
-------
100 , . „, .
I I pi I I
90
80
> 70L-
1 60
S 50
I -
lit
8 ,
10
0.2 0.3 0.40.50.60.8 1.0 2 3 4 5 6 8 10 20
PARTICLE DIAMETER, micrometers
• FIRST STAGE CYCLONE
• SECOND STAGE CYCLONE
£ THIRD STAGE CYCLONE
V FOURTH STAGE CYCLONE
O FIFTH STAGE CYCLONE
Figure 5. Laboratory Calibration for the Five Stage Series
Cyclone System. (472 cm3/sec, particle density—1.0
gin/cm ) .
249
-------
CYCLONE 1
CYCLONE 4
CYCLONE 5
CYCLONE 2
CYCLONE 3
OUTLET
INLET NOZZLE
Figure 6. Five Stage Series Cyclone System.
250
-------
ro
ui
I-1
HEATER
CONTROLLER
FILTER
GAS COOLER
CONDENSATE
COLLECTOR
COLLECTOR
DRY GAS METER
ORIFICE METER
CENTRALIZED TEMPERATURE
AND PRESSURE READOUT
CONTROL MODULE
-1*1-
<>
1MPINGER
T.C.
10 CFM VACUUM PUMP
Figure 7. Schematic of the Source Assessment Sampling System.
-------
samples at suitable surface densities on the collection media, while still
collecting a truly representative sample make the use of microscopy infrequent.
Details of particle size distributions from about 0.01 ym to about 0.5 ym
(ultrafine particles) are desired or needed occasionally, but far less frequently
than for the larger particles. Electron microscopy is sometimes used to pro-
vide information on the ultrafine particulate size distribution. Here again,
difficulties in obtaining suitable representative, uncontaminated, unagglomerated
samples make this an infrequently chosen method. The methods most frequently
used for the determination of the concentrations and size distributions of ultra-
fine particles are diffusional and electrical mobility analyses.
Diffusional methods generally use condensation nuclei counters for measure-
ment of particle concentrations and diffustion batteries for providing the
necessary particle size discrimination. The condensation nuclei counter, il-
lustrated in Figure 8, operates by causing a sample of the aerosol being meas-
ured to become supersaturated in the concentration of some condensible vapor—
usually water or alcohol. The supersaturated vapor will condense on any par-
ticles present in the sample having diameters larger than some critical diam-
eter which is determined by the particle solubility and the amount of super-
saturation achieved. This critical diameter is typically about 0.002 urn.
The condensation of the vapor results in the formation of a rather homogeneous,
monodisperse fog containing one fog droplet for every particle in the sample
which had a diameter larger than the critical nucleation diameter. Light
scattering techniques are then used to determine the concentration of the fog
droplets and hence the original concentration of the particles whose diameters
were larger than the critical size. Size distributions are then determined by
measuring the concentrations of a sample gas stream upstream and downstream of
a series of tubes, narrow parallel rectangular channels, or screens, for which
losses resulting from particle diffusion to the internal surfaces can be pre-
dicted as functions of particle size. Figure 9 illustrates a typical dif-
fusion battery of the parallel rectangular channel type and penetration curves
for a particular set of flow conditions. The condensation nuclei counters
require samples which are essentially at ambient pressure and temperature and
252
-------
PHOTO DETECTOR
A
HUMIDIFIER
CHAMBER
LIGHT STOP
ICOMPARATO^ _f I
SAMPLE
& HOLD
-TTIMER 2k
| TRIGGER \- -ITIMER 3(.
OUTPUT
DIGITAL
PANEL
METER
-JCOUNTERJ—* RANGE
3630-249
Figure 8. Diagram of a condensation nuclei counter. After Haberl and
Fusco.
253
-------
Figure 9a. Parallel plate diffusion battery.
20
0.01
PARTICLE DIAMETER, Aim
Figure 9b. Parallel plate diffusion battery penetration curves for
monodisperse aerosols (12 channels, 0.1 x 10 x 48 cm)."*
254
-------
which have much lower particle concentrations than those typically found in
industrial flue gases. Therefore extractive sampling is required followed
by extensive dilution and sample conditioning to remove undesirable conden-
sable vapors. A schematic diagram of a complete system used for such analyses
by Southern Research Institute is shown in Figure 10. Dilutions from 10:-1 to
4000:1 can be provided with the system. Diffusional analyses typically pro-
vide data over the size range from about 0.01 ym to 0.2 ym.
SAMPLING FUGITIVE EMISSIONS
SAMPLING STRATEGIES
Because they are not confined to a duct, or any particular area,
fugitive emissions are more difficult to characterize than ducted emissions.
The boundaries and velocity of the plume must be determined, and temporal
variations and dilution must be taken into account. An important part of
current research and development programs is the development of sampling
strategies.
Measurements of fugitive aerosols may be made near the source, where
essentially no dilution has occurred; in the air surrounding the source, where
only limited dilution has occurred; or upwind and downwind of the source where
dilution is essentially complete. Measurement methods used in each applica-
tion are called quasi-stack, roof monitor, and upwind-downwind, respectively.
Figure 11 is a schematic showing the quasi-stack technique for sampling
fugitive emissions. In this technique, the emissions are captured at their
source in a temporary hood, and transmitted by means of an exhaust blower,
through a duct where conventional, in situ sampling methods are used to
measure the particulate concentration and the velocity of the gas. The emis-
sion rate of the pollutants can be calculated from these data.
The quasi-stack method is the most accurate technique for measuring
fugitive emissions. Almost all of the emissions are captured with little
dilution. Unfortunately, it may be difficult or impossible to apply in
255
-------
ro
DUMP
BLEED
TIME
AVERAGING
CHAMBER
DILUTION DEVICE
CHARGE NEUTRALIZER
PROCESS EXHAUST LINE
CHARGE NEUTRALIZER
CYCLONE
ORIFICE WITH BALL AND SOCKET
JOINTS FOR QUICK RELEASE
SOX ABSORBERS (OPTIONAL)-
HEATED INSULATED BOX
RECIRCULATED CLEAN, DRY, DILUTION AIR
FILTER BLEED NO. 2
MANOMETER
COOLING COIL
3630-036
PRESSURE
BALANCING
LINE
BLEED NO. 1
Figure 10. Sample Extraction-Dilution System (BEDS).1*
-------
EXHAUST
U 3DMIN. frj U 3DMIN. »-j
MEASUREMENT
/
HOOD
D —
I AIR FLOW
\ PITOT
E
DUCT
3E
3
GAS
SAMPI PR
SOURCE
PARTICLE
SAMPLER
CONTROL
VALVE
BLOWER
BYPASS
AIR
Figure 11. Schematic of a quasi-stack sampling arrangement.
257
-------
cases where the emissions from a source cannot be isolated, where the physical
arrangement will not permit its installation, where the character of the emis-
sions may be changed, or where the process operation might be interrupted.
Figure 12 illustrates the roof monitor method for sampling fugitive
emissions. This method is applicable to methods or processes within buildings
that have a small number of openings to the atmosphere. The structure of such
a building serves as a large hood, confining the emissions to a. limited vol-
ume of air (and hence dilution) before transmitting them through the openings
to the outside.
As in the quasi-stack method, the particulate concentration and gas
velocity are measured and used to calculate the emission rate. In the roof
monitor method, the emission rate represents the combined strength of all
the sources within the building.
The roof monitor method is less accurate than the quasi-stack method
because of the higher dilution ratio and multiplicity of sources and exhaust
openings. On the other hand, it can be applied to almost any process that
operates indoors without interrupting the normal plant operations.
Although it is the least accurate of the three sampling techniques, the
upwind-downwind method can be applied to any source of fugitive emissions.
The particulate emission rate from a source is determined by taking the dif-
ference between particulate concentrations measured in the atmosphere by a
network of samplers located upwind and downwind from a source. The samplers
are normally arranged in a network near the ground, but in some applications
have been attached to masts to form a two-dimensional, vertical array on the
downwind side of the source. The data are used as input into dispersion models
to estimate the strength of the source.
SAMPLING INSTRUMENTS
In using the quasi-stack technique to monitor fugitive emissions, the
concentration of particles is generally high enough to allow the use of
258
-------
CUPOLA
VOLTMETER
TOGAS
TRAVERSE LINE =
PULLEY
DETAIL A
GASEOUS EMISSION
SAMPLE LINE
POWER
LINE
HI-VOL
DETAIL B
Figure 12. Schematic of a roof monitor traversing sampler
arrangement.
259
-------
stack sampling instruments as described above. However, for roof monitor
sampling and upwind-downwind sampling, the aerosol is diluted and it is
desirable to use high volume samplers.
Figure 13 illustrates the high volume filter sampler that is used exten-
sively in the USA to measure particulate concentrations. The sampler is
enclosed in an aluminum enclosure to prevent rain and large particles from
falling onto the filter. A blower is used to draw a gas volume of approx-
imately 1130 £pm through a glass fiber filter (20.3 x 25.4 cm).
If a measure of the particle-size distribution is desired, cascade
impactors are available that fit directly over the Hi-Vol filter, using
the same blower and flow control system, to yield the additional data. Three
models of cascade impactors for use with the Hi-Vol are illustrated in
Figure 14.
A number of multiple jet impactors that operate at flowrates on the
order of 30 £pm are also available for fugitive and ambient sampling.
Figure 15 is a schematic illustrating the operating principle of the
•
dichotomous sampler, or virtual impactor,that has recently been developed
for sampling fugitive and ambient aerosols. This instrument samples at a
flowrate of 50 Jlpm. The separation of particles having aerodynamic diameters
of 2.5 ym and larger from smaller particles is done by impingement into a
dead air space, or virtual surface. In order to optimize the performance of
the instrument, it has been found desirable to have a small flow from the
stagnant volume. The main advantage of the dichotomous sampler is that it
can be used to collect large samples without overloading.
A Fugitive Air Sampling Train (FAST) has been designed to utilize com-
mercially available equipment to obtain a 500 milligram sample of particulate
matter and a 15 milligram sample of organic vapors in an 8-hour sampling period
at locations where the pollutant concentration levels are about 200 micrograms
per cubic meter. The prototype unit has been constructed to contain the sampl-
ing devices in a single unit, as shown in Figure 16, about 2x1x1 meters in
260
-------
AIR FLOW
"FILTER
-BLOWER
•CONTROLS
Figure 13. The standard high volume particle sampler.
261
-------
COLLECTION
PAPERS
/I
1
1
\
iX
i !
?
\ \
*\ <
1 .— I
h
t
~°°^~T
\ i
\ PARTICLE
SLOT DEPOSIT
Figure 14a. The Sierra Instruments Hi-Volume cascade impactor.
\
\
IMPACTION
PLATE
FILTER
Figure 14b. The BGI Hi-Volume cascade impactor.
AIR FLOW
JET PLATE
GASKET'
PARTICLE DEPOSIT
JET PLATE GLASS FIBER SUBSTRATE
Figure 14c. The Andersen 2000 Hi-Volume cascade impactor.
262
-------
INTAKE
(50 Spm)
TER A
(2.5 Spm)
FILTER B
(47.5 8pm)
Figure 15. A schematic of the dichotomous sampler.
263
-------
LOUVER
INLET
IMPACTOR STAGE
TO MAIN
VACUUM BLOWER
CYCLONE
XAD-2
MODULE
" j——IM—
TO VACUUM PUMP |
Figure 16. Fugitive air sampling train assembly.
264
-------
size. This sampling unit contains a single stage cascade impactor to collect
about 90% of all particles larger than 15 ym, a cyclone separator to collect
about 50% of the remaining particles larger than 3 ym, and a glass fiber filter
to collect more than 99% of the particles still entrained in the sampling
stream. The particulate sampling stream flowrate of about 300 m3/hr is pro-
vided by a separately packaged Roots-type vacuum blower. A side stream of
about 8 m3/hr is taken from the main stream below the filter by a separately
mounted vacuum pump and drawn through a canister of XAD-2 adsorbent resin to
extract organic vapors. The component arrangement is shown schematically in
Figure 17.
SAMPLING AMBIENT EMISSIONS
Sampling ambient aerosols is done with the same type of instruments
as those used in sampling dilute fugitive emissions. A large network of
Hi-Vol samplers is used to continuously monitor the total suspended partic-
ulate (TSP) burden in the USA.
Cascade impactors, dichotomous samplers, and the FAST, are all used as
part of research studies or efforts to characterize urban aerosols in more
detail.
265
-------
CYCLONE
INLET
CYCLONE
OUTLET C
T.C.
VACUUM
PUMP
EXHAUST
Figure 17- Fugitive air sampling train components,
266
-------
REFERENCES
1. American Society for Testing and Materials
Standard Method for Sampling Stacks for Particulate Matter. Designation
D2928-71, Annual Book of ASTM Standards, 1977.
2. Federal Register.
Standards of Performance for New Stationary Sources: Revision to
Reference Methods 1-8. 42(160): 41753-51789,1977.
3. Smith, W. B., and R. R. Wilson, Jr.
Development and Laboratory Evaluation of a Five-Stage Cyclone System.
Southern Research Institute, EPA, Research Triangle Park, NC. 66 pp, 1978.
EPA-600/7-78-008
4. Smith, W. B., K. M. Gushing, and J. D. McCain
Procedures Manual for Electrostatic Precipitator Evaluation.
Southern Research Institute, EPA, Research Triangle Park, NC. 430 pp, 1977.
EAP-600/7-77-059
5. Haberl, J. B., and S. J. Fusco
Condensation Nuclei Counters: Theory and Principles of Operation.
General Electric Technical Information Series, No. 70 POD 12, 1970.
6. Smith, W. B., P. R. Cavanaugh, and R. R. Wilson.
Technical Manual. A Survey of Equipment and Methods for Particulate
Sampling in Industrial Process Streams. >:
Southern Research Institute, EPA, Research Triangle Park, N.C. 281 pp, 1978.
EPA-600/7-78-043
7. Helming, E. M.,. Compiler
Symposium on Fugitive Emissions Measurement and Control.
TRC, EPA, Research Triangle Park, NC., 1976.
EPA-600/2-76-246
267
-------
D.L. Zelikson, N.G. Bulgakova
PRESENT-DAY TENDENCIES IN AEROSOL DISPERSION ANALYSIS
Improving methods for protecting the atmosphere from industrial
emissions is based on the reliability of results from aero-mechanical
and physio-chemical measurement. Measurement is strictly the basis
of practically all aspects of environmental protection since it al-
lows quantitative evaluation of efficiency of protection and of end
results - the attainment of sanitary norms. The leading role of mea-
surement in supplying technical process with the means of gas-clean-
ing may be realized only means of advanced development of concrete
types of measuring techniques and of its basis, metrology.
Raising the question of lead development in_measurement tech-
nology should above all accentuate research in an equipment oriented
direction, deciding types of measurement directly necessary for re-
liable evaluation of gas-cleaning efficiency- Dispersion analysis
and a survey of possible means of its development is the purpose of
this report. An outline of literature (1-4) of various schemes of
classifying'"methods for measuring aerosol particle diameter attests
to the lack of a single common opinion on questions of selecting
classifying features. Sufficient]^ arbitrary features are usually
used applicable to this narrow "problem. Disregarding attempts to
set up a substantial classification model because of its complexity
and the fact that it falls outside the limits of the given topic
it is necessary as a consequence to select one basic type of disper-
sion analysis - dynamic dispersion analysis. Classification of mea-
surement methods with regard to "dynamic features" is not found in
the literature. At the same time, dynamic measurement type comprises
a wide area of aerosol dispersion analyses to which in part relate
more known methods of inertial precipitation and sedimenta-
tion.
Dynamic dipsersion analysis uses a fundamental property of
particles - the presence of mass and inertia - and is realized
by utilizing the effect of various forces on the suspended parti-
cle. Under the influence of these forces, particles acquire a giv^n
motion, therefore the problem of dynamic analysis is reduced to cor-
rect measurement of this motion's parameters.
From a metrological point of view, dynamic analysis is close
268
-------
to absolute measurements more preferable in result reliability
when the magnitude of the measured value is found in agreement with
the estimation of the given value's conception.
Prom determination of dynamic analysis follow basic classi-
fication features:
- the nature of the acting forces;
- types of particle motion;
- methods of measuring parameters of motion.
An enlarged classification is introduced in the table. Detailed
classification is connected with clarification of concrete methods
of force interaction with the particles and with each measure-
ment method.
The classification provides five types of motion:
- eguilibrial, when the particle is found at rest or is
moving evenly and in a straight line;
- curvilineal, if when under the influence of applied forces
the particle trajectory deviates in relation to its inertia;
- rotary motion has significance for non-spherical particles„
mainly for those of elongated form and, obviously, it is almost
th*. sole form of motion allowing evaluation of the particle's
form;
- oscillatory, i.e. back-antf-forth motion, whose amplitude and
frequency dtermines the particle mass;
-impulse motion which has been practically unused until now and
has a conditional character.
Active forces may be divided by their nature into aerodynamic,
acoustic, electromagnetic, and distinct from the others, optical.
In the field of aerodynamic forces, the particle participates
in any of the listed types of motion. In sedimentation methods,
counterbalancing gravitational and centrifugal forces, other than aero-
dynamic, act upon the particle, which causes equilibrial motion.
Aerosol stream flow against a barrier such as walls in cyclone
analysers and impactors or against a counter flow of pure gas in
so-called virtual impactors (5) is charaterized by the particle's
curvilinear tra j ectory.
Aerosol oradient flov; in a boundary layer or with stream
vortex limits particle revolution. The submersed supersonic stream has
periodic structure and and the stream velocity flow axis perio-
dically changes from sonic to higher. The spatial modulation of
269
-------
DYNAMIC DISPERSION ANALYSIS OF AEROSOLS
EQUILIBRIAL
CURVILINEAR
ROTARY
OSCILLATORY
IMPULSE
^
1
1
1
i
1
•
1
1
•
•
•
•
1
I
>
V
\
METHODS
'!•'
:
|:
OF
"
N
\
V
t
••
|
I
I
•
i
•
•
•
•
•
•
•
' k
*
1
1
(
1
•
1
I 1
'.
I •
1
'
1 •
* *
| »
T
i.
1
1
»
!:
i V 1
1"
]
1
s-
i
i
i
I
L
l
;i
!i
I'
i
»i
'
i
1
MEASUREMENT
g
H
H
o
£3 O
O
s!
«
Si
USTIC
o
o
<
o
w
3
H
|
1
3
H
EH
&
O
CJ
H
«
H
1
H
y
«
ACTIVE FORCES
AERODYNAMIC
t.' • ] ACOUSTIC
I- 1 ELECTROMAGNETIC
!•• • I OPTICAL
-------
velocity results in particle separation (6) . The shock wave transmits
impulse motion to the particles.
In the acoustic standing wave, the particles, moving in loop
motion, form an eguilibrial spatial lattice(7), whose period of
origin is a function of dispersion composition. Acoustically linear
effects, accompanied by directed flows, perpendicular to the aero-
sol stream, may be used for fractional precipitation of particles
moving curvilinearly. The travelling sonic wave draws particles into
oscillation and revolution. The latter motion occurs due to the
presence of velocity gradients in the wave (7). Sonic radio pulse
or video pulse has an effect on the particle that is analogous
to .a shock wave.
Electrical, magnetic and optical forces influence the parti-
cle identically in equilibrial or curvilinear motion. In the first
case, these forces counterbalance gravitational and/or aerodynamic
forces, and in the second, the appearance of phoresis is used, act-
ing perpendicularly to the aerosol stream (8) . Revolution of elongated
particles may be carried out by electro- or photophoresis. The spe-
cific action of the electromagnetic field appears in mass-specto-
mcury. Mass-spectral analyzers used for high-dispersion charged
aerosols separate particles according to curvilinear trajectories
(9). The rotary electromagnetic field of the multi-phase . flow
holds wide possibilities for particle rotation (10). Spherical par-
ticle oscillation in the plane condenser's variable field is used
for the most exact diameter measurements. Electrical discharge is
the source of impulse movement.
In the standing optical wave are oscillating subatomic char-
ged particles, the ions that form the spatial lattice analogous to
the acoustic lattice. Irradiation of particles by a high-intensity
laser source resulting in a piercing of the dispersion medium aids particle
visualization (11) and may be conditionally related to impulse
types, accompanied by transfer to the second aggregate state.
Methods of measuring movement parameters are based on
partial usage of active field energies or on registration of
the results of their influence.
Mechanical measurement methods reduce to determination
of mass of the particles settled from the aerosol, the rate of
their settling or revolution, the path travelled by the parti-
cles or the time "taken to cover a given distance, and the
271
-------
amplitude of oscillations etc. in various combinations of types
of motions and active forces aiven above. Aerodynamical measure-
ment methods preclude those using aerosol flow parameter regi-
stration as a function of dispersion phase concentration. This
hydraulic pressure variation of filters and other hydraulic sys-
tems, in particularly, in conjunction with changes in the
viscosity of the dual-phase turbulent flow. Acoustic measurement
methods include recording absorption or dispersion doses, which
includes energy diffracted by particles of thetsonic-'field. Mea-
surement of the floppier effect of frequency in relation to par-
ticle speed is possible, as well as recording sound pressure
during particle impact.
Electrical discharge/ transfered by particles directly to the
collector electrode or induced, is measured in the electrical meth-
ods. For measurement of mass of settling, tuning fork or piezoreso-
nance weights have been developed in which oscillation frequency of
the resonance contour is set in relation to the amount of settling.
The piezoelectrical effect is used also in pressure converters re=
cording impulses during single particle impacts against the sensory
element surface (12). Measurement of electrical beam absorption in
the aerosol "flow, magnetic and dielectric permeability dielec-
trical loss coeeficient, etc. is possible. Optical measurement me-
thods present the widest possibilities. Equipment based upon these
methods are universal, frequently automated, probeless means of
measurement introducing the least interference in the measurement
process and having maximum accuracy and sensitivity. Cinematogra-
phic, photographic and visual recording of movement parameters -or
fixation of particle position is related to these methods. Light absorp-
tion, scattering and difraction, polarization charactersistic vari-
ations, emission frequency Doppler effect and charactersitic illu-
mination of exploding particles are used in this equipment.
Radiometric methods rival optical measurement. These methods
include absorption, scattering and difraction of emissions, inher-
ent emission of radioactive aerosols, and Doppler signal modula-
tion down to the use of the Messbauer effect in special cases.
The considered methods of measurement are used for recording
all tyoes of particle movement in different combinations of active
forces in wide ranges of fractional concentrations and dimensions
272
-------
particles of various forms. In aerosol measurement technology and
industrial gas cleaning, particle size in the range of 5-6 orders '
of magnitude from 1 nanometer at concentrations also in the range
of 5-6 orders of magnitude from 1 microgram/m are of interest, and
in experimental studies from one particle in the measured volume. The maximally
attained* accuracy of particle measurement on the order of 1 micron
10 (13). Dynamic analysis methods of aerosol dispersed
composition altogether meet the above demands and provide to a
considerable degree the possibility for automating analysis in
various conditions of use.
Two ways of of developing aerosol measurement techniques are
possible. The first calls for creating universal means of measure-
ment favorable for use in any, even extreme, .conditions. Ac-
cording to the second, units that have been developed are special-
ized according to their branch of industry or according to gas-
cleaning or dust removal systems.- Universal equipment is more expen-
sive, complex in its design and maintenance, requires more quali-
t
fied operating personel during installation at experimental sites,
but then is automated in a large degree and therefore simple in
operation under conditions of sufficient reliability. Specialized
equipment is distinguished by the simplicity of its design and
mounting but practically excludes the possibility of changing
the study site. Apparently, the technology of measurement in the
area of dispersion analysis should be developed with determined
relationship of these types of equipment with regard to the
necessity of establishing calibration means to ensure accuracy.
BIBLIOGRAPHY
l.Klimenko, A.P. Methods and Equipment for Measuring Dust Concentration,
Moscow, "Khimiia", 1978.
2. Methods, Equipment, and Systems for Monitoring the Industrial Envir-
onment ; Interinstitutional Collection, Leningrad, LETI, 1976.
3. Larchenko, V.I., Filippov, V.P., Kulenev, Y.P., Modern Methods for
Controlinq Aerosol Concentration ; Survey Information TS-4 "Analytic
Equipment and -Equipment for Scientific Investigation", Moscow, TsNIITEI
equipment construction, 1977.
273
-------
4. Soprunyuk, P.M., Pau, A.A., Semen, N.P., New Dust Monitoring Equip-
ment ; Survey Information TS-4 "Analytic Equipment and Equipment for
Scientific Investigation", Moscow, TsNIITEI equipment construction, 1977
5. A Method for Measuring Particle Diameter, Japanese patent 113DO
(COIN 15/02) No.52-22560, 1977.
6. Alkhimov, A.P., Papyrin, A.N., Predein, A.L., Soloukhin, R.I.,
Experimental Study of the Speed Lag Effect of Fine Particles in
a Supersonic Gas Flow, PMFT Journal, 1977, No.4.
7. Mednikov, E.P., Acoustical Coagulation in Aerosol Settling,
Moscow, USSR Academy of Sciences, 1963.
8. Kutukov,V.B.-, Ostrovskii, Y.K., Yalamov, Y.I., A Method for
Separating Aerosol Particles from Gas Flow, USSR Author's Certificate,
Patent no.558200, COIN 21/22, 1977.
9. Susoev, A.A., Chupakhin, M.S., Introduction to Mass-Spectometry,
Moscow, Atomizdat, 1977.
10. Tolstoi, N.A., Spartakov, A.A., Trusov, A.A., Electrooptical
Properties of Lyophobic Colloids, "Colloid Journal", 1966, 28, No.5,
p. 735.
11. Zakharchenko, S.V., Kolomiets, S.M., Skripkin, A.M., Breakdown
of Dispersion Medium By Laser Emission,"PMFT Journal", 1977, 3, No.
24, p.1339-1343.
12. Kozelkina, E.V., Kozelkin, V.V., Otroshko, N.G., Methods of Deter-
mining Solid Particle Concentration in Gas Suspension Flows, "Factory
Laboratory", 1971, No.3, 307-308.
13. Ashkin, A., Recent Experiments with Optical Levitation, "Optical
Communications", 1976, 18, No.l, 190-191.
274
-------
INSTRUMENTS FOR AUTOMATICALLY MONITORING
PARTICLE-SIZE DISTRIBUTIONS
W. M. Farthing
W.B. Smith
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35205
W. B. Kuykendal
Industrial Environmental Research Laboratory
Environmental Protection Agency
Research Triangle Park, N.C. 27711
(This document has been reviewed and approved
for publication by the U. S. EPA.)
275
-------
INSTRUMENTS FOR AUTOMATICALLY MONITORING
PARTICLE-SIZE DISTRIBUTIONS
INTRODUCTION
Although the majority of size distribution measurements are still made
using manual techniques, there is a clear need for automatic systems with
real-time data output. Examples of such applications are: monitoring processes
with periodic particle generation, process control, and optimizing control
device performance.
The ideal measurement method is one which yields particle-size data con-
veniently, without perturbing the sample. Also, it is usually desirable to
measure the diameter of the particles in terms of their aerodynamic behavior
so that an estimate can be made of potential health effects. Some hybrid
systems have been developed wherein the particles are size-segregated by im-
paction, with optical;piezoelectric, and $-particle monitor sensors after each
stage; but, the greatest emphasis is now on the development of optical systems.
The optical systems appear to offer greater convenience, speed, and reliability,
but the accuracy will ultimately depend upon some knowledge of the optical
properties of the aerosol being measured, and solving the problem of relating
optical measurements to aerodynamic behavior.
In studies of process streams, the particle concentration, size range,
and composition may present experimental difficulties, as well as the corro-
siveness and temperature of the gas.
This paper includes descriptions of particle measuring systems now in
regular use in the U.S.A., and some prototype systems that have been recently
developed for in situ applications.
276
-------
SYSTEMS NOW IN REGULAR USE l
CASCADE IMPACTORS WITH PIEZOELECTRIC CRYSTAL SENSORS
Piezoelectricity is a property of certain crystals, such as quartz,
which involves the production of an electrical charge on certain faces of the
crystal when the crystal becomes mechanically stressed. The converse process
also occurs; that is, a piezoelectric crystal becomes mechanically stressed
where an electrical charge is placed on certain faces. This two-way capability
is responsible for the ability of a piezoelectric crystal to cause an oscillating
electric circuit to oscillate at the natural vibrational frequency of the crystal.
When foreign material, such as aerosol particles, adheres to the surface of a
vibrating piezoelectric crystal, the natural frequency of vibration of the
crystal decreases. The magnitude of the frequency change is directly propor-
tional to the mass of the added material.
Carpenter and Brenchly and Chuan have developed and tested multiple-stage
cascade impactors with piezoelectric crystals on each stage to monitor the rate
and amount of mass collected. Chuan's impactor is now sold commercially by
Berkeley Controls, Inc. (2700 Du Pont Drive,Irvine, CA 92714). Chuan's
impactor has ten stages, with the cut points reported to be from 0.05 to
about 25 ym. Because of the extreme sensitivity of the instrument (and upper
limit on mass accumulation), it is more suitable for ambient than stack work,
where sample extraction and dilution would be required. The impactor must be
disassembled, and each crystal cleaned before the limits of the linear range
are reached. In a typical urban atmosphere, this limit is reached in about
two hours. When interfaced with a programmable calculator, the system will
print out the measured size distribution at chosen time intervals. The
piezoelectric impactor is shown in Figure 1.
ELECTRICAL MOBILITY ANALYZER
Figure 2 illustrates the relationship between the diameter and electrical
mobility of small aerosol particles. If particles larger than those of minimum
mobility are removed from the sample, the remaining particles exhibit a monotonically
277
-------
10-6
* 10-7
0
10-8
0
0
a o
Q
D D
O E - 5.0 x 105 V/m
Nt - 8.0 x 1011 sec/m3
D E - 1.5 x 105 V/m
Nt - 3.2 x 1012 sec/m3
SHELLAC AEROSOL K = 3.2
0.0 0.2 0.4 0.6 0.8 1.0
PARTICLE DIAMETER, pm
1.2
1.4
3630-251
Figure 2. Particle mobility as a function of diameter for shellac
aerosol particles charged in a positive ion field. K
is the dielectric constant of the aerosol particles.
278
-------
decreasing mobility with increasing diameter. Several aerosol spectrometers,
or mobility analyzers, have been demonstrated that employ the diameter-
mobility relationship to classify particles according to their size, and Figure
3 illustrates the principle on which these devices operate. Particles are
charged under conditions of homogeneous electric field and ion concentration,
and then passed into the spectrometer. Clean air flows down the length of
the device and a transverse electric field is applied. From a knowledge of
the system geometry and operating conditions, the mobility is derived for any
position of deposition on the grounded electrode. The particle diameter is
then readily calculated from a knowledge of the electric charge and mobility.
Difficulties with mobility analyzers are associated primarily with charg-
ing the particles to a known value with a minimum of loss by precipitation and
obtaining accurate analyses of the quantity of particles in each size range.
The latter may be done gravimetrically, optically, or electrically.
The concept described above has been used by Whitby, Liu, et al, at the
University of Minnesota, to develop a series of models of the Electrical Aerosol
Analyzer(EAA). A commercial version of the University of Minnesota devices is
now marketed by Thermosysterns, Inc., as the Model 3030 (Figure 4). The EAA is
designed to measure the size distribution of particles in the range from
0.0032 to 1.0 ym diameter. The concentration range for best operation is 1 to
1000 yg/m3, and thus dilution is required for most industrial gas aerosols.
The EAA is operated in the following manner. As a vacuum pump draws
the aerosol through the analyzer (see Figure 4), a corona generated at a high
voltage wire within the charging section gives the sample a positive electrical
charge. The charged aerosol flows from the charger to the analyzer section as
an annular cylinder of aerosol surrounding a cone of clean air. A metal rod,
to which a variable, negative voltage can be applied, passes axially through
the center of the analyzer tube. Particles smaller than a certain size (with
highest electrical mobility) are drawn to the collecting rod when the voltage
corresponding to that size is on the rod. Larger particles pass through the
analyzer tube and are collected by a filter. The electrical charges on these
particles drain off through an electrometer; giving a measure of current.
279
-------
CHARGED PARTICLES
CLEAN AIR
LAMINAR FLOW
fc
A
\
\
SMALLER PARTICLES OF
HIGH ELECTRICAL MOBILITY
LARGER PARTICLES OF
LOW ELECTRICAL MOBILITY
1
3630-252
Figure 3. The electric mobility principle.
280
-------
CONTROL MODULE
ANAL*Zen nutria I
DAT* RCAO COMMAND -
CTCLE STMT I
CVCLC KKT COMMAND -
ACROMX. flOWMITiR RIAOOUT
CHAHJt* CURRENT MUWUT
— CMAdOtd VOLTMC MADOUT
AUTOMATIC HIGH VOLTAOC COHTWIL AMD MAOOUT
OfCTKOMCTCR IANALrl£R CUMEMTI WACOUT
•TOTAL fLWKTfK KAOOIT
TO VACUUM
to
00
Figure 4. Flow schematic and electronic block diagram of the Electrical
Aerosol Analyzer.
-------
A step increase in rod voltage will cause particles of a larger size to
be collected by the rod with a resulting decrease in electrometer current.
This decrease in current is related to the additional number of particles be-
ing collected. A total of eleven voltage steps divide the 0.0032 to 1.0
micron size range of the instrument into ten equal logarithmic size intervals.
Different size intervals can be programmed by means of an optional plug-in
memory card.
The electrical aerosol analyzer can be operated either automatically or
manually. In the automatic mode, the analyzer steps through the entire size
range. For size and concentration monitoring over an extended period of time,
the analyzer may be intermittently triggered by an external timer. The stand-
ard readout consists of a digital display within the control circuit module,
although a chart recorder output is available. It is almost always advanta-
geous to use a strip chart recorder to record the data. This allows the
operator to identify a stable reading that may be superimposed on source varia-
tions and also gives a permanent record of the raw data. The EAA requires only
two minutes to perform a complete size distribution analysis, which generally
makes it advantageous to use, especially on stable sources.
OPTICAL PARTICLE COUNTERS
Figure 5 is a schematic diagram illustrating the principle of operation
for optical particle counters. A dilute aerosol stream intersects the focus
of a light beam to form an optical "view volume." The photodetector is located
so that no light reaches its sensitive cathode except that scattered by par-
ticles in the view volume. Thus, each particle that scatters light with enough
intensity will generate a current pulse at the photodetector, and the amplitude
of the pulse can be related to the particle diameter. Optical particle counters
yield real-time information on particle size and concentration.
In an optical particle counter, the intensity of the scattered light and
the amplitude of the resulting current pulse depend on the viewing angle, par-
ticle refractive index, particle shape, and particle diameter. Different view-
ing angles and optical geometries are chosen to optimize some aspect of the
282
-------
LIGHT TRAP
LAMP
SAMPLE AEROSOL
TO PUMP
PHOTOMULTIPLIER
3630-243
Figure 5. Schematic of an optical single particle counter.
283
-------
counter performance. For example, the use of near forward scattering will
minimize the dependence of the response on the particle refractive index, but
with a severe loss of resolution near 1 ym diameter. The use of right angle
scattering smooths out the response curve, but the intensity is more dependent
on the particle refractive index.
Figure 6 illustrates some of the optical configurations that are found
in commercial particle counters. The pertinent geometric and operating constants
of the counters are summarized in Table 1.
The commercial optical counters that are available now were designed for
laboratory work and have concentration limits of a few hundred particles per
cubic centimeter. The lower size limit is nominally about 0.3 ym diameter.
For use in studies of industrial aerosols, dilution of the sample is required
and the useful upper limit in particle size has been limited by losses in the
dilution system to about 2.0 ym diameter. In addition, the particle diameter
that is measured is not aerodynamic, and some assumptions must be made in order
to compare optical with aerodynamic data. However, it is possible to "calibrate"
an optical counter, on a particulate source, to yield aerodynamic data. This
is done by using special calibration impactors, or settling chambers.
DIFFUSION BATTERIES WITH CONDENSATION NUCLEI COUNTERS
Diffusion batteries with condensation nuclei counters have been standard
equipment for measuring the size distribution of submicrometer particles for
many years. Data acquisition with these systems, however, has been tedious
and time consuming. Recently, Knutson and Sinclair developed and demonstrated
an automatic diffusional sizing system.2 Figure 7 shows their system. The
aerosol is drawn through a 60JI tank to damp short-term fluctuations and into
the miniature, screen-type diffusion battery. A motor driven valve having
15 ports is used to switch from port to port on the diffusion battery and to
sample filtered air. The aerosol concentration at the inlet and outlet of
the diffusion battery is measured using a continuous-flow condensation nuclei
counter which has alcohol as a working fluid. The data are recorded on punched
tape for later analysis by computer. Knutson and Sinclair's system has been
284
-------
CALIBRATOR
CLIMET
02*
SCATTtRING
FHOTOOETECTOR
CURVED MIRROR
gam REFLECTIVITY
«0mm F.L.
CYUNOER LENS
HE-NE 6mW LASER
REFERENCE
PHOTOOETECTOR
MODULE
PMSLAS-200
5 mm F.L.
PARABOLIC MIRROR
90« REFLECTIVITY
O" B1NO SEAL
DUMP WINDOW
AEHODYNAMICALLV
FOCUSING INLET
W DIAGONAL FLAT
M + % REFLECTIVITY
! £ SHEATH AIR
V SAMPLE AIR
02b
COLLECTION PUPIL
UGHT LENS LENS
TRAF
FHOTOMULTIFUER
FHOTOMULTIFLIER
ROYCO 220
ROYCO226
02c
02d
PHOTOMULTIFUER
TUBE,
CONDENSER LENSES
PHOTOMULTIPLIER
LIGHT
TRAF
VOLUME
ROYCO 248
8 AND L 40-1
02*
02f
Figure 6. Optical configurations for six commercial particle
counters.
285
-------
TABLE 1.
CHARACTERISTICS OF COMMERCIAL, OPTICAL, PARTICLE COUNTERS
Bausch & Lorab Model 40-1
820 Linden Avel
Rochester, NY 14625
Climet Models 201, 208
Climet Inst. Co.
1620 W. Colton Ave.
Badlands, CA 92373
•Model LAS-200
Particle Measuring Systems
1855 S. 57th Ct.
Boulder, CO 80301
Illuminating Cone
Half Angle, y
13°
15
Light Trap Half Collecting Aperture Inclination Between
Angle, a Half Angle, 3 Illuminating And
Viewing
Collecting Cone Axis, i|> Volume
33°
35
53"
90
0°
0.5
0.5
Sampling
Rate
170 cn'/"oln
7,080
NJ
OO
CT>
Climet Model 150
Royco Model 218
Royco Inat.
41 Jefferson Dr.
Menlo Park, CA 94025
Royco Model 220
Royco Model 245
Royco Model 225
Tech Ecology Model 200
Tech Ecology, Inc.
645 N. Mary Ave.
Sunnyvalle, CA 94086
Tech Ecology Model 208
Particle Measurement Systems
12
5
24
5
5
5
5
0.5
18
11
-
16
7
8
10
35
28
30
24
25
25
20
20
120
0
0
90
0
0
0
0
0
0.4
0.25
2.63
4.0
2.0
0.46
2.5
0.003
472
283
2,830
28,300
283 or 2,830
283
2,830
120 or 1,200
•632.8 mm laser ilium., all others are white light.
-------
ROTARY
15-PORT VALVE
AEROSOL
IN
NJ
OO
4 Ipiti J LJ
SINCLAIR
NUCLEUS
COUNTER
TO PUMP
SINCLAIR
DIFFUSION
BATTERY
STEP AND
HOME
SIGNALS
4 Ipm TO OTHER
INSTRUMENTS
CONVERT
SIGNAL
601
BALLAST
TANK
TIMING AND
CONTROL
CIRCUIT
ANALOG TO
DIGITAL
CONVERTER
Figure 7. Block Diagram of aerosol measurement system.
-------
used to obtain over 1000 size spectra of urban aerosols. The diffusion
battery and rotary switch are now sold commercially by Thermo-Systerns, Inc.,
St. Paul, Minnesota.
SAMPLE EXTRACTION-DILUTION SYSTEMS FOR PROCESS STREAMS
The majority of instruments and systems for measuring particle-size
distributions are designed to measure aerosols at conditions near ambient
air with respect to temperature, gas composition, and pressure. Further-
more, the instruments are limited to aerosol concentrations much lower than
those typically found in process streams. Therefore, some means of extract-
ing samples, conditioning them, and diluting them, is required for measure-
ments in process streams.
Figure 8 shows one sample extraction-dilution system (SEDS) that
has been used by Ensor to measure aerosol properties in process streams.1*
It uses an in-stack impactor to remove particles larger than about 2.7 pm
and a diffusional dryer for removing moisture, before the sample enters the
dilution system. Dilution is done by a three-stage process employing
venturi flow meters to measure the sample flowrate and orifice meters to
measure the flowrate of the dilution air. Flow control is accomplished by
manipulating the dilution air control valves. The dilution can be adjusted
from a factor of 20:1 to 1000:1. Particle sizing can be done using the
electrical aerosol analyzer or diffusional systems.
Figure 9 is a block diagram of the sample extraction-dilution system
developed by Southern Research Institute.3 The sample is removed from the
flue by means of a heated probe which is connected to an oven by means of a
heated, flexible hose. In the oven the sample gas flows through a cyclone,
an orifice for metering sample flow, and an optional SO... absorber assembly.
A
Charge neutralization is done in the cone of the diluter by two 500 yc
polonium-210 strips mounted as shown.
The sample gas enters the dilution chamber at the apex of a perforated
cone into which clean, dry air is pumped through the perforations, creating
a highly turbulent mixing zone. At a downstream point after adequate mixing
288
-------
HEAT TRACED
SAMPLE LINE
SOURCE
r
AEROSOIl
'J
DIFFUSIONAL
DRYER
1
MODEL 1502
CASCADE
IMPACTOR
PRECUTTER
ENCLOSED
/PACKAGE
PI
76-096
T1, T2 DIAL THERMOMETERS
01-03 ORIFICE METERS
V1-V4 VENTURI METERS
CV1-CV6 DILUTION AIR CONTROL VALVES
SV1 SELECTOR VALVE
PI. P2 PRESSURE GAGES
Figure 8. Meterology Research, Inc. - SEDS.
-------
to
VO
o
DUMP
TIME
AVERAGING
CHAMBER
=J>^ BLEED DILUTION DEVICE
CHARGE NEUTRALIZER
PROCESS EXHAUST LINE
CHARGE NEUTRALIZER
CYCLONE
ORIFICE WITH BALL AND SOCKET
JOINTS FOR QUICK RELEASE
-ft-.-:-— --.: LI -:
SOX ABSORBERS (OPTIONAL)
n
innv —I
HfcATED INSULATED BOX
RECIRCULATED CLEAN, DRY, DILUTION AIR
BLEED NO. 2
MANOMETER
COOLING COIL
3630-036
PRESSURE
BALANCING
LINE
BLEED NO. 1
Figure 9. Southern Research Institute SEDS.
-------
has occurred, the dilution sample is extracted and conveyed to the sizing
instrument. The diluted sample passes through a diffusional dryer where
any remaining moisture is removed. The major drying action is accomplished
by using dilution air which has been passed through an ice bath condenser.
The sample then passes through an optional time averaging chamber and into
the sizing instrument. The instrument exhaust lines are returned to the
diluter to maintain a minimal pressure change across the instruments.
Further drying of this recycled air is accomplished by passing this instru-
ment exhaust gas through an absorption dryer.
Changing the sample air flow and the dilution air flow causes changes
in the dilution ratio<. Sample air and dilution air flowrates are controlled
by two bleed valves on the dilution air pump, one upstream of the pump (#1)
and one downstream (#2). Manipulation of these valves changes the internal
pressure of the diluter which, in turn, sets the sampling rate. As the
pressure in the diluter is reduced, the sample flowrate is increased.
In practice, the operation of these valves changes the dilution air
flow only slightly (approximately ±10%) for a many-fold change in sample
flow, or backflush condition, to the maximum flow value attainable with any
particular sample metering orifice.
Particle sizing is done using optical particle counters, the electrical
aerosol analyzer, and diffusional systems.
NEW SYSTEMS FOR AUTOMATIC PARTICLE SIZING
A variety of new systems has been developed for particle sizing work,
some for in situ applications to process streams. Most of the systems employ
light scattering from particles as their basic analytic principle. In this
section some of the new instruments are described. None of the units de-
scribed here are in widespread commercial use, and their performance may
not have been fully evaluated.
291
-------
INTENSITY RATIO AT TWO ANGLES
Hodkinson5 was the first to point out that the ratio of the light
intensity scattered by particles into two small angles is relatively in-
sensitive to the particles' index of refraction.
Shofner et al,6 Gravatt7 and Chan8 have developed prototype systems for
particle sizing which are based on the intensity ratio concept of Hodkinson.
Figure 10 is a schematic of Shofner's system which is called the "PILLS-IV".
The intensities of the scattered light puses at the angles 61 and 82 are
normalized to the reference pulse at 6 » 0 for synchronization and to
account for fluctuations in intensity of the laser source. The optics and
sensors are kept clean and cool by the use of a purge air system.
The laser used in this system is a semiconductor junction diode (X -
0.9 ym) . The useful size range for particle sizing is from 0.2 to 3.0 ym
although the ratio analysis is abandoned at radii less than 0.7 ym and the
magnitude of scattered intensity is employed. Shofner states that the view
volume of his system is approximately 2 x 10~7 cm3. The upper concentration
limit for single particle counters is determined by the requirement that the
probability of more than one particle appearing in the view volume of a given
time be much less than unity. For Shofnerfs system this would set the concentra-
tion limit at approximately 106 particles/cm3, a value much higher than for
conventional single particle counters.
Kuykendal9 and Farthing10 have done tests to evaluate the performance
of the PILLS IV and compared its response to other instruments. The results
of these tests indicate that further refinement will be required to achieve
an acceptable level of accuracy.
OPEN CAVITY LASER
An optical particle sizing device developed by R. G. Knollenberg1*' 12
for atmospheric application might be applicable to industrial emission measure-
ments. The device is, in some respects, similar to conventional near forward
scattering optical single particle counters except that the sensing zone is
292
-------
10
M)
CO
CONTROL LASER AND
ELECTRONICS OPTICS
OPTICS'
DETECTOR,
AMPLIFIER
SIZE
ANALYSIS
CIRCUIT*
READ-OUT
SAMPLING
VOLUME
p(d)
*• DESIGN - ESC PROPRIETARY
+• DIGITAL PRINT-OUT
PAPER OR MAGNETIC TAPE
CRT HISTOGRAM
LED DISPLAY OF COUNTS
MINI-COMPUTER INPUT
RATE METER
TYPICAL SIZE RANGE-0.2-3.0Mm
Figure 10. The PILLS IV optical Particle counter. After Shofner.'
-------
contained within an open cavity laser as illustrated in Figure 11. This
configuration results in very high illumination levels,, permitting the
detection of particles smaller than those sized by most light-scattering
instruments, and has been used in vacuum applications for sizing particles
in the range from 0.1 to 30 ym.
An interesting characteristic of laser cavity scattering is the fact
that the particle or particles in the beam effects the power incident upon
it. This is due to the detuning effect of the laser because of the particles'
presence, which is a function of the particle size. However, the scattering
is normalized with the reference detector signal which is proportional to
the actual intensity incident upon a particle. As with other particle
counters which determine from the magnitude of scattered power, the inferred
particle size from a measurement depends upon the refractive index.
The open cavity laser spectrometer is available commercially from
Particle Measurement Systems, Boulder, Colorado.
SCATTERING AT FORWARD ANGLES FROM A COLLECTION OF PARTICLES
If a sufficient number of moments of the size distribution can be
obtained, measurements of optical parameters of a collection of particles are,
of course, more desirable than sequential single particle analysis. The
latter is usually more restricted by concentration and particle size limits
and more complicated mechanically.
Wertheimer and Wilcock13 developed an approximate technique based upon
diffraction theory to determine the second, third, and fourth moments of the
size distribution. The technique utilizes three masks of different shapes
to spatially filter the detected signal. The optical system is shown in
Figure 12 and the three masks in Figure 13. With the a2 mask the detected
signal from a single particle is proportional to the second power of the
particle radius. Similarly with the a3 (a1*) mask the detected signal is pro-
portional to the thrid (fourth) power of a.. For many such particles in the
field of view of the detector then the three detected signals are proportional
294
-------
Scattering
Signal
Photodetector
Module
Curved
Mirror
0%T
t
Spike
Filter
He-Ne Hybrid Laser
Particle
Plane
Reference
Photodetector
Module
Note: Mirrors are reversed for large particle size ranges.
Figure 11. Schematic of open cavity laser active scattering
particle counter. After Knollenberg.1l
295
-------
Lens 2
Fraunhofer Plane
Rotating Spatial Filter
Lens 1
N>
VD
CTi
Particles on
Microscope Slide
Detector
Spatial Filter
and Collimator
Chart Recorder
Figure 12. Optical system used to verify the response of the Fraunhofer plate
filter. (After Wertheimer and Wilcock.13)
-------
K)
a Region
Fraunhofer Plane
Mask
Rotating Disk
a3 Region
(A)
a1* Region
Detector
(B)
Lens 2
Fraunhofer Plane
and Fixed Sector Disk
Lens 1
Figure 13. (A) Schematic diagram of Fraunhofer plane mask with three response regions.
(B) Instrument configuration for the three region mask. (After Wertheimer
and Wilcock.13)
-------
to the second, third, and fourth moments of the distribution. These signals
give the volume mean radius, the area mean radius, and the standard deviation
of the area distribution.
The authors of this technique demonstrated its usefulness using parti-
cles from 4 to 83 ym in diameter. Since then the method has been incorporated
into a commercial sizing system still being evaluated. From consideration of
the principle upon which this method is based, it is sensitive to refractive
index to the same degree as forward-scatter particle counters.
Figure 14 illustrates an optical particle counter that has been designed
by R. Knollenberg for real-time, in situ particle sizing in process streams.2
The instrument covers a particle size range of 0.3 to 10 ym with 60'channels
of resolution. The counter uses small angle light scattering, and the un-
certainty in sizing spherical particles is due to variations in the detected
particle's index of refraction. Sizing errors of ±20% are possible, with
±10% expected to be more common. The concentration range for accurate meas-
urements limited by coincidence counting is 5 x 101* particles/cm . The main
effect of higher concentrations is said to be a decrease in the effective
sizing range. An optical velocimeter is also included in the design. The
prototype design allows temperatures up to 250°C and gas velocities to 30m/sec.
The results of an initial in situ test at a coal-fired power plant were reason-
able. The opacity calculated from the measured particle-size distribution
was about 15%, while the measured opacity was 17%. The calculated mass load-
ing was 0.01 to 0.02 g/m3 with a volume median diameter of 1.3 ym.
Wertheimer et al13 are developing another portable light scattering
instrument for in situ particle sizing. This device measures the flux scat-
tered from an ensemble of particles at three small angles relative to the
forward direction, 4°, 8°, and 11°, and between 80°-100°. Each measurement
is performed at two wavelengths, 0.45 and 0.9 ym, and the large angle scat-
tering is measured at two orthogonal polarizations. The instrument relates
the small angle signals dominated by Fraunhofer diffraction to the volume of
particles in three size ranges, centered at 1.0, 3.5, and 7.0 ym. For the
lower end of the size distribution, the differences in the two 90° signals,
at two orthogonal polarizations, obtained with the 0.9 and 0.45 ym wavelengths
298
-------
IS3
V£>
\O
STAINLESS STEEL BOOM
REFERENCE DETECTOR
LASER MODEL 80-5T
(COHERENT RADIATION)
FIBERGLASS INSULATION
HIGH REFLECTIVITY FLAT
MIRROR (2)
HIGH VOLTAGE
LASER LEAD _
SIGNAL
CABLES
PURGE AIR
-z__—-t ^ ra=in4*"iSf—— = 'i~I==^==: =^in^ ^gfe^ll'^r
INPUT
WATER
UNE
STAINLESS
STEEL INSULATION
RETAINER AND COVER
HEATER
HIGH REFLECTIVITY
MIRROR 7.5mm
RADIUS
PHOTO DETECTOR
ASSEMBLY
BEAM
SPUTTER
LOW LOSS AH COATED
WINDOW
Figure 14. Sketch of the Knollenberg, in situ particle counter.
-------
is related to the volume of particles in a size range centered about 0.4 ym
and 0.2 ym, respectively. The size range, mass loading, and temperature
ranges are 0.1 to 10 ym, 4 to 400 ppb by volume, and 0° to 260°C.
These principles have been incorporated into a prototype, portable
stack monitor. The probe is 1.5 meter long and 9 cm in diameter, optical
signals are carried through fiber optics cables contained in the probe.
An arc source and silica photodetectors are outside the stack at the end of
the probe, while a digital microprocessor is used to analyze the data. The
microprocessor, air purge system, lamp power supply, and printer are housed
separately from the probe.
Theoretical and experimental programs are continuing to investigate
various approaches to using light scattering as an analytical tool for par-
ticle-size analysis. As progress is made it is anticipated that the data
can be acquired and reduced more conveniently, quickly, and accurately.
300
-------
REFERENCES
1. Smith, W. B., P. R. Cavanaugh, and R. R. Wilson
Technical Manual: A Survey of Equipment and Methods for Particulate
Sampling in Industrial Process Streams.
Southern Research Institute, EPA, Research Triangle Park, NC, 1978. 281 pp.
EPA-600/7-78-043
2. Smith, W. B., Editor
Proceedings: Advances in Particle Sampling and Measurement.
Southern Research Institute, held in Asheville, NC, 1978. 1979, 389 pp.
EPA-600/7-79-065
3. Ragland, J. R., J. D. McCain, and W. B. Smith
Design, Construct, and Test a Field Usable, Prototype System for Sizing
Particles Smaller than 0.5 Micrometers Diameter.
Southern Research Institute, EPA, Research Triangle Park, NC, 1978. 320 pp.
4. Ensor, D. S., R. G. Hooper, and R. W. Scheck
Determination of the Fractional Efficiency, Opacity Characteristics
Engineering and Economic Aspects of a Fabric Filter Operating on a
Utility Boiler.
EPRI-FP-297, Electric Power Research Institute, Palo Alto, Calif., 1976.
220 pp.
5. Hodkinson, J. R.
The Optical Measurement of Aerosols.
Aero. Sci., pp. 287-357, 1966.
6. Shofner, F. M., G. Kreikebaum, H. W. Schmitt, and B. E. Barnhart
In Situ, Continuous Measurement of Particle Size Distribution and Mass
Concentration using Electro-Optical Instrumentation.
Presented at the 5th Annual Industrial Air Pollution Control Conference,
Knoxville, TN, April, 1975.
7. Gravatt, C. C.
Real Time Measurement of the Size Distribution of Particulate Matter by
a Light Scattering Method
J. of APCA 23(12): 1035-1038, 1973
8. Chan, P. W.
Optical Measurements of Smoke Particle Size Generated by Electric Arcs.
Colo. State University, EPA, Washington, D.C., 1974, 49 pp.
EPA-650/2-74-034
9. Kuykendal, W. B., and C. H. Gooding
New Techniques for Particle Size Measurements.
Presented at the Workshop on Sampling, Analysis, and Monitoring of Stack
Emissions, Oct. 2-3, 1975, Dallas Texas, pp. 183-208.
301
-------
10. Farthing, W. E., and W. B. Smith
Evaluation of the PILLS IV.
Southern Research Institute, EPA, Research Triangle Park, NC., 1978.
51 pp.
EPA-600/7-78-130
11. Knollenberg, R. G.
Active Scattering Aerosol Spectrometry.
Proceedings of Seminar on Aerosol Measurement, National Bureau of
Standards, Washington, D.C., 1974.
12. Particle Measuring Systems, Inc.
Classical Scattering Aerosol Spectrometer Probe Model CSASP-100.,
Boulder, Colo., 1976. 2 pp.
13. Wertheimer, A. L., and W. L. Wilcock
Light Scattering Measurements of Particle Distributions.
Appl. Optics 15(6): 1616-1620, 1976.
302
-------
METHODS AND DEVICES FOR MEASURING INDUSTRIAL GASES'
DEW POINT
Measuring the dew point of industrial gases has great sig-
nificance both in industrial processes as well as in gas clean-
ing. For example, operative 'monitorina ofrthe dew point is necessa-
ry in the coal industry for retrieval of rare elements from i-.
coal during its burning in systems equipped with filter bags
made from "nitron" fabric. The work of these systems is impos-
sible without dew point monitoring due to the high sensitivity of
the fabric to concentrating sulfuric acid. Continuous "monitoring of
the dew point is also necessary in coal-enriching plants for
efficient control of corrosion in the elements of the gas-clean-
ing equipment.
In many branches of the industry, drying proceeds accor-
ding to the temperature of the waste gases. Continuous monitor-
ing of the dew point allows -control of the drying process with
regard to the degree of the stack gases' moisture content,
which, apparently, is significantly more accurate than controlling
the exhaust gases by temperature.
It has been established that corrosion most markedly ap-
pears when the surface temperature is close to the exhaust gases'
dew point temperature for gases containing 803.
Implemetation of dew point operative monitoring permits the
introduction of efficient corrosion control means such as, for
example, slightly raising exhaust gas temperature, burning fuel
with a low excess air ratio and introduction of an admixtures
neutralizina SO-?.
303
-------
Chemical methods for determining the SO, concentration are complicated
and cumbersome, and are therefore unusable for operational moni-
toring, whereas the method of evaluating 803 content in stack
gas according to dew point temperature, which is highly sensi-
tive even to insignificant changes of 803 concentration, is
very simple and quick-responding.
Dew point temperature measurement consists of simulta-
neously measuring two values, one of which allows fixing the
moment when condensation begins, and the other determnies the
temperature of the condensation surface at the moment of its
moistening.
Measurement of the condensation surface's temperature
may be done by the usual monitoring and measuring devices
working as a unit with a thermocouple, thermosensitive resistor,
and thermistor, depending upon the range of dew point measure-
ments and the unit's accuracy and speed requirements.
According to the prir.-iple of fixation of the moment
of dew precipitaion on condensation surface, the "'direct'
method of dew point measurement" and correspondingly, the
hygrometer's "dew point" it is possible to make classifi-
cations in the following manner (1):
1. Optical, where occurence of the condensate film
is judged according to variations of the optical proper-
ties of a polished condensational surface at the moment
that dew precipitates on it (2,3);
2. Electrical, where changes in the electrical pro-
perties of a section of the condensational surface, for ex-
ample, conductivity, capacitance, dielectric permeability,
serve as the basis for this method.
3. Electrochemical, based on the occurence of galvanic
304
-------
electromotive force during condensation of electrolyte
vapor in the inter-electrode space, which is at that time
the condensational surface;
4. Thermal, using temperature difference between the
cooler and condensational surface, occuring at the moment of
condensation;
5. Hygroscopic, based on physical regularities of absorb-
tion and release of water by a salt-saturated aoueous solution.
6. Radioactive, based on differences in the dispersion
characteristic of particles in the presence or absence of
condensate.
Analysis of possibilities and expediency for using the
methods of measurement named above for purposes of continu-
ous automatic monitoring of gas dew point temperature shows
that use of the optical method is essentially limited
because of the gas dust content.
The "thermal and hygroscopic methods cannot be
adapted for monitoring hot gases with complicated variable
chemical composition.
The electrical (conductometrie) method, using changes
of ohmic resistance between two electrodes, received
at the condensation surface, had the most widespread
distribution.
The basis of this method gave rise to several domestic
designs of equipment.
Research has shown that the electrical method princi-
pally allows measurement of dew point of hot, multi-component,
gas mixtures, however it is not useful for work in a dusty
environment (dust shorts out its contacts).
305
-------
The most expedient means for measuring the dew point
of hot dusty gases is the electrochemical method of fix-
ing the moment of condensation. However, at the present time,
series-produced equipment is not available.
Dew point hygrometers are also classified according
to their method of cooling the condensing element.
Presently, the following methods of cooling have aained
use:
1. Thermoelectric, based on the Peltier effect;
2.Gas expansion, for example, of C02» freon, etc;
3. Liquid nitrogen, propane or other coolants, for
example, solid CO.;
4. Cooling by use of the vortex effect of power energy
separation. The complexity of accurately feeding the cooling
agent lowers the operational characteristics and practically
excludes the possibility of fully automating cooling by
various cooling agents.
Gas expansion methodology is also sufficiently compli-
cated, since it demands cumbersome equipment for increasing
pressure on the expanded gas.
It is more reasonable to use the vortex effect of
compressed gas power_ separation for cooling. A self-evacua-
ting tube with slit-type over-speed diffusor, providing, by
comparison with vortex tubes of different design, deep refri-
geration is often used. This allows measurement of the dew
point within a +20- -60°C range with a -2°C error.
Using thermoelectrica! units based on the Peltier effect
152, from the point of view of automation, the most conven-
ient method for cooling limited volumes. Multistage
306
-------
thermopiles produce significant (up to 50-70°C) tempera-
ture drops (fig.l.). This method of refrigerating conden-
sing elements is the most widely used one in the USSR (3,6).
Indirect methods are used for determining dew points,
when the dew point in determined by calculation on the basis
of the moisture and pressure of the studied gas (4,5).
Recently, infrared methods for determining dew point
have become widely used, due to their increased sensitivity
and quick response.
An automatic infrared hygrometer, the IFO-459, is pro-
duced in the USSR (7). This is a dual-frequency absorption
modulation (25 hertz) unit with compensation in the'-optical-canal.
It works within a wave length range of 1.4-2.1 u using in-
terference light filters. A lead-sulfur sensory element is
used as a photodetector.-This unitj.is-iused..in various'.bran-
ches: of lindus try during dew point determination, in determin-
ing moisture content in gases and in preparing industrial
air for drying. The unit's high sensitivity and quick response
in all ranges of water steam concentration are its advanta-
ges.
Dew point characteristics for hygrometers used in the
USSR are represented in table 1.
For increasing hygrometer accuracy and operational stability, the
differential and compensational methods are often used. This
is done, for example, when two photodetectors, reacting corre-
spondingly to the light stream reflected from and scattered
by the condensation surface, are used, which provides auto-
matic error compensation against contamination of the conden-
sinq element surface.
307
-------
Under temperatures lower than -30°C, exact determina-
tion of phase condition of the condensate , which is often
not considered and therefore introduces additional errors
in determining dew point, is especially important (8). It is
necessary under these temperatures to consider the different
values of water vapor cressure^ver water and over ice (fig.
2). This feature of water steam determines to a great extent
the metrological characteristics of dew point hygrometers
in the negative temperature range in as much as, for example,
temperature difference reaches 5°C at a saturation temn#»rature
levelC7f3k<*fc£**£-60°C. ice formation always passes through a
liquid phase formation stage As an example, we can cite the photoelect
system for stabilizing the thickness of the condensate. Thanks to
this system, it is possible to obtain a diagramic record (fig. 3) of t
mirror tempera tureen determining the phase diagram of the condensate
and formation of the solid ph?»se in the negative temperature dew point
range forr slow changing~moisture content of the gas under analysis
(the sloping nature of the curve). The points
(1) correspond to dew point temperature at the moment of conden-
sate precipitation in the liquid phase, and points (2) correspond
to the frost point settled temperature at water vapor saturation
over a constant ice layer. Any states between these two points
are characterized by by the presence of the condensate's mixed
phase. The absence of temperature difference 'between the points
1 and 2 (i.e. the absence Tjfara "peak" in the diagrammatic record) in-
dicates the existence on the mirror of a condensate liquid phase on the
surface during the entire period of observation, which occurs at
temperatures above -30°C.
308
-------
Difficulties in determining dew point at temperatures higher
than 100°C are commonly known (1). In this temperature range
absorption-type automatic hygrometers without sampling
using seirdLfconductor-emitters can be widely applied.
The latter are characterized by lew cost, reliability, compactness,
working convenience, possibility for modulating emission by
feed current, large dynamnic range for re-tuning out-put
capacity, long period of service (104 hrs.), small size
of emitting area (less than 1mm2), feasibility of coverage with
the aid of selection of composition of semi-conductive material
of wide spectral range (0.56-20 microns), relatively large
out-put capacity (l.Omilliwatts-10 watts) (9). It is also
necessary to consider the appearance of new types of semi-con-
ductor lasers working at room temperature. Multi-wave matrices
of semi-conductor emitters that may significantly simplify hygro-
meter design, especially for dual-frequency hygrometers, find a
special place in use. Their hj.gh response ("10 newton .seconds)
allows operation in impulse conditions or with use of high
frequency narrow-band modulation ( 30megahertz), which may
significantly increase the signal-noise ratio.
The absorptional optical method for determining dew point
without sampling may be accomplished in the unit whose
lay-out is shown in fig.4.
In this instrument the amount of water vapor in the gases
is determined by comparison with *a standard amount in the cali-
brating cell. Also measuring pressure in the gas conduit by con-
version exactly .'determines its dew-point. The optics are protected
from pollution by traditional means used for optical dust measure-
ment.
309
-------
It is necessary to note the common features of optical
absorptional hygrometers without sampling: quick response,
due to using quick-response optic-electronic elements; high measure-
ment exactness from the system's lack of sampling.
Water vapor molecule concentration value, measured by the
optical absorption method is described by the following expression:
o.- 2
n" * j.
-------
* • i
measuring head
Fig? 1
Dew-point hydrometer -with -thermoelectric cooling.
1. Copper rod. 2;. nichrome spiral. 3. metal mirror. 4. resis-
tance thermometer. 5. thermocouple. 6. thermopile. 7. radiator.
8,9. lenses. 10,ll.photoresistors. 12. light source. 13. eyepiece.
14. magnifier. 15,18. rectifiers. 16. chamber. 17. rotameter.
19. microampermeter. 20. dew point register. 21. control elements.
*
22. thermopile supply unit.
311
-------
i i i I i i
- " r- ; — r-v f _ e
' l /
• 4. ! /
7
/
i j YW \ i
• i
-2
Fig.2. Dependence of water vapor pressure
on temperature.
1. over water
2. over ice
312
-------
o •
vo j
JJ
•H
O
o
rH
0)
cooling
eating
cooling
eating
oolina
heating
r- cooling
-30 -3) -fG O^C
Fig.3. Diagrammatic record of mirror temperature.
1. Corrsponds to dew point temperature at the
nioiaent .oi condensate release in the fluid phase.
2. Corresponds to justained frost point temperature
during wa-cer_vapor saturation under a constant ice
313
-------
Pig. 4. Unit diagram - dew point hygrometer without sampling.
1,11. reflectors.
2. calibrating tray.
3. gas duct.
4. pumping generator.
5.- semi-conductor, emitter.
6. modulator.
7. beam splitter.
8. recorder.
9. amplifying unit.
10. photoreceptor.
314
-------
IU'UC-J>LH J'CU' ' of dew
• 1 point iMMHuri'rrnt t
| J "C
V'i* 1 double licnm fj.in rx|MMr< *c l_ _i'irar.inuri'mi-nt , mil
0- i m Kirn ! C.« ! k.i's j
! ! v ! ! _
^2 0.5-6 i 50 0,0o'10~3
i O.* - 30 * * tO
IT M> - 150 "' 52m'/hr
4 BtcAm 2 w /hr
3i» 30 - 100
m
iH
en
i 0.5 1-2 >/m»n 2
slnqlv bc.im
m-if 1 -c-ondiu-t lv«
thormooliTK'nt
- 20 * + 20
100
1-3 I/Mm
3-5
iro-4'.l
t loril
option! (tenuity
0.002
0.05
0.2
451
continuous
-------
BIBLIOGRAPHY
1. Berliner, M.A., Moisture Measurement, Moscow, Energiia, 1973.
2. Merkulov, A.P., Kolyshev, N.D., Netsetaev, V.A., Karshin,
A.I., "Basic Results of the Development and Testing of Dew
Point Hygrometers with Vortex Cooling", in "Equipment and Means
of Automation", 1974, no.10, p.21.
3. Dew Point Hygrometer M-116. Prospectus of the Achievements
of the Soviet Economy.
4. Automatic Psychrometric APG-203 Hygrometer. Information leaflet
no. 215-76, Interdisciplinary information, TsNIITEI Institute of
Instrument Building, 1976.
5. APV-201-type Psychrometric Automatic Moisture-Gage, Prospectus
of the Achievements of the Soviet Economy, TsNIITEI Institute of
Instrument Building, 1977.
6. Karpov, A.K. et al., "The Automatic Dew Point Hygrometer
GA-1", in the collection "Industrial and Plant Processing of
Natural Gas", Moscow, 1976, p.63-72.
7. Togulev, V.P., Ustimov, IU.N., Gross, L.G., Petrov, IU.A.,
"Automatic Infrared Hygrometer11, "OMP", 1977, no.3, p.25.
8. Petukhov, V.P.,."Determination of the Condensate's Phase
State in Automatic "Dew Point" Hygrometers", "Measurement
Technique", 1977, no.l, p.72."
9. Maniurov, A.I., "Perspectives on the Use of Semi-Conductor
Emitting Diodes Made From Solid Solutions for the Quantitative
Evaluation of Harmful Contaminants", in the interinstitutional
collection "Ecology and Work Safety in Radioelectronics", Moscow,
1977-
10. Zuev, V.E., Propagation of Visible and Infrared Waves in
the Atmosphere, Moscow, "Sov. Radio", 1970.
316
-------
DESIGN OF AN IMPROVED IMPACTOR
W.B. Smith
K.M. Gushing
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35205
D,B. Harris
t
Industrial Environmental Research Laboratory
Environmental Protection Agency
Research Triangle Park, N.C. 27711
(This document has been reviewed and approved
for publication by the U. S. EPA.)
317
-------
DESIGN OF AN IMPROVED IMPACTOR
Over the past ten years cascade impactors have become
commonly used measurement devices for the determination of
particulate size distributions of industrial effluents. Data
obtained with impactors are used to characterize particulate
emissions from sources, to determine the performance of particu-
late control devices, and in the selection and design of control
devices for specific sources.
Figures 1 and 2 show schematic diagrams of five commercial
cascade impactors which are commonly used in stack sampling.l' z' 3'"'s'6'7
Particle separation by size interval takes place within these
impactors by passing the sample gas stream sequentially through
a series of dry impingement type inertial classifiers. The classi-
fiers operate by impingement of the aerosol stream in an air jet
against a plate, causing the air in each jet to sharply change
direction and flow around the plate. (See Figure 3) Because of
inertia, some particles cannot follow the flow streamlines and are
deposited on the collection plate. Each cascaded impingement
stage in the series operates at a higher impingement velocity
(or as a higher energy separator) than the previous stage. Thus
finer and finer particles are removed by the various stages as the
sample gas stream proceeds through the impactor. An absolute fil-
ter is usually positioned after the final impaction stage to
collect all remaining particulate.
Recently many researchers have undertaken investigations
to better characterize cascade impaction operation. Both the
development of better theories of operation8'9'10'11 and the
explicit calibration of existing commercial and prototype de-
vices12'13'11* has led to a better understanding of the interpre-
tation and accuracy of cascade impactor data. The basic
318
-------
INLET J6T
STAGE NO. 1
^RECOLLECTION
CYCLONE
FILTER
MPACTOR BASE
JET ITAGE
(7 TOTAL!
COLLECTION
PLATE
MRI MODEL 1502
MODIFIED BRINK
INLET
COLLECTION
PLATE 17 TOTAL)
COLLECTION PLATE
FILTER HOLDER
UNIVERSITY OF WASHINGTON MARK III
Figure 1. Schematics of three commercial cascade impactors
319
-------
INLET CONE
SIERRA MODEL 220
JET STAGE 19 TOTAL)
ANDERSON MARK III
Figure 2. Schematics of two commercial cascade impactors.
320
-------
PATH OF
I SMALL PARTICLE
3630-230
Figure 3. Schematic diagram, operation of cascade impactor
321
-------
limitation of cascade impactors is that they provide data with
only relatively low resolution and do not permit the exact re-
construction of the aerosol being sampled, even over the limited
range of sizes normally covered by most impactors (approximately
0.5 to 10.0 microns). Improved methods of data reduction may
reduce the reconstruction error.
The theory of the impaction process has been developed by
several researchers to a state where the efficiency, E, of im-
paction can be estimated as a function of the particle size
(D ) , Reynolds Number (Re), the jet diameter or width (D. ,W),
XT J
the jet to plate distance (S) , and the jet throat length (T) .
(Marple9 specifically uses the dimensionless reduced quantities
(S/D. ,S/W) and (T/D.,T/W) when considering the jet to plate
distance and the jet throat length, respectively.)
E = E (Dp, Re, S/Dj, T/Dj)
It is common practice to relate the particle diameter, D ,
to the square root of the inertial impaction parameter (/\JJ) ,
which is the ratio of the particle stopping distance (S,) to the
jet diameter or width (D. or W) . Thus,
or
* -
where C = Slip Correction Factor,
VQ = Jet Velocity (cm/sec) ,
U = Gas Viscosity (poise) ,
p = Particle Density (gm/cm ) , and
D. = Jet Diameter (cm).
The square root of the inertial impaction parameter,
is used in impaction theories as a dimensionless quantity pro-
portional to particle diameter,
322
-------
-D
P 18 |i D..
The inertial irapaction parameter is useful in graphing impac-
tor calibration data because information from all stages of an
impactor can be placed on a single graph, and under many circum-
stances would, in theory, lie along a common curve. The value of
/ijJ at 50% collection efficiency, /1JJT7, defines the impaction stage
D50, which is the particle diameter for which half of the particles
are collected and half are passed to the next stage. For data
reduction purposes the Dso is used as the effective stage cut
diameter.
Recently Marple9 has been able to calculate theoretical im-
paction efficiency curves for several values of the jet to plate
distance, jet Reynolds Number, and jet throat length. Figure 4
shows the results of these calculations for both round and rec-
tangular jet impactors. Using the data from Figure 4, a design
chart for round jet cascade impactor can be formulated as shown
in Figure 5.
Several basic problems have become areas of concern for re-
searchers developing new cascade impactors, as well as users of
current commercial devices. Several of these problems are dis-
cussed below:
1. Non Ideal Size Discrimination - Ideally each stage of a cascade
impactor would remove all particles larger than a critical size
and pass to the next stage everything smaller than that size.
In reality the fractional efficiency curve for stage collection
is a smoothly varying function which passes through the 50% effi-
ciency point at a particle size designated as the stage cut point
or D50. Log normal approximations to stage efficiency curves
generally have standard deviations on the order of 1.3. The
overlapping stage efficiency curves result in particles of one
particular size being collected on several impaction stages.
This drastically complicates the data analysis procedure and the
recovery of a reliable particle size distribution.
323
-------
100
9? 80
tu
>
u
60
UJ
- 40
u.
20
S/Dj(T/Dj=1/2)
Round
i
— — — Rectangular
i l i
0.1
0.2 0.3
0.4
0.5 0.6
0.7
0.8
(a) EFFECT OF JET TO PLATE DISTANCE (Re=3.000)
0.8
VT"
(b) EFFECT OF JET REYNOLDS NUMBER (T/W=1)
0.8
(c) EFFECT OF THROAT LENGTH (Re=3,000)
Figure 4. Theoretical stage collection efficiency versus
for round and rectangular jet cascade impactors
(After Marple.9)
324
-------
I
cf
ui
oe
W - Jet Diameter
Re - Reynolds Number
C - Cunningham Slip Correction
Particle Aerodynamic Dia.
at 50% Cut Point
10°
101
10° L
50 100
NUMBER OF ROUND JETS PER STAGE, n
500 1000
3630-232
Figure 5. Design chart for round impactors (Dso = aerodynamic
diameter at 50% cut point.) After Marple.
325
-------
2. Particle Bounce and Reentrainment - In an ideal impactor re-
moval at each stage would become 100% at a certain size and remain
constant for all larger particles. However, recent data12'1" has
shown most stage collection curves approach but never reach 100%,
and roll over after a certain size and become less efficient for
larger sizes. The fact that the particles do bounce and become
reentrained can be attributed to several reasons including the
type of collection surface, the nature of the particle, the jet
velocity, and jet air flow interference on the collection surface.
Each of these mechanisms contribute to particles not being cap-
tured initially and to subsequent particle reentrainment. This
is evident in the calibration data shown in Figure 6 for an
Andersen Mark III Cascade Impactor. Research has shown that these
problems can be minimized by the use of collection substrates16
(glass fiber mats or thin grease layers), the lowering of stage
jet velocities, and the careful arrangement of multiple jets
to reduce interference effects which cause blow off of the cap-
tured particles and their subsequent reentrainment. Because the
collection of fine particles requires high jet velocities, lower-
ing of jet velocities to improve capture efficiency reduces the
size resolution of the device. Usually an attempt is made to
reduce the jet velocities on the upper stages where particle
bounce and reentrainment is more severe and causes a greater
deformation of the recovered size distribution. The improvement
of operation resulting from more jets and/or smaller jets is
also bounded by the physical limits of placing many jets on a
small jet plate and the ensuing increase in machining and tool-
ing costs.
3. Wall Loss - Particle loss to the internal walls of impac-
tors is an unavoidable situation which can be minimized by re-
ducing surface area. Maintenance of proper flow streamlines
and avoiding abrupt bends in the particle paths are also nece-
ssary. Particle wall loss is also attributable to settling and
326
-------
u
tu
g
u
tu
_i
8
99.8
99
98
95
90
80
70
60
50
40
30
20
10
5
2
1
0.5
I I
I TT
0.2
0.03
STAGE SYMBOL
I I I
0.05 0.08 0.1
0.2
0.4 0.6 0.8 1.0
2.0
Figure 6. Stage collection efficiency versus /\|> for Andersen
Mark III Stack Sampler. (Flow rate = 14.2 1pm,
particle density = 1.00 gm/cm3, temperature = 20°C)
327
-------
particle bounce, Van der Waal's forces, and, to a smaller degree,
diffusion. Frequently particle bounce results in a dust pattern
forming on the underside of the jet plates. This type of wall
loss is an indication of poor impactor design, stage overloading,
or high jet velocities. Recent wall loss data12 for commercial
cascade impactors is shown in Figure 7.
Based on the current knowledge of impactor theory and the
result of field and laboratory testing of currently available
commercial and prototype cascade impactors, high and low flow
rate impactors are being designed and fabricated that attempt
to conform to our criteria for optimized cascade impactors list-
ed in Table 1. An optimized cascade impactor can be defined as
one which is designed to operate both in a laboratory or field
environment in such a way that the effects of particle bounce,
particle reentrainment, wall loss, overlapping stage collection
efficiency curves, etc. are minimized. Mechanical reliability
and case of operation are also considered in this design program.
Design parameters for 0.1, 0.5, and 2.0 ACFM, optimized,
cascade impactors have been generated by a special computer
program written at Southern Research Institute. Six impaction
stages with cut points of 9.6, 4.8, 2.4, 1.2, 0.6, and 0.3 mi-
crometers aerodynamic diameter were chosen as a design goal for
these cascade impactors. The actual theoretical stage D50's as
calculated by the computer program were very close to the desired
stage cut points, but varied slightly to limit the number of holes
per stage and to allow the use of standard drill sizes. The
final design parameters are presented in Tables 2, 3, and 4.
An assembly drawing of the 0.5 ACFM impactor is shown in Figure
8.
A prototype of the 0.5 ACFM optimized cascade impactor is
currently being fabricated by Southern Research Institute for
calibration and testing purposes.
328
-------
2
O
IV
60
50
40
30
20
10
5 ^
1 1 1
—
—
—
••••»
_
a
o
u * $
* *
1
0.5
0.2
w
—
—
•••••»
III
1 1 1 1 1 1 1 d>
ff -
O
V A —
V A 0
oo —
A A
0 0 ""
D —
*
• __
.a • A . 9
a
^^^m
—
__
1 1 1 1 1 1 1 1
1.5
PARTICLE DIAMETER, micrometers
20
O ANDERSEN MARK Ml STACK SAMPLER. NONISOKINETIC SAMPLING.
D MODIFIED BRINK MODEL BMS-II CASCADE IMPACTOR. GLASS FIBER SUBSTRATES. NONISOKINETIC SAMPLING.
A MODIFIED BRINK MODEL BMS-II CASCADE IMPACTOR. GREASED COLLECTION PLATES. NONISOKINETIC SAMPLING.
V MRI MODEL 1502 INERTIAL CASCADE IMPACTOR. NONISOKINETIC SAMPLING.
O SIERRA MODEL 226 SOURCE CASCADE IMPACTOR. 14LPM. NONISOKINETIC SAMPLING.
• SIERRA MODEL 226 SOURCE CASCADE IMPACTOR. 7LPM. ISOKINETIC SAMPLING
A U. of W. MARK III SOURCE TEST CASCADE IMPACTOR. NONISOKINETIC SAMPLING.
Figure 7. Recent cascade impactor particle wall loss data.
After Gushing, et al.12
329
-------
TABLE 1. CRITERIA FOR IMPACTOR DESIGN
1. The jet Reynolds number should be between 100 and 3000.
2. The jet velocity should be 10 times greater than the settl-
ing velocity of particles having the stage Dso.
3. The jet velocity should be less than 110 m/sec.
4. The jet diameter should not be smaller than can be attained
by conventional machining technology.
5. The ratio of the jet-plate spacing and the jet diameter or
width (S/W) should lie between 1 and 3.
6. The ratio of the jet throat length to the jet diameter (T/W)
should be approximately equal to unity.
7. The jet entires should be streamlined or countersunk.
330
-------
co
TABLE 2
0.1 ACFM OPTIMIZED CASCADE IMPACTOR DESIGN SPECIFICATIONS
Impactor Sampling Rate =0.1 ACFM
Stage D50 - Design Goal (ym)
Jet Inlet Pressure (atm)
(Ambient=0.987 atm)
Best Jet Drill Diameter (cm)
Number of Jets on this Stage
Actual Jet Reynolds Number
Jet Velocity (m/sec)
Actual Square Root of Stokes #
Cunningham Correction Factor
Computed Stage D50
Gas Temperature = 204.0°C Particle Density =1.0 gm/cc
9.6 4.8 2.4 1.2 0.6 0.3
0.987
0.47
1
362.6
2.72
0.478
1.017
9.474
0.986
0.30
1
568.0
6.68
0.476
1.034
4.776
0.985
0.19
1
895.3
16.65
0.475
1.070
2.359
0.975
0.098
2
859.2
31.29
0.475
1.145
1.195
0.953
0.0254
31
209.0
30.05
0.480
1.297
0.589
0.921
0.0254
19
396.9
71.05
0.477
1.777
0.297
-------
CO
Co
TABLE 3
0.5 ACFM OPTIMIZED CASCADE IMPACTOR DESIGN SPECIFICATIONS
Impactor Sampling Rate =0.5 ACFM Gas Temperature = 204.0°C Particle Density = 1.0 gm/cc
Stage D50 - Design Goal (pro) 9.6 4.8 2.4 1.2 0.6 0.3
Jet Inlet Pressure (atm) 0.987 0.986 0.981 0.970 0.931 0.901
(Ambient=0.987 atm)
Best Jet Drill Diameter (cm) 0.794 0.518 0.206 0.107 0.025 0.016
Number of Jets on this Stage 114 8 155 235
Actual Jet Reynolds Number 1073.6 1643.4 1030.1 981.5 204.4 206.9
Jet Velocity (m/sec) 4.77 11.19 17.75 32.99 30.05 49.95
Actual Square Root of Stokes # 0.474 0.471 0.474 0.474 0.480 0.480
Cunningham Correction Factor 1.017 1.035 1.070 ' 1.146 1.304 1.707
Computed Stage Dso (pm) 9.221 4.797 2.372 1.212 0.588 0.316
-------
co
CO
CO
TABLE 4
2.0 ACFM OPTIMIZED CASCADE IMPACTOR DESIGN SPECIFICATIONS
Impactor Sampling Rate =2.0 ACFM Gas Temperature = 204.0°C Particle Density = 1.0 gm/cc
Stage Dso - Design Goal (ym) 9.6 4.8 2.4 1.2 0.6 0.3
Jet Inlet Pressure (atm) 0.987 0.986 0.983 0.973 0.934 0.849
(Ambient=0.987 atm)
Best Jet Drill Diameter (cm) 0.794 0.404 0.206 0.107 0.045 0.022
Number of Jets on this Stage 4 8 17 32 119 355
Actual Jet Reynolds Number 1073.6 1054.1 971.0 984.3 602.5 375.6
Jet Velocity (m/sec) 4.77 9.21 16.71 32.99 49.88 69.95
Actual Square Root of Stokes # 0.474 0.474 0.474 0.474 0.476 0.478
Cunningham Correction Factor 1.017 1.035 1.070 1.145 1.324 1.844
Computed Stage D50 (ym) 9.221 4.694 2.448 1.212 0.598 0.300
-------
J-^y^j./ J&f
zf't 9f>3 .r^A-r1
sounra BUKH~L
M, AlAMHA ISMS
'. ? *•'-* fy/wr~J -3'*w'f'S'
ftTj-0 ^ s/
Figure 8. Assembly drawing of the 0.5 ACFM, six-stage,
optimized cascade impactor.
-------
REFERENCES
1. Andersen A. A.
A Sampler For Respiratory Health Hazard Assessment.
Amer. Ind. Hyg. Assoc. J., pp. 160-165, 1966.
2. Brink, J. A., Jr.
Cascade Impactor For Adiabatic Measurements.
Ind. And Eng. Chem., 50(4), pp. 645-648, 1958.
3. Jackson, B., and D. Ensor
Preliminary Calibration Of The M.R.I. Inertial Impactor
Meteorology Research, Inc., M.R.I. Research & Development,
10 pp., 1974.
4. Lundgren, D. A.
An Aerosol Sampler For Determination Of Particle Concentration
As A Function Of Size And Time.
J. of APCA, 17(4), pp. 225-259, 1967.
5. May, K. R.
A Cascade Impactor With Moving Slides.
A.M.A. Archives Of Ind. Health, 1956, Vol. 13, pp. 481-188.
6. Pilat, M. J., D. S. Ensor, and J. C. Bosch
Cascade Impactor For Sizing Particulates In Emission Sources.
Amer. Ind. Hyg. Assoc. J., 32(8), pp. 508-511, 1971.
7. Willeke, K.
Performance Of The Slotted Impactor.
Amer. Ind. Hyg. Assoc. J., 36(9), 683-391, 1975.
8. Cohen, J. J., and D. N. Montan
Theoretical Considerations, Design, And Evaluation Of A Cascade
Impactor.
Amer. Ind. Hyg. Assoc. J., pp. 95-104, 1976.
9. Marple, V. A.
A Fundamental Study Of Inertial Impactors.
Dissertation, Univ. of Minn., University Microfilms, High
Wycomb, England, 1970, 243 pp.
10. Mercer, T. T. , and R. G. Stafford
Impaction From Round Jets.
Ann. Occup. Hyg., 12, pp. 41-48, 1969.
11. Ranz, W. E., and J. B. Wong
Impaction Of Dust And Smoke Particles On Surface And Body
Collectors.
Ind. And Eng. Chem., 44(6), pp. 1371-1381, 1952.
335
-------
12. Gushing, K. M., G. E. Lacey, J. D. McCain, and W. B. Smith
Particulate Sizing Techniques For Control Device Evaluation:
Cascade Impactor Calibrations.
Southern Research Institute, EPA, Research Triangle Park, NC,
1976, 94 pp.
EPA-600/2-76-280
13. Dzubay, T. G., L. E. Hines, and R. K. Stevens
Particle Bounce Errors In Cascade Impactors.
Atmos. Environ. 10, pp. 229-234, 1974.
14. Rao, A. K.
An Experimental Study Of Inertial Impactors.
Dissertation, Univ. of Minnesota, 1975, 194 pp.
15. Marple, V. A., and K. Willeke
Impactor Design.
Atmos. Environ., 10, pp. 891-896, 1976.
16. Felix, L. G., G. I. Clinard, G. E. Lacey, and J. D. McCain
Inertial Cascade Impactor Substrate Media For Flue Gas
Sampling.
Southern Research Institute, EPA, Research Triangle Park, NC,
1977, 89 pp.
EPA-600/7-77-060
336
-------
TECHNICAL REPORT DATA
(Please read Inunctions on the reverse before completing)
• REPORT NO
EPA-600/9-85-011
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE "
Third U. S. /U. S.S. R. Symposium on Particulate
Control*
5. REPORT DATE
April 1985
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Miscellaneous
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
NA
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
NA (Inhouse)
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings; September 1979
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES AEERL project officer is Jaroslaw Pekar, Mail Drop 60, 919/541-
3995. (*) Symposium held in Suzdal, U. S. S. R., September 10-12, 1979.
16. ABSTRACT
The proceedings document the Third U. S. /U. S. S. R. Symposium on Parti-
culate Control, September 10-12, 1979, in Suzdal, U. S.S.R. Papers covered such
topics as: predicting back-corona formation and fly ash resistivity, improved elec-
trostatic precipitator (ESP) mathematical modeling, calculating effects of back coro-
na in wire-duct ESPs, chemical conditioning of flue gas before an ESP on a 500-MW
power unit, sodium conditioning tests with EPA's mobile ESP, fly ash dielectric
properties and critical current density, general industrial application of precipitating
electrodes for ESPs, performance analysis of a hot-side ESP, dynamics and strength
of corona electrodes in industrial ESPs, ESPs for use at high temperatures and high
pressures, cleaning hot aspirated air from sinter machines and clinker coolers,
ESP modeling with the TI-59 calculator, removing fly ash at electric power stations,
particle collection by granular bed filters and dry scrubbers, filter materials for
cleaning gases and possible areas of their application, high temperature ceramic
filters, measuring the size of finely dispersed particles less than 0. 3 micrometer
for process gases, particle measurement in the U. S., aerosol dispersion analysis,
automatic monitoring of particle-size distributions, measuring the dew point of
process emissions, and designing an improved impactor.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Particles
Dust
Aerosols
Coronas
Fly Ash
Mathematical Models
Electrostatic Precipita-
tor s
Flue Gases
Measurement
Ceramics
Pollution Control
Stationary Sources
Particulate Matter
Russia
Back Corona
13B
14G
11G
07D
20 C
21B
12 A
131
14B
11B
Z. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
343
20. SECURITY CLASS (Thispage)
Unclassified
22.'PRICE
EPA Form 2220-1 (9-73)
337
------- |