v>EPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
EPA-600/7-79-189
August 1979
Electrostatic Precipitators
for Collection of High
Resistivity Ash
Interagency
Energy/Environment
R&D 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 afriaximum 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-600/7-79-189
August 1979
Electrostatic Precipitators for Collection
of High Resistivity Ash
DY
D.H. Pontius, P.V. Bush, and W.B. Smith
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
Contract No. 68-02-2193
Program Element No. EHE624
EPA Project Officer: Leslie E. Sparks
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
EXECUTIVE SUMMARY
This research program included as principal objectives the comparison of
various types of electrode systems for charging fine, high-resistivity dusts,
the investigation of techniques for charging high resistivity dusts in a high
current density corona systeu, performance of a laboratory scale study to
determine the technical feasibility of selected charging systems, and finally
the design, fabrication and testing of a 0.47 m3/sec (1000 acfm) pilot-scale
precharger applicable to a two-stage system for electrostatic precipitation
of high resistivity particulate materials. As a preliminary step, the liter-
ature was reviewed for indications of previous attempts to control back coro-
na resulting from the presence of high resistivity dust in an electrostatic
precipitator. Limited theoretical and experimental investigations were car-
ried out to eliminate impracticable techniques and to develop novel approaches
to the solution of the problem. This work resulted in the derivation of a
new three-electrode particle charging device (precharger) upon which further
developments in this project were based.
The general concept of the three-electrode precharger is that a properly
biased, open mesh screen electrode placed near the grounded plate electrode
in a wire plate system will serve to remove a large portion of the ions
resulting from back corona, while permitting a reasonably high primary corona
current to pass. This concept was tested in a small laboratory device, where
it was found that back corona effects could be controlled sufficiently well to
permit charging of dusts having electrical resistivity"above 1012 ohm-cm to_
levels that could be achieved for low and moderate resistivity dusts
(<5 x 10 ohm-cm) in a conventional corona geometry.
As a consequence of the laboratory scale work, a pilot scale system was
designed and fabricated for testing at a gas volume flowrate of approximately
1000 ACFM. The tests performed on that device demonstrated good charging
results, but also revealed the necessity for improvements in the mechanical
design. Hence a second generation, ruggedized version of the 0.47 m3/sec
charger was designed, constructed and tested. Charging results remained con-
sistent with those of previous tests.
Used with a modified conventional pilot-scale ESP as a second stage
(collector) the precharger was tested as a part of a two-stage system. Meas-
urements of particle size distributions and mass loadings at the inlet and
outlet of the system showed overall collection efficiency above 90% when op-
erated at a specific collection area of 25.2 m2/m3/sec (128 ft2/1000 acfm)
where the dust resistivity was above 1012 ohm-cm.
-------
These tests indicate the feasibility of making substantial size reduc-
tions, with concomitant economic savings, in the fabrication of electrostatic
precipitators applied to the collection of high resistivity dusts.
This report has been submitted in fulfillment of Contract No. 68-02-2193
by Southern Research Institute under the sponsorship of the U. S. Environ-
mental Protection Agency. This report covers a period from September 30,
1976 to July 31, 1978, and work was completed as of September 30, 1978.
ii
-------
CONTENTS
Executive Summary i
Figures iv
Tables . x
Acknowledgements xi
1. Introduction 1
2. Summary and Recommendations 5
Laboratory studies 5
Pilot scale program 6
Recommendations 6
3. Preliminary Studies 8
Description of three-electrode concept 8
Bench-scale tests 9
Laboratory scale precharger 17
Electrode geometry studies 21
4. Pilot Scale Precharger 43
Initial pilot test program 43
Installation of automatic screen voltage control .... 67
Pilot tests at IERL/RTP 70
Summary of tests results 75
Second generation pilot precharger 77
5. Charged Particle Collector 87
6. Engineering and Cost Analysis . . . 99
Estimated costs of full scale precharger —
collector systems 99
Comparisons of costs with conventional precipitators . . 105
References 112
Appendices
A. Investigation of alternate methods A-l
B. Theoretical study of space charge effects B-l
C. Precharger C-l
D. Collector D-l
iii
-------
FIGURES
Number
1 Apparatus used for preliminary evaluation of three-
electrode corona geometry concept 10
2 Current as a function of time for each electrode.
Corona discharge electrode is at -25 kV, and
screen electrode is at -9 kV. Dust laden air
is injected at 1.2 1/min 11
3 Current as a function of time for each electrode.
Corona electrode is at -25 kV. The screen electrode
is at 8.5 kV, which is below the magnitude of potential
required to accept ions resulting from back corona at
the plate. Screen voltage was shifted momentarily to
-9 kV at t = 6 min 13
4 Current as a function of time for corona discharge
electrode and plate electrode as a function of time,
with dust injection. Screen electrode was removed, and
voltage is -25 kV. Losses to oven walls account for
current difference 14
/
5 Current as a function of time for each electrode, with
corona discharge electrode at -22 kV. The screen
electrode was initially set at 8.5 kV, and reduced
to 8.0 kV after 30 min. running time 15
6 Current for each electrode in the three-electrode system.
The vertical dashed lines denote times at which voltage
changes were made to the values indicated. Initially
the discharge electrode voltage was 15 kV and the
screen voltage was 8 kV 16
7 Laboratory scale precharger assembly in the three-
electrode configuration 18
8 Bare plate I-V characteristics at ambient and at 130°C
with corona electrode-to-plate spacing = 8.89 cm 19
9 Comparison of the I-V characteristics of the three-electrode
configuration with grid-to-plate spacings of 1.0 and 2.5 cm.
The corona electrode-to-plate spacing = 8.89 cm, temperature
= 130°C, and the grid current = 0 pA 20
iv
-------
Number
10 Test results of. a three-electrode charger used for
back corona suppression. Temperature = 125°C,
corona electrode-to-plate spacing = 8.9 cm, grid
electrode-to-plate spacing = 2.6 cm, and dust
loading ~ 3.4 g/m3 22
11 Test results of a three-electrode charger used for
back corona suppression. Temperature = 125°C,
corona electrode-to-plate spacing = 8.9 cm, grid
electrode-to-plate spacing =2.6 cm, and dust
loading ~ 3.4 g/m3 23
12 Test results of a three-electrode charger used for
back corona suppression. Temperature = 130°C,
corona electrode-to-plate spacing = 8.9 cm, grid
electrode-to-plate spacing =2.6 cm, and dust
loading = 6.8 g/m3 24
13 Comparison of theoretical and experimental I-V
characteristics for a wire-plate configuration with
11 cm plate width and 9 cm electrode separation 26
14 Comparison of theoretical and experimental I-V
characteristics for a wire-plate configuration with
11 cm plate width and 3 cm electrode separation 27
15 Comparison of theoretical and experimental I-V
characteristics for a wire-plate configuration with
11 cm plate width and 2 cm electrode separation 28
16 Comparison of theoretical and experimental I-V
characteristics for a wire-plate configuration with
5.5 cm plate width and 5 cm electrode spacing v- 29
17 Comparison of theoretical and experimental I-V
characteristics for a wire-plate configuration with
2.75 cm plate width and 9 cm electrode separation 30
18 Comparison of theoretical and experimental I-V
characteristics for a wire-plate configuration with
2.75 cm plate width and 2 cm electrode separation 31
19 Ratio of apparent to actual plate width used to provide
"best theoretical fit for various values of plate
width to electrode separation ratio 32
20 Electrical characteristics of a parallel wire-plate corona
electrode system for rive values of electrode spacing.
Plate width is 11 cm, and wire diameter is 0.25 mm . 33
-------
Number Page
21 Electrical characteristics of a parallel wire-plate corona
electrode system for five values of electrode spacing.
Plate width is 5.5 cm, and wire diameter is 0.25 mm 34
22 Electrical characteristics of a parallel wire-plate corona
electrode system for five values of electrode spacing.
Plate width is 2.75 cm, and wire diameter is 0.25 mm 35
23 Electrical characteristics of a parallel wire-plate corona
electrode system for five values of electrode spacing.
Plate width is 11 cm, and wire diameter is 0.79 mm 36
24 Electrical characteristics of a parallel wire-plate corona
electrode system for five values of electrode spacing.
Plate width is 5.5 cm, and wire diameter is 0.79 mm 37
25 Electrical characteristics of a parallel wire-plate corona
electrode system for five values of electrode spacing.
Plate width is 2.75 cm, and wire diameter is 0.79 mm 38
26 Total corona current for fixed length of barbed wire
discharge electrode as a function of separation
between barbs 40
27 Corona current per disc for a discharge electrode consisting
of discs at various spacings with axes aligned 41
28 Comparison of I-V characteristics for a 0.02 cm wire, a
barbed wire and an array of disc discharge electrodes.
The passive electrode is a 14 cm diameter cylinder for
all three curves 42
29 View of the pilot scale charger on its side 44
30 View of the pilot scale charger electrode configuration 45
31 Corona current vs. corona voltage characteristics for the
precharger with the screen voltage adjusted to maintain
zero screen current 46
32 Results of the dust loading characterization at the pre-
charger inlet with the sandblasting gun at 138 Pa (20 psi) ... 47
33 Results of the dust loading characterization at the pre-
charger inlet with the sandblasting gun at 138 Pa (20 psi) ... 48
34 Results of the dust loading characterization at the pre-
charger inlet with the sandblasting gun at 138 Pa (20 psi) ... 49
VI
-------
Number Page
35 Results of the dust loading characterization at the pre-
charger inlet with the sandblasting gun at 138 Pa (20 psi) . . 50
36 Back corona suppression test results. Test conditions
were: T = 130°C, %H20 =1.2 (by volume), average gas
volume flowrate = .71 m3/sec (1500 ft3/min), dust load-
ing =1.88 g/m3, MMD = 25 pm, E = 3.26 x HTV/m, and
Nt = 1.4 x 1013 sec/m3 51
37 Back corona suppression test results. Test conditions
were: T = 130°C, %H20 =1.2 (by volume), average gas
volume flowrate = .71 m3/sec (1500 ft3/min), dust load-
ing = 1.88 g/m3, MMD = 25 pm, E = 2.72 x 105V/m, and
Nt = 1.3 x 1013 sec/m3 52
38 Back corona suppression test results. Test conditions
were: T = 130°C, %H20 = 1.2 (by volume), average gas
volume flowrate = .71 m3/sec (1500 ft3/min), dust load-
ing = 1.88 g/m3, MMD = 25 Vim, E = 2.61 x 105V/m, and
Nt = 8.8 x 1012 sec/m3 53
39 Back corona suppression test at 130°C, 1.2% H20, p - 1.2 x 1012
fl-cm, j = 9.4 x 105 nA/m2, and the dust loading = 1.88 g/m3. . 55
40 Corona current vs. corona voltage for clean and dirty
electrodes at 130°C, 1.2% H20, and p = 1.2 x 1012ft-cm.
Grid current was held to zero for these measurements 56
41 Back corona suppression test and Q/m measurement at 75°C,
1.2% H20, p = 1.4 x 1012 fi-cm, j = 9.4 x 105 nA/m2,
dust loading = 1.88 g/m3, and Q/m = 9.6 x 10~6 C/g 57
42 Back corona suppression test and Q/m measurement at 100°C,
1.2% H20, j - 9.4 x 10s nA/m2, dust loading = 1.88 g/m3,
and Q/m = 3.0 x 10~6 C/g 60
43 Corona current vs. corona voltage for clean and dirty
electrodes at 100°C and 1.2% H20. Grid current was
maintained at zero for these measurements 61
44 Back corona suppression test at 75°C, 1.2% H20, p = 1.4
x 1012 fi-cm, j = 9.4 x 10"5 nA/m2, ash injection at
276 Pa (40 psi) and manual rapping 62
45 Back corona suppression test and Q/m measurement at 75°C,
1.2% H20, p = 1.4 x 1012 fi-cm, j = 9.4 x 10~5 nA/m2,
ash injection at 276 Pa (40 psi), pneumatic rapping at
552 Pa and Q/m = 8.7 x 10~6 C/g 63
vii
-------
Number Page
46 Back corona suppression test at 100°C, 1.2% H20, j = 9.4 x
10~5 nA/m2, ash injection at 414 Pa (60 psi) and
pneumatic rapping at 552 Pa (80 psi) 64
47 Corona current vs. corona voltage for clean and dirty
electrodes at 100°C and 1.2% HzO. Grid current was
held to zero for these measurements 65
48 Back corona suppression test and Q/m measurement at 100°C,
1.2% H20, j = 9.4 x 10~5 nA/m2, ash injection at 345 Pa
(50 psi) pneumatic rapping at 552 Pa (80 psi) and Q/m
= 2.86 x 10" C/g 66
49 Schematic diagram of the electronic circuit designed to
provide automatic adjustment of the precharger screen
voltage in response to changes in the primary corona
current 69
50 Back corona suppression test with T = 93°C, flowrate =
5.19 m3/sec, and mass loading -1.0 g/m3. The Q/m value
obtained in this test is 1.36 x 10~6 C/g 71
51 Current-voltage characteristics for the precharger clean
and dirty at 93°C 72
52 Back corona suppression test with T = 107°C, flowrate -
5.19 m3/sec, and mass loading = 1.0 g/m3 73
53 0.47 m3/sec (1000 acfm) precharger assembly 77
54 I-V curves of the downstream collector section with dirty
wires and plates, no dust flow, and 149°C (300°F) 79
55 I-V curves of sections 1 and 2 of the downstream collector with
dirty 2,5 cm mesh discharge electrodes, dirty plates, no dust
flow, and 149°C. 80
56 I-V curves of sections 3 and 4 of the downstream collector with
dirty wires, dirty plates, no dust flow and 149°C 81
57 Precharger corona electrode I-V curve with the grid current held
at zero, temperature = 158°C, and gas flowrate = 0.47 m /sec
(1000 acfm) 82
58 Optical particle counter measurement system 84
59 Number of counts vs. particle diameter as observed with the
Climet optical particle counter for the three conditions:
1 - precharger off and collector off, 2 - precharger off
and collector on, and 3 - precharger on and collector on ... 85
viii
-------
Number Page
60 The current-voltage characteristic of five 0.64 cm
diameter wires spaced 9.5 cm from a grounded plate
with a wire-to-wire spacing = 3.81 cm 88
61 The current-voltage characteristics of 0.32 cm
diameter wires spaced 9.5 'cm from a grounded plate
at three wire-to-wire spacings 89
62 Comparison of electrical behavior for various types of
corona discharge electrodes. The wires are in arrays
of five in parallel, spaced at 3.8 cm 91
63 Small pilot scale precipitator assembly. . . 92
64 I-V characteristics of section 1 of the downstream
collector with no gas flow and ambient conditions 93
65 I-V characteristics of section 2 of the downstream
collector with no gas flow and ambient conditions 94
66 I-V characteristics of section 3 of the downstream
collector with no gas flow and ambient conditions 95
67 I-V characteristics of section 4 of the downstream
collector with no gas flow and ambient conditions • • 96
68 I-V characteristics of section 1 through 4 of the
pilot scale downstream collector at ambient condi-
tions. The two gas passages in each section were
electrically connected for these tests 97
69 Theoretical collection efficiency of the pilot scale
precharger-collector combination, plotted as a
function of particle diameter. Charging parameters
are Nt = 8.63 x 1012 sec/m3, Ep - 3.15 x 10s V/m
and T - 348°K. These data correspond to charging
experiments where the dust resistivity was greater
than 1011 fl-cm 103
ix
-------
TABLES
Number Page
1 EAA data analysis for the following test conditions:
T = 75CC, % H20 = 1.2% (volume), p= 1.4 x 1012 ft-cm,
j , = 9.4 x 10s nA/m2, E , , . =2 kV/cm,
Jprecharger ' collector '
dust loading » 1.88 g/m3 .................... 58
2 Test conditions for which successful back corona
suppression was maintained ................... 68
3 Estimated performance of the pilot scale precharger-
collector system ....... . ................ 102
4 Cold ESP cost model ........................ 106
5 Hot ESP cost model ........................ 108
6 Cold S03 conditioned ESP cost model ................ 109
7 Detailed average costs for new cold ESP. ....... ...... 110
8 Comparison of average costs for electrostatic precipitators
collecting high resistivity fly ash .............. • HI
-------
ACKNOWLEDGEMENTS
Technical assistance in the design of an automatic grid voltage control
circuit was provided by Mr. W. J. Steele and Mr. R. N. Coker. Precharger
fabrication was accomplished principally by Mr. R. H. Leopard, under the
supervision of Mr. T. D. Hughes. Lodge-Cottrell Operations, Dresser Indus-
tries, Inc. provided design assistance in connection with the ruggedized ver-
sion of the 1000 ACFM precharger.
The work of Mr. G. Ramsey and Mr. B. Daniel of IERL in helping with the
testing program is appreciated. Finally, we gratefully acknowledge the con-
tinued encouragement and assistance of Dr. L. E. Sparks.
XI
-------
SECTION 1
INTRODUCTION
The presence of high resistivity particulate material in an electrostatic
precipitator tends to degrade the collection efficiency of the system by re-
ducing the rate of particle charging. The problem occurs as a result of a
phenomenon known as "back corona", which arises from electrical breakdown in
the dust layer on the precipitator collection plates. Breakdown of the dust
layer leads to localized field effects capable of producing ionization of gas
molecules near the breakdown sites. Under these conditions, corona discharges
occur at both the corona wire and the grounded plate electrode, resulting in
a bipolar ion current throughout most of the space between electrodes. When
both positive and negative ions are present the particle charging mechanisms
become ineffective, leading to very poor performance of the ESP.
The electric field strength in the dust layer on precipitator collection
plates depends upon the electrical resistivity of the dust and the current
density passing through it. The thickness of the layer is not a primary factor.
Thus, electrical breakdown and the consequent back corona can occur even where
only an extremely thin dust layer exists. It has been demonstrated that me-
chanical cleaning by rapping, scraping or brushing cannot provide a clean
enough metal surface to prevent back corona. -It is therefore unrealistic to
attempt to defeat the effects of high resistivity dust by application of
mechanical plate cleaning techniques.
Several approaches have been developed to deal with the high resistivity
problem in conventional precipitators. Among these are the use of chemical
additives, operation of precipitators at elevated temperatures and the use of
extraordinarily large collecting surfaces relative to the gas volume flowrate.
There are, however, disadvantages associated with each of these methods. The
use of chemical additives entails the expense of providing an injection system,
as well as the cost of maintaining a regular supply of the reagent to be used.
Operation of a precipitator at elevated temperature (350°C to 450°C) presents
engineering difficulties due to thermal stresses and materials considerations.
Insulation costs for "hot-side" precipitators increase the capital outlay
required relative to the expense of installing a conventional ESP. The use of
a very large specific collection area (SCA, ratio of total collecting plate
area to total gas volume flowrate) is a fairly reliable approach, since the
overall effect of back corona is to reduce the efficiency of an ESP. But, for
a given application, the installation cost of an ESP is roughly proportional
to the value of the SCA. Thus, each of the techniques currently employed for
the collection of high resistivity dusts by electrostatic precipitation entails
substantial installation or operating costs above those associated with the
collection of dusts having moderate electrical resistivity.
-------
The principal objective of this project was to investigate possible solu-
tion to the problems associated with the precipitation of high resistivity
particulate matter and to evaluate the results in terms of applicability to
the control of industrial air pollution. The work is based on previous studies
of particle charging with regard to the development of an effective two-stage
electrostatic precipitator system.1
In a two-stage system the charging and collecting functions are separated
inasmuch as it is possible. The first stage, or precharger, is operated at a
relatively high current density to provide a dense ion field for effective
particle charging. The precharger is, physically, a relatively small part of
the system, so it is possible to resort to unusual and relatively expensive
techniques for controlling back corona in the precharger without incurring
prohibitively high costs for the system as a whole. The second stage of the
system serves as a particle collector. The desirable operating parameters in
the collector are high electric field strength and low, uniform, ion current
density at the plate electrodes. The high field provides for maximum migra-
tion velocity, and the low current density permits operation below the thres-
hold for back corona. Operation at zero current density is impractical,
since collection of reentrained particles may require some additional charg-
ing in the collecting stage.
The focus of this research work has been on the development of an effec-
tive precharger. Several approaches were examined and compared theoretically
and experimentally with regard to feasibility of controlling the effects of
back corona in an environment where the ion current density was well above
the threshold for back corona. As a result of this investigation a three-
electrode system was devised, which, after preliminary study, appeared'to I 'be
superior to the other concepts under consideration. Laboratory scale tests
of the three-electrode system supported the preliminary work, and led to the
development of a small pilot scale precharger capable of handling a gas flow-
rate of 0.47 - 0.94 m3/sec (1000 - 2000 acfm).
Tests of the pilot scale precharger were carried out at Southern Research
Institute and at the U.S. Environmental Protection Agency's Industrial Envi-
ronmental Research Laboratory at Research Triangle Park, North Carolina. The
results of these tests showed that the precharger could attain charge levels
on high resistivity particles (>1012 ohm-cm) comparable to those achieved for
particles of moderate resistivity (<5 x 1010 ohm-cm) in a conventional ESP.
Measurements of collection efficiency, using a wire-plate device for the
collector stage, showed a marked improvement in efficiency with the pre-
charger energized, compared with operation of the collector alone.
Since the precharger was designed to demonstrate the feasibility of the
concept under consideration, durability was not emphasized in the design.
Thus, having shown that the three electrode precharger could perform well
with high-resistivity dust, the program was concluded by designing and test-
ing an improved device. The new precharger was made more ruggedly, and all
insulating materials were removed from regions through which the dust laden
gas could flow. Results of the tests were favorable. Charging results were
consistent with those achieved by the first pilot precharger. By using
closely-spaced wires and screen discharge electrodes in the collector stage
-------
to maintain low current densities, it was possible to achieve a collection ef-
ficiency above 90% with a collection area less than 25.6 m2/m3/sec. (130 ft2/
1000 ACFM) for dust having electrical resistivity greater than 1012 ohm-cm.
-------
-------
SECTION 2
SUMMARY AND RECOMMENDATIONS
LABORATORY STUDIES
An investigation of electrode designs and methods for overcoming back
corona was carried out, with the objective of developing a two-stage electro-
static precipitator system capable of collecting high resistivity dusts with
greater efficiency than can be achieved by conventional precipitators. Sev-
eral alternative approaches were subjected to limited theoretical and exper-
imental studies. Among the ideas considered were heated passive electrodes,
a novel technique for injecting liquid or gaseous chemical reagents directly
into the active corona region and various electrode geometries and energiza-
tion schemes. The most promising approach appeared to be a three-electrode
geometry in which a screen electrode is used to trap ions originating from a
back corona discharge, thus preventing those ions from interfering in the
particle charging process.
The electrode arrangement in the new system consists of parallel plates,
between each pair of which is a corona discharge electrode (a barbed wire in
the prototype system), and a pair of open mesh screens, each located in a
plane parallel to a plate and much closer to the plate than to the corona
wire. The screen electrodes are energized at a voltage having the same polar-
ity, but much lower magnitude than the potential of the corona wire. Ions
originating at the corona wire are thus deflected away from the metallic part
of the screen, and pass through the holes on the way to the plate electrode.
When high resistivity dust is introduced into the system some small fraction
of the particles will be deposited on the plate electrode, and back corona
can be generated as in a conventional system. The ions originating from the
back corona discharge are, however, attracted to the screen electrode. Since
most of those ions are trapped by the screen, they are not permitted to inter-
fere with the normal particle charging processes in the principal gas stream.
The back corona effects are thus controlled. No back corona discharge occurs
at the screen electrode because virtually none of the primary corona current
is accepted by the screen.
After the principal features of the three-electrode system were tested
in a bench scale mock-up, a small laboratory scale device was designed and
fabricated for the purpose of evaluating control of back corona and high re-
sistivity particle charging effectiveness in a realistic configuration. In
the laboratory scale precharger Teflon spacers and insulators were used to
maintain the electrodes in their proper relative positions. The device was
exposed to redispersed fly ash heated sufficiently to raise the resistivity
-------
above 1012 ohm-cm. Measurements of charge to mass ratio were made on particles
that had passed through the energized precharger. The results compared favor-
ably with calculated values based on primary corona current only, ignoring
back corona. That is, the charging effectiveness of the device was similar to
what would be expected for an ash having low resistivity in a conventional
wire-plate corona system.
PILOT SCALE PROGRAM
In the next phase of the investigation a larger scale precharger was con-
structed for testing in combination with an existing pilot scale electrostatic
precipitator, which could be used as a downstream collector following the pre-
charger. The system was designed to handle approximately 0.71 m3/sec (1500 acfm)
of simulated flue gas. Included was an automatic control circuit for the volt-
age applied to the screen electrodes. When the precharger was brought into opera-
tion with high resistivity (~1012fi cm) dust loading the effects of the screen
electrodes in controlling back corona were clearly evident. The primary co-
rona current could be maintained at a constant level as the screen current
fluctuated over a wide range in response to the back corona current. Charg-
ing measurements made on particles sampled on the exit side of the precharger
indicated a charge to mass ratio of the order of 2 x 10~6 C/g. Comparisons
were made between the performance of the system in particle collection with
the precharger energized versus the results obtained with the precharger
turned off. Conditions in the downstream collector were maintained as similar
as possible for comparative tests. Collection efficiency was markedly im-
proved by action of the precharger over what could be acieved by the down-
stream collector alone.
Because of mechanical problems in the precharger a second, ruggedized
version was designed and fabricated. The duct-dimensions and electrical spac-
ings were kept the same as in the original device. The testing program for
this device was similar to that described in the above paragraph. Results
of collection efficiency measurements made by use of optical particle counters
and mass trains showed improvements in collection efficiency with the pre-
charger on equivalent approximately to doubling the specific collection1 area
of the ESP serving as the downstream collector, in comparison with similar mea-
surements made on the ESP operating alone. In particular, collection effi-
ciencies above 90 per cent were recorded for the two-stage system operating
at an SCA of 128 on a dust having resistivity of approximately 1012& cm. The
efficiency for the ESP alone under the same conditions was measured at about
70%.
RECOMMENDATIONS
Further developments are required in order to demonstrate that the two-
stage concept can be applied successfully to the requirements of industries
and utilities. The precharger must be complemented by an optimal collecting
device. The small pilot scale system should be tested on an actual pollution
source where electrical resistivity is a problem, and a larger scale system
should be designed and tested to ensure that a practical scale-up is feasible.
-------
Further fundamental studies are also in order to determine whether the
concept can be modified in any way to provide still better particle charging,
and to explore the applicability of such a system to a variety of air pollu-
tion control problems.
-------
SECTION 3
PRELIMINARY STUDIES
In order to evaluate various alternative techniques for counteracting
back corona and space-charge effects in a high current density corona field,
preliminary theoretical studies and limited laboratory tests were carried out.
Among the more promising of the concepts considered in detail were heated pas-
sive electrodes, introduction of chemical conditioning material through a
porous passive electrode, and injection of chemicals directly into the active
corona region at the discharge electrode. The general premise for this study
was that extraordinary means for the control of back corona could be used in
a particle charging device, which could serve as the first stage in a two-
stage electrostatic precipitator (ESP) system. Since saturation charging is
generally reached within a distance of a few inches at ordinary gas velocities
in an ESP the particle charging device, or precharger, would be, physically,
only a small part of the overall two-stage system. Thus the costs of applica-
tion of special techniques or materials in the precharger might be more than
offset by the reduction in collecting area required in the second stage (col-
lector) of the system, due to the enhanced charge on the particles.
In the course of the investigation a novel approach to the control of
back corona was developed. The new concept was based on the use of a third
electrode, whose purpose was to act as a sink for ions generated as a result
of back corona. Because this technique appeared to comprise a more practi-
cable approach to the solution of the back corona problem, it was given prec-
edence for further research, and an application for patent was initiated (U.S.
Serial Number 882,673, dated March 2, 1978). The results of preliminary work
done on the other approaches, mentioned in the above paragraph, are summarized
in Appendix A.
DESCRIPTION OF THREE-ELECTRODE CONCEPT
The basic idea underlying the three-electrode corona system is to capture
the ions resulting from back corona near their source, rather than attempting
to prevent back corona from occurring. Two of the electrodes used in the sys-
tem are the conventional corona discharge and passive electrodes. The third
is a screen electrode placed near the passive electrode.
Separate power supplies are provided for the corona discharge and screen
electrodes. The passive electrode is set at ground potential. Consider, for
example, a two electrode system where the corona discharge electrodes is at a
high negative potential with respect to the grounded passive electrode. Now,
locate an equipotential surface near the passive electrode and insert a con-
ducting screen coincident with that equipotential surface. If the screen volt-
age is set equal to the original potential on the surface the electric field
will be practically undisturned on comparison with the original field. Only the
8
-------
non-zero thickness of the wires in the screen will cause very localized modi-
fications to the field. A corona current originating at the discharge elec-
trode will be distributed such that a fraction of the total current equal to
the ratio of open area to total surface of the screen will reach the passive
electrode. The remainder of the current will be intercepted by the screen.
Now, if the potential on the screen electrode is made more negative the
field near the screen will become distorted in such a way that negative ions
from the discharge electrode will be repelled from the screen wires and forced
toward the open area, through which they can proceed to the plate. If we
introduce high resistivity particulate material into the system it is certain
that depositions will occur on both the plate and the screen electrodes.
Since negative ions from the discharge electrode are being repelled by the
screen it must have a lower current density than the plate, and hence corona
from the screen electrode would probably not occur. If back corona occurs,
the positive ions from the passive electrode would be attracted to the screen
electrode, where many would be captured and removed from the system. If most
of the positive ions resulting from back corona can be captured by the screen
electrode, the ion field between the screen and the discharge electrode would
be essentially unipolar, providing an effective particle charging region.
BENCH-SCALE TESTS
Experimental tests were run to verify the basic concepts involved in the
three-electrode system discussed in the preceding paragraphs. The apparatus
used was as shown in the schematic diagram, Figure 1. The system was enclosed
in an oven maintained at 150°C, and a continuous flow of redispersed fly ash
was introduced. The current for each electrode was monitored separately.
(Discrepancies in current sums can be accounted for by losses to the oven
walls.) Effectiveness of the concept is interpreted in terms of the relative
magnitudes of the three current measurements. When back corona occurs the
plate current should rise significantly. If the screen grid is effective in
removing ions resulting from back corona, there should be a rise in screen
current consistent and commensurate with the rise in plate current, and the
discharge electrode current should remain nearly constant.
Figure 2 shows the results of an experiment where the behavior was near
that predicted. After an overall initial drop in current at all electrodes
(cause uncertain, possibly because of development of a space charge) back co-=
rona apparently set in rapidly. The plate current rose from 50 yA to 350 yA
in about six minutes. The screen current increased quite consistently from
about 5 PA to 250 yA. The corona discharge current rose also, but by less
than 50%, compared with a seven-fold increase in plate current. A disturbing
aspect of the experiment is that there appears to be no tendency toward ap-
proaching a steady-state operating condition. The average current density at
the plate was very high, however, being well over 1000 nA/cm .
In a second experiment, the negative voltage on the screen was reduced
by about 6% so that the screen tended to repel positive ions and accept nega-
tive ions. The behavior of the system was virtually inverted, in agreement
with theoretical expectations. The screen current was opposite from its di-
rection in the previous experiment, and with the apparent onset of back
-------
Ovtn
fly ash Injection
Corona Electrode
Screen
Plate
I
Microammeter
Microammeter
Microommeter
High Voltage
Power Supply
High Voltage
Power Supply
Figure 1. Apparatus used for preliminary evaluation of three-electrode
corona geometry concept.
-------
400
350
300
< 250
0-CORONA CURRENT
D-PLATE CURRENT
A-GRID CURRENT
-50 I 1 1
02468
TIME , min
10
12
Figure 2. Current as a function of time for each electrode. Corona
discharge electrode is at -25 kV, and screen electrode is
at -9 kV. Dust laden air is injected at 1.2 1/min.
11
-------
corona, the plate and discharge electrode currents rose together, while the
screen current remained constant. These effects are shown in Figure 3.
Figure 4 shows the system behavior with the grid electrode removed, which
is quite similar to the result depicted in Figure 3. The rate of increase in
current as back corona apparently develops is approximately the same in the two
experiments.
Another experiment in which the screen voltage was adjusted to accept posi-
tive ions repeated the results obtained in the first. This test was of longer
duration. Again, the screen and plate currents increased simultaneously, as
shown in Figure 5. After about 30 minutes sparking occurred between the screen
and plate, forcing a reduction in screen voltage. When that adjustment was
made the screen and plate currents continued to increase consistently but a
more rapid increase in primary corona current also occurred.
The experiments performed with a three-electrode corona system thus indi-
cated possible utility under some conditions where back corona is present. The
additional degree of freedom resulting from the addition of a third electrode
might complicate electrical control of the system. Improved behavior may be
achieved by optimizing screen wire spacing and screen-to-plate separation. In
a more realistic system the current density at the plate would be more uniform
than in a point-plane apparatus. Under such conditions the peak current den-
sity would be smaller, and back corona effects easier to control.
Further tests of the three-electrode concept were carried out in order to
investigate the possibility of operating at steady-state conditions after the
onset of back corona. The discharge electrode was a sharp point, spaced 3 cm
from a plate electrode. A wire screen electrode, 84% open area and 0.62 cm
wire spacing was located parallel to the plate electrode at a distance of
1.0 cm.
The experiment progressed as shown in Figure 6. Fly ash was injected over
the plate in a dry oven at a temperature of 150°C. With 15kV on the discharge
electrode and 8kV on the screen, a sharp rise in both screen and plate current
occurred after approximately 4 minutes. After about 8 minutes the screen and
plate currents had risen by a factor of about 8, and occasional sparking
occurred. A relatively small change in the current at the discharge electrode
occurred. At this time the screen voltage was reduced to 7.8kV. At t = 10
min. the screen voltage was further reduced to 6.5kV and the corona discharge
electrode voltage was reduced to 13kV. During the following 20 minutes the
primary corona current remained essentially constant, and the current at both
of the other electrodes drifted slowly toward a steady value.
Finally, at t = 30 min., the discharge electrode voltage was returned to
its original value of 15kV. The primary corona current rose slightly and the
current at the other two electrodes settled to a lower value, approximately
three times the discharge electrode current.
Throughout the experiment the screen current followed the variations in
the plate current quite consistently, indicating that ions resulting from back
12
-------
400
350 -
0-CORONA CURRENT
D-PLATE CURRENT
A-GRID CURRENT
9kV
I
6 8
TIME, min
10
12
Figure 3.
Current as a function of time for each electrode. Corona
electrode is at -25 kV. The screen electrode is at 8.5 kV,
which is below the magnitude of potential required to accept
ions resulting from back corona at the plate. Screen volt-
age was shifted momentarily to -9 kV at t = 6 min.
13
-------
600
O-CORONA CURRENT
D-PLATE CURRENT
4 6
TIME , min
Figure 4. Current as a function of time for corona discharge electrode
and plate electrode as a function of time, with dust injec-
tion. Screen electrode was removed, and voltage is -25 kV.
Losses to oven walls account for current difference.
14
-------
0-CORONA CURRENT
D-PLATE CURRENT -
A-GRID CURRENT
TIME, min
Figure 5. Current as a function of time for each electrode, with corona
discharge electrode at -22 kV. The screen voltage was initially
set at 8.5 kV, and reduced to 8.0 kV after 30 min. running time.
15
-------
250-
200-
UJ
or
a:
i>
o
«crttn
V. n = 7.5kV,ond Vdj§c|l=l3KV
,-Vdiich«l5kV
i
i
PASSIVE ELECTRODE
SCREEN ELECTRODE
.DISCHARGE ELECTRODE
i I i i i i
10
20 30
TIME, min
40
50
60
Figure 6. Current for each electrode in the three-electrode system. The
vertical dashed lines denote times at which voltage changes
were made to the values indicated. Initially the discharge
electrode voltage was 15 kV and the screen voltage was 8 kV.
16
-------
corona were intercepted by the screen electrode. Variations in current levels
show a strong dependence upon the voltage applied to the screen electrode.
LABORATORY SCALE PRECHARGER
In order to examine the three-electrode concept in a more realistic con-
figuration a small laboratory scale device was designed in a parallel wire-
plate arrangement. A photograph of the precharger is shown in Figure 7. The
height of the plates in this device is about 30 cm (12 in.), the enclosure is
made of Teflon to isolate the electrodes from external effects. The plate-to-
plate and screen-to-plate spacings were made variable so that the effects of
changing those parameters could be examined. The screen electrodes were per-
forated plates with 0.635 cm hexagonal openings (79% open area).
After preliminary current-voltage (I-V) measurements were made at ambient
conditions, the precharger assembly (Figure 7) was installed in the test sec-
tion of an existing dry wall pilot scale electrostatic precipitator in order
to evaluate its performance under conditions of elevated temperature and dust
loading. The effect of elevated temperature (130°C) is indicated in Figure 8.
The increase in current is probably a direct result of the increased mobility
of ions at higher temperatures.
The three-electrode system with a corona electrode-to-plate spacing of
8.89 cm and a screen electrode-to-plate spacing of 1.0 cm was the initial pre-
charger configuration studied under conditions of both high temperature and
dust loading. The current-voltage relationships for this geometry, when sub-
jected to a dust loading of approximately 3.5 g/m3 of redispersed fly ash
(resistivity of ^1013 Si-cm) at 130?C, revealed that back corona was not con-
trolled. The occurrence of back corona is indicated by a significant rise in
the plate current. If the screen electrode is effective in removing ions re-
sulting from back corona, there should be a similar rise in the screen current,
and the discharge electrode current should remain nearly constant. Failure to
suppress back corona also occurred in an experiment with the corona electrode-
to-plate spacing reduced to 3.81 cm and all other parameters held constant.
The performance of the system seems to be quite sensitive to the position
of the screen relative to the other electrodes. The screen-to-plate separa-
tion was increased from 1.0 cm to 2.5 cm, with the corona electrode-to-plate
separation held at 8.89 cm. The difference in I-V characteristics at the two
spacings is shown in Figure 9. With a dust loading of approximately 3.4 g/m
of redispersed fly ash and at a temperature of 130°C this configuration
controlled back corona temporarily until the screen voltage required to main-
tain a constant discharge electrode current exceeded the value obtainable
with the screen electrode power supply. . Similar results had been encountered
in the intermediate stages of the earlier investigation of the three-electrode
system.
Investigation of the three-electrode geometry laboratory scale charger
continued under various conditions. Tests were conducted at 125°C or 130°C.
In all cases, the corona electrode-to-plate separation was held at 8.89 cm,
and the grid electrode-to-plate spacing was 2.6 cm. The corona electrode
17
-------
Figure 7. Laboratory scale precharger assembly in the three-electrode
configuration.
18
-------
lOV—r
I05
UJ
QC
OC
O
IO
I I I
I '
O D
o a
a
o a
o a
-Ambient
- |30° C
I
I
10
20 30 40
VOLTAGE, kV
I L
50 60 70
Figure 8. Bare plate I-V characteristics at ambient and at 130°C with
corona electrode-to-plate spacing - 8.89 cm.
19
-------
10"
10
u
or
cr
13
O
U
10
Grid-to-plate spacing , cm
o * 1.0
a = 2.5
_L
I
10
20 30 40 50
CORONA VOLTAGE, kV
60
70
Figure 9. Comparison of the I-V characteristics of the three-electrode
configuration with grid-to-plate spacings of 1.0 and 2.5 cm.
The corona electrode-to-plate spacing =8.89 cm, temperature
= 130°C, and the grid current = 0 yA.
20
-------
used was a 0.028 cm diameter stainless steel wire and the grid electrodes used
were perforated plates with 79% open area.
Figure 10 shows the results of one test where the charger was subjected
to a dust loading of approximately 3.4 g/m3 at a temperature of 125°C. The
charger grids and plates were continually rapped with pulsed solenoids at a
rapping frequency of 2 sec~!. Back corona is evident in the sharp increase in
the plate current. The grid voltage was adjusted throughout the experiment to
maintain the corona current at its initial value. In this case, the grid elec-
trode effectively suppressed the back corona for the duration of the test. The
random fluctuations in the grid and plate currents could be a result of uneven
dust feeding, an effect related to the rapping of plates and grids, or some
combination of the two.
A series of experiments was conducted which included a determination of
the charging effectiveness, as well as the back corona suppression capability
of the laboratory scale charger. The particle charging measurements were made
by collecting fly ash on an isolated silver filter which was placed immediately
downstream from the charger. The filter was connected to an electrometer so
that the integrated charge could be monitored for a sample of fly ash which had
passed through the charger. The collected fly ash was then weighed and the
charge/mass ratio was calculated.
An example of the results of an experiment in which the charging effec-
tiveness measurement was made is shown in Figure 11. In this test, a dust
loading of 3.4 g/m3 and a temperature of 125?C were the conditions under which
the charger was operated. The corona discharge electrode current was held
constant throughout the test, indicating successful back corona suppression.
The charge/mass (Q/m) ratio obtained in this experiment was 4.35 x 10~5 C/g.
This compares to a Q/m value of less than 1 x 10~6 C/g obtained in previous
experiments with a conventional wire-plate precharger at a similar dust load-
ing and a comparable resistivity.
The dust loading was increased to approximately 6.8 g/m3 and the above
experiment was performed with all other parameters the same. The results
(Figure 12) show much higher grid and plate currents. The corona electrode
current began to increase after eight minutes, which indicates an increasing
difficulty to suppress the back corona generated at this higher dust loading.
ELECTRODE GEOMETRY STUDIES
The general electrode configuration used in the laboratory scale studies
proved successful in achieving control of back corona, but only limited work
was done in seeking an optimum geometry. Thus, as a preliminary step leading
to the design of a pilot scale device, a combined theoretical and experimental
investigation was made in order to provide a data base for selecting design
parameters.
A computer simulation comparing the electrical performance of wire-plate
corona systems having a wide range of geometric parameters was successfully
employed to provide a set of theoretical current-voltage characteristics which
21
-------
io4r
ui
IT
CC
O
o Corona Current
A Grid Current
Plate Current
E= 2.93 x I05 V/m
Nt = 1.31 x I013 sec/m3
15 20
TIME, min
Figure 10. Test results of a three-electrode charger used for back corona
suppression. Temperature = 125°C, corona electrode-to-plate
spacing = 8.9 cm, grid electrode-to-plate spacing = 2.6 cm,
and dust loading * 3.4 g/m3.
22
-------
o Corona Current
Grid Current
a plate Current
Charging
Measurement
E= 2.70 x I06 V/m
Nt = 1.53 x I013 sec/m3
Q/m = 4.36 x I0~5 C/g
10 15 20
TIME , min
Figure 11. Test results of a three-electrode charger used for back corona
suppression. Temperature = 125°C, corona electrode-to-plate
spacing = 8.9 cm, grid electrode-to-plate spacing = 2.6 cm,
and dust loading « 3.4 g/m3.
23
-------
o Corona Current
Grid Current
a Plate Current
Charging Measurement
E * 2.70 x I05 V/m
Nt * 1.53 x I015 sec/m3
l.2lxlCT5 C/g
TIME, mm
Figure 12. Test results of a three-electrode charger used for back corona
suppression. Temperature = 130°C, corona electrode-to-plate
spacing =8.9 cm, grid electrode-to-plate spacing = 2.6 cm,
and dust loading ~ 6.8 g/m3.
24
-------
match a set of experimental electrical measurements obtained from a single
laboratory setup. In order to find a theoretical current-voltage curve which
would match a particular experimental result, the effective plate width was
adjusted. Other geometric parameters including wire-to-plate spacing and
wire diameter remained fixed at actual experimental values. An ion mobility
of 2.4 x 10"1* meter2/volt-second was used in the simulation and coincides with
previous experimental determinations of mobility. Figures 13 through 18 pre-
sent the results of matching theoretical and experimental electrical charac-
teristics for six wire-plate configurations. The decreasing accuracy of the
theoretical curve fits as the current increases may result from variations in
the ion mobility due to the changing electric fields which were not accounted
for in the computer model. The ratio of the effective plate width to the
actual plate width was compared with the ratio of the actual plate width to
the wire-to-plate separation, as shown in Figure 19. This curve was used to
predict the effective plate widths required to match the theoretical and ex-
perimental current-voltage curves for two wire-plate geometries for which
experimental data existed. The effective plate widths which produced the best
fits to the experimental curves differed by 0% and 20% from the predicted
values. Thus, an approximate computer model may be obtained for any wire- .
plate configuration by using the effective plate width indicated in Figure 19.
The I-V characteristic of a wire-plate system with a single wire dis-
charge electrode is not strongly dependent upon plate width. The sparkover
voltage becomes smaller, however, if the plate width is reduced to less than
approximately the wire-plate separation. Figures 20 through 25 are an experi-
mentally generated family of curves for wire-plate corona systems. In gen-
eral a drop in maximum current occurs, often sharply, as the wire-plate sepa-
ration is increased beyond the distance equal, to the plate width.
Discharge Electrodes
In order to produce large electric field strengths necessary for corona
generation, field lines must converge strongly at the corona discharge elec-
trode. A very thin wire thus serves effectively as a discharge. But in ap-
plications in a severe environment a fine corona wire does not have the struc-
tural strength to perform for long periods of time.
Barbed wire electrodes have been employed in many electrostatic precip-
itators in order to provide an electrode with both good structural strength
and strong field convergence regions for good corona production. Since each
barb serves as a corona point, a maximum corona current can be achieved by
using as large a number of points as possible. If, however, the barbs are
too closely spaced an interference will occur which can reduce the total co-
rona current in the following manner: convergence of electric field lines at
a corona point causes a reduction in field strength on the discharge electrode
a short distance away from the corona point. Upon inception of a corona cur-
rent the transverse component of the current causes a further reduction in
field in the region outside the corona on the discharge electrode due to the
space charge associated with the corona current. In a linear array of corona
points, it would thus be possible for a corona discharge to occur only at
every other point if the points are too closely spaced.
25
-------
icT
10
.o2
UJ
cr
10
11 11 11 11 I 11 11 11 1 I 11 11 11 I I I II I 111 I I 11111 11 11 I I I I 11 I 11 I I I I I 1 ( I I I
O.I »' i' 11 <» 1111
PLATE WIDTH = Mem
ELECTRODE SPACING =9 cmJ
IO 20 30 40
60 70
APPLIED VOLTAGE, kV
Figure 13. Comparison of theoretical and experimental I-V
characteristics for a wire-plate configuratio'n with
11 cm plate width and 9 cm electrode separation.
26
-------
io4™
I0a
I-
Z
UJ
cc
0
-I 2
< 10
o
11 11 I I 1 I I I 111 I M M I I | 111 I I II
PLATE WIDTH = 11 cm
ELECTRODE SEPARATION - 3cm
|0 I 111 i 11 11 i 11 i i 111 i 11 11 11 M i 1111 i ii i 1111 i I i ii i i 11 i i 11 M n i 111 1111 i
0 10 20 30 40 50 60
APPLIED VOLTAGE, kV
70
Figure 14. Comparison of theoretical and experimental I-V
characteristics for a wire-plate configuration
with 11 cm plate width and 3 cm electrode separation,
27
-------
10* fr-
o
103
10
<
a.
UJ
or
K
Z)
o
o
10
PLATE WIDTH * II cm
ELECTRODE SPACING = 2cm
0 I I ' ' i
Figure 15
i i
K)
20
APPLIED VOLTAGE , kV
30
Comparison of theoretical and experimental I-V
characteristics for a wire-plate configuration
with 11 cm plate width and 2 cm electrode sepa
ration.
28
-------
10'
< 2
a. IOZ
LU
ac
10
o
I-
OjL-i-L"
I i i i i I i i i I ji i i i i i i i I i i I I i I r
PLATE WIDTH * 5.5 cm
ELECTRODE SPACING = 5 cm
I i i i i I i i i i I i i i i I i i
I
i i
10 20 30 40 50
APPLIED VOLTAGE t kV
60
70
Figure 16.
Comparison of theoretical and experimental I-V
characteristics for a wire-plate configuration
with 5.5 cm plate width and 5 cm electrode spacing.
29
-------
id*
:nTmin|MiMiiii|M INN iijiii i iiMi|iiMMiii|iiiiMimi i mm i:
10
10
-------
10
I03
10s
UJ
a:
oc
h-
o
I-
10
0.1
I -
i i i I I I I I I i I [ I I i i i i I i i i i r
O
PLATE WIDTH =2.75 cm
ELECTRODE SPACING * 2 cm
41 0
i i i i i i i i i i i t i i i t i i i I i i i
10
1
20
30
APPLIED VOLTAGE , kV
Figure 18. Comparison of theoretical and experimental I-V
characteristics for a wire-plate configuration with
2.75 cm plate width and 2 cm electrode separation.
31
-------
T — I — I
1 — TT
T—I I I I 1 I I
1 I I I I I I I I
i
I2
CL
I-
O
g
*
UJ
a.
H
UJ
OS
t
a.
i I
I I
I
£ ACTUAL PLATE WIDTH
WIRE-PLATE SEPARATION
Figure 19.
Ratio of apparent to actual plate width used to
provide best theoretical fit for various values
of plate width to electrode separation ratio.
32
-------
10
-1 i i i i i i i TI n n i i n i \\ i TT
i_
10
100
z
UJ
K
DC
U
I
10
WIRE-PLATE SEPARATION:
V 2cm
0 3 cm
A 5cm
O 7cm
n 9cm
O.M i i i i I i i i i i
i
i i i i i t
Figure 20,
10 20 30 40 50
APPLIED VOLTAGE, kV
60
_L_L
70
Electrical characteristics of a parallel
wire-plate corona electrode system for five
values of electrode spacing. Plate width
is 11 cm, and wire diameter is 0.25 mm.
33
-------
ICF
10
100
h-
UJ
£ 10
o
0.1
Figure 21,
WIRE-PLATE SEPARATION:
V 2 cm
O 3cm
A 5cm
0 7cm
D 9cm
10 20 30 40 50
APPLIED VOLTAGE, kV
60
70
Electrical characteristics of a parallel
wire-plate corona electrode system for five
values of electrode spacing. Plate width
is 5.5 cm, and wire diameter is 0.25 mm.
-------
Itf1
- I I I I
1 I I I I I I I I I I I I
I03
100
cc.
or
o
<
o
10
WIRE-PL ATE SEPARATION:
y 2cm
O 3 cm
A 5 cm
O? cm
O 9cm
O.I
I i
10
60
70
Figure 22
20 30 40 50
APPLIED VOLTAGE, kV
Electrical characteristics of a parallel
wire-plate corona electrode system for five
values of electrode spacing. Plate width is
2.75 cm, and wire diameter is 0.25 mm.
35
-------
10 ti i i i ii i i i i
100
h-
z
tu
o
10
O.I
Figure 23.
WIRE-PLATE SEPARATION:
2cm
O 3 cm
A 5cm
0 7cm
D 9cm
10 20 30 40 50
APPLIED VOLTAGE, kV
60
70
Electrical characteristics of a parallel
wire-plate corona electrode system for five
values of electrode spacing. Plate width is
11 cm, and wire diameter is 0.79 mm.
36
-------
10
4
I03
100
UJ
i 10
:D
o
0.1
I II
WIRE -PLATE SEPARATION:
o3 cm
A 5cm
0 7cm
9cm
i
10 20 30 40 50
APPLIED VOLTAGE, kV
60
70
Figure 24.
Electrical characteristics of a parallel
wire-plate corona electrode system for
five values of electrode spacing. Plate
width is 5.5 cm, and wire diameter is 0.79 mm.
37
-------
10"
' I
,o3
100
tu
QC
oc
Z)
o
t-
o
10
WIRE-PLATE SEPARATION
V 2 cm
O 3cm
A 5 cm
0 7cm
D 9 cm
0.1
Figure 25.
fiii
i i i I iiit
10 20 30 40 50
APPLIED VOLTAGE, kV
60
70
Electrical characteristics of a parallel
wire-plate corona electrode system for
five values of electrode spacing. Plate
width is 2.75 cmf and wire diameter is
0.79 mm.
38
-------
As the corona points on a barbed wire electrode are brought closer to-
gether, it would therefore be expected that the total corona current per unit
length of electrode would increase to some maximum value. Bringing the points
closer together would force a quenching of the corona on alternate points,
thus reducing the total current. This conjecture has been borne out by exper-
iment, as shown in Figure 26. A barbed wire electrode along the axis of a
14 cm diameter cylindrical conductor was used as a corona discharge electrode.
A maximum total corona current is found for a barb spacing of approximately
0.7 cm.
A similar effect is shown in Figure 27 for a discharge electrode made up
of a set of discs spaced along a rod on the cylinder axis. In this example
the current per disc is plotted as a function of the disc separation distance.
The rapid change in current per disc where the separation is close is similar
to the barbed wire results.
The number of possible configurations for discharge electrodes is prac-
tically limitless; however, most can be derived from a set of sharp points or
sharp edges supported by a rigid structure. Since many discharge electrode
structures are complicated geometrical figures or contain discontinuities at
the surface analytical treatment is generally impracticable. Comparative
empirical studies using a fixed passive electrode offer the best means of
evaluating discharge electrode types.
Among those tested, including helix, ribbon, barbed wire and disc elec-
trodes those producing the greatest total current were barbed wire and stacked-
disc electrodes. Figure 28 shows a comparison of these two types, along with
a straight wire electrode. Sharpedged discs, 5 cm in diameter and spaced ap-
proximately 5 cm apart produced the largest total current. The barbed wire
electrode also exhibits better performance than the straight wire. The barbed
wire also has the advantage of causing less obstruction to the flow of gas
through the system than does a system of disc electrodes.
39
-------
ICT
10
z
HI
cc
cc
r>
o
I-
o
10
i i i i i i i
APPLIED VOLTAGE:
O 20 kV
D 40 kV
A 60 kV
i 11111 11 11 11 i 111 I
111
Figure 26
123
BARB SEPARATION,
cm
Total corona current for fixed length of
barbed wire discharge electrode as a func-
tion of separation between barbs.
40
-------
10
I-
UJ
tr
o
I I I I I i i i
i 1111 i 11 11
APPLED VOLTAGE :
O 40 kV
n 50 kV
A 60 kV
40 I ' ' ' 11 i 11 11 i i i i 11 i i i I i i 11 11 i i i I i i i 11 i 111 I i 11 i i i i i i 11 11 11 i i 1
Figure 27
12345
DISC SEPARATION , cm
Corona current per disc for a discharge electrode
consisting of discs at various spacings with axes
aligned.
-------
10'
icr
UJ
OC. 9
o: IOZ
o
i-
o
10
= " 'I'n'l" "I "• i| .•••[ IM.M
i i i
DISCHARGE ELECTRODE:
O.02cm WIRE
A BARBED WIRE
[] MULTIPLE DISC
I i i i I I i i i i I i i i i I i i i i I i i i i I i i i i I i i i
10 20 30 40 50 6O 70 80
APPLIED VOLTAGE, kV
Figure 28.
Comparison of I-V characteristics for a 0.02 cm
wire, a barbed wire and an array of disc dis-
charge electrodes. The passive electrode is
a 1.4 cm diameter cylinder for all three curves.
42
-------
SECTION 4
PILOT SCALE PRECHARGER
The results of the laboratory scale work were sufficiently encouraging to
permit proceeding with the development of a pilot scale three-electrode pre-
charger capable of handling a minimal gas flowrate of approximately 0.71 m3/sec
(1500 acfm). Such a device was designed and fabricated by Southern Research
Institute. Figure 29 is a photograph of the precharger lying on one side.
The top of the precharger is shown at the left-hand side of the picture. The
three thick rods protruding from the top of the device are rapping rods, which
are welded to the top edge of the plate electrodes. A pair of spring-loaded
supports can be seen flanking each of the rapping rods. The arrangement of
the electrodes can be seen in more detail in Figure 30, which also shows the
shape of the corona discharge electrodes. The discharge electrode-to-plate
spacing is 9.2 cm and the screen electrode-to-plate spacing is 2.0 cm.
INITIAL PILOT TEST PROGRAM
The device was installed in the test section at the inlet of an existing
conventional pilot scale ESP at SoRI. Current-voltage (I-V) characteristics
for the charger at ambient conditions and at elevated temperature were deter-
mined (Figure 31). There is approximately a two-fold increase in current with
elevated temperature and a lOkV decrease in breakdown voltage. This result is
expected because of the increased ion mobility at higher temperatures.
Tests of precharger performance under conditions of high resistivity dust
loading were undertaken using redispersed fly ash. Temperatures ranging from
75°C to 130°C were used, and measured values of dust resistivity were in the
range of 1012 to 1013 ohm-cm. The fly ash was injected into the system by
means of a sandblasting gun, at a rate controlled by the air pressure applied.
Impactor measurements were made to determine the. size distribution of the
redispersed fly ash. Operating the sandblaster at a pressure of 138 Pa (20 psi)
injects approximately 1.88 g/m into a 0.71 m3/sec (1500 ft3/min) stream of gas.
The particulate mass median diameter at the inlet of the precharger has been
determined to be about 25 ym. Figures 32-35 show the results of particle size
distribution measurements.
Preliminary tests were conducted to determine the effectiveness of the
screen electrode in the prevention of back corona, using a particulate loading
of 1.88 g/m3 at a temperature of 130°C, and humidity controlled at 1.2% (by
volume). Ash resistivity under these conditions was determined to be approx-
imately 10 ohm-cm. These tests were performed without plate rapping. Fig-
ures 36, 37, and 38 illustrate three back corona suppression tests. The
43
-------
Figure 29. View of the pilot scale charger on its side,
-------
V
\_
(Full scale)
Detail of the Corona Discharge Electrode
Figure 30. View of the pilot scale charger electrode configuration.
-------
K)4
10s
UJ
-------
DIET PKDWCER TEST FEED WOE 3D fSL 10-S/S&-77
Wl" Z»ff m/GC DELUDE
loSr
103,:
IJOS:
1O
'1
10
1
H 1 I I I Hlf 1 1 I I I lll| 1 1 I I I lll|
10° 101 ID5
PARTICLE DIAMETER (MICROMETERS)
Figure 32. Results of the dust loading characterization at the precharger inlet
with the sandblasting gun at 138 Pa (20 psi).
-------
Q.
M
INLET PREOMGER 1ST FEED RWE M PH 1D-5/S-77
DCUff W95 LESS THW .5111006
99. B-
99.5-
95
90
BO
70
BO
50
30
EO
1O
5
S
0.5
O.E
0.01
10
"1
-I—I I I I lll( 1 1 Mill
10°
101
H MM MM
IflP
PARTICLE DIAMETER (MICROMETERS)
Figure 33. Results of the dust loading characterization at the precharger inlet
with the sandblasting gun at 138 Pa (20'psi).
-------
DIET nBMHB) TBT FEED HUE a F3 11HWH7
w = e.ff w/tr none us use ww & HODS
104-::
10s-
10°
I « *
I «
10
rl
I I I IIM| H—I I I Mill 1—I I I Mill
IflP
101
PARTICLE DIAMETER (MICROMETERS)
Figure 34. Results of the dust loading characterization at the precharger inlet
with the sandblasting gun at 138 Pa (20 psi) .
-------
XET PREOWER TEST FEED RATE SI PH lfl-S/S-77
WUK IMS US mm.SKIDDS
/-s
u
d
-UJ]
«
•
•
•
:
«
••
i
4
M :
h-
f\/ '
^^? /i f\lLO
i
•
«
r-j
*— f i
•
§
I
1
10^1
1 :
i
i
<
i
-in?
•
•
»
:
***»»
r •
: •
I *
* ••
* •
r 9
t
»
•
I
•
' I
:
i
»
»
*
:
•
»
»
»
•
r
*
•
•
»
m
m
t 1 1 1 1 lUl 1 1 1 1 1 1 111 1 1 1 1 1 1 III
icr1 ±cP lo1
PARTICLE DIAMETER (MICROMETERS)
Figure 35. Results of the dust loading characterization at the precharger inlet
with the sandblasting gun at 138 Pa (20 psi).
50
-------
to8
ui
a:
oc.
ID
O
I02
10
Corona"
-o—o—a—o-
Screen
r*~
- Screen Current < 0
10
20
TIME , minutes
30
Figure 36. Back corona suppression test results. Test conditions were: T = 130°C,
%HaO = 1.2 (by volume), average gas volume flowrate «= .71 m3/sec
(1500 ft3/min), dust loading - 1.88g/m3, HMD = 25 ym, E - 3.26 x 105V/m,
and Nt - 1.4.x 1013 sec/m3.
51
-------
O
- Screen Current < 0
30 40
TIME, minutes
Figure 37. Back corona suppression test results. Test conditions were: T = 130°C,
%HaO = 1.2 (by volume), average gas volume flowrate = .71 m3/sec
(1500 ftVmin), dust loading - 1.88g/m3, HMD - 25 ym, E - 2.72 x 105V/m,
and Nt - 1.3 x 1013 sec/m 3.
-------
10
10s
t
«•
I
tr
o
—i—i—i—r
- Screen Current < 0
\r/
TIME, minutes *'
Figure 38. Back corona suppresstion test results. Test conditions were: T = 130°C,
%HaO = 1.2 (by volume), average gas volume flowrate = .71 m3/sec
(1500 ft3/min), dust loading « 1.88g/m3, HMD = 25 ym, E = 2.61 x 105V/m,
and Nt - 8.8 x 1012 sec/m3.
53
-------
procedure in each of these experiments is as follows: 1) the voltage
on the corona discharge electrodes is adjusted to give the total corona current
corresponding to the desired current density with the screen electrodes voltage
simultaneously adjusted to give zero screen current; 2) the ash injection sys-
tem is turned on at time t = 0; and 3) the screen electrodes voltage is adjusted
throughout the duration of the experiment in order to maintain the total corona
current constant. In each of the tests shown, the total corona current could
not be kept constant for the entire time due to excessive sparking at the screen
electrodes' voltage required to hold the corona current down. When this condi-
tion occurred the corona current was held at the lowest stable value.
Further tests were performed to determine the charging effectiveness of
the precharger, using the same general procedure described in the above para-
graph to control the voltages on the precharger electrodes. The measure of
charging effectiveness was taken to be the ratio of charge to mass, Q/m on a
sample of particulate matter extracted downstream from the precharger. The
particles are collected on a silver mesh filter mounted in an insulated plas-
tic filter holder, fitted with a nozzle for isokinetic sampling. The filter
is connected to an electrometer so that the charge accumulation can be moni-
tored during the collection process. A foil shield, grounded through a 10
Megohm resistor, is wrapped around the body of the plastic filter holder to
prevent a buildup of surface charge on the insulating material. The mass of
the collected particulate is determined at the conclusion of the experiment
and the Q/m value is calculated.
Tests of the precharger's back corona suppression capability and charging
effectiveness were conducted with the gas stream temperature equal to 130°C,
the moisture content in the gas stream equal to 1.2% (by volume), a dust load-
ing of approximately 1.88g/m , and an average volume flowrate of 0.71 m3/sec
(1500 ft /min). The fly ash used in the experiments under these conditions
has a measured resistivity of 1.2 x 1012 fi-cm. The result of a typical test
at these conditions is shown in Figure 39. The passive electrodes were rapped
manually every five minutes during this experiment. The average value of Q/m
obtained from four tests at 130°C is 3.8 x 10~6 C/g. The current-voltage
(I-V) characteristics of the precharger at 130°C with the electrodes clean and
the electrodes dirty (Figure 40) show that back corona is being produced at
the current level maintained during the experiments.
Additional tests of the precharger's performance were conducted at 75°C,
with the other parameters constant. The fly ash used in these experiments had
a measured resistitivy of 1.4 x 1012 ft-cm at test conditions. The passive
electrodes were rapped manually every two minutes during the experiments. The
result of one test at 75°C is shown in Figure 41. Control of back corona was
achieved, as shown by the successful maintenance of the corona current at its
initial value. An evaluation of charging effectiveness was made during the
test, as indicated in the figure, and the Q/m value was measured to be 9.6 x
10~6 C/g. Successful back corona suppression was maintained during another
test at these conditions which lasted two hours.
The Southern Research Institute ultrafine particle sampling system with
a Thermosystems, Inc. Electrical Aerosol Analyzer (EAA) was set up at the out-
let test section of the EPA-SoRI pilot scale ESP. Table 1 shows the results
54
-------
10*
01
o:
•r
•u o o-
°-Corona Current
o- Grid Current
_L
_L
10
20 30 40 50
TIME, minutes
60
Figure 39.
Back corona suppression test at 130°C, 1.2% HaO, p = 1.2 x 10 12 f2-cm,
j = 9.4 x 10s nA/m2, and the dust loading « 1.88g/m3.
55
-------
H
Z
UJ
o:
c?
<
o
cc
o
o
10s
I01
1 r
o--plates Clean
o—Plates Dirty
10
20 30 40 50
CORONA VOLTAGE, kV
60
Figure 40. Corona current vs. corona voltage for clean and dirty electrodes at
130°C, 1.2% H20, and p = 1.2 x 1012 fi-cm. Grid current was held to
zero for these measurements.
56
-------
I03
Z
UJ
a:
-------
Collection Efficiencies
Particle Diameter
Collector
Ul
CXI
0.013 *
0.022 *
0.031
0.050
0.092
0.150
0.220
0.310
- 1.01
- 8.22
13.10
17.98
22.51
21.58
21.92
21.89
Precharger + Collector
(ft)
22.22
52.38
41.17
24.81
47.38
49.55
49.34
49.71
Large standard deviations were recorded for the collection efficiencies
corresponding to these two size ranges.
x 1012fl-cm,
Table 1. EAA data analysis for the following test conditions:
T = 75°C, % H2O = 1.2% (volume), p = 1.4
j . = 9.4 x 105nA/m2
Jprecharger
dust loading = 1.88g/m3.
2 ., . =2 kV/cm,
collector ' '
-------
of the EAA data analysis for a test conducted at 75°C. The comparative effi-
ciencies of the precharger-collector system and the collectors alone indicate
a more than two-fold average increase in ESP performance with the addition of
the precharger. Optical particle counter data were also acquired during the
test documented above, showing a similar percentage of performance enhancement
for particles of diameters greater than 0.3 micrometers.
The result of a back corona suppression test at 100°C is shown in Figure
42. In this case, the corona current was maintained constant throughout the
test by momentarily turning off the screen and corona discharge electrodes'
power supplies during rapping (the plates were rapped every two minutes). The
absence of the electric field improves the rapping efficiency enough that suf-
ficient fly ash is removed from the plates to allow continuous operation.
The Q/m measured during this test was 3.0 x 10~6 C/g. The I-V characteristics
of the precharger at 100°C with the electrodes clean and dirty show in Figure
43 that back corona is evident at the operating current level.
In further experiments a back corona control test was conducted with the
gas stream temperature equal to 75°C and the moisture content equal to 1.2% by
volume. The fly ash had a resistivity of 1.4 x 1012 fi-cm at these conditions.
The ash was injected with the sandblaster set at 276 Pa (40 psi) pressure,
or twice the pressure used in previous tests. This corresponded to a dust
loading of 5.97 g/m3. The precharger plates were rapped manually every two
minutes. The results of this experiment are shown in Figure 44.
Automatic pneumatic rappers were installed on the precharger and placed
in service. The test described in the above paragraph was repeated with this
modification made to the system (Figure 45). The rappers were operated at
552 Pa (80 psi) air pressure (25 ft/lbs energy per impact). A charging effec-
tiveness measurement was made during the test, yielding a Q/m value of 8.6 x
10~6 C/g. In addition, a Climet model 208B optical particle counter and a
Tracor-Northern model TN-1705 pulse height analyzer were used to qualitatively
evaluate the downstream collector efficiency enhancement with the precharger
on. The collector was operated at -30 kV potential throughout the experiment.
Data were taken for particles in the range .75 ym diameter to 3.3 ym diameter.
The penetration for particles in this diameter range was decreased by 33.7%
with the precharger turned on.
A back corona suppression test was then performed with the gas stream
temperature equal to 100°C and the ash injected at the sandblaster setting
of 414 Pa (60 psi) pressure. All other parameters remained the same as in
the test described above. As can be seen in Figure 46, control of back corona
is more difficult in this case. This is evidenced by the instability of the
corona electrode current. The I-V characteristics of the precharger at 100°C
before (clean) and after (dirty) this test iridicate back corona is produced
by the dust layer deposited on the passive electrodes (Figure 47).
Figure 48 illustrates a back corona suppression test with a gas temperature
of 100°C and fly ash injected under 345 Pa (50 psi) pressure on the sandblaster.
Other parameters, including moisture content and plate rapping were unchanged
from the preceding test. The decrease in dust loading allowed more stable
59
-------
UJ
-------
I04
10*
10*
bJ
flC
CC
D
O
10'
10°
1 T
o-plates Clean
a- Plates Dirty
10
20 30 40
VOLTAGE, kV
Figure 43. Corona current vs corona voltage for clean and dirty electrodes at
100 C and 1.2% H20. Grid current was maintained at zero for these
measurements. UUCBB
61
-------
0 O O O O -0
10
20
30 40 50
TIME , minutes
60
70
Figure 44. Back corona suppression test at 75°C, 1.2% H20, p = 1.4 x 1012 ft-c
j = 9.4 x 10~5 nA/m2, ash injection at 276 Pa, and manual rapping
62
-------
10
20
30 40
TIME , minutes
50
60
70
Figure 45.
Back corona suppression test and Q/m measurement at 75°C, 1.2%
p • 1.4 x 1012 fl-cm, j - 9.4 x 10~5 nA/m2, ash injection at 276 Pa,
pneumatic rapping at 552 Pa, and Q/m » 8.7 x 10~s C/g.
63
-------
10
10
20
30 40
TIME , minutes
50
60
70
Figure 46. Back corona suppression test at 100°C, 1.2% H20, j - 9.4 x 10~5 nA/m2,
ash injection at 414 Pa, and pneumatic rapping at 552 Pa.
64
-------
10
10
20 30 40
CORONA VOLTAGE ,
kV
Figure 47. Corona current vs. corona voltage for clean and dirty electrodes at
100°C and 1.2% H20. Grid current was held to zero for these
measurements.
65
-------
10
10
30 40
TIME , minutes
Figure 4'8. Back corona suppression test and Q/m measurement at 100°C, 1.2%
j « 9.4 x 10 nA/m , ash injection at 345 Pa, pneumatic rapping at
552 Pa, and Q/m = 2.86 x 10~6 C/g.
66
-------
control of the back corona. A charging effectiveness measurement made during
the test gave a Q/m value of 2.86 x 10~6 C/g.
The test conditions for successful back corona suppression experiments
are summarized in Table 2. As in the case depicted in Figure 32, there were
other conditions for which back corona control was marginal.
Stress cracks developed on the precharger's passive electrodes due to the
rapping force applied during the experiments. Modifications were made to
remedy this problem and prevent its reoccurrence.
INSTALLATION OF AUTOMATIC SCREEN VOLTAGE CONTROL
An automatic screen voltage control circuit was devised and built to en-
able "hands-off" operation of the precharger. Figure 49 is a schematic dia-
gram of the electronic circuit used for controlling the screen voltage in
response to fluctuations in the primary corona current. The output voltage of
the screen power supply (Spellman model RHR15PN225/RVC/TP/FG) can be controlled
over its entire range by applying a low voltage signal to pin 6 of the remote
voltage control terminal board.
Since adjustments required in the screen voltage depend upon variations
in the primary corona current, the input signal to the control circuit is de-
rived from the ground return line on the corona power supply by means of a
4N25 opto-isolator. The signal is then amplified by a factor of ten by means
of one section of an LM747 dual operational amplifier. The other section of
the LM747 is used as an integrating circuit to even out rapid transient volt-
ages in the control signal. A dc bias voltage, derived from a voltage divider
network, is added to the control signal at the input of the integrating
circuit.
In order to set the system for automatic control, both power supplies are
turned on, and the screen supply is set in the automatic mode. The corona
power supply is set at the desired voltage and current level for clean pre-
charger operation. The lOkft potentiometer in the control circuit is then ad-
justed to the point where the screen current falls to zero. No further adjust-
ments are necessary under normal operating conditions.
When high resistivity dust is injected into the precharger, the effects
of back corona may tend to increase the primary corona current. Such a change
is sensed by the automatic control circuit, which increases the screen voltage
until the primary corona current returns to its original value. The screen
voltage is thus caused to follow the fluctuations resulting from back corona
in such a manner that the primary corona current remains constant.
The automatic control circuit was installed in the screen power supply
cabinet and tested using two different types of corona power supplies. In
both cases the circuit performed as described in the preceding paragraphs.
The primary corona current was held at a very steady level as large fluctua-
tions occurred in the screen current.
67
-------
TEST CONDITIONS
Gas Stream
Temperature
(°C)
130
100
100
75
75
75
1
1
1
1
1
1
.2
.4
.4
.4
.4
.4
X
X
X
X
X
X
Fly ash
Resistivity
(ft-cm)
1012
1012 6 75°C
1012 @ 75°C
1012
1012
1012
Air Pressure to
Corona Ash Dispersion Rapping
Current Density Device Mechanism
(nA/cm2) (Pa)
94
94
94
94
94
121
138
138, 276
276, 345
138, 276
276, 414
414
manual
manual
pneumatic
manual
pneumatic
pneumatic
(552 Pa)
(552 Pa)
(552 Pa)
oo
Table 2. Test conditions for which successful back corona suppression was maintained.
-------
TO CORONA WIRE
Cl
rh rrr
4N25
I,, .^ i
P:~ M-&
T~T
/T7 N.C.
-6V
I j
REMOTE VOLTAGE
CONTROL TERMINAL
BOARD
Figure 49. Schematic diagram of the electronic circuit designed to provide
automatic adjustment of the precharger screen voltage in response
to changes in the primary corona current.
-------
PILOT TESTS AT IERL/RTP
A series of tests was conducted at the Environmental Protection Agency
Industrial Environmental Research Laboratory in Research Triangle Park, North
Carolina. The pilot scale precharger was installed in the place of the inlet
test section of the in-house precipitator for these tests. Telescoping.duct
sections with sampling ports had been fabricated and were used to fit the pre-
charger into the ESP inlet test section space.
Tests were performed to make sure the precharger's behavior had not
changed during transportation and/or set-up, and also to check the automatic
screen power supply control circuit's ability to maintain the corona current
constant. An example of a back corona suppression test conducted for these
purposes is illustrated in Figure 50. This test was made with a gas tempera-
ture of 93°C (200°F), gas volume flow rate of 5.19 m3/sec (1100 ACFM), pneu-
matic rappers operated at about 310 Pa air pressure, and fly ash, having a
resistivity p ^ 1.4 x 1012 ft-cm, injected at a rate equivalent to 1.0 g/m3
(0.44 gr/ft3). The figure shows that despite large variations in the screen
current, the corona current was held constant by the automatic controller.
The screen current rise corresponds to increasing back corona activity on the
plates. The presence of back corona is confirmed by the difference between
clean and dirty plate current-voltage curves shown in Figure 51. A charge-to-
mass ratio measurement was conducted during the back corona test illustrated
in Figure 50 with the resulting value of Q/m » 1'.35 x 10-6 C/g.
Another back corona suppression test is depicted in Figure 52. The gas
temperature equalled 107°C (225°F) for this test (all other parameters remained
the same as described in the above paragraph). Very large fluctuations in the
screen current are evident with only two minor disturbances occurring in the
corona current.
Experiments were conducted in conjunction with the downstream collector
in order to determine the efficiencies of the precharger, collector, and the
precharger-collector system. Particles penetrating the device were extracted
through a sampling nozzle inserted in a sampling port on the outlet test sec-
tion of the ESP, reduced in number concentration with a diluter, and analyzed
with a Climet optical particle counter. A Tracer-Northern multi-channel ana-
lyzer was used to count the particles and provide their size distribution.
The information was then recorded on a DECwriter printer.
The first efficiency test was performed with the collector plates spaced
38 cm (15 in.) apart and the wire-to-wire spacing equal to 17.8 cm (7 in.).
Other test parameters were fixed at the following values: gas temperature =
100°C (213°F), gas flow rate = 5.19 m3/sec (1100 ACFM), moisture content =
0.6%, and fly ash, (p £1.4 x 1012 fl-cm), injection rate = 1.0 g/m3 (.44 gr/ft3).
The particle diameter range sele'cced to provide the particle number counts for
the efficiency computation was 1.8 to 5.0 Jim. Three values of efficiency, or
decrease in penetration of particles in the selected diameter range, were
measured: 1) the collection efficiency of the precharger; 2) the collection
efficiency of the precharger-collector system; and, 3) the collection effi-
ciency of the collector. The results of this test and a repeat of this test
are summarized beginning on page 75.
70
-------
I04
I03
I O 0 O O
UJ
cr
(T
O
I02
10
Q/m
10 20
J_
a o
Screen
Corona
I i l
30 40
TIME, mm
50 60 70
Figure 50. Back corona suppression test with T = 93°C, flowrate - 5.19 m3/sec,
and mass loading =1.0 g/m3. The Q/m value obtained in this test
is 1.36 x 10~6 C/g.
71
-------
I04
10s
QL
O
I02
10
10
I »I
Dirty Plates
Clean Plates
_L
ZO 30 40 50
CORONA VOLTAGE , kV
60
70
Figure 51. Current-voltage characteristics for the precharger clean and dirty
at 93°C.
72
-------
O4
T \ 1 1 1 T
UJ
I
o
o - Corona
o-Screen
i I i I ' 1 I 1 1 1 1 1
10 20
30 40 50 60
TIME , min
70
Figure 52. Back corona suppression test with T = 107°C, flow rate =5.19 m3/sec,
and mass loading = 1.0 g/m3.
73
-------
Another set of efficiency tests was performed with the collector plate-
to-plate spacing » 30.5 cm (12 in.), wire-to-wire spacing = 17.8 cm (7 in.),
gas temperature » 92°C (198°F), gas flow rate =9.44 m3/sec (2000 ACFM), mois-
ture content = 0.6%, and fly ash (p > 1.4 x 1012 ft-cm) injection rate =
1.0 g/m3 (.44 gr/ft3). Again, the 1.8 - 5.0 ym diameter particle range was
used to calculate the following efficiencies: 1) the collection efficiency of
the precharger; 2), 3), and 4) the collection efficiencies of the precharger-
collector system at three different collector field strengths; and 5), 6), and
7) the collection efficiencies of the collector at three different electric
field strengths. The results of these measurements are also included in the
summary.
A third set of experiments was made with a collector plate-to-plate spac-
ing of 20.3 cm (8 in.). All other conditions remained the same as in the
previous tests. The collection efficiencies of the precharger alone, the pre-
charger-collector system at three different collector field strengths, and the
collector alone at three different field strengths were calculated for the
particle size range 1.8 - 5.0 ym diameters. The summary also includes these
results.
As indicated by the tabulated results of the efficiency measurements,
the relative electrode positions in the downstream collector have a marked
effect on the system behavior. In the 38 cm plate-to-plate spacing case, the
percentage of collection efficiency enhancement due to the precharger was
found to be negligible. The 30.5 cm plate-to-plate condition showed a very
significant improvement in collection when the precharger was on. The 20.3 cm
plate-to-plate spacing yielded a significant enhancement due to the precharger,
but less of a performance boost than with a 30.5 cm spacing. The differences
in performance of the precharger-collector system at the three plate-to-plate
spacings was possibly due to anomalous electric field effects as the ratio of
plate-to-plate to wire-to-wire separation was varied. These data indicate
that additional experiments on the effects of plate-to-plate and wire-to-wire
spacing are needed.
A trend in the efficiencies is evident. The percentage enhancement due
to the precharger is less on the second, or repeat experiment at each plate-
to-plate spacing. This is probably caused by the absence of rapping in the
collector and the resultant degradation of performance by the deposited dust
layer on the plates.
It should be noted that the collection efficiencies given in the tables
were calculated from comparisons of dust concentration at the outlet of the
precipitator with the precharger on and off, and collectors on and off. The
real system collection efficiency, i.e., outlet concentration versus inlet
concentration, was determined by collecting a mass sample at the ESP outlet
and comparing the dust loading to the predetermined fly ash feed rate. This
measurement was made with the collector plate-to-plate spacing of 20.3 cm.
The decrease in particulate penetration resulting from action of the precharger,
with the collectors operated at MO kV was 22% and 9% for the two tests,
which agrees with the number concentration percentages. The overall system
collection efficiency was determined to be 92% and 86% respectively at these
conditions.
74
-------
SUMMARY OF TESTS RESULTS
I. Plate-to-plate spacing = 0.38 m (15")
Gas temperature = 100°C (212°F)
Gas flow rate = 5.19 m3/sec (1100 acfro)
Moisture content = .6 v/o
Dust loading - 1.0 g/m3 (0.44 gr/ft3)
Particle size range observed = 1.8 - 5.0 ym dia.
a. Collection efficiency of precharger (27 kV, 700 yA) = 9.8%, 3.1%
b. Collection efficiency of precharger (27 kV, 700 yA)
plus collectors (38-41 kV, 0.00-0.04 mA) - 33.9%, 35.8%
c. Collection efficiency of collectors (36-41 kV,
0.00-0.08 mA) = 32.4%, 37.1%
d. Percentage decrease in penetration due to precharger = 2.3%, -2.2%
II. Plate-to-plate spacing = 0.305 m (12")
Gas temperature = 92°C (198°F)
Gas flow rate = 944 m3/sec (1000 acfm)
Moisture content = .6 v/o
Dust loading = 1.0 g/m3 (0.44 gr/ft3)
Particle size range observed = 1.8 - 5.0 ym dia.
a. Collection efficiency of precharger (28 kV, 700 yA) = 18.0%, 10.3%
b. Collection efficienty of precharger (28 kV, 700 yA)
plus collectors (29.8 - 30.3 kV, 0 mA) = 28.9%, 23.2%
c. Collection efficiency of precharger (28 kV, 700 mA)
plus collectors (29.6 - 40.6 kV, 0.00 - 0.03 mA) = 39.7%, 29.9%
d. Collection efficiency of precharger (28 kV, 700 yA)
plus collectors (44.0 - 46.0 kV, 0.05 - 0.18 mA) = 50.8%, 35.4%
e. Collection efficiency of collectors (29.9 - 30.3 kV,
0 mA) = 13.0%, 9.4%
f. Collection efficiency of collectors (40 kV,
0.00 - 0.01 mA) = 25.4%, 21.3%
g. Collection efficiency of collectors (43.8 - 45.1 kV,
0.03 - 0.29 mA) = 26.8%, 23.9%
h. Percentage decrease in penetration due to precharger
1) with collectors at 30 kV =18.0%, 15.2%
2) with collectors at 40 kV - 19.5%, 10.9%
3) with collectors at 45 kV - 32.7%, 15.1%
III. Plate-to-plate spacing - 20.3 cm (8")
Gas temperature = 92°C (198°F)
Gas flow rate =9.44 m3/sec (2000 acfm)
Moisture content = .7 v/o
Dust loading = .44 g/m3 (1.0 gr/ft3)
Particle size range observed = 1.8 - 5.0 ym dia.
a. Collection efficiency of precharger (27.8 kV, 700 yA) = 19.9%, 8.8%
b. Collection efficiency of precharger (27.8 kV, 700 yA)
plus collectors (30.0 - 30.3 kV, 0 mA) = 24.9%, 23.2%
75
-------
c. Collection efficiency of precharger (27.8 kV, 700 A)
plus collectors (34.8 - 35.1 kV, 0.05 - 0.35 mA) = 26.8%, 19.2%
d. Collection efficiency of precharger (27 kV, 700 A)
plus collectors (36 - 40 kV, 0.5 - 2.25 mA) = 45.5%, 28.3%
e. Collection efficiency of collectors (30.0 - 30.1 kV,
0.00 - 0.04 mA) = 11.4%, 12.7%
f. Collection efficiency of collectors (34.8 - 35.0 kV,
0.08 - 0.63 mA) = 13.6%, 16.3%
g. Collection efficiency of collectors (31.0-40.0 kV,
0.93 - 2.25 mA) = 30.3%, 19.0%
h. Percentage decrease in penetration due to precharger
1) with collectors at 30 kV =15.3%, 12.0%
2) with collectors at ^35 kV = 15.2%, 3.5%
3) with collectors at ^0 kV = 21.7%, 11.4%
The test results detailed in the above summary show that the use of the
precharger can produce a substantial improvement in the collection efficiency
of the system. The overall values of collection efficiency were, however,
quite low. It was concluded from these results that improved performance of
the system would require optimizing the electrical configuration of the pilot
scale ESP that served as the downstream collector.
SECOND GENERATION PILOT PRECHARGER
Although the electrical performance of the prototype pilot scale pre-
charger was good, there were some design problems that required correction if
adequate performance was to be expected in a field environment. The spacers
holding the screen electrodes were fabricated of glass-filled Teflon. Be-
cause these spacers were located in the gas stream, they became coated rapidly
with fly ash, which tends to degrade the insulating properties of the material.
Teflon is also susceptible to heat damage.
A redesign of the precharger was undertaken with the objective of remov-
ing all insulating materials from the gas stream and providing a generally
more rugged structure. The design features of the precharger are shown in
Figure 53. The gas flow baffles serve to inhibit gas sneakage around the
precharger electrodes. The passive electrodes, supported and edged all around
by .95 cm (.375 inch) diameter rod, are rapped with pneumatic springless im-
pactors. The screen electrodes are made of .635 cm (.25 inch) hexagonal open-
ing, 79% open area perforated sheet steel and are framed and mounted on tubu-
lar supports. The corona discharge electrodes are barbed wire with a 2.5 cm
(1.0 inch) barb-to-barb spacing.
The precharger was taken to the IERL precipitator facility.at Research
Triangle Park, where it was installed, along with a sampling section, in the
test section location of the in-house ESP. The objectives of the tests con-
ducted with the system were to examine various downstream collector electrode
geometries for their potential application, to evaluate the precharger's charg-
ing effectiveness, and to determine the effect of the precharger on the collec-
tion efficiency of the precipitator.
The precipitator sections were set up with 22.86 cm (9 in.) plate-to-
plate spacings. Section 1 was initially configured with .3175 cm (1/8 in.)
diameter wires spaced 22.86 cm apart. Section 2 had a 2.54 cm (1 in.) mesh
76
-------
GAS FLOW
BAFFLE
PRECHARGER
HOUSING
HOPPER
CORONA DISCHARGE ELECTRODE
SUPPORT AND BUS BAR
SCREEN ELECTRODE
SUPPORT AND BUS BAR
CERAMIC HIGH
VOLTAGE INSULATORS
PRECHARGER
HOUSING (REF)
PASSIVE ELECTRODES
SUPPORTS
SCREEN ELECTRODES
CORONA
ELECTRODES
Figure 53. 0.47 m3/sec (1000 acfm) Precharger Assembly.
-------
discharge electrode. Section 3 was set up with .3175 cm diameter wires spaced
2.54 cm apart. Section 4 was arranged with .635 cm (1/4 in.) diameter wires
with a 5.08 cm (2 in.) wire-to-wire spacing. This variety of collector dis-
charge electrode configurations was selected so that comparisons between the
voltage-current characteristics of the different designs might contribute to
the determination of the most appropriate discharge electrode for the col-
lector section. The I-V curves for the four sections are shown in Figure 54.
The discharge electrodes in Section 1 were replaced with a 2.54 cm mesh
discharge electrode. The mesh electrode provides the high electric field and
low current density combination desirable in the collector of a two-stage pre-
cipitator. The small wire diameter and large wire-to-wire spacing configura-
tion is clearly inferior in this respect. Figures 55 and 56 show the I-V
curves of the downstream collector sections as configured for the collection
efficiency tests and at the operating temperature.
The current-voltage characteristic of the precharger is shown in Figure
57. The alignment of the screen and passive electrodes required several ad-
justments before problems with the relative spacing of the electrodes were
eliminated as operating limitations.
The precharger-collector system was operated at 150°C (302°F) for the
tests. Steam injection was not used and the moisture content measured at the
operating conditions was 1.14% by volume. These values of temperature and
moisture content of the gas stream contribute to a resistivity of the redis-
persed fly ash of approximately 5 x 1012 ohm-cm. The fly ash was injected
into the system at a rate of approximately 1.15 g/m3 (.5 gr/ft3). The total
gas volume flowrate through the system was held to 0.47 m /sec (1000 ACFM)
for the tests.
The precharger charging effectiveness was tested under the conditions
described above. The precharger was operated with corona voltage = 21 kV,
corona current = 700 yA, grid voltage = 7.2 to 8.2 kV, and grid current = 2500 to
20,000 uA. The fluctuations in the grid voltage and current were the result
of intense back corona from the passive electrodes. The passive electrodes
were rapped 21 times per minute with 5.6 x 10k kg/m2 (80 psi) air pressure on
the pneumatic impactors. Sparking from the grids to the passive electrodes
occurred throughout the charging test. A charge/mass ratio was determined
under these conditions to be -1.89 x 10~6 C/g.
Another test of particle charging was made with continuous sparking on
the grid; 6-11 kV applied voltage on the grid, 10-25 mA grid current, 16 kV
corona voltage, and 1000 yA corona current. The Q/m value measured in this
case was -2.06 x 1C"6 C/g and -2.46 x 10~6 C/g. These values of charge-to-mass
ratio have an average of -2.1 x 10~6 C/g, which is equivalent to the Q/m
values obtained in tests with the original 0.47 m3/sec precharger.
The next phase in the precharger evaluation was to determine the effect
of the precharger on the efficiency of the downstream collector-precharger
system. The collector sections were set up as described earlier and operated
with a total current of 0.01 - 0.05 mA per section. This current setting cor-
responded to an applied voltage of 20-35 kV per section.
78
-------
Z
UJ
a:
ac
3
10
20 30
VOLTAGE , kV
50
Figure 54.
I-V curves of the downstream collector section with dirty wires
and plates, no dust flow, and 300°F.
1) 0.312 cm diameter wires spaced 22.9 cm apart
2) 2.54 cm mesh
3) 0.312 cm diameter wires spaced 2.54 cm apart
4) 0.635 cm diameter wires spaced 5.08 cm apart
79
-------
Ul
DC
oc.
o
J_
10
20 30
VOLTAGE, kV
40
50
Figure 55. I-V curves of sections 1 and 2 of the downstream collector with
dirty 2.54 cm mesh discharge electrodes, dirty plates, no dust
flow, and 149°C.
80
-------
Ul
3
JL
_L
10
20 30
VOLTAGE , kV
40
50
Figure 56. I-V curves of sections 3 and 4 of the downstream collector with
dirty wires, dirty plates, no dust flow, and 149°C.
3) 0.318 cm wire diameter and 2.54 cm wire spacing
4) 0.635 cm wire diameter and 5.08 cm wire spacing
81
-------
I04
I03
10
3
<
O
OC
8
10
o
_L
10 20
CORONA VOLTAGE , kV
30
Figure 57. Precharger corona electrode I-V curve with the grid current held at
zero, temperature = 158°C, and gas flowrate = 0.47 tn3/sec.
82
-------
An optical particle counter system (OPCS) was used to monitor the particle
concentration as a function of particle diameter at the outlet of the precharger-
collector system. The measurement system is shown schematically in Figure 58.
The concentrations of particles in the range of 1.5 to 5.0 ym diameters were
monitored during the efficiency tests.
The first attempt to evaluate the precharger-collector system was made
by monitoring the particle concentration for the particle diameter region of
interest with the precharger on and off and the collector sections held to
0.01 - 0.05 mA current in both cases. An average of 7234 particles/sec was
observed with the collector on and the precharger off. With the collector on
and precharger on the number of particles/sec measured was 2566. Therefore,
the precharger effectively decreased the penetration of particles in the 1.5 -
5.0 ym diameter range by 64.5%.
The results of another test of particle concentration vs. particle diam-
eter for the precharger-collector system is shown in Figure 59. The curves
indicate concentration vs. diameter for particles in the range of 1.5 to
5.0 ym diameters for three conditions: 1) precharger off/collector off,
2) precharger off/collector on, and 3) precharger on/collector on. The dif-
ferences in particle concentration for the three conditions are consistent
for all particle diameters in the region of interest. Condition 1 measurements
yielded an average value of 15,936 particles/sec for the entire particle diameter
region of interest. The average for condition 2 tests was 6,632 particles/sec.
From these two values, the collector alone accounts for a decrease in penetra-
tion of particles in the range of interest of 58.4%. The addition of the
precharger in condition 3 measurements further decreased the particle concen-
tration in the region of interest to 2,797 particles/sec. This corresponds to
an overall system penetration decrease of 82.4% over the condition 1 case, or
an improvement in performance attributable to the precharger of 57.8%.
It should be emphasized that the measurements of particle concentration
in condition 1 tests were taken at the outlet end of the collector. There-
fore, the decreases in penetration due to the collector and the collector with
precharger do not necessarily represent collection efficiencies. Mass train
measurements taken at the outlet and inlet of the precipitator indicated a
mass collection efficiency of approximately 70% for the collector alone.
Assuming this is directly related to the particle concentrations measured with
the OPCS, an additional 12% collection efficiency due to settling can be added
to the OPCS measured penetration decrease of 58.4% for the collector. Further,
adding the settling percentage to the penetration data for precharger with
collector gives a system collection efficiency of 94%.
The tests of the 0.47 m3/sec precharger in conjunction with the IERL in-
house precipitator show a significant precipitator enhancement capability for
the precharger. The extremely high fly ash resistivity (5 x 1012 ohm-cm),
very low collector S.C.A (25.6 m2/m3/'sec, or 130 ft2/1000 acfm), and the non-
uniformity of the collector sections' discharge electrodes would tend to degrade
the performance of the two-stage system below the normal operation expectations.
Even so, a collection efficiency greater than 90% was obtained with this non-
ideal two-stage system; with a contribution to the overall efficiency of approx-
imately 60% directly attributable to the action of the precharger.
83
-------
NOZZLE
DUCT
DtLUTER
OPTICAL PARTICLE
COUNTER
MULTI-CHANNEL
ANALYZER
TELETYPE
Figure 58. Optical particle counter measurement system.
84
-------
I0
10
£
o
I02
10
15
2.8
3.4
4.0
4.4
4.6
5.2
DIAMETER , pm
Figure 59. No. of counts vs. particle diameter as observed with the Climet
optical particle counter for the three conditions:
1 - precharger off and collector off,
2 - precharger off and collector on, and
3 - precharger on and collector on.
85
-------
86
-------
SECTION 5
CHARGED PARTICLE COLLECTOR
In order to complete the precipitation process it is necessary to provide
a suitable mechanism for collecting the charged particles emerging from the
precharger. Since high resistivity particles may be encountered the problem
of back corona must again be dealt with. It is not necessary, however, to
maintain a high number density of ions in the downstream collector, because
the particles are already charged. On the other hand, if the current density
in the collector were reduced to zero, particles reentrained into the gas
stream by rapping might not be recollected due to loss of charge during con-
tact with the grounded collecting surfaces. It is thus assumed that the opti-
mum current density in the collector should be slightly less than that which
would bring about back corona.
The electric field strength in the collector should be as high as can be
achieved within the constraints imposed by limiting the corona current density.
The maximum field strength would result from the use of a parallel plate ar-
rangement of electrodes in the collector, but that would produce, ideally, no
corona current at all. A conventional wire-plate configuration would have to
be operated at a relatively low applied voltage because of the limit on cur-
rent density imposed by the presence of high resistivity materials. A wire-
plate system could be modified from conventional practice, however. By the
use of large diameter corona wires or by a much reduced spacing between wires
the current-voltage characteristics can be adjusted to provide more desirable
operating parameters.
Because a rectangular geometry offers significant advantages in flexibil-
ity and convenience in design and fabrication as compared to cylindrical or
other type configurations, emphasis in this investigation was placed on a
model employing parallel plate passive electrodes with corona discharge elec-
trodes arranged in the plane midway between adjacent passive electrodes. The
discharge electrodes could be parallel wires, an array of sharp points in the
plane, a screen, or any of several other conceivable constructions. The
screen and parallel wire arrangements were considered most attractive from an
engineering and economic viewpoint.
Computer models of several corona wire diameters and wire-to-wire spac-
ings in the conventional wire-plate precipitator configuration were executed.
Further, a series of bench-scale experiments was performed to evaluate var-
ious corona electrodes. Corona wires of 0.65 cm and 0.32 cm diameters were
tested at several wire-to-wire spacings (see Figures 60 and 61). Also tested
were 2.54 cm and 1.27 cm square mesh, and 2.54 by 5.08 cm rectangular mesh.
87
-------
10"
o:
o
10
20
40 60
VOLTAGE , kV
80
100
Figure 60. The current-voltage characteristic of five
0.64 cm diameter wires spaced 9.5 cm from a
grounded plate with a wire-to-wire spacing =
3.81 cm.
88
-------
I0a
-I01
UJ
oc
oc
o
10
Wire-to-WiTB Spocing.cm .
o « 3.81
A « 7.62
a • oo (one wire)
20
40 60
VOLTAGE , kV
80
too
Figure 61. The current-voltage characteristics of 0.32 cm
diameter wires spaced 9.5 cm from a grounded
plate at three wire-to-wire spacings.
89
-------
Figure 62 shows a comparison of current density as a function of spacing
between active and passive electrodes for several wire and screen electrode
configurations. It was found that the 2.54 cm square mesh screen electrode
performed at the highest attainable electric field strength, at a low, con-
trollable current density.
The use of a screen-type discharge electrode provides a periodic struc-
ture for corona activity. It tends to avoid a potential problem that might
exist for closely spaced wires — the development of localized regions of
enhanced corona discharge spaced unpredictably, and unevenly, along the wire,
resulting in an overall poor distribution of corona current.
In preparation for testing the two-stage concept a downstream collector
was designed for use with the pilot scale precharger. . An assembly drawing of
the collector is shown in Figure 63. Discharge electrodes were 2.54 cm
square mesh screen.
The device was fabricated and subjected to a preliminary testing program.
Current-voltage relationships for the four collector sections were made. The
two parallel gas passages in each section were independently energized to
check the electrical, and thus the mechanical consistency of the electrodes
in each section. Discrepancies between the I-V characteristics of two gas
passages or two sections could be due to the existence of local surface dis-
continuities on the electrodes, electrode misalignment, or the varying prox-
imity of the discharge electrodes to hopper baffles or other structural
grounds. All of the I-V characteristics taken in this set of tests were at
ambient conditions.
The current-voltage curves corresponding to the two gas passages in sec-
tion 1 are shown in Figure 64* A maximum of nearly fivefold difference in
current values between the two gas passages occurs in the mild-range of the
voltage values (7.2 yA to 34 yA at 32 kV applied). In the projected operat-
ing range of 50-60 kV applied the difference is markedly less. Careful at-
tention to smoothing the electrode surfaces may alleviate this inconsistency.
Also, some exposed ends of the wire mesh discharge electrode may be present
(in gas passage 2 especially) and leading to atypical I-V characteristics.
Figure 65 shows the current-voltage curves for the two gas passages in
section 2 of the downstream collector. The similarity between the two curves
is much greater in this case.
The I-V curves for section 3 are shown in Figure 66. The agreement
between gas passages is good with the exception of a large difference in
breakdown values (12 kV difference). This disparity may be due to the same
factors discussed in conjunction with section 1 curves. Section 4 I-V char-
acteristics are shown in Figure 67. There is very good agreement between
the curves of the two gas passages in this section.
Figure 68 shows the current-voltage curves for all four sections of the
downstream collector where the two gas passages in each section were electri-
cally connected, as will be the case in actual operation. The sections vary
90
-------
3000
2000
00.64 cm DIAMETER WIRES
• 0.32 cm DIAMETER WIRES
02.5cm SQUARE MESH SCREEN
A2.5X5.I cm RECTANGULAR SCREEN
• 1.3cm SQUARE MESH SCREEN
1000
800
600
500
400
300
=. 200
in
LU
01
a:
100
80
60
50
40
I
_L
I
468 10 12
ELECTRODE SEPARATION, cm
14
Figure 62. Comparison of electrical behavior for various types of
corona discharge electrodes. The wires are in arrays
of five in parallel, spaced at 3.8 cm.
91
-------
vo
4»V»F 41* «'** 4
tiff*** C»-, «~»™- ft*"*"
i-1* jr**-t #r**f**e
JH t **«-. >r MM r« *- r
* B~Ik mmrr ^
J777-D- 34
Figure 63. Small pilot scale precipitator assembly.
-------
I03
t io2
UJ
a:
QL
10
-Gas passage #1
o-Gas passage #2
10
20 30 40
VOLTAGE , KV
50
60
70
Figure 64. i-v characteristics of section 1 of the down-
stream collector with no gas flow and ambient
conditions.
93
-------
I05
nTIO2
z
cc
o
10
A-Gas passage
o - Gas passage # 2
10 20 30 40
VOLTAGE, kV
50
60
70
Figure 65. I-V characteristics of section 2 of the down-
stream collector with no gas flow and ambient
conditions.
94
-------
- io2
UJ
(T
tr
o
10
A-Gas passage*!
o-Gas passage #2
10
20 30 40
VOLTAGE , kV
50
60
70
Figure 66. I-V characteristics of section 3 of the down-
stream collector with no gas flow and ambient
conditions.
95
-------
I04
I03
UJ
DC
IT
O
10
A-Gas passage #1
o-Gas passage*2
_L
20
30 40 50
VOLTAGE, kV
60
70
Figure 67. i-v characteristics of section 4 of the down-
stream collector with no gas flow and ambient
conditions.
96
-------
LJ
CC
ce
10
o-Section I
a-Section 2
• -Section 3
A-Section 4
10
20 30 40
VOLTAGE , kV
50
60
70
Figure 68. I-V characteristics of section 1 through 4 of
the pilot scale downstream collector at ambient
conditions. The two gas passages in each section
were electrically connected for these tests.
97
-------
considerably in their electrical characteristics. Mechanical differences
between the four sections will be carefully eliminated, inasmuch as it is pos-
sible. This should normalize the electrical behavior of the sections.
The particle collector, in combination with the pilot scale precharger
will be tested in the field, on a slip-stream taken from the exhaust ducting
upstream of existing control devices at a coal-fired electric power plant.
That test program will be carried out in connection with a separate research
project under EPA Contract No. 68-02-2683, which supports work leading toward
optimization of the downstream collector design.
98
-------
SECTION 6
ENGINEERING AND COST ANALYSIS
ESTIMATED COSTS OF FULL SCALE PRECHARGER - COLLECTOR SYSTEMS
The criterion commonly used to estimate the cost of electrostatic pre-
cipitators is the number of square feet of collecting surface required to
meet the design efficiency. Although the cost of the collection electrodes
generally represents only 15-20 percent of the total precipitator cost, it
is a reliable yardstick for estimating the capital investment required for
a conventional precipitator installation. In order to approximate the cost
of a full scale precharger-collector system the same criterion will be used
as a base cost factor with additional factors added to estimate the extra-
ordinary expenses required by the hybrid system.
The determination of the collecting surface area required in a full
scale hybrid ESP is made using the data acquired in the pilot scale precharger
performance tests and the collector study. The corona current densities main-
tained in all of the experiments are greater than found in the range of cur-
rent densities (5-75 nA/cm2) used in conventional full-scale fly ash precip-
itators operating under best conditions. The effect of ash resistivity on
the current-voltage characteristics of conventional precipitators has been
described by White.1 He indicates that the current density would have to be
reduced from .1% to .01% of clean plate values in the presence of a 2 mm
thick deposited layer of high resistivity fly ash (10 -1013 fi-cm) to prevent
breakdown, i.e., before back corona formation. Thus, the current densities
maintained in the pilot scale charger tests are 2 to 4 orders of magnitude
greater than could be expected in a conventional precipitator handling
1012 n-cm resistivity fly ash.
To determine whether valid theoretical estimates of particle charging
behavior could be obtained with the data from the pilot scale experiments, a
calculation of charge to mass ratio, Q/M, (see Appendix B) for a polydisperse
aerosol simulating the particle size distribution encountered in the experi-
ments was made and the theoretical value compared with experimental Q/m values.
A log-normal particle size distribution was assumed, although this was only an
approximation to the actual size distribution with which the charger was
tested. A HMD, or Dso. of 19 pm was used, and a geometric standard deviation,
Cg, of 3.0 was derived from the particle size distribution obtained from actual
impactor data, with the approximation CfgSsDso/Die. The physical conditions
used in the Q/m calculations correspond to those in the precharger test where
the gas stream temperature = 75°C, dust loading =7.65 g/m3, fly ash resis-
tivity - 1.4 x 10 ohm-cm, corona current density » 94 nA/cm2, and Nt =
99
-------
8.63 x 1012 sec/m3. An ion mobility of 2.2 x lO'V/V-sec, mean thermal veloc-
ity of 500 m/sec, particle relative dielectric constant of 5, and particle den-
sity equal to 2.47 g/cm3 were assumed. The theoretical Q/m calculated for this
set of conditions equals 2.90 x 10~6C/g. This compares to an average measured
value of Q/m = 2.69 x 10~6C/g at these conditions, or a difference of approx-
imately 7%. Since there is adequate agreement between theoretical and experi-
mental charging values, estimates of expected performance for the pilot scale
precharger-collector system can be drawn from the experimental values of phys-
ical parameters and theoretical values of particle charge.
The predicted performance of a pilot scale system with a given collecting
surface area was determined. In order to give a conservative evaluation of per-
formance, the particle diameter which gives the poorest charging characteristic,
.2 ym diameter, was used in calculating the expected efficiency of the precharger-
collector system. A calculation of the efficiency of collection of 19 ym diam-
eter particles was also made.
In order to determine the collection efficiency of the pilot scale pre-
charger-collector system the following calculations were made:
1) charge on .2 ym and 19 ym diameter particles,
2) mobility of charged .2 ym and 19 ym diameter particles,
3) migration velocity of .2 ym and 19 ym diameter particles at four dif-
ferent collector field strengths, and
4) efficiency of collecting .2 ym and 19 ym diameter particles with the
migration velocities determined in 3).
The charge on the particles was determined from the combined theoretical
field and diffusional charging effects and is given by
| yaE'-Nt
q(a) = ™4e0 |^V4£
0
where Nt = ion concentration-time product (sec/m3),
Ep = electric field strength in the precharger (V/m),
a = particle radius (m),
e0 = permittivity of free space (fd/m),
e = electronic charge (C),
k = particle dielectric constant,
y = ion mobility (m2/V-sec),
T » temperature (°K),
K = Boltzman's constant (j/°K),
V = mean thermal ion speed (m/sec), and
q(a) = charge on a particle of radius a (C).
Values of charging parameters from the pilot scale charging experiments
used in the calculation were Nt = 8.63 x 10li sec/m3, E = 3.15 x 10s V/m, and
100
-------
T = 348°K. Other values used were k = 5, y = 2.2 x 10~'*m2/V-sec, and v =
500 m/sec. The charge accumulated on the particles was determined to be
q (0.1 x 10~6) = 2.09 x 10~18C and q (9.5 x 10~6) = 6.49 x 1Q-15C.
After determining the charge on the particles, their mobility was calcu-
lated using the expression
where q(a) = charge on a particle of radius a(C),
ri = viscosity of the gas (kg/m-sec) ,
C = Cunningham slip correction factor, and
M(a) = particle mobility (m2/V-sec).
The value of viscosity, 2.1 x 10~5kg/m-sec, is that for air at 348°K. The
mobilities for the two particle sizes of interest at these conditions are
M(.l x 10~6 ) = 9.86 x 10-V2/V-sec and M(9.5 x 10~6) = 1.74 x 10~6m2/V-sec.
The mobility is related to the particle migration velocity by the follow-
ing relationship:
W = MEc, (3)
where Ec = electric field strength in the collector (V/m) ,
M = particle mobility (m /V-sec) , and
W = particle migration velocity (m/sec).
The migration velocity was calculated for four values of Ec corresponding to
applied voltages of 30, 40, 50, and 60 kV and an electrode spacing of 9.5 cm
(19 cm duct width) in the collector. The values of the migration velocity
for .2 ym and 19 ym diameter particles at these four conditions are listed
in Table 3.
Now that the migration velocity is known, predicted values of efficiency
for monodisperse particles can be determined from the Deutsch-Anderson equation.
. -(SCA)W ,.,
n = 1 - e v ' (4)
In this equation W = particle migration velocity (m/sec),
SCA = A/V (sec/m),
A = effective collection surface area (m2),
V = gas f lowrate (m /sec) , and
D = fractional efficiency.
The collection surface area of the pilot scale collector is given to be
23.78 m2. Assuming a gas volume f lowrate of .71 m3/sec (1500 f t3/min) , the
SCA used in the calculations of efficiency is 33.62 sec/m (170 ft2/1000 ACFM) .
The collection efficiencies for the various calculated migration velocities
are given in Table 3. The projections of collection efficiency were expanded
in detail over the range of particle diameters between 0.1 and 10 ym. Figure
62 shows collection efficiency curves for two values of collecting field
101
-------
PILOT SCALE PRECHARGER-COLLECTOR SYSTEM
PERFORMANCE PREDICTION
Particle
diameter
(pm)
.2
19
(v/B)
3.15 x 10s
3.15 x 105
Nt
(sec/m3)
8.63 x 1012
8.63_x 1012
q
(C)
2.09 x 10~ls
6.49 x ID"45
M
[m2/V-sec)
9.86 x 10-"
1.74 x 10~6
SCA
(sec/m)
33.62
33.62
EC
(V/m)
3.16 x 10s
4.21 x 10s
5.26 x 105
6.32 x 10s
3.16 x 105
4.21 x 10s
5.26 X 10s
6.32 x 10s
w
(m/sec)
3.12 x 10-2
4.15 x 10~2
5.19 x 10-2
6.23 x 1CT2
5.50 x 10-1
7.33 x 10-1
9.15 x 10"1
1.10 x 10°
Collection
Efficiency
%
65.0
75.2
82.5
-100
-100
-100
-100
Table 3. Estimated performance of the pilot scale precharger-collector system.
-------
99.99
99.9
I I I I I I 111 |II1IIIH1| 1li
E «5.26xio5v/m
E « 3.16x10^/111
C
1 ii ih in mil I i 1 1 1 In ill i I i I 1 1 1 1 •
Ji iiimtil i M ill i ill i I I 1 1 1 1 1 • I i i i
0.01
0.4 0.6 1.0 2
PARTICLE DIAMETER, pm
Figure 69. Theoretical collection efficiency of the pilot
scale precharger-collector combination, plotted
as a function of particle diameter. Charging
parameters are Nt = 8.63 x 1012sec/m , E =
3.15 x 105V/m and T = 348°K. These data*corre-
spond to charging experiments where the dust
resistivity was greater than 10l ftcm.
103
-------
strength: 3.16 x 105V/m, corresponding to an applied voltage of 30 kV, and
5.26 x 105V/m for 50 kV applied voltage. These curves cover the region of
minimum particle mobility. For particles greater than 10 ym in diameter the
theoretical collection efficiency is above 99.99%.
If a collection efficiency requirement is specified, the effective col-
lection surface area needed to meet the design efficiency can be determined
from Equation 4 by using the particle migration velocity obtained with the
pilot scale precharger data and the design gas volume flowrate. Values of
Ep = 3.15 x 105V/m and Nt = 8.63 x 1012sec/m3 for the precharger can be used
to determine the charge acquired by a 2 ym diameter particle. This diameter
particle is chosen to give an estimated effective migration velocity for the
particles passing through the precharger-collector system. The charge is cal-
culated using Equation 1 to be q = 9.3 x 10~17C.
With an operational field strength in the downstream collector of Ec =
4.0 x 10s V/m, the particle migration velocity is given by Equations 2 and 3
to be W = 1.02 x 10-1m/sec. The average gas velocity in the collector is
Vc = 1.5 m/sec and the total gas volume flowrate is given to be 940 m3/sec
(2.0 x 106 acfm). The collection surface area required in order to give 99.95%
collection efficiency for 2 ym diameter particles can be determined to be
7.0 x lO1*™2 (7.6 x 105ft2). This corresponds to as SCA of 74 m2/m3/sec
(380 ft2/1000 cfm).
For a collection efficiency of 99.5% the collection surface area required,
with all other conditions the same, is reduced to 4.9 x 10V2 (5.3 x 105ft2).
This is an SCA of 52 m2/m3/sec (265 ft2/1000 cfm).
Although costs among precipitator vendors vary greatly, an average for
the erected cost of conventional precipitators as a function of collecting
surface area is $108/m2 ($10/ft2) collecting area2'3. Using a conservative
value of SCA of 59 m2/m3/sec (300 ft2/1000 cmf) for a design efficiency of 99.5%
and a gas volume flowrate of 9.4 x 102 m3/sec gives a total collection surface
area of 5.55 x 101* m2 (5.97 x 105 ft2). The basic cost for an erected precip-
itator of this size is approximately $6 x 106.
It is necessary to add to this base cost the extraordinary expenses incur-
red by having as the first electrical section the three-electrode precharger
geometry, and by using an unusual discharge electrode in the collector sections
of the precipitator. The estimates of these extra costs are absed on the de-
tailed proposal for fabrication and erection of a 14.2 m3/sec (30,000 cfm) ver-
sion of a two-stage precipitator of this design made by Lodge-Cottrell Division
of Dresser Industries. The precharger costs represented approximately 5% of the
erected precipitator total cost. A very conservative adjustment to the base
cost of the hybrid system to account for the precharger is 20%. The discharge
electrodes in the collector are expected to cost less than 1% more than con-
ventional electrodes. These adjustments in the base cost of $6 x 10 yield a
total capital investment estimate for a 99.5% efficient two-stage ESP using the
SoRI precharger and high field, low current density collector treating 9.4 x
102 m3/sec of flue gas of $7.26 x 106.
104
-------
COMPARISONS OF COSTS WITH CONVENTIONAL PRECIPITATORS
In order to accurately determine the cost effectiveness of a full scale
two-stage ESP using the three-electrode precharger concept, the costs of con-
ventional electrostatic precipitators for use in the same application and with
the required collection efficiency must be determined. Three sources have
been relied upon to provide cost information for conventional ESPs.
A cost model for ESPs burning low sulfur Western coal was developed by
David V. Bubenick of Research-Cottrell, Inc.1* Models for calculating capital
investment and annual operating costs for cold, hot, and cold SOs conditioned
ESPs were derived. The only parameters required for computation of costs are
volume of gas treated and area of the collection electrodes. For the purposes
of this analysis, the following conditions were applied: 1) low sulfur (.5 -
.7%) coal, 2) gas volume = 940 m3/sec at 163°C (2,000,000 acfm at 325°F), 3)
collection efficiency = 99.5%, and 4) a precipitation rate parameter of 4 cm/sec.
These conditions and the Deutsch-Anderson equation
1 - Aw /CN
n = 1 - e — (5)
where r| = collection efficiency,
A = area of collection surface,
Q = volume flowrate, and
w = migration velocity,
give a value for the collection surface of A = 1.25 x 105m2, or 2.25 times
as large as predicted for the SoRI two-stage ESP. Note that this value for
the collection area does not Include such non-ideal effects as rapping, non-
uniform gas flow, sneakage, and aerosol polydispersity. However, the cost
models compensate for this inaccuracy.
Upon substituting A = 1,25 x 105m2 arid Q=940 m3/sec into the
cost model for cold electrostatic precipitators, the total capital investment
and annual operating costs are $13.6 million and $2.5 million, respectively.
This is consistent with the $108/m* of collecting area cost estimate used in the
two-stage ESP case. A detailed cost breakdown is shown in Table 4. The capital
investment is 86 percent higher than is expected with the precharger-collector
hybrid system. Annual operating costs for the two-stage system are not expected
to be greater than the conventional ESP. The power consumption should be no more
that equal to the conventional ESP due to the operating mode of the collecting
sections.
In the case of the hot precipitator model, adjustments to the volume
flowrate and collection area must be made. Due to the temperature increase
from 163°C to 371°C there is a corresponding increase in the gas volume
treated: QH = -j^- x Q = 1336 m3/sec. The Research-Cottrell model
also makes an allowance for the collection area of hot precipi'tators. A
decrease in specific collection area, SCA = A/Q, of \ is achieved over that
of a cold ESP. The collection area required by the hot ESP to achieve collec-
tion efficiency of 99.5% is thus determined to be AH «= 9.22 x 10V1. Sub-
stituting these values of AH and QH into the hot precipitator cost model
105
-------
TABLE 4, COLD ESP COST MODEL
Capital Investment, $*
1. Complete collector, flange-to-flange 5.130 x 106
2. Typical precipitator accessories 7.695 x 105
3. Structural support 8.726 x 105
4. Tax and freight 5.420 x 105
5. Engineering 3.000 x 105
6. Erection and installation 5.367 x 106
7. Contingencies 6.490 x 105
8. TOTAL CAPITAL INVESTMENT 13.63 x 10'
Annual Cost, $*
9. Labor 8.88 x 103
10. Maintenance 6.00 x 10**
11. Power 3.76 x 105
12. Administration 8.88 x 102
13. Overhead 4.45 x 10**
14. Capital charges 2.04 x 106
15. TOTAL ANNUAL COST 2.53 x 106
* 1977 dollars
106
-------
yields an estimated total capital investment of $11.4 million and an annual
operating cost of $2.2 million (see Table 5 for a detailed cost breakdown).
The cost model for a cold SO3 conditioned ESP is calculated with an SCA
credit allowance of ^. That is, the collection surface area required in a gas
conditioned precipitator is 3 less than an unconditioned ESP requires. There-
fore AG = 8.34 x 10** m2. A cost breakdown for a cold S03 conditioned ESP
with this collection area, a gas volume flowrate of 940 m3/sec, and
achieving a collection efficiency of 99.5% is shown in Table 6. The total
capital investment is $10.5 million and the annual operating expense is $2.11
million.
Another estimate of the cost of conventional cold-side electrostatic pre-
cipitators was obtained from a report of the Electric Power Research Institute
(EPRI).2 Data from four ESP manufacturers describing ten large modern precip-
itators served as a base for cost estimates. "The system parameters (the same
as were used for the Research-Cottrell cold ESP model) and cost breakdown are
shown in Table 7. Only initial capital investment is tabulated.
A third source of information concerning costs of electrostatic precip-
itators is a Southern Research Institute repott5 which includes data for cold-
side, hot-side, and flue gas conditioned precipitator costs. These data are
for larger installations than references 4 and 2; however, comparisons can be
drawn. The design parameters and cost information art: shown in Table 8.
The costs of conventional electrostatic precipitators for handling high
resistivity particulate matter, whether cold-side, hot-side, or SOs conditioned,
are significantly higher, as reported from the sources quoted above, than the
estimated cost of an SoRI precharger-collector hybrid precipitator designed
for the same efficiency and same conditions. A very considerable savings in
capital investment seems possible by installing the two-stage ESP where par-
ticles of high resistivity need to be efficiently collected. Specifically,
there appears to be a very substantial advantage at large utility boiler
installations burning low sulfur coal.
Based on the information compiled here, the EPA/SoRI precharger-
collector electrostatic precipitator is very cost competitive with conven-
tional technology in high resistivity particle collection. Tests of the
two-stage system in the field will allow for more accurate estimates of
the system's performance.
107
-------
TABLE 5. HOT ESP COST MODEL
Capital Investment, $*
1. Complete collector, flange-to-flange 4.536 x 106
2. Typical precipitator accessories 6.804 x 10s
3. Structural support 6.605 x 105
4. Tax and Freight 4.700 x 105
5. Engineering 2.598 x 105
6. Erection and installation 4.244 x 106
7. Contingencies 5.425 x 105
8. TOTAL CAPITAL INVESTMENT 11.39 x 10(
Annual Costf $*
9. Labor 1.038 x 10*
10. Maintenance 8.850 x 10 *
11. Power 3.509 x 105
12. Administration 1.037 x 103
13. Overhead 4.497 x 10**
14. Capital charges 1.709 x 106
15. TOTAL ANNUAL COST 2.20 x 106
* 1977 dollars
108
-------
TABLE 6. COLD SO3 CONDITIONED ESP COST MODEL
Capital Investment, $*
1. Complete collector, flange-to-glange 3.578 x 106
2. Typical precipitator accessories 5.368 x 105
3. Structural support 5.893 x 10s
4. Tax and freight 3.763 x 105
5. Engineering 2.084 x 105
6. Erection and installation 3.713 x 106
7. Contingencies 4.501 x 105
8. Total SO3 conditionirg system
investment 1.043 x 106
* 1977 dollars
9. TOTAL CAPITAL INVESTMENT 10.495 x 10
Annual Cost, $*
10. Labor 8.880 x 103
11. Maintenance 6.000 x 10"
12. Power 2.497 x 10s
13. Labor (SO3 conditioning) 3.000 x 10"
14. Maintenance (30$ conditioning) 3.129 x 10"
15. Utilities (SO3 conditioning) 3.880 x 10"
16. Sulfur (SO3 conditioning) 6.600 x 10"
17. Administration 3.888 x 103
18. Overhead 4.847 x 10"
19. Capital charges 1.574 x 106
20. TOTAL ANNUAL COST 2.111 x 106
109
-------
TABLE 7. DETAILED AVERAGE COSTS FOR NEW COLD ESP
Item Cost ($*)
High-voltage power 838/925
Control panels 698,961
Ext. high-voltage system 206,082
Electrical devices 2,576
Casing 1,608,298
Hoppers 1,037,279
Collecting system 1,403,934
High-voltage system 589,051
Rapper system 861,251
Inlet plenum 283,363
Outlet plenum 266,189
Internal Access 55,814
External Access 121,932
Superstructure 356,350
Ventilation system support 23,184
Operating floor insulation 76,422
Hopper dust control 75,563
Safety interlocks 81,574
Support structure 1,078,725
Access facilities 501,237
Contingencies 2,033,342
TOTAL 12,200,052
*Cost corrected to 1977 dollars by assuming 7% inflation for
1975 and 1976.
110
-------
TABLE 8. COMPARISON OF AVERAGE COSTS FOR ELECTROSTATIC PRECIPITATORS
COLLECTING HIGH RESISTIVITY FLYASH
Design Factors
Hot-side
Cold-Side
TOTAL ANNUAL COST
4,185
4,007
* Cost corrected to 1977 dollars by assuming 7% inflation for 1976
Flue-Gas
Conditioned
Gas volume, 103ACFM
Temperature, °F
Migration velocity, cm/sec
Collection area, 10 3 ft2
SCA, ft2/103ACFM
Plant area required for site, 10 3
Cost, 103 $*
Base, accessories, and plenum
Flues
Support structure
Erection
Insulation
Gas conditioning
Ash handling @ $5350/hopper
Power ($856/kW)
Land @ $10,700/acre
TOTAL INVESTMENT
Annual Cost, 10 3 $*
Fixed charges @ 18% investment
Heat loss <§ $1.75/106Btu
Energy loss @ $.02/kwh
S03
Maintenance
3,640
750
8.5
1,169
321
ft2 28
6,420
1,051
708
6,146
2,806
257
2,009
6
19,403
3,493
306
304
82
2,500
300
4.75
1,411
564
35
6,420
396
842
7,013
2,248
289
2,977
9
20,194
3,635
282
90
2,500
300
8
848
339
19
4,280
368
499
4,302
1,388
1,872
193
2,310
4
15,216
2,739
367
535
104
3,745
-------
REFERENCES
1. Pontius, D. H., L. G. Felix, J. R. McDonald, and W. B. Smith.
Fine Particle Charging Development. Southern Research Institute,
EPA, Research Triangle Park, N.C., 1977. 239 pp. EPA-600/2-77-173.
2. Air Pollution Systems, Inc. Development Program for an lonizer-
Precipitator Fine Particle Dust Collection System as Applied to Coal-
Fired Utility Steam Generators: Final Report, Volume 1: Technical
and Economic Summary. EPRI FP-291. Palo Alto, CA. October 1976, p. 3-6.
3. Southern Research Institute. A Review of Technology for Control of Fly
Ash Emissions from Coal in Electric Power Generation. SORI-EAS-77-243,
Birmingham, AL. July 1977. p. 32-
4. Bubenick, D. V. Economic Comparison of Selected Scenarios for
Electrostatic Precipitators and Fabric Filters. J. of APCA, 28(3):
279-283. 1978.
5. Browne, W. R., and E. E. Stone. Sulfur Dioxide Conversion Under Corona
Discharge Catalysis. PH 86-65-2, U.S. Department of Health, Education
and Welfare. 1965. 21 pp.
6. Matteson, M. J., A. L. Stringer, and W. L. Busbee. Corona Discharge
Oxidation of Sulfur Dioxide. Environmental Science and Technology.
6(10): 895-901, 1972.
7. Parker, J. D., J. H. Boggs, and E. F. Blick. Introduction to Fluid
Mechanics and Heat Transfer. Addison-Wesley Publishing Co., Reading,
Mass. 1974. 612 pp.
8. Melcher, J. R., and K. S. Sachar. Charged Droplet Technology for Removal
of Particulates from Industrial Gases. Final Report to Air Pollution
Control Office, Durham, N. C. Contract No. 68-002-0018, Task 8. August
1971.
9. White, H. J., Industrial Electrostatic Precipitation. Addison-
Wesley Publishing Co., Reading, Mass. 1963. pp. 322-324.
112
-------
APPENDIX A
INVESTIGATION OF ALTERNATE METHODS
Limited theoretical studies and laboratory tests were applied to the eval-
uation of various techniques for controlling back corona and space charge
effects in a high current density corona field. Practical and economic factors
led to the development of the three-electrode system as described in detail in
the text of this report. Since the alternative methods explored were not pur-
sued beyond the preliminary tests, and therefore did not contribute substan-
tially to the principal goals of the project in a direct manner, the results of
these investigations are included in this appendix rather than in the main
text.
INJECTION OF CONDITIONING REAGENTS
A method for control of high resistivity fly ash by means of injecting
conditioning reagents directly into an electrical corona at the discharge elec-
trode was investigated. The rationale behind this technique is as follows:
In a charging device, both current density and electric field strength must be
high in order to provide adequate particle charging. In the presence of high
resistivity fly ash, back corona problems are most severe where the current
density at the passive electrode is greatest. Any method used to control back
corona must therefore be most effective in the precharging section of a two
stage precharger-ESP system. A controlled amount of conditioning material in-
jected through the corona region would become highly charged and driven by the
electric field to the passive electrode. The greatest amount of conditioning
agent would naturally be applied to the region of highest current density on
the passive electrode. Conditioning against high resistivity would thus be
limited principally to the particulate material actually collected in the pre-
charging section. The total amount of conditioning material required should
therefore be much smaller than that required to treat the entire influx of
particles.
Experiments were done in the laboratory to test the injection technique
with the application of various reagents of both liquid and gaseous types.
The theoretical considerations and experimental results are presented in the
following paragraphs.
Gas Injection
Initial experiments involved the introduction of a mixture of SOz and 02
into the corona region. The high electric field strength in the corona should
A-l
-------
provide the most favorable environment for conversion to SOs. Furthermore,
the ionization of 80s molecules would provide for their transport to the pas-
sive electrode. The SO 3 ions would also be directly involved in the charging
process.
Previous studies of the conversion of SOa to SO 3 and the effects of this
reaction on high resistivity fly ash have dealt with the S02 as a constituent
of the gas stream.2'3 The present study is based upon the premise that more
efficient use of the reagent can result from direct injection into the corona
region.
Theory of Ion Injection:
In order to provide enough gas to account for a corona current made up
entirely of SOs ions the minimum requirement of SOz is that which would pro-
vide one molecule for each elementary unit of charge in the corona current.
Thus for a corona current I the required number of moles per second of S02
is
n = ^ , (1)
where e is the electron charge in coulombs, and A is Avogadro's number. Since
one mole of gas occupies 22.4 liters under standard conditons, the required
volume flowrate is
IT 22.4 nT
U = 273 p '
where T is the temperature (K) , and p is the pressure in atmospheres. Thus,
from equation (1)
- 22.4 IT
~ 273 eAp '
or
U = 8.501 x 10~7 — liters/sec (2)
In previous pilot scale particle charging experiments a precharger was
used to charge particles in a total gas flow of about 400 cfm (189 £/sec)
while operating at a total current of approximately 15 mA. The minimum SOa
gas injection in such a device at 300K and one atmosphere is, from equation
(2), U = 3.83 x 10~6 A/sec. Comparing this value with the total gas flow U ,
we have
U/U = 2.02 x 10~8,
6
or about 20 parts per billion. This is an extremely small amount compared to
a level of several parts per million used in conventional conditioning methods
with S03.
A-2
-------
The feasibility of applying an ion injection scheme such as that described
in the above paragraphs depends upon S02 conversion to 80s, ionization and
transport in the corona current, as well as the nature of the effect of SO3 on
the conduction mechanism in a deposited layer of fly ash. Although the amount
of S02 injected would be extremely small, almost all of it would be restricted
to the precharger section of a two stage precipitator, where dust accumulation
would be minimized by aerodynamic design or continuous rapping.
Experimental Verification
A simple point-plane corona apparatus was set up in order to test the
theoretical concepts discussed in the above. The discharge electrode was de-
signed to admit a flow of gas into the corona region. A length of 0.2 mm
nichrome wire was secured to the inside of a piece of 2.4 mm i.d. dielectric
tubing so that the wire extended about 2 mm beyond the end of the tube. The
gas flowing through the tube thus drifts directly into the corona region sur-
rounding the end of the wire. During operation a mixture of two parts SOa to
one part 62 was forced through the tube at a controlled rate with a syringe
pump.
In order to measure the amount of injected material carried over by the
corona current, a petrie dish containing distilled, deionized water was used
as a collecting electrode. A platinum wire immersed in the water served as
a connection to the system ground.
In each experimental run a corona current was established and gas was in-
jected into the corona region. After typically 100 min. running time the
water used as the collecting electrode was analyzed chemically for sulfates by
titration against a barium perchlorate solution.
Two experiments were run with the total gas flowrate set at 0.06 ml/min
at room temperature. According to our theoretical interpretation the excess
gas would remain un-ionized and drift out of the system. After each experi-
ment two aliquots of the solution produced were analyzed for sulfate content.
In addition, control experiments were run under three conditons: (1) corona
current on with no gas flow, (2) corona current on with ambient air replacing
the mixture of SOa and Oa, and (3) injection of SOg and Og with the corona cur-
rent turned off. Only trace amounts of sulfates were found in the control
experiments. The experimental results shown in Table A-l indicate fairly
effective transport of the selected ion in the corona current. Measured sul-
fate concentrations were 74 to 84 per cent of the theoretical values.
Preliminary tes,ts of fly ash conditioning were made using this system, but
with a flat metal plate used as a collecting electrode. The system was en-
closed in an oven to produce conditions favoring high resistivity in fly ash.
An experiment was run to compare the operation of the system with and without
gas injection at the discharge electrode. First, the clean plate I-V charac-
teristic was determined for a 2.5 cm electrode spacing. Oven temperature was
maintained at 150°C. The result is shown in Figure A-l. The system was then
set to operate at a total current of approximately 300 JJA, and a mixture of
two parts S02 and one part 02 was injected at a flowrate of 0.33 y£/sec.
A-3
-------
TABLE A-l. COMPARISON OF THEORETICAL AND
EXPERIMENTAL DETERMINATIONS OF
SULFATE ION TRANSPORT
Corona Solution Volume Calculated Normal Measured Normal
Current Recovered Sulfate Concentration Sulfate Concentration
50 yA 25 ml 2.48 x 10~4 N 1.92 x 10~4 N
50 yA 25 ml 2.48 x 10~4 N 2.08 x 10~4 N
100 yA 40 ml 3.11 x 10~4 N 2.43 x 10~4 N
100 yA 40 ml 3.11 x 10~4 N 2.32 x 10~4 N
-------
700
500
300
200
QC
a:
ID
O
100
70
50
30
20
10
I I I
GAS INJECTION OFF
GAS INJECTION ON
(SAME AS CLEAN PLATE)
10 15 20
APPLIED VOLTAGE, kV
25
30
Figure A-l. Comparison of I-V characteristics for a corona system
under dust loading conditions with gas injection on
and similar conditions with gas injection turned off.
A-5
-------
Fly ash was blown into the system in puffs through a tube by an elutriator.
When changes in the current occurred after a puff a brief period, about five
to fifteen seconds, was allowed before the next puff so that the conditioning
effect of the gas could restore the original operating condition. Continuing
this procedure for several minutes it was possible to deposit a total of 323
mg of ash on the grounded electrode with essentially no change in the I-V
characteristic. There was no indication that the procedure could not be con-
tinued to deposit still more ash without producing back corona effects. After
the gas flow was turned off and no further ash was applied the I-V character-
istic remained unchanged.
The procedure was begun again with a clean plate, but with no gas injec-
tion. After the first few puffs of flyash the current rose sharply. The
current was left on, but application of fly ash was suspended. No tendency
toward a drift in current at a fixed voltage was observed over a period of
fifteen minutes; the change in electrical behavior was permanent. The I-V
characteristic was measured for comparison with the previous result. The
curve, shown in Figure A-l, exhibits a rapid rise in current with a voltage
above approximately 12 kV, indicating the presence of back corona. The total
ash accumulated was 67.5 mg, about one fifth of the amount in the previous run.
Further experiments were run after modifications in the apparatus were
made in order to provide a continuous flow of ash laden air and to control the
humidity in the oven containing the corona system.
A mixture of S02 and 02 was injected by syringe pump into the corona
region as previously described. An elutriator, operated by a continuous flow
of air, was used to introduce fly ash into the system when required. Humidity
was controlled in the oven by circulating air from an external bubbler or
desiccator source.
The general procedure used was to operate the corona system as a precip-
itator while observing any changes in current or voltage. Onset of back
corona is normally accompanied by an abrupt increase in corona current.
Injection of a mixture of SOz and 02 had little effect on back corona
when desiccated air was circulated in the system. A measurable conditioning
effect was observed, however, when moisture was present.
In a particular experiment the moisture level in the oven was maintained
at approximately 10% by volume, and the corona current was held at 50 yA with
an electrode spacing of 5 cm. Air, loaded with approximately 4.3 g/m3 redis-
persed fly ash, was blown into the interelectrode space at a rate of 1.2 1/min.
The experiment was run for 20 min. with a mixture of two parts S02 to one part
02 injected at the corona discharge electrode at a rate of 0.02 ml/min. No
change in electrical behavior was observed. The I-V characteristic measured
at the end of the experimental run is presented in Figure A-2." A total of
103 mg of fly ash was precipitated during this run.
The experiment was repeated, starting with a clean plate, but without
injection of gas at the corona electrode. After operating for 20 min. the
I-V characteristic was measured, and the precipitated dust was weighed. The
A-6
-------
1000
800
600
400
300
200
100
1 80
•*
£ 60
LJ
IT
DC
3 40
30
20
10.
NO GAS INJECTION
INJECTED
10 20 30
APPLIED VOLTAGE, kV
40
50
Figure A-2.
Comparison of I-V characteristics of a point-plate
electrode system under two conditions with high resist-
ivity ash deposited on the plate electrode. For the
lower curve a mixture of SOa and Oa is injected into
the corona region. The upper curve represents the
case for no gas injection. The higher current in the
latter case is taken to he a result of back corona.
A-7
-------
I-V characteristic, shown in Figure A-2 displays substantially higher current
values than the curve representing the experiment with gas injection. The
mass of fly ash precipitated in the experiment without gas injection was 79 mg.
In this experiment it has been demonstrated that at least some condition-
ing against back corona can be accomplished by injecting SOz and Oz into the
ionizing region of a corona electrode system in the presence of moisture.
Liquid Reagent Injection
Theory
It has been demonstrated that the injection of a gaseous conditioning
reagent into a corona region can provide at least some control of high resist-
ivity fly ash in an electrical corona system. There is, however, a funda-
mental limitation on the rate at which a reagent in the gaseous state can be
delivered to the passive electrode. That limitation is based on the fact that
no more than one ionized molecule of the conditioning agent can be delivered
to the passive electrode for each elementary unit of charge in the corona cur-
rent. If, on the other hand, a liquid were injected into the corona region
the formation of fine charged droplets might be expected to provide a greater
quantity of conditioning reagent at the passive electrode.
The formation of fine, charged liquid droplets in an electrical corona
has been termed electrohydrodynamic spraying. Investigation of the phenomenon
with regard to various applications has led to an explanation of the mechan-
isms for droplet formation. Melcher and Sachar5 have presented a discussion
of the phenomenon, summarized in the following paragraphs:
Consider a simple system consisting of a tube with small bore diameter
through which a liquid may be passed at a slow flowrate, opposite a conduct-
ing plate electrode. A voltage is imposed between the tube and plate elec-
trodes. At low voltages, large drops are formed at the end of the tube, as
shown in Figure A-3(a). With the onset of corona, sporadic spitting from the
tip of the liquid stream begins. As the field is increased the droplets
become smaller, but still remain within about a factor of ten of the diameter
of droplets formed with no applied field.
As the voltage is raised still further a significant change in the flow
configuration occurs. Dripping and spitting cease, and a steady stream
appears, which narrows to a very fine jet at the tip of the stream. Figure
A-3(b) illustrates the shape of the flow configuration with a large field ap-
plied. This stream has been observed to be as small as one micrometer in
diameter.
The shape of the stream with a high voltage applied may be explained in
terms of the tangential component of electric field at the liquid surface,
resulting from an electrical current through the stream. When electrical
corona occurs, charge conservation requires an electrical current to flow
through the liquid stream. The field direction in the stream is the same as
the direction of current flow, hence a tangential component of the field exists
A-8
-------
(a) NO CORONA CURRENT
(b) CURRENT DRIVEN STREAM
Figure A-3.
(a) Liquid droplet formation at the tip of a tube with
no corona current present, and
(b) stream formed by current passing through liquid to
corona point at tip.
-------
at the liquid surface. The presence of electrical charge at the liquid sur-
face produces an acceleration which draws the stream out into a fine jet. It
can also be shown that the polarization forces in the presence of an electrical
current tend to stabilize the stream.
At the point where the stream becomes sufficiently fine, electrical break-
down occurs and the resulting corona discharge carries away the stabilizing
current. The jet then breaks up into highly charged droplets which are driven
by the electric field to the passive electrode. The current-driven jet is a
source of both ions and charged droplets. The droplets are, generally, in the
range of 1 to 50 ym diameter.
An orifice is not required to produce a spray of droplets. They may even
be formed from a liquid film on a discharge electrode. Since the corona activ-
ity occurs at the air-liquid interface, clogging and deterioration of the dis-
charge electrode may be avoided.
Experimental Work
Several different liquid reagents were used in experiments where injection
at the corona discharge electrode was tested for conditioning effects on fly
ash. Because some of the liquids used were electrical conductors the corona
discharge electrode and the associated reagent injection system were connected
to electrical ground, and the plate electrode was connected to the high volt-
age terminal of the corona power supply.
The reagents used in the experiments included distilled water and aqueous
solutions of HzSOii, NaOH, NHi»OH and NaaCOa. Solution concentrations between
0.1 and 1.0 Normal were used. In all experiments the temperature of the system
was maintained at 150°C.
The initial tests of conditioning effects on fly ash were done by first
precipitating a quantity of ash onto the passive electrode without application
of a reagent. The operating voltage was then adjusted to a level where strong
back corona was evidenced by a current much greater than that which occurred
with a clean plate at the same voltage. The conditioning agent was then in-
jected through the discharge electrode at a continuous, controlled rate, and,
with the corona voltage held constant, the current was monitored. A decrease
in current at a fixed voltage is interpreted as a decrease in back corona
current.
Solutions of HaSOij more than 0.5 N produced detectable reduction of corona
current when applied at the rate of 0.04 ml/min for several minutes. A 0.1 N
solution required approximately 30 min. to produce a similar effect at the same
flowrate. The total amount of dust treated in these tests was 50-100 mg.
Desiccated air was circulated through the oven during the experiments.
Corrosive effects of HaSOi, were noted on both electrodes following exper-
imental runs. Formation of a residue at the end of the discharge electrode
tended to cause clogging, and the surface of the passive electrode was etched.
Lining the discharge electrode with teflon tubing prevented the clogging, but
the problem of erosion at the passive electrode remains.
A-10
-------
Solutions of NaOH and NHi»OH, in concentrations of 0.3 to 1.0 N produced
conditioning effects similar to those observed with foSOt,, but with much less
corrosive effect.
Experiments run with H20 and a 1.0 N solution of NaCl produced no condi-
tioning effect. In the latter test back corona became more pronounced as
salt crystals formed in the passive electrode. These tests confirmed that
the mere presence of water or an electrically conductive solution is not suf-
ficient to produce the desired reduction of back corona.
Further experiments were done using dust injection at a continuous flow-
rate. An elutriator operated by air from a continuous supply served as the
source of particulate matter. A settling chamber in the line between the
source and the oven removed most of the larger particles. At an air supply
rate of 1.2 1/min the system supplied approximately 5 g/m3 of redispersed fly
ash. Microscopic examination showed few particles greater than 5 urn in dia-
meter.
Using 0.5 N foSOi, injected at 0.04 ml/min back corona was controlled
throughout a 60 rain, test. I-V characteristics taken at the conclusion of
the experiment were similar to the clean plate characteristic, but with a re-
duced sparkover voltage. Solutions of less than 0.5 N HiSOi* were not success-
ful in preventing back corona in similar tests. With a 0.1 N solution back
corona began within 5 min. after dust injection was started.
A 0.3 N solution of NH^OH also reduced the effects of back corona in tests
similar to those run with HaSOi*. The I-V characteristic was nearly the same as
the clean plate characteristic except for a reduction in sparkover voltage.
The general result of these experiments indicates that a conditioning
procedure based on injection of reagents at the corona discharge electrode
may be feasible as a method for controlling back corona in a precharger
system.
HEATED PASSIVE ELECTRODE
In general, the resistivity of fly ash increases as temperature rises up
to a maximum resistivity value. Further increases in temperature result in
decreased resistivity.
Experiments done by White9 indicate that good particle charging perfor-
mance can be accomplished by maintaining the ground electrode temperature
near 300°C. The principal difficulty with this method is the requirement for
a large amount of energy to heat the ground electrodes. White reports a value
of about 4000 to 6000 joule/m3 of flue gas treated (2 to 3 kW/1000 cfm) .
The amount of heat required to maintain an electrode at a given tempera-
ture depends principally upon the heat losses to the flue gas blowing past
the electrode. Intuitively, it is clear that the electrode surface area in
contact with the flue gas should be made as small as practicable in order to
minimize heat losses. If the surface of the passive electrode is made too
A-ll
-------
small, however, the field adjacent to the electrode may become strong enough
to support a corona discharge.
Let us consider, for example, a corona system consisting of parallel rods
and wires with axes perpendicular to the direction of gas flow. The wires
serve as corona discharge electrodes and the rods are at electrical ground.
The rate at which energy must be supplied to the rod electrodes in order to
maintain them at a given temperature above that of the flue gas may be esti-
mated by calculating the rate of heat transfer from a rod to the gas. The
radiative losses are small in comparison with conductive losses.
The heat transfer problem is treated in the standard texts. ** First,
computing the Reynolds number associated with rod diameter D and gas velocity
V, we have
R - ^V
R " V
where v is the kinematic viscosity. The rate of heat transfer per unit area,
q'/A, from a rod at temperature T into a gas having temperature T is
= C n
A D R Ur L) '
where k is the thermal conductivity of the gas, and C and n are empirical
parameters depending upon the numerical value of R. Since A = TrDl for a rod
of circular cross section, we can express the heat transfer per unit length
of the rod as
•3- = TTCkRn(Tr-T)
or
) (Tr-T) .
Now, using a temperature of 300°C at the surface of the rod, the thermal con-
ductivity of the air is 4.5 x 10~2 joule/sec m°C, and the gas viscosity is
approximately 5.4 x 10~5m*/sec. For a gas velocity of 3 m/sec and rod dia-
meter between 0.1 and 6 cm we may use C = 0.615 and n = 0.466. Inserting
these values into the above equation yields
^1- 14.2 (Tr-T)D°-466
In a practical device the surface of the grounded electrodes might be required
to be 150°C or more hotter than the flue gas temperature. In order to provide
sufficient passive electrode surface area to maintain a reasonable ion current
without developing a corona discharge at the ground electrode, the cylinder
A-12
-------
diameters must be several times greater than the corona wire diameter. It is
unlikely that a cylindrical ground electrode less than about 1.5 cm in diameter
would be useful. The required rate of energy input to maintain the ground
electrode at approximately 300°C in the presence of a flue gas at 150°C would
thus be above 300 watts per meter of electrode length.
If a wire-cylinder separation of 10 cm is used the total gas flow U in
the space between a cylinder one meter long and the adjacent pair of corona
wires at a velocity of 3 m/sec is
U=2x0.1mxlmx3 m/sec
or
U - 0.6 m3/sec.
This flowrate is equivalent to about 1270 cfm. Dividing the power input per
unit length of electrode by the gas volume flowrate we obtain a value of
500 joule/m3 (or approximately 240 watts/1000 cfm). The calculated, value is
about one tenth of the power input requirements indicated by White. These
calculations, however, are based upon a minimal temperature differential
between gas and electrode, and surface area of the electrode has been taken to
be quite small. The calculated value thus represents an extreme lower bound
to energy requirements for heating the passive electrode. It is thus con-
cluded that probable energy expenditures are too great to make the technique
feasible.
OTHER TECHNIQUES
Among the other approaches considered for control of back corona effects
were the use of porous collection electrodes, semiconductor coatings on col-
lection electrodes, and wet-wall methods. These methods were generally con-
sidered to provide a low probability of success because of associated tech-
nical complications.
Porous electrodes made of sintered metal or similar material could pro-
vide a means for delivering chemical conditioning agents into a deposited dust
layer. Such chemicals, in gaseous or liquid form, would be forced through the
porous plate into the collected high resistivity material. Problems with this
concept include probable clogging of the porous plates and reduced strength of
the collecting electrodes.
The use of a semiconductor layer might provide a ballasting mechanism to
distribute current density more evenly in the presence of incipient breakdown
of a high resistivity dust layer. It is likely, however, that the semiconduc-
tor layer would degrade quickly in the harsh flue-gas environment. This tech-
nique would also be very expensive in comparison with others considered.
Wet-wall methods offer a high probability of success in a precharger.
There are, however, some undesirable aspects of such an approach. The liquid
must be well dispersed over the collecting surfaces, leaving no dry areas.
A-13
-------
Depending on the characteristics of the dust, concrete-like deposits might
occur due to the additional moisture concentration in the precharger. For
these reasons further consideration of wet-wall techniques were set aside
against the eventuality that no more practicable methods could be found.
SUMMARY
The search for a practical method for controlling the effects of back
corona in the presence of a large ion current density and high resistivity
dust included a search of the literature, theoretical studies and small scale
laboratory investigations. The most promising techniques evaluated included
the following:
A. Chemical conditioning
1, Injection of chemicals into the active ionization region,
2. Injection of chemicals through porous collection electrodes.
B. Electrical methods
1. Use of a screen electrode to remove ions resulting from
back corona
2. Application of ballasting by means of a semiconductor
layer on collecting electrodes
C, Wet-wall techniques
D. Heated collecting electrodes
Each of these approaches was considered with regard to effectiveness in
controlling back corona, technical feasibility and cost. As a result of
these studies the method involving the use of a screen electrode (described
fully in the text of this report) was developed. The other concepts are
not considered valueless, however. It is possible that, for certain applica-
tions, one or more of these techniques may be worthy of further study.
A-14
-------
APPENDIX B
THEORETICAL STUDY OF SPACE CHARGE EFFECTS
The electrical conduction properties of a corona discharge system depend
very strongly upon the distribution of charge in the interelectrode space.
The principal current carriers are ions derived from the gas molecules in the
conducting region. If the corona discharge electrode is negative with respect
to the passive electrode, a free electron contribution to the conduction pro-
cess also exists. Because the electrical mobility of the ions and free elec-
trons is very large and the electrical field strength is, generally, very high
in a corona system, the effects of macroscopic motions of the gas can be
ignored in determining ionic drift velocities.
When suspended particles are present in a corona system they become
charged by ion attachment in accordance with well-known principles. If the
number density of the particles in the interelectrode region is large the
charge on the particles can make a significant contribution to the overall
charge distribution. In contrast with the ions, however, the drift, or migra-
tion velocities of the particles, due to the forces exerted by the electric
field, are ordinarily much smaller than the velocities associated with the
motion of the gas. The motion, and hence the spatial distribution of partic-
ulate matter in an electrostatic precipitator is, therefore, strongly domin-
ated by the presence of any turbulent motion of the gas. This conclusion is
supported by empirical studies which show that the concentration of dust in an
ESP decreases exponentially along the path of the gas flow through the system,
as expressed by models of the form of the Deutsch equation.
An effect of gas turbulence in an ESP is to produce a constant mixing of
the aerosol, which tends to promote a homogeneity in the spatial distribution
of the particles. Statistically, the charging conditions for the particles
may be taken to be uniform. Thus, to the extent that complete and continuous
mixing of the aerosol can be assumed, the space charge associated with the
charged particles can be considered to be uniform along a cross-section of an
ESP. The overall charge distribution in the interelectrode region therefore
consists of a superposition of the mobile ion charge distribution on the rela-
tively fixed and nearly uniform (in a given cross-section) charge distribution
associated with the particles.
PARTICULATE SPACE CHARGE CALCULATION
Given a particle size distribution and a set of physical conditions for
charging, a calculation can be made to yield the ratio of charge to mass for
a polydisperse aerosol. Let T(a) be the probability amplitude of the distri-
bution as a function of particle radius a, and let q(a) be the charge per
B-l
-------
particle of radius a, for given values of electric field strength, ion density,
residence time, ion mobility, gas temperature and particle dielectric constant.
The ratio of charge to mass for the aerosol is
Q/M - -{I ?(aWa)da , (B-l)
J 0 ;,TTa3pT(a)da
where p is the mass density of the particles. Multiplying the ratio Q/M by the
mass loading yields the space charge density due to the charged particles.
Equation B-l cannot generally be evaluated in closed form, but will require
the application of numerical integration techniques. The function T(a) may it-
self be difficult to express algebraically for an actual aerosol. However, in
order to demonstrate the application of Equation B-l to find reasonable space
charge characteristics, we will evaluate Equation B-l for several log-normal
particle size distributions, defined by
where a „ is the geometric mean of the numerical distribution and Og is the geo-
metric standard deviation. In those cases where the mass median radius agm is
specified, agjj is derived from
log agN = log agm - 6.908 Iog2ag . (B-3)
There is a number of charging theories available for prediction of q(a).
A fair approximation is the sum of the classical field charging and diffusion
charging expressions:
(B-4)
C 1 JVT^./ t I J.TII11LCHJ
and
/I +
\ -
f , _ akT I + avTTe2Nt\ ,R ex
3D(a) - IT ln - 2kf - (B"5)
respectively, where
e = electron charge
E = electric field strength
k = dielectric constant of the particulate material
K = Boltzmann's constant
N = number density of the ions
t = residence time of the particles
T = absolute temperature
v = mean molecular velocity
y = ion mobility.
B-2
-------
Now, using q(a) = qp(a) + q^Ca) along with Equation B-2, we can evaluate the
overall particulate space charge defined by Equation B-l. This has been done
for several examples, using an HP-65 calculator to carry out the calculation.
(HP-65 program listed at the end of this appendix.)
Figure B-l shows values of Q/M calculated as a function of Og for four dif-
ferent values of mass median diameter. For all of the curves shown in Figure
B-l the electric field strength was taken to be 4 x 105V/m, the ion density-
residence time product (Nt) was set at 1.0 x 1013sec/m3, the ion mobility was
1.8 x 10~Mm2/Vsec, absolute temperature was 295K and the mean molecular velocity
was 500 m/sec. The particles were assumed to have a relative dielectric con-
stant of 5 and density of 2.25 g/cm3. As the distribution broadens from mono-
dispersity (og = 1) there is a monotonic increase in Q/M. The significance of
the shape of a particle size distribution can be illustrated by the observation
that the value of Q/M for a particle distribution with mass median diameter of
20 ym and Og = 3 is approximately the same as that of a monodisperse aerosol
consisting of 5 ym diameter particles.
In Figure B-2 the ratio of charge to mass is plotted as a function of mass
median diameter for log-normal distributions of particles. Five values of Og
were used in the calculations.
EFFECTS OF PARTICULATE SPACE CHARGE IN AN ESP
The number density N of the ions in the interelectrode region of a corona
system can be calculated in terms of the current density j , the electric field
strength, and the ion mobility as follows:
N - (B-6)
Because the drift velocity of the particles is much smaller than that of the
ions the charged particle contribution to j can almost invariably be ignored,
even though the space charge associated with the charged particles may be a
substantial part of the overall charge distribution.
In order to compare the ion number density with the particulate space
charge, we shall define the quantities ?i and £p as the charge per unit volume
in the conducting region due to ions and particles respectively. For singly-
charged ions
Ci = Ne
or i
Si=^ . (B-7)
The value of £p is the product of Q/M by the total mass of particulate material
L per unit volume of aerosol,
5P - ||L CB-8)
By way of example, let us consider a full-scale ESP operating at an average
current density of 1.2 x 10~3 A/m2, and an electric field strength of 4 x 105V/m.
B-3
-------
o
o
a.
I I I I I
E=4.0x|08V/m
= l.0xio'3sec/m5
i I i i i i I i i i i * i i
i i i i i i I i i
1.5 2.0 2.5 3.0
GEOMETRIC STANDARD DEVIATION
Figure B-l. Overall ratio of charge to mass for log-normal dis-
tributions of particles as a function of geometric
standard deviation. All charging parameters were
kept the same for all calculations. (Dm is mass
medial diameter.)
B-A
-------
g
o
0.
E = 4.0X I05 V/m
Nt = I.0xi0l3sec/m5
3 4 5 6 8 10 20 40
MASS MEDIAN DIAMETER , pm
100
Figure B-2.
Overall ratio of charge to mass for log-normal dis-
tributions of particles as a function of mass median
diameter. All charging parameters were kept the
same for all calculations.
B-5
-------
-Using a value of 1.8 x 10~um2/Vsec for the ion mobility, we obtain from Equation
B-7 an ionic charge density of 1.33 x 10~5 C/m3 . These electrical quantities
are consistent with those used for calculated values of particulate charge in
Figures B-l and B-2.
Suppose the aerosol passing through the ESP had a dust burden of 3g/m3 with
mass median diameter of 10 ym and Oo = 2.5. Using Equation B-8 and reading
Q/m = 4.6 x 10~6 C/g from Figure B-l, we find that the space charge per unit
volume due to the charged particles is 1.38 x 10~5 C/m9, or just slightly
larger than the ionic space charge density.
In attempts to predict the behavior of an ESP on a theoretical basis a
proper accounting of the space charge effects has proven difficult. One approach
is to use an "effective mobility" derived by taking the. measured current to be a
sum of ion and particulate contributions. The effective mobility is then used
in the computation of voltage-current relationships. This approach is based on
the assumption that the motion of the particles depends principally upon the
electrical forces, while ignoring the gas turbulence effects.
In the opposite extreme case the contribution to the measured current due
to the presence of charged particles can be ignored. The space charge may then
be calculated directly from the particle charging theory for a given or assumed
particle size distribution. If the resulting particulate space charge is sig-
nificant the electric field strength may have to be recalculated. Since the
charging conditions depend upon the space charge, and vice versa, a self-con-
sistent solution must ultimately be sought.
HP-65 PROGRAM: Q/M FOR LOG-NORMAL DISTRIBUTION OF PARTICLES
In order to determine the overall ratio of charge to mass in a log-normal
distribution of particles, given the total particulate mass per unit volume of
aerosol, the geometric mean radius ao, the geometric standard deviation (Jg, and
the charging conditions as defined for Equations B-4 and B-5, a numerical inte-
gration scheme is applied. The particle distribution is divided into logarith-
mic increments, each defined by limits aj and 3a j , where 3 > !• The charge per
particle is then computed for the midpoint in each increment (a? = /$aj) by
+ i}.
in accordance with Equations B-4 and B-5 (MRS units), where Ci = 4e./e and
k/e.
Now, the number of particles in increment j is
Nj = Y(aJ)(0-l),
where (a?) is the amplitude of the distribution,
1 = fa log a exp f-(1°8 ^^l -
g I 2 (log ag)*J
B-6
-------
Thus, we have for the total contribution to the charge resulting from par-
ticles in the j increment,
Qj = qj (aj) (0"1)
The resulting values of Q. are summed, as are the contributions to mass from
each increment, and a ratio of the sums over the entire distribution is taken.
User Instructions
Part A
STEP
1
2
3
A
5
6
7
8
9
INSTRUCTIONS
Enter dielectric constant
Enter ion density-time
Enter ion mobility
Enter electric field strength
Enter temperature
Enter mean molecular speed
Enter mean particle radius
Enter geo. standard deviation
Read in Part B
INPUT
DATA/UNITS
K
Nt(sec/m3)
y(m2/V-sec)
E(V/m)
T(K)
v(m/sec)
a (m)
°g
new card
KEYS
1 A | 1
fRTTil 1
| R/S | |
rwsir i
i — II — I
1 II — 1
l ll I
1 — II — 1
rni i
OUTPUT
DATA/UNITS
Part B
STEP
1
2
INSTRUCTIONS
Read in Program B
Go to Part C
INPUT
DATA/UNITS
KEYS
A
1
mi i
OUTPUT
DATA/UNITS
Tart C
STEP
1
2
INSTRUCTIONS
Enter particle density
Read Q/M
INPUT
DATA/UNITS
P(kg/m3)
KEYS
All
Jl
OUTPUT
DATA/ UN ITS
Q/M (Coul/kg)
B-7
-------
Program Listing
Part A
CODE
23
11
41
41
01
51
35 07
02
61
81
02
71
01
61
33 01
02
83
02
01
00
07
43
08
33
09
33
71
01
08
83
06
01
07
43
KEYS
LBL
A
ENT
ENT
1
-
g,x=^y
2
+
•
•
2
X
1
+
STO 1
2
•
2
1
0
7
EEX
8
STO
9
STO
X
1
8
•
6
1
7
EEX
CODE
42
05
33 08
84
33 07
84
71
41
41
81
71
35 07
34
09
61
81
33
71
01
35
02
33
71
01
01
83
06
00
02
42
42
01
09
KEYS
CHS
5
STO 8
R/S
STO 7
R/S
X
ENT
ENT
R/S
X
g,x^y
RCL
9
+
t
•
STO
X
1
9
IT
STO
X
1
1
•
6
0
2
EEX
CHS
1
9
CODE
33
71
01
35
02
71
34 08
71
34
09
71
84
33 06
71
33 02
34 07
34 06
81
34 08
81
34
09
81
84
71
33 03
00
33 08
84
33 05
35 08
33 06
24
KEYS
STO
X
1
a
TT
X
RCL8
X
RCL
9
X
R/X
STO 6
X
STO 2
RCL 7
RCL 6
i
•
RCL 8
•
RCL
9
•
R/S
X
STO 3
0
STO 8
R/S
STO 5
9 -V
STO 6
RTN
B-8
-------
Part B
CODE
23
11
00
33
09
33 08
33 07
34 05
03
42
35
05
34 06
71
33 04
23
12
34 04
34 05
83
01
35
05
71
33 04
34 06
81
31
08
34 05
31
08
81
41
KEYS
LBL
A
0
STO
9
STO 8
STO 7
RCL 5
3
CHS
g
yX
RCL 6
X
STO 4
LBL
B
RCL 4
RCL 5
•
1
g
yX
X
STO 4
RCL 6
-
t
log
RCL 5
*
log
-
ENT
CODE
71
02
81
42
32
07
34 05
31
08
81
41
34 05
83
01
35
05
01
51
71
41
41
34 04
03
35
05
71
33
61
08
35 08
41
34 04
34 03
KEYS
X
2
•
T
CHS
r1
In
RCL 5
f
log
•
i
ENT
RCL 5
•
01
g
yX
1
-
X
ENT
ENT
RCL 4
3
g
yX
X
STO
+
8
g, R+
ENT
RCL 4
RCL3
CODE
71
01
61
31
07
34 04
71
34 02
71
34 04
41
71
34 01
71
61
71
33
61
07
34 05
03
35
05
34 06
71
34 04
35 22
12
35 01
24
KEYS
X
1
+
f
In
RCL 4
X
RCL 2
X
RCL 4
ENT
X
RCL 1
X
+
X
STO
+
7
RCL 5
3
g
yX
RCL 6
X
RCL 4
g,x-y
B
g, NOP
RTN
B-9
-------
Part C
CODE
23
11
41
35
02
71
04
71
03
81
34 08
71
35
04
34 07
71
24
KEYS
LBL
A
ENT
g
IT
X
4
X
3
•
•
RCL 8
X
9
1/x
RCL 7
X
RTN
B-10
-------
APPENDIX C
PRECHARGER
C-l
-------
n
I
ro
1 —
1
1
!
L
1
I-I.-VB-1K-.V-J
T
1
i
~ ~~ ~ ~\
1
i
I
4
I
»•
I
„ Jl^_
•~1
t
!
-I
1
i
\
t
•
. j
B
SECTiON C-C"
SOVTMM BJUIOI MlM
il«MIMOMA*>. ALABAMA MJ
-------
n
i
1
s
1
1
®I
-4-
1 —
1
Jt,
^i
"•t
~~
•ft
-
1
.ft
. (ite,
~J*\ '
! x^
c
*
ft
— 1
- e ^ e <
GENERAL NOTES
ALL MATLRIAL SM^LL Bt u &AG£
M£TAL SKi.tT,tJHuLt4 NOTED.
SHARP CORNEBi. BuBH 11,
iPATTER «UiT Bt REMOVt
')M-
.| f ' L
— 'D
j (£_} T^
L*; |i
1 ;^r
'V ! ;
x\ 1 1
n - ( r - i u..
DETAIL "C1-/
SECTION'B-B"
DETAIL "C"
GAS
^
FLOW
BAFFLE fTY^
ONL1THIS
"P--
5ECTION 'AA"
4|" - D r,.
-------
n
[~»AIKET -ALL
ANGLES FOR FRONT ACCESSES
MATCH 0»«.L S/«(0.y?S)»Ol£S IN ANGLEi 1Q APPROPRIATE
OOOH
n n
2 BM'D i !.9»l.5»iOflAfttl Ift &HCQ-O i-'-JiLa «OtAB[ll5
ANGLES FOR TOP ACCESS
T9
4
7.1S I l.ll
3
ALL PLATtS IO GAGC % UNLC*» NOTfO.
*ioc-_ -P 05
-------
DETAIL :
DtvEimD v«*
n
i
turn »««IC*TW
Td nx,,-
r* (wrrni
i* OACC PLATIS
I/MO ro*
oCVftdlhto Vfkn
f
* X_
.—
rr
Sjpr
-------
TYPICAL MESH
COW. (DETAIL AA)
NOTE;
«,2S MEI
LLEVATION : SCREENS AND TUBES
PLAN
NOTES:
C*M-^ t^JT
HL*T BC. T
CLKT PWIT ON
C vtlj> SOKTrtR ftUST
*BlvE i fcuQiii^lum , ^TIC-M II
BOTTOM END,DETAIL,ITEM
^? 6LOT -rt^^OTIT in vfeftK Pwt.Tt <
MffijjVn * i oc 31 -U -
-------
I a» D'* BCXJI
) WITHt_OCj(
JTS
5 _
w
JL
ftrTi
t ifij
/
•
i T
i
' fr
V
1 1
I
I
jk
ftp
" r
.'WiAD '»-
AtSt^ec
L il"
SECTION
^r±;^ n
^,. -^« *f* frT^B -^ I
* NOTFS :
VIBRATOR PLATE
SCREEN ELECTRODE SUPPORT
CONNECTION TO INSULATOR
CORONA ELECTRODE SUPPORT
CONNECTION TO INSULATOR©
tiCC-31 -I-Qfe
-------
i._ j?5 -^I_.£.-.Lfc*_*i±
n
oo
r^fc^_**:' kfva
fyj^^^.r
•m;4?J£
t-X** t^«.H*t JLlfelJiS.^
CM.^^* ft^fetVA-,
ft. 7 $ L, 33 !?1f 11 ill °BT ha>IT
4IOO-3I-D-07
-------
! i\
dt
I
M /
11 /
n i/
DEVELOPED VIEW. ITEM
FORMING SEQUENCE
3=
ETlTVn
-------
n
I
3.0 DIA.
2.0 WA.
II
SECTION) A-A
TCHERANGES UNLESS
OTHERWISE NOTED
FRACTIONS
I- g
DECIMALS
i o«o
ANGLES
FINISH
AFfMOVEO
CHECKED
DRAWN
REVISIONS
ZONE MO.
SOUTHERN RESEARCH INSTITUTE
iHMINGHAM, ALABAMA 35205
PLATE SUPPORT GUIDES
HEMS
SCALE
DATE
-B-
-------
APPENDIX D
COLLECTOR
D-l
-------
-------
- P-3
SOUTNBdl KSLHKH IMSIIIUli
MUUNOHAM. AIAkAMA ISMS
ST,f!^"°"°- 3 77 7- C -//
-------
(«».)«ft.S
\
-»
...
(w/)
*5^
r A-***.S (f f 9
. - . . i* - 4 I... i i; ' -
,=..-™4™..s .,»,,-flm«_-.
/ t" ~ - .,4::.----.'.:
/ ' <1
,
•
ij-
«•
H*
4
J i
f *
J _
*-
j
••:•<••':::'••'??.
^*~ ./n
"^ &Ptt-& *Y P
-------
' —» I ITf*1 I3O\
-. -» a. j-rwKr (J»'»-
. —» «. ITf* II* *
SOUTHEM KSUKH
iltMINOHAM. MA1AMA »5205
3777.
-------
-------
j , I
1*1
, ivn.t.. It f-4f* (.cfo)
* 1*610;
I*
*-
<
r <-' nrvs
4
».
»
' I»M> vr IP* -^
/~1
iL
5
» r
CV &
SOUTHEIN KSEAKH MSTirUIE
iHMIXOHAM, AiAiAMA 3M05
PILOT PKt an
3777
-------
0
03
PLATf.
fg G**r .
tf-t t
.j J a/4f
ITfr HOT.
SOUTHEM KSEAKH MSmifTi
iHMIMGHAM. ALA1AMA 35105
PILOT
Cat-i. gc-T/o A;
- lie.
-------
(£)
ycffftj ft-*cnn>pr^B)
t-
s4 L 2
f -\__— -W* *r'a
f|:s ?::::::: ::::
S^ |E iEi::iEEE!:Eii
f^:S:::l:§
i [ 1 1| 1 1 1 1 1 1 1 1 j 1 1 1 [ ) i
8 ::::::::::::::::::
y IIHIHI IM 1 1 Mil
*<.
ft I
^
11! I5l!|![!i!!:i!j|
j -d J L^_ Xft -i -*•
LJ | j 1 .< — 1_ -j- -
ft -"ir — ^-••"ti5±
-r-1 : — ^-r-l r , ' ' j ] -
P '; 1 ! 1 1 ' ' ' it11 -
--H-J — ' — HT-| H'T
1 , "|" . .
, • J .
P - - - — --- i
llll!llllllllllll\111111^
J® 6
cofjftef. of-mn-
PLU& wti-o AT
i/on <*"* Pitff
SfvaTH
170
/t,
-fop (fill. 30+ T.f I "Tvee *.O3f IU4O.
t-oue-
SAftf A$ i-ifl ffftf-r
,f3f t
p/ef.e_, **+ s.s.ji'
t. X. 1ft L0MO-
>f- j
i . CUT
r^ fi/t/uisf lot S.S.
t-o/us-
lT£M 110 1
ITff If l
'' \ Pivs, u,fLP p**. PTT,H.
:U J
jc Ha-erl 3 «-5
10'AT/HIT'. Top cue
iitjm mm
AT t+AH&fR. SOC/tfrnA
Iff t-
/ 'rff 5
(ft vtt-D M+ta&fK BLOCKS Iffftf it/ro -nf
KAH- 1101 € fff CV/LL. BLfcxro H/M?« H6t£S
M" oteP Mi"-
(l) WfLO TVO cexve*. PtrrfS /•»•» c/*t -TO
»f> K*IL. /-rof.
sii-ire* soi-ofn JCffftv tier /v-r0
f*et.fs t" -r~f>f if A a, i-ro I.
IfJSSKT 3C"ee-f "mm* /MTU S'Or
i-rej t Sii-vfK ioi-ocfi. A, rf*s i
Ct-AMf f?Alt-S TOIfAec AHO /* T rue
Ol riff S'af AHO H'fi.e i*J
fr) CaM
0)
ALL. SHARP
TO1EHAMXS UNLESS
CHECKEB 0. P.
sourmm KSURCH
iltMIMOHAM. AIA1AMA 3ii
-------
a
-------
^ sru-r
_ 3O3.
Hft. CO.. Wftvitu, /*A-iS.
SOtfTHEM KStAKH MSTITUn
il«MINGHAM, AlAiAMA 35205
-------
u
i
(->
to
*.«.
JR*
1
i.7
//.
3.
J
^»
^3
\
9S
f
\
>,
-
8
^r-T
) 1 !|
' H
1 H
1 II
' 'I
1 1
1 -4r4
\
Ik
»
1 - t
1 II
i '!
1 ''
11
1 II
, \ i
Lull
•r
— ^~T
u
u „,
~J
d
2fc
.
1
f \ J
T /
<,
/^
if.
1
i
t~
i
>
3
6
1
i
i
i
1
I
)'i
1
_ /¥«*. ». * SfL '- '(r S3
r, //ffx J4^ 5^ -// UH , If-t sS
_ 1/gsl g OSf-
. Pve""*-ric- j cvn "feet, life i I
(,*(,* to, 31 iff /Tl PT.I.B, /M*ft.r
-» S.f.. fff* //*-*
I-4
l O
i-rri t.no
FRACTIONS » 1.-
SOUTHUN KSEAKH MSfHUTE
tltMIMOHAM. ALAJAMA
Pfeci
-------
r
*JU*|
-l*k
(Z»/) AS~Sf*o*-/u
a
-h
(a* W (ir/t)
/
o
~, Afi $ t-£
As
'
ftl LAJt
gOLT j fffx tte*0, S-J £-10*3
HISTJ /HVO p-utT w*tff£r-S.
ts •
ITffl tit I
I Tff 3. 10 ~i~
IT CM 7t«'3
ITffl t-i-ff
eon TO ~ioimT ne
\3A AID 'M ft- '*"-*
t r
CMtOctD P. f.
SOUTHIM HSEARCH BCTITUn
JUMINOHAM. AIA»AMA 3S2M
-------
I
lELstjlal-ftl
© @"-
X \
t-$
/
"1*
_.l*
1C
\
\
PGlveCASE ('•317)
r TJTJF 7-
iff — "3 j-
- -
v -TfiaoG-H , t-"
««.T iSt
UMf. BtLT
CKS / ''
? j.
.- «<»*OCIr
£)/*/<./* "6. V.f,,7/b Tf
eos rcvu /* *• s ff /ff -
3 o/j»/^
09ST&M
t PITCH JOT, f-.1t
, s. y /-t r#*u'^ i"
JO —/
. 91
* *»o
(10 t
"*
-1 \
V
TOLERANCES UNLESS
t 010,
SOUTHERN USEAKH
HKMINOHAM, ALAiAMA M205
DfiT
37 7 f _c_ 23
-------
li
o
I
it
i
k
1
1
1
i.
2
.^
/"
DX
-1-
/r
/7
«
*J
p
i
P
,
M-
/<
* ^
I
>•
/
tor
\
<
j
x
\
^ '
\
\
f
t
1
L
I
^-,
<
i
/•
\
, •
7~(ie
—\
tr
"
»T.I
j
T
i~*IM*;~l^h~
///i
_L
1010
P.y, 6.
SOUTHEM KSEAKH itSTHUH
»I»MINGHAM. AlAiAMA 352O5
-------
THHT
(-•
cr.
US'
* z
• ,. I f-fr S
i. p. 2 " "• C
It
CUT
,0f* * t,X (
1 isoz -m
TOLERANCES UNLESS
OTHERWISE NOTED
FRACTIONS +
- g
DECIMALS
ANGLES
1 010
APPROVED
DATE
REVISIONS
ZONE NO
SOUTHERN RESEARCH INSTITUTE
BIRMINGHAM. ALABAMA 33303
TITLE
PILOT f#f a? i rvf T2v^
Yl-
DATE ,
3 777 -B-
-------
f/ffiS
tux
fCj It/O }.S. j
f, TOP, It/f S.t. j
s/ofi- ^£ t it/t T.t.j 1 9
T*f A- , it/rs.S^ I? S-^ffo) x /Z x/
S Plpt; tt/t f.S., SCMfO. 10,
4-.05- ' ' '
*
JC !«>./
A ;r v #/
I, ffr/t S.f. j *ib r- f
its* ne»o j r*-«-i,
IT£"I
i -ret ife r~
/ren ifcoff
»fT7-C-
lOUMHCffUMBI
frfcOj-|®
I — C:
—' i ~.-~» I mt rn.
SOUTMOM KSEAKH Msnran
tltMIXOMAM, AlAtAMA HIM
PILOT
HITTD •»"«.
377V
-------
a
M
00
ai
•i::i
-.-& 4
r/n it (Sift
t-r^
•Ot
1 A-'- '
it H*tfcl
!«,<•»"
hH
^
e/«s. M«)
~i*
'• i
LI
re^j^/^tf^ov^j^tjuga M.
P &&
-------
GKIT.
See "rs i
lt»f
• oro
E if-tf-s.
3ff , v*">
~
to*
: ptArr, it-r s.S. ,it t-*f.r
8 t it* , 0FM> to' TO /:
ITf* 2WC» OLiff
iff J
SfoT
j If-8 s.S.j A/u6ia l
t. ROMD 8' ft>
OPTIONAL. : us
is \e**y otosr TO
TDi£MANCCK UNL
OTIIC
SOUTWM RESEAKH MSIITUTE
JUMINOHAM. AlAtAMA 3S20S
f-3o-~V
3 777 -C- 2 y
-------
aa ram mag
ft-ccp /t-ys.S.,
*•
K-/T-// K, e;
SOOTHiM RESEAKH MSTITUTE
iltMIMOHAM, ALAiAMA MM5
/V«. (7 7" fHtC(fl TA Tff*.
-------
Ke OAi
l
; i
^
i
i '
1 1
i i
1 1
i i
i ,
i|
1 1
1 1
1 1
1 1
1 !
| 1
1 '
1 1
I 1
i
1 1
i !
1
^
— r-
^
- i
•-
-_-_
_>
i
|
i
-4-
x^
•1
-4
<"
.-1
I .
I
I
i »
'
I
I »
i
1 •
i
1
1
ij
2.?
feflw
si
J
aft
.£-.
TP Fir
IHHOS
tlELOHCHT.
IS"
Nan.
ft
f
-*15W^**(U.
Cff-3l fXP
U-^ £* o*e>-
» *"-$•
'y it 0
»vS, you- f.f.
W
If/ OUT H-apfSK 3177-O-lt-A
SOUTHEM KSEAKH
JIMtlMOMAM, AtAIAMA 35M5
PILOT
r TIOlu of
-------
o
i
ro
tsJ
-------
^
G>
10
1 1
.1 J
It.
u 5i_j_$-{ j
•
z
«_
*"
5roT HOLD
•jm.
JX*
31*7
nor
•
/ r 6-*f-cr , c/fo x 3. If x
tt+rr , tws- c ****** fM>r. ir-t- tj.
/r
»*>»
T»«r/** tHfr
niiencM
{•»-Tr T&f urrm J. f Fltuft
SOUTHEM KSEAKM MSIIIUli
•IftJUI
fffffif /•*
-------
13. F r+f BfHO TC //.<«/
.ore* 13. g x- 1ST
.Sfo K/f.fx fi t}f*u re ni xz.
-rap
.os-o x is-.a
, it--rs.r.,
«rwo n n,i » 2.
SOUTHUM KSEAKM
ilUUHQHAM, AlAiAMA MM5
-------
U IU
1'
IrT
-jLiL-
1
[f*^"<«* ASH*
f7 rifo utrrm
»J'V ZJ-X stf*** ivt.vfv
Svrr*t.y *
*e~rx
-------
to
IB tan
t -t- T.S,
•( 11,1
r-t-4t.se. t-atje- /t-g-S.S.
2 * X
rt.4--re, it- r S. S. it
,Ofo K i.t(. < /J~, it f tie /tj-r o
AH Si-E ».SV « S.«yj- llui C,S
SfoT yaet,o CLIPS
n soemrvre f*>e f»rri
__BW« TO*-^
SOUTHIBI KSEAKH MSTITIITE
, AlAIAMA
f/i. O T P&ECIPI r* TO/%
-------
ro/| / I Ff-+»&e. 4fw-r*x. . rrt
\-tf*l(ffy} P'A<"
X NO
3 (fur I rf*t ISOf
SOUTWW RESEARCH MSTITUTE
irKMINOHAM, ALA1AMA 3S1OJ
flt-OT
-------
NJ
CO
•m-o vot-r
6 ^w AT
r fim-rtt. y.f,
«*7- f
X -f to«/s_ ftc
.
P#e&f>fr+-rr/t 7-0
Tc.
Ltx-41.
SOUTHEM KSEAKH MSTITOTE
HtMINOHAM, AtAtAM* 35IM
-------
e
to
VO
\__.
^
/
1
\
':
/
/
/
L
r** fee
y-m--0- IO
-------
TECHNICAL REPORT DATA
(Please read InUructions on the reverse before completing}
1. REPORT NO.
EPA-600/7-79-189
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE _ , „,,..,
Electrostatic Precipitators for Collection of High
Resistivity Ash
5, REPORT DATE
August 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
D.H. Pontius, P. V.Bush, and W.B.Smith
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
68-02-2193
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
Final: 11/76 - 1/79
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES IERL-RTP project officer is Leslie E. Sparks, Mail Drop 61,
919/541-2925.
16. ABSTRACT
rep0rt gives results of a. research program to: (1) compare various
electrode systems for charging fine high- resistivity dusts; (2) investigate techniques
for charging the dusts in a high current density corona system; (3) perform a labora-
tory scale feasibility study of selected charging systems; and (4) design, fabricate,
and test a 0. 47 cu m/sec (1000 acfm) pilot-scale precharger for application to a two-
stage system for electrostatic precipitation of high resistivity particulates. A litera-
ture review of previous attempts to control back corona caused by high resistivity
dusts , and limited theoretical and experimental investigations : eliminated the imprac
ticable and evaluated potentially useful approaches to the development of charging
systems for high resistivity dust, and resulted in the derivation of a new three-elec-
trode particle precharger , upon which further developments were based. The three-
electrode concept, tested in a small laboratory device, charged high resistivity dusts
to levels achievable only on low and moderate resistivity dusts in conventional sys-
tems. Charging results remained good for a pilot scale system designed, built, and
tested at a gas volume flowrate of 0.47 cu m/sec. A rugged version of the pilot
scale precharger was tested as a part of a two-stage system, where the collector
was a modified pilot scale ESP. The new technique has economic potential.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution Coronas
Electrostatic Precipitators
Dust
Ashes
Electrical Resistivity
Electrodes
Pollution Control
Stationary Sources
High Resistivity Dusts
13B
131
11G
21B
20C
09A
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
187
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
E-l
------- |