EPA
EPA-600/7-78-037
U.S. Environmental Protection Agency Industrial Environmental Research
Office of Research and Development Laboratory . OTQ
Research Triangle Park. North Carolina 27711 MflTCn 197o
SECOND US/USSR SYMPOSIUM
ON PARTICULATE CONTROL
Interagency
Energy-Environment
Research and Development
Program Report
-------
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 grouping 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 INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, 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 Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-037
March 1978
SECOND US/USSR SYMPOSIUM ON
PARTICULATE CONTROL
Franklin A. Ayer, Compiler
Research Triangle Institute
P. O. Box 12194
Research Triangle Park, N.C. 27709
Contract No. 68-02-2612
TaskS
Program Element No. EHE624
EPA Project Officer: Dennis C. Drehmel
Industrial Environmental Research Laboratory
Office of Energy, Minerals and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
ABSTRACT
The Second U.S./U.S.S.R. Symposium on Particle Control was sponsored by the Par-
ticulate Technology Subgroup of the U.S./U.S.S.R. Stationary Source Air Pollution
Technology Working Group under the U.S./U.S.S.R. Environmental Agreement. Papers
were presented by Soviet specialists from research design institutes and industry, by
representatives of U.S. government agencies, and representatives of the private sec-
tor. Topics included: electrostatic precipitator research and application, electrostatic
precipitator gas flow modeling, electrostatic precipitator rapping and reentrainment
studies, electrostatic precipitator reliability studies, electrostatic precipitator modeling,
flue gas conditioning, high temperature electrostatic precipitator modeling, flue gas
conditioning, high temperature electrostatic precipitator application, use of fabric
filters in the U.S. cement industry, emission standards, state-of-the-art of mist
eliminators, mist eliminator testing, particle size distribution measurement in the
micron and submicron ranges, dust resistivity and back corona formation in elec-
trostatic precipitators, and fly ash composition and its effect on resistivity.
FOREWORD
The proceedings for the U.S./U.S.S.R. Symposium on Measurement and Control
of Particulate Emissions is the final report submitted to the Industrial
Environmental Research Laboratory for the U.S. Environmental Protection
Agency Contract No. 68-02-2612, Task 5. The symposium was held at the
Research Triangle Park, North Carolina, on May 25-28, 1977.
The primary objective of the symposium was to dicuss work done by American
and Soviet specialists on particulate abatement technology.
Mr. James A. Abbott, Chief, Particulate Technology Branch, Utilities and
Industrial Power Division, Industrial Environmental Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina, was
the symposium General Chairman.
Dr. Dennis C. Drehmel, Particulate Technology Branch, Utilities and Industrial
Power Division, Industrial Environmental Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina, was
the symposium Vice-Chairman.
Mr. Franklin A. Ayer, Manager, Technology and Resource Management
Department, Center for Technology Applications, Research Triangle Institute,
Research Triangle Park, North Carolina, was the symposium Coordinator and
Compiler of the proceedings.
11
-------
Table of Contents
PAGE
September 26,1977
Welcoming Remarks 1
N. A. Jaworski
Overview of Electrostatic Precipitator Applications and Research
in the United States 2
L E. Sparks
The Present Status and Prospects of Development of Studies
of Electrical Gas Cleaning in the U.S.S.R 6
I. K. Reshidov (speaker),
I. A. Kizim,
I. V. Yermilov, and
V. M. Tkachenko
Techniques of Modeling Gas Flow in Electrostatic
Precipitators in the U.S 11
Samuel Bernstein
Investigation of Gas Distribution Systems in Electrostatic
Precipitators Collecting Dust and Fly Ash 21
I. E. Idechik,
JA. L. Ginzburg,
M.O. Shteinberg,
V. P. Alecsandrov, and
V.S. Korjagin
N. G. Bulgakova (speaker)
Studies of Particle Reentrainment Resulting From
Electrode Rapping 30
Leslie E. Sparks,
Owen J. Tassicker (speaker), and
John P. Gooch
Investigation of Dust Carryover and Optimization of Regimes
of Electrode Rapping in Electrostatic Precipitators 47
Yu. I. Sanayev,
I. K. Reshidov (speaker), and
I. A. Kizim
September 27,1977
Investigations Aimed at the Increasing of Reliability and Time of
Service of Electrostatic Precipitator Units and Gas Cleaning
Installations as a Whole 53
V. A. Guzayev,
I. V. Yermilov,
L S. Ryhzov, and
V. N. Saksin
V. M. Tkachenko (speaker)
ill
-------
Table of Contents (con.)
PAGE
Study of Operating and Maintenance Requirements of Electrostatic
Precipitators in the U.S.A 59
Dennis C. Drehmel and
Michael Szabo (speaker)
Mathematical Modeling of Electrostatic Precipitator Filters
and Its Application to Industrial Units in the U.S.S.R 66
I. V. Yermilov and
T. I. Dimitrieva
V. M. Tkachenko (speaker)
Mathematical Modeling of Electrostatic Precipitators in
the United States 79
Jack R. McDonald
Grady Nichols (speaker)
The Study of the Components of the Electric Field in
Electrostatic Precipitators 100
V. I. Levitov,
G. F. Mustafin, and
V. M. Tkachenko (speaker)
Flue Gas Conditioning in Coal-Fired Power Plants in
the United States 113
Edward B. Dismukes
The Problem of Collecting Dusts That Cause Reverse
Corona Formation in ESPs 124
S. N. Panev,
V. M. Tkachernko (speaker), and
T. Y. Chelombitko
September 28,1977
Dynamical Behavior of Collecting Plates of Electrostatic
Precipitators and Prediction of Rapping Efficiency 129
V. B. Mescheryakov and
A. I. Zavyalov
I..K. Reshidov (speaker)
Electrical Gas Cleaning at Higher Pressures 140
A. Yu. Valdberg,
V.V. Danilin,
A. G. Ljapin, and
V. M. Tkachenko (speaker)
Air Pollution Control of Emissions From Kilns and Coolers Utilizing
Fabric Filters in the Cement Industry 150
Frank R. Culhane
Eric Brandt (speaker)
iv
-------
Table of Contents (con.)
PAGE
Removal of Harmful Particles From the Exhaust of the Cement Plants 168
G. I. Vodolazskiy
Y. Izmodyendov (speaker)
Emission Standards and Background Document on Control Technology
in the Cement Industry in the U.S.A 174
Gary McCutchen,
George E. Weant, III, and
George Walsh (speaker)
State-of-Art Survey of Mist Elimination in the U.S.A 180
Seymour Calvert
Mist Eliminator Testing at the Shawnee Prototype Lime/Limestone
Test Facility 194
John E. Williams
September 29.1977
Field Measurements of Particle Size Distribution With Inertia!
Sizing Devices 215
N. G. Bulgakova (speaker),
L. Ya. Gradus, and
S. S. Yankovskiy
Sizing Techniques for Submicron Particles 223
Wallace B. Smith and
D. Bruce Harris (speaker)
Dust Resistivity and Reverse Corona Formation in Electrostatic
Precipitators 237
I. K. Reshidov (speaker),
Yu. I. Sanayev, and
I. A. Kizim
A Review of the Influence of Fly Ash Composition on
Resistivity 248
R. E. Bickelhaupt
-------
WELCOMING REMARKS
N. A. Jaworski*
On behalf of the United States Environmental Protection Agency, we
would like to welcome both our U.S. participants and our Soviet partici-
pants to the 2nd U.S./Soviet Symposium on Particulate Control. Many of
you who are attending this symposium were at the first joint symposium
in San Francisco, and while I personally was not at that symposium, I
have heard that it was indeed a success. Since that symposium, consider-
able research effort has been accomplished both in the U.S.S.R. and in
the U.S.
In addition to the research, there has been joint testing by the
Soviets and Americans both here in the U.S. and in the U.S.S.R. The
information that we have gathered through our research programs and our
joint testing programs is valuable to both countries. I am very pleased,
being a member of the Working Group involved in this particular area of
research, that we can have the opportunity to host this symposium.
The papers to be presented in this program reflect the excellent
work being conducted both in the U.S.S.R. and in the U.S. The program
also reflects some of the cooperative effort that has been conducted
through the U.S./U.S.S.R. Joint Committee on Cooperation in the Field of
Environmental Protection. The program has been in existence for the
past 5 years and has recently been extended for an additional 5 years.
Many of the American participants and Soviet participants in this symposium
have been involved in the particulate abatement project of the Working
Group on Stationary Source Air Pollution Control Technology since its
inception.
After this symposium, research efforts for future joint U.S./Soviet
cooperation will be discussed and developed. It is tentatively planned
to hold the third Symposium on Particulate Control in the Soviet Union
in 1979.
Again, I would like to welcome both the American and Soviet partici-
pants of the joint symposium to the Research Triangle Park, North Carolina,
and especially to our laboratory, the Industrial Environmental Research
Laboratory.
/
The general chairman of the program is Mr. Jim Abbott and he has
done an excellent job working with our Soviet colleagues in organizing
this symposium. I am sure we will have a successful exchange of information
during the symposium and afterwards in planning our efforts for the next
symposium and future cooperative endeavors.
*Deputy Director, Industrial Environmental Research Laboratory,
Research Triangle Park, N.C.
-------
OVERVIEW OF ELECTROSTATIC PRECIPITATOR APPLICATIONS
AND RESEARCH IN THE UNITED STATES
L. E. Sparks*
Abstract
A broad overview of major electrostatic precipitator applications and
-research program in the United States is presented in this paper. Major
emphasis is on coal-fired utility voiler applications and research sponsored
by the Environmental Protection Agency 's Industrial Environmental Research
Laboratory, Research Triangle Park, N. C. Other applications and research
sponsored by Electric Power Research Institute are discussed, but in less
detail.
INTRODUCTION
The purpose of this paper is to provide a broad overview of the major
electrostatic precipitator (ESP) applications and research programs in the
United States. The major emphasis of the paper will be on coal-fired utility
boiler applications and on the research sponsored by the Environmental Pro-
tection Agency's (EPA) Industrial Environmental Research Laboratory at Re-
search Triangle Park, N. C. (IERL-RTP). The reason for this emphasis is:
coal-fired utility boilers are the largest single application of ESP and IERL-
RTP is the major sponsor of ESP research.
Many of the topics touched on in this paper will be discussed in detail
in other papers presented at this symposium.
APPLICATIONS
Coal-Fired Utility Boilers
More than 95 percent of all particulate control devices now in operation
on large utility boilers are electrostatic precipitators. The reason for this
dominance is that electrostatic precipitators generally have lower operating
and maintenance costs than scrubbers and filters. Also, in cases where the
electrical resistivity of the fly ash is not excessive (say 5 x 1010 ohm/cm
or less at 150° C), the capital costs of electrostatic precipitators may be
less than those of competing devices.
The increased use of low-sulfur coal in utility boilers has resulted in
increased interest in hot-side electrostatic precipitators and fly ash con-
ditioning as techniques for overcoming the adverse effects of the high elec-
trical resistivity of fly ash from low-sulfur coal. Several firms are selling
conditioning systems and conditioning agents. Sulfur trioxide is still the
major conditioning agent, but several proprietary agents such as those de-
veloped by Appollo chemical are being used. Available data, e.g., Sparks
*Particulate Technology Branch, Environmental Protection Agency, In-
dustrial Environmental Research Laboratory, Research Triangle Park, N.C.
-------
1976 (ref. 1), indicate that conditioning agents can improve collection of
high-resistivity fly ash. It appears that conditioning is a reasonable solu-
tion for retrofit situations.
Hot-side electrostatic precipitators (the electrostatic precipitator is
located ahead of teh air heater) have been successfully used on several boil-
ers firing low-sulfur coal. In general, the success of hot-side electrostatic
precipitators located in the Eastern U.S. has been greater than the success of
those located in the Western U.S. However, hot-side precipitators capable of
meeting EPA standards have been successfully operated in both parts of the U.S.
At present, there is no clear cut preference among all utilities and vendors
for either hot-side or large cold-side electrostatic precipitators for low-
sulfur coal applications. Both types of electrostatic precipitators have been
able to achieve very high particulate removal efficiencies and both have their
proponents (ref. 1).
Other Applications
Although large utility boilers are the major application of electrostatic
precipitators, they are also widely used in other industries.
Hot electrostatic precipitators are used to control the emissions from
cement plants. Nichols and McCain, 1975 (ref. 2), report performance data
for such an application which show that very high particulate removal effi-
ciencies are possible. High-efficiency electrostatic precipitators are used
in the pulp and paper industry with good success, (see -Gooch et al., 1976
[ref. 3] and Paul, 1975 [ref. 4]). Electrostatic precipitators are widely
used in both the ferrous and nonferrous metal industries. Their use in these
industries was the topic of an EPA sponsored symposium in 1977. Wet electro-
static precipitators have been successfully applied to several industries,
(see Gooch and Dean, 1976 [ref. 5]). Other applications of electrostatic
precipitators are reviewed by White 1975 (ref. 6).
Two-stage electrostatic precipitators have successfully applied to many
sources where hydrocarbon mists must be collected. These two-stage electro-
static precipitators have been able to eliminate the blue,haze from many
operations.
ELECTROSTATIC PRECIPITATOR DESIGN
The design and selection of electrostatic precipitators for specific ap-
plications is difficult and it is generally based on emperical relationships.
Each vendor, consultant, and user has his own formulas. Often, each of these
formula predicts a different specific collector area for the same application.
Several investigators have attempted to develop mathematical models to bring
some order to the existing chaos. The most successful of these models (suc-
cess being defined as an ability to predict both overall mass and fractional
efficiencies) is the EPA model developed by Southern Research Institute and
described by Gooch et al (ref. 7).
-------
RESEARCH AND DEVELOPMENT
Because dusts with high electrical resistivity are difficult to collect
in electrostatic precipitators, most of the research sponsored by EPA and
others is aimed at one or more aspects of the resistivity problem. Research
on high resistivity generally falls into three areas—prediction of resistiv-
ity, reduction of resistivity by conditioning, and elimination of resistivity
problems by redesign of electrostatic precipitators.
Bickelhaupt (refs. 8,9) has had some success in predicting electrical
resistivity of fly ash from coal and ash chemical composition data. However,
the role of sulfur trioxide in resistivity has not yet been successfully in-
corporated in his work. The Electric Power Research Institite (EPRI) is
sponsoring resistivity prediction work based on minicombustors.
EPA has developed a fly ash conditioning system based on sodium carbon-
ated injection. Several companies, as mentioned earlier, have developed pro-
prietary conditioning agents. The overall environmental impact of fly ash
conditioning is a major concern. EPA is sponsoring research to determine what
the environmental impact may be. This program is just underway and only one
test reported by Sparks (ref. 10) has been completed.
Perhaps the most exciting research is that aimed at defeating resistivity
caused problems by redesign of electrostatic precipitators. EPA, EPRI, and
others are sponsoring work in this area. EPRI has announced plans for large
pilot scale tests of their system this year. The EPRI system is based on a
high intensity ionizer developed by Air Pollution Systems in Seattle, Wash-
ington.
The system being developed by EPA is based on particle charging work con-
ducted at Southern Research Institute. The EPA system is being evaluated at
small pilot plant scale which will be followed by large pilot scale demonstra-
tion. Successful completion of these two projects will significantly reduce
the cost of controlling particulate emissions from sources that produce high
resistivity dusts.
EPA and EPRI have both sponsored research on rapping-reentrainment in
electrostatic precipitators. This work is discussed in a paper to be given
later during this symposium.
EPA is sponsoring research to allow prediction of the electrical opera-
ting point of full-scale precipitators. EPA is also sponsoring research to
determine the applicability of electrostatic precipitators for'control ling
particulate in advanced energy processes. Most of these processes require
particulate cleanup at high temperatures and pressures. This work is not yet
complete, but preliminary results indicate that the basic physics are fav-
orable.
CONCLUSIONS
Electrostatic precipitators are an Important tool in keeping the environ-
ment clean. The overall outlook is that they will continue to be chosen for
high-efficiency particulate control of large gas volumes.
-------
The successful completion of ongoing research is likely to improve the
performance and reduce the cost of electrostatic precipitators--especially for
high-resistivity dusts.
REFERENCES
1. L. E. Sparks, "Electrostatic Precipitator Options for Collection of High
Resistivity Fly Ash," Proceedings Conference on Particulate Collection
Problems in Converting to Low Sulfur Coals. G. B. Nichols, Compiler,
EPA-600/7-76-016, NTIS PB 260-498/AS, 1976.
2. G. B. Nichols, and 0. D. McCain, "Particulate Collection Efficiency
Measurements on Three Electrostatic Precipitators," EPA 600/2-75-056,
NTIS PB 248-220/AS, 1975.
3. J. P. Gooch, 6. H. Marchant, and L. G. Felix, "Particulate Collection
Efficiency Measurements on Electrostatic Precipitator Installed on a
Paper Mill Recovery Boiler," EPA-600/2-76-141, NTIS PB 255-297/AS, 1976.
4. J. E. Paul, "Application of Electrostatic Precipitators for the Control
of Fumes from Low Odor Pulp Mill Recovery Boilers," Symposium on Electro-
static Precipitators for the Control of Fine Particles, C. E. Feezel,
Editor, EPA-650/2-75-016, NTIS PB 240-440/AS, 1975.
5. J. P. Gooch and A. H. Dean, "Wet Electrostatic Precipitator System Study,"
EPA-600/2-76-142, NTIS PB 257-128/AS, 1976.
6. H. J. White, "Role of Electrostatic Precipitators in Particulate Control--
A Retrospective and Prospective View," Symposium on Electrostatic Precip-
itators for the Control of Fine Particles. C. E. Feazel, Editor, EPA-
600/2-75-016, NTIS PB 240-440/AS, 1975.
7. J. P. Gooch, J. R. McDonald, and S. Oglesby, "A Mathematical Model of
Electrostatic Precipitation," EPA 650/2-75-037, NTIS PB 246-188/AS, 1975.
8. R. E. Btckelhaupt, "Influence of Fly Ash Compositional Factors on Elec-
trical Volume Resistivity," EPA 650/2-74-092, NTIS PB 237-698/AS, 1974.
9. "Effect of Chemical Composition on Surface Resistivity of Fly Ash," EPA-
600/2-75-017, NTIS PB 244-885/AS, 1975.
10. L. E. Sparks, "Effect of a Fly Ash Conditioning Agent on Power Plant
Emissions," EPA-600/7-76-027, NTIS PB 262-602/AS.
-------
THE PRESENT STATUS AND PROSPECTS OF DEVELOPMENT OF STUDIES
OF ELECTRICAL GAS CLEANING IN THE U.S.S.R.*
I. K. Reshidov, I. A. Kizim, I. V. Yermilov, V. M. Tkachenko (U.S.S.R.)
Abstract
Research trends in the U.S.S.R. are presented in the area of electric gas
cleaning. The solution of problems on preventing nonpermissible pollution of
the atmosphere is being realized through the study and development of new
types of electrostatic precipitators, the development of units for electric
gas cleaning for various branches of industry, the development of methods for
intensifying the operation of these units, and increasing the efficiency of
operating electrostatic precipitators by redesigning them. NIIOGAZ conducted
a number of studies of electrostatic precipitalor electrode systems, estab-
lished rational geometric dimensions for needle-type coronal electrodes, and
showed their essential advantage in comparison to wire-type electrodes. The
coronal discharge characteristics were determined for various systems of
electrodes with respect to a variation in air temperature from 20° to 400° C
and pressure from 0. 1 to 0.6 MPa. It was shown that it is possible to achieve
high efficiency cleaning of blast furnace gas at higher pressure and temper-
ature in a dry plate type electrostatic precipitator. Procedures have been
worked out for calculating the dynamic characteristics of electrode systems
and conditions for their rapping. Studies are being conducted on intensifying
the operation of high resistivity dust. Studies are also being conducted on
predicting back corona manifestations, mathematical modeling of electrostatic
precipitators, and electric power supply, etc.
INTRODUCTION
Growth of industrial production is accompanied by increased gas volumes
to be cleaned before they are discharged to the atmosphere.
A principal control device used in the U.S.S.R. for particulate col-
lection is the electrostatic precipitator. This device treats about 50-60
percent of the total gas volume to be cleaned. ESPs have found wide usage in
utilities, ferrous and nonferrous metallurgy, rock product applicaton, chemi-
cal and other industries.
A wide range of measures are carried out to solve air pollution problems.
The following are of considerable importance:
1. Study and development of new ESP types.
2. Extend application of ESPs to new sources and improve collection
efficiency.
3. Reconstruction of existing ESPs to improve collection efficiency.
*For further information regarding the material in this paper, please
contact I. K. Reshidov, U.S.S.R."113105, Moscow, M-105, 1st Nagatinsky Pass.,
6, NIIOGAZ.
-------
Study and Development of New Types of ESPs
To meet the goal of environmental protection under the conditions of
growth of industrial production, it is necessary that the newly installed ESPs
provide persistent cleaning efficiency at the level of 99-99.5 percent. In
perspective, there is a need for ESPs with collection efficiency up to 99.9
percent, mainly for industries with high inlet dust concentration and for
high-capacity units. The requirement for installation of more efficient ESPs
makes necessary constant development of more effective and reliable systems.
This development is conducted on the basis of the investigation of electrical
and dynamic characteristics of electrode systems and the study of reliability
and operating life of the system.
Investigation of the characteristics of ESP electrical fields by method
of probe characteristics with and without reverse corona, study of breakdown
voltages and particle drift velocities, and examination of efficiency of new
electrode systems in the industrial conditions makes it possible to find
scientific grounds that account for expediency of using needle electrodes with
rational geometric dimensions. It was shown that electrodes of this type
surpass by far other types of electrodes. They provide low voltages of corona
ignition, higher values of corona currents, and particle drift velocity.
Needle electrodes permit even distribution of corona currents on the surface
of precipitating electrodes, and this decreases the probability of reverse
corona formation.
Based on the results of this investigation, uniform horizontal plate ESPs
UG-type with needle corona electrodes and C-shape precipitating electrodes
were developed and are being manufactured. They have found their application
in various industries. In order to find other ways to improve ESP collection
efficiency, studies of combined electrode systems that form interchanging
fields of corona discharge and electrostatic field are being undertaken. To
investigate such complex fields, a new experimental method to determine vector
component of field intensity was developed. It is based on photographing the
trajectory of a preliminary charged metal ball in the investigated field.
NIIOGAZ has studied characteristics of corona discharge for various elec-
trode systems when the temperature changed from 20 to 100° C and pressure from
0.1 to 0.6 MPa. The possibility of high-efficiency cleaning of blast furnace
gases in a dry plate ESP at a gas pressure of about 0.3 MPa and temperature
230° C was proven at a commercial pilot plant. This is of extreme importance
for utilization of energy from blast furnace gas in utility gas-expansion
turbines.
To achieve reliable and effective performance of ESPs, it is really
important to determine correctly the mass of the rapping hammers in relation
to physical and chemical properties of the dust and geometry of electrodes.
For several years, theoretical and experimental studies of acceleration
and frequency of vibration of electrodes when rapped were conducted in NIIOGAZ.
On the basis of these investigations, a method of calculating dynamical charac-
teristics of electrode systems was developed. This method is now used when
new electrode systems are designed.
-------
Practical application shows that the most important characteristics of
ESP performance is reliability of its instrument operation. Highly reliable
equipment provides stable collection efficiency within the whole period of ESP
operation between overhauls. Studies of the reliability of ESPs in actual
field operating conditions allows development of recommendations for improving
the existing apparatus.
On the basis of these investigations, new types of electrostatic precip-
itators are being developed, i.e., horizontal ESP for industrial application
at temperature up to 330° C. Uniform vertical UV-type ESPs providing an effi-
ciency of about 99.0-99.4 percent are going to be put on serial production.
Hot-side ESPs (T up to 450° C) are planned for use in power generation. An
ESP has been installed before the air preheaters on a power plant burning
high-sulfur, heavy oil. Application of such precipitators, besides reducing
the emission level, increases the efficiency of power unit utilization by
reducing contamination of the air preheater surface. In the near future, a
hot-side ESP will be installed to work out problems connected with collection
of high-resistivity fly ash of ekibastuzsk coak, in which the U.S.S.R. is
rich.
Development of ESPs for Various Applications and Methods to Increase
Parti oil ate Collection Efficiency
Experience in ESP application shows that the auxiliary equipment asso-
ciated with the ESP for most industrial gases is comparatively uniform and
varies mostly in geometrical dimensions. However, the auxiliary equipment for
a specific application must almost always be individually designed.
The installation of an ESP besides actually installing the ESP itself
comprises:
1. gas-distributing system,
2. dust collecting system,
3. gas conditioning system, and
4. power supply system.
These systems can vary even for the same industrial applications due to the
peculiarities of gas feeding and dust removal systems.
While designing the ESP, it is necessary to determine overall dimensions
and capacity of the precipitator for the required collection efficiency. Due
to the fact that installations with the same design requirements are quite
rare, design based on parameters of analogous ESP installations can cause
serious mistakes. Design can be done more precisely with the help of method-
ologies where all the factors influencing the efficiency of cleaning are
reflected. The U.S.S.R. ESP mathematical model recently developed begins to
find application in design practice. To have a wide use of this method it is
necessary to develop procedures forecasting electrical parameters of ESP
performance, physical and chemical properties of dust-laden gas flow, and
their interconnection.
For effective operation of the ESP installation, questions of gas dis-
tribution are of vital importance. In the U.S.S.R., the choice of a gas
8
-------
distributing system is done on the basis of modeling of systems in the scale
1:10 and 1:20. However, now the atlas on utilized gas distributing diagrams
has been developed. It permits selection of the optimum variant of diagram
for typical ESP arrangement.
It is known that when specific resistivity of collected dust is high,
reverse corona appears in the ESP, and that reduces the efficiency of gas
cleaning. As it follows from numerous studies conducted in the U.S.S.R. and
abroad, specific resistivity of dust is considerably dependent on temperature,
humidity, and chemical composition. Therefore, to eliminate reverse corona in
the ESP, conditioning systems are employed, i.e.:
1. systems controlling gas humidity and gas temperature, and
2. systems controlling chemical conditioning of gases.
Systems for chemical conditioning of gases are not widely spread in the U.S.S.R.
due to the complexity of their performance. More widely used are systems for
controlling humidity and temperature of gases, particularly in cement applica-
tion. Systems for reducing temperature and humidity of gases are now planned
for application in heat- and power-engineering to remove fly ash of ekibastuzsk
and kuznetzk coal.
So, the question of forecasting reverse corona formation^turns out to be
the most important one when choosing optimum working temperature and humidity
of gases. One cannot always answer this question correctly by measuring spe-
cific resistivity of dust. That is why methods directly forecasting reverse
corona are being developed.
Obvious increase in efficiency of the ESP can be achieved by optimization
of the rapping schedule. Studies undertaken in the U.S.S.R. make it possible
to predict the relationship of collection efficiency to mass of dust precipi-
tated on electrodes (dust volume). It turned out that when dust volume in-
creased to a certain limit, collection efficiency increased too. But further
growing of the dust layer results in its self-collapsing and reducing of
efficiency. The relationship of optimum mass on electrodes to specific resist-
ivity of dust has been found. On the basis of these investigations, methodology
for determining optimum intervals of electrode rapping was developed. Presently,
this methodology has been applied in the cement industry, heat- and power-engineer-
ing, and other industry branches. Optimization of rapping intervals alongside
of raising cleaning efficiency makes it possible to increase reliability of
performance of ESPs.
Practice shows that optimization of working regimes of power supply units
is a reserve to increase ESP efficiency. Optimization of regimes means set-
ting up an optimum spark rate that is supported by a control diagram. Here, a
method developed in NIIOGAZ of adjusting units by a subsequent field is used.
Transducers for determination dust concentration in gas are supposed to be
also used for this purpose.
To take the best advantage of reserves of ESP power supply, NIIOGAZ is
conducting studies concerning optimization of voltage control, impulse power
supply of ESP, etc.
-------
Development and Realization of Out-Of-Date ESP Models Reconstruction
The above mentioned methods for increasing collection efficiency have
found wide application while modernizing out-of-date ESPs. In the cement
industry, power-engineering, and other branches of industry, out-of-date
precipitators are being systematically replaced. The following technical
innovations account for increasing cleaning efficiency of modernized ESP
installations:
1. substitution of out-of-date, low-efficiency equipment for more
modernized equipment,
2. increase of ESP plate area,
3. improvement of gas distribution,
4. optimization of electrode rapping intervals and supply unit working
regimes,
5. increase of reliability of power supply unit performance, and
6. technological preparation of gases to be cleaned.
The reconstruction of gas cleaning installations at the Cherepetskaya,
Pribaltijskaya, and Troitzkaya hydroelectric power stations, and Bryansk and
Novorossijsk cement plants have shown that gas cleaning efficiency can be
raised up to 98-98.5 percent. Further increase of efficiency can be achieved
due to the application of 12-meters-high precipitating electrodes instead of
those 8-meters-high that were installed on the out-of-date precipitators.
Application of high-efficiency ESPs at the industrial sites under construction
and modernization of out-of-date apparatus of electrical gas cleaning makes it
possible to prevent atmospheric pollution above the standards.
10
-------
TECHNQUES OF MODELING GAS FLOW IN
ELECTROSTATIC PRECIPITATORS IN THE U.S.
Samuel Bernstein*
Abstract
The construction of very large precipitators that satisfy high efficiency
requirements for the collection of fine particles poses a challenging design
problem. Flow modeling studies, when combined with appropriate analysis, pro-
vide a cost-effective design tool. In this paper the modern approach to flow
modeling is reviewed by a discussion of the objectives and requirements of flow
modeling, modern experimental methods, and advanced techniques. Proper flow
simulation requires the appropriate selection of model scales and instrumenta-
tion. Experimental techniques have been refined through the use of laser light
for smoke flow visualisation, through the measurement of velocities with Laser
Doppler Anemometry, through automated data acquisition systems, and through
computational techniques.
INTRODUCTION
The construction of very large precipitators (typically designed to treat
several million cubic feet of gas per minute) and the requires its for fine par-
ticle collection have stimulated interest in careful gas flow modeling. Although
gas flow modeling studies were introduced into the precipitator field approxi-
mately 30 years ago (ref. 1), relatively little research on this subject has been
published in the open literature, and there are conflicting reports on the useful-
ness of flow modeling studies (refs. 2,3,4). The purpose of this paper is to
review modern approaches and techniques in flow modeling for precipitator studies.
The objectives and requirements of flow modeling studies and modern experimental
methods are reviewed and advanced techniques are discussed.
OBJECTIVES AND REQUIREMENTS OF FLOW MODELING FOR
INDUSTRIAL ELECTROSTATIC PRECIPITATORS
Flow modeling studies provide flow simulation in a scaled-down model of
the precipitator installation. Simulation may be used for precipitator develop-
ment work or for modeling an industrial installation. In precipitator develop-
ment work the simulation requires modeling the fundamental physical parameters
of the precipitator: the electric forces, the fluid dynamic forces, and the
particle dynamics. In the modeling of an industrial installation, the emphasis
is on the effect of the particular ducting arrangement on the gas flow distribu-
tion, and, consequently, the parameters of interest are the fluid dynamic forces
and the particle dynamics. This type of simulation is unique in every installa-
tion because each installation requires a different system for ducting around
the precipitator box.
The objectives of flow modeling studies for an industrial installation are
to simulate the flow distribution and the total pressure drops in a given in-
stallation. Flow nonuniformities reduce the efficiency of the precipitator since
regions of high or low velocities do not have the optimum flow conditions. In
*Flow Research Company, Kent, Washington.
11
-------
addition, the regions of relatively high velocity provide locally high shear and
therefore enhance reentrainment of collected particles and increase the pressure
drop through the precipitator. If the flow is also separated, the resulting
structural vibration may reduce the life cycle of the precipitator.
Recognition of the importance of flow uniformity and flow models is often
stated in the contract for the precipitator work, as has been suggested by
the Industrial Gas Cleaning Institute (IGCI) (refs. 5,6). However, the IGCI
suggests specific values for gas uniformity, and, based on the above discus-
sion, the requirement in each installation should be reviewed on a case-by-
case basis.
The requirement for relatively uniform flow is relevant for precipitators
with either laminar flow or turbulent flow; however, in practical installations,
the flow is always turbulent. Unfortunately, the effect of the turbulence on
the precipitator is not completely understood, and only some general observa-
tions are possible.
The turbulent shear force near the collector plate, Tt, may be expressed
by (see, for example, Launder, et al. [ref. 7]):
tt = Meff
where 3u/9y is the mean velocity gradient near the collector plate and ugff is
the effective turbulent viscosity. The effective viscosity of the turbulent
flow is much larger than the molecular viscosity in laminar flow, and it is
proportional to the turbulent kinetic energy of the fluid and to the turbulence
length scale. Consequently, low reentrainment losses will be associated with
flows of relatively low turbulence intensity and small-scale turbulence.
Usually, in flow modeling studies, internal-flow correction devices are
introduced to improve the flow distribution and total pressure losses. How-
ever, flow modeling studies, if performed early enough in the design, could
also suggest a more cost-effective ducting arrangement.
MODERN EXPERIMENTAL METHODS OF FLOW MODELING
Effective flow modeling is usually combined with a theoretical analysis of
the flow. A preliminary analysis identifies the appropriate scale for the model,
the appropriate instrumentation, and the approximate errors that would result
from imperfect simulation and measurements.
Model Scales
As suggested in the previous section, the common flow modeling for indus-
trial installations requires the simulation of the fluid dynamics and particle
dynamics. The relevant similarity requirements for the fluid dynamic simulation
are geometric similarity and similarity of Reynolds number. (The similarity
parameters may be derived by a dimensional analysis: see, for example, Johnstone,
1957 [ref. 8].)
12
-------
Typical plant conditions for a cold-side precipitator include a temperature
of 150° C (300° F) and flow rates of several million cubic feet per minute. The
resulting range of Reynolds numbers is typically between 104 and 105, based on
the inlet duct diameter. In model studies it is not possible to substantially
increase the gas flow velocity above that of the full-scale case because of
compressibility effects, and, consequently, it is not practical to match the
Reynolds number of the model with that of the full-scale plant.
The consequence of having a lower Reynolds number in the model study may
be estimated as long as the flow in the scale model is turbulent. The boundary
layer along the model walls and along the collector plates will be thicker than
in the full-scale installation, and it might be necessary to reduce the number
of collector plates to avoid relaminarization of the flow between the plates.
The measured pressure drop coefficients, which are also functions of the Reynolds
number, will be higher in the laboratory and should be corrected accordingly.
The accepted practice is to use 1/16 scale models; however, the particular choice
of scale depends on the complexity of the installation, and to some extent, on
the required accuracy.
Simulation of particulate matter in the gas flow model requires particles
with sizes proportional to the model scale. The behavior of these very fine
particles may be simulated in the model by the use of particles with larger
diameters but lower densities. The use of cork of a given particle size to
simulate fly ash of smaller size provides the desired simulatior of a lower
settling time in the ducts and a shorter relaxation time to fluiu forces.
Flow Visualization
Flow visualization provides only a qualitative picture of the flow; how-
ever, the early identification of separated regions and flow angularities may
significantly reduce the necessary testing time. One of the most common tech-
niques for flow visualization is the use of tufts (ref. 9). In this technique,
lightweight yarn tufts are mounted on a rake inside the model to identify local
flow patterns.
Another common method of flow visualization is the introduction of smoke
flow into the channel. Several chemical agents are used to generate the smoke,
including titanium tetrachloride, tin tetrachloride, heated kerosene, and smoke
bombs, most of which either have objectionable odors or are poisonous (ref. 9).
We have obtained high-quality smoke flows without the above limitations in flow
tests with the use of a smoke generator designed at the University of Washington
(ref. 10).
Velocity Distribution
The velocities of interest for electrostatic precipitator studies range
from those in the high-speed region in the ducts, with a typical velocity of
17 m/s (approximately 50 ft/s) to the velocity in the low-speed region in the
box, which is between 1.5 and 5 m/s (approximately 5 to 15 ft/s). Pitot tubes
and yaw heads (ref. 9) are often used in the high-speed range. There are
several designs of yaw heads with several arrangements of pressure taps on a
small sphere. The flow direction may be computed by comparing the pressure
difference from several locations on the sphere. In practice, the yaw heads
13
-------
and pitot tube are limited to velocities above 3 m/s (10 ft/s), and their
response time does not allow the measurement of fluctuations of the turbulent
flow at a given location.
Hot-wire anemometers are used for measurements of flows at low velocities.
The major advantages of the hot wire are its sensitivity at low speeds and its
fast response (typically, the hot wire provides frequency responses of 100 kHz
or better [ref. 11]). Hot wires require careful handling and frequent calibra-
tion. The more common single wire is sensitive to the flow magnitude only;
however, a crosswire arrangement, which allows limited determination of the
flow direction in a plane, is also available.
Pressure Measurements
Total pressure losses are measured by traversing a section of the model
with a total pressure probe and micromanometer. As discussed above, the
measured pressure loss coefficient should be corrected to account for the
variation in Reynolds number. (Idelchick [ref. 12] gives some empirical data
that contain the effect of the Reynolds number.)
ADVANCED EXPERIMENTAL TECHNIQUES
Recent developments in laser technology and minicomputer systems offer
significant improvements in flow visualization techniques, remote velocity
measurements, and automatic data acquisition methods. Modern computational
techniques for turbulent flows have also recently attained a high degree of
refinement to support flow modeling work as a flexible and cost effective
design tool.
We recently obtained significantly improved smoke flow visualization pic-
tures when the smoke flow was illuminated by a laser light. In this technique,
a laser beam is directed to a cylindrical glass rod, which scatters the beam.
This scattered beam forms a thin sheet of light, which selectively illuminates
a thin sheet of smoke. By moving the beam, we can move the illuminated plane
to view selected planes of the model with fine detail. This arrangement is
illustrated in figure 1.
A second implementation of the laser beam is the remote measurement of
the gas velocity with Laser Doppler Anemometry (LDA). The basic principle of
operation for LDA is to direct two laser beams to a point in the model where
the velocity will be measured. The two beams form a dark and light fringe
pattern, as shown in figure 2. As the particles in the gas pass through the
fringe pattern, they scatter light pulses at the Doppler frequency, a frequency
proportional to their travel speed. The scattered light is then collected by
a photodetector, and the data is analyzed to compute the flow velocity at the
desired point.
There are several arrangements of LDA systems for measuring either single
or multiple components of the flow velocity. These arrangements are available
as off-the-shelf items from several manufacturers (i.e., DISA Electronics,
Spectron, Thermosystems). LDA is a remote sensor of the gas flow (it measures
the gas flow by measuring fine-particle trajectories), and it offers fast re-
sponse; however, the system is quite costly compared to hot-wire anemometers.
14
-------
Rreciprjator Model
CJl
Figure 1. Smoke flow visualization using laser beam illumination.
-------
Fringe Pattern From Transmitted Beams
Laser
Model Wall
Figure 2. Schematic illustration of a laser doppler anemometry setup.
-------
Data Line to Cornpu
Figure 3. Automated traverse system for measuring velocity in an
electrostatic precipitator model.
17
-------
0.6096 meters
Flow
2.89 m/«
--—_--*-.
00
Computation Time: ~ 50 sec on a CDC 6600
E
0.1016
meter
1
__ -V _
J.. ..
Figure 4. Computed velocity field of turbulent flow around a precipitator structure.
-------
The effectiveness of laboratory modeling has been further improved by
automating velocity data acquisition. In one such system, two hot-wire ane-
mometers are mounted on the rod of an automated traverse system, as shown in
figure 3 (ref. 13). The two wires reduce the distance the traverse system
must travel to measure the flow across the model. The traverse rod position
is set by a stepping motor, which is controlled by a Nova 800 minicomputer.
The traversing rod is stopped at each stepping interval for several tenths of
a second while a few thousand data points are recorded. The minicomputer
stores the calibration curves for the hot-wire anemometers, and the data from
the hot-wire anemometers are collected, reduced, plotted, and printed out by
the computer in several minutes. Data analyses and velocity histograms may
also be generated by the minicomputer at a substantial savings of manpower.
Recent developments in computational techniques for turbulent flows have
increased their utility in supporting flow modeling work. In this approach,
the turbulent flow around a complex precipitator component is computed using a
finite difference scheme for the complete flow equations including an experi-
mentally verified turbulence model. Such a computer code may be easily modified
to predict the effect of changing the local geometry of the duct, and the
addition of flow correcting devices near a bend. Additional computation may
be performed to compute particle trajectories using the detailed flow solution
to calculate the forcing functions on the particles. An example of flow field
calculation around a complex geometry is illustrated in figure 4. The arrows
in the figure represent the magnitude and direction of the averaged velocity
around the structure, and the grid represents the number of computed points in
the finite difference scheme. The versatility and economy of the computational
method may save considerable effort in evaluating a precipitator during flow
modeling studies.
CONCLUSIONS AND RECOMMENDATIONS
The assessment of the status of experimental techniques for modeling flow
in electrostatic precipitators may be summarized as follows: Modeling is a
cost-effective design tool when the model is properly designed and analyzed.
The data generated in flow modeling require a careful assessment of their
experimental limitations.
The accuracy of the experimental predictions will continue to improve
with the introduction of remote sensors (i.e., IDA) and modern automated data
acquisition systems. These systems reduce the sensors' interference in the
flow and also increase the reliability of the data.
Numerical techniques will play a greater role in supporting flow modeling
work as a cost effective tool.
More research is needed to identify the role of fluid dynamic turbulence
in electrostatic precipitators although it may be argued, on basic physical
principles, that the effect of turbulence is significant and the turbulence
quantities should therefore be measured and documented during model studies.
19
-------
REFERENCES
1. H. J. White, Industrial Electrostatic Precipitation, Addison-Wesley
Publishing Co., Reading Massachusetts, 1963.
2. C. L. Burton and D. A. Smith, "Precipitator Gas Flow Distribution," pre-
sented at EPA and Southern Research Institute Symposium on Electrostatic
Precipitators, Pensacola, Florida, September 20 - October 2, 1974.
3. J. B. Opfell and W. T. Sproull, "Limitations of Model Studies in Predicting
Gas Velocity Distribution in Cottrell Precipitators," Ind. Eng. Chemical
Proc. Design Develop.. Vol. 4, No. 2 (1965), pp. 173-177.
4. D. J. Gibson, "Scale Model Flow Testing Upgrades Cost-Effectiveness,"
Research/Development, Vol. 27, No. 11 (1976).
5. Industrial Gas Cleaning Institute, "Criteria for Performance Guarantees,"
Publication No. EP-3, Stamford, Connecticut, 1973.
6. Industrial Gas Cleaning Institute, "Gas Flow Model Studies," Publication
No. EP-7, Stamford, Connecticut, 1973.
7. B. E. Launder and D. B. Spalding, Mathematical Models of Turbulence,
Academic Press, New York, 1972.
8. R. E. Johnstone and M. W. Thring, Pilot Plants, Models, and Scale-up
Methods in Chemical Engineering, McGraw-Hill, New York, 1957-
9. A. Pope and J. J. Harper, Low Speed Wind Tunnel Testing, John Wiley and
Sons, Inc., New York, 1966.
10. 0. Shindo and 0. Brask, "A Smoke Generator for Low Speed Wind Tunnels,"
Univ. of Wash., Dept. of Aero, and Astro., Technical Note No. 69-1, 1969.
11. G. Comte-Bellot, "Hot-Wire Anemometry," in Annual Review of Fluid Mechanics,
M. Van Dyke, W. G. Vincetti, and J. V. Wehausen, Eds., Vol. 8, Annual Re-
views Inc., Palo Alto, California, 1976.
12. I. E. Idelchick, "Handbook of Hydraulic Resistance, Coefficients of Local
Resistance and of Friction," AEC-TR-6630, 1966, (Russian Translation
available through NTIS).
13. S. Bernstein, L. Piper, and E. Murman, "Flow Modeling Studies for the EPRI
Advanced Particulte Control Facility," Flow Research Note No. 105, 1976.
20
-------
INVESTIGATION OF GAS DISTRIBUTION SYSTEMS IN ELECTRO-
STATIC PRECIPITATORS COLLECTING DUST AND FLY ASH*
I. E. Idelchik, JA. L. Ginzburg, M. 0. Shteinberg,
V. P. Alecsandrov, V. S. Korjagin (U.S.S.R.)
Abstract
The objective of studying gas distribution systems is the develop-
ment of simple and reliable units that ensure the uniform distribution
of gas in the active zone of an electrostatic precipitator and in the
parallel operating units. This also includes the development of methods
for calculating and designing such systems. In accordance with this ob-
jective, the authors describe original designs of gas distribution systems
along with their characteristics (degree of gas distribution, nonuniformity,
and a coefficient of hydraulic drag). Results are given of a study on the
distribution of dust along the collector branches and a collector design
is proposed that makes it possible to achieve a more uniform distribution
of dust. It is noted that the basic results of the studies are generalized
in "The Atlas of Gas Air Distribution Units of Various Industrial Equipment"
under the authorship of I. E. Idelchik.
Theoretical and experimental works have shown that the efficiency of dust
and fly ash collection by electrostatic precipitators depends on the uni-
formity of gas flow distribution over the apparatus cross section, and if
there is a series of electrostatic precipitators connected in parallel, it
depends on the uniformity of gas flow distribution over all precipitators.
If the velocities of a gas flow are nonuniform across the precipitator
section, the degree of nonuniformity can be described by the coefficient of
momentum in the German's equation:
In this case, the German's equation has the following form:
— '— 1
C- &
where Vm = velocity of gas flow in the section, and
K = a coefficient taking into account the electrical and other
characteristics of a precipitator.
The dependences in equation (1) were experimentally tested on the pilot-
plant electrostatic precipitator, the electrodes of which had the height 1.8
m, and the total length of two electric fields was 5 m. Perforated lattices
* For further information regarding the material in this paper, please
contact I. K. Reshidov, U.S.S.R. 113105, Moscow, M-105, 1-st Nagatinsky
Pass., 6, NIIOGAZ.
21
-------
with various perforation disposition along their height and rotatable plates
in the gas conduit were used at the entry to the field 1 to provide various
values. The obtained results are given in figure 1. One can see that they
confirm the justice of the formula in equation (1) with precision sufficient
for practical purposes.
The main volume of work was done with geometrically similar models pro-
duced to a scale 1:10-1:20 relative to a full size. The region of the gas
conduit feeding the gas to an electrostatic precipitator, the gas distribution
system, and the first 1-2 fields of the apparatus were modeled.
We investigated the gas distribution systems in separate electrostatic
precipitators with the active height being from 7.2 to 12 m and the active
area from 30 to 265 m2, which were used in dust collectors for large power
blocks (from 300 to 800 Mw), in nonferrous*metallurgy, and sulfuric acid pro-
duction.
One of the typical variants of the gas conduits and the gas distribution
systems for large electrostatic precipitators used in thermal power production
is a vertical pit with baffles and a system of plate perforated lattices
(figure 2). The aerodynamics of such a system, which was preliminarily stud-
ied by means of model experiments, has been investigated using a pilot-plant
electrostatic precipitator of 12 m height. By means of model experiments we
obtained Mk = 1.02; by means of plant experiments, Mk = 1.09.
Thus, plant measurements gave results similar to model measurements.
This confirmed the competence of gas distribution modeling in electrostatic
precipitators.
Besides, the measurements of concentration and particle-size distribution
across the active section were made in this electrostatic precipitator. The
measurements were made in the mixing chamber between first and second lattices
in points 1-6 along three vertical lines (see figure 2). The velocity field
is shown in figure 2a, and the dust loading field in figure 2b. These fields
are shown in respect to average fields over the active section of the velocity
Vm and the dust load (jm. One can see that the considered construction of a
gas distribution system provides not only a practically uniform velocity
field, but also a dust loading field.
In addition to dust loading fields, an appropriate particle-size distri-
bution was determined in each point from 1 to 6. The comparison of particle-
size distributions in points 1 to 6 allowed a conclusion to be made about a
practically constant fractional composition of particulate matter up to 100 (jk
over the height of the apparatus at its inlet.
For large degrees of a flow expansion n. (ratios of the areas of cross
section of active precipitator region and gas feeding conduit), in the range
20 < n < 50, the simple and reliable gas distribution system with one per-
forated lattice and a perforated screen has been developed (figure 3). For a
small resistance coefficient related to the velocity at inlet, ^= 1, such
construction provides a satisfactory degree of gas distribution nonuniformity
in practice (M = 1.08-1.15).
22
-------
ro
CO
Figure 1. Dependence of precipitator efficiency n on the degree of flow
nonunifirmity M •
experiment,
calculation
{5 M
K
-------
0 /
ro
Figure 2. Distribtuion of gas flow velocities and fly ash size in the
electrostatic precipitator with 12-m electrodes.
a - distribution of velocities V\/m
b - distribution of fly ash concentrations /v/ti»l
-------
(2+3)2).
rvs
en
Figure 3. Gas distribution system for electrostatic precipitators
with central flow inlet.
-------
rxj
CTl
5
r
Figure 4. Gas flow rate distribution r over electrostatic precipitators
connected in parallel.
a - 10 electrostatic precipitators
b - 14 electrostatic precipitators
-------
ro
,! ' It !.i
Ji
>• 1
1
_J
•
_
\
. •
J ^
•
^
X
1
a
II U II M
!L
1.0
0,6
V
^nf
—&r
*
-------
1 —
f
!
f
n
i
i
n
\
f
i
111 ft ft t
w
w
as
0,6
-Kn
£?.,.„ u.nfM
;
V
Vm
A
x
r
2X
€
>•
1
I
1
1
^
/
/
Jgi ^Jl
M/jfrek
4%
1,0
JC
1,0
U
1,T
iO
1,6
jn^
t,UJ
no.
(Jfi£
0,6
^
/^m
\**&£2
\/\
f \
I \
\ >
V
Vm
/gv
r
2-
•
r^^,.
•*
... 5\. .,
1-
1
1
1
1
I
^
T
/
L;
N,
M#«k
3 I) 7 9
Figure 6. Distribtuion of velocities y and flow dust load
>/« across manifold offsets with variable cross section.
a - seal ari form
b - wedge-form. 1 -
; 2 -
28
-------
In addition, to study the gas distribution over the electrostatic pre-
cipitator cross section, gas and dust distributions over the apparatus instal-
led in parallel in one installation have been investigated. Some of such
installations involved complex manifold systems containing up to 14 precipi-
tators installed in parallel (with an active area of one apparatus from 30 to
60 m2). The diagram of installation and gas flow distribution over the appa-
ratus that is the result of investigations is given in figure 4 as an example.
Experimental investigation of dust distribution by means of manifold models
shows the significant dependence of distribution on a manifold construction.
Tests were carried out on the coal ash loading air flow: The density of
the coal ash was 2.4 g/cm3, the velocity of the air flow was in the range 15
to 25 m/s. The medium particle size distribution of coal ash was 25 pk. Dust
concentration increased in manifold offsets of constant cross section (figure
5b) from the 1st to 8th offset. In the last offset, the concentration is
about twice as high as in the first one.
The dust distribution in manifolds of arbitrary cross section is given in
figures 5, 6a, and 6b. In the last offset of a scalariform manifold (figure
6a), the dust concentration is 3.5 times as high as in the first one. One can
see the tendency towards a decrease in the nonuniformity of dust distribution
over the offsets in a wedge manifold with an inclined wall at the side of
offsets (figure 6b). A more uniform dust distribution was obtained in a
manifold of uniform-variable cross section with smooth offsets (figure 5a).
An increase of dust concentration in the last lateral offsets in a down-
stream direction can be explained by the inertial separation of large par-
ticles from the manifold flow. As the comparison of considered manifold
variations has shown, the indicated effect of inertial separation can be
reduced and the uniform dust distribution can be obtained creating more fa-
vorable conditions for entry of the dust loading gas flow to manifold offsets.
The results of aerodynamics modeling of various inlet (outlet) flows and
distributors conforming to the definite types of electrostatic precipitators
and apparatus of wide technological type have been symmarized by Professor I.
E. Idelchik in the "Atlas of Gas-Air Distributors at Various Plant Apparatus."
Diagrams of inlet and outlet manifolds designated for a flow inlet to appara-
tus installed in parallel and a flow outlet from the apparatus and recommen-
dations to the most beneficial arrangement of apparatus from the angle of
maximum uniformity of the flow distribution are also given in the "Atlas." At
the same time, the necessary data for calculation of basic manifold dimensions
are given to provide an allowable degree of flow nonuniformity over individual
apparatus of this series.
There are not only optimal variants of inlet, outlet, and flow distribu-
tion over the apparatus cross section and the apparatus installed in parallel,
but also a successive improvement of aerodynamic characteristics, as better
variants of inlet and outlet regions and steadying devices are used. A great
variety of such variants and their aerodynamic characteristics allows the most
suitable schemes of apparatus and construction to be selected in each specific
case.
29
-------
STUDIES OF PARTICLE REENTRAINMENT
RESULTING FROM ELECTRODE RAPPING
Leslie E. Sparks,* Owen J. Tassicker,1" and John P. GoochT
Abstract
Measurements of rapping reentrainment on a pilot-scale prec^,p^tator
indicated that rapping emissions decreased with increasing time between raps.
Mie percentage of the collected electrode also increased with ^ncreased ti-me
between raps. Particle size distribution data for the rapping puffs suggested
that thicker dust layers produce larger reentrained particles upon rapping. A
pronounced vertical stratification of the rapping emissions was found for the
pilot-scale studies.
Measurements of fractional efficiency with and without rapping at full-
scale precipitator installations show that rapping efficiency losses occur
primarily for particle diameters greater than 2.0 ym diameter. The largest
rapping losses were measured on hot-side installations. Mass emission data
suggest a correlation between the dust removal rate in the last field of the
precipitator and the emissions due to rapping.
INTRODUCTION
Collection of participate matter by the electrostatic precipitation
process consists of three separate operations: (1) particle charging, (2)
particle collection, and (3) removal and disposal of the collected material.
In ideal circumstances, all material collected on the grounded electrodes
would be transported to a collection hopper without reentrainment into the gas
stream. While this ideal situation is approached in the collection of liquid
particles, the process of removing dry particulate from collecting electrodes
is usually accompanied by a significant reintroduction of the collected material
into the flue gas. This paper discusses particle reentrainment caused by
rapping of the collection electrodes in conventional wire-plate electrostatic
precipitators.
The purpose of an electrode rapping system is to provide an acceleration
to the electrode that is sufficient to generate inertial forces in the collected
dust layer that will overcome those forces holding the dust to the electrode.
A successfully designed rapping system must provide a proper balance between
electrode cleaning and minimizing emissions resulting from rapping, reentrainment,
As part of an overall program to gain a better understanding of the electro-
static precipitation process, measurements with the objective of quantifying
and size-characterizing losses due to-electrode rapping have been conducted on
one pilot-scale and several full-scale electrostatic precipitators. The
pilot-scale study was funded by the U.S. Environmental Protection Agency, and
the full-scale tests were sponsored by the Electric Power Research Institute.
^Environmental Protection Agency, Research Triangle Park, N.C. 27511
^Electric Power Research Institute, Palo Alto, California 94304
tSouthern Research Institute, Birmingham, Alabama 35205
30
-------
EXPERIMENTAL STUDIES
Most of the previous work on rapping reentrainment has not included
direct measurements of the effects of rapping on overall efficiency and ef-
ficiency as a function of particle size. Accordingly, measurement programs
have been conducted with the objective of quantifying rapping reentrainment in
terms of the percentage and particle size distribution of the reentrainment
material.
Methods of Measurement
The quantification of rapping reentrainment requires methods of measuring
the mass and particle size distribution of particulate exiting the precipitator
with and without rapping. During both the pilot- and full-scale precipitator
test programs, optical real-time systems and integrating mass systems were
used. For the full-scale tests, particle size measurements were obtained
using a method based on electrical mobility analysis for particle diameter
less than about 0.20 urn.
Mass Concentration Measurements. Mass measurements were obtained with in-stack
filters. The sampling probes used at the inlet and outlet were heated and
contained pitot tubes to monitor the velocity at each sampling location for
the full-scale tests. Glass fiber thimbles were used at the inlet to collect
the particulate and Gel man 47-mm filters were used at the outlet. Different
procedures were employed at the pilot-scale unit compared to the full-scale
units.
At the pilot plant facility, two outlet sampling trains were used (ref. 1):
(1) the upper sampling train for the upper 68 percent of the precipitator
outlet and (2) the lower sampling train for the lower 32 percent of the precipi-
tator. The outlet sampling locations were about 1 meter from the plane of the
outlet baffles, and only one lane of the precipitator was sampled. Both
outlet mass trains were modified to consist of two systems: one of which was
used to measure emissions between raps and the other to measure emission
during raps. One of the two filters on each of the two outlet probes was
designated the between rap sampler and the other the rapping puff sampler.
After stable conditions were obtained, the between rap sampling systems were
started. Before rapping the plates, sampling was discontinued and the probes
were rotated so that both nozzles on each probe pointed downstream. The dust
feed was turned off, and after a clear flue was obtained, the second filter
was rotated into the gas stream. Sampling was resumed and the plates were
rapped. When dust had settled, sampling with this second set of filters was
discontinued and the nozzles to the filters were again pointed downstream.
The dust feed was then turned on and the sampling was resumed again with the
between rap system.
For the full-scale precipitator installation, one would expect to be able
to measure rapping reentrainment simply by obtaining data with either a mass
train or an impactor sampling system, with a rapping system energized and
subsequently deenergized, and then comparing these measurements. However, it
was found that during the test program at the first installation (Plant 1) the
sensitivity of the electrostatic precipitator to changes in resistivity and
other process variables could overshadow the differences in total emissions
caused by energizing and deenergizing the rappers. The variation in precipi-
31
-------
tator performance caused by the resistivity and other process variable changes
made it impossible to determine rapping reentrainment losses from a direct
comparison of data obtained one day with rappers in the normal mode ana rap
pers deenergized on subsequent days.
In order to minimize this difficulty, a revised sampling strategy was
adopted for the remaining installations. This strategy consisted^of sampling
with mass trains and impactors dedicated to designated "rap and no rap
periods. Data with a rapping system energized and deenergized were obtained
by traversing selected ports with dedicated sampling systems in subsequent
periods on the same day. This procedure, while necessarily distorting the
frequency of the rapping program being examined, minimized the effects of
resistivity and other process variable changes.
The use of this sampling strategy leads to two possible procedures for
calculating the fraction of losses attributable to rapping reentrainment. The
first procedure calculates the ratio of emissions obtained with rappers off to
rappers on and subtracting from unity. The emissions data utilized in this
procedure were obtained during the time in which alternating sampling periods
for rap and no rap sampling trains were employed. The second procedure con-
sisted of subtracting the mass emissions obtained with the rappers deenergized
from those of the previous day with normal rapping, and dividing by the emis-
sions obtained with the rappers operating normally. It could be argued that
if the alternating on-and-off procedure for sampling did not distort the
results obtained, and if there were no other variations in parameters affect-
ing the precipitator performance, that data obtained from the "rap" period
should be approximately equal to that obtained during periods in which the
rappers were operating in a normal fashion. In this paper, we have calculated
percentage of rapping emissions using both of these procedures.
Particle Sizing. Three sizes of selected sampling systems were used in the
measurement programs, two of which were real time extractive systems (a large
particle system and a fine particle system) while the third (cascade im-
pactors) provided time integrated in situ data. The large particle (diameter
range 0.6 to 20.2 urn) extractive system was employed only for outlet measure-
ments to provide qualitative information on the relative fractions of the
emissions that could be attributed to rapping losses in the precipitator. In
addition, this system also provided data on particulate concentration changes
with time.
The fine particle system (0.01 to 0.3 urn) was employed at both the inlet
and outlet of the full-scale precipitators for purposes of providing fractional
efficiency data and to give quantitative information on the contribution of
rapping, if any, to emissions in this particle size range.
Description of Installations and Results
Pilot-Scale Precipitator. The pilot-scale rapping tests were conducted on a
nearly full-scale pilot precipitator owned and operated by Fluidyne Engineer-
ing. Figures 1 and 2 illustrate the features of the test facility. This
pilot unit effectively represents one electrical section in a full-scale
precipitator. The plate height is 6 m and the plate length is 2.7 m. The
total collecting area is 167 m2, and wire to plate spacing is 11 cm. In the
original design, the plates were constructed from expanded metal. For this
32
-------
PILOT PRECIPITATOR
NOMINAL 1.2 M WIDTH
FOR (5) PASSAGES
BURNER SECTION
H2SO4
Figure 1. Pilot precipitator fluidyne.
33
-------
VARIOUS VIEWS OF FLUIDYNE PILOT PRECIPITATOR
CO
TYPICAL FLOWl
9 FT/SEC
300°F
35000 ACFM
SCAB 1.4
WEST
CHANNEL NO.
PLATE NO.
9ACCELEROMETERS
ARE MOUNTED ON PLATE 4
PLATE ROWS 1. 2, & 6 EXPANDED METAL
PLATE ROWS 3. 4, & 5 SOLID
PLATE ROWS 2-1, 2-3 WERE 0.5 M
Figure 2. Various views of fluidyne pilot precipitator.
-------
rapping reentrainment study, three of these plates were replaced to provide
two lanes with solid plates on each side of the lane. Outlet sampling was
confined to the lanes with solid plates.
Table 1 presents a summary of results obtained from the experiments on
the Fluidyne Pilot Unit. These results indicate that rapping emissions de-
creased with increasing time between raps. Figure 3 shows the effect of the
rapping interval on efficiency. The percentage of the collected dust removed
from the collecting electrode also increased with increased time between raps,
as figure 4 illustrates. These results are consistent with a theory of dust
removal which indicates that the product of the normal plate acceleration and
the dust surface density must be greater than the tensile strength of the
layer (ref. 2).
Figure 5 presents particle size distribution of rapping puffs for the
indicated rapping interval. These data suggest that thicker dust layers
produce larger reentrained particles upon rapping. An inspection of the
impactor substrates at the outlet sampling locations 2 and 3 revealed that the
majority of the large particles in the rapping puffs were agglomerates.
Producing relatively large agglomerates instead of individual particles is
desirable because the larger agglomerates are recollected faster than discrete
particles or smaller agglomerates.
The measurement of the vertical distribution of the rapping loss at the
Fluidyne pilot unit indicated that 82 percent of the rapping emission occurred
in the lower 32 percent of the precipitator (ref. 1). This effect was appar-
ently due to both hopper boil-up and gravitational settling of the reentrained
material. Figure 6 illustrates the vertical stratification as a function of
particle size. All of the particle size bands show a decrease in concentration
with increasing distance from the bottom baffle.
Full-Scale Precipitators. Six full-scale electrostatic precipitator installa-
tions were tested under the EPRI-sponsored research program (ref. 3). In
terms of location in the power plant system and type of fuel burned in the
boiler, the units may be classified as follows:
Plants 1 and 2 - Cold-side ESPs collecting ash from low-sulfur
Western coals
Plant 3 - Hot-side ESPs collecting ash from low-sulfur
Eastern coal
Plants 4 and 5 - Cold-side ESPs collecting ash from high-sulfur
Eastern coals.
Plant 6 - Hot-side ESPs collecting ash from low-sulfur
Western coal
Table 2 summarizes the important design parameters and the results ob-
tained for the six installations. The installations were characterized by
relatively high overall mass efficiency. Rapping losses as a percentage of
total mass emissions ranged from over 80 percent for one of the hot-side units
to 30 percent for the cold-side units. The high rapping losses at Plant 3 are
probably due both to reduced dust adhesivity at high temperatures and the
relatively short rapping intervals.
35
-------
Table 1. Pilot-scale Rapping Experiments
Plate .
Acceleration Rap Gas Avg. Plate Total Penetration Due to
Type of G's Intervals Velocity Current Density Penetration Rapping Reentramment
Test x,y,z axis Min m/sec nA/cm^ % '°
Rap 1t1615 12 0.87 23.3 11.4 53
Rap 32
Rap 52
Rap 150
No Rap --
too
ss
EFFICIENCY,
00 CO CO
01 o en
7.6 32
6.
1 18
6.9 25
5.2
PRECIPITATOR EFFICIENCY
i
~~7
y
«
1
WITH
OUT RAPPI
\
WITH RAPPING
•
IMG
"•*«,»
'
40 60 100 140 180
TIME INTERVAL BETWEEN RAPS, MINUTES
Figure 3. Average efficiency for fluidyne pilot ESP for
various rapping intervals.
36
-------
DUST REMOVAL EFFICIENCY
MASS/AREA GAINED BETWEEN RAPS, kg/m2
0.26 0.78 1.3 1.8 2.3 2.9 3.4
100
3> BO
I 60
UJ
U
ul 40
u.
Ul
20
1 I ' I ' I
COLLECTED BETWEEN RAPS
sr i . \ . i . i . i . i . i
u 20 40 60 80 100 120 140 160
TIME INTERVAL BETWEEN RAPS, MINUTES
Figure 4. Dust removal efficiency vs. the time
interval between raps.
CUMMULATIVE PERCENT
DISTRIBUTION RAPPING PUFFS
10
o
oc
o
UJ
N
Ul
_J
o
1.0
ill!
0.01 0.1 1 10 20 40 60 80
PERCENT LESS THAN
INDICATED SIZE, BY MASS
Figure 5. Cumulative percent distribution for rapping puffs,
37
-------
SPATIAL DISTRIBUTION
OF RAPPING PUFFS
10°
-J 1
^ 2 t
ir ~ CO
O O
H uj °- 1Q5
<^^ p™*
£? UJ
SF£W
u-z<
20 .
-°5 104
2J"CO
D^"-
o§^
o ** ^
QC Q-
< UJ d 103
^ N 2
O CO Q.
1- CL
DC
102
I ' i
• A
Q ®
_ is e
*^ HI 1
0 i
^
-v v v ^
""A A
A ^
1 . 1
20 60
i j Uj i j 1
DIAMETERS, wm
e 1.5-3.0
H 3.0-6.0
S> A •*• 6.0-12.0 —
. w V 12.0-24.0
• A > 24.0
(^
' ^
SB
—
^^
&
, 4MI , I , X
100' '330 370 410
DISTANCE FROM BOTTOM BAFFLE, cm
Figure 6. Spatial distribution of rapping puffs.
38
-------
Table 2. Summary of Results From EPRI Test
GO
ID
Plant
Boiler Load, MH
ESP Vendor
Design Gas Flow, ACFM
Design Temperature, *f
Total Plate Area, ft1
Design SCA, ftVlOOO ACFM
Design Efficiency, %
Number of Electrical Fields in
Direction of Gas Flow
Plate-to-plate Spacing, in.
Emitting Electrode Design
Rapper Design
Portion of ESP Tested
Boiler Load During Test, MW
Gas Flow During Test, ACFM
Temperature During Test, *f
SCA During Test, ftVlOOO ACFM
Measured Efficiency, %
Dust Resistivity at operating
Temperature, ft-cm
Rapping Frequency
Field
1
2
3
4
5
6
% of Mass Emissions
Attributed to Rapping*
1
150
Lodge Cottrell
800,00ft
270
403,200
534.19
97.0
6
12
Mast with Square
Twisted Wires
Drop Hammer
Total
128
699,622
306
576.31
99.92
1.4xl011
Raps/hr
6
6
3
3
1
1
4
160
Lodge Cottrell
906,000
340
158,750
175.22
99.0
3
11
Mast with Square
Twisted Wires
Drop Hammer
1/2
160
328,762
311
241.44
99.55
1.7xl01'
Raps/hr
10
6
1
-
-
-
5
122
ELEX/AAF
575,000
325
127,032
220.93
98.7
4
10
Rigid Barbed
Wires
Tumbling Hammers
1/2
122
248,291
315
255.81
99.80
2x10 l«
Raps/hr
10
10
5
5
-
-
3
275
Research Cottrell
1,250,000
650
336,960
270.00
99.2
4
9
Hanging Round
Wires
Magnetic Drop
Hanuter
1/2
271
432,071
610
389.94
99.64
3.2x10"
Raps/hr
30-60
30-60
30
30
-
-
2
500
SF- Carborundum
2,330,890
250
1,. 136, 256
487.5
99.33
5
9 3/4
Electrode Frame
With Spiral Wires
Tumbling Rammers
1/6
508
316,566
223
598.28
99.85
4.6xl011
Raps/hr
10
5
5
2
1
-
6
800
Joy- Western
3,940,000
662
1,209,600
307.00
99.5
6
9
Hanging Round
Wires
Magnetic Impulse
Hammers
1/16
800
268,696
678
281.36
98.98
Raps/7 3mln
8
8
3
3
1
1
31
65-33
30
85
36-29
63-44
Indicating range of values from two methods of calculation.
-------
Figure 7 contains the collection efficiency as a function of particle
size at Plant 1 for sampling periods in which the rapping system was energized
and deenergized. Due to the previously discussed difficulty with process
variations at this location, data from the large particle real-time system
were used to estimate the collection efficiency for particles greater than 1.0
pm. Figure 8 presents analogous data from Plant 2, with the exception that
the rap and no rap data were obtained with the alternating sampling plan using
the ultrafine sizing system and inertia! impactors. Fractional efficiency
data with the rapping system operating normally are given in figure 9. For
particles larger than about 1.0 pm diameter, the fractional efficiency data
for the "normal" and rap sequence data sets indicate reasonable agreement.
The fractional efficiency data sets for one of the hot-side units are
presented in figure 10. The normal (not shown) and rap sequence data sets
again show reasonable agreement for sizes greater than about 1.0 pm. For
these locations, it is apparent that rapping losses occur for the most part in
the larger particle sizes, primarily as particles larger than 2.0 pm diameter.
Thus, it appears that rapping reentrainment does not cause a significant
change in fine particle emission for these three installations. However, the
rapping losses for both the hot- and cold-side precipitators provide a major
contribution to the overall penetration, and illustrate that significant
improvement in overall mass collection efficiency may be possible by optimi-
zation of the rapping system design and operation.
Figure 11 contains rapping emission for the six installations as a function
of the dust calculated to have been removed in the last field of the precipitator.
The data suggest a good correlation exists between rapping losses and the dust
removal rate in the last field. Data for the two hot-side installations
tested show higher rapping losses, which would be expected due to reduced dust
adhesivity at higher temperatures. The simple exponential relationships shown
are suggested for interpolation purposes. Note that reduction of input power
at Plant 4 caused a pronounced increase in rapping emissions.
CONCLUSIONS
Measurements of rapping reentrainment on a pilot-scale precipitator
indicated that rapping emissions decreased with increasing time between raps.
Particle-size distribution data for the rapping puffs suggested that thicker
dust layers produce larger reentrained particles upon rapping.
Measurements of fractional efficiency with and without rapping at full-
scale precipitator installations show that rapping efficiency losses occur
primarily for particle diameters greater than 2.0 pm diameter. The largest
rapping losses were measured on hot-side installations. Mass emission data
suggest a correlation between the dust removal rate in the last field of the
precipitator and the emissions due to rapping. The electrostatic precipitator
with the highest overall mass collection efficiency exhibited a minimum collec-
tion efficiency of 99.2 percent at 0.20 pm diameter.
40
-------
RHO = 2.27
PENETRATION-EFFICIENCY
• EFFICIENCY • NORMAL 10/8/76,10/11/76
102T
101
cc
HI
z
UJ
0.
I-
Ul
u
cc
UJ
Q.
10°
10-
-54
I 2
0.0
90.0
99.0
uu
u
GZ
u.
UJ
H
Z
LU
u
cc
01
o.
99.9
99.99
10-
10°
PARTICLE DIAMETER (MICROMETERS)
10s
Figure 7. Plant no. 1 rap-no rap fractional efficiency, including
ultrafine and impactor measurements.
41
-------
RHO = 2.34 PENETRATIOIM-bhl-IUICiMOT
102
101
PERCENT PENETRATION
_>
o
o
1(T1
1 \J
10-2
1
-
*i*
I A T -A
i i 9 *
f . '%
• 5 §
V- **
•; }* -
£
}
i i i
0~2 10~1 10° 101 10
-I 90.0
H 99.0
u
z
UJ
UJ
Z
UJ
U
CC
UJ
a.
H 99.9
99.99
PARTICLE DIAMETER (MICROMETERS)
Figure 8. Plant no. 2 rap-no rap ultra-fine and impactor fractional efficiencies.
42
-------
RHO = 2.34
PENETRATION-EFFICIENCY
10*
„
O
^
DC
LU
Lu 10°
Q.
PERCENT
NT1
ID-2
1(
A Rap
& No rap
-
i
T
H * $
" tj§} V'H "
|
u.u
90.0
u
z
LU
O
99.0 t
LU
PERCENT
99.9
99.99
r2 10~1 10° 101 102
PARTICLE DIAMETER (MICROMETERS)
Figure 9. Plant no. 2 normal condition ultrafine and impactor
fractional efficiencies.
43
-------
PENETRATION-EFFICIENCY
RHO = 2.27
g
<
cc
K
UJ
Z
LU
O.
K
Z
UJ
O
QC
LU
Q.
0
o
10
-2
-> 0.0
90.0
(O
(0
b
o
Z
UJ
O
LL
U.
LU
I-
LU
U
QC
LU
a.
99.9
99.99
10
,-1
10°
101
10s
PARTICLE DIAMETER (MICROMETERS)
Figure 10. Plant no. 3 ultrafine and impactor rap-no rap fractional
efficiencies, duct Bl.
44
-------
100 r
10
to
O
UJ
O
I
y2 = .618X-894
0.155X-905
0.1
10
100
CALCULATED MASS REMOVAL BY LAST FIELD
mg/DSCM
Figure 11. Rapping emissions vs. dust removal by last field,
4a. - Plant 4, Normal Current Density
4b.- Plant 4,1/2 Normal Current Density
-------
REFERENCES
1. Herbert W. Spencer, III, "Rapping Reentrainment in a Nearly Full
Scale Pilot Precipitator," Environmental Protection Agency Publication
No. EPA-600/2-76-140, May 1976.
2. 0, J. Tassicker, "Aspects of Forces on Charged Particles in Electro-
static Precipitators," Dissertation, Wollongong University College,
University of New South Wales, Australia, 1972.
3. Southern Research Institute, Final Report on EPRI Project RP413, in
preparation.
46
-------
INVESTIGATION OF DUST CARRYOVER AND OPTIMIZATION OF REGIMES
OF ELECTRODE RAPPING IN ELECTROSTATIC PRECIPITATORS*
Yu. I. Sanayev, I. K. Reshidov, I. A. Kizim
Abstract
The authors present the results of studies on dust entrainment -in an
electrostatic preoipitator. The process of entrainment during precipitation
and rapping of precipitator electrodes is studied and possibilities indicated
of reducing entrainment during rapping by optimizing rapping conditions. Cal-
culations have been worked out for determining optimal rapping conditions. A
programmable unit has been developed for controlling the rapping of precipitator
electrodes in order to realise the calculated operating conditions. The intro-
duction of this procedure and the programmable unit makes it possible to save
from 1.5 to 25,000 rubles per year in the case of one electrostatic precip-
itator.
Modern industries, especially power plants, develop by intensifying per-
formance of individual process units. This leads to the growth of production
and at the same time to the increase of industrial wastes. Fly ash emissions
at large thermal power stations amount to hundreds of tons per day. This
deteriorates performance of I.D. fans and heat exchangers, but the main problem
is that of environment pollution.
A common device to clean large volumes of waste gases is an electrostatic
precipitator. The theory of electrostatic precipitation predicts the possi-
bility of building electrostatic precipitators with high dust removal efficiency,
However, design practice encounters many problems. At the present time, dust
reentrainment is the main problem that decreases efficiency of electrostatic
precipitators.
In this country and abroad, many authors state that dust reentrainment in
electrostatic precipitators has been inadequately studied although its influ-
ence on precipitator efficiency is significant.
To date, the dust reentrainment factor has not been accounted for in
precipitator design. This leads to inadequate design. Continuous rapping of
collecting plates is commonly used in practice. This decreases gas cleaning
efficiency and precipitator lifetime.
Increasing of dust removal efficiency by way of increasing precipitator
size leads to higher cost of equipment. One must look for less expensive
means.
The main goal of the present work is to study the physical nature of dust
reentrainment and to find parameters that influence it. Obtained information
*For further information regarding the material in this paper, please
contact I. K. Reshidov, U.S.S.R. 113105, Moscow, M-105, 1st Nagatinsky Pass.,
6, NIIOGAZ.
47
-------
will lead to the development of ways and means to increase precipitator effi-
ciency at considerably lower costs.
Theoretical evaluation of dust reentrainment at the present time is im-
possible. So the best way to study reentrainment is by experiment. The
number of factors that influence the reentrainment of dust is quite high and
to make them constant in the field tests is impossible. So it was decided
first to study the physical nature of the phenomena in the laboratory. Studies
distinguished two main kinds of reentrainment: reentrainment that occurs
during particle deposition and reentrainment due to rapping of collecting
plates. It is assumed in defining reentrainment that particles impacting an
adhesive conducting surface are not reentrained. The amount of reentrained
dust during deposition was defined as the difference between the quantity of
dust collected on the plate coated with adhesive compound and that collected
on the clean plate. The amount of dust reentrained during rapping of collec-
tion plates was defined as the difference between the quantity of fly ash
collected by the precipitator with a rapping device operated and not operated.
The interval of work without rapping of the plates was short enough so
that there was no dust removal efficiency decrease as a result of excessive
fly ash deposition on collecting plates. To account for the reentrainment,
the coefficients Kx and K2 were added to the German equation. This permitted
evaluation of the influence of reentrainment on the precipitator efficiency.
The laboratory apparatus for studying reentrainment .consist of precipi-
tators equipped with the dust feeder especially designed for their purpose and
capable of highly uniform dust feeding and continuous weighing of dust. The
dusts used in the experiments are various kinds of fly ash and cement.
Analysis and allowance for measurement errors combined with adjustment of
the experimental apparatus held the relative error of weighing of collected
dust to less than 5 percent. This made it possible to study the quantitative
correlation of dust reentrainment.
As the processes taking place during rapping of the collecting plates are
short-lived, rapid film shooting was used to study them.
For dust reentrainment studies in the field, a probe was developed for
deposition of dust on clean and adhesive discs directly from the active zone
of a precipitator.
The studies showed that dust reentrainment during deposition is due to a
jarring loose of dust from the dust layer by large particles, forming craters
in the dust layer. The process was confirmed by photographs showing particle
motion in strobe light. Calculations revealed that layer erosion takes place
when particle motion energy is not less than 108-1010 joules.
For studies of flow and gravitational forces effect on the knocking out
of particles, the dust layer on the collecting plates was blown with clean
air. The tests showed that these forces do not cause dust reentrainment.
The results of the experiments on deposition of dust on clean and adhesive
discs are presented as curves for dust reentrainment by deposition in, relation
to the gas velocity. It can be seen that for the laboratory electrostatic
48
-------
precipitator, the dust reentrainment by deposition at a gas velocity of 2 m/s
amounts to 15 percent; decrease of unevenness of gas stream from 0.28 to 0.17
m/s lowers reentrainment by 5 times. If the field length is 1.5 m, then dust
reentrainment by deposition in the laboratory precipitator can lower the
effective drift velocity by nearly 2 times.
The dust reentrainment by deposition in an industrial precipitator in the
first field can amount to 70 percent.
Studies on an industrial precipitator showed that the reentrainment value
by deposition decreases along the length of the precipitator because the par-
ticle size diminishes.
In addition to the type of reentrainment discussed above, there exists
one due to rapping of electrodes.
The analysis of films taken at the speed of 1,200 shots/s distinguished
three main stages of the rapping process.
1. The motion of the main dust mass that takes place at the vicinity of
electrodes.
2. Disaggregating of dust taking place during separation of the dust
from the surface and contact of moving aggregates with electrodes
and with other aggregates.
3. The formation and upward motion of a dust cloud.
The disaggregating of dust takes place over the whole surface of the
collecting plate so this process makes the main contribution to the quantity
of reentrained gas.
Taking into account the gained impression relative to the nature of reen-
trainment, the studies were carried out concerning the basic factors effecting
dust reentrainment. Dust reentrainment at the velocity of 1 m/s is minor; at
the velocity of 3 m/s it increases up to 30 percent. The quantity of shaken
dust considerably affects the dust reentrainment. To characterize the mode of
shaking, the authors introduce in this study a special term "dust capacity"
(m) i.e., the quantity of dust on 1 m2 of the surface of the collecting plate
just prior to rapping.
The studies showed that the collected dust distributes unevenly along the
length of the electrode. That is why the term "dust capacity" characterizes
the mean value of amount of dust that is equal to the relation of total amount
of dust to the area of the electrode.
The relationship of collection efficiency of rapped-off dust is derived
and shows that the increase of dust capacity leads to the increase of collec-
tion efficiency as a whole.
In precipitators, the spontaneous falling of dust sometimes takes place.
It starts at the place of maximum collection of dust when the dust capacity is
close to 2 mg/m2 and gradually propagates over the whole length of the electrode,
The total reentrainment due to the spontaneous falling of dust and to the
rapping has some minimum value corresponding to the optimal interval of rapping.
49
-------
The relation of rapping intervals to dust capacity was analyzed during
investigations on the pilot plant of the Cherepet thermal power station.
During the tests, the relationships were obtained for the dust reentrainment
coefficient K2, the collection efficiency during the period of no rapping, and
the collection efficiency during various modes of rapping characterized by
dust capacity m.
In the range of m = 0.75 - 1.3 kg/m2, the collection efficiency reaches
its maximum. During continuous rapping, the collection efficiency decreases
by 5 percent.
The experimental studies were carried out to determine the relationship
of collection efficiency versus dust capacity and K2. Obtained relationships
confirm the existence of a functional relation between collection efficiency
and dust capacity m.
When investigating precipitators collecting various kinds of fly ash and
dust, the dependence of optimal dust capacity on dust resistivity was established.
The high resistivity of the dust layer requires the decrease of its thickness
in order to avoid considerable voltage drop in the gap.
Investigations on individual precipitators developed a technique of
evaluation of rapping intervals. On the basis of the testing of a number of
precipitators, the applicability of this technique is shown.
Taking into account the experimental relationship m=f (pv), a formula is
derived for evaluation of the absolute duration of the rapping interval. The
latter is the function of precipitator geometry and technological parameter of
precipitator performance such as gas velocity, dust load, dust resistivity,
collection efficiency, field length, and interelectrode gap. Correlation of
rapping intervals, as to the fields, is determined by collection efficiency
and the number of fields in the precipitator.
To evaluate the mode of rapping by precipitators with often changing
modes of operation, a nomogram has been derived.
For realizing optimal intervals, a special regulating instrument has been
developed. This instrument has been passed through the Interdepartmental
Commission and recommended for mass production. The application of the said
technique and the regulating instrument increased collecting efficiency of
industrial precipitators and reliability of rapping system.
At the Ladyzhin thermal power station (precipitator EGZ-3-177), the de-
signed collection efficiency has been reached and the emissions decreased as
much as three times. At the Refta thermal power station where fly ash with
high resistivity (from Ekibastuzk coals) is collected, the emissions are now
decreased by 30 percent. The Lurgi precipitator is used here for gas cleaning.
At the Voskresensk cement plant (precipitator GP-75-3), a three times decrease
has been reached (see table 1).
Optimization of rapping intervals increased the reliability of rapping
system units as much as three to ten times. This is accompanied by electric
power savngs. Cost savings due to this new way of operation per each pre-
cipitator are from 1.5 to 23 thousand rubles per year.
50
-------
Table 1. Modes of Rapping Before and After Optimization
Ladyzhin power
station, e.s.p.
EGZ-3-177 node
of rappina
Parameters
Gas velocity,
ffl/sec
Inlet dust load,
g/rn3 (norm.
cond. )
Outlet dust
load, g/ra3
(norm, cond.)
Collection
efficiency
percent
Parameter
m.kg/m2
Time of
pause ,
inin.
Dust rentr.
coef f . , K2
I
II
III
I
II
III
before
optiniz.
1.4
18
0.43
97.5
0.07
0.02
0.006
3.3
3.3
3.3
0.75
after
optim.
1.4
16
0.17
99.0
0.86
0.86
0.86
40
138
478 ,
1
0.92
Refta power
plant, Lurgi
e.s.p. mode of
rapping
before
option* z
1.2
56
2.7
95.1
0.1
0.03
0.008
3.3
3.3
3.3
0.80
after
. optim.
1.1
55
1.9
96.5
0.1
0.1
0.1
3.3
9
38
0.88
SI antsy cement
plant, e.s.p.
PGDS-24-3 mode
of rapping
before
optimiz.
1.0
34.7
1.6
95.0
0.1
0.04
0.01
3.3
3.3
3.3
0.77
after
optim.
1.0
34.7
1.2
96.5
0.14
0.14
0.14
47
14
33
0.8
Voskresensk
cenent plant
e.s.p. GP-75-3
•ode of rapping
before
optiniz.
1.3
18
25
0.62
96.3
0.03
0.012
0.004
0.75
0.75
0.75
0.68
after
optim.
1.3
25
0.23
98.6
1.1
1.1
1.1
25
70
200
0.93
Yaroslavl
refinery
e.s.p. SG-15
mode of rapping
before
optiniz.
1.4
15
0.3
98.0
0.9
0.2
0.04
30
30
30
0.86
after
optimiz.
1.4
15
0.15
99.0
1.5
1.5
1.5
48
122
1020
0.96
Electric conditions:
volt, field
kV
corona
current,
mA
I
II
III
I
II
III
32
31
33
800
644
760
31
32
32
940
800
900
32
31
30
1310
1280
1290
32
31
30
1320
1290
1310
50
47
44
75
110
119
50
47
44
75
110
119
49
54
60
41
57
50
49
54
60
41-
57
50
51
50
49
50
60
120
51
50
49
50
60
120
-------
CONCLUSIONS AND RECOMMENDATIONS
The main results of investigations are as follows:
1. It is shoton that reentrainment of dust deposited on electrodes
is one of the main factors considerably decreasing dust collection
efficiency.
2. The physical mechanism of dust reentrainment in precipitators has
been studied.
3. The technique of evaluation of optimal intervals for rapping of
collecting plates has been developed.
4. The realization of optimal rapping intervals made it possible at the
Ladyzhin thermal power station (precipitator EGZ-3-177) to reach the
designed collection efficiency of 99 percent. The emissions have
dropped from 0.47 to 0.7 g/Am3. At the Refta power station where
the Lurgi precipitator collecting fly ash from Ekibastuzk coals is
used, the emissions have dropped by 30 percent. At the Slantsy
cement plant (precipitator PGDS 24-3), the emissions from a rotary
kiln (dry method of production) decreased by 5 tons per day.
5. Cost sayings from the application of the new technique and the
regulating instrument for optimization of rapping intervals are from
1.5 to 23 thousand rubles per precipitator per year.
Our experience in the Semibratovo branch of NIIOGAZ concerning application
of new methods shows that the said cost savings are reached at comparatively
low expenses (of the order of 300 rubles, which consists of the price of the
instrument and the cost of its mounting).
The developed technique of rapping intervals evaluation has been tested
at the cement plants, thermal power stations, regional power plants, oil
refineries,, and can be applied to the majority of precipitators in other
industries.
52
-------
INVESTIGATIONS AIMED AT THE INCREASING OF RELIABILITY AND TIME
OF SERVICE OF ELECTROSTATIC PRECIPITATOR UNITS AND GAS CLEANING
INSTALLATIONS AS A WHOLE*
V. A. Guzayev, I. V. Yermilov, L. S. Ryhzov, V. N. Saksin
Abstract
Basic trends of studies on the reliability and durability of gas cleaning
equipment in the U.S.S.R. are presented in this paper. Problems in the develop-
ment of various procedures for determining the reliability of the mechanical
components of electrostatic precipitators in light of a number of problem?
that come up in reliability studies are discussed. Basic reliability criteria
are given that find widespread adoption and on the basis of these reliability
studies under industrial conditions and bench models3 the authors present
basic 'reasons for the failure of mechanical components and relate them to the
results of studies concerned with the production of reliable electrostatic
components.
INTRODUCTION
At the present time, the problem of reliability and time of service of
mechanical equipment units of ESPs is of greatest importance in the field of
electrical gas cleaning.
Reliability of ESP operation is directly related to the efficiency of
dust precipitation of the installation. Indeed, the rupture of discharge
electrode elements, the breakdown of shaft-insulators, and some other kinds of
damage can put out of operation whole fields of an ESP. This causes an
essential drop of gas cleaning efficiency. Failures of mechanical equipment
units located inside the installation, as a rule, can be eliminated only dur-
ing periods of current and heavy repairs, because technological plants usually
are not stopped for elimination of minor defects in an ESP.
Increasing reliability of mechanical equipment units of an ESP is not an
easy task. ESP units operate in very hard conditions that strongly interfere
with their reliability.
The main factors decreasing reliability are as follows:
1. High temperature (up to 500° C);
2. Abrasive medium;
3. Corrosion;
4. Vibrations caused by rapping of emitting and collecting elec-
trodes.
There are no precedents of operation of equipment in similar conditions.
Exact simulation of mechanical equipment unit operation on a laboratory stand
*For further information regarding the material in this paper, please
contact I. K. Reshidov, U.S.S.R. 113105, Moscow, M-105, 1-st Nagatinsky
Pass., 6, NIIOGAZ.
53
-------
is very difficult. Hence, the necessity arises to carry out special tests
directly in the field or on special stands.
Investigations of reliability and time of service of ESPs in this country
are carried out in the following directions:
1. Development of means of estimation of reliability of individual
units and precipitators as a whole, as well as finding out relations
between ESP efficiency and its reliability. Investigation of these
problems is necessary for compilation of scientifically founded
methods of testing mechanical equipment units on the stand, for
determination of required levels of individual unit reliability, as
well as for estimation of reliability when making decisions as to
the overall dimensions of an ESP.
2. Investigation of reliability and service time of units in field
conditions with the aim of finding out unreliable units and probable
causes of their failure.
3. Investigation of mechanisms of damage of equipment operating in
conditions of high temperature, abrasive medium, corrosion, and
vibrations; development of methods of increasing reliability and
service time of equipment.
4. Development of highly reliable mechanical equipment units.
Development of Methods of Determination of Reliability of ESP Mechanical
Equipment Units
Reliability of mechanical equipment units of electrostatic precipitators
in this country is commonly characterized by such criteria as the parameter of
failure frequency and the probability of faultless operation.
The parameter of failure frequency is estimated by the following equa-
tion:
/.
s-
where An = number of units put out of operation in the time interval; and
n = number of units tested.
The parameter of failure frequency gives sufficiently full characteristics of
repairable units, including those in ESPs.
The intensity of failures is another frequently used criterion of
reliability. It is estimated by the equation:
(n -
where n = number of running units by the beginning of tests.
54
-------
If the intensity of failures is constant in time and the number of units put
out of operation is small in comparison to the total number of units, then the
parameter of failure frequency and the probability of faultless operation can
be estimated by the equation:
P -
(-
(3)
)
where t = time elapsed from the beginning of tests.
Clarity and simplicity of estimation are the advantages of these criteria
of reliability. But, unfortunately, they cannot be used for solution of some
important problems encountered in the course of investigations of reliability.
Such problems include:
1. Determination of "responsibility" of each unit for the decrease of
ESP efficiency;
2. Evaluation of the required level of unit operation reliability;
3. Correlation of reliability of operation of ESPs of various design;
and
4. Determination of required overall dimensions of ESPs with accounts
of possible efficiency drops due to unreliable operation of units.
The above mentioned problems can be solved using reliability criteria
directly related to the degree of gas cleaning efficiency.
An ESP consists of a series of parallel operating channels formed by col-
lecting plates. Gas cleaning takes place in the electrical field of these
channels. Generally, a failure of responsible units with time leads to the
decrease of the number of running channels or, in other words, to the decrease
of the precipitation area of the ESP (but not the area of collecting plates).
The relative decrease of the number of running channels due to the fail-
ure of some units can be classified as the coefficient of responsibility
A //c
where ANi = number of channels out of running order due to Ani failures of
units of one and the same type during the time At;
N = total number of channels.
Similar to equation (1), one can describe an expression for the parameter
of failure frequency of working channels due to the failure of units of one
and the same type
<»
rr-
where —rr- can be found by equation (4).
The parameter of failure frequency of channels due to the faults of units
of various types can be estimated by the equation:
55
-------
where X, = total number of unit types.
For estimation of the parameter of failure frequency of working channels,
the mutual influence of failures was not taken into consideration as it is
negligibly small because the number of units put out of operation in periods
between repairs is relatively small.
The probability of faultless operation of an ESP P can be, similar to
equation (3), found by the equation:
- -- -ear/.
L
where N = number of working channels in running order by time r .
Relative reentrainment of dust through a field of precipitators with the
probability of faultless operation by time i equal to/7 can be easily found by
where 0^ = gas cleaning efficiency of c field with account for reliabil-
ity;
?. = gas cleaning efficiency of the same field by the beginning of
*c the test;
= gas cleaning efficiency during the period of failure ("mechani-
• cal" cleaning).
The first member of the equation (8) shows relative reentrainment through
operating gaps of an ESP, the second, through those out of running order. By
simple transformation of the equation (8) and taking into consideration that
total reentrainment through an ESP is equal to the product of relative reen-
trainment through individual fields, one can derive the general equation of
dust precipitation efficiency, which accounts for reliability of mechanical
equipment units:
where m = number of fields
The equation (9) can be expediently used for precise calculations. Gas
cleaning efficiency of individual fields of ESP can be, for example, estimated
by the methods described in (1). Taking into consideration that the effi-
ciency of mechanical gas cleaning is about 0.2-0.3, one can use for approxi-
mate estimations the simpler equation:
where = tne general gas cleaning efficiency of an ESP in running order
It should be noted that the probabilities of faultless operation of a field
and of an ESP generally coincide.
When deriving equation (10), it was supposed that gas cleaning efficien-
cies of individual fields are equal, and that the second number of equation
56
-------
(2) is negligible compared to the first one. Analytical data have shown that
the estimation of reentrainment by the equation (10) differs from that by the
precise equation (9) by not more than 30 percent with variation of gas clean-
ing efficiency from 97 to 99 percent, and with broad variations of particle
size distribution (mean-root-square error was in the range from 2.2 to 5.6).
The error grows with the range of particle size distribution and gas cleaning
efficiency.
The results of calculations by the above equations have shown that the
main responsibility for operation of an ESP lies with discharge electrodes.
Their contribution into the reliability of an ESP is from 65 to 75 percent.
The developed technique allows a sufficiently simple estimation of re-
quired unit reliability level on the basis of tolerable decrease of gas clean-
ing efficiency during the period between repairs. It also allows during
designing to provide for necessary reserves of overall dimensions of the ESP
to compensate any possible efficiency decrease due to unit failures. Besides,
the technique correlates reliability of various design ESPs to their faultless
operation.
Field Tests of Reliability and Service Time of Mechanical Equipment Units
Initial information as to reliability of ESPs was obtained mainly through
questionnaires sent to industrial plants where precipitators were installed.
The relevance of this information was verified by the direct control tests of
ESPs when they were stopped for repair. Agreement of the failure frequency
parameters obtained by different ways was satisfactory. Analysis of mechan-
ical equipment unit defects has shown that they are due mainly to the follow-
ing causes:
1. Inadequate assembling of the equipment;
2. Design defects of the units;
3. Low quality of completing sets;
4. Inadequate maintenance of the equipment.
Study of Physical Mechanism of Mechanical Equipment Units
Our practice has shown that field tests of ESP unit reliability cannot
give complete answers to some questions related to a kind of fault, a cause of
failure, etc. The answers to these questions require special investigations
in the laboratory or on the stand to understand the physics of the processes
taking place in the ESPs.
Operation of ESP units is investigated on special stands installed in
NIIOGAZ. The stands completely simulate operating conditions of an industrial
installation, such as high temperature, dust-laden gas flow, etc. For in-
stance, a special stand was installed to find out the causes of fissures in
thrust-partition quartz insulators of the type XY-105Y, which appeared in
precipitators of the third dimension. This stand can change temperature in
the range from 20° C to 250° C, weight loading on the insulator from 0 to 5.5
tons, voltage from 0 to 80 kV. It was found that fissures in the thrust-
partition quartz insulators are caused by the great difference of temperature
expansion coefficients of quartz and steel.
57
-------
The investigations have shown that increasing of thickness of the as-
bestos gasket between the insulator and the bearing flange sharply lowers the
probability of insulator mechanical destruction during warming up of the ESP
casing. Recommendations were elaborated for prevention of fissure formation.
Development of New Highly Reliable Mechanical Equipment Units
On the basis of the investigations carried out by NIIOGAZ, modified units
of EG-type ESP were developed (shafts-insulators, rapping mechanism hammers,
bearings, etc.). They are applied in serial production of precipitators.
Further increase of reliability and service time of ESPs can be reached
on the basis of new principles:
1. Development of nonmechanical means of electrode cleaning without
rotating parts that cannot operate reliably in conditions of abra-
sive medium, corrosion, and high CaO content in dust;
2. Development of new highly reliable and effective systems of dis-
charge electrodes; and
3. Application of special coatings to prolong service time of discharge
and collecting electrodes as well as of other equipment of precipi-
tators.
At the present time, the equipment developed on the basis of new solu-
tions has been tested on the stand and in industrial ESPs in this country.
REFERENCES
1. I. V. Yermilov, "Study in Calculation of Gas Cleaning Processes in Plate
Type Electrostatic Precipitators,11 (authors's abstract of candidate
dissertation), Moscow, MEI, 1974, 24 p.
58
-------
STUDY OF OPERATING AND MAINTENANCE REQUIREMENTS
OF ELECTROSTATIC PRECIPITATORS IN THE U.S.A
Dennis C. Drehmel,* Michael Szabo.t
Abstract
The purpose of studying operating and maintenance requirements of elec-
trostatic preoipitators (ESPs) is to establish procedures that will give the
highest collection efficiency over a period of time. That is, the ESP will
be maintained at its best performance level with a minimum of lost time be-
cause of malfunctions. Since it is presumed that the reader is familiar with
ESP technology3 no description of theory or basic design will be reviewed.
This paper describes operating considerations, maintenance requirements, and
malfunction problems of modern precipitators. Much of the discussion is based
on fly ash application of the wire weight type precipitators. However, much
of the information can be applied to the use of precipitators on other proc-
esses.
OPERATING CONSIDERATIONS
Power Supply
The optimization of applied power to the precipitator is accomplished by
automatic power controls that basically vary the input voltage in response to
a signal generated by the sparkover rate. Provisions are also included to
make the circuit current sensitive to protect against overload and to allow
control in the event that spark level cannot be reached.
The automatic control circuit is divided into three functions: spark
control, current limit control, and voltage limit control. In general, spark
control is based on storing electrical pulses in a capacitor for each spark
occurring in the precipitator. If the voltage of the capacitor exceeds the
preset reference, an error signal will phase the mainline silicon controlled
rectifier (SCR) back to a point where the sparking will stop. Usually this
type of snap action control will tend to ov,er correct, especially at low spark
rates.
Proportional control, another type of spark control, is also based on
storing of electrical pulses for each spark occurring in the precipitator.
The phase back of the mainline SCR, however, is proportional to the number of
sparks in the precipitator. The main advantages of proportional control over
spark control is that the precipitator determines its own optimum spark rate
based on four factors: temperature of the gas, ash resistivity, dust concen-
tration, and internal condition of the precipitator.
*Industrial Environmental Research Laboratory, Environmental Protection
Agency, Research Triangle Park, N. C.
fPEDCO-Environmental Specialists, Atkinson Square, Suite 13, Cincinnati,
Ohio.
59
-------
The voltage limit control feature of the automatic control module limits
the primary voltage of the high-voltage transformer to its rating. For cur-
rent limit control, if the primary current exceeds the unit's rating, a signal
is generated, as with the spark control, which retards the firing pulse of the
firing circuit and brings the current back to the current limit setting.
With all three control functions properly adjusted, the control unit will
energize the precipitator at its optimum or maximum level at all times, deter-
mined by conditions within the precipitator. This will result in any one of
the three automatic control functions operating at its maximum, i.e., maximum
voltage, maximum primary current, or maximum spark rate. Once one of the
three maximum conditions is reached, the automatic control will prevent any
increase in power to reach a second maximum. If conditions require, due to
changes within the precipitator, the automatic control will switch from one
maximum limit to another.
RAPPERS
The magnetic impulse gravity impact rapper is a solenoid electromagnet
consisting of a steel plunger surrounded by a concentric coil, both in a
watertight steel case.
The electrical controls provide a number of separate adjustments so that
all rappers can be divided into a number of different groups. Each group may
be independently adjusted according to transmissometer readings. They are ad-
justed manually to provide adequate release of dust from collecting plates
while preventing undesirable stack puffing.
In some applications, the magnetic-impulse, gravity-impact rapper is also
used to clean the precipitator discharge wires. In this case, the blow is im-
parted to the electrode supporting frame in the same manner, except that an
insulator insolates the rapper from the high voltage of the electrode support-
ing frame.
VIBRATORS
The vibrator is an electromagnetic device, the coil of which is energized
by alternating current. The vibration set up by the coil is transmitted to
the high tension wire frame through a rod.
For each installation there will be a certain intensity and time period
of vibration that will produce the best collecting efficiency. Insufficient
intensity of vibrating will result in heavy buildups of dust on the discharge
wires. This condition causes the following adverse operating conditions:
• It reduces the sparkover distance bewteen the electrodes, thereby
limiting the power input to the precipitator.
• It tends to suppress both the formation of negative corona and the
production of unipolar ions necessary in the precipitation process.
60
-------
• It alters the normal distribution of electrostatic forces in the
treatment zone. Unbalanced electrostatic fields can cause the dis-
charge wires and the high-tension frame to oscillate.
Maintenance Requirements
A preventive maintenance schedule should be established for each instal-
lation, detailing the precipitator parts to be checked and maintained daily,
weekly, monthly, quarterly, semiannually, annually, and on a situational
basis. Ideally, one individual with engineering training, or equivalent
training with precipitators, should have overall responsibility for the main-
tenance program. This person should participate in study programs, seminars,
and symposiums to broaden his knowledge. Operators and maintenance personnel
should also be carefully selected and trained.
Daily. It is obvious that gross departures from normal readings on the trans-
former-rectifier meter and transmissometer indicate trouble. It is not so
widely recognized that small variations, often too slight to be noticed with-
out checking of daily readings, can indicate impending trouble.
Problems that usually have a gradual, rather than sudden, influence on
precipitator performance include: (1) air inleakage at air heaters or in
ducts leading to the precipitator, (2) dust buildup on precipitator internals,
and (3) deterioration of electronic-control components. Such problems can be
indicated by a small, but definite, drift of daily meter readings away from
baseline values.
For example, sudden tripout of an apparently normal electrical set prob-
ably indicates a short or ground in the secondary circuitry. A low but steady
voltage reading indicates a high resistance ground—such as that from discharge
wires to ground through ash accumulating above a plugged hopper, or from
clinker formation on a wire.
Fluctuating voltage, dipping to low values, suggests a broken and swing-
ing discharge electrode. Fluctuation of spark-rate meter readings does not
necessarily indicate a problem unless there is confirmation by fluctuating
voltage and/or current readings.
Probably 50 percent of all electrical set tripouts are caused by ash
buildup. Short of set tripout, buildup above the top of hoppers can cause
excessive sparking that erodes discharge electrodes. Further, the forces
created by growing ash piles can push internal components out of position,
causing misalignment 'that may drastically affect performance.
Although various indicators and alarms can be installed to warn of hopper-
ash buildup and of ash-conveyor stoppage, the operator can doublecheck by
testing skin temperature at the throat of the hopper with the back of the
hand. If the temperature of one or more hoppers seems comparatively low, the
hopper heaters may not be functioning properly. Generally, however, low
temperature indicates that hot ash is not flowing through the hopper and that
bridging, plugging, or failure of an automatic dump valve has held ash in the
hopper long enough for it to cool. The ash collected subsequently will pile
up at the top.
61
-------
If the temperature of all hoppers appears low to the touch, the operator
should check the ash-conveyor system to see if it has stopped or if dust
agglomeration is so great that the conveyor can no longer handle all of the
fly ash.
Daily check of the control-room ventilation system minimizes the possi-
bility of overheated control components, which can cause drift of control set
points and accelerated deterioration of sensitive solid-state devices.
Weekly. Solenoid-coil failures, fairly common when high voltage was used, are
rare with modern low-voltage equipment. Still, a weekly check of all units is
advisable. Rapper action should be observed visually, and vibrator operation
confirmed by feel. In addition, since rapping accelerations of 30 g are often
required for proper collection, an accelerometer mounted on the plates should
be checked to verify that rapping acceleration is adequate. This is best done
on a pretest check.
Control sets must be checked internally for deposits of dirt that may
have penetrated the filter. Accumulation of dirt can cause false control
signals and can damage such large components as contractors and printed cir-
cuits.
Finally, filters in the air supply lines to control cabinets and the pre-
cipitator top housing should be checked and cleaned, if necessary, to prevent
plugging.
Monthly. Most new precipitators incorporate pressurized top housings that
enclose the bushings through which high-voltage connections are made to the
discharge electrodes within the precipitator box. This prevents leakage from
the precipitator into the housing, which could cause ash deposits and/or
moisture condensation on the bushings, with a risk of electrical breakdown at
the typical operating potential of 45 kV d.c.
Inspect bushings visually and by touch for component vibration. Check
differential pressure to be sure that the fan that pressurizes the housing is
in good operating condition. Also, operate manually the automatic standby fan
to make sure it is service-ready.
Quarterly. Quarterly maintenance includes inspection of electrical-distribu-
tion contact surfaces, which should be cleaned and dressed and the pivots
lubricated, if this is not done even more frequently. These could cause false
signals. Further, since transmissometer calibration is subject to drift,
calibration should be verified to avoid the possibility of false indications
of precipitator performance.
Semiannual1y. Routine inspections, cleaning, and lubrication of hinges and
test connections should be done semiannually. If this task is neglected,
extensive effort eventually will be required to free-up test connections and
access doors, involving expensive downtime.
62
-------
Exterior inspection for corrosion, loose insulation, exterior damage, and
loose joints can identify problems while repair is still possible. Special
attention should be given to points at which gas can leak out as fugitive
emissions.
Annually. Scheduled outages must be of sufficient duration to allow thorough
internal inspection of the precipitator. Checks should be made for: (1)
possible leakage of oil, gas, or air at gasketed connections, (2) corrosion
where heat loss is great or gas temperatures are low, and (3) possible misalign-
ment of internal components. Also, high-voltage switchgear should be inspected
for possible binding, misalignment, or defeated interlocks—defects that
create a safety hazard in addition to reducing performance.
All insulator support bushings, rapper insulators, and antisway insula-
tors should be cleaned and inspected for hairline cracks and evidence of
tracking. Faulty insulators can cause excessive sparking and voltage loss,
and can fail abruptly, possibly even explode, if allowed to deteriorate.
Situational. Certain preventive-maintenance and safety checks are so impor-
tant that they should be performed during any outage of sufficient length,
without waiting for scheduled downtime. Air flow readings should be compared
with baseline values to detect possible performance deterioration. Further,
meter readings taken immediately upon restoring the precipitator to service
can serve as a check on any changes that may have resulted from maintenance
done during the outage.
Critical internal alignments should be checked whenever an outage allows
and immediate corrective action taken if misalignment is discovered. Control-
cabinet and top-housing interiors should be checked during any outage of 24
hours or more and cleaned if necessary. Any outage of more than 72 hours pro-
vides an opportunity to check grounding devices, alarms, interlocks, and other
safety equipment, and to clean and inspect insulators and bushings.
Malfunctions
Discharge Wire Breakage. Probably the most common problem associated with
suspended wire electrode-type ESPs is wire breakage, which typically causes an
electrical short circuit between the high-tension discharge wire system and
the grounded collection plate. This electrical short trips the circuit breaker,
disabling a section of the ESP, which will remain disabled until the broken
wire is removed from the unit.
Electrical erosion, the predominant cause of failures, occurs when re-
peated electrical sparkovers or arcs occur in a localized region. A sparkover
causes localized heating and vaporization of a minute quantity of metal with
each spark. If the sparkover occurs at random locations, no serious degrada-
tion of the discharge electrode occurs. If the sparkover occurs repeatedly at
the same location, however, significant quantities of material can be removed,
with subsequent reduction of cross-sectional area and ultimate failure at this
point.
63
-------
Localized sparking can be caused by misalignment of the discharge elec.~
trodes during construction or by electric field variations caused by "edge
effects where the discharge and collection electrodes are adjacent to each
other at the top and the bottom of the plates. Corrective measures for elimi-
nating failure at these points are adding shrouds and providing a rounded
surface at the end of the collection electrode to reduce the tendency for
sparking.
Electrical erosion can also be caused by "swinging" electrodes, which can
occur when the mechanical resonance frequency of the discharge wire and weight
system is harmonically related to the electrical frequency of the power supply.
Poor workmanship during construction can also cause electrical failures
of the discharge electrode. If pieces of the welding electrode remain attached
to the collection plate, localized electric field deformation can lead to
sparking and ultimate failure of the discharge electrode.
Mechanical fatigue occurs at points where wires are twisted together and
a continued mechanical motion occurs at one location. This situation is found
at the top of a discharge electrode where the wire is twisted around the
support collar. Methods of reducing mechanical fatigue include selection of
discharge electrode material that is less susceptible to cold-work annealing
after attachment or modification of the design of the corona wire attachment.
COLLECTION HOPPERS AND ASH REMOVAL
Hoppers and ash systems often constitute problems in precipitator opera-
tion. If the hoppers become full, the collected dust may short-circuit the
precipitator. The power through the dust may fuse the dust, forming a large
clinker-type structure called a "hornet's nest." This structure further
interferes with ash removal and must be removed. Most problems associated
with hoppers are related to providing for proper flow of the dust. Improper
adjustment of the hopper vibrators or failure of the conveyor system are the
usual causes of failure to empty the hoppers. It may be necessary to provide
heat and/or thermal insulation for the hoppers to prevent moisture condensa-
tion and resultant cementing of the collected dust.
Malfunctions of the evacuation and removal system include ash water pump
failure, water jet nozzle failure, disengagement of vacuum connections, and
failure of sequencing controls.
INSULATOR/BUSHING FAILURE
Suspension insulators are used to support and isolate the high-voltage
parts of an ESP. Inadequate pressurization of the top, housing the insulators,
can cause ash deposits and/or moisture condensation on the bushing, which may
result in electrical breakdown at the typical operating potential of 45 kV
d.c.
Corrective or preventive measures include inspection of ventilation fans
for the top housing and availability of a spare fan for emergencies. Frequent
cleaning and checking for damage of the fans by vibration is also necessary to
ensure trouble-free operation.
64
-------
CONCLUSION
Performance of the electrostatic precipitator can be ensured by mainte-
nance practices on daily, weekly, monthly, quarterly, semiannual, and annual
schedules. Particular attention should be given to hopper and discharge
electrode problems which lead to the most common malfunctions. Proper opera-
tion of automatic power supplies, rappers, and vibrators is essential for high
collection efficiency.
REFERENCES
1. M. F. Szabo and R. G. Gerstle, "Operation and Maintenance of Particulate
Control Devices on Coal Fired Utility Boilers," Pedco Environmental Inc.,
EPA-600/2-77-129, July 1977.
2. P. P. Bibbo and M. M. Peaces, "Defining Preventative Maintenance Tasks
for Electrostatic Precipitators," Research Cottrell, Inc., Power,
August 1975, pp. 56-58.
3. H. L. Engelbrecht, "Plant Engineer's Guide to Electrostatic Precipitator
Inspection and Maintenance," Air Pollution Division of Wheelabrator Frye
Inc., Plant Engineering. April 1976, pp. 193-196.
4. M. F. Szabo and R. G. Gerstle, "Electrostatic Precipitator Malfunctions in
the Electric Utility Industry," Pedco Environmental Inc., EPA-600/2-77-006,
January 1977.
5. Handbook for the Operation and Maintenance of Air Pollution Control
Equipment, Frank Cross Jr., P.E., and Howard E. Hesketh, Ph.D, P.E.,
Eds., 1975.
6. Sabert Oglesby, Jr., "A Manual of Electrostatic Precipitator Technology,"
Southern Research Institute, August 1970.
7- "Electrostatic Precipitator Manual," The Mcllvaine Co., Chapter VI, Sec-
tion E, copyright 1976.
65
-------
MATHEMATICAL MODELING OF ELECTROSTATIC PRECIPITATOR FILTERS AND
ITS APPLICATION TO INDUSTRIAL UNITS IN THE U.S.S.R.*
I. V. Yermilov and T. I. Dimitrieva
Abstract
A Soviet mathematical model for designing industrial electrostatic
precipitators (ESPs) is described in this paper. The Soviet model is
compared with the corresponding American model. "Engineering and experi-
mental values of gas cleaning efficiency are given for a number of indus-
trial ESPs in the U.S.S.R. along with engineering and experimental values
for the fractional efficiency of dust removal for the case of an industrial
pilot ESP at the Cherepetsk State Regional Power Station. An expression is
given for the generalized coefficient that takes into consideration the
correction for reentrainment3 back corona, etc.
INTRODUCTION
The current stage in the growth of industry in developed countries is
characterized by greater attention to problems dealing with environmental
protection, particularly air pollution from industrial enterprises. The laws
of many countries contain rigid standards for permissible levels of dust
emission and effective measures have been taken to control these emissions and
comply with the given standards. In conjunction with this, requirements have
increased for more precise predictions of the design for the gas cleaning
efficiency of units, since errors in design result either in noncompliance
with standards for dust removal or excessive expenditures for gas cleaning
equipment.
Calculating electrostatic precipitator (ESP) efficiency is associated
with many difficulties. This is related to the complexity of processes as-
sociated with the electrodeposition of dust in the corona discharge field of
the ESP and also to a great extent by the probable nature of the gas stream
parameters of the gas entering the gas cleaning unit. For this reason, during
design, semi empirical procedures for calculation are used that are based on
the utilization of experimental data obtained for the efficiency of operating
units (analogs). One of the best examples of this type of procedure is given
in reference 1.
Semiempirical procedures give good results in designing ESPs that do not
differ significantly from the analogs. The calculation of ESPs intended for
new production having a higher cleaning efficiency than the analog and using
different technology for their basic production is associated with serious
errors.
More accurate predictions of cleaning efficiency can be achieved by
setting up a scientifically substantiated procedure for calculating ESPs; this
procedure is based on knowledge accumulated from the electrodeposition proc-
*For further information regarding the material in this paper, please
contact I. K. Reshidov, U.S.S.R. 113105, Moscow, M-105, 1-st Nagatinsky
Pass., 6, NIIOGAZ.
66
-------
esses for dust in the coronal discharge field. After the fundamental works of
Deutsch and Pauthenier (refs. 2, 3), the major contribution in setting up a
scientifically substantiated method for calculating ESPs was made by Soviet
and American scientists (refs. 4, 5, 6). Nevertheless, the design procedures
described in the above literature have not received broad application in
design practice. This is because these procedures only very approximately
reflect such important physical processes as reentrainment, particle charging,
drift velocity, etc. As a consequence, this results in an imprecise calcula-
tion of efficiency.
MATHEMATICAL MODELS FOR CALCULATING ESPs
In recent years, both in the U.S.S.R. and in the United States, mathe-
matical models have been developed for calculating the efficiency of indus-
trial ESPs (refs. 7, 8, 9) which take into consideration the latest informa-
tion associated with calculating dust electro-deposition processes.
The models have a number of common characteristics. First of all, in the
case of both models the overall gas cleaning efficiency is calculated on the
basis of fractional efficiency. Secondly, both the Soviet and American models
take into consideration the reentrainment of dust, gas distribution nonuni-
formity, and the particle charge component that is conditioned by ion dif-
fusion more precisely than the earlier models.
However, the method of considering these processes is not identical.
Reentrainment in the American model is taken into consideration during calcu-
lation of fractional efficiency. The correction factor for reentrainment is
determined by comparing the calculated and experimental values for fractional
efficiency. In the Soviet model, the correction for reentrainment is derived
after calculating the overall efficiency. This method of determining reentrain-
ment from our point of view is advantageous because the entrainment factor can
be easily determined from experimental data on the efficiency of the unit
without having to resort to some specific kind of calculation.
The gas distribution in the American model is taken into consideration
with the aid of the correlation factor:
t »
where: fa = is rated gas cleaning efficiency;
£ = is the root-mean-square deviation for gas velocity.
Gas distribution in the Soviet model is accounted for by an analogous
factor:
*U> 62*1- (2)
Calculations using formulas (1) and (2) for values d and
-------
American model, charge is calculated by a more complex model, which ultimately
yields results close to those obtained by the Soviet model.
The Soviet and American models calculate ESP field intensity differently.
In order to calculate the field in the Soviet model, the operating voltage and
corona onset voltage must be known. While in the case of the American model,
the voltage current and ion mobility must be known.
In the case of the Soviet model, a number of additional processes are
taken into consideration:
1. Increased particle entrainment through the inactive (or more accurately,
semi active) zones of the ESP;
2. Particle charging in a higher intensity field of the corona electrodes;
3. Dust concentration profile nonuniformity in the ESPs.
The above factors were introduced into the model as the result of study-
ing the electrodeposition processes with the studies carried out both on lab-
oratory and industrial ESPs in the U.S.S.R. (refs. 11, 12).
Both the Soviet and the American models utilize a single experimental
factor called the reentrainment factor. In the case of the Soviet model, this
factor equals 0.7 for thermal electric power stations with continuous elec-
trode rapping (one blow on the electrodes in 3 minutes for each field) and 0.8
in the case of optimal rapping intervals for each field. In the case_of ESPs
used in cement plants, this factor equals 0.80 and 0.95 respectively.
The Soviet mathematical model was verified using experimental data associ-
ated with the efficiency of industrial ESPs in the U.S.S.R. (refs. "/, 8).
Nevertheless, the quantity of experimental data used for comparison and the
range of parameter variation for the gas stream did not make it possible to
evaluate the accuracy and universality of the computational model. An addi-
tional circumstance was that the comparison was carried out only on domestic
(Soviet) ESPs that were similar in their design. For this reason it would be
interesting to verify the computational model on ESPs with designs other than
the Soviet one, and also under conditions of a wider range of gas stream
parameter variation, in particular, dust resistivity. Results are presented
below from such a comparison.
SOVIET INDUSTRIAL ESPs
Table 1 gives the calculated IP/ along with the experimental fa for the
overall gas cleaning efficiency in ESPs established in the U.S.S.R., along
with the initial data necessary for calculation. The procedure described in
references 7 and 8 was used for carrying out the calculations.
As can be seen from table 1, the calculated efficiency values are in
satisfactory agreement with experimental values in the absence of back corona
( *-^ < 5-109 ohm/m). At large resistance values, experimental values for
efficiency are lower than the calculated values. This is related to the fact
that back corona shows up large resistance values for the trapped fly ash.
This back corona is not taken into consideration in a computational model.
68
-------
Table 1. Experimental and calculated gas cleaning efficiencies of industrial ESPs in the USSR.
CD
Plant
Ladyzhinsk GRES
Tripol'sk GRES
Berezovsk GRES
Novosibirsk TETs
Reftinsk GRES
Troitsk GRES
Troitsk GRES
Troitsk GRES
Cherepetsk GRES
Cherepetsk GRES
Cherepetsk GRES
Cherepetsk GRES
Ulyanovsk
Cement Plant
Balakleysk
Cement Plant
Vol'sk
Cement Plant
Slantsev
Cement Plant
Type of
ESP
EGZ-3-177
YG2-4-53
PGD-3-50
UG2-4-53
"Lurgi"
UGZ-4-265
Pilot
Pilot
DGPN-2-55
Pilot
Pilot
Pilot
PGD-4-38
PGD-4-50
DGPN-3-35
PGD-2-20
17
KV
U.
3S.7 19.2
M
L
M
d.
M
I,
T
•K
P
HM
U.
JL.
f° M
*-V
0.1375 12,0 0,18 0,294 308 745 1,09 0,098 0,248
41,1 21,7 0,1375 10.0 0,10 0,294 410 740 1,93
45,8 22.4
52,0 28,6
.0.1625 7,5 0.19 0,75
0.1375 10,0
3!,025".50,I50
3C.3
33.2
32,8
38,2
'40,0
35,0
30,0
59,3
58.6
57,0
55,3
25.0
23,5
23,0
20,0
22,0
22,0
22,0
28,2
27,3
27.1
26,3
0,1375
13.65
16,0
0,1375 10,0
0.1375
0,1625
0.1375
0,1375
0,1375
0.1625-
0,1625
0,1625
0,1625
10,0
6,6
5.0
5,0
5.0
10,0
10.0
9k96
5.0
O.JO 0,294
O..I6
0,18
0.18
0.18
0,18
0.18
0.18
0,16
0.19
0.19
OJ85
0,19
0,75
0,294
0,294
0,294
0,294
0,294
0,294
0,294
0.75
0,75
0,75
0.75
427 750
410
406
400
399
394
422
403
403
401
460
483
463
323
750
1.12
1,71
743 1.04
748 1,38
744
746
732
740
740
742
750
748
749
758
2.09
1,87
1,24
1,50
1,50
1,50
1,20
1,12
1,42
1.07
0.093
0.104
0.151
0.056
0^060
0,069
0,080
0,045
0,180
.0.090
0,043
0,071
O.OG5
0,055
0,140
0,272
0,290
0,265
0.272
0.246
0,260
0,254
0.276
0.259
0,259
0,259
0.31S
0,3-12
0,318
0,163
l&
23,0
29,0
20,7
18.0
14,0
30.0
30,0
30,0
30,0
19,0
19.0
19,0
14
12
fl,o
3,8
6
€
4
3,15 6000 1,1
4,63 5000 1,4
2,88 5000 1,4
3,16
3,12
3.62
3,62
3,62
3,3
3,66.
3.66
3.66
3,38
2,75
2.96
1.9
5000
5000
5000
5000
5000
5.2
o.o
8,0
8.0
5000
5000
5000
5000
1.4
I.I
I.I
1,1
I.I
1. 55
1.05
1,05
1.05
I.I5
I, IS
'•
V
0,915 0,7
0,89
0,88'
0,89
0,95£
0,8
0,7
0,7
O.fl
0,315*0.?
0,915
0,915
0,86
0.9
0,9
0,9
0.88
0,98
1,15 0,88
1,15 0,91
0,7
0,7
1,0
J.o
I.D
1,0
0,8
f
'ATM
2,15'ID9
1,5 -I09
\\
*«
99,6 99,1
96,7 97,4
ID^IO10 95.6 95,4
I.8.I09
97.3 90,1
I.35-I011 95,2
I. MO11
97,2
I.OS'IO11 91,2
7,7-lOW
8.3 -I09
2.9-109
2.3-I09
2,9'I09
A
3-IQ8
0.8 2-10°
'
rt
0,8 3-10°
0,8
o
I08'
93,4
95,7
98,5
55,4
95.5
99,0
99.S
97.9
98,0
97,0
98.8
">4,9
D5.4
9P,5
9S,0
95.5
93,1
9ti,9
K*
0.65
0,70
0.67
0,63
0.73
o.sr
0,60
Q,57.
0,79
99,2.0.82
97.6 0,79
95,7 Ot88
7£
88.9
95", 9
95.9
97-S
95.4
S6.9
92,0
91,9
se.a
99,3
97.5-
95.8'
Note: I/ Experimental data were obtained by workers of SFNIIOGAZ under the guidance of I.A. Kizim,
V.N. Nesin, V.D. Pustovoyt, V. G. Razumnyi, A.I. Pol'shchikov, I.V. Ermilov.
2/ The degree of gas cleaning of the ESPs at the Cherepetsk State Regional Electric Power Plant
was attained with no electrode rapping.
3/ The specific dust resistivity was measured in the laboratory under a working gas temperature
and current density of ,Q^ Q_Q ^\ ^Q-3 A2
M
-------
In order to calculate the cleaning efficiency over a broader range of
dust resistance values, gas velocity, electrode height, etc., it is advisable
to take into consideration reentrainment, back corona, and the correction for
the dielectric permeability of particles by means of the generalized cor-
rection coefficient K .
i 4 - /- a (Sf* }Y \YJL\f/£
,»*,** - f af('s.) fa) fa) (-%) (%
where: fi - the specific volumetric resistance of dust,
a. = gas velocity,
•/ = precipitator electrode height,
3 - gas moisture content,
£ = average median particle size,
m = dust mass on the precipitator electrode,
/> ,u tf &0ldflme = base parameter values,
° Xy* = reentrainment factor,
.^ p = factor taking into consideration the back corona, and
-g = factor taking into consideration the relative charge
e reduction of dielectric particles in comparison with
the conducting particles.
Table 1 shows the calculated values for the correction factor KQ and
overall cleaning efficiency tpi , calculated with consideration given to this
correction factor. The agreement of the calculated and experimental results
in this case is significantly better than in the first instance.
Figure 1 shows the experimental and calculated fractional gas cleaning
efficiency for the industrial pilot-type ESP at the Cherepetsk State Regional
Electric Power Station. The experimental values for efficiency were obtained
in the absence of electrode rapping. The satisfactory agreement of the cal-
culated and experimental values obviously attests to the fact that the basic
factors are reflected quite adequately in the model where these factors affect
gas cleaning in the absence of electrode rapping.
AMERICAN INDUSTRIAL ESPs
Table 2 presents the calculated^ and the averaged experimental values
for the test period of the total cleaning efficiency's of some American in-
dustrial ESPs. The experimental values necessary for the calculations and the
cleaning efficiency values were obtained during the course of a joint study of
the operation of the ESPs at the Allen Electric Power Station, which was
conducted by a group of Soviet and American specialists (ref. 14).
The calculations, the same as in the first case, were carried out using
the procedure presented in references 7 and 8. The initial parameters that
were necessary for the calculation were averaged for the test period (table
2).
Some difficulties were encountered in averaging the dispersion com-
position of the trapped fly ash. The dispersion composition of the fly ash at
the inlet to the unit, as presented in reference 13, is not given in relative
units but in absolute units (^3). In order to calculate the dispersion compo-
sition in relative units, it is necessary to know the total fly ash concen-
70
-------
0,999
0,99(3
0,995
0950
0900
°'500 o.i
0.2 0.3 0.4 0,5
4 -y =35*;
2 -17
. calculated
«?.0 3.0 4,0 5.0
experimental
.JO 40 50 #?0/
Figure 1. Experimental and calculated fractional efficiencies of dust particle collection
of the industrial-pilot ESP at the Cherepetsk State Regional Power Station.
-------
Table 2. Experimental and calculated gas cleaning efficiencies of industrial ESPs in the USSR.
U (/«
Kg KB
It.
/« a
_
ti'r
a,.
Comments
ro
"GORGAS" Electric
Power Plant ESP (13) 32,5 ja.O 04IJ13 10,97 0,175 I.3M 127 715 1.8 0,0710.268 25,0 5,0 50001,15 I 0,799,6959.82
High-
temperature ESP (13) 23,1 12,00,111318,2 0,175 1.381 575 750 1,23 0,1600.165 9.0 1,2 50001,15 I 0.7 99.30 53.«
ESP 29,6 16,0 0,II'i3 10.97 0,229 1,381 602 750 1,26*0.1650.5 29,6 3,95 5000 1.J5 1 0,7 99,66 99.71 AccordinSto U-s-
measurement data
ESP at the Allen 29.6 16,0 0, HfJ 10,97 0,229 I,3e
-------
tration in the gas. Nevertheless, the fly ash concentration that was measured
in work by a dust probe was higher than that measured by the impactor (ref.
13). From our point of view, this was related to the precipitation of large
dust fractions on the nozzle and partially on the walls of the impactor, which
decreased the overall concentration. In conjunction with this, the dispersion
composition of dust in relative units was calculated on the basis of the
concentration that was measured by the dust probe. Under these conditions,
the obtained distribution contains an error in the large fraction range (d >
10 mkm). However, this does not have to reflect on the accuracy of the calcu-
lations for the overall efficiency, since the fly ash entrainment is determined
primarily by the fine fractions.
The dispersion composition of fly ash at the inlet and outlet to the ESP
at the Allen Station was also averaged for the test time. The data that were
averaged separately were obtained with the aid of Soviet equipment (impactor 2
and the 2-cyclone separator) and by American equipment (table 2). In all
cases, the dispersion composition of dust at the inlet was satisfactorily
described by a log-normal distribution. At the outlet in the fine fraction
range (d < 1 mkm) and in the large fraction range, deviation was observed from
the lognormal law.
In conjunction with the fact that the coronal electrodes in the above
described industrial ESPs are longer than the precipitator electrodes, the
on-active zone was taken to be equal to 0 in conjunction with A&*£ A
gas distribution grid with guide vanes was mounted at the ESP inlet at the
Allen Electric Power Station in order to level the gas stream velocity field.
Since accurate information concerning the guide vanes and the connecting duct
work was not known, it was not possible to measure the field velocity nonuni-
formity coefficient on the model. An analysis of analogous gas distribution
systems with a single grid (with an active cross section of 0.45-0.5) and
guide vanes with gas coming up from below showed that the nonuniformity coef-
ficient varies within the following limits AM = 1.06-1.25 (ref. 15). For
this reason -6^. was taken in the calculations as equal to the average
value--!.15.
Taking into account the above-mentioned observations, calculations were
made that showed good agreement between calculated and experimental gas clean-
ing efficiency values. The discrepancy between the calculated and experimen-
tal values may be explained, on the one hand, by errors in the initial data
used in the calculations (dispersion composition, ^ ) and, on the other hand
by errors in measuring the degree of cleaning.
Attempts were also made to compare the calculated and experimental values
of fractional efficiency of entrainment with rapping. It should be noted that
the Soviet model does not provide for calculating the fractional efficiency of
entrainment with rapping, since secondary entrainment is taken into account
after calculating the total efficiency. However, this represented a special
interest from the point of view of broadening the understanding of the elec-
trodeposition process.
Calculation of the fractional efficiency taking into account entrainment
with rapping was made on the assumption that the coefficient of entrainment
for the various fractions is constant. Comparison of calculated and experi-
mental values for the fractional efficiency should have reflected the accepta-
bility of such an assumption.
73
-------
t^yyyy
OtQ995
0.999
0.998
0.995
0.99
0.98
0.9
0.
Fi
Eff.
-
\
\
\
\
A
\
~ i i i i
\
05 O.J
o — e>
gure 2. Experin
\^
*"r
•— *.
i
— I
1_
•^
^'
+^
i i i i
/
/
/ O
/
/
O
/
/
f
/
/
]/
o
/
y
/
o-
)
d
i t i i
0.2. 0.3 O.V 0.5 J,0 £S 2.0 3.O ^0 50 tt
ilculation [fs&2 Stfv • jJoZ/P/Sv Sci-S'
cperiment Z/-/^M/Sfcc T-fy27°M' $ -/f£
lental and calculated fractional gas cleaning efficiencies
at the "GORGAS" Electric Power Plant, USA (13).
-------
0.999B
0.9995
0.999
0,998
0.995
0.99
0.98
0.95-
0.9
o.i
I. Eff .
\
\
\
;\
\
\
\
\
\
; +
r~ 1 1 1 I
\
s
\
\,
•\
\
^
f
***,
i
^^
c
OA- — "
A
i i i i
\
c
^
/
\
A
/
A
i
/
i
*
1
I
/
/
/•A
/
rf
">* O.i 0.2 0,3 OJ 0.5 J.O tf- 2.0 3.0 qo 5.0 JO.O J,M
•— r ~ calculation U '- 23.4 kv Vo = 12 kv $£ - /;
A/ — experiment £7-/^m/sec. TZS75°H' "Ru. *$**•
Figure 3. Experimental and calculated fractional gas cleaning efficiencies in
a high-temperature ESP in the USA (13).
-------
o»
0.9399F
0.9998
0.999S
Fff.
0.999
0.998
0.995
0.99
0.98
0.95
0.9
O-
7
7
d
I i i i
0.2 0.3
S
1,0 1,5 2.0 3tO %O £0
JO
— experimental US data
- calculation
U =
Figure 4. Experimental and calculated gas cleaning efficiencies in
a high-temperature ESP at the Allen Electric Power Plant.
-------
The value of the fractional coefficient of entrainment * °p (in the for-
mula for fractional degree of cleaning of gas is multiplied by the fractional
speed of drift) was found on the basis of the equation of the level of gas
cleaning found using 4. and 4^n . It was found that for all three instru-
ments dp equals approximately 0.5.
Figures 2, 3, and 4 show the calculated and experimental values for the
fractional efficiency of entrainment. As a rule, the calculated values for
fractional efficiency are slightly lower for fine fractions and higher for
larger ones. Apparently, this shows that during rapping of electrodes coagu-
lation of fine particles takes place which causes the instability of -Acp .
CONCLUSION
The conducted comparison of calculated values of ESP efficiency with ex-
perimental values from industrial ESPs in the U.S.S.R. and in the United
States shows that the mathematical model makes it possible to calculate the
total gas cleaning efficiency of ESPs with an accuracy acceptable for prac-
tice. In order to use the model in designing industrial ESPs it is necessary
to supplement the model with methodology for calculating the electrical and
dust-gas stream parameters.
REFERENCES
1. A. B. Walker and V. W. Frisch, "Experience in the USA in the Area of
Designing and Operating Hot-Side Electrostatic Precipitators on
Waste-Heat Boilers Operating on Coal," reports of Symposium on
Controlling Particulate Emissions Into the Atmosphere and Related
Problems, May 26-28, 1976.
2. W. Deutsch, "Bewegung und Ladung der Elektrizitastrager in Zylinder-
kondersator," Annalen der Physik, 1922, Bd. 68, N. 12, S. 335-344.
3. M. Pauthenier and M. Moreau-Hanot, "La charge des Particules Spheriques
dans un champ jonise," Journal de Physique et le Radium, Vol. 3 (No. 7),
1932, pp. 590-613.
4. V. N. Uzhov, "Industrial Gas Cleaning Using Electrostatic Precipitators,"
Moscow, Khimiya, 1967, p. 344.
5. H. White, "Industrial Electrostatic Precipitation," London, 1963, p. 337.
6. I. P. Vereshchagin, V. I. Levitov, G. Z. Mirzbekyan, and M. M. Pashin,
"Basis for the Electro-gaso-dynamics of Dispersed Systems," Energiya,
Moscow, 1974, p. 480.
7. I. V. Yermilov, "Calculating the Gas Cleaning Efficiency of Plate Type
Electrostatic Precipitators," Industrial and Sanitary Gas Cleaning, No.
1, 1977, pp. 5-6.
77
-------
8. I. V. Yermilov, "Calculating the Efficiency of Electrostatic Precipi-
tators for Thermoelectric Power Stations," reports of Symposium on Con- (
trolling Particulate Emissions Into the Atmosphere and Related Problems,
May 26-28, 1976, Kishinyov.
9. S. Oblesby and J. Gooch, "Mathematical Electrostatic Precipitator Model,
Analysis of Parameter Functions Which Have an Effect on ESP Operating
Efficiency," reports of Symposium on Controlling Particulate Emissions
Into the Atmosphere and Related Problems, May 26-28, 1976, Kishinyov.
10. G. Z. Mirzabekyan, "Aerosol Particle Charging in a Coronal Discharge
Field," collection of works: Strong Electric Fields in Technological
Processes, Moscow, Energiya, 1969, p. 240.
11. I. V. Yermilov, "Dust Concentration Distribution in the Coronal Field of
an Electrostatic Precipitator," Elektrichestvo. No. 7, 1974, pp. 27-31.
12. I. V. Yermilov and G. Z. Mirzabekyan, "Calculating the Gas Cleaning
Efficiency in ESPs," Elektrichestvo. No. 3, 1976, pp. 36-40.
13 "Particulate Collection Efficiency Measurements on Three Electrostatic
Precipitators," Environmental Protection Technology Series,
EPA-600/2-75-056, October 1975.
14. Report on Joint Soviet/American Tests of Dust Removal and Measurement
Equipment in Order to Study the Physical and Chemical Properties of
Aerosol Conducted in the U.S.S.R. and the U.S. in 1976 (Topics A-4 and
A-6), Moscow, NIIOGAZ, 1976.
15. I. Ye. Idel'chik, "Atlas of Gas Air Distribution Devices of Various
Industrial Units. Industrial and Sanitary Gas Cleaning," No. 6, 1976,
pp. 11-12.
78
-------
MATHEMATICAL MODELING OF ELECTROSTATIC
PRECIPITATORS IN THE UNITED STATES
Jack R. McDonald*
Abstract
Recent work in the United States toward the development of a mathemati-
cal model of electrostatic precipitation is discussed. This work includes
the calculation of electrical conditions in wire-plate geometries and the
calculation of particle charging rates. A mathematical model is described
that calculates collection efficiency in an electrostatic precipitator as a
function of particle size and operating conditions. The model determines
the electric field, particle charge, and removal efficiency as functions of
position along the length of the precipitator. Procedures for estimating
collection efficiency losses caused by nonuniform gas velocity distributions,
gas bypassage of electrified regions, and particle reentrainment are dis-
cussed. Model predictions of fractional collection efficiency are compared
with field data from a hot-side precipitator used to collect fly ash.
INTRODUCTION
The approach to modeling precipitator performance in the United States
has been to theoretically describe the individual processes involved in elec-
trostatic precipitation and to incorporate them into a complete and compre-
hensive model. Although the model still contains empirical relationships, an
objective of the modeling research has been to eliminate the use of such rela-
tionships, since it is difficult to establish their validity over the wide
range of conditions found in precipitators. Due to the large number of high
efficiency precipitators in use in the United States and increasingly more
stringent particulate emissions standards, the collection of fine particles
(0.1 to 2.0 ym in diameter) has been of great interest. In support of the
modeling research, a large number of fractional efficiency measurements have
been made under essentially idealized conditions in a laboratory precipitator
and under conditions existing in full-scale industrial precipitators, parti-
cle charging experiments have been performed, and data relevant to electrical
operating conditions have been obtained. The most recent modeling research
has involved the development of mathematical descriptions of the electrical
conditions in wire-plate geometries and particle charging by unipolar ions.
DESCRIPTION OF THE MATHEMATICAL MODEL
Calculation of Electrical Conditions
The electrical conditions in wire-plate geometries are calculated by
using a technique developed by McDonald et al. (ref. 1). In this numerical
technique, the appropriate partial differential equations that describe the
electrodynamic field are solved simultaneously and subject to a suitable
*Southern Research Institute, Birmingham, Alabama.
79
-------
choice of boundary conditions existing in a wire-plate geometry. The proce-
dure yields the voltage-current curve for a given wire-plate geometry and
determines the electric potential, electric field, and charge density distri-
butions for each point on the curve. Figures 1 through 6 show comparisons of
the predictions of this technique with available experimental data (refs. 2
and 3). The agreement between theory and experiment is within 15 percent.
Calculation of Particle Charge
Particle charge is calculated from a unipolar, ionic-charging theory for-
mulated by Smith and McDonald (ref. 4). In this approximate theory, particle
charge is predicted as a function of particle diameter, exposure time, and
electrical conditions. The charging equation is derived based on concepts
from kinetic theory and determines the charging rate in terms of the proba-
bility of collisions between particles and ions. The theory accounts simul-
taneously for the effects of field and thermal charging and accounts for the
effect of the applied electric field on the thermal charging process. For
large particles and high applied electric fields, the theory predicts essen-
tially the same charging rate as the classical field charging equation. For
low applied electric fields the charging equation reduces to the classical
diffusion equation.
The agreement between the results predicted by the theory and experi-
mental data (refs. 5 and 6) for cases where electron-charging can be ignored
is within 25 percent over the entire range of available data and is within
15 percent for practical charging times in precipitators. Figures 7 through
10 show comparisons between the theory and experiment. The theory agrees
well with experimental data on the charging of fine particles where particle
charging is difficult to describe physically and mathematically.
Ideal Calculation of Particle Collection Efficiency
The details of the mathematical model for electrostatic precipitation
and a computer program, which performs the various operations, are described
elsewhere in the literature (refs. 7, 8, and 9). In the following paragraphs,
the mathematical framework of the model is briefly outlined.
Using probability concepts and the statistical nature of the large num-
ber of particles in a precipitator, particle collection efficiencies can be
expressed in the form of an exponential-type relationship. In the model,
the collection fraction (n- J for the 1th particle size in the jth increment
1 >j
of length of the precipitator is given by
n1fj-l -exp (-w^j
where w. , (m/s) is the migration velocity of the ith particle size in the
' 9 J
jth increment of length, A..(m2) is the collection plate area in the jth
increment of length,and Q(m3/s) is the gas volume flow rate. The derivation
of equation (1) is based on the simplifying assumptions that (1) the gas is
flowing in a turbulent pattern at a constant mean forward velocity; (2) turbulence
80
-------
18
16
?14
x
(M
S /
12
U
ai
Q
£ 8
111
o
CC
£ 6
Sx = 0.1143m
Sy = 0.07348 m
b = 1.8 x 10'4 m2/volt-sec \
6 = 1.0 •
f= 1.0 f
• THEORETICAL
O PENNY AND MATICK'S
OPERATING CONDITIONS
• ®
/ /
; ;
a = 3.048 x 10'4 m / / ,
/ ' a = 1.016 x 10-3 m
7 /
/ /
• •
i j
i !
10 20 30 40 60 60
APPLIED VOLTAGE, kV
Figure 1. Theoretical (ref. 1) voltage-current characteristics for the
geometries used by Penney and Matlck (ref. 2).
81
-------
40
30
2 20
10
Sx = 0.1143 m
Sy = 0.07348
b = 1.8 x 10'4 m2/volt-sec
6 = 1.0
O EXPERIMENTAL) = 3
• THEORETICAL /
ACVDCDIMCMTAI ^
0
PLATE
468
DISPLACEMENT, meters x 10'2
WIRE
Figure 2. Theoretical (ref. 1) and experimental (ref. 2) potential profiles
along a line from the corona wire to the plate.
82
-------
40
30
1 "
LU
10
Sx = 0.1143 m
Sy = 0.07348 m
b = 1.8 x 10'4 m2/volt-sec
8 = 1.0
f = 1.0
O EXPERIMENTAL } „ _ ,.0,6 x 10* m
• THEORETICAL /
^EXPERIMENTAL! 4m
^THEORETICAL f
0 2
I
•PLATE
468
DISPLACEMENT, meters x 10'2
10
WIRE
J"
Figure 3. Theoretical (ref. 1) and experimental (ref. 2) potential profiles
along a line from a point midway between the wires to the plate.
83
-------
CN
UJ
3
O.
1
UJ
Q
LU
QC
OC
O
UJ
cc
UI
r~—i r
10
m
b
6
f
1.5 x 10'4
0.115 m
0.05 m
1.60 x 10'4 m2/wolt-$ec
1
1
OPERATING VOLTAGE FOR
TASSICKER'S DATA
I
20
APPLIED VOLTAGE, kV
30
Figure 4. Theoretical (ref. 1) voltage-current characteristic for the geometry
used by Tassicker (ref. 3).
84
-------
0.07
0.06
0.05
CM
u
lit
< 0.04
a.
fc
V)
Q 0.03
UJ
cc
oc
3
0.021
0.01
THEORETICAL
TASSICKER MEASURED
-0.05
0.0
DISPLACEMENT, m
0.05
Figure 5. Theoretical (ref. 1) and experimental (ref. 3) current density at
the plate as a function of displacement.
85
-------
TASSICKER MEASURED
(WITH DISCHARGE)
3.0
THEORETICAL
2.0
a.
X
Q
HI
iZ
E
o
TASSICKER MEASURED
(WITHOUT DISCHARGE)
1.0
THEORETICAL
(WITHOUT DISCHARGE)
-0.05
Figure 6.
0.0
DISPLACEMENT, m
0.05
Theoretical (ref. 1) and experimental (ref. 3) electric field at
the plate as a function of displacement.
86
-------
30
25
P
20
z
111
5
LU
_l
Ul
15
C3
cc
<
o 10
114
_1
O
K-
CC
PARTICLE DIAMETER - 0.18
AE= 1.8 x 106 V/m
D E = 3.6 x 105 V/m
O E = 3.0 x 104 V/m
——THEORY
I
1.0 2.0 3.0
N0t, (sec/m3 x 1013)
4.0
5.0
Figure 7. Theoretical (ref. 4) and experimental (ref. 5) particle charge as a
function of "exposure time" for a 0.18 ym diameter particle.
87
-------
PARTICLE DIAMETER
A E = 9.0 x 105 V/m
D E = 3.6 x 105 V/m
O E = 3.0 x 104 V/m
— THEORY
0.28
0.0 1.0 2.0 3.0 4.0
N0t, (sec/m3 x 1013}
5.0
6.0
Figure 8. Theoretical (ref. 4) and experimental (ref. 5) particle charge as a
function of "exposure time" for a 0.28 ym diameter particle.
-------
400
350
300
z
D
250
UI
UI
-I
UI
200
(S
<£
<
O
UI
_l
O
BC
150
100
50
PARTICLE DIAMETER - 0.92
9.0 x 105 V/m
D E = 3.6 x 10s V/m
O E = 6.0 x 105 V/m
THEORY
I
I
0.0 1.0 2.0 3.0
N0t, (sec/m3 x 1013)
4.0
5.0
Figure 9. Theoretical (ref. 4) and experimental (ref. 5) particle charge as a
function of "exposure time" for a 0.92 um diameter particle.
89
-------
103C
D E = 1.08 x 106 V/m
N0t = 1.0 x 1013 sec/m3
THEORY
0.4 0.6 0.8 1.0
PARTICLE DIAMETER, (urn)
Figure 10. Theoretical (ref. 4) and experimental (ref. 5) particle charge as a
function of particle size for low, moderate, and high electric fields.
90
-------
is small scale, fully developed, and completely random; and: (3) particle
migration velocities are small compared with the gas velocity.
Equation (1) gives a reasonable representation of fine particle collec-
tion provided that the migration velocities can be determined accurately.
The migration velocities are calculated from
w. . = q. .E.C./6ira.y ,
1>j •>u J i i
where
q. . = charge on the ith particle size in the jth increment of
1§J length (coul),
E. = electric field adjacent to the collection electrode in the
jth increment of length (V/m),
C.. = Cunningham correction factor for the ith particle size,
a.. = radius of the ith particle size (m), and
y = gas viscosity (kg/m.s).
The collection fraction n^ for a given particle size over the entire
length of a precipitator is given by
i iNi i /Ni 1 '
i.J i,jy Li (3)
where N. . is the number of particles of the ith particle size per cubic
meter of gas entering the jth increment.
The overall mass collection efficiency n for the entire polydisperse
aerosol is obtained from
n = L*i n,Pn- ,
1 ' (4)
where P. is the percentage by mass of the ith particle size in the inlet
size distribution.
Recent field tests conducted in the United States, where particle emis-
sions and size distributions were obtained with and without rapping, indicate
that the ratio of the measured migration velocity without rapping to that
ideally calculated from the model is nearly constant for particle sizes from
0.2 to 4.5 ym in diameter. The measured migration velocities are higher than
91
-------
those calculated from the model and the experimental data provide correction
factors given in figure 11, which are used as multipliers on the ideally cal-
culated migration velocities.
Methods for Representing Nonideal Effects
The nonideal effects of major importance are: (1) gas velocity distri-
bution, (2) gas bypassage of electrified regions, and (3) particle reentrain-
ment. These nonideal effects will reduce the ideal collection efficiency
that may be achieved by a precipitator operating with a given specific col-
lecting area. Since the model is structured around an exponential -type equa-
tion for individual particle sizes, it is convenient to represent nonideal
effects in the form of correction factors which apply to the exponential
argument. These correction factors are used as divisors for the ideally cal-
culated migration velocities. The resulting "apparent" migration velocities
are empirical quantities.
A numerical measure of the degradation of precipitator performance can
be obtained by calculating the ratio of the migration velocity determined by
the summation of point values of penetration to that determined by equation
(1). An expression for this ratio may be obtained by setting the penetration
based on the average velocity equal to the corrected penetration obtained
from a summation of the point values of penetration and solving for the re-
quired correction factor, which will be a divisor for the migration velocity
given in equation (1). A limited number of traverse calculations which have
been performed indicate a correlation between the correction factors (F.)
and the normalized standard deviation (a ) of the velocity traverse. Based
upon a pilot plant study (ref. 10), the following empirical relationship be-
tween the F.., o , and the ideal collection efficiencies (n.-) has been obtained:
F1 = 1 + 0.766 V' + 0.0755 o In (1/1-r) . (5)
g
As of the present, relationships of the form F. = F. (ru ,a ) have not been
established for full-scale precipitators. i 1 9
The effect on precipitator performance of gas bypassage of electrified
regions can be estimated if the simplifying assumption is made that perfect
mixing occurs following each baffled section. An expression for the pene-
tration (PN ) of a given particle size from the last baffled section, which
is corrected for gas bypassage, can be derived in the form
where S is the fractional amount of gas bypassage per baffled section and
NS is the number of baffled sections. Correction factors (B^) are obtained
by taking the ratios of the effective migration velocities under ideal
92
-------
_ 3
O
cc
o
o
in
o
UJ
OC
II
"• 1
i i i i i i i i \ i r
i i i i i i i i i iii
0.2 0.3 0.4 0.5 0.6 0.8 1.0 1.5 2.0 2.5 3.0
DIAMETER, Mm
5.0
Figure 11. Empirical correction factors for the ideal migration velocities.
93
-------
conditions to the "apparent" migration velocities under conditions of gas
seepage so that
In (1-n.)
Bi
.In [s + (l-S)(l-ni
Although it is difficult to quantify the complex mechanisms associated
with particle reentrainment due to (1) the action of the flowing gas stream
on the collected particle layer, (2) sweepage of particles from hoppers
caused by poor gas flow conditions or air inleakage into the hoppers, and
(3) excessive sparking, the effect of these nonideal conditions on precipi-
tator performance can be estimated if some simplifying assumptions are made.
If it is assumed that a fixed fraction of the collected material of a given
particle size is reentrained and that the fraction does not vary with length
through the precipitator, an expression can be derived that is identical in
form to that for gas bypassage:
PN = [R + (l-R)(l-n,)1/NR]NR , (8)
NR L i J
where PM is the penetration of a given particle size corrected for reentrain-
?
ment, R is the fraction of material reentrained, and NR is the number of stages
over which reentrainment is assumed to occur.
The results from several field tests on high efficiency precipitators
indicate that the emissions due to rapping reentrainment can be accounted
for by assuming that rapping losses come primarily from the last section of
the precipitator and that, due to rapping, 5 percent of the mass per cubic
meter of gas collected in the last section should be added to the total
emissions obtained without rapping. The experimental data indicate that the
emissions due to rapping reentrainment can be approximated by a log normal
distribution with a mass median diameter of 4.0 ym and a geometric standard
deviation of 2.5.
COMPARISON OF MODEL PREDICTIONS WITH FIELD DATA
FROM A HOT-SIDE PRECIPITATOR
The mathematical model of electrostatic precipitation has been used to
simulate operating conditions for a hot-side precipitator from March 12
through March 19, 1976, during a joint U.S.-U.S.S.R. field test program in
the United States. Figure 12 shows the particle size distributions at the
inlet that were obtained from measurements with modified Brink cascade im-
pactors on various days. The solid line is a representative curve which was
used to obtain an inlet size distribution for use in the model. Table 1
contains all the data that are necessary for use of the model. The individ-
ual entries were obtained by averaging all measurements of that variable over
the test period.
Figure 13 shows the fractional efficiencies obtained from the experi-
mental data and two theoretical curves predicted by the model. The theoret-
ical overall mass collection efficiency based on 0 = 0.25, S and/or R = 0.05,
94
-------
10.0
CO
o
o
c
D)
CD
Z
5
CO
1
Ul
O
a
0
o
1.0
0.1
D
o
B1-5 (3/12/76)
B1-7 (3/13/76)
B1-3 (3/15/76)
B2-4 (3/15/76)
B1-5 (3/16/76)
B2-7 (3/16/76)
B2-2 (3/17/76)
B1-4 (3/18/76)
B2-6 (3/19/76)
0.01
0.1 1.0 10.0
PARTICLE DIAMETER (PARTICLE DENSITY = 2.27 gm/cm3).
100.0
Figure 12. Measured inlet particle size distributions.
95
-------
Table 1. Data Used in Model Predictions for a
Hot-Side Precipitator
Precipitator length = 10.98 m
Particle dielectric constant = 100.0
Estimated efficiency = 99.6%
1/2 wire-to-wire spacing = 0.1143 m
Gas flow rate = 208.5 acm/s
Temperature = 601.7° K
Dynamic viscosity = 2.8 x 10~5 kg/m.s
Ion mean thermal speed = 631.0 m/s
Inlet mass loading = 0.00488 kg/acm
Density of dust particles = 2270.0 kg/m3
Wire-to-plate spacing = 0.1143 m
Corona wire radius = 0.01384 m
Gas velocity = 1.181 m/s
Pressure = 760 mm of mercury
Ion mobility = 4.85 x 10~* m2/V,s
Section # Plate Area (m2) Applied Voltage (kV) Current (A)
1 3.917 x 103 33.5 1.45
2 3.917 x 103 30.0 1.29
3 3.917 x 103 28.7 1.45
4 3.917 x 103 27.4 1.45
96
-------
99.9
„ 99.8
I
0)
a
o
114
5 99.5
u.
a.
g
m 99
8
I I
0.1
THEORETICAL BASED ON
ADJUSTED co's
THEORETICAL BASED ON
ADJUSTED u's, —
<7g = 0.25, s = 0.05,
with rapping losses
1.0
PARTICLE DIAMETER (BASED ON p = 2.27 gm/cm3)
10.0
Figure 13. Experimental and theoretical fractional efficiencies.
97
-------
and rapping losses as described earlier is 99.65 percent as compared with
the measured value of 99.67 percent.
SUMMARY
In its present form, the mathematical model for electrostatic precipi-
tation provides a basis for indicating performance trends caused by changes
in specific collecting area, electrical conditions, and particle size distri-
bution. Procedures, based on simplifying assumptions, can be used to esti-
mate the effects of nonuniform gas flow, gas bypassage of electrified regions,
and particle reentrainment. An empirical relationship is used to better pre-
dict ideal migration velocities and work is in progress to describe the physi-
cal mechanisms inherent in this relationship. Comparisons of the model pre-
dictions with field data obtained from a full-scale, hot-side precipitator
collecting fly ash show good agreement.
REFERENCES
1 0. R. McDonald, W. B. Smith, H. W. Spencer, III, and L. E. Sparks, "A
Mathematical Model for Calculating Electrical Conditions in Wire-Duct
Electrostatic Precipitation Devices," J. Appl. Phys., Vol. 48, No. 6,
2231 (June, 1977).
2. G. W. Penney and R. E. Matick, "A Probe Method for Measuring Potentials
in DC Corona," AIEE Preprint (1957).
3. 0. J. Tassicker, "Aspects of Forces on Charged Particles in Electro-
static Precipitators." Ph.D. Dissertation, Department of Electrical
Engineering, Wallongong University College, University of South Wales,
July 1972.
4. W. B. Smith and J. R. McDonald, "Development of a Theory for the Charging
of Particles by Unipolar Ions," J. Aerosol Sci.. Vol. 7 (1976), pp. 151-166.
5. G. W. Hewitt, "The Charging of Small Particles for Electrostatic Pre-
cipitation," AIEE Trans. 76_ (1957), p. 300.
6. D. H. Pontius, L. G. Felix, J. R. McDonald, and W. B. Smith, "Fine
Particle Charging Development," draft prepared for the Environmental
Protection Agency, Industrial Environmental Research Laboratory, Research
Triangle Park, North Carolina, Contract No. 68-02-1490.
7. J. P. Gooch, J. R. McDonald, and S. Oglesby, Jr., A Mathematical Model
of Electrostatic Precipitation. Environmental Protection Technology
Series, Publication No. EPA-650/2-75-037 (April 1975).
8. J. P. Gooch and J. R. McDonald, "Mathematical Modelling of Fine Particle
Collection by Electrostatic Precipitation," Atmospheric Emissions and
Energy-Source Pollution, AICHE Symposium Series No. 165, Vol. 73 (1977),
p. 146.
98
-------
9. J. P. Gooch and J. R. McDonald, "Mathematical Modelling of Fine Parti-
cle Collection by Electrostatic Precipitation," Conference on Particu-
late Collection Problems in Converting to Low Sulfur Coals, Interagency
Energy-Environment Research and Development Series, Publication No.
EPA-600/7-76-016, 68 (October 1976).
10. L. Preszler and T- Lajos, "Uniformity of the Velocity Distribution
Upon Entry into an Electrostatic Precipitator of a Flowing Gas,"
Staub.. 32 (11) (November 1972), pp. 1-7.
99
-------
THE STUDY OF THE COMPONENTS OF THE ELECTRIC FIELD IN
ELECTROSTATIC PRECIPITATORS*
V. I. Levitov, G. F. Mustafin, V. M. Tkachenko
Abstract
The authors present a research method and derived formulas for
determining the intensity vector components in a field of a unipolar
coronal discharge, in an electrostatic field, and in their combinations
with respect to the trajectory of a freely moving test sphere. A de-
scription on the experimental unit and the test equipment is given for
registering the motion trajectories of spheres. Results are presented
for the experimental treatment of the proposed method for an electrode
system ( a number of wires between two planes). It is shown that the
difference between experimental and theoretical values does not exceed
IS percent in the case of the primary intensity vector component perpen-
dicular to the precipitator electrodes. On the basis of the developed
method, the authors studied the electric fields set up by needle-strip
electrodes and combination coronal electrodes.
Electric field intensity is a principal factor governing high performance
and sizing of electrostatic precipitators (ESP).
By comparison with wire electrodes, needle corona electrodes produce
stronger electric fields. Combined discharge electrode assemblies are also
known. They induce complex fields with corona discharge and electrostatic
field.
Generally accepted environmental methods (refs. 1-3) of determining elec-
tric fields using a freely suspended test body (ref. 4) do not provide meas-
urements of the components of these complex fields. Knowledge of these com-
plex fields is of great importance in evaluating electrode assemblies and ESP
collection efficiencies.
The developed experimental method (ref. 5) permits one to evaluate the
components of intensity in the field with unipolar corona discharge, in the
electrostatic field, and in their combination.
The essence of the method is as follows. Freely suspended conducting
balls (which have been previously charged) are introduced into the discharge
gap. The charge on the balls is high enough to prevent additional charging.
The trajectory of the balls should have a peak, where initial velocity af-
fected by the field and gravity are absent. The projected trajectories of
ball motion in two mutually perpendicular planes are recorded simultaneously.
*For further information regarding the material in this paper, please
contact I. K. Reshidov, U.S.S.R. 113105, Moscow, M-105, 1-st Nagatinsky
Pass., 6, NIIOGAZ.
100
-------
Parameters of the trajectory in the proper points are estimated by photograms.
Components of electric field for these points are computed from the equation
of motion:
where M, q, and V = mass, charge, and migration velocity of a ball, re-
spectively;
3 = gravity acceleration;
_ E = electric field intensity;
Fdf = drag force.
From equation (1), it is clear that the electric field is computed in the
simplest way if the mass and the charge of the balls are constant and if the
effects of the drag force and the electric wind can be neglected.
Uniform size and density of balls provide mass stability. Constant
charge can be obtained by inducing an electric field in the charging region
that is stronger than the electric field for the studied corona discharge
field. Environmental effects, as test data with a falling ball have in-
dicated, can be ignored if a ball radius is more than 0.3 mm.
Then for the rectangular coordinate system, equation (1) becomes:
Mo BT x = %Ex (2)
Mo dT 9 = %Ey
Mo dT 2 = Mo9 ± %Ez
where M and q are constant values of mass and charge on a ball, re-
spectively. Equations (2-4) show that the components of the electric field
EX, Ey, and EZ can be
proper axis is known.
E , E , and E can be computed when acceleration of a ball moving along the
x y z
Direct double differentiation of the data leads to large errors in cal-
culations. Therefore it is reasonable to find a solution using the angle of
arrival.
Let us integrate equation (3) at i+1 and i-1, where i is an arbitrary
point of the trajectory:
+,1 *i +.1
E.dt
/•
% /
m
i - 1
Near the point i in a small segment of the trajectory corresponding to
(i - 1) -r (i + 1), the E value can be assumed constant and equal to E .
101
-------
Then:
mo Vyi + 1 - Vyi -
"
Using the formula of a free-falling ball, and taking into account that
the initial velocity at the peak of the trajectory is absent, i.e., V = VzTga
(figure 1), we finally have:
_ mod Vhi+1 tgon'+l - Vhi-i tgcn'-l
yl qo VhHl - N/FFl (5)
where h = height of falling of a sphere, and
a = angle of arrival in a vertical plane.
For another component:
E _ "1og v/ivKftgcn+l tgpi+1 - Vhi-1 tgai-1 tgpi-1
X1 qo VhT+1 - VlvKL (fi)
where p = angle of arrival in a horizontal plane.
For the third electric field component E directed along a corona elec-
trode, the expression will be the simplest when a horizontal position of
corona electrodes is considered. Then the expression for E will be similar
to equation (6).
Widely used horizontal plate ESPs use an electrode assembly with a series
of vertical corona electrodes located between two collecting electrodes (fig-
ure 1). With this electrode assembly, the electric field component, E , is
perpendicular to the collecting electrode and mainly accounts for charging and
particle collection. Another component, E , directed along the gas flow
/\
affects particle charging. The vertical component, E , parallel to the corona
electrode, occurs in three-dimensional fields, for example, in a needle corona
electrode apparatus. E takes par
very closely to corona electrodes.
electrode apparatus. EZ takes part in charging only those particles passing
The present investigation was concerned with the distribution of the -
first two electric field components, EX and E , in electrode assemblies having
a set of cylindrical strip-needle and combined electrodes located between two
planes. Experiments were conducted on a laboratory installation, which in-
cluded a batching and charging device, a model electrostatic precipitator, a
high voltage supply, and apparatus to take photos of trajectories of ball
motion at pulse illumination of the interelectrode region.
The specially developed batching and charging device blew out separate
balls by command and provided stable charging outside the field created in the
102
-------
c
C
Figure 1. Projections of the trajectory of moving the test body in two mutually
perpendicular planes.
103
-------
cylindrical-geometry corona assembly. A more accurate method, i.e., recording
trajectories of ball movement in the well-defined electrostatic field, was
used to measure charge. The experiments had shown that the maximum deviation
of the charges on the balls from an average value did not exceed 6 percent.
Therefore, the need to measure the value of charge for each ball before its
introduction in the studied field was eliminated. It simplified the task of
the investigation.
Synchronization of illuminating the test gap by a pulse light source with
movement of particles in the gap was provided in the installation. A start
button of the camera fed a signal to begin operation of the particle batching
device and the pulse light source.
Steel balls of 0.67-mm diameter were used in experiments. Size instabil-
ity and nonsphericity of the balls did not exceed 1 percent. The balls se-
lected were of a small enough size that sufficient information on the studied
field could be gained. Simultaneously, with small sizes, the fulfillment of
the main condition of charging a breakdown body was simplified with small
balls:
Eo > Ech > E
where E . = intensity of a charging field, E = intensity of the studied elec-
tric field, and E = critical intensity when the ball surface began to exhibit
corona.
For 0.67-mm balls, the critical intensity is 16.5 kV/cm. Therefore, with
them it is possible to study fields having a peak intensity value of about 16
kV/cm. This value is sufficient to study fields in the outside region of
corona discharge.
With negative corona, experiments were conducted under atmospheric con-
ditions at 50 kV. Cylindrical wires, strip-needle, and combined electrodes
were spaced plates with needles being stamped on edges. Arrangement of elec-
trodes and geometrical dimensions of the discharge gap are shown in figure 2.
The cylindrical wires had 3-mm diameter. The height of needles for strip
-needle and combined electrodes was 12 mm with 40-mm spacing between needles.
In all cases, the needles were oriented perpendicularly to the flat collecting
electrodes.
The method had been further developed experimentally for an assembly com-
prising a set of wires between two planes. This electrode assembly is used in
plate ESPs and it can be designed by several methods (ref. 6), which had been
verified by experimental data obtained both from probe measurements and by the
method using a freely-suspended test body.
Experiments were made in the cases of a middle wire along the neutral
force line (x = b/2), in the intermediate plane (x = b/4), and along the
central force line (x = 0), as well. Experimental and calculated values of E
for this electrode assembly are given in figure 3. Discrepancies between y
measured and theoretical values do not exceed 15 percent, which is quite
acceptable for practical purposes.
104
-------
X
CM
0=180
j3=180
2^=3
\ II III! \M
J
II
CM
0=190
170
85
B/4=47,5
B/2=95
II
\M
Figure 2. Schemes of electrode position. 1-collectingelectrode;2-cylindricalwire;
3-strip-needle electrode; 4-flat combined electrode.
105
-------
105 V/M E
3.0
2.5
2.0
1.5
1.0
0.05
0.10
3.0 M
Figure 3. Dependence of E component on field coordinates for the electrode
assembly including^a series of wires between two planes.
2H = 275 mm
106
-------
Investigation of possible simultaneous estimation of the electric field
components constituted the greatest interest. Figure 4 represents experi-
mental and calculated values of E and E as a function of field coordinates.
y *
The data were obtained for the same trajectory of a moving test body.
It is obvious from the figure that the measured data qualitatively agree
with the computed data. Deviation of the measured data from the computed data
for E is less than 13 percent. The E component (the field directed along
y *
the collecting electrodes) is markedly different.
In the literature, the experimental data for the E component are absent.
X
Therefore, its deduction will be of great interest for the qualitative field
characteristics. For example, considering that E rapidly decreases with in-
/\
creasing the distance from a corona electrode one can establish when the field
becomes one-dimensional. For the given case, the field is assumed to be
practically one-dimensional at Y = 2/3 h T h , i.e., in the region adjacent
to the collecting electrode. The region is 1/3 of the discharge gap.
Attention was paid to investigations of the electric field generated by
widely used strip-needle electrodes and by combined corona electrodes as well.
Figures 5 and 6 represent experimental values of E and E , respectively, for
these electrode assemblies. ^
Distribution of the E component had been obtained for the most typical
regions of the discharge gap. With strip-needle electrodes, these were planes
(against needles) of the central and neutral force lines and the intermediate
plane. With combined electrodes, these were sections passing through the cen-
ter of a flat portion of the electrode, through needles, along the neutral
wire, and in the intermediate region as well.
From figure 5 it is clear that E values for planes of neutral force
lines (curves 4 and 7) are practically the same for both types of electrodes.
With the needle plane (curves 2 and 5) and in the intermediate plane (curves 3
and 6) values in the test region of the discharge gap were higher for combined
electrodes than for strip-needle electrodes. In both cases, the difference
diminished as the collection electrode was approached.
Highest electric field values were obtained for a flat portion of com-
bined electrodes, i.e., the region with electrostatic field. In this case, E
values were 1.5-2 times more than the similar values for a plane of neutral
force lines in the strip-needle electrode assembly (curves 1 and 7). There-
fore, the mean value of E within the whole discharge gap would be substan-
tially higher for combined electrodes than for the strip-needle electrodes.
The distribution of the E component, figure 6, was obtained for the re-
/\
gion closest to the intermediate plane for both combined (curve 1) and strip
-needle electrodes (curve II) as well as for a region adjacent to the needle
plane of combined electrodes in the electrostatic side (curve III). The
107
-------
10° V/M
V
\
_ ^r % ^^ ^ ^^
1 * ^ ^^ ^^*A
\ X X X X
\
\
\
1
r
2 Ji
^
1 C
2
r
> c
\
X X X
9 10 11 12 13 J*.
15 1
"* *~"\s
^ O
*X o
1^"*»>ll. O
12 ••^.. o
13 ^»— ,
14 15
3 O O-> S
x
1 1
'y
|3 = 180 mm
2H = 300mm
2 Zt ~ 3 mm
x— x c
Experiment
o — o
• • Calculation
17 1fi
5J«--*
1 ^^
No
points
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
10'2X,M
2.195
2.32
2.45
2.595
2.75
2.88
3.05
3.195
3.28
3.47
3.66
3.82
4.00
4.20
4.42
4.65
4.88
5.15
y
10
15
Figure 4. Comparison of test values of simultaneously obtained E and E components
with theoretical data for 50 kV volta§e. y
1 • plane;
2 • corona wire
108
-------
105 V/M E
4.0
3.0
2.0
1.0
0.05
0.10
M
Figure 5. E component as a function of field coordinates for electrode assemblies:
a^series of strip-needle corona electrodes and a series of combined corona
electrodes between two planes. 2H = 275mm.
Combined Electrodes
1 -plane 11; 2- plane II - II; 3-plane I
I; 4-plane IV IV.
5 - x = o,
Strip-Needle Electrodes
6 - x = b/4, 7 - x = b/2.
109
-------
105 V/M
2.0
1.5
1.0
0.5
0
Figure 6
0.05
0.10
M
E component as a function of field coordinated for electrode assemblies:
a series of strip-needle and a series of combined corona electrodes between
two planes. 2H = 275mm.
I, II - combined electrodes; III, IV, V - strip-needle electrodes.
110
-------
figure shows that for both types of electrodes, the E component slowly de-
/V
creases with increasing the distance from a corona electrode and that it tends
to zero approximately in the middle of the discharge gap.
Thus, for both strip-needle and combined electrodes, the electric field
is practically one-dimensional in the region spreading from the center of the
discharge gap to the collecting electrodes.
CONCLUSIONS
1. The developed method permits one to evaluate the components of the elec-
tric field of a complicated configuration including unipolar volume
charge, electrostatic field, or their combination.
2. Test data, obtained for electrode assemblies comprising a set of wires
located between two planes, are in satisfactory agreement with theo-
retical values: the maximum deviation for the major E component does
not exceed 15 percent. ^
3. With the electrode assemblies studied, the electric field is practically
one-dimensional in the region adjacent to the collecting electrode. The
field is 1/3 of the discharge gap for the cylindrical wires and a half of
the discharge gap for strip-needle and combined corona electrodes.
4. Due to the electrostatic zones, corona electrodes with needles perpen-
dicular to the collection electrode surface provide a markedly higher
mean value of the E component outside the discharge gap, as compared
with strip-needle electrodes.
REFERENCES
1. V. D. Kravchenko and V. I. Levitov, "To the Theory of the Sataugh Probe,"
Izvestiya of the Acadamy of Sciences, No. 10, 1955.
2. V. I. Levitov, A. G. Lyapin, V. I. Popkov, and Tsin-Izyan-Yan, "Investi-
gation of DC Corona Field With Probe Technique Using an Oscillograph,"
Izvestiya of the Acadamy of Sciences, OT N, No. 2, 1962.
3. V. I. Levitov and I. K. Reshidov, "Peculiarities of Electric Fields in
Plate Electrostatic Precipitators," A collection "Intensive electric
fields in technological processes," Energy, 1971.
4. I. P. Vereshchagin and V. A. Babashkin, "Measurement of the electric
field of corona discharge by a method using a test body," A Collection
"Intensive electric fields in technological processes," Energy, 1971.
5. V. I. Levitov, G. F. Mustafin, and V. M. Tkachenko, Author's certificate
No. 481003 "A Method of Estimation of Electric Field," Bulletin Found-
ings, Discoveries, Industrial Specimen, Brace Marks, No. 30, 1975.
Ill
-------
6. I. P. Vereshchagin and V. I. Vasyayev: "Procedures of Calculating Corona
Discharge Field," Electricity. No. 5, 1971.
112
-------
FLUE GAS CONDITIONING IN COAL-FIRED POWER PLANTS
IN THE UNITED STATES
Edward B. Dismukes*
Abstract
An overview of flue gas conditioning is given in terms of the reasons for
its utilization, the efforts to develop the technology, and the acceptance of
the technology by the utility industry. Specific conditioning agents now in
use (sulfur trioxide, ammonia, and others) are reviewed. Conditioning mechanisms
associated with various agents are discussed; of these, resistivity modification
is perhaps most generally important, but others are also significant on occasion.
Finally, some of the practical aspects of conditioning are discussed. These
matters range from the problem of deciding when conditioning is likely to be
beneficial to the various new problems that may be encountered as side effects
of conditioning.
OVERALL STATUS OF CONDITIONING TECHNOLOGY
Reasons for Utilization
The primary reason for the utilization of chemical additives for flue gas
conditioning in this country is to maintain high efficiency of electrostatic
precipitators when they operate with high-resistivity fly ash from low-sulfur
coals. Such fuels are being burned increasingly as a means of lowering the
emission of sulfur oxides. Older precipitators that were installed to collect
ash from medium- to high-sulfur coals as a rule are not efficient enough when
the fuel is changed unless some corrective measure is taken.
Use of conditioning is but one of the corrective measures that can be
adopted. As one alternative, a so-called cold-side precipitater operating at
around 150° C may be replaced with a hot-side precipitator operating at around
350° C. As a second alternative, the gas stream may be cooled below the
design cold-side temperature. As a third alternative, the cold-side precipi-
tator may be enlarged to provide a larger ratio of collecting surface to gas
volume (ref. 1).
The foregoing statements carry the implication that lowering of the
electrical resistivity is the central objective of conditioning. Usually this
is true; however, as discussed in this paper, other conditioning mechanisms
are sometimes important.
Development of Technology
Private industry, apart from the utility industry, is the main segment of
commerce now responsible for the development of conditioning technology. The
private industry active in this area consists of chemical manufacturers,
"Southern Research Institute, Birmingham, Alabama
113
-------
precipitator manufacturers, and specialty companies related to the precipi-
tator industry. Chemical companies are interested in exploiting the potential
of a wider market for their products, whereas precipitator companies often are
interested mainly in making available a corrective measure for precipitators
they have installed in the past.
Some of the companies installing conditioning systems employ well-known
chemical agents such as sulfur trioxide and depend upon engineering expertise
in the injection of these agents to gain success in the competitive market.
Other companies have proprietary formulations or are attempting to develop
such formulations to gain a competitive advantage. As a result of efforts to
maintain secrecy insofar as possible about product composition, a full under-
standing of what can be accomplished with some commercial agents is not avail-
able.
Some utility organizations, notably, the Tennessee Valley Authority, pro-
vide exceptions to the general rule that outside industry is mainly responsible
for technical developments. The TVA organization was a pioneer in the use of
ammonia as a conditioning agent (ref. 2).
Acceptance by the Utility Industry
The utility industry, in general, has thus far regarded conditioning as
an acceptable way to upgrade the performance of old precipitators, but not as
an acceptable design feature of new precipitators. There are few precipitators
indeed (maybe only one) that have been built or planned originally with condition-
ing incorporated in their design. The usual policy for a new precipitator
that is to operate on high-resistivity ash is to build a hot-side unit or a
cold-side unit of large capacity.
The current degree of utilization of conditioning is indicated by the
results of a survey being conducted by the author. A total of about 125
privately owned utility companies were asked about their use of the tech-
nology. Out of the companies polled, roughly 25 percent have been identified
as current or former users.
It is difficult to describe how the users of conditioning are distributed
in terms of location, company size, or other factors. The users include
eastern, midwestern, southern, and western companies that burn coal from many
sources. One large company located in the Midwest and another smaller company
in the West have adopted conditioning by sulfur trioxide on a large scale.
Other companies throughout the country have similar power units operating
side-by-side with only one unit in a series equipped for conditioning. One of
the reasons why it is difficult to give an accurate current picture of the use
of conditioning is the frequently transitory nature of conditioning facilities.
There is evidence that one of the most successful vendors of a chemical system
has had about 50 percent of its installations discontinued after initial
trials.
114
-------
CONDITIONING AGENTS AND MECHANISMS
Sulfur Trioxide and Sulfuric Acid
Each of these compounds is widely used as a conditioning agent. If
sulfur trioxide is the compound used, however, it reacts with water vapor in
flue gas at the temperatures of cold-side electrostatic precipitators to pro-
duce sulfuric acid. Thus, although the two compounds maintain separate identities
up to the point of addition to flue gas, they are effectively the same substance
(sulfuric acid) after addition.
Typically, the vapor of sulfuric acid acts as an agent for lowering the
electrical resistivity of fly ash as a result of its condensation or adsorption
on ash surfaces. It undoubtedly acts jointly with water vapor in this connection
because of the high affinity of the substances for each other. Data from the
field illustrating simultaneous reductions of the acid vapor concentration and
the ash resistivity with decreasing gas temperature are shown in figure 1
(ref. 3).
Sometimes, sulfuric acid may act as a conditioning agent by another
mechanism such as increasing the cohesiveness of fly ash and reducing rapping
reentrainment of fly ash in the flue gas. Various qualitative indications of
this effect have been seen in the field (ref. 3), as might be predicted from
the tensile strength measurements made on fly ash in the laboratory (ref. 4).
300
TEMPERATURE, °F
320 340 360
380
u
h-
CO
lit
cc
1012
1011
- I
I
I
I
RESISTIVITY
1
I
1
I
20
15 I
z"
o
10
<
cc
o
2
O
O
PO
o
CO
140 150 160 170 180
TEMPERATURE, °C
0
190 200
Figure 1. Resistivity and SCL concentration as functions of temperature.
115
-------
HOURS
0900
0800
1800
1900
100 80 60 40 20 0 0 20 40 60 80 100
RELATIVE VALUE OF ^
"* ~ LIGHT OBSCURATION
Figure 2. Reduction of rapping reentrainment by ammonia.
Ammonia
This compound is less widely used in this country than sulfunc acid. Major
reasons for this are the lack of consistent field results and the lack of a
clear indication of the mechanisms by which it acts. With respect to the role of
ammonia as a resistivity modifier, in particular, the available data are not con-
sistent. In some applications, a significant resistivity effect is absent (ref. 3);
in other applications that involve different reaction environments, a significant
effect may occur.
Ammonia conditioning by mechanisms other than resistivity modification
sometimes occurs (ref. 3). One alternative mechanism is to increase the cohesive*
ness of ash and lower reentrainment loss as illustrated in figure 2; this
probably involves some type of reaction with naturally occurring sulfuric acid,
perhaps producing ammonium sulfate that acts as a binder when collected with
ash particles. Another mechanism is through a space-charge effect, again
involving a reaction with sulfuric acid to produce ammonium sulfate or bisulfate
as a fume of fine particles; evidence for this mechanism is shown in figure 3.
Ammonium Sulfate
There are strong reasons to believe that this compound is the main com-
ponent of various proprietary commercial formulations that are being widely
used. It is normally injected in an aqueous spray at high gas temperatures
(around 600° C) even when it is intended to improve precipitator performance
at temperatures around 150° C. Thermal decomposition to ammonia, sulfur
116
-------
50
40
N
v>
UJ
Q
H
30
20
10
15
NONH3
20 PPM /
OFNH3 /
/
/J
I
20 25
30
VOLTAGE, kV
35
40 45
Figure 3. Current-voltage relationship showing space-charge effect by ammonia,
117
-------
trioxide, and water is to be expected immediately after addition, followed by
recombination as the temperature decreases, as is shown by the reaction sequence
in figure 4.
Little is known about the mechanisms of action of ammonium sulfate, but
claims have been made that the compound can act as a resistivity modifier, an
agent increasing cohesiveness, or an agent enhancing space charge. Theoretical
consideration indicates that each mechanism may be important, but experimental
data to say which mechanism is predominant are lacking.
Triethyl amine
This compound is not known to be currently in use in this country. It is
worth mentioning, however, for two reasons: (1) the compound is claimed by
Australian investigators to act as an "agglomerating" agent (ref. 5), and (2)
certain proprietary formulations employed in this country are claimed to act
through the same mechanism.
The phenomenon of agglomeration could be a powerful mechanism of condition-
ing, for it offers the potential of overcoming the relative inefficiency of
precipitators for collecting particles below 2 urn in size. There is serious
doubt, however, that the mechanism is operative with any chemical now known,
whether it is triethyl amine or some other compound. The critical experimental
test of observing a change in particle-size distribution usually has not been
applied to triethylamine; in one instance in this country, the test failed to
confirm that agglomeration by a proprietary chemical was operative (ref. 6).
600 °C
H2O (Gases)
150 °C
(NH4)2S04 (Solid fume)
Figure 4. Reaction sequence of ammonium sulfate as a conditioning agent.
118
-------
Compounds of Sodium
Sodium compounds are of particular interest because they are the only
agents discussed in this paper that have been used in connection with both
hot-side and cold-side electrostatic precipitators. Their potential in both
connections is illustrated by figure 5, which shows the importance of the
sodium oxide content of the ash in determining both high-temperature and
low-temperature resistivity (ref. 7).
The plots in figure 5 indicate the advantages that may accrue by feeding
a compound such as sodium carbonate, along with coal, to a boiler. The added
compound decomposes and the sodium constituent is incorporated in the fly ash.
Conditioning by sodium compounds may also be achieved py feeding a solid
powder or an aqueous solution into the gas stream. In this event, a reduction
in resistivity may be expected because the more conductive sodium compound is
coprecipitated with ash rather than being chemically incorporated in the ash.
Compounds of Transition Metals
Certain compounds of transition metals have been added to flue gas on an
experimental basis in the belief that they might catalyze the oxidation of
sulfur dioxide and thus make a higher concentration of sulfur trioxide avail-
able as a natural conditioning agent. One compound investigated was vanadium
pentoxide; it was placed in a bed in a simulated flue gas stream (without
suspended fly ash) and found to be an effective catalyst (ref. 8). The activity
of this compound would not be remarkable, however, unless it could be shown to
have sustained activity despite the threat of poisoning by fly ash. Another
compound investigated was ferric sulfate; this compound has the potential both
of decomposing to ferric oxide and sulfur trioxide and catalyzing the oxidation
of sulfur dioxide. Tests with ferric sulfate do not appear to have given
definitive results at this stage.
PRACTICAL ASPECTS OF CONDITIONING
The use of conditioning in this country has been at best a qualified
success. There have been numerous trials of conditioning that have not yielded
the results hoped for. This has led to abandonment of the technology on
occasion; it currently leads to provisional trials of the technology without
commitment to its permanent adoption.
A basic practical question that any utility company should face is whether
conditioning is a likely remedy for a precipitator that does not operate as
efficiently as required. An answer to this question should be obtained by
performing certain diagnostic tests. Suppose that high resistivity is assumed
to be a source of difficulty. This should be confirmed by making resistivity
measurements directly or by drawing reasonable inferences about resistivity
from voltage-current data on the precipitator power units. It is unfortunately
true that such questions have not always been answered, and it is not surprising
that conditioning has given disappointing results on occasion.
119
-------
1014
1012
¥
I
—
55
UJ
1010
109
108
0.25%
Na2O
I
60 112 182 282
TEMPERATURE
442 °C
Figure 5. Resistivity of fly ash of different sodium contents in an atmosphere
containing 9 percent water vapor.
120
-------
1.
LIQUID
S03
LIQUID
H2S04
S03
EVAPORATOR
H_2S.°4___J
FLUE
GAS
LIQUID
S02
LIQUID
S
SO2
EVAPORATOR
I
CATALYST
*S02
BURNER
S03
FLUE
GAS
Figure 6. Options for Injecting S03 or
121
-------
Suppose that a problem of high resistivity has been definitely established.
A utility company must then make a choice about the type of agent to be employed,
Sulfur trioxide or sulfuric acid is probably the agent of choice from the
point of view of expected efficiency. But this choice means the commitment of
considerable capital and the disadvantages of handling noxious materials. An
alternative choice is an agent that is less certain to work but that requires
a lower investment in equipment and that can thus be discontinued more readily
if it is unsuccessful. Ammonia surely falls into this category; some of the
proprietary formulations probably fall into this category as well.
Successful injection of a conditioning agent is a matter that calls for
considerable ingenuity and expertise. A decision must be made about the types
of injection systems available (figure 6), the location in the gas train where
injection is to be made, the maximum concentration of agent that will be
required, the physical state of the agent that is to be added, and the design
of the injection manifold that will give adequate uniformity of injection.
Once an injection system is installed and satisfactorily meeting its
primary performance requirements, there are still questions to be concerned
about. Will the ash removal system continue to be adequate, or will it suffer
in efficiency from increased cohesiveness of the ash? Will sulfur trioxide
cause corrosion problems in the long run? Will ammonia or related compounds
cause blockage of the air preheater? Will sodium compounds in the boiler
cause a slagging problem? Will a chosen additive lead to emissions of a new
type that make the overall environmental impact of the additive unfavorable?
In connection with this question, it is sometimes stated that injected S03
appears quantitatively on the fly ash; data from the field are in opposition
to this claim (ref. 3).
Much needs to be done in this country to place conditioning on a sound
technical footing. Signs that such steps are being taken are given by the
systematic research and development activities underway with both private and
public funds. Still, research is not keeping pace with the proliferation of
conditioning trials, which often are motivated by the willingness of the
utility industry to examine any potential measure, however uncertain the
outcome, to avoid citation for unacceptable particulate emissions.
REFERENCES
1. L. E. Sparks, "Electrostatic Precipitator Options for Collection of High
Resistivity Fly Ash," Symposium on Particulate Control in Energy Processes,
Report EPA-600/7-76-010, September 1976.
2. J. T. Reese and J. Greco, "Experience with Electrostatic Fly-Ash Collection
Equipment Serving Steam-Electric Generating Plants," J. Air. Pollut. Contr.
Assoc.. Volume 18, 1968, pp. 523-528.
3. E. B. Oismukes, "Conditioning of Fly Ash with Sulfur Trioxide and Ammonia,"
Report EPA-600/2-75-015 or TVA-F75 PRS-5, August 1975.
122
-------
4. J. Dalraon and D. Tidy, "The Cohesive Properties of Fly Ash 1n Electrostatic
Precipitation," Atm. Environ. Volume 6, 1972, pp. 81-92.
5. E. C. Potter and C. A. J. Paulson, "Improvement of Electrostatic Precipita-
tor Performance by Carrier-Gas Additives and Its Assessment Using an Extended
Deutsch Equation," Chero. Ind.. 1974, pp. 532-533.
6. L. E. Sparks, "Effect of a Fly Ash Conditioning Agent on Power Plant Emissions,"
Report EPA-600/7-76-027, October 1976.
7. R. E. Bickelhaupt, "Surface Resistivity and the Chemical Composition of Fly
Ash," J. Air Pollut. Contr. Assoc.. Volume 25, 1975, pp. 148-152.
8. S. Kanowski and R. W. Coughlin, "Catalytic Conditioning of Fly Ash Without
S03 from External Sources," Environ. Sci. Techn. Volume 11, 1977, pp. 67-70.
123
-------
THE PROBLEM OF COLLECTING DUSTS THAT CAUSE REVERSE
CORONA FORMATION IN ESPs*
S. N. Panev, V. M. Tkachenko, and T. Y. Chelombitko
Abstract
Significant difficulties arise in removing high resistance dust by ^elec-
trical means due to the manifestation of back corona in the electrostatic pre-
cipitator. NIIOGAZ has conducted a number of studies on developing a system
for cleaning flue gases from rotary magnesite roasting furnaces where the dust
has a resistivity of 10l* to 1
-------
But lack of vapor in the plant caused abandonment of such method of gas
conditioning, and increased collection efficiency of the high-resistivity MgO
dust was obtained by reducing gas velocity in the active zone of the apparatus.
Application of hot-side ESPs is the other way to solve this problem, but
this method is also associated with an additional increase of cross section of
the apparatus. Thus, for example, installation of an ESP at 400° C instead of
140° C could cause an increase of active cross section of the apparatus of 1.6
times for the same gas velocity. In addition, the capacity of utility boilers
where flue gas is cooled from 600° to 700° C to 230° to 250° C would decrease.
The foregoing reasons were responsible for a new design of dust collection
system for six rotary furnaces (dimensions 90 m x 36 m) on the "Magnezit" plant.
The following apparatus were applied for each furnace: a group of six NIIOGAZ
TCHN-24 type cyclones, diameter 1,600 mm, as a first stage and two UG2-4-37 ESP
type as a second stage.
The Type UG2-4-37 ESP is horizontal, of uniform cross section, and is
equipped with precipitating electrodes with C-shape elements, belt-needle
corona electrodes with 40-mm spacings between the needles and needle length
of 12 mm, and has the following specifications (ref. 3).
1. Number of fields 4
2. Active cross section, m2 37
3. Active length of field, m 2.51
4. Active height of electrodes, m 7.46
5. Space between single field electrodes, mm 275
6. Total precipitating area of precipitating
electrodes, m2 3,150
7. Total active length of corona electrodes, m 8,467
8. Gas temperature, °C 250
Dust is removed from the electrodes by means of rappers.
During testing of the dust collection system, power for each ESP field
was supplied from separate type AUF-400 power supplies each having double half-
wave rectification and a nominal value of rectified (peak) voltage of 80 kV and
mean current value of 400 mA.
Units of AUF type allow control of the electrical regime of an ESP by
three methods, i.e. to achieve extreme, spark, and periodical control. Under
the conditions of extreme control, maximum mean voltage is automatically main-
tained on the ESP electrodes. Spark control provides a definite frequency of
sparkover of the electrostatic precipitator. Optimum frequency of sparking is
determined experimentally.
When the last method of control is used, periodic evaluation of voltage up
to a sparkover level and subsequent lowering of operating voltage to a given
value take place.
125
-------
The above-mentioned system of removal of high-resistivity MgO dust from
process gases has been tested by NIIOGAZ specialists on the "Magnezit" plant
with the aim of determining the ESP efficiency and optimizing the operation
of electrode rapping and electric power supply. The rotary furnace, in this
case, utilized natural gas as fuel, and besides raw materials for the purpose
of secondary calcination, the furnace was fed with dust from hoppers of the
dust collecting apparatus.
Tests of the group of cyclones showed that at a gas velocity of 3.3 m/s
and mean inlet particulate size of 18.5 m, gas cleaning efficiency was 77
percent. Particulate concentration at the outlet of the cyclones was 8 to
10 g/nm3 with a mean particulate size of 12 m.
Table 1 shows mean values of the results of type UG2-4-37 electrostatic
precipitator tests when two electrode rapping regimes and extreme principles
of voltage control are involved.
In the first regime, each electrode of all the fields was rapped every
15 min. In the second regime the intervals of electrode rapping were: for
the first field, 3.5 min; for the second field, 15 min; for the third field,
45 min; and for the fourth field, 110 min. This schedule was calculated with
methods developed by the Semibratovo branch of NIIOGAZ for determining rapping
intervals of ESP precipitating electrodes. When the ESP was tested, moisture
content of the gas was 43 to 48 g/m3 of humid gas and dust resistivity was at
the level of 1012 ohm/cm. Dust resistivity was measured directly in a duct by
using a transducer on which a layer of dust was formed in bulk and later com-
pressed.
The ESP under test was operated with back corona. This indicates not only
high dust resistivity but also allows the immediate detection of back corona by
means of measuring volt/ampere characteristics when the voltage is elevated or
lowered (ref. 4).
All ESP fields showed a considerable excess of corona currents during
voltage reduction (fig. 1).
The table shows that when the gas velocity in the apparatus is 0.5 to
0.6 m/s (gas stays in the apparatus 17 to 20 s) and the temperature is 140° C,
the type UG2-4-37 ESP provides high efficiency cleaning of process gas with
high-resistivity MgO dust. When a nonrational schedule of electrode rapping
Table 1. Mean results obtained when a type UG2-4-37 ESP
was tried for Mg02 dust collection
Rapping
Cycle
1
2
Gas temperature
ESP
(°C)
140
142
Particulate
loading
inlet outlet
(g/nm3) (mg/nm3)
8.03 55
9.60 16
Gas
velocity
(m/s)
0.52
0.57
Collection
efficiency
(%)
99.21
99.81
Particles
drift velocity
(cm/s)
3.48
4.87
126
-------
49
19
21
23
25
27
29
31
33
35
Figure 1. Corona volt-ampere characteristics in an electrostatic precipitator
type UG2 -4-37 during cleaning of magnesite dust from gas.
O- field 1, •- field 2, X - field 3, Q- field 4
voltage increase, - - - - voltage drop
127
-------
is used (schedule 1), cleaning efficiency is 99.21 percent and residual par-
ticulate loading of the gas is 55 mg/nm3. Optimization of the electrode
rapping schedule (schedule 2) allows an increase in effective particle drift
velocity of 40 percent, and lowers dust entrainment from the ESP by a factor
of 3.4.
Studies were also made of the effects of electric regime control on gas
cleaning efficiency when a rational rapping schedule was employed.
It was found that when the frequency of sparking was changed from 50 to
150 sparks/min, collection efficiency increased from 99.66 percent to 99.79
percent but did not exceed the value at extreme control (99.81 percent). Thus
an increase of active cross sections of ESPs and doubling the residence time
in the ESP allowed the achievement of effective collection of high-resistivity
MgO dust without gas conditioning.
CONCLUSION
1. In cleaning process gases from rotary furnaces calcining MgC03,
intense reverse corona is formed in electrostatic precipitators
of the UG type.
2. It is possible to eliminate the harmful influence of reverse
corona without gas conditioning by installing an additional ESP
of the UG type and increasing residence time in the apparatus
up to 17 s.
3. Optimization of the precipitator electrode rapping schedule allows
a decrease in dust entrainment from ESP by a factor of 3.4.
4. A two-staged system that consists of a group of TCHN-24 cyclones
and type UG-4937 ESPs provides cleaning of high-resistivity MgO
dust with collection efficiency of 99 percent.
REFERENCES ,
1. G. M. A. Aliev, "Development and Installation of Dust Collecting System
after Rotary Furnaces for Calcining MgC03," Industrial and Sanitary Gas
Cleaning, No. 4, p. 2, 1972.
2. G. M. A. Aliev, "Metallurgy," in Dust Collecting in Refractory Production.
M, 1971.
3. "Energy," in Reference Book on Dust and Fly Ash Collection, M, 1975.
4. J. Ermilov, B. Zolotarjov, and J. Reshidov, "Method of Detection of
Reserve Corona in ESP."
128
-------
DYNAMICAL BEHAVIOR OF COLLECTING PLATES OF ELECTROSTATIC PRECIPITATORS
AND PREDICTION OF RAPPING EFFICIENCY*
V. B. Mescheryakov and A. I. Zavyalov
Abstract
The authors present the state-of-the-art on the dynamia behavior of
industrial electrostatic precipitator electrodes. A functional relationship
is given that makes it possible to predict electrode rapping efficiency for
periodic impact rapping by calculation. The engineering data support the
results of experiments conducted with up to date equipment under laboratory
and test stand conditions. New designs have been developed based on this
study. These designs are undergoing industrial testing.
Performance of industrial electrostatic precipitators (ESPs) depends on
dust-collecting efficiency and reliable cleaning of collecting plates. To
study the dynamics of collecting plates analytically and experimentally, the
investigations were started in 1969 by the Semibratovo branch of NIIOGAZ and
the theoretical mechanics chair of the Moscow Institute of Railway Engineers.
The program of investigations is described in reference 1. By now it is sat-
isfactorily completed; based on its results the methods have been developed to
predict efficient performance of collecting plates; those of the new design
are being applied to industrial electrostatic precipitators.
First of all, an individual collecting electrode member, which can be
classified as a thin rod of open profile (refs. 2 and 3), was studied ana-
lytically and experimentally. The technique of acceleration measurements has
been tested with the aid of a specially designed laboratory plant. The next
step was to study dynamical properties of a real collecting electrode on the
stand.
Analytically, a collecting electrode was considered as a system of thin-
walled rods of open profile (fig. 1). During rapping of deposited dust by
impact, an impulse of short duration is applied to the collecting electrode.
The process of impulse transfer can be considered in steps. During the first
step the hammer strikes the anvil of the rapping bar. During the second step
the impulse is distributed over the collecting electrode. And, at last, dur-
ing the third step the impulses distributed over the plates form accelerations
in them. To dislodge the dust from the collecting electrodes effectively
these accelerations must reach some definite level. To be able to predict the
necessary level, one must know the relationship of maximum acceleration in a
collecting plate versus the weight of the hammer and the height from which it
falls. In solving the problem, the geometry and physical characteristics of
all bodies taking part in the transfer of impaction energy play an important
role.
*For further information regarding the material in this paper, please
contact, I. K. Reshidov, U.S.S.R. 113105, Moscow, M-105, 1-st Nagatinsky
Pass., 6, NIIOGAZ.
129
-------
/J
4
"(T~r.~F'
JL
.;/-
3
T^J i_j fr^xrj^rrr^
6-
C.
DC
a.
OL.
a. Collecting electrode: 1 - suspension bar
2 - collecting electrode element
3 - rapping bar
4 - anvil
b. Profile of rapping bar and suspension bar
c. Profile of collecting electrode element
Figure 1. Collecting electrode.
130
-------
To determine the contact force acting on the anvil of the rapping bar, a
nonlineal integral equation was used:
where PW^~ contact force,
m = mass of the hammer,
2f0- initial velocity of the hammer,
J° = density of rapping bar material,
F = cross section area of the rapping bar,
C- - velocity of the longitudinal wave propagation in the rod,
H = coefficient depending on the geometry of contacting surfaces.
Equation (1) was solved numerically. Transformation of it to a lineal
form results in decreased values for the time of contact (ref. 4).
The impaction impulse, S , determined from equation (1) propagates be-
tween the tails of collecting electrode members. The impulse values are found
from the following equation (ref. 5):
S
it* .- (2)
Here«C- jj--* , where m is corrected mass of the collecting electrode member
tail (taking into account the mass of that part of the collecting electrode
member, which is involved into motion during the time, t , when the impulse
takes place); M is mass of the rapping bar; and t is time during which the
compression wave propagates through the rapping bar. As is seen from equation
(2), the values of impulses transferred to the collecting electrode member
tails form a decreasing geometric progression. The last collecting electrode
member is situated near the end of the rapping bar, that is why one must take
into account the back stroke for impulse evaluation. This is done by doubling
the value found according to equation (2).
Figure 2 shows impact impulse distribution over the collecting electrode
members, the mass of a hammer being 5 kg.
If the collecting electrode members have no tails and are in direct con-
tact with the rapping bar, then the first term in equation (2) is excluded and
the parameter 0< is expressed as
-7
Here F = cross section area of the rapping bar;
F^ = cross section area of the collecting electrode member;
E,G = elasticity coefficients;
K, = form factor for the cross section of a collection electrode
member.
To estimate the value of maximum accelerations in cross sections of col-
lecting electrode members, the differential equation was used for swinging of
131
-------
= 5 kg
1 - calculation
2 - experiment
Figure 2. Distribution of impact impulse over
collecting electrode members.
132
-------
a thin-walled rod taking into account shear distortion and the inertia of
longitudinal motion. To estimate energy dissipation, the conception of fre
quency-independent inner friction in materials was used. The equation was
solved with the aid of Fourier transformation. As a result, the following
equation for estimation of maximum acceleration (ref. 6) is obtained:
-------
/;
50
0 L
103
m = 5 kg; B = 2.5 m; H = 7.9 m
1 - clean electrode
2 - electrode with fixed air-foam rubber simulating
dust layer
Figure 3. Comparison of calculated and experimental
curves of electrode rapping efficiency.
134
-------
co
tn
Figure 4. The nomogram for selection of collecting electrode design parameters.
-------
Figure 5. The nomogram for selection of collecting electrode design parameters.
-------
u>
20 30 AO SO 60 70
Figure 6. The nomogram for selection of collecting electrode design parameters.
-------
Nomograms based on experimental results were compiled for the selected
construction parameters (figs. 4, 5, and 6). The following equation allows
compilation of these nomograms using regression analysis. The equation gives
the degree of rapping efficiency for the collecting plates:
'- Zff. 5t96 - Z.?2GH+ 4 (ftm + O.S39
where
Here S~ width of electrode in m;
H ~ height of electrode in m;
m= mass of a hammer in kg;
Ciorp- value of the rapping acceleration related to the acceleration of
gravity;
J3- material density (in case of dirty electrode the equivalent den-
sity is used).
The progress in studies of dynamical behavior of collecting plates under
the influence of impaction allowed proposal of new types of collecting plates
(refs. 9 and 10). Patents were claimed and granted (refs. 11, 12). Specimens
of new collecting plates are now under test in a number of industrial ESPs in
different branches of industry. The preliminary results are satisfactory.
REFERENCES
1. V. B. Meshcheryakov, I. K. Reshidov, and V. N. Uzhov, "Problems of Reliabil
ity and Durability of Precipitator Electrodes and Possible Ways of Solv-
ing these Problems," in the collection of works: Industrial and Sani-
tary Gas Cleaning. TsNIITEneftekhim, No. 2, 1972.
2. V. B. Meshcheryakov, A. S. Arkhipov, and A. I. Zav'yalov, "Experimental
Study of the Free Oscillations of the Precipitator Elements of Electro-
static Precipitators," in the collection of works: Industrial and Sam'
tary Gas Cleaning. TsNIITEneftekhim, No. 5, 1972.
3. V. B. Meshcheryakov, I. I. Ivanchenko, and A. I. Zav'yalov., "Using the
Theory of Thin-wall Rods in Studying the Free Oscillation of Electro-
static Precipitator Elements," in the collection of works: Industrial
and Sanitary Gas Cleaning. TsNIITEneftekhim, No. 5, 1973.
4. Ye. N. Kurbatskiy and A. L. Shevelev, "Longitudinal Impact of a Concen-
trated Load Along an Elastic Rod," works of MIIT, issue 509, 1976.
5. Ye. N. Kurbatskiy, "Propagation of Longitudinal Waves Along a Rod With
Attached Concentrated Masses," works of MIIT, issue 476, 1975.
6. V. B. Meshcheryakov and Ye. N. Kurbatskiy, "An Approximate Evaluation of
Acceleration in a Beam Under the Short Term Effect of a Transverse Force,"
works of MIIT, issue 509, 1976.
138
-------
7. V. B. Meshcheryakov, "Computational Evaluation of the Rapping Efficiency
of the Industrial Electrodes of Industrial Electrostatic Precipitators,"
in the collection of works: Industrial and Sanitary Gas Cleaning, No. 1,
1977.
8. A, I. Zav'yalov, "Procedure for Studying the Dynamics of Precipitator
Electrodes," Tsement. No. 3, 1975.
9. V. B. Meshcheryakov and A. I. Zav'yalov, "Problems of Improving ESP
Precipitator Electrodes," in the collection of works: Industrial and
Cleaning of Gases and the Aerodynamics of Dust Trapping Equipment, Yaro-
slavl1, 1975.
10. A. I. Zav'yalov, Selection of Design Parameters for Precipitator Elec-
trodes Which Ensure Efficient Rapping and Reliability in Operation," in
the collection of works: Industrial Gas Cleaning and the Aerodynamics
of Dust Removal Equipment, YaroslaviV, 1975.
11. A. I. Zav'yalov, V. V. Nagornyy, L. P. Smirnov, and S. I. Sobchuk, "ESP
Precipitator Electrodes," U.S.S.R. author certificate, Moscow, Class VOZS
3/40, No. 472690, Announced 06.11.73, No. 1967934/23-26, Published
05.06.75.
12. A. I. Zav'yalov, V. V. Nagornyy, V. I. Panasyenko, and S. Sobchuk,
"The Electrostatic Precipitator," U.S.S.R. author certificate, Moscow,
Class VOZS 3/76, No. 523712, Announced 01.07.74, No. 2039906/26, Pu-
blished 05.08.76.
139
-------
ELECTRICAL GAS CLEANING AT HIGHER PRESSURES*
A. Yu. Valdberg, V. V. Danilin, A. G. Ljapin, V. M. Tkachenko
Abstract
A number of branches of industry require high efficiency cleaning of
large volumes of gas at higher pressures or at both higher pressures and
temperatures. In this report^ using blast furnace gas cleaning as an example^
the authors study problems associated with the use of dry electrostatic
precipitators to meet these objectives. Results are given from studying
the electrical characteristics of coronal discharge in air at a pressure
up to 0.6 MPa and a temperature of up to 400° C. Test results are also
given of a pilot industrial dry-type electrostatic precipitator cleaning
furnace gas at a pressure of 0. 3 MPa at a temperature up to 25° C.
Within the past few years several industries have been faced with the
necessity of treating vast gas volumes at higher pressures or at both higher
pressures and temperatures. For example, the construction of large blast
furnaces in ferrous metallurgy has resulted in considerable volumes of stack
gases at pressures of about 0.35 MPa. This has permitted use of energy from
compressed blast-furnace gases in gas turbines. Good performance of these
turbines is attainable with fine purification (to 5 mg/Nm3) of vast gas
volumes (to 800,000 Nm3/hr per furnace) at a minimum decrease of pressure and
temperature in the gas cleaning system.
Although modern wet processes provide a desirable, efficient means of
gas cleaning, they require large water consumption, a complex and expensive
sludge removal system, and additional preheating of purified gas before dis-
charge into a turbine. Development and application of dry cleaning for blast-
furnace gases has made it possible to eliminate both the sludge handling
system and the gas preheater ahead of the turbine and to increase the turbine
capacity well due to gas feed at higher temperature.
Electrostatic precipitators are a most promising device to treat large
volumes of hot gases. However, experience in operating dry ESPs to handle
blast-furnace gases at higher temperatures is lacking and the literature
contains inconsistent data for corona discharge effects on performance of
ESPs. Therefore, NIIOGAZ had undertaken an investigation to study electrical
characteristics of corona discharge at both higher temperatures and pressures,
as well as to evaluate collection efficiencies of dry ESPs as applied to
blast-furnace gases.
The investigation has been conducted on an ESP model with an electrode
assembly consisting of coaxial cylinders and point-to-plane electrodes. The
temperature of a dry process and its pressure varied in the range of 20°-400°
C and 0.1-0.6 MPa, respectively. Cylinders with diameters of 65, 99, and 140
mm were used as collecting electrodes. The length of the active part of these
*For further information regarding the material in this paper, please
contact I. K. Reshidov, U.S.S.R. 113105, Moscow, M-105, 1-st Nagatinsky
Pass., 6, NIIOGAZ
tGas volume is henceforth assumed to be under normal conditions.
140
-------
cylinders that served as the collecting electrode was 500 mm with a plan of
180-mm diameter, and had edges made in accordance with a profile of the
Pogovsky electrode (for w = Jj at the discharge gap) of S = 4 cm. End shields
of the cylindrical electrodes were separated from the active part by pyroceramic
rings. Wires having diameters of 0.5, 1, 2, 3, 7, and 20 mm were used as
corona electrodes. In addition, edges were made in the form of a hyperboloid
of rotation with a radius of curvature at the tip p = 0.07 and 0.9 mm. Corona
electrodes were also made in the form of rods of 6- and 18-mm diameter and had
semi spherical ends.
Rectified voltage of positive and negative polarities was fed to the
corona electrodes from a high voltage supply, which provided voltage of up to
about 250 kV at current values of 50 mA.
The corona onset voltage (V ) was determined by smoothly increasing the
voltage in the discharge gap by means of stabilizing the voltage at a value
corresponding to an average value of corona current on the level of 1 mkA, or
it was detected by appearance of current pulses on an oscillograph screen.
The input signal was fed from a resistance connected to the grounded
circuit of the measuring part of the collection electrodes. The voltage was
decreased and the voltage value, when pulses and corona were disappearing, was
measured. The minimum value was assumed to be equal to V
Breakdown voltages (Vur) were determined by occurrence of a current spike
in the grounded circuit of the measuring part of the collection electrodes, as
well as by voltage decrease in the discharge gap when the voltage at the exit
of a high-voltage supply was increasing at the rate of about 2kV/s. For each
temperature and pressure range, 25-30 measurements of V. values were made.
Ion mobility, k, and critical voltage, V , were estimated through
C* i
treatment of voltage-current characteristics by the reduction method. The
characteristics were measured for coaxial cylindrical electrodes. The method
is based on the following expression (ref. 3):
-& /?* ft,*/*
.
where C = corona current per unit length of a discharge electrode;
permittivity of free space;
radius of a collecting electrode;
radius of a discharge electrode;
applied voltage;
, '.
K
Ion mobility is described by equation (2) in accordance with the two
cylinder method (ref. 4):
K_ L
* re* fa -14)* (2)
141
-------
where Tj-l/J = difference in potential applied to two cylinders under the
developed corona. R2 and Rx were assumed to be radii of the cylinders at the
similar corona current, •y = radius of the corona electrode.
<"0
THE POSITIVE CORONA
There is much evidence in the literature that operation under positive
corona at higher temperatures is more effective than operation under negative
corona mainly due to the increased puncture voltage of the former (ref. 5).
Therefore, by investigating corona characteristics, one could demonstrate the
possible applicability of positive corona in ESPs.
At temperatures of 20 to 400° C, and at pressure of 0.1 to 0.6 MPa, two
forms of positive corona may occur: continuous (avalanche type) and pulse
(streamer type). The pulse corona is characterized by the absence of current
pulses and it has an image of a uniform luminescent sheath at the surface of
the anode. A slight glow, created by numerous streamers, spreads towards the
cathode. Pulses of corona current are very strong. At normal (atmospheric)
conditions, the continuous corona exists at polished tips and at fine wires
with d = 2 mm diameters as well. An increased air density (6) results in
conversion of a continuous form into a pulse one (figure 1). At the increased
temperatures, the transition of a continuous corona to a pulse one occurs at a
higher density than that at a normal temperature. Pulse corona generation is
accompanied by a sharp decrease of breakdown voltage (by a factor of 2).
Further increase of air density to a critical value (6kp) leads to disappear-
ance of pulse corona discharge and gap breakdown starts without preceding
corona discharge. 6kp values increase with decreasing radius of the discharge
electrode.
Ion mobility was evaluated by reduction (Ki) and two-cylinder methods
(K2). For the temperatures and pressures studied, it was found that the
product of air density and ion mobility was approximately a constant value
(<$!
-------
120
100
80
60
u>
40
20
^
^
*f-
^^
/
7,
/
#&
i $x#'
fx-X-T ..
0*0
lo
XD'
"\/5
/>
fj-
<>*
a
^
s
Figure 1. Breakdown and initial voltages at positive corona in the gap of
the "point-to-plant" electrode-assembly ( p = 0.9 mm, S = 4 cm).
a - breakdown voltage (I - continuous corona, II - pulse corona),
b - initial voltage;
1 - 20° C, 2 - 100° C, 3 - 200° C, 4 - 300° C, 5 - 400° C.
-------
120
100
80
60
40
20
1 Q.
1,4
1,2
1,0
0,8
0,6
0.4
20
100
200
300
400
T,°C
Figure 2. Breakdown voltages in the "coaxial cylinders" electrode assembly
at P = 0.14 mPa, D = 140 mm.
d = 0.5 mm.
a - positive; b - negative polarity.
144
-------
Meanwhile, it has been found that a stable, continuous corona exists only
at fine polished wires with a 2-mm or smaller diameter or at edges with a
fixed radius of curvature. Mechanical damage on the surface of a discharge
electrode (scratches, scores) or erosion due to breakdown of electrodes leads
to creation of pulse corona discharge and a sharp decrease in breakdown
voltage.
Available data allow one to conclude that application of positive corona
in industrial ESPs, particularly for blast-furnace gas control at P = 0.35 and
T = 250°, is not reasonable.
THE NEGATIVE CORONA
For an electrode assembly consisting of coaxial cylinders, the value of
averaged ion mobility (K^ within the whole gap was estimated by the method of
reduced characteristics. The K2 value, corresponding to the region of the
discharge gap between R! and R2, was estimated by the two-cylinder method.
For the range of P and T studied, it was established that the product of
air density and ion mobility 6Ki increases substantially with rising temperature.
So, at T = 400° C and P = 0.14 MPa for D = 99 mm and d = 3 mm, the value of
6K], is equal to 6.55 cm2/V/s, which is 2.5 times more than 6Ki under positive
corona.
The values of K2 determined by the two-cylinder method, for Rj = 50 mm
and R2 = 70 mm are lower at high temperatures than the corresponding values of
KI. So, at T = 400° C and P = 0.14 MPa, we obtained 6K: = 5.4 for d = 0.5 mm
as compared with
-------
Figure 3 shows V. as a function of varying air density,
-------
140
120
100
80
60
40
20
c
D <
*x
*
*
^
0
0?
•
> mw +
>
o
a
. *
•
D '
S7
;•
;
01
D2
tf 3
• 1'
B2'
V3'
f 1"
X2"
03"
0
a.
M
r
0,5
1.0
1.6
2,0
2,5
3,0
Figure 3. Breakdown voltage at negative corona in the "point-to-point"
electrode assembly for edges with varying radius of curvature.
1,2,3 - p = 0.077 mm; V, 2', 3' - 0.9 mm, 1"
2", 3" - 3 mm; I, i; I" - T = 20° C; 2,2'>2" - 200° C
and 3,3-,3" - 400° C.
147
-------
The rapping systems operated intermittently at intervals of 18 min.
Duration of a cleaning cycle was 6 min. To evaluate the reentrainment,
factor measurements of dust content were made during the operation and
between cleaning cycles. At 0.5 - 1.2 m/s gas velocity through the acti 3
zone of the ESP, the rapping reentrainment factor given by: K . =
log (l-n5ezCfciT!p was in the range of °'93 to °'96> sll'9ntly exceeding the
value of reentrainment obtained for industrial ESPs. This behavior may be
associated with improved electrical characteristics of the test ESP and
with lower gas velocity as well.
At 1.2 m/s (residence time of gas in the ESP active zone was t = 2.1
s), dust concentration at the exit did not exceed 60 mg/Nm3 when shakers
were in operation. At 0.5 m/s (t = 5 s), the value decreased to 13.1
mg/Nm3. Exit dust concentration averaged 9.2 mg/Nm3 at 99.8 percent
collection efficiency regardless of rapping reentrainment.
On the basis of the available data, a preliminary design of a dry
plate ESP for high-efficiency gas cleaning had been developed. The instal-
lation is designed to treat 200,000 Nm3/hr of blast-furnace gases at 0.35
MPa and 250° C.
CONCLUSIONS
The main results of the investigation may be summarized as follows:
1. Operation of ESPs both at higher pressures and temperatures
is not reasonable under positive corona.
2. Cooling blast-furnace gases to 120-150° C in a hollow scrub-
ber, or also in a scrubber with converging gas feed ahead of
the dry ESP is possible. Operating difficulties have not
been observed.
3. Application of strip-tooth discharge electrode assemblies to
treat blast-furnace gases at both increased temperatures and
pressures is recommended.
4. High collection efficiency of a dry plate ESP has been
demonstrated. At 0.3 MPa and 230° C, the pilot-plant ESP
has shown 99.8 percent efficiency when residence time was 5
seconds.
REFERENCES
1. A. G. Ljapin and V. V. Danilin, "Application of the Unipolar Positive
Corona for Electrical Gas Cleaning," Izvestija of the Academy of
Sciences, U.S.S.R., Energetics and Transport, No. 6, 1975. pp. 155-
160.
2. A. G. Ljapin and V. V. Danilin, "Negative Unipolar Corona in Air at
Increased Temperatures and Pressures," Izvestija of the Acadeny of
Sciences, U.S.S.R., Energetics and Transport, No. 5, 1976. pp. 148-
154.
148
-------
3. V. I. Popkov, "To the Theory of DC Current Unipolar Corona," Electricity.
No. 1. 1949. pp. 33-42.
4. V. I. Levitov, "A. C. Current Corona, M," Energy, 1969.
5. C. C. Shale, "Progress in High-Temperature Electrostatic Precipitation,"
Contr. Assoc., Vol. 17. No. 3 (1967), pp. 159-160.
6. N. B. Bogdanova and B. I. Popkov, "The Form of Corona Discharge and
Breakdown of Air Gaps," Electricity. No. 8, 1973, pp. 27-33.
149
-------
AIR POLLUTION CONTROL OF EMISSIONS FROM
KILNS AND COOLERS UTILIZING
FABRIC FILTERS IN THE CEMENT INDUSTRY
Frank R. Culhane*
Abstract
This paper covers the application of fabric filters within the cement
industry to control particulate emissions from the cement kilns and clinker
coolers.
The subject matter is presented in tao parts: Part I - Cement Kilns, and
Part II - Clinker Coolers.
PART I - CEMENT KILNS
Particulate emission from rotary cement kilns has long been a difficult
problem, but has increased in severity as our air pollution codes have become
more stringent with the passing of time. From a gas cleaning standpoint, the
difficulties have been due to the large volume of gas involved, heavy dust
loading, fine particle size, high dust resistivity on dry-kiln applications,
and high moisture on wet-kiln applications.
The standard multicyclone collector is capable of removing a high percent-
age of dust by weight, but for nearly complete collection of the particulate,
the only practical device, until recently, has been the electrostatic precipi-
tator and, to a much lesser extent, the wet scrubber. The fabric filter,
which is one of the oldest and simplest collectors capable of a high degree of
stack gas cleaning, for a long while was considered unsuited for cleaning kiln
flue gases because of the temperature limitations of the filter fabric and
generally accepted, but poorly understood, fear of moisture. However, in the
1950's, development of silicone-finished fiberglass, and later graphite- and
Teflon-coated fiberglass, raised the operating temperature limits to 288° C
(550° F) and provided a fiber with extremely low water absorbancy. As a
result, the cement industry in the United States tested a fabric filter uti-
lizing fiberglass filter bags on wet-process kiln gas in 1956, and in October,
1957, the first full-size kiln baghouse went into operation. Today, some 20
years later, there are over 385 cement kilns in the United States, of which it
is estimated that over 33 percent are equipped with baghouses, testifying to
the acceptance of the fabric filter in this industry.
Portland Cement Process
In order to have an appreciation of the variables involved in the proper
application of fabric filters for the cement industry, we should first briefly
review the processes involved in the manufacture of Portland cement. Stated
*Wheelabrator-Frye Inc., Air Pollution Control Division, Pittsburgh,
Pennsylvania 15219.
150
-------
briefly, the raw materials utilized in the manufacture of Portland cement con-
sist of argillaceous* and calcareous! materials that are quarried, crushed,
ground, and thoroughly mixed in either a wet or a dry state. The mix is fed
as a slurry into a rotary kiln in the wet-process plant, and as a dry powder
in the dry-process plant. The kilns in the United States vary between 2.5 m
(8.2 ft) and 5.5 m (18 ft) in diameter, and are from 30 m (98.4 ft) to 230 m
(754.6 ft) in length. The kilns are installed horizontally, refractory lined,
rotate slowly between h and 1 rpm, and are set at a slight incline in the range
of 3 cm/m. The raw materials are fed into the upper end of the kiln and
travel downward toward the firing end of the kiln. The kilns are either fired
with powdered coal, fuel oil, or natural gas, with the burner located at the
lower end of the kiln. At the firing end, the kiln gas temperature ranges
from 760° C to 871° C (1,400 to 1,600° F).
As the feed moves along the upper part of the kiln, it is exposed to
increased gas temperatures as it flows slowly toward the lower end. As the
raw material is heated, the water is evaporated, the organic material burns
away, and the carbonates lose their carbon dioxide and the sulfate is decom-
posed, liberating S03. In addition, chlorides and alkali salts are partially
volatized. About a third of the original weight of the raw mix is lost in the
burning process due to volatilization.
Chemical reactions take place between the materials in the raw mixture at
the elevated temperatures within the kiln. In the course of these reactions,
new compounds are formed and some melt to partially fuse the charge into
clinker. The resultant clinker is then discharged into a cooler. When cool,
the clinker is mixed with a carefully controlled quantity of either calcined
or uncalcined gypsum and the mixture is ground to a very fine powder as Port-
land cement.
Wet Kiln
It is estimated that within the cement industry in the United States,
there are some 385 cement kilns in operation today. Of these 385 kilns, a
total of 214 are wet-process kilns and 171 are dry-process kilns, which means
that 56 percent of the kilns in the United States are of the wet process type.
As stated previously, the first fabric filter installed in the United States
was installed on a wet-process kiln. Generally speaking, the gas conditions
from a wet kiln are readily suitable for electrostatic precipitation without
gas conditioning. Therefore, the electrostatic precipitator has been widely
accepted in the United States within the cement industry as a suitable and
economic solution to kiln emission control. However, wherever the plant
operations are: (1) concerned with "snowfTaking" from electrostatic precipi-
tators, (2) in localities wherein high availability has been a demand of local
air pollution control authorities, or (3) where plant operations have had a pref-
erence for fabric filters because of their inherent simplicity and consistent
*Typical sources of argillaceous materials are: clay, shale, slate, slag,
silica, ashes, sand, and iron ore.
tTypical sources of calcareous materials are: limestone, sea shells,
chalk, and marble.
151
-------
high efficiency, then fabric filters have been installed on kiln exhaust gases
with a high degree of success.
The application of fabric filters in wet kilns does necessitate operating
the baghouse dry. Therefore, special design features must be incorporated
into the equipment in order to keep the recovered dust in a dry, free-flowing
state and to minimize corrosion and wetting of the filter bags.
As a recognition to sound application engineering and good design, the
following installation is cited as typical of many successful installations
of fabric filters effectively handling wet kiln emission in the cement in-
dustry.
Figure 1 shows a typical wet kiln installation utilizing a fabric filter
to control kiln emissions. The kiln is 3.05 m (10 ft) in diameter and 91.4 m
(300 ft) long, and produces 324 tons (1,800 bbl) per day of clinker. A 14.6 m
(48 ft) section of the feed end of the kiln contains chains that serve to pre-
heat the feed. The slurry feed contains 35 to 40 percent water by weight. The
kiln is gas-fired.
In 1965, a 10-compartment Wheelabrator baghouse was installed between
the existing multicyclone and the existing kiln fan on the kiln described
above. Operating details are as listed below:
Data
Gas Volume
Gas Temperature
Gas Dew Point
Dust Load
Dust Size
Pressure Drop
Efficiency
Metric Units
70,688 rnVhr
254° C
73° C
1.37-2.29 g/m3
50-10% urn
152.4 mm
99.9%
English Units
80,000 ACFM
490° F
163° F
0.6-1.0 gr/ft3
50-10% urn
6" WG
99.9%
The baghouse is operating extremely well and is performing beyond expec-
tations of the customer. The unit has been in service since the spring of 1965,
and maintenance has been low. Bag life has been good, availability of the
equipment has been excellent, and the efficiency has been consistently high.
The stack discharge has been optically clear, enabling the plant to discharge
the kiln gases 20 m (65.6 ft) above grade, eliminating the need and expense of
a tall stack.
Dry Kiln
Of the 385 kilns in the United States, 171 are operating on dry process
systems that represent 44 percent of the total kiln number. Even though the
number is not as great as the number of wet-process kilns, the dry kilns re-
main a dominant factor in the American cement industry production because on
a kiln-for-kiln basis, they are by far larger.
In addition, from the standpoint of air pollution control, a dual problem
of higher dust loadings and higher resistivity necessitates the employment of
large water conditioning towers ahead of the electrostatic precipitators when-
152
-------
OJ
Figure 1. Baghouse shown at left controlling emissions from a wet kiln.
-------
ever high-efficiency electrostatic precipitators are specified for emission
control. Of the many baghouse installations in the cement industry on dry
kilns, a typical installation is represented in the photograph listed herein
as figure 2.
This kiln has a nominal diameter of 4.9 and 5.3 m (16 and 17.5 ft), is
179.8 m (590 ft) long, and produces 1,620 tons (9,000 bbl) per day of clinker.
The kiln is coal-fired and the exit gases are cooled by water sprays in the
final section of the kiln.
In 1964, a 16-compartment Wheelabrator baghouse was installed immediately
after the kiln and before the kiln fan. This approach was unique as it re-
presented the first installation of a baghouse on kiln exit gases without a
mechanical collector to serve as a precleaner for the baghouse. Operating
details are as listed below:
Data Metric Units English Units
Gas Volume 249,788 mVhr 300,000 ACFM
Gas Temperature 288° C 550° F
Gas Dew Point 53° C 127° F
Dust Load 25.15 g/m3 11 gr/ft3
Dust Size 40-10% urn 40-10% urn
Pressure Drop 127 mm 5"
Cloth Baghouse Pressure Drop 222 mm 8.75"
Efficiency 99.9% 99.9%
This baghouse is operating extremely well and the unit has been in service
since 1965. The stack discharge meets the EPA requirements of less than 0.15
kg of particulate matter per metric ton (dry basis) of feed (0.3 Ib of partic-
ulate matter per ton of kiln feed). Average bag life has been 2 years.
Recently, the operating temperature has been reduced to 274° C (525° F),
resulting in bag life of 2-1/2 years. Previous operating temperature of 288° C
(550° F) resulted in bag life of 1-1/2 years.
The usual rotary kiln is notoriously wasteful in fuel consumption, requir-
ing upwards of 1,500 kcal/kg to 2,250 kcal/kg of clinker (1,000,000 Btu to
1,500,000 Btu per barrel of clinker), which has lead to many improvements in
the conservation of heat. Heat transferred from the cooling of clinker is
recovered as hot secondary combustion air, and the heat leaving the kiln can
be used to preheat the incoming raw kiln feed.
Air Suspension Preheater
The air suspension preheater method of preheating the dry kiln feed with
the kiln gases was developed in Europe and is known throughout the cement
industry today as the Humboldt preheater kiln. The preheating system consists
of a number of cyclones through which the kiln exit gases pass before reaching
the final gas cleaning equipment. The dry raw feed is introduced into the top
chamber and falls by gravity through four cyclones in series. Each cyclone is
swept by hot ascending kiln gases countercurrent to the descending raw feed.
The kiln gases leave the kiln at about 950° C (1,750° F), and, by direct
transfer, the kiln feed is heated to 750° C (1,380° F) during its contact in
the suspension kiln gas.
154
-------
Ul
01
Figure 2. Baghouse shown at right controlling emissions from a dry kiln.
-------
Preheater Kilns
A preheater kiln is a dry process kiln and a typical fabric filter instal-
lation handling the emissions from this type of dry kiln is shown in the
photograph identified herein as figure 3.
This kiln produces 850 to 950 metric tons of clinker per day (4,722-5,000
bbl/day) from a raw dry feed of 2,100 to 2,200 tons/day. The kiln is 4 m in
diameter and 56 m long, and is fired with Bunker oil containing 2.75 percent
by weight. Flue gases exit the air suspension preheater and pass through the
raw mill before entering the bag filter. There are times, however, when the
kiln gases go directly to the bag filter whenever the raw mill is down for
maintenance or because the raw mix silos are full.
In 1965, an 8-compartment Wheelabrator baghouse was installed following
the final cyclone heater exchange and the raw mill. The process is shown in
the flow chart shown in figure 4. The data on the process are as follows:
Data Metric Units English Units
Volume of gases passing through
raw mill to bag filter 230,000 mVhr 136,000 ACFM
Volume of gas direct to bag
filter 190,000 mVhr 112,000 ACFM
Temperature after heat exchange 330° C 626° F
Temperature after raw mill 200° C 392° F
Moisture by volume 3-6% 3-6%
Dewpoint at B/H inlet with
gases passing through raw mill 43° C 109° F
Dewpoint at B/H inlet with
gases bypassing raw mill 30° C 86° F
Dust load after raw mill 60 g/m3 26 gr/ft3
Size % less than 50 nm 85-95 85-95
Dust recovery 15 tons/hr
Note: Automatic bypass for temperature surges above 260° C (500° F).
The operation of the baghouse serving this kiln has been outstanding,
with the initial set of filter bags lasting 5 years, which has been generally
beyond the expectation of both the manufacturer and the operators. Over the
past 12 to 13 years, the availability of the fabric filter has been in excess
of 99 percent, consistently producing filtration efficiencies of 99.9 percent
by weight.
Kiln Emission Code
In considering a fabric filter to control emissions from cement kilns, it
is necessary to include consideration of the air pollution code requirements
in order to make an equitable economic decision in relationship to other means
of control. The EPA Code states that no owner or operator shall cause to be
discharged into the atmosphere from any kiln, any gases that:
156
-------
Figure 3. Baghouse controlling emissions from air preheater kiln.
Flow diagram for this system is shown in Figure 4.
157
-------
LIMESTONE
CRUSHING
PLANT
RAW
MATERIAL
STORE
PACKING
PLANT
CEMENT
SILO
PLANT
CLINKER
STORE
CEMENT
GRINDING
PLANT
HOMOGENIZING
PLANT
RAW MIX
SILOS
|HOTAHYKILN
WHEELABRATOR FILTER
SCREW COHVEYOR J
I
I
I
t^°* ^
j§SS-»
I
PNEUMATIC PUMP
COOLER BLOWERS
^sro|>—/
HOMOGENIZING
PNEUMATIC WMP
HOT GAS
GENERATOR
Figure 4. Flow sheet for air suspension air preheater kiln.
-------
1. contain particulate matter in excess of 0.15 kg/metric ton of feed
(dry basis) to the kiln (0.30 Ib/ton), and/or
2. exhibit greater than 20 percent opacity.
Examination of the allowable limits on kiln mass rate of emissions equates
to no more than 0.015 percent of the dry kiln feed that can be discharged into
the atmosphere. By translating the maximum allowable emission rate to percent
of kiln feed on a dry basis, you can readily make a comparison to the emission
standards of other countries.
SUMMARY
Comparative analysis of allowable mass emissions from cement kiln gases
in the United States to the governmental standards of other industrialized
countries demonstrates the high standards and stringency of the US/EPA Code
requirements. A code that restricts emissions from cement kilns to no more
than 0.015 percent of the kiln feed rate is a stringent code. It is this
stringency of the US/EPA requirements that explains the high percentage of
fabric filters employed on cement kilns in the United States, as compared with
the practice of the cement industry in other countries. The installations
described herein demonstrate that the fabric filter properly designed, applied,
installed, operated, and maintained can successfully handle the variables of
this application and meet the US/EPA Code requirements within a reasonable
factor of safety. When comparing the economics of a fabric filter versus an
electrostatic precipitator, the selection must be decided on a case-by-case
basis because of the number and range of the variables involved.
159
-------
PART II-CLINKER COOLERS
Some 40 years ago, the cement industry developed the clinker cooler to
improve the quality of the Portland cement process. By quenching the clinker
with cooling air as the cooler leaves the kiln, the desired improved clinker
is obtained. The heated quench air, in terms of quantity, represents some
three times as much air as is needed for kiln combustion. This allows 33 per-
cent of the heated quench air to be directed to the kiln as hot air for second-
ary combustion, and the remaining 66 percent of the quench air to be discharged
to atmosphere. As the clinker is air quenched, the quenching air entrains
particulate matter in the form of clinker dust, which uncontrolled would
create an air pollution control problem.
A typical clinker cooler of the grate type is shown in figure 5. The
cooler is designed to air quench 1,420 kg/min (3,130 Ib/min) of clinker at
1,370° C to a temperature of 65° C (2,500° F to 150° F). To do this, 93.2 m3/s
at 32° C (197,420 ACFM at 90° F) is blown through a 38-cm thick (15 in) moving
bed of clinker utilizing seven forced-draft fans. As stated above, a portion
of the higher temperature quench air serves as preheated secondary combustion
air. In the example shown in figure 5, 84.5 m3/s at 760° C (179,000 ACFM at
1,405° F) of quench air is diverted to the kiln. A smaller quantity of quench
air is diverted to the coal mill which, in this case, represents 11.3 m3/s at
420° C (24,800 ACFM at 790° F). The balance of the quench air is vented to
atmosphere, which totals 85 m3/s at 154° C (180,000 ACFM at 310° F). This
cooler handles a kiln having a daily clinker production of just over 2,000
metric tons. (12,350 bbl/day).
Cooler Upset Conditions
Within the burning zone of the kiln, a coating of kiln feed builds up on
the surface of the refractory and serves as a dam for the feed lying upstream
of the burning zone. Under normal operations, the feed spills over the "dam"
and quickly forms into moderately sized clinker in the burning zone. In ab-
normal operations, the "dam" becomes excessive and, as this ring breaks, or is
broken, excessive quantities flow through the burning zone uncalcined, and
enter the grate cooler as fines. This condition is called a kiln "upset."
Figure 6 shows the gas flows and temperature conditions for the cooler
under "upset" operating conditions. During kiln upset, the following condi-
tions exist:
1. The discharge rate increases 50 percent from 1,420 kg/min to 2,136
kg/min (3,130 Ib/min to 4,700 Ib/min).
2. The clinker temperature from the kiln drops from 1,370° C to 1,200° C
(2,500° F to 1,204° F).
3. The automatic controls maintain the same flow through the cooler
grates, 93.2 ms/s (197,420 ACFM).
4. The same weight of secondary combustion air is required and main-
tained for the kiln.
160
-------
o>
SECONDARY AIR
CLINKER -^ 84.5 m3/ SEC. 760° C
1420kg/min 1370°C
AIR TO COAL MILL
11.3M3/SEC. 420°C
QUENCH AIR FAN
5.1 M 3/ SEC. 32° C
SIX FANS
88.1 M3/SEC. 32° C
AIR TO VENT
85.0 M3/SEC.-154° C
9.5 g/m3 or 48.6 kg/min. OF DUST
CLINKER COOLED TO
65°C
Figure 5. Clinker cooler under design conditions.
-------
CLINKER
2136 Kg/min. 1200°C
SECONDARY AIR
57.8 M3/SEC. 437° C
AIR TO VENT
139.3 M3/SEC.-370° C
55.5 g/M3 or 463 Kg/min. OF DUST
AIR TO COAL MILL
11.7M3/SEC. 421°C
QUENCH AIR FAN
5.1 M3/SEC. 32 °C
SIX FANS
88.1 M3/SEC. 32° C
CLINKER COOLED TO
93°C
Figure 6. Clinker cooler during kiln upset conditions.
-------
5. The same weight of air to the coal mill is required and is main-
tained.
6. The clinker dust emitted from the cooler is increased from 9.5 g/m3
to 55.5 g/m3 (4 gr per ACF to 25 gr per ACF).
7. The quench air out of cooler vent increases from 154° C to 370° C
(310° F to 700° F), and
8. The clinker is cooled to only 93° C (200° F).
This upset condition results in a design requirement for the dust collec-
tor capacity of 139.3 m3/s at 370° C (295,000 ACFM at 700° F), a 50 percent
increase in gas volume and a tenfold increase in the dust loading.
In summary, based on the above, a clinker cooler pollution control system
must handle the following range of conditions:
Normal Upset
Air volume 100% 150%
Dust Load 9.15-13.7 g/m3 34.4-57.2 g/m3
(4.6 gr/ft3) (15-25 gr/ft3)
Temperature 204° C (400° F) 343° C - 538° C
(650° F - 1,000° F)
Thermal Control
The application of a fabric filter makes thermal control of the heated
quench air from the cooler vent necessary to prevent damage of the filter
fabric during periods of gas temperature surges above design conditions. The
design engineers can control these temperature surges by utilizing either:
(1) evaporative water cooling, (2) tempering air, (3) blending or mixing of
air from adjoining clinker coolers, (4) radiation cooling, or (5) a combina-
tion of these four methods. The temperature limitation of the filter fabric
is 288° C (550° F) for fiberglass and 218° C (425° F) for Nomex.
Fabric Filters Using Fiberglass
Fabric filters utilizing si 1 iconized fiberglass cloth cleaned by repres-
suring have proven to be successful in this application. Operating experience
has shown that the clinker dust releases very readily from the cloth. The
modular design has been used widely with some installations utilizing a single
fan on each module, as shown in figure 7. The use of fiberglass minimizes air
cooling requirements and, accordingly, is preferred by many of the plant operat-
ing personnel.
Fabric Filters Using Nomex
An alternate filter fabric to fiberglass is Nomex, which will handle
temperatures up to 218° C (425° F) without damaging the fiber. Nomex is
163
-------
Figure 7. Modular Dustube fabric filter handling emission
discharge from clinker coolers.
-------
commonly utilized on fabric filters utilizing a "burst" or "explosion" of
compressed air to flex the bag for cloth cleaning purposes. The application
of the Ultra-Jet or Pulse-Jet design on clinker cooler emission has been quite
common in the United States, with the initial installations dating back to
1965. A typical installation is shown in the photograph identified as figure
8. Because clinker dust is relatively coarse, and releases quite readily from
the filter bags, sizing the unit has presented few difficulties. Again, the
modular design has been preferred, enabling the plant operating personnel to
isolate a single unit for maintenance without affecting continuity of the
plant operation.
Inasmuch as the dust is abrasive, special care has been taken in the
design of the ductwork and inlet to the units to avoid erosions of the steel
and filter fabric. Also, the dust-handling equipment is of special heavy duty
design to avoid erosion. Fans should, at all times, be located on the clean
side of the filter to avoid erosion of the fan blades and scroll.
In order to meet this requirement, the cement industry in the United
States has successfully applied fabric filter to clinker coolers after full
evaluation of the economics involved. When properly sized, designed, applied,
operated, and maintained, the tubular fabric filter has consistently met the
Code requirements and has operated with a high degree of availability requir-
ing minimal and easy maintenance.
Cooler Emission Code
The EPA Code states that no cement plant owner or operator shall cause to
be discharged into the atmosphere from any clinker cooler any gases which:
a) contain particulate matter in excess of 0.050 kg/metric ton of feed
(dry basis) to the kiln (0.10 Ib/ton), and/or
b) exhibit 10 percent opacity, or greater.
The US/EPA Code for kiln emissions, as stated above, represents a strin-
gent requirement for the cement plant operating personnel. However, the
US/EPA Code for clinker cooler emissions is more stringent than for the kilns.
The kiln allowable emission is up to 0.015 percent of the kiln feed. The
cooler emissions are restricted to 0.005 percent of the kiln feed. When the
cooler allowable figure is equated to mass rate emissions from the vent stack,
the outlet particulate loading oftentimes cannot exceed 10 to 11 mg/m3. This
low outlet residual dust loading requires collection efficiencies of 99.9
percent by weight, which has been successfully achieved by the employment of
fabric filters in the United States cement industry, as described within this
paper.
SUMMARY
It is essential that the particulate emission control equipment be designed
to meet EPA Code requirements during normal and upset operating conditions of
the kiln. To do this, the thermal control and particulate control equipment
must be designed to handle 50 percent increase in volume, 400 percent increase
165
-------
Figure 8. Ultra-Jet modular fabric filter handling
emission from clinker coolers.
-------
in dust load, and temperature surges up to 538° C from a 200° C base. The
code requirements equate to outlet particulate loadings below a range from 50
to 25 mg/m3. This low outlet residual dust loading requires collection effi-
ciencies greater than 99.9 percent by weight. The fabric filter has been
successfully and widely applied in the United States and Canada on clinker
coolers to meet the total conditions of the cooler air pollution control
problem and obtain the desired solution. The typical installations described
herein demonstrate the viability of the fabric filter to handle the stringency
of the U.S. code and the variable operating conditions created by the quench
cooling process.
167
-------
REMOVAL OF HARMFUL PARTICLES FROM THE EXHAUST
OF THE CEMENT PLANTS*
G. I. Vodolazskiy
Abstract
At present more than 3,500 pollution control devices are in use throughout
the cement industry. Of these, 1,000 are of the electrical type, approximately
1,000 are bag filters, and more than 1,500 are other types of dust collecting
units (cyclones, scrubbers, etc.). Design solutions are examined in this re-
port for new process lines for cement plants. Problems associated with rebuild-
ing existing gas cleaning units are also considered. Data are given on the
operating conditions of units in use and their cleaning efficiency in removing
dust during raw materials processing, clinker calcining and milling, etc.
INTRODUCTION
Today more than 3,500 dust collectors are in service in the cement indus-
try. This dust collecting equipment is represented by 1,000 electrostatic
precipitators, about 1,000 bag filters, and by more than 1,500 devices of
other types (cyclones, scrubbers, etc.). Every year the number of dust col-
lectors in the cement industry increases by 80-100 units.
Owing to the great deal of development work accomplished in the last
years, the value of harmful pollutants to the atmosphere has been reduced. As
it is known, dust emissions worsen the sanitary conditions at the plants and
in the adjacent areas. They also hasten service wear of the installations,
making working conditions worse, and causing occupational diseases adversely
affecting labor productivity.
DUST ABATEMENT AT THE RAW MATERIAL PROCESSING AREAS
The great influence on the dust collecting efficiency at the raw material
processing areas has the rational technological developments directed to low-
ering the heights of material overfalls from the conveyors, reducing the sec-
tions of the hoppers and the angles of their slopes, and approaching the
natural slope of the material in movement, arrangement of baffle plates, and
rock cushions in the chutes. In connection with this, the speed of the belt
conveyors must not exceed 0.7-1.0 m/s. Nonloaded belts of the conveyors
should also be equipped with dust collecting devices.
The obligatory condition of dust collection at the raw material process-
ing areas is the installation of airtight enclosures, their design and sizes,
influencing greatly the dust collection efficiency. For the purpose of the
dust abatement at all loading points, especially when handling loose material
from the upper conveyor to the lower one, it is provided to install an
enclosure with air exhaust, the lower enclosure being supplied with double
*For further information regarding the material in this paper, please
contact I. K. Reshidov, U.S.S.R. 113105, Moscow, M-105, 1-st Nagatinsky Pass.
6, NIIOGAZ.
168
-------
walls. Such enclosures are also installed at the discharge openings of Jaw,
cone, and hammer crushers. The Institute of NIPIOTSTROM has designed such
enclosures with double walls for conveyor transfer points of material and
discharge openings of hammer crushers. The similar enclosures are designed by
the Institute of NIIRUDVENTILYATSIYA.
Tests conducted at several enclosures with double walls in the crushing
departments of the cement plant of "Pzoletariy" in Novorossisk showed^that the
volume of the aspiring air to be exhausted had been reduced by 1.5-2 times,
with the dust content in it being lowered at 30-35 percent, compared with the
single-walled enclosures.
According to the analyses, it was stated that about 150-200 ms of aspir-
ing air should be exhausted from all the enclosures of the crushing department
when processing 1 ton of material.
The dust content of the aspiring air depends not only on physical and
chemical properties of the material to be treated, but especially on its
humidity. For example, the department for dolomite crushing of the Vitebsk
group of enterprises of limestone flour (capacity 3 million t/yr) does not
require exhausting because of high moisture content in the raw material,
amounting to 8-10 percent.
One of the most important means of dust suppression is wetting the mate-
rial. Dust content in the aspirTng air can be reduced several times by install-
ing sprayers in the enclosures of the places generating dust and by the opti-
mal wetting of the rock. As a result of this, at some plants in southern
regions, it turned out to be possible to employ a simplified method of dust
collection, consisting of the installation of a hydrosuppression system as the
first stage of dust collection, and cyclones as the second stage.
However, the majority of works do not have these highly efficient exhaust
systems for crushing departments. In this case, the increase of collection
efficiency is attained by means of the wet method of dedusting the aspiring
air as well as the dry one.
At the wet process cement plants it is appropriate to utilize the wastes
from wet-type systems of dust control by means of feeding them for preparing
raw mix. This is the reason for giving preference to the wet-type system of
dust control at the crushing departments. Some difficulties and additional
expenses connected with the employment of wet dust collectors in the winter
time are justified by a simplicity of their design and their efficiency.
Hydrodynamic dust collectors, GDP type, with capacities from 20,000,
30,000 and 50,000 nrVhr, designed by NIPIOTSTROM, and apparatuses, PVM type,
designed by NIISANTECHNIKA, are widely employed by a wet method of dust collec-
tion.
For a dry method of dedusting the aspiring air of crushing departments,
gravel bed filters type ZF-4, ZF-5, ZF-6 with capacities of 1,500, 4,000, and
5,000 mVhr, respectively, are used. The specific capital costs by employment
of the gravel bed filters will make up 1.2-1.5 rubles for 1 nrVhr, collecting
efficiency being 98-99 percent.
169
-------
KILN DUST COLLECTING EQUIPMENT
Reduction of rotary kiln dust emissions is the main task in a general
problem of dust collection at the cement plants. All the new present-day
operation lines are designed and put into service equipped with standardized
4-field type electrostatic precipitators, typical design data being 1-1.2 m/s
gas velocity, 8-10 s retention time. The electrostatic precipitators meet all
the requirements of industry and sanitary standards having inlet dust burdens
not amounting the optimal values in the range of 10-15 g/m3.
However, the available electrostatic precipitators 1-, 2-, and 3-fold got
completely overburdened with gas and dust as well.
After analyzing the condition of dust collection equipment for dedusting
exhausts from the kilns for clinker burning, it was revealed that the dust
losses from the kilns were considerably influenced by performance conditions
of these kilns.
Dust content at the outlet of kilns of the same type, processing the same
raw material at the same plant, may vary greatly depending on conditions of
in-kiln preheating devices. Application of the optimal chains allows attain-
ment of dust content at the outlet of the kiln from 2.5-10 g/m3. This is one
of the ways to considerably improve the process of dedusting rotary kiln
exhausts. The second way consists of raising the capacity of dust collecting
systems by increasing dimensions (volumes) of electrostatic precipitators. So
it is known that in order to increase an electrostatic precipitator efficiency
from 95 to 99 percent, it is necessary to raise the specific volume (1 m3 of
an electrostatic precipitator per 1,000 m3 of gas rate per hour) two times, on
the average. In the practice of cement manufacture, such a situation takes
place when burning mechanisms are intensified, their kilns upgraded, and their
capacity raised, causing the amount of gases to be cleaned to be increased.
As a result of this fundamental renovation of available operating electro-
static precipitators, it is obligatory to clean the increased volumes of gas.
The latter factor entails some difficulties, as successful renovation usually
causes complete disassembly of an old unit and installation of a new one.
The most appropriate way to improve the efficiency of the filters, appli-
cable for control of emissions from rotary kilns, can be achieved by keeping
in service the available electrostatic precipitators and installation of
additional filters with multiple or series connections to the operating units.
Design studies show that in most cases, such solutions are quite possible and
expedient. This is one of the ways of improving dust collecting systems for
control of kiln exhausts.
When choosing one of two possible alternative methods of installation of
additional electrostatic precipitators, parallel or series connection, one
should show a preference to series connection. There are good reasons for
such a choice, as the kiln dust contains 40-50 percent of fine particles of
about 5 microns or less in diameter. Therefore, to promote dust coagulation
and efficient collection of this dust, it is necessary to lengthen a particle
path of movement in the electric field. According to the results of tests,
the series connection of additional electrostatic precipitators resulted in
reduction of a residual dust burden at 3-5 times, compared with multiple
connection. So, by series installation, the outlet dust burden was 50 mg/m3
170
-------
at a gas velocity in the electrostatic precipitators in the range of 1.2-1.5
m/s. By multiple connection of the same filters, the velocity reduced to
0.5-0.6 m/s, and the dust burden increased up to 150-200 mg/m3.
One of the alternatives of the installation of additional electrostatic
precipitators has been developed for the cement plant of "Proletariy." The
specific capital costs by such design solution made up 1.5-2.0 rubles/mVhr.
But the same costs totalled 2-2.5 rubles/mVhr by the installation of elec-
trostatic precipitators without using existing filters.
The additional installation of electrostatic precipitators should be
effected parallel with modernization of the existing units, mainly due to gas
distribution devices, replacement of electrode systems, improvement of con-
veyor systems, and more perfect converter sets for energy supply.
During the last 5-6 years in the cement industry, the equipment of step-
up rectifiers substations of electrostatic precipitators has been completely
renovated. More perfect semiconductor sets have been installed at all the
plants. Automation of all these sets has been carried out due to the intro-
duction into production of about 1,500 controllers of "Tsemes," designed by
NIPIOTSTROM. Since 1974, serial production of sets of type ATF has been
started and sets are being designed by some institutes and companies with the
participation of NIPIOTSTROM. Now these devices (capacity of 250-1, 600 ma)
are being produced by Kuybishev Electroworks in Moscow. A distinctive property
of these sets is that they have a thyrister controlling device and a system of
automatic voltage regulation respecting the intensity of sparking in the
electrostatic precipitators.
To ensure that the electrostatic precipitators operate more efficiently,
it is important to properly prepare the aerosol before processing. This
procedure is being introduced at some dry process kilns. According to the
results of tests conducted in May 1974, irreversible dust losses from the
kilns of the Katav-Ivanovsky Cement Plant were reduced five times by a con-
ditioning system developed by GIPROCEMENT and the Institute of NIPIOTSTROM.
Dust control of emissions from the kilns requires the installation of
dust collectors at cooler vents and at the transfer points of clinker hand-
ling. Taking into consideration physical and chemical properties of kilns for
clinker burning (high abrasion, high electric resistance), it is most advis-
able in this case to use filters with filtering media consisting of spherical
grains or grains of some other form. For this reason, some pilot designs of
filters, types ZF-7 and ZF-8, have been developed. There are plans to test
the units as soon as possible.
DUST CONTROL OF MILLING EQUIPMENT
New cement mills and dry mills are designed and put into operation equipped
with a 3-stage system of dust precipitation. The system consists of a so-called
aspiring shaft, cyclones, and bag filters or electrostatic precipitators.
However, most of the existing milling units have no such system.
As a result of this factor, a high inlet dust burden creates unfavorable
conditions for operation of electrostatic precipitators and bag filters. This
-------
leads to an average dust burden at the inlet of the filters applicable for
control of emissions from the mills of up to 50 g/Nm3-
Hence, the intensification of control of dust emissions from milling
equipment should be effected in two ways. In the first place, it is necessary
to equip all the milling plants with a 3-stage system. This measure does not
require high capital costs, and it can be realized during medium repairs and
major overhauls. The great advantages of the application of the above-mentioned
developments are confirmed by the results achieved at the mills of the Dushanbe
Cement Plant, at the "Proletariy" and "Oktyabr" plants, and other plants where
the inlet dust content in the aspiring air before filters was reduced by 2-5
times.
The second stage of improvement of dust collecting systems in milling
plants consists of gradual renovation of obsolete equipment, and especially,
its replacement by modern equipment. It is first of all the installation of
electrostatic precipitators-'-type UG, bag filters, type SMTS, and other de-
vices.
CONTROL OF DUST EMISSIONS FROM THE SECONDARY SOURCES OF AIR POLLUTION
The level of equipment for the secondary sources of air pollution (cement
silos, clinker storages, stations of cement loading to the road, and railway
transport) with dust collectors is still low.
Proceeding from the specific character of the secondary sources of air
pollution, it is necessary to create simple and highly reliable filters.
Cement silos, for example, may be equipped with frameless pressure bag filters
without cleaning devices and dust handling facilities. However, such filters
require high housings on the silos, which raises the cost of construction, and
besides, it cannot be realized at the conditions of the operating works.
The main dust collecting plants of the above-mentioned sources of pollu-
tion are bag filters and cyclones. In order to make easier the maintenance
work of bag filters for dedusting the aspiring air of silos, it has been
proposed for some developments to be made to install the filters at the levels
close to the ground. Application of bag filters, type SMTS, allows location
in the space between modules of silo bins and also creates the possibility of
cutting down the volume of construction work and improving maintenance and
operating conditions.
An indispensable condition of cleaning the dust-laden air at the cement
loading points is introduction of the appropriate units for ensuring mechani-
zation of the operations parallel with large scale measures of dust'control.
Taking into consideration the value of dust losses from drying drums, it is
necessary to install not less than a 2-stage system of gas cleaning, that is,
cyclones and electrostatic precipitators. Dust control at clinker storages is
provided by construction silos. The possibilities of dust collection at
clinker storages provided with grab-bucket cranes practically comes to dedust-
ing the points of material discharge to the storage. For this purpose, special
dust collection chambers are installed. The high efficiency of these units
was confirmed in the process of their operating at a lot of cement plants.
172
-------
Taking into account t^he experience of maintenance of the existing pneu-
matic systems of dust elimination (Urals Aluminum Plant, Serovsky Plant of
Ferroalloys, Serbryakovsky Cement Plant, and other plants), there have been
elaborate designs of central systems for the elimination of secondary dust
from some industrial enterprises. As boosters, vacuum watering pumps, type
RMK, and sometimes ejectors or double fans are provided for. Cyclones, GTS,
and vacuum bag filters are usually employed for collection. Such systems of
dust elimination are successfully applied at Novosibirsk Thermal Station-3 and
Barabinsk and Belovsk Hydroelectric Stations.
MAINTENANCE OF DUST COLLECTION PLANTS
The great importance in ensuring efficient performance of equipment and
raising the engineering level of its maintenance has resulted in the establish-
ment of a special service on dust control.
Controls for efficient work of dust collecting units affect sanitary
laboratories of industrial enterprises of the Ministry of Building Materials
of the U.S.S.R. As a result of this design, standards have been established
for operation of dust collecting plants.
The estimation of the degree of dedusting exhaust gases and dust-laden
air arising in the cement making process is effected by checking whether the
level of gas cleaning meets the requirements of the health standards that
specify the maximum permissible dust concentrations and the content of other
harmful pollutants in the nearby layers of the atmosphere and in the working
places. While designing a dust collecting plant, this factor is determined by
means of special calculations.
Fulfillment of the projected technical measures and strict compliance
with all the laws on environmental protection by collectives of enterprises
will promote a successful solution of the problem of air pollution control.
173
-------
EMISSION STANDARDS AND BACKGROUND DOCUMENT ON CONTROL TECHNOLOGY
IN THE CEMENT INDUSTRY IN THE U.S.A.
Gary McCutchen,* George E. Weant, III, and George Walsh*
Abstract
Historically, the Portland oement -industry has contributed significantly
to local air pollution problems. This paper summarizes the control techniques ,
emission standards, and compliance with the standards for particulate emissions
from Portland cement plants.
New source Performance Standards have been established for kilns and
clinker coolers (viz, 0.15 and 0.05 kg/metric ton of feed to the kiln, respec-
tively). Opacity standards have also been applied to kilns (20 percent),
clinker coolers (10 percent), and other equipment (10 percent). In a landmark
decision, the United States Court of Appeals for the District of Columbia
Circuit has upheld the use of opacity standards.
Control devices used on the kilns and clinker coolers have consisted
mainly of baghouses and electrostatic precipitators . Baghouses are the most
effective. However, at least one scrubber is being successfully used to control
particulates from a clinker cooler.
A 1976 unpublished study by EPA examined compliance data collected from
EPA regional offices and showed that nearly every facility with adequate con-
trol equipment was in compliance with the NSPS's.
INTRODUCTION
Portland cement is produced by sintering calcium carbonate and aluminum
silicates to produce a nodule known as a clinker. The clinker is then pul-
verized and mixed with small amounts of gypsum to form Portland cement. In
1974, 79.5 x 106 short tons of Portland cement, with a value in excess of $2
billion, were produced in the United States, (ref. 1) The largest production
districts were in the States of California, Pennsylvania, and Texas.
PROCESSES
The process for producing cement has undergone little modification over
the past 2 decades. Limestone from the quarry is introduced into the primary
crusher. The secondary crusher further reduces the size. The material then
passes to storage silos from which it is proportionally blended with the other
raw materials.
*United States Environmental Protection Agency, Emissions Standards and
Engineering Division, Durham, North Carolina
tResearch Triangle Institute, Research Triangle Park, North Carolina
174
-------
From this point there are two types of cement processes; the dry and the
wet. In the former, the mixed material is passed through a dry grinding cir-
cuit. A special predryer is used in some cases. In the wet process, the
materials are ground with water to form a slurry.
The dry or wet material is fed to a rotary kiln where the clinker is
formed. The clinker is removed from the kiln and cooled. The cooled clinker
is then mixed with gypsum, sent to the final grinding mill, and then to
storage silos.
Perhaps the most important process modification to be implemented in recent
years involves the use of suspension preheaters for dry-process operations.
(ref. 2) The preheaters utilize a series of gravity-flow cyclones and ducts
that provide direct-contact heat transfer between the falling materials and
the rising, hot kiln gases. Thus, the temperature of the kiln feed is raised
and the required heat-transfer load of the kiln is reduced. The result is that
either a shorter kiln is required for the same throughput, or a larger through-
put is possible with the same, existing kiln. A shorter kiln means both less
energy use and less pollution.
EMISSION POINTS
Emission points within the production process are numerous. The kiln
and the clinker cooler receive the most attention. Others include the grind-
ing mill circuits; storage facilities for raw materials, clinker, and product;
conveyors; transfer points; bagging operations; and bulk loading and unloading
facilities, (ref. 3) Uncontrolled emissions can exceed 150 kg/metric ton of
cement (300 Ib/ton). In addition, emission of NO and SO are estimated at
r^ A
1.3 and 10.9 kg/metric ton, respectively (2.6 and 21.7 Ib/ton, respectively).
(ref. 4)
EMISSION STANDARDS
Because of the highly visible nature of the particulate emissions and the
severe local problems associated with these emissions, the industry was one of
the first to have standards applied. Existing plants are regulated by State
and local standards, while new and modified plants are regulated by Federal
New Source Performance Standards (NSPS).
One of the top producing States is Pennsylvania, where the particulate
emission standard is based on process weight rates. This standard allows
particulate emissions of 0.69 kg/metric ton (1.37 Ib/ton) of cement or 1.10
kg/metric ton (2.20 Ib/ton) of feed to the kiln for an average size kiln of
23.4 metric tons/hr (25.8 tons/hr). This is equivalent to 99.5 percent con-
trol.
On December 23, 1971, NSPS for particulates were promulgated for Portland
cement plants, based on standards proposed in August 1971. (refs. 3,5) These
standards were 0.15 and 0.10 kg/metric ton of feed to the kiln for kilns and
clinker coolers, respectively. The emission test methods to be employed were
also described. In addition, opacity standards of 10 percent were prescribed
for kilns, clinker coolers, and other sources.
175
-------
Table 1. Stages in the Development of New Source Performance Standards
for Portland Cement Plants
Proposed August 1971
Source
Kiln
Clinker Cooler
Other
Emission
(kg/metric ton)
0.15
0.10
Opacity
(%)
10
5*
5*
Promulgated December 1971
Emission
(kg /me trie ton)
0.15
0.05
Opacity
(X)
10
10
10
Revised and Upheld
November 1974
Emission
(kg/metric ton)
0.15
0.15
Opacity
(%)
20
10
10
*Proposed standard reads "Visible emissions shall not be released to the atmosphere." (pg. 28, reference 2).
Under the paragraph on particulate matter from other equipment it says, "For the purposes of this standard,
visible emissions are considered to be any emission of greater than 5 percent opacity "
-------
Q
K
g
-------
Three tests were used as the basis for the kiln standard: two on wet-
process kilns with electrostatic precipitators (ESP) as controls, and one on a
dry-process kiln using a baghouse as a control. Neither of the wet-process
kilns met the proposed standards, while the dry-process kiln did. Inspections
of three additional kilns using baghouse controls showed no visible emissions.
Also, three tests were used as the basis of the clinker-cooler standard.
Of the three, the one with an ESP and one of the two with baghouse controls
met the standard. Observations at two additional coolers with baghouse con-
trols showed no visible emissions.
A suit brought by the Portland Cement Association against EPA (158 U.S.
App. D.C. 308,486 F.2d 375) sought a review of the NSPS for Portland cement
plants. A response by EPA proposed a revision in the opacity standard for
kilns from 10 to 20 percent, (refs. 6,7) No other revisions were proposed.
The Court of Appeals upheld the revised standards. The stages in the develop-
ment of the standards are shown in table 1.
The Court's ruling on the use of opacity standards to measure pollution
or aid in controlling emissions is considered a landmark decision. It ruled
that ". . . (3) plume opacity was not too unreliable to be used either as a
measure of pollution or as an aid in controlling emissions ..." (ref. 8)
EMISSIONS CONTROL AND COMPLIANCE
Control devices on kilns and clinker coolers are principally electro-
static precipitators and baghouses. At least one scrubber is being used
successfully to control particulates from a clinker cooler, (ref. 9) Cen-
trifugal collectors, along with ESP's and baghouses, are used to control
particulates from other sources. Overall, baghouses are the most common type
of control.
A 1976 unpublished report by ESED examined compliance data collected from
EPA regional offices. The data from eight plants showed that all kilns and
seven of the eight clinker coolers were in compliance, (ref. 9) The results
of the compliance testing, as well as the standards development testing, are
shown in figure 1. These data imply that present particulate control tech-
nology is capable of complying with the Portland cement NSPS.
REFERENCES
1. B. C. Brown, "Cement," Minerals Yearbook. 1974. Vol. 1 - Metals, Minerals,
and Fuels, Bureau of Mines, U.S. Dept. of the Interior, Washington, 1976.
2. N. R. lammaitino, "Cements' Changing Scene," News Features, Chemical Engineer-
ing, June 24, 1974, pp. 102-106.
3. Office of Air Programs, "Background Information for Proposed New Source
Performance Standards: Steam Generators, Incinerators, Portland Cement
Plants, Nitric Acid Plants, and Sulfuric Acid Plants," Tech Report No.
APTD-0711, U.S. EPA, August 1971.
178
-------
4. T. G. Hopper and W. A. Marrone, "Impact of New Source Performance Stand-
ards on 1985 National Emissions From Stationary Sources," U.S. EPA 450/
3-76-017, in print.
5- Federal Register, Vol. 36, No. 247, Part II, December 23, 1971.
6. Emissions Standards and Engineering Division, "EPA Response to Remand
Ordered by U.S. Court of Appeals for the District of Columbia in Portland
Cement Association V. Ruckelshaus (486 F.2d 375, June 29, 1973)," U.S.
EPA 450/2-74-023, November 1974.
7. Federal Register. Vol. 39, No. 219 (November 12, 1974).
8. District of Columbia Circuit Court, "Portland Cement Association V.
Russell E. Train," U.S. Court of Appeals, 158 U.S. App. D.C. 308, 486
F.2d 375, decided May 22, 1975.
9. Ed McCarley, EPA, ESED, Durham, N. C., June 20, 1977, personal communi-
cation.
179
-------
STATE-OF-ART SURVEY OF MIST ELIMINATION IN THE U.S.A.
Seymour Calvert*
Abstract \
Mist or drop entrainment is generated by most types of gas scrubbers
and must be removed from the gas. Mist eliminators3 or entrainment
separators, are incorporated into the scrubber body or in vessels down-
stream from the scrubber. Major types in use are cyclones, radial baffles,
zigzag baffles, flow directional change, beds of massive packing,^ beds ^
of fibrous packing, and tube banks. Primary drop collection efficiencies
for several types are given in the form of an A.P.T. cut/power plot.
Eeentrainment sets the upper limits on gas and liquid flow capacity.
Solids deposition is discussed and a quantitative relationship for
predicting suspended solid deposition rate is given.
i
This paper is a review of the state-of-the-art of mist elimination for
scrubbers. Mist eliminators or entrainment separators are generally
necessary to prevent undesirable rates of liquid drop emission from the
scrubber. An unfortunate consequence of the achievement of thorough and
vigorous liquid-gas contacting in the scrubber is that some of the liquid
is atomized and carried out of the scrubber (entrained) by the gas that has
been cleaned.
The liquid entrainment or mist, as it is commonly referred to, will
generally contain both suspended and dissolved solids. Entrainment drops
can cause corrosion, erosion, and mechanical failure of the fan blaaes or
housing. Liquid or residual solid entrainment can also be deposited in the
ductwork, heat exchange surfaces, and smokestack and cause eventual plugging
and/or corrosion. It can cause problems in the area immediately surrounding
the point of emission due to "rain-out" of liquid drops.
Stack emissions in the atmosphere will include the solid residues of
the dried entrainment drops. The particulate composition can be quite
different than that entering the scrubber, especially where reactive
solutions or slurries are used for gas scrubbing. Thus, a bizarre consequence
of excessive entrainment from a scrubber system can be that more pollutant
is emitted either in total or within a certain size range than entered the
scrubber.
In many cases excessive entrainment imposes a limitation on the
capacity of the scrubber. That is, while the scrubber itself might be
capable of handling a larger gas flow rate, the rate at which entrainment
is considered excessive will dictate a limit on gas flow rate.
*Air Pollution Technology, Inc., San Diego, California 92117
180
-------
GENERAL PRINCIPLES
The flow rate and drop size distribution of the entrainment will
depend greatly on the nature of the gas-liquid contacting within the
scrubber. The drops entrained from the scrubber contacting zone may be
separated from the gas by a number of mechanisms, the most important of
which will be discussed below. The initial collection of the drops is
referred to as "primary collection" and the main mechanisms for this are as
follow:
1. Sedimentation due to gravitational force.
2. Centrifugal deposition caused by changes in the main gas flow
direction.
3. Inertia! impaction upon collection bodies placed in the flowing
gas stream.
After the liquid drops have been collected, they will coalesce into a
sheet of liquid, unless the velocity of impact is so high that some of the
drops shatter. In a properly designed and operated entrainment separator,
the liquid should drain out of the entrainment separator. The capacity of
the entrainment separator is limited by the reentrainment and/or flooding
that will occur when the gas velocity is too high. The mechanisms by which
reentrainment and flooding occur are as follow:
1. Liquid holdup in the separator increases, the cross-sectional
area for gas flow decreases, and the resulting aerodynamic
forces cause massive bodies of liquid to be blown out.
2. At high gas flow rates, the aerodynamic drag forces can cause
tearing of liquid from the films flowing downward in the entrainment
separator.
3. Drops may shatter due to collision with one another or with solid
bodies and form smaller drops that are carried out by the gas.
The deposition of solids within the entrainment separator can be a
major source of difficulties. The most important mechanisms by which this
occurs are as follow:
1. Settling to nonvertical surfaces.
2. Deposition by inertial impaction caused by changes in liquid flow
direction and turbulence.
3. Liquid loss from slurry on the entrainment separator surface
leaves a residue of solid particles.
4. Precipitation of dissolved solids.
APPARATUS TYPES
The apparatus used for entrainment separation can be described in
general terms as shown in the categorical list below:
A. Gravitational sedimentation
1. Within the scrubber and its outlet ducting, sedimentation is
always active and important.
181
-------
2. Entrainment separators using sedimentation are rarely used
following a scrubber.
B. Centrifugal deposition
1. Cyclone separators of various designs are common. "Anti-
creep" skirts to prevent liquid flow into the gas outlet
tube and bottom baffles to prevent liquid swirl at the
liquid outlet are recommended (Nonhebel 1972) (ref. 1).
(See figure 1.)
2. Radial baffles and other types of guide vanes are used to
induce a rotary motion of the gas stream within scrubber
shell. As in the cyclone separator, this causes drop
deposition on the scrubber shell and the baffles, and from
there the liquid film runs down into the scrubber.
3. Zigzag baffles, chevrons, corrugated sheets, and similar
devices are used to cause one or more abrupt changes of gas
flow direction. Drops are deposited on the baffles and the
collected liquid film either runs down the baffles (if their
major axis is vertical) or drips off as large drops (if axis
is horizontal). There are many proprietary designs employing
shapes that are claimed to give better liquid drainage
(Conkle et al., 1976) (ref. 2). Figure 2 shows an example.
4. A directional change in the gas duct causes considerable
deposition of large drops. This is commonly used with
venturi scrubbers that can be oriented with flow downward
into a duct that turns 90° to cause flow in the horizontal
direction.
C. Inertia! impaction
1. Beds of massive packing elements (such as saddles, rings,
and other gas absorption or distillation tower packings) are
used in either vertical or horizontal gas flow configurations.
2. Fibrous packings are used in beds comparable to massive
packed systems. Knitted wire mesh, screen, metal wool,
felted fibers, open-celled polymeric sponge, and glass fiber
have been used for this purpose.
3. Banks of round tubes, streamlined struts, and other shapes
have been used in vertical and horizontal orientations.
4. Trays (or plates) such as perforated plates, valve trays,
impingement plates, and others have been used for special
purposes. Because there are scrubbing devices, they generate
entrainment and their use following another type of scrubber
may be redundant and should be carefully analyzed.
PRIMARY EFFICIENCY
The overall collection efficiency for drops is dependent on primary
collection and on reentrainment. In most cases, the overall efficiency is
low at gas velocities because the primary efficiency is low, then it
increases with velocity. At higher velocities, reentrainment begins and
causes a decrease in overall efficiency even though the primary efficiency
is high.
182
-------
Gas Outlet
Gas
Inlet
30x15
Liquid Outlet
-<—— 70 -
Figure 1. Cyclone separator. All
dimensions are in cm.
Figure 2. Top view of zigzag
baffle arrangement.
183
-------
100
00
50
oi
w
H
w
10
-"-u
U
0.01
B
0.05 0.1
A, B = BAFFLES, 6 ROWS (30°, 45°)
C, D = TUBE BANK, 6 ROWS (1 cm, 0.3
E = PACKING (2.5 cm DIA)
F = WIRE MESH [0.029 cm DIA)
0.51 5
PRESSURE DROP, cm W.C.
10
50
Figure 3. Entrainment separator performance cut diameters.
-------
Methods for predicting primary collection efficiency for various types
of separators were reported by Calvert et al. (1974, 1975, and 1977) (refs.
3, 4, 5). A concise representation of the primary efficiency capabilities
of several types of separators is given by the "cut/power" relationship
developed by Calvert (1974) (ref. 6). This relationship is shown in a plot
of the drop diameter collected at 50 percent efficiency (i.e., the cut
diameter) against the gas-phase pressure drop or power input for the separa-
tor (see figure 3).
Figure 3 was drawn based on design equations and experimental corre-
lations. It shows curves for baffles at two angles of attack to the flow
direction (not inclinations to the horizontal), curves for tube banks with
two different spacings between tubes with a row, packing of one particular
size, and knitted mesh with a certain wire diameter. The curves would be
different for entrainment separators of other collection element dimensions,
but are not much affected by the extent of the separator. For instance,
the position of the lines changes with the wire diameter--it is not affected
by the thickness of the knitted mesh pad over a reasonable range.
REENTRAINMENT
The liquid and gas flow capacity of an entrainment separator is
limited by the occurrence of reentrainment. Experimental data on collection
efficiency for several types of separators have been taken by Bell and
Strauss (1973) (ref. 7), Calvert et al., (1975), and Houghton and Radford
(1939) (ref. 8). A summary of reentrainment characteristics is given below
for several types of separators.
Cyclone
Cyclone separator performance was measured for a straight cylindrical
cyclone 61-cm i.d. and 224 cm long. A serrated anti-creep skirt was used
on the 25-cm i.d. central outlet tube. Figure 4 is a plot of percent
o
E-
W
CYCLONE INLET AREA
O 30.5x15.2 cm2 (no vane)
D 30.5x7.6 cm2 (with vane)
Water,Rate = 50 £/min
- -O-GHSKDO—-aa-^-Ti^
0 10 20 30 40 50 60 70 80
GAS VELOCITY IN CYCLONE INLET, m/sec
Figure 4. Experimental penetration in cyclone.
185
-------
100
90
80
w
w
o
I—I
H
o
CJ
60
40
20
I T^ ' T
Calvert, et al.
O
Houghton 5 Radford
Data, n=6, 0-30°
Bell 5 Strauss
Data for 2 "V"
Baffles in
Series
I I I I
pg = 380 ym
i =1.5
g I I
°12545678
AIR VELOCITY, m/sec
Figure 5. Collection efficiency for vertical zigzag baffle.
186
-------
penetration (of drops) versus inlet gas velocity. It can be seen that no
reentrainment occured below 40 m/s and that less than 0.5 percent was
observed between 40 m/s and 60 m/s.
Zigzag Baffles
Several variations of zigzag baffle arrangements were studied by
Calvert et al., (1975) for 7.4 cm wide baffles set 7.6 cm apart in rows
normal to the gas flow and 2.5 cm between rows. The arrangement included
horizontal gas flow with vertical baffles (having attack angles of 30° or
45°) and vertical, upward gas flow with baffles inclined 0° (horizontal),
30°, and 45°.
Figure 5 shows a comparison of collection efficiency versus air
velocity data by three investigators. Figure 6 is a plot (for the horizon-
tal gas flow arrangement) of the liquid/gas ratio entering the separator
versus the superficial (empty cross sectional) gas velocity in the separator
with reentrainment rate as a parameter. The indication of zones is approxi-
mate and it indicates that above the shaded area some slight reentrainment
was noted.
Figure 7 is a plot of the approximate curves above which heavy reen-
trainment occurred for three baffle inclinations in the upward flow arrange-
ment. It can be seen that the separator capacity increases with the baffle
inclination.
10"
6
"e
o
10"
a
r-
a
KH
cr
10-
o
o
Some reentrainment (<1%)
Reentrainment in part of
duct only
Primary efficiency <100£
No penetration
I I I I I
SUPERFICIAL GAS VELOCITY, m/sec
Figure 6. Reentrainment from vertical baffles,
(horizontal gas flow)
187
-------
o
H
10
- 3
o
H
2
to 10
- If
10
- 5
Parameter is baffle
inclination to horizontal^
i
_'.. .'..„! U
tttt
T~ i I .
:::rfe:r,
-l-fK-i
i TTTl
0
1 234 5678
SUPERFICIAL GAS VELOCITY, m/s
Figure 7. Reentrainment from horizontal and inclined baffles (upward flow),
188
-------
Massive Packings
, Drop collection efficiency data for horizontal and upward flow through
beds of 2.5-cm nominal size Pall Rings are shown in figure 8, a plot of
penetration versus superficial gas velocity. For horizontal gas flow with
liquid/gas ratio ranging up to 1 1/m3, there is no reentrainment until the
gas velocity exceeds 6 m/s. Reentrainment starts at a lower velocity when
the gas flow is vertical and it is comparable to the 3.5 m/s flooding
velocity predicted for a counter-current packed tower.
Fibrous Packing
Knitted wire mesh pads were studied with horizontal and vertical air
flow and figure 9 shows a plot of drop penetration versus superficial gas
velocity. Reentrainment started at a velocity of 4.5 m/s and a liquid/gas
ratio of 0.4 1/m3 for the upward flow arrangement. In the case of horizon-
tal flow, reentrainment did not become significant until the gas velocity
was about 6.5 m/s, although slight reentrainment was noted in part of the
duct at velocities over about 5 m/s.
Tube Banks
Tube banks composed of 6 rows of 1.9-cm o.d. tubes spaced 3.8 cm
center-to-center (therefore the space between tubes was 1.9 cm) and with
tube centers staggered from row-to-row were studied with horizontal and
upward air flow. The measured reentrainment loci are shown in figure 10,
a plot of liquid/gas ratio versus superficial gas velocity. Liquid/gas
ratio up to 0.22 1/m3 at 7 m/s gas velocity did not have a significant
effect on the horizontal flow arrangement.
Solids Deposition
Industrial experience with entrainment separator fouling and plugging
has yielded some empirical observations and guidelines for design. Vertical
collection surfaces stay cleaner than horizontal ones due to better liquid
drainage. Intermittent washing with sprays is beneficial, but the details
of the washing system and procedure depend on the specific case. Precipi-
tation scaling must be controlled through the system chemistry.
Some quantitative experimentation on suspended solids deposition has
been done by Calvert et al., (1975), and their findings are summarized
below:
1. Deposition rate decreases as the slurry flux on the surface is
increased.
2. Deposition rate decreases as the liquid film thickness is increased.
3. Deposition rate is higher on an inclined baffle than on a vertical
baffle due to the increase in settling rate of solids suspensions.
4. Small drops are more susceptible to being caught in eddies, which
would bring them to the back surfaces of the baffles.
5. Small drops that do hit the baffle surface have a higher deposi-
tion rate than larger drops because of their lower localized
slurry flux.
189
-------
o
E-
W
PH
40
30
20
10
0
i 1 i 1 1 r i i
84 ym DIA. DROPS
O 0.78 2/m2-s, VERTICAL FLOW
D 0-.38 £/m2-s, VERTICAL FLOW
— FLOODING PREDICTION
0
ZONTAL
FLOW
1
1
1
1
1
10
12345678 9
SUPERFICIAL GAS VELOCITY, m/s
Figure 8. Experimental penetration through packed bed
190
-------
50
40
•> 30
o
t—i
t->
I 2°
w
P-.
10
I I I I I I
84 ym DIA. DROPS
O 0.65 £/m2-s, VERTICAL FLOW
D 0.32 £/m2-s, VERTICAL FLOW
I
j HORIZONTAL
FLOW
I
I
I
I
1
123456789
SUPERFICIAL GAS VELOCITY, m/s
10
Figure 9. Experimental penetration through mesh.
191
-------
- 3
10 r^:
£=
•\
on
{-
CO
I
o
H
ex
n
^
10'
10
- 5
VERTICAL FLOW p^r^^fe^
^HORIZONTAL FLOWi>
0 1 2 34 5 6 7 8
i
SUPERFICIAL GAS VELOCITY, m/s
Figure 10. Reentrainment from tube bank.
192
-------
The deposition data were correlated by the following empirical equation:
Rs = Vfy exp [- (0.13 + 0.53<|))6] (1)
where R^ = deposition rate of CaCo3 on a vertical flat surface, mg/cm2-s,
W = weight fraction of solids in the slurry,
<|> = slurry flux, mg/cm2-s, and
6 = liquid film thickness, urn.
REFERENCES
1. Nonhebel, Processes for Air Pollution Control, 2nd ed., Chapter 16 -
"Spray Elimination," CRC Press, Cleveland, 1972.
2. H. N. Conkle, H. S. Rosenberg, and S. T. DiNovo, "Guidelines for the
Design of Mist Eliminators for Lime/Limestone Scrubbing Systems," EPRI
FP-327, 1976.
3. S. Calvert, S. Yung, and J. Leung, "Entrainment Separators for Scrub-
bers -Final Report," EPA 650/2-74-119b, NTIS No. PB 248-050, 1975.
4. S. Calvert, S. C. Yung, H. F. Barbarika, and L. E. Sparks, "Entrain-
ment Separators for Scrubbers," presented at the Second Fine Particle
Scrubber Sumposium, New Orleans, 1977.
5. S. Calvert, I. L. Jashnani, and S. Yung, "Entrainment Separators for
Scrubbers," J. Air Pollution Control Association, Vol. 24, No. 971
(1974).
6. S. Calvert, "Engineering Design of Fine Particulate Scrubbers," J. Air
Pollution Control Association, Vol. 24, No. 929 (1974).
7. C. G. Bell and W. Strauss, "Effectiveness of Vertical Mist Eliminators
in a Cross Flow Scrubber," APCA Journal, Vol. 23, No. 967 (November
1973).
8. H. G. Houghton and W. H. Radford, Measurements on Eliminators and the
Development of a New Type for Use at High Gas Velocities," Trans.
American Inst. of ChE, Vol. 35, No. 427 (1939).
193
-------
MIST ELIMINATOR TESTING AT THE SHAWNEE PROTOTYPE
LIME/LIMESTONE TEST FACILITY
John E. Williams*
Abstract
This paper presents the major test results and discusses their
significance in respect to practical application to commercial-'Size
units from the mist eliminator testing conducted at the Shawnee prototype
lime/limestone test facility over a 5-year period from 1972 through
1977. Much of the early work was conducted at a reduced gas velocity
using various combinations of mist eliminator washing techniques and
hardware configurations, including a wash tray upstream of the mist
eliminator. It was not recognized early in the program that scaling and
mud-type solids deposition in the restricted area of the mist eliminator
were indeed separate problems., each with separate and distinct solutions.
Consequently, it was only after this fact was recognized that methods
were developed to adequately control both problems at design gas velocities.
Control of mud-type solid deposits proved to be more difficult than
control of scaling. The real key to successful control of mud-type
solid deposits was not found until tests were initiated to study methods
to improve alkali utilization. A correlation was found between alkali
utilization and accumulation of the mud-type solids. Above about 85
percent alkali utilization, these solids can easily be removed even with
very infrequent (once per 8-hour shift) washing with fresh makeup water.
Tests to confirm this correlation and its significance for both improving
operating reliability and in reducing overall scrubbing costs for commercial-
size applications are also discussed. Selection of operating parameters
(such as liquid-to-gas ratio (L/C), reaction tank residence time, pH, or
percent solids %n the circulating slurry to maintain a gypsum saturation
level in the scrubber below about ISO percent) will normally control
scaling.
Recent test results applicable to mist eliminator performance
obtained during flue gas characterization testing at Shawnee are also
presented.
INTRODUCTION
With dramatic increases in the use of high sulfur coal reserves
being a near certainty, the widespread use of flue gas desulfurization
(FGD) will play an increasingly important role in near term SO control
A
strategies. Exoanded coal use must be accompanied by sound air and
water pollution control measures to minimize threats to human health and
to avoid widespread environmental damage.
*Emission/Effluent Technology Branch, Utilities and Industrial
Power Division, Industrial Environmental Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina
27711
194
-------
As the FGD technology most commonly accepted and used by domestic
utilities for coal-fired boiler applications, lime/limestone processes are
especially important to the success of a policy to rapidly expand coal use.
Based on the most recent estimate, 30 FGD systems are presently installed
and an additional 94 are in various stages of planning or construction.
Most of these 124 units will be in service by the early 1980's. Of these,
78 have already selected lime/limestone processes and an additional 25 have
not yet selected a process type.
For the past 5-1/2 years the Industrial Environmental Research Lab-
oratory-Research Triangle Park (IERL-RTP) has sponsored the operation and
testing of a prototype lime and limestone wet scrubbing test facility
located at the Tennessee Valley Authority's (TVA) Shawnee power plant near
Paducah, Kentucky. This versatile facility allows comprehensive testing of
up to three scrubber types under a variety of operating conditions. Bech-
tel Corporation of San Francisco, as the major Environmental Protection
Agency (EPA) contractor, designed the test facility and directs the test
program. TVA constructed and operates the facility.
The major concerns of the utility industry to date regarding lime/
limestone scrubbing have centered around scaling and plugging potential,
the large quantities of byproduct sludge generated, and the high costs
(capital and operating) of scrubbing. It is toward these areas of concern
that the Shawnee.program has been directed.
A summary which describes in more detail the test facility, the test
program objectives, the significant test results to date, and the future
plans is presented in reference 1. This paper will be directed only to the
mist eliminator testing portion of the Shawnee program. The major mist
eliminator test results and their significance in respect to practical
application to commercial-size units will be discussed.
DISCUSSION OF RESULTS
Throughout the Shawnee program, one of the most persistent problems
has been that of scaling and plugging, especially in the relatively restricted
area of the mist eliminators. Consequently, much of the early effort was
devoted to identifying and learning how to control two separate and distinct
operating reliability problems—scaling and soft, mud-type solids deposition,
especially in the mist eliminators. In fact, identification of the problems
was even more difficult because it was not recognized early in the program
that scaling and mud-type solids deposition were indeed separate problems,
each with separate and distinct solutions. This fact was not really apparent
until methods were found to adequately control both.
In a lime- or limestone-based system, the most frequently encountered
scaling problem is due to sulfate (gypsum) crystals which precipitate on
scrubber internals rather than outside the scrubber in the reaction tank.
Selection of operating parameters (such as liquid-to-gas ratio (L/G),
reaction tank residence time, pH, or percent solids in the circulating
slurry to maintain a gypsum saturation level in the scrubber below about
130 percent will prevent this type of scaling. The use of very infrequent
(once every 4 to 8 hours) top sequential washing with fresh makeup water
and low pressure drop nozzles proved adequate at Shawnee for control of
195
-------
scaling in the mist eliminators. However, identification of those operating
parameters affecting scaling potential, safe operating ranges, and mist
eliminator washing techniques proved time consuming (refs. 2, 3, and 4).
This was especially true with the problem of mud-type solids deposition,
which was superimposed over the scaling problem.
Solution to the problem of mud-type solids deposition proved even more
difficult than control of Scaling, especially in the mist eliminators.
Much of the early work of Shawnee was conducted at a reduced gas velocity
using various combinations of mist eliminator washing techniques and^hard-
ware configurations, including a wash tray upstream of the mist eliminator.
The use of a wash tray upstream of the mist eliminator to transfer the mud-
type deposits to an area where they could more easily be removed by washing
was only marginally successful and required reduced gas flow rates. Ultimately,
washing techniques for the mist eliminator were found which controlled
restriction from the mud-type solids to less than 10 percent of the open
area at satisfactory gas flow rates and without the need for a wash tray.
This level of solid deposits did not hinder operation or cause shutdown for
periodic cleaning. However, the real key to successful control of these
mud-type solid deposits was not found until well into the advanced test
program.
During tests to study methods to improve alkali utilization, a cor-
relation was found between alkali utilization and the accumulation of these
mud-type solids. Above about 85 percent alkali utilization, the solids can
easily be removed even with very infrequent (once per 8-hour shift) washing
with fresh makeup water. In fact, an entirely clean system was maintained
over an extended period using only approximately 25 percent of the available
fresh makeup water. (The available makeup water is defined as that added
to maintain a closed-loop water balance and represents the water lost
through evaporation and that which leaves the system with the purged solids.)
This correlation was subsequently confirmed at Shawnee using three different
methods for improving alkali utilization (ref. 2) and has also been confirmed
on a different system at TVA's 1 MW Colbert pilot plant (ref. 5). The
ability to improve alkali utilization then becomes extremely important for
limestone scrubbing because:
1. It provides a means for existing installed full-scale systems, at ,
low cost, to solve or to reduce the severity of an operating
reliability problem that persisted in many of the early applications.
2. It permits the use of limestone costing about $4 to $6/metric ton
rather than lime costing about $30 to $40/metric ton at greater
than 90 percent utilization (or about l.lx stoichiometric quantity),
which is comparable to that normally obtained with lime.
3. It further reduces the overall costs of scrubbing by substantially
reducing the quantity of byproduct sludge produced.
The methods developed for improving alkali utilization apply equally
as well for lime scrubbing as for limestone scrubbing. However, since
alkali utilization of greater than 85 percent is normally obtained with
conventional lime scrubbing, further improvement of the lime utilization is
not as significant as improved limestone utilization.
196
-------
Scaling and plugging are no longer a major concern at Shawnee. Avoiding
these types of problems has become almost routine, although lessons learned
are not ignored in selecting the operating conditions for current tests.
The detailed discussions contained in references 2, 3, and 4 help to illustrate
that finding solutions to these problems was not as simple and straightforward
as it may appear from the brief summary presented here. Schematic drawings
showing the scrubbers, the overall scrubber systems, the various types of
mist eliminators tested at Shawnee, the wash tray, and the various types of
wash configurations used are also included in references 2, 3, and 4.
Another important facet of the mist eliminator testing at Shawnee of
interest for this paper are the flue gas characterization tests that address
the mist eliminator efficiency. In order to maximize the amount of useful
data for the funds expended and to obtain long-term continuous data (as
opposed to having an outside group come in for one or two spot checks), it
was decided to train onsite personnel to conduct mist eliminator emission
characterization tests routinely. Through a separate contract, TRW Corporation
was chosen to develop the best testing methods for particulate mass loading,
particle size distribution, sulfate measurement, and mist eliminator liquid
entrainment; to write the appropriate operating manuals; and to train the
onsite personnel in the procedures developed. Most of these tests are now
being run routinely by the Shawnee onsite personnel.
The test methods and preliminary test results covered briefly here are
discussed in detail in references 4 and 6. Initial intensive test series
on both the venturi/spray tower and the turbulent contact absorber (TCA),
systems were run to determine the effect of major operating variables on
particulate mass removal, emitted particle size distribution, and S03
removal. Variables investigated on the venturi/spray tower system were gas
flow rate,, slurry circulation rate, MO addition, venturi pressure drop,
y
mist eliminator wash configuration, percent solids recirculated, and inlet
particulate mass loading. Variables investigated on the TCA system were
gas flow rate, slurry circulation rate, mist eliminator4 wash configuration,
and inlet particulate mass loading.
The limited data available thus far indicate:
1. Outlet particulate mass loading appears to be unaffected by
operating variable levels over the ranges investigated except for
venturi pressure drop and inlet particulate mass loading.
Outlet mass loading was always less than the inlet mass loading
even for very low values of the inlet mass loading. Outlet mass
loading was directly proportional'to inlet mass loading and was
inversely proportional to the venturi pressure drop.
2. Scrubber reaction product emission represented only a small
percentage of the total mass emission (usually less than 10
percent of the total) with the bulk being fly ash particles that
entered with the flue gas. Size distribution results and scanning
electron microscope photographs indicated that most of the scrubber
reaction product emission was confined to the fractions larger
than 5 u. This also suggests that the type of mist eliminator used
at Shawnee is reasonably efficient. However, this is based on a
limited number of tests and further tests are planned to verify
these observations.
197
-------
3. Measurement of mass penetration through the scrubbers as a
function of particle size indicated that there was no massive
generation of particulate in any size fraction (i.e., outlet
greater than inlet). Penetration became significant in the range
of 1 u and smaller where penetration varied from 25 to 90 percent.
Under comparable operating conditions, the relative mass penetration
for the TCA system was greater than for the venturi/spray tower
system.
4. The S03 removal efficiency of the scrubbers appeared to be
fairly constant at about 58 percent over the range of variables
tested.
SUMMARY AND CONCLUSIONS
Mist eliminator testing at Shawnee has made a significant contribution
toward improved reliability for lime and limestone scrubbing systems.
Furthermore, the results have helped in overcoming other major utility and
vendor concerns as well. These include equipment and process variations
that can lead to substantial reductions in both the quantity of sludge
produced and in the costs for installing and operating the lime and limestone
scrubbing sytems. The most significant results of the Shawnee mist eliminator
tests to date include:
1. Demonstration has shown that conventional lime/limestone systems
can be operated reliably. Two separate reliability problems were
identified—scaling and soft, mud-type solids deposits, and
methods to control each were demonstrated.
2. Soft, mud-type solids deposition was shown to be a strong function
of alkali utilization. At high alkali utilization (greater than
about 85 percent) these solids are much more easily removed, and
very infrequent intermittent freshwater wash is adequate for
their complete removal in restricted areas such as the mist
eliminator where accumulation can lead to plugging problems.
3. Several equipment or process variations were also demonstrated
for improving alkali utilization. This is particularly signifi-
cant for limestone where alkali utilization is typically about 60
to 70 percent and can be increased to 85 to 95 percent, or
comparable to that normally obtained with lime. This not only
improves reliability, but also reduces costs by permitting the
use of a much less expensive alkali feed material and by substantially
reducing the quantity of byproduct sludge produced.
4. Initial results from a limited number of tests indicate that when
a reasonably efficient mist eliminator is used, scrubbers do not
generate or contribute to either particulate or sulfate emissions
in any particle size fraction. Mass emissions were unaffected by
changes in operating variables over the ranges investigated
except for venturi pressure drop and inlet mass loading. Reaction
products were concentrated in those fractions greater than 5 |j
and were usually less than 10 percent of the total mass emissions.
The S03 removal efficiency was relatively constant at about 58
198
-------
percent over the range of variables investigated. Further tests
are planned to verify these observations and to better quantify
the initial data.
REFERENCES
1. J. E. Williams, unpublished paper*, "Summary of Operation and Testing
at the Shawnee Prototype Lime/Limestone Test Facility," April 1977.
2. Bechtel Corporation, "EPA Alkali Scrubbing Test Facility: Advanced
Program - Second Progress Report," EPA-600/7-76-008 (NTIS No. PB 258-
783/AS), September 1976.
3. Bechtel Corporation, "EPA Alkali Scrubbing Test Facility: Advanced
Program - First Progress Report," EPA-600/2-75-050 (NTIS No. PB-245-
279/AS), September 1975.
4. Bechtel Corporation, "EPA Alkali Scrubbing Test Facility: Advanced
Program - Third Progress Report," EPA-600/7-77-105, September 1977.
5. Tennessee Valley Authority, "TVA's 1-MW Pilot Plant: Final Report on
High Velocity Scrubbing and Vertical Duct Mist Elimination," EPA-
600/7-77-019 (TVA PRS-19) (NTIS No. PB 269-850/AS), March 1977.
6. R. G. Rhudy and H. N. Head, paper presented at the Second Fine Particle
Scrubber Symposium in New Orleans, Louisiana, "Results of Flue Gas
Characterization Testing at the EPA Alkali Wet-Scrubbing Test Facility,"
May 1977.
*A copy of this paper is included here as Attachment A.
199
-------
SUMMARY OF OPERATION
AND TESTING AT THE
SHAWNEE PROTOTYPE LIME/LIMESTONE
TEST FACILITY
John E. Williams
Emissions/Effluent Technology Branch
Utilities and Industrial Power Division
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
April 1977
ATTACHMENT A
200
-------
INTRODUCTION
EPA's Industrial Environmental Research Laboratory-RTF and its predecessors
over approximately the past 15 years have undertaken a program of research
and development to establish the technical and economic feasibility of
promising processes for removing sulfur dioxide and particulates from
coal-fired boiler flue gases. An important part of this effort involves
the operation and testing of a prototype wet lime and limestone scrubbing
test facility located at the Tennessee Valley Authority's (TVA) Shawnee
power plant near Faducah, Kentucky. This versatile facility allows
comprehensive testing of up to three scrubber types under a variety of
operating conditions. Bechtel Corporation of San Francisco, as the major
contractor, designed the test facility and directs the test program. TVA
constructed and operates the facility.
Purposes of this report are to describe the Shawnee test facility, to
discuss the test program objectives, and to summarize major test results
to date. More detailed information is presented in published EPA reports
listed in the reference section at the end of this report. Two additional
reports are presently being reviewed and should be published by July 1977.
These are: 1) the third progress report covering testing from February
through November 1976 under the advanced program; and 2) the second annual
report covering the Shawnee field* evaluation of the disposal of flue gas
cleaning wastes.
The test facility consists of three parallel scrubber systems: a venturi
followed by a spray tower, a Turbulent Contact Absorber (TCA), and a
Marble-Bed Absorber. Each system is capable of treating approximately
10 MW equivalent (30,000 acfm @ 300°F) of flue gas from Shawnee unit 10
(nominally 150 MW total) containing 1000 to 5000 ppm sulfur dioxide
(S02) and 0.2-11 gms/m3 of particulates (fly ash). Operation of the
Marble-Bed Absorber was discontinued in July 1973 (see Reference 1 for a
discussion of the reasons), and subsequent operation has been limited to
the two remaining scrubber systems. Each of the systems can be operated
in a variety of configurations. Typical system configurations depicting
lime nesting with the venturi/spray tower and limestone testing with the
TCA scrubber are illustrated schematically in Figures 1 and 2, respectively.
Major goals of the original two-year test program were:
• Characterization of the effect of important process variables on
sulfur dioxide and particulate removal.
• Development of mathematical models to allow scale-up to full-size
scrubber facilities.
201
-------
AIR
FUEL
FLUE GAS
VENTURI
SCRUBBER
> LIME >-
> ,
>—I
REHEATER
MAKEUP
XA A A«g
««ft\VV»it\
V V V
A A A
A A A
A A A
A A A
V
' r ^ '
SCRUBBER
EFFLUENT
HOLD
TANK
i
SPR
TOW
-«
f
~*^
'*
AY
ER
ET
— 1—
(
^i^
)
^ WATER X
SAMPLE POINTS
O Gas Composition
® Particulate Composition & Loading
© Slurry or Solids Composition
. — Gas Stream
__ Liquor S-ream
SETTLING POND
Figure 1. Venturi/Spray Tower-Typical System Configuration
-------
rvs
o
co
SCRUBBER REACTION TANKS
SAMPLE POINTS
O Gas Composition
® Participate Composition & Loading
0 Slurry or Solids Composition
_ _ Gas Stream
— Liquor Stream
SETTLING POND
Figure 2. Turbulent Contact Absorber-Typical System Configuration
-------
• Development of information to study the technical and economic
feasibility of lime/limestone scrubbing.
• Demonstration of long-term reliability.
Although for the most part these goals were met at the completion of the
original test program, it became apparent that additional Shawnee testing
was needed to supply further information and to optimize lime and limestone
systems in the areas of: sludge disposal, improved reliability, variable
load operation, and improved process economics. This conclusion was based
on comments of need from utilities and flue gas desulfurization system ven-
dor representatives, and particularly the encouraging results from IERL-RTP
pilot plant support studies which indicated several potentially attractive
new scrubber operating concepts and process variations that could favorably
affect both the economics and the operating reliability of scrubbing.
Therefore, in 1974 the Shawnee program was extended and the scope was
expanded. Additional objectives of the expanded program were:
• Investigation of advanced process and equipment design variations for
improving system reliability and economics.
• Evaluation of process variations for a substantial increase in alkali
utilization for limestone systems.
• Evaluation of the effect of increased magnesium ion concentration or
other additives on reducing gypsum saturation and increasing S02
removal efficiency.
• Determination of the efficiency and reliability of lime and limestone
scrubbers under conditions of widely varying flue gas flow rate and
inlet SO- concentration.
• Evaluation of system performance and reliability without fly ash in
the flue gas.
• Assessment of the effectiveness of forced oxidation or other process
variations for producing an improved throwaway sludge product.
• Determination of the practical upper limit of S02 removal efficiency
for both lime and limestone scrubbing systems.
• Development of a design/economic study computer program for comparison
of both initial investment and lifetime operating costs for full-scale
lime and limestone systems.
Two smaller scrubbing systems (0.1 megawatt each) have also been operated
at the IERL-RTP pilot plant in support of the Shawnee program. These
small, pilot-scale scrubber systems are capable of simulating the Shawnee
scrubber systems with excellent agreement in the lime/limestone wet-scrub-
bing chemistry. Preliminary data are generated on the pilot-scale systems
to verify and guide the selection of those promising concepts that should
logically be investigated on the larger size Shawnee units. Indeed, results
obtained from operation of the IERL-RTP pilot plant were very instrumental
in the decision to continue testing at Shawnee under the advanced test
program. 2Q4
-------
DISCUSSION OF RESULTS
The test program schedules based on the defined objectives for the
initial program and the advanced concept testing are presented in Figures
3 and 4, respectively. As shown by the schedules, much of the early
effort was devoted to identifying and learning how to control two separate
and distinct operating reliability problems—scaling and soft, mud-type
solids deposition, especially in the relatively restricted area of the
mist eliminators. Much of the early work was also conducted at a re-
duced gas velocity using various combinations of mist eliminator washing
techniques and hardware configurations, including a wash tray upstream
of the mist eliminator. In fact, it was not recognized early in the
program that scaling and mud-type solids deposition were indeed separate
problems, each with separate and distinct solutions. This fact was not
really apparent until methods were found to adequately control both.
In a lime or limestone based system, the most frequently encountered
scaling problem is due to sulfate (gypsum) crystals which precipitate on
scrubber internals rather than outside the scrubber in the reaction
tank. Selection of operating parameters (such as liquid-to-gas ratio
(L/G), reaction tank residence time, pH, or percent solids in the circu-
lating slurry to maintain a gypsum saturation level in the scrubber
below about 130%) will prevent this type of scaling. The use of very
infrequent (once every 4-8 hours) top sequential washing with fresh
make-up water and low pressure drop nozzles proved adequate at Shawnee
for control of scaling in the mist eliminators. However, identification
of those operating parameters affecting scaling potential, safe oper-
ating ranges, and mist eliminator washing techniques proved time con-
suming. This was especially true with the problem of mud-type solids
deposition which was superimposed over the scaling problem.
Solution of the problem of mud-type solids deposition proved even more
difficult than control of scaling, especially in the restricted area of
the mist eliminators. The use of a wash tray upstream of the mist
eliminator to transfer the mud-type deposits to an area where they could
more easily be removed by washing was only marginally successful and
required reduced gas flow rates. Washing techniques for the mist elimi-
nator were ultimately found without requiring a wash tray and at design
gas velocities which limited restriction from these mud-type solids to
less than 10% of the open area. These solid deposits did not hinder
operation or cause shut-down for periodic cleaning. However, the real
key to successful control of these mud-type solid deposits was not found
until well into the advanced test program.
205
-------
TEST PROGRAM FUNCTIONS
SYSTEM CHECK-OUT
AIR/ WATER TESTS
SODIUM CARBONATE TESTS
L IMESTONE WET-SCRUBBING TESTS:
Factorial Tests
Reliability Verification Tests
Reliability Tests (TCA)
LI ME WET-SCRUBBING TESTS:
Reliability Tests (Venturi/Spray Tower)
1972
MAMJJASOND
1 23456789 10
BOILER OUTAGE
SYSTEM MODIFICATI
J F
1112
.4
1
!j
& —
ONS
1973
MAMJJASOND
[13 14 15 16 17 18 19 20 21 22
1
1
1
1
1974
J F M A M J J A S 0
23242526272829303132
Figure 3. Shawnee Test Program Schedule
-------
L miaul wutf i.
Figure 4. Shawnee Advanced Test Program Schedule
SCHOMUFOR
ADVANCED TESIHtOGMM
•nmnoposasixiMiaH
CXIENSION -NOVEMBO 1W
-------
During tests to study methods to Improve alkali utilization, a correlation
was found between alkali utilization and the accumulation of these mud-type
solids. Above about 85% alkali utilization, the solids can easily be
removed even with very infrequent (once per 8-hour shift) washing with
fresh make-up water. In fact, an entirely clean system was maintained over
an extended period using only approximately 25% of the available fresh
make-up water. (The available make-up water is defined as that added to
maintain a closed-loop water balance and represents the water lost;through
evaporation and that which leaves the system with the purged solids.) This
correlation was subsequently confirmed at Shawnee using three different
methods for improving alkali utilization and has also been confirmed on a
different system at TVA's 1 MW Colbert pilot plant. The ability to improve
alkali utilization then becomes extremely important for limestone scrubbing
because:
• It provides a means for existing installed full-scale systems, at
minimal cost, to solve or improve an operating reliability problem
that persisted in many of the early applications.
• It permits the use of limestone costing about $4-6/metric ton rather
than lime costing about $30-40/metric ton at greater than 90% utiliza-
tion (or about l.lx stoichiometric quantity), which is comparable to
that normally obtained with lime.
• It further reduces the overall costs of scrubbing by substantially
reducing the quantity of by-product sludge produced.
Another major area being studied at Shawnee during the advanced test
program is that of forced oxidation of sulfite to sulfate (gypsum) . This
concept was developed in the IERL-RTP pilot plant and includes both a
staged scrubber configuration and forced oxidation within a single scrubber.
Both configurations have the potential for obtaining essentially complete
oxidation of the solids to gypsum and an accompanying improvement in solids
settling characteristics and dewatering properties. However, each method
has possible advantages and disadvantages compared with the other, and both
will be evaluated further on the larger units at Shawnee. Forced oxidation
in a single scrubber is certainly a more simplified approach requiring less
equipment than staged scrubbing and can probably be more easily controlled
from an operational point of view. However, this method applies only to
limestone systems while the staged scrubbing configuration can be used- with
either lime or limestone. The staged scrubber configuration has the added
advantage of also being able to obtain very high (about 95%) alkali utili-
zation simultaneously with the forced oxidation to gypsum. The benefit-S
and significance of high alkali utilization, especially for limestone
operation, were pointed out in the preceding paragraph.
It has been estimated that the application of forced oxidation in the
staged scrubber configuration (compared to conventional limestone scrub-
bing) can reduce limestone requirements by roughly one-third and can
reduce the volume of by-product sludge generated by more than 50%. An
208
-------
economic sensitivity analysis made earlier in the program by TVA indicated
that sludge disposal costs represent a significant portion of the overall
costs for lime/limestone scrubbing, and the large quantities of waste
sludge that must be disposed of have long been one of the major objections
of the utility industry to these types of flue gas desulfurization (FGD)
processes. Obviously, reductions of this magnitude in both limestone
requirements and waste sludge generation would have a large impact on the
economics of scrubbing and, hopefully, on wider acceptance and application
by industry of lime/limestone scrubbing.
Other promising areas* which have been or will be studied at Shawnee during
the advanced test program include:
1. Addition of MgO to increase the SO. removal efficiency and to
force operation into the gypsum unsaturated mode. From the test
results to date, it appears virtually certain that the S02
removal efficiency can be substantially improved by MgO addition
using either lime or limestone feed. Higher MgO addition rates
are required for limestone than for lime, but the S0» removal
efficiency can be increased using either alkali from a normal 80%
removal to 95% removal or greater, with other operating para-
meters (L/G, percent solids recirculated, scrubber type, pressure
drop, etc.) remaining the same. However, MgO addition did not
always result in a gypsum subsaturated operation, and under some
conditions the scaling potential was actually increased. The
reasons for this, certain other apparent inconsistencies in the
data, and how these types of systems can be controlled are not
yet fully understood. Efforts are being continued to satisfac-
* torily resolve these important questions.
2. Demonstration of variable load operation on the venturi/spray
tower system with lime. The gas rate was adjusted hourly to
follow the boiler load over a range of 60-150 MW while the inlet
SO„ also varied from 1400 to 5000 pptn. Sustained runs were made
at conditions probably much more severe than would be encountered
at most operating plants with no problems of reliability or
control of the system. Variable load operation will also be
included during the long term reliability runs planned on each
scrubber system (one with lime and one with limestone) in which
the most promising concepts will be combined for the extended
reliability demonstration runs near the end of the current program.
3. Operation with and without fly ash in the system has also been
made with both lime and limestone. The major differences noted
were in the solids settling rate and the solids dewatering
properties. This was true for both lime and limestone operation
*These are discussed in more detail in References 1-3.
209
-------
with filter cake solids averaging 40-50% without fly ash compared
to 50-60% with fly ash in the system. When operating without fly
ash, there was also a greater tendency for the clarifier overflow
to be turbid, and on occasion flocculent addition to the clari-
fier was even required to maintain a clear overflow for return to
the scrubber loop. Some difference was also noted in the pH and
SO. removal efficiency at a given stoichiometric feed rate. The
direction of the difference was opposite to that expected based
on the high pH of the effluent from the regular plant fly ash
pond which normally has a pH of 10.5-11.0. Without fly ash in
the system the pH averaged about 0.2 pH units higher, and there-
fore the SO. removal was also about 5% higher. This is attributed
to a higher leaching rate for the acidic components in the fly
ash compared to the basic components. In separate tests it was
shown that upon standing, Shawnee fly ash/water mixtures first
exhibit a lower pH, and with time the pH gradually increases. It
is probable that operation without fly ash would lead to lower
erosion rates of nozzles, piping, TCA spheres, etc. However,
operation has not yet been for extended periods to properly
assess this possibility.
4. Factorial tests conducted with both lime and limestone have been
very useful in developing and checking the accuracy of the design
models for scale-up to full-scale units. The models have also
been useful in developing the computerized Shawnee data base and
the design/economic study computer program. When completed the
design/economic study computer program will be one of the most
useful tools for industry, vendors, and engineering firms pro-
duced from the Shawnee program. It will readily permit compari-
sons of capital and lifetime operating costs for essentially all
of the equipment and process variations studied during the Shawnee
program. The computer program may also be modified to include
other types of scrubbers and processes to even further expand its
applicability and usefulness.
5. Field evaluation studies to assess techniques for the disposal
of power plant flue gas cleaning wastes^ are also being conducted
at Shawnee in parallel with the regular program. Waste material
(sludge) produced during operation of the test facility using
both lime and limestone as absorbents has been placed in six
small disposal ponds on the plant site. Three of the ponds
contain untreated wastes; each of the three remaining ponds
contain waste that was chemically treated by one of the three
contractors (Dravo, Chemfix, and IU Conversion Systems) who
presently offer such commercial fixation processes. Test samples
of treated and untreated wastes, ground water, surface water,
leachate, and soil cores are being analyzed to evaluate the
environmental acceptability of current disposal technology.
210
-------
Results to date indicate that the total dissolved solids in
leachate from untreated wastes increase after filling, reaching a
peak value equivalent to the input liquor, and then steadily
diminish. Leachates from treated wastes exhibit similar charac-
teristics, except the peak value is approximately half the level
of the total dissolved solids (TDS) in the input liquor. In
addition to a reduction in the leachate TDS, the treated materials
also exhibit greatly improved structural stability and, for at
least one of the processes, greatly reduced permeability. (These
improved physical properties would allow management of disposal
operations such that leachate generation can be minimized or
eliminated.) Ground water monitoring has shown no change at-
tributable to the ponds. Monitoring will be continued during the
balance of the Shawnee program.
6. Peripheral types of studies are also underway or are planned
during the Shawnee program. These will not be discussed in
detail but include the following (reports listed in the reference
section are available for more detail):
a.
A corrosion/erosion study and a mechanical components
evaluation program conducted by TVA. (ongoing)
b. Flue gas characterization study to determine the effect of
changes in major operating variables on particulate mass
loading and particle size distribution, gaseous and solid
sulfates, and on mist eliminator efficiency. (ongoing)
c. Future studies of alternative scrubber internals, automatic
feed control, minimum energy requirements, maximum S07
removal efficiency, and variation of limestone type and
size. (planned)
211
-------
SUMMARY AND CONCLUSIONS
With dramatic increases in the use of high sulfur coal reserves being a
near certainty in the Administration's developing energy policy, particu-
larly in the eastern half of the country, the widespread use of FGD will
play an important role in near term SO control strategies. Expanded coal
use must be accompanied by sound air and water pollution control measures
to minimize threats to human health and to avoid widespread environmental
damage.
As the FGD technology most commonly accepted and used by domestic utilities
for coal-fired boiler applications, lime/limestone processes are especially
important to the success of a policy to rapidly expand coal use. Based on
the most recent estimate, 30 FGD systems are presently installed and an
additional 94 are in various stages of planning or construction. These
124 units represent almost 50,000 MW of electric generating capacity, most
of which will be in service by the early 1980's. Of these, 78 have already
selected lime/limestone processes and an additional 25 have not yet selected
a process type.
The major concerns of the utility industry to date regarding lime/limestone
scrubbing have centered around scaling and plugging potential, the large
quantities of by-product sludge generated, and the high costs (capital and
operating) of scrubbing. It is toward these areas of concern that the
Shawnee program has been directed.
The Shawnee program has made major contributions toward improvement of lime
and limestone scrubbing technology in the areas of reliability, variable
load operation, system control, sludge disposal techniques, and process
economics. The most significant results to date include:
• Demonstration has shown that conventional lime/limestone systems can
be operated reliably. Two separate reliability problems have been
identified—scaling and soft, mud-type solids deposits, and methods to
control each have been demonstrated.
• Soft, mud-type solids deposition was shown to be a strong function of
alkali utilization. At high alkali utilization (greater than about
85 percent) these solids are much more easily removed, and very infre-
quent intermittent fresh water wash is adequate for their complete
removal in restricted areas such as the mist eliminator where accumu-
lation can lead to plugging problems.
212
-------
• Several equipment or process variations were demonstrated to improve
alkali utilization. This is particularly significant for limestone
where alkali utilization is typically about 60-70% and can be
increased to 85-95%, or comparable to that normally obtained with
lime. This not only improves reliability, but also reduces costs
by permitting the use of a much less expensive alkali feed material
and by substantially reducing the quantity of by-product sludge
produced.
• The ability to operate during widely varying boiler load and inlet
SO. concentrations for extended periods was demonstrated on the
venturi/spray tower system using lime with no reliability or system
control problems. This has long been a major concern of utilities
for lime/limestone FGD systems.
• Addition of MgO to either lime or limestone systems has shown that
a substantial increase in S0_ removal efficiency can be obtained,
and also indicates a good potential for forcing operation into the
gypsum unsaturated mode. However, further work is needed to fully
understand how to design and control such a system for subsaturated
operation to avoid potential problems.
• The initial results of staged scrubber forced oxidation studies at
Shawnee appear very promising. The ability to simultaneously
obtain both high alkali utilization (over 90%) and essentially
complete oxidation of sulfite to sulfate (gypsum) opens up numerous
possibilities for further reducing costs and quantity of sludge
produced while maintaining or even improving operating reliability.
• Data generated during the factorial testing have been very useful
in developing and checking the accuracy of the design models for
scale-up to full-scale lime/limestone systems. The models have
also been useful in developing other valuable tools for industry
such as the computerized Shawnee data base and the design/economic
study computer program. Furthermore, all of these will have increased
value as they are expanded to include additional data and other
equipment and process variations.
Although substantial progress and significant improvement have been made
over the past several years in lime/limestone scrubbing, to be of practical
use to industry, results of research and development efforts must be
accepted and applied by the utilities and FGD system vendors. Acceptance
and application of the Shawnee results by several of the major FGD
system vendors has recently been more apparent, but this area of techno-
logy transfer to commercial application by utilities has, unfortunately,
been sluggish. Consequently, a positive, applications-oriented program
involving a higher degree of participation and coordination by EPRI, the
utility industry, and FGD system vendors is now being considered as a
means of overcoming the apparent reluctance to accept and apply pilot
plant and prototype results to commercial size units. The Shawnee pro-
totype test facility will in all likelihood play an important role in
this program continuation.
213
-------
REFERENCES
1. Bechtel Corporation, "EPA Alkali Scrubbing Test Facility: Summary of
Testing Through October 1974." EPA-650./2-75-047 (NTIS No. PB 244-901),
June 1975. The results of the initial test program from March 1972 to
October 1974 are presented in this report.
2. Bechtel Corporation, "EPA Alkali Scrubbing Test Facility: Advanced
Program - First Progress Report." EPA-600/2-75-050 (NTIS No. PB 245-
279/AS), September 1975. The results of advanced testing from October
1974 through April 1975 are presented in this report.
3. Bechtel Corporation, "EPA Alkali Scrubbing Test Facility: Advanced
Program - Second Progress Report." EPA-600/7-76-008 (NTIS No. PB
258-783/AS), September 1976. The results of advanced testing from
June 1975 through mid-February 1976 are presented in this report.
4. Aerospace Corporation, "Disposal of Flue Gas Cleaning Wastes: EPA
Shawnee Field Evaluation Initial Report." EPA-600/2-76-070 (NTIS No.
PB 251-876/AS), March 1976. The results of the field evaluation
program from September 1974 through July 1975 are presented in this
report.
5. EPA Technology Transfer Capsule Report, "First Progress Report:
Limestone Wet-Scrubbing Test Results at the EPA Alkali Scrubbing Test
Facility."
6. EPA Technology Transfer Capsule Report, "Second Progress Report:
Lime/Limestone Wet-Scrubbing Test Results at the EPA Alkali Scrubbing
Test Facility."
7. EPA Technology Transfer Capsule Report, "Third Progress Report:
Lime/Limestone We*-Scrubbing Test Results at the EPA Alkali Scrubbing
Test Facility," EPA-625/2-76-010.
214
-------
FIELD MEASUREMENTS OF PARTICLE SIZE DISTRIBUTION WITH
INERTIAL SIZING DEVICES*
N. G. Bulgakova, L. Ya. Gradus, and S. S. Yankovskiy
Abstract
The authors describe the design, principal of operation, and calculation
of units used in the U.S.S.R. for determining the dispersion composition of
particles under industrial conditions. The following are considered: im-
pactors, two-cyclone separator, and rotational analyser. Procedure is given
for impactor calibration that is based on the graphoanalytic method for de-
termining the parameters of the fractional characteristic of stages starting
with the known parameters of polydispersed dust. Results are presented from
the comparison of three methods with respect to each other under laboratory
and industrial conditions. These are also compared with the classical methods
for dispersion analysis involving samples taken beforehand. It is shown that
the error of inertial methods in determining particle size under industrial
conditions does not exceed +_ 15 percent with respect to the average median
diameter. A method is recommended for evaluating the reproducibility of the
results obtained with one unit under identical conditions.
Efficiency of. dust collection by electrostatic precipitators is influ-
enced by particles 'having a diameter less than 5 ym. Traditional methods of
particle sizing use a dust sample on a filter and then disperse it in a liquid
media (sedimentation column or in Coulter counter) or in a gas stream (in
Gonnel device or BAHCO centrifuge). In the course of these processes aggre-
gates may disintegrate into component particles and this distorts test results.
On the contrary, very small particles (less than 1 or 2 ym) may aggregate in
the liquid media in the course of analysis, which leads to overestimating of
real particle sizes.
In order to avoid the above-mentioned faults of the traditional particle
sizing methods, several devices have been developed recently. In these devices
an aerosol is fractionated during the sampling itself. These devices include:
impactors, in which separation process takes place in the space of small dimen-
sions, each stage consisting of a nozzle with deposition surface placed opposite
it; cyclone separators, which allow considerable amount of fractional partic-
ulate to accumulate; and the rotary dust analyzer, in which aerosol sample is
separated into fractions when passed through rotating cylinder, with particles
depositing on the inner walls of this cylinder.
There are two main types of impactors: the first has converging nozzles
(Brink impactor) and the second has nozzles in the form of openings in a thin
plate (Andersen impactor and NIIOGAZ impactors used in the U.S.S.R.).
The advantage of the Andersen Impactor lies particularly in the fact
that according to Stokes Law, particles are exposed to a considerably lesser
blowing effect of flow due to the small opening diameter. On the other hand,
*For further information regarding the material in this paper, please
contact I.K. Reshidov, U.S.S.R. 113105, Moscow, M-105, 1-st Nagatinsky Pass.,
6, NIIOGAZ.
215
-------
these devices are suitable only for low dust concentration (up to 1 g/m3) be-
cause of the clogging of small diameter openings by particulate matter. The
Brink impactor permits sampling at dust concentration up to 10 g/m3-
The construction of the NIIOGAZ impactor is based on the Andersen prin-
ciple of a set of perforated disks provided with openings serving as nozzles
for the next stage and with substrate-filled grooves relating to the previous
stage to minimize dimensions of the device and to minimize the weight of the
particulate settled, both opposite the nozzles and on the Internal surfaces.
In order to retain particles deposited opposite a nozzle, the surface is
usually greased with a thin silicon oil layer. In NIIOGAZ impactors a layer
of lubricant (ref. 1) consisting of a mixture of fine powder with highly-dis-
tilled oil Can allowable temperature of heating is 60° to 70° C) is used.
Under higher temperatures a layer of thermostable fibrous material (ref. 2)
able to retain deposited particles is inserted in the groves on the plates.
A particle size distribution is determined from the deposit mass on dif-
ferent stages of a device. When calculating this distribution a fractional
efficiency curve for each stage is approximated by a vertical line passing
through cut-off diameter d™. Value of d™ is determined from Stokes rela-
tion (ref. 3):
.2 _ MD _.R _ MD3^ R
a50 " V~ btK50 " ~Q btK50
where StRcn = value of Stokes number referred to dcn,
bu bU
V = velocity of gas flow in the nozzle,
P = density of particles,
M = gas viscosity,
D = nozzle diameter,
Q = gas flow rate through the device.
As d5Q is constant for impactor stages of a fixed geometry, equation (1) can be
used for caluclation of (d50)exp> referred to experimental conditions, by the
formula using a value of (d50)cal being found as a result of calibration:
exp = Cd50)cal A M_
where A = = const.
(Here Qcal , Mca] , and Pca] are gas flow rate, its viscosity, and density par-
ticles under calibration conditions.)
Fractional characteristics and d5Q values for separate impactor stages,
which were obtained by calibration and with the use of a thin grease layer,
cannot immediately coincide with that obtained for impactors provided with
216
-------
glass fiber substrate or oil-impregnated powder. A method for comparing the
calculated and experimental values of d™ was developed. According to this
method particles are deposited on the same substrate and in the same amount as
during field tests. The method is based on placing two identical calibrated
disks in pairs and on the assumption that both fractional characteristics and
particle size distribution curve are governed by lognormal law.
Two parameters of fractional characteristic (i.e., cut-off diameter d™
and standard deviation afr) can be obtained with the aid of two parameters of
particle size distribution (i.e., mass median diameter dp and standard devi-
ation a ) and of the efficiency values found for the first (ru) and second
(no) identical stages. For this a special nomogram is used. On this nomo-
gram two families'of curves corresponding to equal values of (n-i) and of (rip),
respectively, are plotted and with abscissa and ordinate axes values of
Igo llgofr and Ig(d50/dp/)lgo , respectively, placed. The unknown parameters
of the fractional characteristic d5Q and af are found as the abscissa and
ordinate values of the crossing point of two curves corresponding to ru and
n2 measured during experiments.
The cyclone separator includes two impactor stages for preliminary col-
lection of the coarsest particles and two cyclones of 32 and 16 mm diameter
respectively with a filter at the outlet. For the calculation of particle
size distribution presented as two parameters of the lognormal distribution
(d and a ), three nonograms with equiefficiency curves for each separation
stage are used (ref. 4). On the nomogram axes, corresponding values of d
and o are put. Parameters of the particle size distribution can be found as
an abscissa and an ordinate of the crossing point of the three curves corre-
sponding to the values of the efficiency for each stage, being measured in
the course of experiment.
Nomograms were plotted by calibrating the device stages.
The rotationary dust analyzer consists of the rotating cylinder enclosed
in a casing. For convenient usage, cylindrical thin-walled sections whose
length increases downstream are inserted. These sections, made of aluminum
foil, are to be weighed before and after sampling. Coarse particles are
settled on the first stages and the smaller ones on the next stages.
At the device outlet a fiberglass filter is placed. Particle size dis-
tribution is calculated using the weight gains of the dusted sections with the
aid of the Avdoev method (ref. 5). Rotational device calibration was per-
formed by passing monodispersive glass spheres of different diameters through
the device.
The comparison of results obtained by the different inertlal devices as
well as by common methods based on preliminary dust sampling was carried out
under laboratory and field conditions.
217
-------
In one series of such investigations the inertia! devices were placed
into stack and, in another one, the sample was air-suspended by a fluidized
bed laboratory dust feeder followed by an orifice. The feeder is filled with
a mixture of sand and powder. Air-suspended dust was additionally disaggre-
gated when passing through the narrow slit orifice with the velocity of_150
m/s to 200 s/s. The unbranched aerosol stream flowed through the fraction-
ating device.
Under field conditions, samples were taken by impactor, cyclone separator,
and/or rotationary analyzer simultaneously. The inlet sections of sample
probes placed in a duct were situated approximately on the same plane at the
distance of 100 mm from one another. The results of comparative measurements
of particle size distribution are presented as points with coordinates corre-
sponding to mass median diameter d and standard deviation a (in all cases
size distribution was approximated by lognormal law).
Cyclone separator, impactor, rotating analyzer, air separation (in Gonnel
apparatus), and liquid sedimentation (including pipette method) were compared;
the results are shown in figure 1.
Four illustration points referring to the same sample having been analyzed
by different methods are connected by straight lines.
As can be seen in figure 1 there is a good agreement between mass median
diameter and standard deviation values, which was measured by inertial sizing
devices and air separation method (in Gonnell device for fly ash of Chere-
petskaya GRES (power station) and for magnesite dust).
On the other hand, disagreement between mass median diameter values for
dolomite and zinc oxide dusts having approximately the same values of standard
deviation is large. The agreement in the first two cases can be explained by
the fact that the dust is loose and its mean particle size is greater than
10 urn. Such dust easily disaggregates during the air separation.
The mentioned disagreement between the analyses resulting for the dust
sampled directly from the stack and for resuspended dust confirms the assump-
tion that coarse dust moves in ducts in aggregated form. The disagreement of
the result for zinc oxide and dolomite dust is explained by the particle ag-
gregate in Gonnell apparatus. Thus, the Gonnell method gives overestimated
particle sizes. This confirms a well-known fact that only particles larger
than 5 to 7 ym may be processed by Gonnell separation.
An agreement between results obtained by liquid sedimentation and cyclone
device in Nizhne-Turinskaya GRES is explained by the fact that coarse fly ash
particles with diameters of more than 40 ym remain unagglomerated both in the
gas stream and in liquid sedimentation.
For carbide furnace dusts, sedimentation analysis for equal geometrical
mean size leads to more narrow particle size distribution (the small standard
deviation value) than three cyclone separators and impactor. The narrowing of
distribution of the finest and coarsest fractions reflects an underestimating.
Underestimated content of the smallest fraction can be explained by its co-
agulation in the course of sedimentation and the underestimated content of the
218
-------
i
0,8
i ,
0.6
0,4
0,2
I
\
I
6 8 10
20
40
60 80 d.
Figure 1. Comparison of size distribution data for particulates and mists
by several methods.
1 - dolomite dust (at ESP outlet)
2 — particulates of blast-furnace gases
(at carside furnace outlet)
3 — Cherepetsk power station ash
4 — Lower-Turin power station ash
5 — zinc particulates of sulfur acid production
6 — magnesia dust (at cyclone outlet)
7 - quartz particulate matter
8 - coal dust
9 — tarry mist
o - cyclone separator
A— air separation
— pipette method
o - NIIOGAZ impactor
— liquid sedimentation
A— microscopy
219
-------
0,8
0,6
0,4
0,2
c>o
/
1-
N
8
1 2 4 6 8 10 20 40 80 (f
)
Figure 2. Comparison of results of aerosol particle size distribution
using inertial devices.
1 — electrolysis particulates
2 - dolomite dust (at the ESP inlet)
3 — Baltic power station ash
4 — hydrophobic chalk
5-corund M10
6 — mixture of spores and likopodia
7 — dried droplets of sodium salt solutions
8 — mist of transformator oil
9 - dioctilphtalate mist
10 —corund M7
11-corundM20
o - cyclone separator
o — rotary analyzer
o — impactor
220
-------
largest fractions is due to the destruction of the aggregates in the course
of sedimentation. In cyclone separator and impactor these aggregates remain
unchanged.
In figure 2 the results of simultaneous measurements of particle size dis-
tribution by impactor, cyclones, and rotary analyzer conducted in ducts are
shown. According to the results obtained, mass median diameters differ from
one another by not more than +_ 15 percent. The disagreement in standard devi-
ation is not more than 20 percent.
In order to evaluate the reproducibility of results obtained by the same
inertial device, measurements with the corund powder were conducted under lab-
oratory conditions. As an example, the results of five impactor measurements
are presented in figure 3. A scattering area can be limited by two straight
lines having the same angle of incline. In this case a scattering zone can be
presented by only one parameter, i.e., d (a = const). The latter can be con-
sidered as a measure of results scattering. Thus, an error can be presented
as a deflection +_ Adp from the mean value d . For example, according to fig-
ure 3, we have d,, + Ad., = 5.7 + 0.5 ym. This method can be used for evalua-
p — p —
tion of reproducibility of results obtained by the same device under the same
conditions.
95
90
80
60
40
20
10
10
Figure 3. Variation of results for 5 size corund powder
distributions obtained under equal conditions
using the NIIOGAZ impactor.
221
-------
REFERENCES
1. A. A. Rusanov and S. S. Yankovskiy, "Greasing in Order to Hold Particles
on a Hard Surface," Author Certificate No. 197344 issued 17,04,67.
2. Yu. V. Abrosimov, A. A. Rusanov, and S. S. Yankovskiy, "Multistage
Impactor for Weight Dispersion Analysis," Author Certificate No. 261772
issued 28,06,68.
3. A. Berner, "Staub-Reinhalt," Luft, 1972, No. 8.
4. N. 6. Bulgakova and S. S. Yankovskiy, "Graph Method for Determining the
Fractional Efficiency Curve of Cyclones and Dispersion of Dust," Theo-
retical Bases of Chemical Technology, Moscow, 1971, No. 5.
5. A. A. Rusanov and S. S. Yankovskiy, "Impactors for Determining the Dis-
persion Composition of Industrial Dust," Literature Survey of TSNIITEnef-
tekhim, Moscos, 1970.
BIBLIOGRAPHY
Bulgakova, N. G., and S. S. Yankovskiy, "Cyclone Separator for the Dis-
persion Analysis of Industrial Dust, Express-Information," Series KhM-14,
No. 2, TSINTIkhimneftemash, Moscow, 1976.
Skryabin, G. M., "Rotational Analyzers of the Dispersion Composition of
Industrial Dust," Literature Survey of TSINTIkhimneftemash, Moscow, 1973.
Yankovskiy, S.S., and L. Ya. Gradus, Calculating and Calibrating NIIOGAZ
Impactors with Flat Stages, Scientific Technical Collection of Works
"Industrial and Sanitary Gas Cleaning," Moscow, TSINTIkhimneftemash,
1975, No. 3.
Bulgakova, N. G., L. Ya. Gradus, and S. S. Yankovskiy, Low Rate Dust
Feeder, Scientific-Technical Collection of Works, "Industrial and Sani-
tary Gas Cleaning," Moscow, TSINTIkhimneftemash, 1973, No. 3.
222
-------
SIZING TECHNIQUES FOR SUBMICRON PARTICLES
Wallace B. Smith and D. Bruce Harris*
Abstract
The size distribution of submicron particles suspended in process streams
can be measured with, cascade impactors3 electrical mobility analyzers,, or dif-
fusion batteries and condensation nuclei counters. If samples are to be taken
for analysis of composition^ then the impactor would be used. On sources where
the size distribution is stable3 the electrical mobility analyzer yields data
quickly and conveniently. If temporal variations in the particle-size distri-
bution are of interest^ the diffusional system is most useful.
The electrical mobility analyzer and diffusional system both require that
most process streams be cooled and diluted before the sample is taken. The
sample conditioning procedure is difficult and special care must be taken to
avoid serious errors. Line losses due to diffusion and turbulent deposition
limit the useful size range to 0.01 to 2 ym diameter.
Data are shown for several pollution sources where these methods were
used in conjunction with ordinary cascade impactors to obtain particle-size
distributions from 0. 01 to over 10 wn diameter.
INTRODUCTION
The size distribution of submicron particles suspended in process streams
can be measured with cascade impactors, electrical mobility analyzers, or dif-
fusion batteries and condensation nuclei counters. If samples are to be taken
for analysis of composition, then the impactor would be used. On sources
where the size distribution is stable, the electrical mobility analyzer yields
data quickly and conveniently. If temporal variations in the particle-size
distribution are of interest, the diffusional system is most useful.
The electrical mobility analyzer and diffusional system both require that
most process streams be cooled and diluted before the sample is taken. The
sample conditioning procedure is difficult and special care must be taken to
avoid serious errors. Line losses due to diffusion and turbulent deposition
limit the useful size range to 0.01 to 2 (j"> diameter.
In this paper, each of the three sizing methods is described; a sample
extraction and dilution system is also described; and some experimental re-
sults are shown.
*Dr. Smith is with the Southern Research Institute in Birmingham, Ala-
bama, and Mr. Harris is with the Environmental Protection Agency, Industrial
Environmental Research Laboratory, Research Triangle Park, North Carolina.
-------
SAMPLING PROCEDURES
Cascade Impactors
It is possible to extend the sizing capability of cascade impactors to
submicron particles by operating the device at pressures of 0.01 to 0.1 atmos-
pheres. If all operating parameters except the pressure are held constant, the
cut point, or D50, is inversely proportional to the slip correction factor:
'so
a L - %.
Since C increases rapidly with pressure, cut points of 0.02 urn or less can be
obtained. Pilat (ref. 1) has developed and tested a low pressure impactor for
sampling from process streams.
Figure 1 shows the sampling train used by Pilat in his experiments. Two
impactors are operated in series. The first impactor is a conventional design
with cut points from about 0.3 to 20 urn diameter. The second impactor is
operated at reduced pressure with cut points from about 0.03 to 0.2 pm dia-
meter. The sampling train contains a low pressure drop condenser; a .double
vane, leakless, high vacuum pump; a control box with pressure gauges, thermo-
couple pyrometers, and valves; and a dry gas meter. A 90-mm diameter filter
holder is used downstream from the second impactor. The maximum flow rate is
approximately 50 1/min.
Figure 2 shows typical data reported by Pilat for tests made at the
outlet of an electrostatic precipitator installed on a coal-fired electric
generating plant.
Sample Extraction and Dilution
Although in situ sampling is preferred, this is not possible with the
electrical and diffusional apparatus, which are designed for laboratory use.
Figure 3 is a schematic diagram of an aerosol extraction-dilution system which
has been used successfully on several field tests. A 15 1/min sample flow is
removed nonisokinetically from the process stream through a rigid probe, and a
heated, flexible hose into a heated chamber that contains a cyclone, sample
metering orifices, and diffusional adsorbers. Large particles are removed by
the cyclone and the excess air is vented to atmosphere. The low flow rate
sample, which is to be diluted, passes through a calibrated orifice into the
diluter. If the formation of a fine aerosol by condensation is observed, the
sample must first be passed through a series of diffusional absorbers contain-
ing activated charcoal. A Polonium-210 (500 mC) ionizer is used in the dilu-
ter to neutralize any electric charge that may exist on the particles. 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 turbulent
mixing zone. At a downstream point, after adequate mixing has occurred, the
diluted sample is extracted and conveyed to the sizing instrument. Sample and
dilution airflow rates are controlled by means of two bleed valves.
224
-------
40
10
o
E
of
LU
o
cc
1.0
0.1
0.02
i i i i
I ' ' I
I ' ' ' I '
i I , , ! . . I . . .
0.1 1.0 10.0 40 80
PERCENTAGE SMALLER, by weight
99.0
Figure 1. Impactor sampling train for sampling from 0.02 to 20
diameter particles (ref. 1).
225
-------
TIME
AVERAGING
CHAMBER
ISJ
r--------:Li;
-|H.-----------v.k
PROCESS EXHAUST LINE
CHARGE NEUTRALIZER
CYCLONE
ORIFICE WITH BALL AND SOCKET
JOINTS FOR QUICK RELEASE
SOX ABSORBERS (OPTIONAL)
HEATED INSULATED BOX
RECIRCULATED CLEAN, DRY, DSLUTION AIR
FILTER BLEED NO. 2
COOLING COIL
PRESSURE
BALANCING
LINE
BLEED NO. 1
MANOMETER
Figure 2. Size distributions measured at outlet of electrostatic
precipitator on 175 MW coal-fired utility power genera-
ting plant (ref. 1).
-------
10,000
1,000 —
O
p
_l
5
UJ
O BASE AT POINT 1
O BASE AT POINT 1
COLLISON
SPRAYER
O BASE AT POINT 2 - COLLISON
O BASE AT POINT 2 • SPRAYER
SAMPLE ORIFICE DESIGNATIONS
.059 . .042 . .029
100
100 1,000
CALCULATED DILUTION FACTOR
10,000
Figure 3. A sample extraction-dilution system for submicron particles.
227
-------
Figure 4 is a graph showing calibration data for the sample extraction-
dilution system. It can be seen that the calculated and measured dilution
agree very well for 0.15 urn diameter particles. Aerosol dilution is required
for both the diffusional and electrical mobility systems described in the
following paragraphs.
Sizing Particles by Diffusion
Diffusion bateries may consist of long, narrow, rectangular channels in
parallel, a cluster of small diameter tubes (ref. 2), or a series of screens
(ref. 3). Figure 5 shows a typical rectangular channel diffusion battery
which contains 98 channels, each 0.1 cm wide, 10 cm high, and 48 cm long. As
the sample aerosol moves in streamline flow through the channels, the parti-
cles diffuse to the walls at a predictable rate, which depends on the particle
size, flow rate, and diffusion battery geometry.
Variations in the number of channels and aerosol flow rate are used as
means of measuring the number of particles in a selected size range. Figure 6
is a block diagram showing a typical field test system. It is only necessary
to measure the total concentration of particles at the inlet and outlet of
each diffusion battery configuration in order to calculate the particle-size
distribution. Figure 7 shows the theoretical performance curves for several
configurations at an aerosol flow rate of 6 1/min.
Particle concentration measurements are made with condensation nuclei
(CN) counters. These counters are essentially automated cloud chambers. They
operate by drawing the aerosol through a humidification chamber, where it
becomes saturated for ambient temperature and pressure conditions. The aero-
sol is then pulled through a rotary'valve into an expansion chamber where
adiabatic expansion produces a super-saturated condition causing water drop-
lets to grow on the particulate in the aerosol. The droplet growth rate of
the smaller droplets is more rapid than the rate of increase for the larger
particles. Thus, after several milliseconds, a more or less monodisperse
aerosol of water droplets is formed in the chamber. These droplets are
greater than 1 urn diameter and are detected by light scattering techniques.
The response of the CN counter is independent of particle size; thus, the
instrument is said to have a linear sensitivity with respect to particle size.
This uniform sensitivity is also linear in concentration for particle concen-
trations up to about 10s particle/ cm3.
Figure 8 shows typical data taken at a power plant with cascade impactors
and the diffusional system.
Sizing Particles by Electrical Mobility (ref. 4)
The electrical aerosol analyzer (EAA) operates on the concept that there
is a monotonic decrease in electrical mobility with increasing particle size
for submicron particles that are exposed to a homogeneous ion field.
The EAA is shown schematically in figure 9. As a vacuum pump draws the
aerosol through the analyzer, a corona generated by a high voltage wire within
the charging section gives the sample a positive electrical charge.
228
-------
CHANNEL DIMENSIONS
MULTI CHANNEL BATTERY
Figure 4. Calculated dilution versus true dilution for the sample extraction-
dilution system, 0.15 ym diameter particles.
229
-------
ANTI- PULSATION
DEVICE
SAMPLE FROM
DILUTER _
ANTI-
PULSATION
DEVICE —*
COUNTER
^
I
D.B.
CN COUNTER
O.B. 2
D.B. 3
RETURN TO
DILUTER
O.B. 4
O.B. 5
RETURN
TO OILUTER
Figure 5. A typical rectangular channel diffusion battery.
230
-------
100
90
80
70
Z 60
O
< 50
cc
uj 40
m
°- 30
20
10
0
0.
01
J_
0
10
20
30
40
50 <
60 S
70
80
90
100
0.02 0.03 0.04 0.05 0.1
PARTICLE DIAMETER, jum
0.2
0.3 0.4 0.5
Figure 6. Block diagram of a diffusional sizing system.
231
-------
OJ
fSJ
1012
I
a"
§"
z
•a
1010
109
108
I I I
I I I
A A A
0.01
OPTICAL DIFFUSIONAL
IMPACTOR
I I I
I I I
0.1
1.0
PARTICLE DIAMETER,
"I I I-
I I I
10.0
Figure 7. Theoretical diffusion battery penetration curves. D.B. 1
consists of 13 channels, 0.1 x 10 x 48 cm. D.B. 2 consists
of 98 channels of the same dimensions.
-------
CONTROL MODULE
ANALYZER OUTPUT SIGNAL-
OATA READ COHmttD - -
CYCLE START COMMAND -
CYCLE RESIT COMMAND-
AEROSOL FLOWttETER READOUT
CHARJtR CURRENT READOUT
CHARGER VOLTAGE READOUT
ITOMATIC HtGH VOLTAGE CONTROL AND READOUT
ELECTROMETER (ANALYZER CURRENT) RCACOUT
•TOTAL FLOWfCTCR READOUT
N>
OJ
Otl
-•> EITCRNAL
-fc DATA
—• ACQUISITION
— SYSTEM
TO VACUUM PUMP
Figure 8. Particle-size distribution measured at a power plant by
diffusional analysis and cascade impactors (E.S.P. Outlet).
-------
1013::
in
i? i
MIO11-:
.ID"5-
a
a Iti^r
§ 10*4
icf= =
'!'
COPPER SMELTER
O ELECTRICAL MOBILITY DATA
• CASCADE IMPACTOR DATA
POWER PLANT
A ELECTRICAL MOBILITY DATA
A CASCADE IMPACTOR DATA
50% CONFIDENCE INTERVALS ARE
SHOWN WHEN LARGER THAN PLOTTING
SYMBOL
i
F ^™
'I,
10
,-5
i i i i i in| 1—i M i nil 1—i i i i mi
10
-1
10P
101
PARTICLE DIAMETER (MICROMETERS)
Figure 9. Schematic of the electrical aerosol analyzer (ref. 4).
234
-------
BCURA
CYCLONE
MARK III
IMPACTOR
MARK IV
IMPACTOR
STAGE
PRESSURE
TAPS
DRY GAS
METER
Figure 10. Particle-size distributions measured, by the electrical mobility
method and cascade impactors (E.S.P. Outlet).
235
-------
The charged aerosol flows from the charger to the analyzing tube section,
which is an annular volume of aerosol surrounding a cylinder of clean air. A
metal rod, to which a variable, negative voltage can be applied, passes axial-
ly 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.
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 is related directly to the number of particles in the aerosol
between the two discrete particle sizes. A total of 11 voltage steps divide
the 0.032 to 1.0 p size range of the instrument into 10 equal logarithmic size
intervals. Different size intervals can be programmed via an optional plug-in
memory card.
The concentration limits of the EAA are 1 to 1,000 pg/m3.
Figure 10 shows field test data taken at the outlet of electrostatic pre-
cipitators operating on the effluents of a coal-fired power boiler and a
copper smelter.
SUMMARY
Three methods are now used in the United States to size submicron parti-
cles for control device evaluation and emission studies. Any of these methods
can be used in conjunction with ordinary cascade impactors to define particle-
size distributions in industrial aerosols from 0.01 to over 10 pm. Although
the methods are limited, sufficient data have been obtained to study the
efficiency of several control devices for collecting submicron particles.
REFERENCES
1. M. J. Pilat, G. M. Fioretti, and E. B. Powell, "Sizing of 0.02 to 20
Micron Particles Emitted From Coal-Fired Power Boiler With Cascade Im-
pactors," presented at APCA-PNWID Meeting, November 20, 1975, Vancouver,
B. C.
2. A. J. Breslin, S. F. Guggenheim, and A. C. George, "Compact High Effi-
ciency Diffusion Batteries," Staub (in English), 31(8), 1971, pp. 1-5.
3. D. Sinclair. "A Portable Diffusion Battery," Am. Ind. Hygiene Assoc. J.,
36(1), 1975, pp. 39-42. "
4. B. Y. H. Liu, K. T. Whitby, and D. Y. H. Pui, "A Portable Electrical
Aerosol Analyzer for Size Distribution Measurements of Submicron Aero-
sols," presented at the 66th annual meeting of the Air Pollution Control
Association, Paper No. 73-283(1973).
236
-------
DUST RESISTIVITY AND REVERSE CORONA
FORMATION IN ELECTROSTATIC PRECIPITATORS*
I. K. Reshidov, Yu. I. Sanayev, I. A. Kizim
Abstract
In this paper the authors study the reasons for the appearance of back
oorona "In electrostatic precipitators. 'Procedures are described for measuring
dust resistivity under industrial conditions. Two modifications are given
of a device for measuring resistivity developed at the Semibratov branch of
NIIOGAZ that make it possible to form a dust layer in the clearance between
comb-type electrodes in a gas duct or directly on the precipitator electrode.
Basic methods are studied for reducing back corona intensity by using design
factors and conditioning.
One of the major problems in electrical pas cleaning is to increase col-
lection efficiency of high resistivity dust. A great number of studies are
being undertaken by Soviet and foreign scientists in this respect.
It is known that passing of corona current through an electrode covered
with a high resistivity dust layer is accompanied by a voltage drop across the
layer Ui. Ui can be approximately expressed as:
Ui = EI • d = Pv-J-d (1)
where E! = field intensity, v/m
d = thickness of ash layer, m
P = ash specific resistivity, ohm/cm
J = density of corona current passing through ash layer, a/m2.
Due to the voltage drop in the dust layer, the voltage in the interelec-
trode space decreases. This results 'in a reduction of the corona current and
in a deterioration of dust precipitation. When dust specific resistivity is
high enough, EI reaches such a value, at which sparkover occurs and corona
discharge of the reverse sign appears in the layer.
A method to detect reverse corona in ESP was suggested in the U.S.S.R.
(ref. 1). It lies in comparing volt/ampere characteristics taken by
increasing and decreasing of the voltage on the electrodes.
Criterium for the existence of back corona is the excess of corona
current of descending curve over the current of ascending curve. Reverse
corona time lag can account for the hysteresis character of volt/ampere
characteristics.
*For further information regarding the material in this paper, please
contact I. K. Reshidov, U.S.S.R.'113105, Moscow, M-105, 1-st Magatinsky Pass.,
6, NIIOGAZ.
237
-------
This method is now widely used when industrial electrostatic precipitators
collecting high resistivity dust are investigated.
Because back corona is highly dependent on dust specific resistivity, it
is very important to study factors influencing the value of specific resistivity.
Besides, investigations of high resistivity dust precipitation aimed at fore- ^
casting ESP performance efficiency require the true value of specific resistivity
to be known.
The value of dust specific resistivity is influenced by the particle size
distribution and the density of the ash layer, the electric field intensity,
and conditions of taking measurements.
Presently, there are several methods to measure specific resistivity
value that basically differ in the way the dust layer is deposited. Dust
layer is brought on the electrode usually by hand or in the field of corona
discharge. There are installations where dust deposition is done mechanically
by filtration through metalIceramics or by high-efficiency cyclones of small
diameter.
To measure dust specific resistivity under industrial conditions, an
apparatus of the "cyclone" type is applied (ref. 2). It consists of a small-
sized, high-efficiency cyclone and resistivity transducer located in cyclone's
hopper. The transducer's measurement system consists of coaxial cylindrical
electrodes. An instrument for measuring resistivity is used as a secondary
one.
It was found experimentally that the value of specific resistivity is
considerably dependent on the degree of dust layer packing, level of applied
voltage, and density of corona current that passes through the dust layer
(ref. 3).
Simm (ref. 4) thinks that dust specific resistivity should be measured at
currents proceeding sparkover of the layer. Experimental data relating the
specific resistivity of an ash from near-Moscow coal to the corona current
density passing through a layer (ref. 5), shows that high measurements errors
can appear when the value of P is determined by the instrument for resistivity
measurements when the electrode is put on the ash layer. Such measurements do
not allow one to judge the critical value of specific resistivity, in the best
case they can give an idea about the order of magnitude of the measured value.
That is why measurement to determine the critical value of P , at which reverse
corona can appear, should be conducted at current densities close to values
typical for industrial ESP.
As a method to determine specific resistivity of dust, the one dealing
with dust deposition due to electric field in corona discharge is preferable.
To reduce measurement errors connected with the fact that electrodes are put
on the layer and the layer packs, the layer should be of a given thickness, as
is done, for example, in a Lurgi instrument (ref. 6).
For research purposes, NIIOGAZ is using two modifications of the instru-
ment for dust resistivity measurements.
238
-------
The measuring part of the instrument for specific resistivity measurement
(figure 1) consists of two electrodes located in one plane forming a comb-type
gap. When high voltage is supplied, the gap is filled with dust. The value
of specific resistivity is measured when high voltage is cut off.
The second modification of the instrument (figure 2) allows one to measure
specific resistivity of the dust layer formed directly on the precipitating
electrode of the ESP under investigation.
Figure 1. The measuring part of the device for
evaluation of electrical dust resistivity.
239
-------
VIEW - A
B • B
Figure 2. A scheme of the device for measuring electrical
resistivity in the active zone of the ESP.
240
-------
Application of these instruments allowed us to obtain comparable data
when the electrostatic precipitators utilized in various industry branches
were tested. Table 1 gives an example of measuring specific resistivity.
It was found out experimentally that the most intense reverse corona occur
in the last electric fields of ESP. This can be connected with the fact that
the largest or conductive particles (underburnt matter) are collected in the
first ESP fields. That is why it is very interesting to see the relationship
between particles size distribution or other physical and chemical properties
of dust and its conductivity. For that purpose, experiments were undertaken
to determine dust conductivity at the outlet of pilot-commercial ESP where re-
verse corona is not formed and at the outlet of ESP DGPN-55-2 type under the
conditions of reverse corona.
Comparison of data given in table 1 shows that dust resistivity at the
outlet of pilot precipitator (OPE) exceeds the ash resistivity at the inlet
by more than an order of magnitude.
Table 1. Measuring Specific Resistivity
Place and con-
ditions of
measurements
Mean specific
resistivity
value, ohm/m
Mean-square Coefficient
error, of variation,
ohm/m %
At the inlet to the
pilot ESP (OPE)
t = 135° C
At the outlet of the
pilot ESP (OPE)
t = 130° C
At the inlet to ESP
DGPN-type
t = 150° C
At the outlet of
ESP DGPN-type
t = 145° C
2.87 x 109
4.4 x 1010
8.3 x 109
2.8 x 1010
0.25 x 109
0.37 x 1010
0.7 x 109
0.3 x 1010
8.9
8.4
8.5
10.7
241
-------
Figure 3 gives particle size distribution at the outlet and inlet to ESP.
Here one can see considerable difference in particle size. But in this case
it is rather difficult to account for the difference in dust resistivity due
only to the difference of particle size distribution while it s known that
there is a considerable quantity of underburnt matter in ash A-III-type. The
measurement gave such values of underburnt matter:
at the inlet - 16.9 percent,
collected - 20 percent,
at the outlet - 24 percent.
Since there is a considerable difference in the underburnt matter content
at the inlet and outlet of the apparatus, conductivity of ash could not be
identically connected with particle size distribution.
From this point of view, interesting data is found with the measurements
of resistivity of ash from near-Moscow coal. The content of underburnt matter
in this ash at the inlet is 0.78 percent, in the collected portion it is 0.8
percent, i.e., practically the same and very small. Mean values of ash resis-
tivity at the inlet and outlet of ESP were respectively 8.3 x 1011 ohm/cm and
2.8 x 1012 ohm/cm.
From the above data, one can see a clear connection between granulometric
composition of dust and the value of its specific resistivity. Thus, change
of particles' median diameter from 20 to 5 (jm caused more than a threefold
growth of dust resistivity.
Let us consider the main methods to reduce reverse corona intensity.
Critical field intensity E in the dust layer is determined by the product of
ash specific resistivity and corona current density. Therefore, there are two
main ways to reduce E , and thus reverse corona intensity: (1) reduce ash
specific resistivity, or (2) reduce corona current density.
The value of ash specific resistivity is greatly influenced by tempera-
ture, humidity of gases, and S03 content. It is seen from the relationship of
ash specific resistivity to temperature of exhausted gases (relationship has a
steep maximum), that ash resistivity can be reduced both by lowering and ele-
vating temperature.
Figure 4 shows how the temperature influences gas cleaning efficiency
when the ash from Ekibastuzsk coal is collected. This method is not always
suitable in the industrial conditions, since the optimal temperature of the
gases exhausted from the boiler is not determined only by ESP performance.
Therefore, it is expedient to consider other methods intensifying ESPs1 per-
formance when high-resistivity ash is collected.
Conditioning can be provided by feeding steam, spraying water, or moist-
ening raw material, and also by introducing chemical agents. Gas conditioning
by moistening is widely spread for the cases when natural moisture content in
the gas is low and the gas temperature is relatively high. It was found out
that introduction of moisture increases the electrical stability.
242
-------
98
95
90
80
70
JS 60
o
V 50
+••
* 40
•3 30
20
10
5
4 5
10
20
30 40 50
Particle diameter 6'ju
Figure 3. Size distribution of ash particles in the pilot ESP-
1 - at inlet/ 2 - at outlet
243
-------
rss
£
0.998
O997
0.996
0.995
0.994
0.992
099
OMB
098
0.97
0.96
0.95
0,94
0.92
09
OS -
O7 -
O6 -
OS -
O4 -
O2 .
95
100
105
110
115
120
125
130
135 140
145
150° C
Figure 4. Gas cleaning degree as a function of temperature at collection
of ekibastuzsk coal ash.
-------
But conditioning by moistening at power stations causes difficulties be-
cause it requires a considerable consumption of steam or water. Besides, at
the power stations at sufficiently low temperatures of exhaust gases it is
impossible to obtain complete evaporation of water. That causes fallout of
liquid droplets in gas ducts, ESPs, fume pumps, and sealing of their surfaces,
causing corrosion.
An effective means to increase gas cleaning efficiency is to use chemical
conditioning to reduce specific resistivity value. The way that an electro-
static precipitator operates to remove ash from Ekibastuzsk coal when gas is
conditioned with ammonia was investigated. When gas velocity was 2.45 m/sec
and 30 ppm of ammonia was introduced, cleaning efficiency increased from 96.5
to 98 percent and outlet dust concentration decreased 1.6 times. When in flue
gases before the pilot ESP, 8-32 ppm of ammonia was added at a gas velocity of
1.5 m/sec and cleaning efficiency was raised from 88.4 to 94.4-98.3 percent
(figure 5).
Experiments have shown that beyond 20-25 ppm, a further increase in ammo-
nia concentration does not lead to an increase in ash precipitation efficiency,
i.e., that this concentration should be optimal to condition flue gases for
ESP that remove high resistivity ash of Ekibastuzsk coal.
At this ammonia concentration, outlet dust concentration decreased more
than sixfold from the normal 0.8 g/m3. Gas cleaning efficiency increased due
to reduction of reverse corona intensity. That can be confirmed by the change
of electrical parameters of ESP and by the ash specific resistivity measured
in the course of tests. When ammonia was introduced, ash resistivity at the
inlet to the ESP reduced by three orders to 6 x 107 ohm/m. Gas conditioning
by ammonia allowed a twofold increase in the precipitation rate parameter.
Economic expendiency of ammonia conditioning was proven at the electro-
static precipitators on a 300-MW unit of Troitzkaya Power Station.
Performance of precipitators for removal of high-resistivity dust from
clinker kilns for dry cement production is accompanied by intense reverse
corona. Conditioning of gases by water was applied to increase collection
efficiency. When the gases' humidity increased from 4 to 10 percent, voltage
in the ESP became 10-12 kV higher. Specific resistivity measured by an instru-
ment of ICP-type/dust precipitation on a comb-shaped electrode in the electric
field decreased, when moistened by 1-1.5 orders, to 9 x 108 ohm/m.
Reverse corona intensity that is indirectly determined by loop square
between ascending and descending volt/ampere characteristics decreased. Gas
cleaning efficiency increased from 93 to 98.5 percent.
Although gas conditioning was effective in increasing gas cleaning effi-
ciency in this case, the required efficiency was not obtained. To increase
collection efficiency of high resistivity dusts, design methods were studied,
alongside the above mentioned ones, that could be called technological.
An investigation of belt-needle electrodes with various needle lengths
was conducted at a pilot-commercial ESP. This investigation showed that re-
duction of needle length caused a decrease of discharge current, elevation of
working voltage, and reduction of corona discharge power. One of the possible
245
-------
ro
£»
cr>
o
c
(0
0)
10 20
Rate of ammonia consumption
30
•jrf
Figure 5. Dependence of gas cleaning degree, corona discharge power,
current and corona voltage on ammonia rate.
-------
ways to even the electric field and reduce specific corona currents, besides
the above-mentioned, is to turn belt-needle electrodes 90° so that the needles
are oriented in the direction of gas flow.
Comparison of field equipotentials shows that when needle elements are
turned in the direction of gas flow, considerable decrease of maximal and mean
corona current density occurs. That allows reduction of corona current in-
tensity.
Complex application of various methods is an operative measure to increase
efficiency of ESPs collecting high resistivity dust.
REFERENCES
1. I. K. Reshidov, I. G. Yermilov, and B. V. Zolotaryev, Author's Certificate
No. 284973.
2. A. A. Rusanov, I. I. Urbakh, and A. V. Anastasiadi, "Cleaning of Flue
Gases in Industrial Energetics," Energy, 1969.
3. I. K. Reshidov, "Peculiarities of Electrostatic Precipitators1 Performance
for High Resistivity Dust Collection," Proceedings, Scientific Technical
Conference on Industrial Gas Cleaning, Yaroslavl1, 1969.
4. W. Simm, Untersuchungen Uber das Ruckepruhen bei der elektrischen
Staubabscheidung, Che.-Ing.-Ted.n.. No. 1, 1959, pp. 43-49.
5. V. I. Levitov and I. K. Reshidov, "How to Apply the Automatically Con-
trolled Method of Probe Characteristics to some Problems of Electrical
Gas Cleaning," Strong Electric Fields in Technological Processes, Ed.Ac.
Popkov, V. I., Energy, 1969.
6. H. G. Edshold, Eine Messerrichtung zur Bestimmung des spezifischen
elektrischen Staubwiderstands, Staub, No. 1, 1966, pp. 11-14.
247
-------
A REVIEW OF THE INFLUENCE OF FLY ASH COMPOSITION
ON RESISTIVITY
R. E. Bickelhaupt*
Abstract
This paper reviews the influence of fly ash composition on surface
and volume resistivity. The pertinent literature of the past fifteen
years regarding this subject is included in the review. Both practical
and academic viewpoints with respect to the resistivity—ash composition
relationship are considered.
The author subscribes to the idea that the conduction mechanism for
fly ash is analogous to that for silicate glass. Therefore with respect
to the ash composition^ the alkali metals present are of greatest importance
as charge carriers. To emphasise this point and to illustrate that the
presence of sulfuric acid in the environment or an adsorbed film containing
charge carriers is not required to produce a reasonable level of conduction3
additional laboratory data are presented.
INTRODUCTION
Interest in the subject of resistivity is maintained because of the
importance of this material property with respect to the dry electrostatic
collection of fly ash. It is a key characteristic utilized in sizing
precipitators, troubleshooting inefficient operation, and anticipating
problems associated with changes in coal supply.
Several factors either independently or in combination affect the
magnitude of the resistivity value. These include: specific surface,
particle size distribution, ash layer porosity, environmental chemical
composition, ash layer field strength, temperature, and ash chemical
composition. When considering a general case within a narrow temperature
range for ashes produced from low-sulfur coals, the fly ash chemical
composition is the dominant factor. For a given set of conditions, a
four order of magnitude change in resistivity as a function of fly ash
composition has been observed. The relationship between resistivity and
ash composition is also important with respect to developing an under-
standing of the conduction mechanism, the selection of conditioning
agents, and in the prediction of resistivity for ashes from heretofore
unburned coals and cleaned coals.
In this paper, fly ash is defined as essentially an inorganic,
amorphous, chemically heterogeneous substance produced from the fusion
and volatilization of the accessory coal minerals. Physically, the
material is primarily spherical in shape, has a typical mass median
diameter of 5 to 15 y'm, and has a pycnometric density value of 2.0 to
3.0 g/cm3.
*Southern Research Institute, Birmingham, Alabama.
248
-------
Presentation and Definition of Resistivity Data
Data are usually illustrated by the logarithm of resistivity as a
function of reciprocal absolute temperature. This is desirable because
over a large temperature range, resistivity is very sensitive to this
variable and pragmatically one is interested in both cold - and hot-side
precipitators. This type of presentation is also suitable for discussion
of conduction mechanisms and for distinguishing between volume and
surface resistivity. Figure 1 is an example of laboratory resistivity
data shown in the above manner.
The linear portion of the curve FE and its extension to D is usually
referred to as volume resistivity. That part of the curve AB and its
extension to C is defined as surface resistivity, and although plotted
as a function of temperature, it is dependent on the combined effect of
temperature and environment. Phenomenologically one can accept the
segment BE as the resistivity resultant from a parallel circuit comprised
of the volume and surface resistivities. Several authors (refs. 1, 2,
and 3) have advanced mathematical expressions that define the shape of
the curve ABEF. These expressions emphasize the particulate nature of
the fly ash, define the effect of environmental partial pressures and
temperature, and point out the need to know the fly ash material parameters
to be able to calculate the magnitude of resistivity. It was suggested
above that the chemical composition of the ash is the most important
material parameter. In this paper an attempt is made to review the
relationship between resistivity and fly ash composition from a mechanis-
tic as well as an empirical viewpoint.
Volume Resistivity
Until several years ago, the volume conduction mechanism was thought
to be electronic and particulates were considered intrinsic semiconductors
(refs. 4 and 5) or impure insulators. Efforts were made to identify the
specific electronic mechanism and little attention was given to fly ash
composition. This viewpoint was probably encouraged by the assumption
that the ashes were principally crystalline, by the observation that log
p vs 1/T is linear and that the numerical value of experimental activation
energy is commensurate with semiconduction, and by the tendencies of
investigators to explain conduction for all amterials subjected to
electrostatic precipitation using one mechanism.
About 1970 there was evidence that investigators (refs. 6, 7, and
8) were observing the pronounced reduction in resistivity for fly ashes
containing large amounts of sodium. In late 1971 and early 1972 an
effort (refs. 9 and 10) was made to determine the mechanism for volume
conduction. It was hypothesized that the conduction mechanism was ionic
and that the charge carriers were alkali metal ions.
The high temperature, linear portion of the resistivity-temperature
curve can be expressed in terms of an Arrhenlus equation,
P = P0 exp (0/kT), (1)
249
-------
1013
1000/T(°K)
°C
°F
TEMPERATURE
Figure 1. Typical laboratory resistivity—temperature data.
250
-------
where
p = resistivity
PQ = a complex material parameter,
0 = experimental activation energy.
K = Boltzmann constant, and
T = absolute temperature.
The denominator of the expression for the complex material parameter,
p contains a term for the number of mobile charge carriers. Therefore,
it was expected that the resistivity would be inversely proportional to
the atomic concentration of alkali metal ions at a specific temperature.
Figure 2 shows this relationship to be true. Resistivity was determined
at -v350% C using several fly ashes having generally similar chemical
composition but differing significantly in amount of Na20. The best
illustration of the data was shown by a plot of log resistivity versus
the log of the combined atomic concentration of lithium and sodium.
Transference experiments were used to demonstrate the mass transfer
associated with ionic conduction and to identify the charge carriers.
Some of the results of these experiments are shown in figure 3. An
exuded mass of material collected in the region of the negative electrode
was found to contain about 50 percent sodium and crystallographically
contained compounds associated with the reaction of sodium and the
ambient laboratory environment. Chemical analyses of the individual ash
layers qualitatively indicate that the three alkali metal ions migrated.
Because of the combined effect of concentration present and mobility,
the sodium transfer was most obvious. Gravimetrically it was observed
that a weight loss occurred at the anode and a gain at the cathode. The
values are of reasonable magnitude for the amount of electrical energy
passed, assuming that sodium was the principal migrating species. Tha
data and observations shown in figures 2 and 3 were offered as proof of
the hypothesis that the volume conduction mechanism is ionic and that
the alkali metal ions, principally sodium, serve as charge carriers.
The following mechanism (ref. 11) satisfying electrostatic neutrality
was suggested. Alkali metals migrate to the cathode accepting electrons
which causes a depletion of these ions at the anode. At the anode in
the region of depleted alkali, unbridged and unassociated oxygen ions
under the influence of an intense electric field migrate to the anode to
form oxygen molecules with the release of electrons.
In the subsequent years, Selle et al., (refs. 12 and 13) and Bickel-
haupt (ref. 14) have conducted extensive research programs associated
with the resistivity-fly ash composition relationship. Selle has shown
the pronounced effect of sodium on resistivity for lignite and subbitumi-
nous coals mined in the western United States. These studies have led
to the development of a correlation that can be used to determine resis-
tivity from ash composition for these coals.
251
-------
(0
101
o
I io«
DC
108
1 FLY ASH A
OB
(FROM REF. 10)
OF
0.1 1.0
CONCENTRATION, LITHIUM + SODIUM
ATOMIC PERCENT
10.0
Figure 2. Representation of high temperature—long time transference results,
252
-------
THREE CONTIGUOUS
ASH LAYERS
CHEMICAL ANALYSIS
EXTRUDED MASS AT
CATHODE - 50% Na
GRAVIMETRIC DATA
Disk
1.
2.
3.
Li2O
0.018
0.01
0.006
Na2O
3.0
2.3
1.4
K2O
0.8
0.7
0.7
+44 mg
- 1 mg
-34 mg
Figure 3. Volume resistivity versus lithium and sodium concentration for
ashes of generally simtlar composition.
253
-------
Bickelhaupt attempted to expand the earlier work on volume resistiv-
ity to include a general cross section of ash compositions produced
throughout this country. These results are summarized in figure 4. The
data show the resistivity at -v-3500 C as a function of the atomic concen-
tration of lithium and sodium for a large number of ashes having diverse
chemical composition. It was observed that the data in the region near
the illustrated curve were produced from ashes having a relatively low
and uniform iron concentration. Those data lying well below the line
between 0.2 and 0.5 atomic percent sodium were related to ashes of
moderate to high iron concentration. If the data were normalized to
constant sodium concentration, it was found that an excellent correlation
existed between resistivity and iron concentration. Others (refs. 6 and
15) have noted this effect. A number of experiments were conducted to
elucidate the effect of iron. The obvious suggestion that iron introduces
an electronic contribution to the total conduction process could not be
demonstrated. Transference experiments using ashes having four levels
of iron concentration yielded useful information. It was found that for
a constant amount of electrical energy passed, the ratio of the alkali
metal migrated to that amount present in the ash increased with increas-
ing iron concentration. As a result, it was speculated that the iron in
the glassy structure of the ash permitted a greater number of the alkali
metals ions to be mobile. This point has never been additionally studied.
As a result of this research, an empirical expression was developed to
estimate resistivity from ash composition for high temperatures. Corre-
spondence has indicated that the expression is reasonably useful.
Emphasizing the role of sodium in the volume conduction process,
evidence (ref. 16) has been recently presented that demonstrates the
potential for sodium as an injected conditioning agent used in a hot-
side precipitator.
Surface Resistivity
Until recent years, the role of ash composition with respect to
surface resistivity received little attention. This was true principal-
ly because the generally accepted mechanism (ref. 17) for surface conduc-
tion does not require the ash composition specifically to be operative.
This electrolytic mechanism suggests that conduction takes place through
an adsorbed film, one or more molecular layers thick, of water and/or
sulfuric acid. In a manner analogous to that for bulk aqueous solutions,
conduction occurs by electrolysis of the adsorbed agents or by a proton
jump processes (ref. 18). In these cases, the hydrogen ion would be the
principal charge carrier. With the exception of the effect a given ash
might have on the adsorption process and the availability of the adsorbed
species to be functional, the chemical composition of the ash was not
considered important to the extent that the material was sometimes
described as inert.
Many investigations have shown that the composition of the fly ash
has a vital role in the surface conduction process. Included are obser-
vations (refs. 7 and 19) of improved precipitator performance when coals
high in sodium concentration were burned, tests (refs. 17 and 20) demon-
strating the effect of sodium as a conditioning agent, and laboratory
studies relating resistivity and ash composition.
254
-------
108
10.0
ATOMIC PERCENT LITHIUM PLUS SODIUM
Figure 4. Volume resistivity versus lithium and sodium concentration for
ashes of generally dissimilar composition.
255
-------
The previously cited works of Bucher (ref. 8) and Selle (refs. 12
and 13) et al. are equally applicable to a discussion of the effect of
ash composition on surface resistivity. These authors have shown that
surface resistivity under constant environmental test conditions decreases
over 3 orders of magnitude as the sodium concentration increases within
the normal limits for ashes produced from Western coals. High coefficients
of correlation were found for relationships between resistivity data and
expressions derived to predict resistivity from ash composition. It was
pointed out that, in the cases of high sodium coals, volatilized sodium
appears on the surface of fly ash as Na2S04 readily available for partici-
pation in the conduction process (ref. 20).
Bickelhaupt (ref. 21) studied surface resistivity as a function of
a broad range of ash compositions to determine a correlation between
these two factors and to establish the role of ash composition with
respect to the conduction process. An illustration of the data acquired
in this study is shown in figure 5. This figure shows the general
dependency of surface resistivity on the atomic concentration of lithium
and sodium. Those data points significantly below the curve in the
region of 0.3 percent lithium and sodium were found to contain 10 to 25
percent Fe203 and 2.3 to 3.9 percent potassium. It was subsequently
determined that resistivity values normalized to a constant lithium plus
sodium concentration could be correlated with either the atomic concen-
tration of iron or iron plus potassium. The choice between these two
correlations was dependent on the iron concentration. Based on these
data, a graphical method of estimating maximum fly ash surface resistivity
was offered.
The effect of iron and potassium was partially clarified using
chemical transference experiments. It was observed that after extended
periods of time under applied electrical field at a temperature at which
only surface conduction was operative, a concentration gradient for the
alkali metals lithium, sodium, and potassium had been established. This
gradient showed the migration of these ions to the negative electrode.
The greater participation of potassium in the conduction process was
found to be sensitive to the greater concentrations of iron in the ash.
It was speculated that, as in the case of volume conduction, the role of
iron is indirect. Ancillary tests revealed an enhanced release of
alkali metal ions for aqueous solutions of ash suggesting that the
presence of iron may decrease the resistance of the glassy ash to attack
by environmental agents. The influence of adsorbed H2S04 and other
condensed agents cannot be overlooked, because the high iron ashes also
possess the higher concentrations of soluble sulfate and the experimental
procedure would not have destroyed these agents. From the transference
experiments, it was estimated that the migration o,f alkali metal charge
carrying ions would account for the major portion of the electrical
energy passed.
In reference 21, an extended description of a surface conduction
mechanism is given that accentuates the role of ash composition. It was
suggested that the conduction mechanism for glassy ash is analogous to
that recorded for silicate glasses (ref. 22). For the general case, one
can visualize an ion exchange process in which a H replaces a Na in
the surface structure of the ash thereby freeing the sodium ion to
256
-------
1014
55
ui
ee
iu
,20
(FROM REF. 21)
Q EASTERN ASH
O WESTERN ASH
H20- Sv/o
SURFACE AREA • 2000 cm'1
0.1 1.0
ATOMIC PERCENTAGE LITHIUM + SODIUM
10.0
Figure 5. Maximum surface resistivity versus lithium and sodium concentration
for ashes of generally dissimilar composition.
257
-------
O WASHED
O WASHED AND ANNEALED, 460°C
A WASHED AND ANNEALED, 750°C
O ANNEALED, 750°C
O ANNEALED, 460°C
1000/T(°K)
°C
°F
TEMPERATURE
Figure 6. The effect of water washing and/or thermal annealing on the
resistivity of an ash.
258
-------
migrate. Also, any condition that would provide a source of mobile
alkali metal ions on the as)i surface will enhance conduction. The
foregoing does not preclude the role of sulfuric acid in the surface
conduction process. However, the role of this natural conditioning
agent within in the scope of the conduction process needs additional
study.
To support the proposal that surface conduction is principally
dependent on the charge carrying alkali metal ions, it was believed
necessary to demonstrate the surface conduction phenomenon in the presence
of air and water alone and in the absence of any natural or inadvertent
surface films or deposits on the ash. Several ashes were washed three
times using distilled water. Each cycle involved extended soak followed
by a rinse during filtration. The ashes were dried, returned to powder
form, and subdivided for various tests. The results were very similar
for all ashes studied, and the data for one are presented in figure 6.
These curves show the data for a given ash subjected to a variety of
treatments and tested under one set of conditions with the exception
that the water-washed specimen, of necessity, had to be tested using
ascending temperature.
The curves suggest that water washing the ash eliminates any conducting
surface films and/or leaches out the alkali metal ions in the immediate
surface of the ash. By thermally annealing in dry air, the washed
specimens seemingly are rejuvenated by either developing an activated
surface with which the resistivity test environment can readily react
and/or by the replacement of the leached alkali metal ions by diffusion
from within the particles. This experiment suggests that a preexisting
conducting film or deposit on the ash is not necessary for the operation
of the suggested mechanism. However, in those cases in which a substantial
amount of the surface conduction is related to adsorbed or condensed
phases, one would not expect the resistivity to return to the value of
an unwashed ash as a result of the thermal anneal.
It is concluded that the influence of chemical composition on the
resistivity of fly ash is quite important, affecting all facets of the
problems associated with resistivity and the dry collection of fly ash.
REFERENCES
1. S. Masuda, "Effects of Temperature and Humidity on the Apparent
Conductivity of High Resistivity Dust," Electrotechn J. of Japan, 7
(108): 1963.
2. K. J. McLean, "Factors Affecting the Resistivity of a Particulate
Layer in Electrostatic Precipitators," J. Air Poll Control Assoc..
26 (9): 866-870, 1976.
3. Ditl Pavel and Robert W. Coughlin, "Improving Efficiency of Electro-
static Precipitation by Physico-chemical Modification of the Electro-
static Resistivity of Flyash," AIChE Journal. 22 (4): 730-736,
1976.
259
-------
4. Harry J. White, Industrial Electrostatic Precipitation. Reading,
Massachusetts: Addison-Wesley Publishing Company, Inc., 1963, pp.
306-309.
5. 0. J. Tassicker, Z. Herceg, and K. J. McLean, "Mechanism of Current
Conduction Through Precipitated Fly Ash," Bulletin #10, Department
of Electrical Engineering, Wallongong University College, The
University of New South Wales, February 1966.
6. C. C. Shale, J. H. Holden, and G. E. Fasching, "Electrical Resistivity
of Fly Ash at Temperatures to 1500MV1 Report 7041, Bureau of
Mines, U.S. Department of Interior, Washington, D.C., March 1968.
7. R. A. Durie, "Investigation of the Electrostatic Precipitation of
Fly Ashes from Coals to be Supplied to the Liddell Power Station,
Part 2, Fuel Research Investigation Report 72. June 1968.
8. W. E. Bucher, "A Study of the Bulk Electrical Resistivity of Fly
Ash from Lignite & Other Western Coals," M. S. Thesis, University
of North Dakota, Grand Forks, North Dakota, December 1970.
9. R. E. Bickelhaupt and G. B. Nichols, "Investigation of the Volume
Electrical Resistivity of Fly Ash from the Sundance and Wabamun
Power Stations," Final Report Project 2865 submitted by Southern
Research Institute, Birmingham, Alabama, to Calgary Power, Ltd.
Calgary, Alberta, Canada, June 30, 1972.
10. R. E. Bickelhaupt, "Electrical Volume Conduction in Fly Ash,"
APCA Journal 24 (3): 251-255 (1974).
11. D. E. Carlson, K. W. Hang, and G. F. Stockdale, "Electrode polar-
ization in alkali-containing glasses," J. Am. Cer. S., 55 (7): 337,
1972.
12. S. J. Selle, P. H. Tufte, and G. H. Gronhovd, "A Study of the
Electrical Resistivity of Fly Ashes from Low-Sulfur Western Coals
Using Various Methods," Paper 72-107 presented at the 65th Meeting
of the Air Pollution Control Association, Miami Beach, Florida,
1972.
13. S. J. Selle, L. L. Hess, and E. A. Sondreal, "Western Fly Ash
Composition as an Indicator of Resistivity and Pilot ESP Removal
Efficiency," Paper 75-02.5 presented at the 68th Meeting of the Air
Pollution Control Association, Boston, Massachusetts, June 15-20,
1975.
14. R. E. Bickelhaupt, "Volume Resistivity - Fly Ash Composition Relation-
ship," Environmental Sc. & Tech.. 9 (4): 336-342, 1975.
15. J. Dalmon and E. Raask, "Resistivity of Particulate Coal Minerals,"
J. Inst. Fuel. 46 (4): 201-205, 1972.
16. A. B. Walker, "Operating Experience with Hot Precipitators on
Western Low Sulfur Coals," A Paper Presented at the American Power
Conference, April 18-20, 1977, Chicago, Illinois.
260
-------
17*. ft. J. White, "Resistivity Problems in Electrostatic Precipitation,"
APCA Journal 24. (4): April 1974.
18. R. A. Robinson and R. H. Stokes, Electrolyte Solutions. London:
Butterworths Scientific Publications, 1955, pp. 116-117.
19. R. H. Holyoak, "Burning Western Coals in Northern Illinois," Paper,
73-WA/Fu-4, presented at the Winter Annual Meeting, ASME, Detroit,
Michigan, November 11-15, 1973.
20. S. J. Selle, "Factors Affecting ESP Performance on Western Coals
and Experience With North Dakota Lignites," Symposium on Particu-
late Control in Energy Processes, EPA - 600/7-76-010, September,
1976, pp. 105-125.
21. R. E. Bickelhaupt, "Surface Resistivity and the Chemical Composition
of Fly Ash," APCA Journal. 25 (2): 148-152, 1975.
22. Robert H. Doremus, Glass Science. John Wiley & Sons, New York,
1973, chapters 12 and 13.
261
-------
TECHNICAL REPORT DATA .
/P/ease read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-78-037
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Second US/USSR Symposium on Particulate Control
5. REPORT DATE
March 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHORIS)
Franklin A. Ayer, Compiler
8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
10. PROGRAM ELEMENT NO.
EHE 624
11. CONTRACT/GRANT NO.
68-02-2612, Task 5
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings; 6-12/77
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES
919/541-2925.
project officer is Dennis C. Drehmel, Mail Drop 61,
^..ABSTRACT
procee(jmgS include papers presented during the symposium, spon-
sored by the Particulate Technology Subgroup of the US/USSR Stationary Source Air
Pollution Technology Working -Group under the US/USSR Environmental Agreement.
The symposium was held at Research Triangle Park, North Carolina, September 25-
October 2, 1977. Papers were presented by Soviet specialists from research design
institutes and industry, and by representatives of US government agencies and the
private sector. Topics included: electrostatic precipitator (ESP) research and appli-
cation, ESP gas flow modeling, ESP rapping and reentrainment , ESP reliability,
ESP modeling, flue gas conditioning, high-temperature ESP application, use of
fabric filters in the US cement industry, emission standards, state-of-the-art of
mist eliminators, mist eliminator testing, particle size distribution measurement
in the micron and submicron ranges , dust resistivity and back-corona formation in
ESPs , and fly ash composition and its effect on resistivity.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution, Dust Control, Gas Flow
Electrostatic Precipitators, Reliability
Mathematical Models, Entrainment
Flue Gases, Treatment, Fabrics
Filters, Cements, Mist, Measurement
Size Determination, Concentration
Electrical Resistivity, Coronas. Fly Ash
Air Pollution Control
Stationary Sources
Particulate, USSR
Rapping
13B, —, 20D
—, 14D
12A, 07D
21B, 14B, HE
13K,11B,04B,-
20C, -. -
FEMET
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (Thispage)
Unclassified
267
22. PRICE
EPA Form 2220-1 (9-73)
262
------- |