EPA-60C/2-77-002
January 1977
Environmental Protection Technology Series
COMPARATIVE U.S./USSR TESTS OF A
HOT-SIDE ELECTROSTATIC
PRECIPITATOR
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development,
U.S. Environmental Protection Agency, have been grouped into
five series. These five broad categories were established to
facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in
related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed
to develop and demonstrate instrumentation, equipment and
methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the
new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection
Agency, and approved for publication. Approval does not signify that
the contents necessarily reflect the views and policies of the Agency, nor
does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
This document is available to the public through the National
Technical Information Service, Springfield, Virginia 22161.
-------
EPA-600/2-77-002
January 1977
COMPARATIVE U.S. /USSR TESTS
OF A HOT-SIDE
ELECTROSTATIC PRECIPITATOR
by
Charles H. Gooding, Joseph D. McCain, and Diane K. Sommerer
Research Triangle Institute
P. O. Box 12194
Research Triangle Park, NC 27709
Contract No. 68-02-1398, Task 33
ROAPNo. 21ADL-029
Program Element No. 1AB012
EPA Project Officers: D. C. Drehmel and D. B. Harris
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
-------
ABSTRACT
In 1972 the United States of America and the Union of Soviet Socialist
Republics signed a bilateral agreement pledging cooperation on environmental
protection. As a part of this agreement, the Working Group on Stationary Source
Air Pollution Control was subsequently formed by the U.S. Environmental
Protection Agency and the U.S.S.R. Research Institute of Industrial and Sanitary
Gas Cleaning to conduct cooperative programs in several areas of air pollution
control technology, including particulate emission control.
This report describes the cooperative test program that was conducted to
quantify and characterize the particulate emissions from a U.S. coal-burning
power plant boiler, which is equipped with a hot-side electrostatic precipita-
tor. The tests were conducted at Duke Power Company's Allen Steam Station in
March 1976. U.S. and Soviet equipment and procedures were used to determine
the flue gas composition and velocity, total particulate mass concentration of
the gas stream, particle size distribution, electrical resistivity of the
particulate entering the precipitator, evidence of back corona in the precipi-
tator, S02 and S03 concentrations in the flue gas, and chemical composition of
the fuel and fly ash. The test site and test procedures are described. The
results of the comparative tests are presented and discussed.
The tests were conducted by York Research Corporation through the support
of EPA Contract No. 68-02-1401, Task 27, and by Southern Research Institute
under subcontract to York. This final report is submitted in fulfillment of
EPA Contract No. 68-02-1398, Task 33, by Research Triangle Institute. This
report covers a period from December 2, 1974, to July 31, 1976, and work was
completed as of August 31, 1976.
-------
iv
-------
CONTENTS
Abstract iii
Figures ..... vii
Tables viii
Abbreviations and Symbols . ix
Unit Conversion Factors x
Acknowledgments xi
1. Introduction 1
2. Description of the Test Site 3
General description of the power company and station . 3
Description of Allen Station Unit 3 3
Description of the unit 3 hot-side electrostatic
precipitator 5
Description of the test facilities 11
3. Test Proceedings 17
Summary of test procedures .... 17
Daily preliminary tests ..... 18
Determination of particulate mass concentration of
the flue gas 19
Determination of particle size distribution 30
Determination of fly ash resistivity 34
Determination of sulfur dioxide and sulfur trioxide
concentration of the flue gas 35
Evaluation of back corona in the precipitator .... 37
Determination of flue gas molecular weight 38
Coal sampling and analysis 40
Ash sampling and analysis 40
Plant data acquisition 41
4. Results and Discussion 43
Gas flow and total mass concentration determinations . 43
Determination of particle size distribution 50
Fly ash resistivity determination 58
Sulfur dioxide and sulfur trioxide determinations . . 58
-------
Results of back corona testing 61
Gas and coal analysis-theoretical gas volume 61
Fly ash analysis 63
Appendixes*
A. Orifice and meter calibration
B. U.S. pitot tube calibration
C. U.S.S.R. pitot tube calibration
D. U.S. data sheets-preliminary velocity determination and
mass determination
E. U.S. test data printout-mass determination
F. U.S. test summaries-mass determination
G. U.S.S.R. data sheets-velocity determination
H. U.S.S.R. test summaries-velocity determination
I. U.S.S.R. data sheets-mass determination
J. U.S. and U.S.S.R. data-particle size determination
K. U.S. data sheets-sulfur oxide determination
L. U.S. test data printout-sulfur oxide determination
M. U.S. test summaries-sulfur oxide determination
N. Boiler operating data
0. Test summaries-theoretical gas volume
P. Precipitator operating data
Q. Precipitator voltage/current data
R. Laboratory data sheets
*The appendixes to this report are contained in a separately bound
volume, copies of which may be obtained from the EPA Project Officers.
VI
-------
FIGURES
Number Page
1 Location of Duke Power Company service area and Allen Steam
Station 4
2 Schematic of Allen Unit 3 flue gas system 6
3 Configuration of hot-side electrostatic precipitator 9
4 Schematic of Allen Unit 3 hot-side electrostatic precipitator . 12
5 Elevation of inlet ducts viewed from underneath hoppers .... 13
6 Elevation of outlet ducts viewed from underneath hoppers ... 14
7 Details of sample port position in ducts 15
8 U.S. total particulate mass sampling train 20
9 Soviet particulate mass sampling train 25
10 Soviet mass sampling probe detail 26
11 Soviet pi tot tube 27
12 Brink cascade impactor . 31
13 Andersen cascade impactor 32
14 U.S. sulfuric acid mist sampling train 36
15 U.S. integrated gas sampling train 39
16 Inlet size distribution on a differential mass distribution
basis - Brink data 52
17 Inlet size distribution on a differential mass distribution
basis - Soviet Model I data 53
18 Inlet size distribution on a differential mass distribution
basis - Soviet Model II data 54
19 Outlet size distribution on a differential mass distribution
basis - Andersen data 55
20 Outlet size distribution on a differential mass distribution
basis - Soviet Model I data 56
21 Precipitator fractional efficiency from Brink and Andersen
data 60
vii
-------
TABLES
Number Page
1 Typical Gas Conditions at Inlet and Outlet of Hot-Side
Electrostatic Precipitator—Allen Unit 3 7
2 Typical Ash Characteristics from Duke Power Coal-Burning
Plants 8
3 Hot-Side Electrostatic Precipitator Specifications 10
4 Summary of Comparative Gas Flow Determinations 44
5 Summary of Comparative Total Mass Concentration Determinations 45
6 Results of Soviet Nozzle and Probe Washes 49
7 Comparative Results of Particle Sizing Devices 51
8 Average Particulate Mass Loadings by Sampling Device .... 59
9 Fly Ash Resistivity Results 61
10 Sulfur Oxides Test Results 62
11 Coal Analysis 62
12 Measured and Theoretical Gas Volumes 62
13 Fly Ash Chemical Analyses 63
viii
-------
ABBREVIATIONS AND SYMBOLS
The following is a list of abbreviations and symbols which are not
explicitly defined where they appear in the text of this report and which
may not be familiar to all readers.
ABBREVIATIONS
Nm3/s
ASTM
DNCMM
DSCFM
ACMM
ACFM
NCM
SCF
ACM
ACF
Ppm
V - I
SYMBOLS
Q
normal cubic meters per second (21° C and 760 mm Hg)
American Society for Testing Materials
dry normal cubic meters per minute
dry standard cubic feet per minute (70° F and
29.92 in. Hg)
actual cubic meters per minute
actual cubic feet per minute
normal cubic meters
standard cubic feet
actual cubic meters
actual cubic feet
parts per million
voltage/current
electrical resistance in ohms
IX
-------
UNIT CONVERSION FACTORS
To Convert From
inch (in.)
foot (ft)
mile
2 2
pound-force/inch (Ibf/in. or psi)
pound-mass (Ibm)
British thermal unit (Btu)
British thermal unit/hour
(Btu/hr)
foot-pound force (ft-lbf)
degrees Farenheit (°F)
Multiplication Factor
1,000,000 = 10C
1,000 « 10:
0.01 = 10
0.001 = 10
-2
-3
To
meter (m)
meter (m)
meter (m)
f
newt on/meter''
gram (g)
joule (J)
(N/nf)
watt (W)
joule (J)
degrees Celsius (°C)
Multiply By
2.540 x 10~2
3.048 x 10"1
1.609 x 103
6.895 x 103
4.536 x 102
1.055 x 103
2.931 x 10"1
1.356
°C = (°F-32) x
Prefix
mega
kilo
centi
mi Hi
Symbol
M
k
c
m
-------
ACKNOWLEDGMENTS
The successful completion of this project was made possible by the follow-
ing individuals, whose participation in various phases of the work is grate-
fully acknowledged.
Dr. Dennis C. Drehmel and Mr. D. Bruce Harris, EPA Project Officers,
competently supervised the coordination and overall progress of the project.
Mr. Jaroslav Pekar and Ms. Elizabeth Koniuskow, also of EPA, provided invaluable
assistance to the written and verbal communications between U.S. and Soviet
personnel.
Drs. Yurii S. Milovidov, Natalia G. Bulgakova, and Gelii S. Chekanov of
the Soviet Union provided technical direction and assistance in the use of the
Soviet test equipment at the Duke Power site.
The management and staff of Duke Power Company displayed outstanding
cooperation throughout the-planning and execution of the test program. Key
personnel were Messrs. Robert R. Carpenter, R. A. Johansen, and Bruce L. Jenkinson
of the Duke Power General Offices and Messrs. James R. Park and Al Saunders of
Allen Steam Station.
More than 30 engineers, scientists, technicians, and support personnel
from York Research Corporation, Southern Research Institute, and Research
Triangle Institute made significant contributions to various phases of the
cooperative program.
xi
-------
SECTION 1
INTRODUCTION
Over the past few decades the United States of America and the Union of
Soviet Socialist Republics have independently developed pollution control methods
to protect the environment from liquid, solid, and gaseous contaminants. In
1972, with recognition of the mutual benefits that could be gained from tech-
nology exchange, the United States and the U.S.S.R. signed a bilateral agree-
ment pledging cooperation on environmental protection. As a part of this
agreement, the Working Group on Stationary Source Air Pollution Control was
subsequently formed by the U.S. Environmental Protection Agency and the U.S.S.R.
Research Institute of Industrial and Sanitary Gas Cleaning.
In March 1973, the First Meeting of the U.S.-U.S.S.R. Working Group on
Stationary Source Air Pollution Control was held in Moscow to begin a coopera-
tive exchange of air pollution control technology and information between the
two countries. A second meeting was held in the United States in.April 1974.
The Protocol of the Second Meeting was signed by the heads of both delegations
on April 25, 1974, in Washington, D.C., establishing eight cooperative projects
for immediate action and 10 proposed areas of cooperation.
The planned cooperative programs encompass several areas of air pollution
control technology, including particulate emission control. High mass-collection
efficiencies are now achieved on particulate emissions from industrial processes
in both countries by utilizing electrostatic precipitators, baghouses (fabric
filters), wet scrubbers, and other innovative devices. Growing concern for the
health and environmental effects of fine particulate emissions (3 microns or
smaller) has resulted in a need for further improvement of conventional control
techniques and for the development of new techniques for fine particulate
control.
Projects A4/A6 of the Second Protocol established a cooperative test
program to quantify and to characterize physically and chemically the particu-
late emissions from selected industrial plants in the United States and in the
1
-------
U.S.S.R., using the sampling and analysis techniques commonly employed by both
countries. These projects were further developed during a joint meeting in
"foscow in October 1974. In December 1974 Duke Power Company's Allen Steam
Station (located near Charlotte, North Carolina) was surveyed, and Allen Unit 3
was selected as the site for the U.S. tests. Allen 3 is a coal-burning, steam-
electric generator whose particulate emissions are controlled by a hot-side
electrostatic precipitator.
A process description and a U.S. test program were developed for Allen 3
and were transmitted to the Soviet Union. In April 1975 the test program was
reviewed and updated to include the Soviet test plans during a joint meeting of
Soviet and American specialists in Moscow. The planning culminated in March
1976 when the following tests were actually conducted on Allen 3.
1. Measurement of flue gas velocity and pressure at the inlet
and outlet1 of the precipitator,
2. Determination of electrostatic precipitator collection effi-
ciency by simultaneous measurement of the inlet and outlet
particulate mass concentration,
3. Measurement of the gas humidity and molecular weight at the
inlet and outlet (using U.S. methods only),-
4. Measurement of the particle size distribution at the inlet
and outlet,
5. Measurement of electrical resistivity of the fly ash at the
i nl et,
6. Measurement of SCL and SO, concentrations at the inlet (using
U.S. method only),
7. Measurement of back corona (using Soviet method only),
8. Determination of the chemical composition of collected fuel
and fly ash samples.
This report describes the Allen power plant and the hot-side electrostatic
precipitator of unit 3. The procedures that were used during the test program
are recounted, and the results of the comparative tests are presented and
discussed.
-------
SECTION 2
DESCRIPTION OF THE TEST SITE
GENERAL DESCRIPTION OF THE POWER COMPANY AND STATION
Duke Power Company provides electric power for the central portion of
North Carolina and the northwestern corner of South Carolina. The approximate
boundaries of the Duke Power service area are shown in Figure 1. Duke Power's
total generating capacity is over 13,000 megawatts (MW), of which approximately
60 percent can be produced by coal-burning units.
Allen Steam Station is located approximately 16 km (10 miles)* southwest
of Charlotte, North Carolina (see Figure 1). Plant Allen has five coal-burning,
single-reheat, steam-electric generating units. Units 1 and 2 have nameplate
capacities of 165 MW each, and units 3, 4, and 5 are rated at 275 MW each.
2
Each-of the five units has a rated main steam pressure of 16.65 MN/m (2,415
lbf/in.2) gauge, a superheat temperature of 566° C (1,050° F) and a reheat
temperature of 538° C (1,000° F). The condensers of all five units are cooled
by once-through flow of water from the Catawba River. The cooling water is
discharged to the South Fork River, which joins the Catawba about 3 km (2
mi) downstream from the plant. When the tubes are clean, the condenser
steam-side pressure is approximately 38 mm (1.5 in.) of mercury absolute.
DESCRIPTION OF ALLEN STATION UNIT 3
Units 3, 4, and 5 at Allen Steam Station are identical, including the
precipitator installations. Unit 3 was chosen as the test unit after consid-
eration of maintenance outage schedules and test area access of the three units.
Commercial operation of Allen Unit 3 began in 1959. Although the unit
has a nameplate rating of 275 MW, it has frequently been operated at a gross
_..«,_, f
throughout the text of this report metric units are preferentially used, and
the commonly used English equivalent is shown in parentheses if appropriate.
For further clarification a list of abbreviations and symbols and a table of
conversion factors are also included at the beginning of this report.
-------
VIRGINIA
TENNESSEE
BOUNDARIES OF DUKE POWER
SERVICE AREA-
• RALEIGH
NORTH CAROLINA
. ALLEN
) STEAM STATION (
©CHARLOTTE
APPROXIMATE SCALE
I cm =37 km
(I in =58 miles)
COLUMBIA
SOUTH CAROLINA
UNITED
STATES
GEORGIA
Figure 1. Location of Duke Power.Company service area and Allen Steam Station.
-------
load of 300 MW or slightly greater. The gross load during the tests varied
from 276 to 279 MW. The unit auxiliaries utilize approximately 6.5 percent of
the generated power so that the net efficiency of Allen Unit 3 is approximately
35.9 percent (heat rate of 9,500 Btu/net kWhr). The expected thermal input to
the boiler is therefore 729 MW (2,487 x 106 Btu/hr) at 280 MW gross electrical
output. The typical coal analysis at the plant is given below.
Higher heating value 25.2 - 27.3 MJ/kg
(10,850 - 11,750 Btu/lbm)
Ash content 15-18 percent
Sulfur content approximately 1 percent
Moisture^content 6-9 percent
Depending on the heating value of the coal, the coal-firing rate at
280 MW ranges from approximately 96 to 104 Mg/hr (212,000 to 229,000 Ibm/hr).
About 25 percent of the ash falls out in the dry-bottom boiler as bottom ash.
The remaining 75 percent of the ash leaves the boiler with the hot flue gases.
As shown schematically in Figure 2, the gases flow first through the hot-side
electrostatic precipitator at about 343° C (650° F), then through the air
preheater where the gas temperature is lowered to approximately 138° C (280° F)
by preheating the incoming combustion air. The flue gases then flow through
the cold-side electrostatic precipitator before the pressure is boosted by the
induced draft fan and the gases exit to the atmosphere through the 77 m (252
ft) stack. The typical gas conditions at the inlet and outlet of the hot-side
electrostatic precipitator as estimated prior to these tests are given in Table
1. Table 2 lists typical ash characteristics from Duke Power coal-burning plants.
DESCRIPTION OF THE UNIT 3 HOT-SIDE ELECTROSTATIC PRECIPITATOR
Allen Unit 3 was equipped at startup with a cold-side electrostatic
precipitator designed to remove approximately 97 percent of the fly ash from
the flue gases. The precipitator efficiency was tested twice in 1961 at
96.9 and 96.4 percent. In the late 1960's, Duke Power began upgrading its
particulate emission controls by adding additional precipitators in a series
or parallel configuration on many of its units. At that time, the estimated
precipitator efficiency of Allen Unit 3 was 87 percent. A hot-side electro-
static precipitator was designed and installed on unit 3 in series with the
existing cold-side precipitator (see Figure 2). Startup of the new precipitator
-------
FEEDER
PULVERIZER
— — —• Indicates gas flow
STACK
FD FAN
Figure 2. Schematic of Allen Unit 3 flue gas system.
-------
TABLE 1. TYPICAL GAS CONDITIONS AT INLET AND OUTLET OF
HOT-SIDE ELECTROSTATIC PRECIPITATOR—ALLEN UNIT 3
Flow^rate (total for four ducts)
Nnr/s (actual ft3/min)
Temperature, °C
(°F)
Pressure, mm Hg gauge
(in H20 gauge)
Ash concentration, g/Nm
(grains/actual ft*)
Moisture (H20), % by volume
Oxygen (02), °/» by volume
Carbon dioxide (C02), % by volume
Sulfur dioxide (S02), % by volume
Sulfur tri oxide (S03), % by volume
Nitrogen oxides (NOV), % by volume
A
Inlet
226
(1,100,000)
343
(650)
-11 to -13
(-6 to -7)
13.1 to 17.3
(2.5 to 3.3)
6
3.5
15.6
0.09
Unknown*
Unknownt
Outlet
226
(1,000,000)
338
(640)
-11 to -13
(-6 to -7)
0.11 to 0.14
(0.020 to 0.027)
6
3.5
15.6
0.09
Unknown*
Unknownt
*Typical values for this type of boiler and fuel are 10 to 50 ppm S03
(by volume).
tTypical values for this type of boiler and fuel are 300 to 400 ppm NOX
as N02 (by volume).
-------
TABLE 2. TYPICAL ASH CHARACTERISTICS FROM DUKE POWER
COAL-BURNING PLANTS
Ash fusibility
Initial deformation temperature
Softening temperature
Fluid temperature
1,315 - 1,480° C
1,370 - 1,480° C
>1,480° C
(2,400 - 2,700° F)
(2,500 - 2,700°" F)
(>2,700° F)
Ash composition
Si02
A12°3
Fe203
CaO
MgO
Na20
i\oU
2
TiO,
-------
CHAMBER A-2 CHAMBER A-l CHAMBER B-l CHAMBER B-2
-
-------
TABLE 3. HOT-SIDE ELECTROSTATIC PRECIPITATOR SPECIFICATIONS
Manufacturer
Startup date
Design gas flow
Design gas velocity
Design specific
collector area
Design efficiency
Overall configuration
Plates
Wi res
Electrical
Research Cottrell, Inc., Boundbrook, N.J.
March 5, 1973 .
590 actual m3/s (1,250,000 actual fr/min)
1.81 ra/s (5.94 ft/s)
53 m2 per actual m3/s (270 ft2 per 1000 actual ft3/min)
99.2%
4 parallel chambers
4 sections in series per chamber
39 parallel gas passages per chamber
40 plates per chamber (cold rolled steel sheets)
plate height is 9.14 m (30 ft)
plate length each section is 2.74 m (9 ft) for total
length in direction of flow of 10.97 m (36 ft)
plate-to-plate spacing is 0.229 m (9 in)« ?
total surface area of plates is 31,305 m (336,960 ft )
48 equally spaced wires per gas passage (hand drawn
Bessemer steel with coppered surface)
wire diameter is 2.77 mm (0.109 in)
wires are hanging type, placed in the center ± 6.35 mm
(1/4 in) of the plate-to-plate space
8 transformer-rectifier sets
16 electrically isolatable bus sections
transformer rating is 96 kVA
rectifier rating is 1500 mA
wave form is double/half full
normal power consumption is approximately 580 kW,
720 kW is maximum consumption
Ash deposits are removed from the wires by vibrators, which have an
adjustable cycle of operation. Each vibrator is normally operated twice every
half-hour with approximately a 90-second delay between the two vibration
periods. Each vibration period lasts 6 seconds. The plates are cleaned by
solenoid-activated hammer-type rappers. Each rapper is activated at least once
every 2 minutes, and some are activitated twice every 2 minutes. The approximate
rapping intensity is 32.5 J (24 ft-lbf). The collected ash falls into hoppers
beneath the precipitator. It is periodically removed from the hoppers by a
dry, pressurized ash-handling system and flows to a collecting tank from which
it is water-sluiced to an ash-settling basin.
10
-------
DESCRIPTION OF THE TEST FACILITIES
During the tests, operational data from the steam-electric generating unit
and the electrostatic precipitator were monitored from inside the plant. In
the control room from which units 1, 2, and 3 are operated, charts continuously
recorded operational parameters such as electrical load, fuel flow, air flow,
steam flow, and flue gas temperatures and pressures. The oxygen concentration
of the gas was also continuously recorded in the control room and periodically
was manually checked with a portable recorder at several duct sample lines.
The precipitator control panels are also in the boiler building. There
are eight control panels for the unit 3 hot-side precipitator (one for each
transformer-rectifier set). Instruments on each panel continuously display
the transformer primary voltage (a.c.), the transformer primary current (a.c.),
the precipitator average current (d.c.)s and the precipitator spark rate.
These instruments were utilized in the back corona tests.
Coal samples were manually collected during the test from the hoppers
located above the coal pulverizer feeders (see Figure 2). Ash samples
were collected downstream of the economizer section of unit 3 as indicated
in Figure 2.
The identical inlet and outlet ducts of the Allen 3 electrostatic pre-
cipitators are separated by the precipitator and fly ash hoppers (Figure 4).
The sampling ports are located on the hopper side of the ducts. Only two
inlet and two outlet ducts were sampled, as indicated in Figures 5 and 6.
The eight sampling ports in each duct have an inside diameter of approximately
154 mm (6.06 in.). The ports are equally spaced and are 0.84 m (2 ft 9 in.)
apart. The outside ports of each duct are 0.41 m (1 ft 4.5 in.) from the
duct wall (see Figure 7). The ports are 0.457 m (18 in.) above a handrail
which is 1.07 m (42 in.) above the platform. The horizontal distance from
the ports to the fly ash hoppers is approximately 3 m (10 ft). A beam near
the inlet duct designated during these tests as Bl prevented some tests from
being conducted in one port.
Two sets of sampling ports were installed for the resistivity
tests. The ports were located on a horizontal segment of the hot gas duct
downstream from the economizer and about 3 m (10 ft) upstream of the 90°
turn which leads to the inlet test ports.
11
-------
EQUALLY SPACED
TURNING VANES
•"6m
(-20ft)
1.83m
(6ft)
Figure 4. Schematic of Allen Unit 3 hot-side electrostatic precipitator.
-------
GAS IS PHYSICALLY SEPARATED HERE
co
TEST PORTS-
A
INLET DUCT B2
GAS FLOW
f
87654321
A
INLET DUCT Bl
GAS FUDW
A
8765432 I
INLET DUCTS SAMPLED
A
NOT TO SCALE
Figure 5. Elevation of inlet ducts viewed from underneath hoppers,
-------
GAS IS PHYSICALLY SEPARATED HERE
\
/
\
fcl^vr-
/
r
\
'N r*
•
OUTLET DUCT Bl
i
GAS FLOW
12345678
V
/Mil IT
/
~-~.
\
"-*^^.
OUTLET DUCT B2
GAS FLOW
1 2345678
J^>OUT
/
^
^fc
LEI
mi i
TEST PORTS
Figure 6. Elevation of outlet ducts viewed from underneath hoppers.
-------
10.22m
(33'6.5")
0.41m
0'4.5")
\
APPROXIMATE
SCALE 100:1
0.84 m
HANDRAIL-
PLATFORM--*
559m
,0.41m
i
0.457m (18")
1.07m (42")
6.71m
(22'}
Figure 7. Details of sample port position in ducts.
15
-------
16
-------
SECTION 3
TEST PROCEEDINGS
SUMMARY OF TEST PROCEDURES
The test program at Plant Allen involved the measurement of several
parameters using U.S. and Soviet equipment and procedures. Replicate runs
were made over a period of 8 days from March 12, through March 19, 1976,
inclusive. Before each day's tests began, ash was removed from the unit 3
precipitator hoppers and boiler soot blowing was conducted. During the actual
sampling both of these operations were suspended. Boiler and precipitator
operating parameters were monitored at half-hour intervals during the sampling
periods. Daily coal and ash samples were collected for analysis.
The overall scope of the tests is summarized below, and the individual
procedures are described in more detail in the remaining paragraphs of this
section.
1. The flue gas velocity and static pressure were measured at
the inlet and outlet using calibrated pi tot tubes supplied
by both countries. Preliminary moisture and molecular weight
determinations were made concurrent with the pitot traverse
with U.S. equipment.
2. To determine the precipitator collection efficiency, mass
sampling was conducted at the inlet and outlet using both
U.S. and Soviet equipment. The standard EPA Method 5 was
used at the inlet, and a hi-volurne EPA Method 5 was used
at the outlet. The Soviet method utilized zero-type tubes
with stainless steel filters.
3. Gas humidity was measured at the inlet with the U.S. equip-
ment concurrent with the mass sampling. Flue gas molecular
17
-------
weight was determined from samples extracted with a separate
Orsat probe attached to the mass sampling probe.
4. Particle size distributions were determined on the inlet
and outlet. For the U.S. tests Brink impactors were used
at the inlet and Andersen impactors were used at the outlet.
A Soviet cascade impactor was also employed for size sampling
at the outlet, and at the inlet two types of Soviet impactors
and a Soviet series cyclone apparatus were used. Outlet samples
for both countries were obtained by complete traverses of the
two outlet ducts using 24 sampling points per duct. Because
of a combination of short sampling times and poor inlet velocity
distributions, the inlet samples were obtained from individual
ports, extracting one sample from each of four ports in each
inlet duct.
5. Electrical resistivity of the fly ash particles was measured at
the inlet by a U.S. method only, using a point-to-piane
resistivity probe. Attempts to obtain resistivity data with
the Soviet equipment were thwarted by various equipment and
weather difficulties.
6. Sulfur dioxide and sulfur trioxide concentrations of the inlet
gas were determined by the U.S. only (EPA Method 8) since the
Soviet method is identical.
7. Back corona was measured using the Soviet method, which is
based on measurement of the precipitator voltage-current
relationships during voltage increase and decrease.
8. Fuel analyses were performed to determine the composition of
ash, sulfur, hydrogen, carbon, moisture, nitrogen, and
oxygen. Heating value was also determined. The collected
ash samples were subjected to quantitative analysis to
determine their chemical composition.
DAILY PRELIMINARY TESTS
Prior to each day's comparative testing, preliminary tests were conducted
with U.S. equipment. Flue gas velocity measurements were made with U.S. pitot
18
-------
tubes before the first few tests. These measurements provided a starting point
for the operating procedure followed in the U.S. mass concentration sampling.
Later tests relied on the previous day's results as a starting point.
Preliminary determinations of flue gas moisture and molecular weight were
conducted each day to provide information for the conduct of the comparative testing,
Data from the preliminary tests are not discussed in Section 4 but are in-
cluded for reference in Appendix D.
DETERMINATION OF PARTICULATE MASS CONCENTRATION OF THE FLUE GAS
Particulate mass concentration was determined by the standard U.S. method
(EPA Method 5) and by the Soviet method.
U.S. EPA Method 5
Flue gas was extracted isokinetically from the duct through a heated probe
and a fiberglass filter. This filter was enclosed in an oven and kept at a
constant temperature of 160° C (320° F). Gas velocity and gas temperature were
monitored continuously at each test point. Sample flow rate was computed
utilizing a preset nomograph, which was adjusted whenever flue gas conditions
of velocity and/or temperature changed with respect to time. A measurement of
moisture content was performed coincident with particulate determination.
The particulate sampling apparatus consisted of a probe, pitot tube,
filter holder, four Greenburg-Smith type impingers, vacuum pump, dry gas meter,
and flow meter in a configuration indicated in Figure 8. The 316 stainless
steel buttonhook-type probe tip (1) was equipped with a 16-mm (5/8-in.)
diameter fitting connected to the probe by a teflon packed stainless steel
coupling. The probe itself (4) was a 16-mm (5/8-in.) diameter medium wall
pyrex glass tube fitted with a ground glass ball joint on one end. It was
wrapped with heater tape capable of maintaining a minimum sample-gas stream
temperature of 160° C (320° F) during sampling to prevent condensation from
occurring within the probe. There was a reverse type pitot tube attached to
the probe in order to provide instantaneous differential pressure readings at
each sampling point. In addition, situated next to the nozzle was a type-K
chrome!-alumel thermocouple (3) connected to a pyrometer (14) for direct
measurement of flue gas temperature.
The probe assembly was sealed to prevent ambient air from leaking into
the duct and diluting the sample.
19
-------
SAMPLE BOX COMPONENTS
METER BOX COMPONENTS
Figure 8. U.S. total participate mass sampling train.
20
-------
The probe was connected to a glass elbow which, in turn, was connected to
a very coarse fritted glass filter holder (6). This holder contained a glass
filter which had been previously numbered and weighed. The filter holder was
contained in an electrically heated enclosed box, which was thermostatically
maintained at a temperature of 160° C (320° F) to prevent moisture condensation.
Attached to this heated box were a series of four impingers connected with
glass ball joints. The first impinger (9) was of the Greenburg-Smith design,
modified by replacing the tip with a 12.7 mm (1/2 in.) inside diameter glass
tube extending to within 12.7 mm (1/2 in.) from the bottom of the flask. This
impinger was initially filled with 100 ml of water. The second impinger (9)
was a Greenburg-Smith with a standard tip and was similarly filled with 100 ml
of water. The third impinger (10) was modified like the first, but without
water. The fourth impinger (11) was also a Greenburg-Smith type modified like
the first and contained 300 grams of dry, indicating, 6-16 mesh silica gel, which
had been previously dried at 175° C (347° F) for 2 hours. Both the impingers
and heated box were housed together to facilitate removal of all sample compo-
nents to a "clean" area. After the fourth impinger (11), the sample gas stream
flowed past a dial thermometer (17), through a check valve to flexible rubber
vacuum tubing (13), vacuum gauge (16), a valve (18), a leakless vacuum pump
(20), which was connected in parallel with a bypass needle valve (19), and a
dry gas meter rated at 0.1 ft per revolution (21). A calibrated orifice
(22) completed the train and was used to measure instantaneous meter flow
rates. The three thermometers (17) were dial type with a range of -4° to
52° C (25° to 125° F). A fourth thermometer in the heated portion of the box
had a range up to 260° C (500° F). A dual manometer (23) measured pressure
drop across the calibrated orifice. This manometer was an inclined-vertical
type graduated in hundredths of an inch of water from 0 to 1.0 inch.
Procedure-
Two separate ducts (Bl and B2) were sampled during each test period. Each
duct was 1.83 m x 6.705 m (6.0 ft x 22.0 ft) and was divided into twenty-four
equal areas with dimensions of 0.61 m x 0.84 m (2.0 ft x 2.75 ft). Test 1
through Test 6 included samples from both ducts Bl and B2. Test 7 included
sample only from duct 82 and Test 8 included sample only from duct Bl. Each
test was performed at the inlet and outlet ducts simultaneously.
21
-------
The center of each equal area was sampled for 5 minutes. Total test
time was 120 minutes per duct. Port number 4 on inlet duct Bl was inaccessible
due to structural steel supports, which obstructed the port opening. Accordingly,
this port was not sampled during any tests.
A sampling system leak rate not-to-exceed 0.02 ft per minute at
a vacuum of 15 inches of mercury was required. This was checked by plugging
the probe nozzle while the sampling train was in operation. In order to prevent
water back-up in the train, the nozzle was slowly unplugged prior to shutting
off the pump. Crushed ice was then placed around the impingers in an insulated
box to insure that gases exiting that section were less than 21° C (70° F).
For each test run the following data were recorded every 5 minutes or
when changes occurred in the flue gas conditions:
1. Point designation.
2. Clock time.
3. Dry gas meter reading (cubic feet).
4. Velocity head (AP in inches of water).
5. Desired pressure drop across orifice (AH in inches of water).
6. Actual pressure drop across orifice (AH in inches of water).
7. Dry gas temperature (°F) gas meter inlet.
8. Dry gas temperature (°F) gas meter outlet.
9. Vacuum pump gauge reading (in. Hg).
10. Filter box temperature (°F).
11. Dry gas temperature (°F) at the discharge of last impinger.
12. Stack temperature (°F).
13. Stack pressure (in inches of water).
At the start of testing, the probe nozzle was positioned directly into the
gas stream and sampling immediately started. When inserting the probe into
highly negative pressure, care must be taken. The pump was running to prevent
water backup. Isokinetic sampling was maintained by the use of a nomograph
which incorporated the relationship of the difference of pitot tube differential
pressure (AP) and the pressure drop across the orifice meter (AH).
The relationship of AP reading with the AH reading is a function of the
following variables:
1. Orifice calibration factor.
2. Gas meter temperature.
22
-------
3. Percent moisture in the flue gas.
4. Ratio of flue gas pressure to barometric pressure.
5. Stack temperature.
6. Sampling nozzle diameter.
7. Pitot tube correction factor (e.g., other than 0.85).
The use of the nomograph allowed for the direct relationship to be
determined within approximately 15 seconds, thus allowing isokinetic conditions
to be maintained throughout the test.
Sample Recovery and Analysis-
Careful handling of the sampling apparatus was necessary in moving from
the sampling location to the cleanup site. The water collected was first
measured volumetrically and then discarded. Samples were placed in designated
containers and analyzed in the following manner in the laboratory at the
power plant:
Container 1: Each filter was sealed in a covered petri dish and placed
in a plastic zippered bag. In the lab the filters were dessicated
to a constant weight. The previously recorded tare weight was
subtracted and the result was reported to the nearest 0.01 mg.
Container 2: All loose particulate was acetone-washed from all surfaces
upstream of each filter, and the wash was sealed in a glass jar with
a Teflon lid liner. The acetone washings and particulate matter were
later transferred to a tared beaker and allowed to dry at room
temperature and pressure. The constant weight was recorded to the
nearest 0.01 mg.
i
Container 3: Silica gel from the fourth impinger of each test was sealed
in a plastic jar. The silica gel was later weighed to the nearest
gram. The tare weight recorded prior to the tests was subtracted and
the difference was converted to volume of water and added to the
volume collected in the other impingers.
U.S.S.R. Balanced Tube Method
Each of the ducts was divided into equal areas. A velocity traverse was
performed before and after each test using a Soviet pitot tube. The velocity
pressure and gas temperature at the center of each equal area were recorded.
23
-------
To obtain the total mass samples a filtration medium was enclosed in a
"zero" balanced tube apparatus, and the entire assembly was inserted into the
flue gas duct. The ducts were traversed by placing the nozzle in the center
of each area sequentially and adjusting the sampling rate to make it iso-
kinetic. The gas at each point was sampled for a specific period of time.
At the end of the test the parti oil ate mass collected on the filter medium
was representative of the average particulate concentration at the areas
sampled in the duct. The inlet duct and outlet ducts were sampled using
identical methods and equipment.
The equipment used in determination of particulate mass concentration is
shown in Figure 9. The balanced tube assembly with the internal filter is
shown in detail in Figure 10. A partial vacuum was induced in the sample train
by a sliding-vane type air pump. Sample flow was actuated by a ball valve
located in the sample train between the condenser and the vacuum pump. A
needle valve bypass across the pump was used as a fine control for sample
flow rate.
Sample gas entered the nozzle under induced vacuum and was drawn through
a thimble filter fabricated from sintered stainless steel. Located at a
point between the nozzle opening and the flow distributor inside the thimble
filter were static pressure taps. The nozzle at this point had an inside
diameter of 9 mm (5/16 in.). One set of pressure taps was located inside the
nozzle while a second set was exposed to the flue gas static pressure. The
pressure taps were extended via one-piece stainless steel tubes to the back
end of the probe where they were connected with leak-proof seals to two inde-
pendent lengths of plastic tubing. The plastic tubes were then connected
across an inclined oil manometer. During sampling the flow rate was adjusted
so that the static pressures inside and outside of the nozzle were equal.
The assumption was made that when the static pressures were equal, isokinetic
sampling was in effect.
The probe used in the performance of the pi tot traverse is shown in
Figure 11. Impact pressure was transmitted through a single tube facing the
gas stream. Static pressure was transmitted through the annular space around
the single tube. Two pressure tubes were located at the back end of the probe
for connection across an inclined manometer by flexible hoses.
24
-------
ro
en
to
c
-s
n>
to
o
CD
rt-
"0
O>
O
ET
Qi
rh
fD
O)
-.1.
IO
O>
ra
Imp angers in ice bath
Orifice
Manometer
11 n lanced
Tube
Mnnometei
Holfinced
Tube
Assembly
Dry Gas Meter
-------
Flue Gas Static/
Pressure Tap //
Nozzle
Sample Gas Static
Pressure Tap
Static Pressure
Lines
Flow Distributor
Sintered Stainless
Steel Thimble
Figure 10. Soviet mass sampling probe detail
26
-------
I
M
II
i!
Ij
ii
H
it
n
n
n
i
i
i
Z«
20
OK
CO
n
1 1
tr
Figure 11. Soviet pi tot tube.
27
-------
Procedure--
Each test was comprised of two separate ducts, Bl and B2. Each duct was
divided into 24 equal areas with dimensions of 0.61 m x 0.84 m (2.0 ft x
2.75 ft). The overall inside dimensions of the ducts were 1.83 m x 6.705 m
(6.0 ft x 22.0 ft). The ducts Bl and B2 were both sampled during each test,
with the exception of Test 7 and Test 8. During the last two tests only
one duct was tested, with duct B2 sampled during Test 7 and duct Bl sampled
during Test 8. Each test was performed with inlet and outlet samples taken
simultaneously.
The center of each equal area was sampled for five minutes; total test
time was 120 minutes per duct. Port number 5 on inlet duct Bl was obstructed
by a structural steel I-beam and, therefore, was not sampled during any of the
tests.
During the first four tests, the sample flow rate was controlled so that
the static pressure inside the nozzle was equal to the static pressure outside
the nozzle. The sample rate was increased during Tests 5 through 8 to 0.1 in.
of water above zero on the null manometer.
When moving the probe to the adjacent port, a shutoff valve in the sample
line prevented the vacuum in the flue gas duct from creating a backflow in the
train and removing particulate from the filter. However, the direction of the
gas flow in the inlet duct was from floor to ceiling, which necessitated that
the nozzle be pointed down into the gas flow. In the first few tests this
condition is believed to have caused the loss of some particulate through the
nozzle when the shutoff valve was actuated. Tests 6 through 8 were performed
with an extension on the flow distributor within the balanced tube apparatus.
This apparently prevented loss of particulate sample as demonstrated by the
higher results obtained in these three tests.
Prior to Test 6 a leak was detected in one of the static pressure tubes
on the outlet train. This train was replaced at that time and a leak-test was
performed. The remainder of the tests were performed with the proven sampling
apparatus.
An attempt was made during Test 5 to filter the sample gas at the back
end of the probe with a 47-mm fiberglass filter. Condensation of water
caused the filter to plug immediately.
28
-------
Sample Recovery and Analysis—
At the end of each test the balanced tube apparatus was placed with the
nozzle pointing upward while the apparatus cooled. The apparatus was gently
tapped to dislodge particulate from the inside surfaces of the nozzle. The
thimble filter was carefully removed and placed in a clean glass jar with the
open end of the thimble facing upward.
At the end of Tests 5 through 8, the inside nozzle surfaces were washed
with acetone and the rinse was saved in a glass jar, sealed with a Teflon-
lined lid.
At the end of Tests 4 through 6, the inside surfaces of the probe were
washed with acetone and the rinse was saved in a glass jar, sealed with a
Teflon-lined lid.
The thimble filters were placed in tared beakers and desiccated overnight.
The beakers were then weighed and desiccated repeatedly until the weight was
constant. The samples were weighed on an analytical balance with a sensitivity
of 0.01 mg.
The acetone rinses were placed in tared beakers and allowed to evaporate
at ambient temperature and pressure. The beakers were then desiccated and
weighed until weight was constant on the same analytical balance used for
thimble filters.
All analyses were performed at the laboratory of the Duke Power Company's
Allen Plant.
Particulate Mass Calculations
Results of both the U.S. and U.S.S.R. particulate mass tests were calcu-
lated on a ratio of areas method. Because isokinetic sampling was used, it
was assumed that the mass of particulate entering the nozzle at each sampling
point was representative of the average particulate mass in the flue gas duct.
The net particulate mass collected in each test was multiplied by the ratio
of the total area of the flue gas duct to the area of the nozzle. The result
of this calculation is the total mass of particulate passing through the duct
during the test period. Division by the test period in minutes is necessary
to compute mass rate. The equation is as follows:
c - m v Ad v 1 y ] v 60
E~mxArTxTx 1000 g/kg x hr
29
-------
where:
E = mass rate (kg/hr),
m = mass filter catch (g),
2
Ad = Total area of duct sampled (m ),
2
An = Area of nozzle (m )5
T = Test time (min).
•
The participate mass concentration was then determined by dividing the mass
flow rate by the volumetric flow rate of gas.
DETERMINATION OF PARTICLE SIZE DISTRIBUTION
Particle size distribution was determined using two U.S. and three Soviet
instruments. All of these instruments operated on the similar principle of
inertia! size classification, but the design and operation of the individual
instruments varied.
U.S. Particulate Size Determinations
The U.S. particle sizing devices used during these tests were modified
Brink-type cascade impactors (Fig. 12) for precipitator inlet measurements and
Andersen cascade impactors (Fig. 13) for outlet measurements. The Brink
impactors included a cyclone precutter as well as "0" and "6" stages. Both
instruments were used with glass fiber mat impaction substrates and final
filters. All collection media used during the tests (impaction substrates and
final filters) were preconditioned at the test site to minimize potential
interferences resulting from uptake of vapor phase components upon exposure to
the flue gases. All sampling was done through the ports which were described
previously.
Outlet samples were obtained by complete traverses of the two outlet
ducts with each impactor using 24 sampling points per duct. The gas velocity
profile in each duct was quite uniform with most point velocities in each
duct differing by only a few percent from the combined average velocity for
both ducts. Nozzles and sampling flow rates were chosen for isokinetic sam-
pling rates to be obtained at the overall average velocity for the combined
ducts. "Buttonhook" nozzles were used and the sample rates and gas volumes
were determined with orifice meters and a dry gas meter. A total of nine
Andersen impactor outlet runs were obtained during the test series, excluding
30
-------
NOZZLE
JL
t
1
A
<
1
i
t.
~
L.
1
1
J.
.1
r4
j.
B
PRECOLLECTION
CYCLONE
JET STAGE
(7 TOTAL)
COLLECTION
PLATE
SPRING
B
Figure 12. Brink cascade impactor.
31
-------
•JET STAGE (9 TOTAL)
SPACERS
GLASS FIBER
COLLECTION
SUBSTRATE
BACKUP
FILTER
PLATE
HOLDER
NOZZLE
INLET
CORE
Figure 13. Andersen cascade impactor.
32
-------
blanks. In addition, eight Andersen blank (or control) runs were made to
determine the extent to which interferences from gas phase components might be
affecting the outlet data. Outlet sampling times ranged from 120 to 360 minutes.
A total of nine Brink inlet samples were obtained. In addition, three
blank or control runs were made with Brink impactors. Because of a combination
of short sampling times (0.5 to 20 minutes) and poor inlet velocity distribution,
each inlet sample was obtained from a single port. Four of the eight ports
in each inlet duct were sampled. Sampling was accomplished using a combination
of nozzle diameter and sampling rate to provide isokinetic conditions for the
particular location at which the run was made. Straight sampling nozzles were
used, and the sampling rates and gas volumes were determined with orifice meters.
U.S.S.R. Particulate Size Determinations
The Soviet particle-sizing devices included two cascade impactors: Model I,
an impactor using multiple round jets at each stage with substrates of glass
fiber mats or similar materials; Model II, a hybrid cylindrical slot/round hole
impactor using the metal walls of the impactor as substrates; and a series
cyclone system. All three devices were designed to use a backup filter plug
of a glass wool or "Fiber-Frax" type material. These plugs were replaced
during some tests by conventional flat glass fiber filter media. All three
Soviet devices were used to obtain precipitator inlet size distributions but
only Model I was used at the outlet. All filter media used during the tests
were preconditioned at the test site to minimize potential interferences
resulting from uptake of vapor phase components upon exposure to the flue gases.
Operation of the Soviet equipment was done in accordance with instructions
from the Soviet test delegation. The operation of the Soviet equipment departed
from their standard practices only in the case of Impactor Model II. In the
case of the Model II their normal practice is to determine tare weights of the
individual stage assemblies before sampling and then to reweigh the assembly
with the dust attached or contained within it after sampling—the weight
difference being that of the collected dust. The balances available during
these tests were not suitable for this procedure; consequently, the collected
particulate from each component of the Model II impactor was removed as com-
pletely as possible and transferred to low tare weight aluminum foils in
order to determine the weight of the particulate matter caught by each stage.
33
-------
Some small portion of the collected particulate was inevitably lost during this
transfer; however, the losses are generally believed to be small compared to
the total catch.
A total of seven Soviet Model I impactor outlet runs were obtained during
the test series, excluding blanks. As with the Andersen complete traverses
were made in both outlet ducts using sample rates which were calculated to be
isokinetic for the overall average velocity of the combined ducts. Soviet
orifice meters were used to determine the sample rates and gas volumes. Sample
times ranged from 120 to 360 minutes. Straight sample nozzles were used. In
addition to the sample runs, two blank or control runs were made to determine
the extent to which interference from gas phase components might be affecting
the outlet data.
A total of five Model I's, eight Model II-'s, and eight series cyclone sam-
ples were obtained at the inlet. In the same procedure used with the Brink im-
pactor, each sample was drawn from a single port using what was thought to be
an isokinetic sample rate. Soviet oriftce meters were used to determine the
sampling rates and gas volumes. Straight sample nozzles were used.
Calibrations, of the Soviet orifice .meters with a dry gas meter at the
test site subsequent to the actual testing.indicated that the Soviet orifice
meter used at the inlet indicated 9 percent less than the true flow and the
outlet orifice meter, as used, indicated 19 percent less than the true flow.
The flow rates and all concentrations and size cuts reported herein have been
adjusted for this discrepancy.
DETERMINATION OF FLY ASH RESISTIVITY
Fly ash resistivity was measured with a U.S. in-situ point-to-piane-type
resistivity probe. Separate ports, located in the horizontal segment of the
inlet ducts about 3 m (10 ft) from the other inlet ports, were used for the
resistivity testing. These special ports were installed to accommodate the
Soviet resistivity measurement device which is similar to what is described
in the U.S. as a Lurgi-type probe. However, attempts to obtain resistivity
data with the Soviet equipment were thwarted by various equipment and weather
difficulties.
To use the U.S. point-to-plane resistivity probe, which was designed by
Southern Research Institute, the probe was lowered into the duct and a thin
34
-------
layer of dust was deposited on a small plate by electrostatic precipitation.
The resistivity of the dust was then determined by measuring the voltage-
current relationship across the thin layer. A total of eight resistivity
determinations were made on four different days at the Allen tests.
DETERMINATION OF SULFUR DIOXIDE AND SULFUR TRIOXIDE CONCENTRATION OF THE FLUE
GAS
The U.S. EPA Method 8 and the Soviet method of determining the sulfur
dioxide and sulfur trioxide of flue gases are identical. Both methods are
based on the absorption of S02 and S03 from an extracted flue gas sample,
followed by separate measurement using the barium chloride/thorin titration
method. Since the methods are identical only the U.S. train was utilized. A
total of six S0«/S07 determinations were made.
C. 3
Equipment
The S02/S03 sampling apparatus consisted of a probe, pitot tube, filter,
four Greenburg-Smith impingers, dry gas meter, vacuum pump, and flow meter as
shown in Figure 14. The probe assembly was of the same configuration as the
U.S. total particulate mass probe described previously.
The probe was connected to a standard Greenburg-Smith impinger (24) which
was filled with 100 ml of 80% isopropanol (CH3CHOHCH3). After the first
impinger was a very coarse fritted pyrex filter holder (6) which held a tared
glass fiber filter. This was connected to a second Greenburg-Smith impinger
(25), which was modified by replacing the standard tip with a 12.7 run (1/2
inch) ID glass tube extending to 12.7 mm (1/2 inch) from the bottom of the
impinger flask. The second impinger was filled with 100 milliliters of a 3%
hydrogen peroxide (H202) solution.
The third impinger was also a standard Greenburg-Smith like the first.
However, it was filled with 100 ml of a 3% hydrogen peroxide (H202) solution.
The fourth and last impinger was modified like the second and was filled with
300 g of dry, indicating, 6-16 mesh silica gel, which had been previously dried
at 175° C (347° F) for 2 hours.
Sampling Procedure
The sampling procedure was identical to that described in the U.S. total
particulate mass concentration procedure. Two separate inlet ducts were
35
-------
1MPINGER TRAIN
METER BOX COMPONENTS
Figure 14. U.S. sulfuric acid mist sampling train.
.16
-------
sampled during each test period. Each duct was 1.83 m x 6.705 m (6.0 ft x
22.0 ft) and was divided into 24 equal areas with dimensions of 0.61 m x
0.84 m (2.0 ft x 2.75 ft). The center of each area was sampled for five
minutes; total test time was 120 minutes per duct. Port number 4 on inlet
duct Bl was inaccessible due to structural steel supports which obstructed the
port opening. This port was not sampled during any of the tests.
Sample Recovery
The 80% isopropanol absorbing solution from the first impinger was trans-
ferred to a 250 ml graduated beaker. Everything upstream of the filter was
rinsed with an 80% isopropanol solution and added to the 250 ml beaker with the
impinger rinse. The beaker contents were then diluted to 250 ml with 80%
isopropanol. The filter was added to the solution, which was mixed and stored.
The solutions from the second and third impingers were transferred to a
500-ml graduate beaker. All glassware downstream of the filter and upstream
of the silica gel was rinsed with deionized, distilled water which was added
to the 500-ml beaker. This solution was then diluted to 500 ml with deionized
distilled water and stored.
Analysis
The container holding isopropanol and the filter was shaken. If the filter
broke into pieces, the fragments were allowed to settle for a few minutes
before the sample was removed. A 100 ml aliquot of sample was pipetted into a
250 ml Erlenmeyer flask and 2 to 4 drops of thorin indicator were added. The
sample was titrated with barium perchlorate to a pink end point. This titration
procedure was repeated with the samples from the second and third impingers.
The S03 concentration was calculated from the titration results of the
solution from the first impinger and filter. The S02 concentration was calcu-
lated from the titration results of the second and third impinger solution.
EVALUATION OF BACK CORONA IN THE PRECIPITATOR
Back corona or reverse ionization is a frequent operating problem in
electrostatic precipitators used to collect high resistivity dust. Back corona
is a condition where the electrical breakdown strength of the gas in the inter-
stitial regions of the dust deposited on the plates is exceeded. The electric
field in the deposit is proportional to the current density and the resistivity
37
-------
of the dust. If the electric field exceeds a critical value, a corona glow
forms on the dust deposit and inhibits precipitator operation.
The Soviets have devised a method of evaluating the occurrence of back
corona in electrostatic precipitators. Basically the procedure involves vary-
ing the secondary voltage to the precipitator and recording the corresponding
secondary current. This procedure was followed on two occasions at Allen in
the morning before the other tests were begun.
DETERMINATION OF FLUE GAS MOLECULAR WEIGHT
The carbon dioxide, molecular oxygen and carbon monoxide content of the
flue gas were measured primarily to determine the molecular weight of the stack
gas. These parameters were also used along with fuel analysis in determining
the theoretical gas volume.
Equipment
Figure 15 illustrates the integrated gas sampling train utilized for
molecular weight determination. It consisted of:
1. A stainless steel or pyrex glass probe fitted with a glass wool
filter to remove heavier particles from the gas stream. This was
followed by:
2. An air-cooled condenser for removal of excess moisture.
3. The sample was evacuated using a leak-free diaphragm pump. The
flow rate was adjusted by means of a needle valve and measured by
3
a rate meter which had a range of 0 to 0.035 ft per minute.
4. This system was attached by a "quick disconnect" fitting to a
flexible Tedlar bag housed in a rigid container.
Procedure
The equipment was set up as shown in Figure 15. After verifying that all
the equipment was leak free, the sample bag was then evacuated to within three
inches of mercury absolute pressure. The probe was then inserted into the flue
and purged. The bag was connected and the sample was drawn in at a rate
proportional to the flue gas velocity.
38
-------
GLASS WOOL
QUICK
DISCONECT
CONDENSER
BAG a BAG CONTAINER
Figure 15. U.S. integrated gas sampling train,
39
-------
Analysis
Analysis was performed inmediately by passing the collected gas through
an "Orsat" apparatus as many times as was necessary to arrive at a constant
reading. The absorption solution was replaced when more than 10 passes were
required. This analysis was repeated until three consecutive analyses of the
sample varied no more than 0.2 percent by volume for each component.
COAL SAMPLING AND ANALYSIS
During each test, Duke Power chemical technicians obtained coal samples
from the hoppers located above the coal pulverizer feeders (see Figure 2). In-
dividual samples were collected at each of 12 sample points at 20-minute inter-
vals throughout each test. The coal corresponding to each test was then com-
positeds systematically divided into smaller samples, pulverized, and air dried.
The samples were then transported to York Research Corporation's Laboratory in
Stamford, Connecticut, for analysis. The following standard procedural methods
were used in the performance of the coal analyses:
BTU value: ASTM D2015 - Adiabatic bomb calorimeter method.
Moisture: ASTM D3173 - Weight loss of sample measured when sample
was heated under controlled conditions.
Fixed carbon: ASTM D3172 - By difference of summation of percent
moisture, ash, volatile matter from 100.
Sulfur: ASTM D3177 - Method B - Barium sulfate precipitated from oxygen-
bomb calorimeter washings, and precipitate was filtered, ashed, and
weighed.
Carbon, hydrogen, and nitrogen: Sample was burned in an oxygen ••
atmosphere in a closed system, while products of combustion passed
through an absorption medium. Gases were analyzed by a Perkin-
Elmer Model 240 Elemental Analyzer.
ASH SAMPLING AND ANALYSIS
Fly ash samples were also collected by Duke Power chemical technicians
during each test. One sample was collected from each of the two boilers
associated with unit 3. The sample point for each boiler was located down-
stream of the economizer where the flue gas turns from vertical downflow to
40
-------
horizontal flow. The two samples from each test were composited and analyzed
by York Research.
PLANT DATA ACQUISITION
To insure a record of important test influences, boiler and precipitator
operating data were recorded throughout the tests at approximately half-hour
intervals. The only noteworthy incidents were several electrical trips in
individual sections of the precipitator. These trips were of short duration
and none was thought to have significantly influenced test results.
Complete records of the boiler and precipitator operating data are
contained in Appendixes N and P, respectively.
41
-------
42
-------
SECTION 4
RESULTS AND DISCUSSION
GAS FLOW AND TOTAL MASS CONCENTRATION DETERMINATIONS
Eight separate tests were conducted using both U.S. and Soviet equipment.
In the first six tests, both U.S. and Soviet equipment was used to traverse
ducts Bl and B2 on the inlet and outlet. The only difference in the sampled
areas was that the U.S. train was unable to sample port 4 of duct Bl on the
inlet because of a physical obstruction, and the Soviet train was unable to
sample port 5 of duct Bl on the inlet for the same reason. With both sample
trains, test 7 was a traverse of duct B2 only, and test 8 was a traverse of
duct Bl only.
For direct comparison of the U.S. and Soviet determinations, the results
of gas flow and total mass concentration tests are presented in Tables 4 and 5,
respectively. The results are expressed in metric and English units and in
actual gas conditions (as measured in the ducts) as well as dry standard or
normal gas conditions (defined here as dry gas corrected to 21° C (70° F) and
760 mm (29.92 in.) of mercury absolute pressure).
The.precipitator collection efficiency as measured by the U.S. train
averaged 99.69 percent with a standard deviation of 0.08 percent. The pre-
cipitator collection efficiency as measured by the U.S.S.R. train averaged
99.03 percent with a standard deviation of 0.57 percent. It should be noted
that the last three U.S.S.R. tests, which were conducted after several equip-
ment and procedural modifications, agreed more closely with the U.S. tests
than did the previous Soviet tests.
The following paragraphs discuss significant features of the comparative
tests. Complete records of the gas flow and mass concentration data are
contained in Appendixes D, E, F, G, H, and I.
43
-------
TABLE 4. SUMMARY OF COMPARATIVE GAS FLOW DETERMINATIONS
Test no.
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
7
7
7
7
8
8
8
8
Type of
train
U.S. -5
U.S. -5
U.S.S.R.
U.S.S.R.
U.S. -5
U.S.-5
U.S.S.R.
U.S.S.R.
U.S. -5
U.S. -5
U.S.S.R.
U.S.S.R.
U.S.-5
U.S. -5
U.S.S.R.
U.S.S.R.
U.S. -5
U.S.-5
U.S.S.R.
U.S.S.R.
U.S. -5
U.S. -5
U.S.S.R.
U.S.S.R.
U.S. -5
U.S. -5
U.S.S.R.
U.S.S.R.
U.S. -5
U.S. -5
U.S.S.R.
U.S.S.R.
Location
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Date
3/12
3/12
3/12
3/12
3/13
3/13
3/13
3/13
3/15
3/15
3/15
3/15
3/16
3/16
3/16
3/16
3/17
3/17
3/17
3/17
3/18
3/18
3/18
3/18
3/19
3/19
3/19
3/19
3/19
3/19
3/19
3/19
Time
start
1425
1405
1425
1405
1030
1030
1030
1030
1000
1004
1000
1004
0950
0948
0950
0948
0934
0918
0934
0918
1115
1046
1115
1046
1009
1010
1009
1010
1400
1405
1400
1405
Gas
Standard
DNCMM
5,061
5,700
5,859
6,002
4,950
5.902
5,376
5.945
4,982
5,690
5,200
5.502
4,749
5,520
5,292
5,469
4,905
5,722
5,257
5,778
5,084
5,743
5,335
5,688
2,685
3,322
2,601
3,306
2.592
2,459
2,390
2,329
DSCFM
178,737
201,310
206,913
211,973
174,793
208,424
189,854
209,945
175,923
200.928
183.642
194,319
167,717
194,936
186,883
193,144
173,217
202,086
185,643
204.055
179,529
202,794
188,420
200,865
94,529
117,314
91 ,866
116,735
91,520
86,821
84.404
82,233
flow
Actual
ACMM
11,658
12,757
13,089
13,221
11,548
13.374
12.548
13,477
11,355
12.953
11.683
12,510
11,133
12,841
12,226
12,758
11.375
12,958
12,147
13,081
11,572
12,979
12,214
12,886
6,191
7,444
5,964
7,467
6,041
5,555
5,479
5,253
ACFH
411,708
450,500
460,480
466,889
407,796
472,280
443,120
475.928
401 .009
457,430
412,594
441,796
393,142
453,467
431,754
450,547
401,685
457,615
428,988
461 ,954
408,660
458,330
431,823
455,065
218,633
262,894
210,632
263,696
213,332
196,156
193,485
185,526
Gas
H20
8.0
7.4
8.0
7.4
8.2
7.2
8.2
7.2
7.2
7.8
7.2
7.8
7.6
8.6
7.6
8.6
7.7
7.1
7.7
7.1
6.7
7.9
-
-
8.3
8.0
-
-
8.0
7.6
-
-
Composition
02
3.5
4.0
3.4
4.0
3.3
3.8
3.3
3.7
2.9
3.7
2.9
3.7
3.2
3.7
3.2
3.7
3.5
3.9
3.5
3.9
2.8
3.7
2.8
3.7
3.3
3.8
3.3
3.9
3.2
3.9
3.3
3.9
C02
14.4
13.9
14.7
13.8
14.7
14.1
14.7
14.1
15.2
14.0
15.2
14.0
15.1
14.5
15.1
14.5
15.0
13.5
15.0
13.5
14.9
14.1
14.9
14.1
15.5
14.0
15.4
13.8
15.2
13.6
15.4
13.8
-------
TABLE 5. SUMMARY OF COMPARATIVE TOTAL MASS CONCENTRATION DETERMINATIONS
in
Test no.
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
7
7
7
7
8
8
8
8
Type of
train
U.S.-5
U.S.-5
U.S.S.R.
U.S.S.R.
U.S. -5
U.S. -5
U.S.S.R.
U.S.S.R.
U.S.-5
U.S.-5
U.S.S.R.
U.S.S.R.
U.S. -5
U.S. -5
U.S.S.R.
U.S.S.R.
U.S. -5
U.S. -5
U.S.S.R.
U.S.S.R.
U.S.-5
U.S. -5
U.S.S.R.
U.S.S.R.
U.S. -5
U.S.-5
U.S.S.R.
U.S.S.R.
U.S.-5
IJ.S.-5
U.S.S.R.
U.S.S.R.
Total mass concentration
Location
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Date
3/12
3/12
3/12
3/12
3/13
3/13
3/13
3/13
3/15
3/15
3/15
3/15
3/16
3/16
3/16
3/16
3/17
3/17
3/17
3/17
3/18
3/18
3/18
3/18
3/19
3/19
3/19
3/19
3/19
3/19
3/19
3/19
Time
start
1425
1405
1425
1405
1030
1030
1030
1030
1000
1004
1000
1004
0950
0998
0950
0998
0934
0918
0934
0918
1115
1046
1116
1046
1009
1010
1009
1010
1400
1405
1400
1405
Standard
mg/NCM
11,325
44.98
5,883
64.75
12,172
33.92
6,341
127.16
10,792
49.33
6,936
73.87
10,943
28.84
7,697
53.60
12,063
33.31
6.223
54.69
10.858
25.23
11.005
46.29
12,598
26.61
13.560
60.40
10.630
30.74
12.890
78.01
gr/SCF
4.949
.0197
2.571
.0283
5.319
.0148
2.771
.0556
4.716
.0216
3.031
.0323
4.782
.0126
3.364
.0234
5.271
.0145
2.720
.0239
4.745
.0110
4.809
.0202
5.505
.0116
5.926
.0264
4.645
.0134
5.633
.0341
Actual
mg/ACM
4,917
20.10
2.643
29.40
5,217
14.97
2.717
56.10
4,734
21.67
3,087
32.49
4,668
12.40
3,332
22.98
5.202
14.71
2.693
24.16
4.770
11.16
4.508
20.43
5,464
11.88
5,914
26.74
4,560
13.61
5.623
34.58
gr/ACF
2.148
.0088
1.155
.0128
2.280
.0065
1.187
.0245
2.069
.0095
1.349
.0142
2.040
.0054
1.456
.0100
2.273
.0064
1.177
.0106
2.084
.0049
2.101
.0089
2.388
.0052
2.584
.0117
1.993
.0060
2.457
.0151
Mass Rate
kg/hr
3,439
15.38
2,068
23.32
3.615
12.01
2.045
45.36
3,226
16.84
2.164
24.39
3.118
9.55
2.444
17.59
3.550
11.44
1,963
18.96
3,312
8.69
3,523
15.80
2.030
5.30
2,117
11.98
1,653
4.53
1.848
10.90
Ib/hr
7.582
33.91
4,559
51.41
7,969
26.48
4,509
100.00
7,111
37.13
4,771
53.77
6.875
21.06
5,388
38.78
7,827
25.21
4,327
41.80
7,301
19.17
7,767
34.83
4.475
11.69
4,666
26.41
3.644
10.00
4.075
24.03
Free Ip ita tor
efficiency X
99.59
98.87
99.71
97.78
99.54
98.86
99.74
99.28
99.72
99.03
99.76
99.55
99.78
99.43
99.70
99.41
-------
U.S. Test Results
The inlet total mass concentration as measured by the U.S. train averaged
4,941 milligrams per actual cubic meter (nig/ACM) with a standard deviation of
319 mg/ACM. The outlet concentration as measured by the U.S. train averaged
15.06 mg/ACM with a standard deviation of 3.85 mg/ACM. These results show
very good reproducibility with the standard deviation being less than 10 per-
cent of the mean at the inlet and about 25 percent of the mean at the outlet.
The high-volume sampling train used at the outlet during the first two
tests was designed for use with metric measurement units. Since the computer
interface used in calculating the results was not compatible with metric units,
an alternate sample train designed for English units was used for the remain-
der of the tests.
A structural steel brace in front of port 4 on inlet duct Bl prohibited
that port from being tested with U.S. equipment. For the purpose of calcula-
tions the assumption was made that the mass concentration at that port was
equal to the average concentration of the particulate mass in the entire duct.
However, there were two other ports in inlet duct Bl with zero gas velocity;
hence the center ports including port 4 might have had velocities somewhat
higher than the average velocity; and, hence mass flow rates of dust somewhat
higher than the average. In fact, the preliminary velocity traverses, which
did include port 4, showed it to have a velocity 22 percent higher than the
average duct velocity. It is, therefore, possible that the inlet results are
biased to the low side because of the uncertainty concerning port 4 in duct
Bl.
For the first six tests, which involved both ducts, the average inlet
gas flow measured using the U.S. train was 11,440 actual cubic meters per
minute (ACMM) with a standard deviation of 191 ACMM. The outlet gas flow
measured by the U.S. train averaged 12,977 ACMM with a standard deviation of
212 ACMM. These results yield a consistent increase of gas flow at the out-
let averaging 13 percent more than the inlet. One possible reason for this
discrepancy is air leakage into the inlet ports or into some other openings
between the inlet and outlet test ports. This hypothesis is supported by the
consistently higher oxygen readings measured at the outlet. Leakage from
outside would also tend to dilute the particulate concentration at the outlet.
46
-------
The one exception to the higher outlet flows occurred in Test 8. Test 7
utilizing duct B2 follows the previous data of higher outlet versus inlet gas
flow; however, Test 8 utilizing duct Bl indicates lower outlet than inlet gas
flow. This result indicates the possibility of flue gas leakage from one duct
to the other.
An alternate explanation of the higher flow rates observed at the outlet
is the error introduced by not sampling port 4 on inlet duct Bl. Since this
port is known to have had a velocity higher than the duct average, the calcu-
lated gas flow of the entire inlet was reduced by the simplifying assumption
made concerning port 4.
U.S.S.R. Test Results
The inlet mass concentration determined by the U.S.S.R. train averaged
3,815 mg/ACM with a standard deviation of 1350 mg/ACM. The outlet mass con-
centration determined by the U.S.S.R. train was 30.86 mg/ACM with a standard
deviation of 11.27 mg/ACM. The standard deviation expressed as a percentage
of the mean, is approximately 36 percent for both of these determinations.
At the inlet the deviation between these results and the U.S. test results is
understandable since the tests were performed with the nozzle facing down, and
no precautions were taken until Test 6 to prevent loss of particulate catch
during port changes. Commencing with Test 6, an extension was added to the
balanced tube apparatus, which transported sample gas higher inside the thimble
and made the loss of dust less probable. Also, prior to Test 6, a leak was
found in one static tube of the U.S.S.R. balanced tube apparatus at the outlet.
The leaking apparatus was replaced with a new unit at that time, and Tests 6,
7, and 8 were performed with the new unit.
After the first four tests, when poor agreement between the U.S. results
and the U.S.S.R. results was apparent, the Soviet delegation suggested operating
the test equipment at a higher sample rate, i.e., 0.1 in. of water above zero
on the null manometer. When the balanced tube apparatus is calibrated for the
particular gas stream, adjustment of the sampling rate is sometimes required.
The Soviet delegation had hoped to spend more time calibrating the instruments
during the initial phase of the test program; however, the scheduling of the
program did not permit more time. The only calibrations performed during the
initial phase of the program were pitot tube calibrations, and the balanced
tube apparatus was operated but not calibrated. The results of the last three
47
-------
tests, which were performed after the equipment modification discussed pre-
viously and with the higher sampling rate, show particulate mass concentra-
tions which are considerably higher than the first five Soviet tests and also
higher than the corresponding U.S. tests. At the inlet the last three Soviet
tests averaged 5,348 mg/ACM with a standard deviation of 742 mg/ACM. At the
outlet the average was 27.25 mg/ACM with a standard deviation of 7.09 mg/ACM.
Part of the reason that the Soviet results are higher than the U.S. results
is probably that the balanced tube apparatus was operated at the higher
sampling rate, and therefore, more gas volume was pulled through the train,
resulting in a higher net filter weight. Since the U.S.S.R. calculations
involve a simple relationship between the net thimble weight and test time, a
superisokinetic sample rate can result in high test results.
The inlet Tests 1 through 4 were performed without testing port 5 on
duct Bl. This port was obstructed by a structural steel brace, which prevented
turning the probe to insert it into the port. It was agreed by the Soviet
delegation to calculate the results using the test time as if that port had
been tested, and by modifying the filter weight as if the port had been tested.
The following is an adjustment equation for the net filter weight and was used
for calculation of results from Tests 1 through 4:
gmc = ^a
where:
gm = grams of catch corrected,
gm, = grams of catch actual,
a
N = number of ports tested.
During each of Tests 5 through 8 the sample time of each of the two test
ports adjacent to the inaccessible test port was increased by 50 percent,
compensating for not testing the inaccessible port.
According to the Soviet delegation, normal cleanup procedure for the
balanced tube apparatus is to turn the apparatus with the nozzle opening
pointing upward. The apparatus is allowed to cool in this position. The
inside surface of the nozzle is brushed with a soft brush and adhering partic-
ulate is allowed to fall down into the filter. The filter is then removed
carefully.
48
-------
It was decided during the field testing to determine whether any partic-
ulate adhered to the inside nozzle surface after brushing. Tests 5 through 8
ended with cleaning the nozzle surface with acetone. The acetone was then
evaporated in a tared beaker and weighed. Although these results were not
used in calculation of the mass rate, they are presented in Table 6.
An additional study was undertaken to determine how much dust leaked
through the thimble holder cases and adhered to the inside surface of the
probe. Before and after Tests 4, 5, and 6, the probe was rinsed with acetone,
and the final rinse was evaporated in a tared beaker and weighed. These
results are also reported in Table 6. The flow of gas leaking through the
thimble holder cases was not monitored by the static pressure taps and, there-
fore, should not have affected the test results.
For the first six tests, which involved both ducts, the inlet gas flow as
measured by the U.S.S.R. train averaged 12,317 ACMM with a standard deviation
of 469 ACMM. The outlet gas flow as measured by the U.S.S.R. train averaged
12,984 ACMM with a standard deviation of 336 ACMM. Like the results obtained
with the U.S. train, the outlet gas flows were consistently higher than the
inlet gas flows, although the difference between inlet and outlet was not as
great with the U.S.S.R. train. The average difference of 5 percent between
the outlet and the inlet gas flow as measured by the U.S.S.R. train was prob-
ably caused by air inleakage through the inlet ports or elsewhere between the
TABLE 6. RESULTS OF SOVIET NOZZLE AND PROBE WASHES
Test no.
4
4
5
5
6
6
7
7
8
8
Location
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
Outlet
Date
3/16
3/16
3/17
3/17
3/18
3/18
3/19
3/19
3/19
3/19
Nozzle
wash
g
«•
-
.00820
.02382
.10725
.01109
.03667
.02186
.02693
.00951
Probe
wash
g
2.71665
.03436
.20444
.02980
.38368
.12777
-
-
-
-
Met.
filter weight
g
23.39200
.18237
20.35294
.19663
36.53119
.16382
21.94611
.12424
19.16634
.11305
49
-------
inlet and outlet sample ports. One significant difference existed between the
U.S. and U.S.S.R. procedures. Although an inaccessible center port on duct Bl
was omitted from particulate mass testing by both the U.S. and the U.S.S.R.
train, the U.S.S.R. pitot tube traverse, which was a separate part of the
Soviet test, included all ports. Since both center ports had significantly
higher than average velocity, the U.S.-determined gas flows had a tendency to
be biased to the low side.
DETERMINATION OF PARTICLE SIZE DISTRIBUTION
As explained in Section 3, three U.S.S.R. and two U.S. devices were used
to determine the particle size distribution. At the precipitator inlet both
Soviet cascade impactors (designated as Model I and Model II) and the Soviet
series cyclones were used. The Brink impactor was used for U.S. inlet testing.
'At the outlet the Soviet Model I and the U.S. Andersen impactor were used.
Comparative results of the three devices used at the inlet are presented
in Table 7. The mass median diameter (HMD) and geometric standard deviation
(ag) estimates were obtained from best judgment fits of the data to log-normal
cumulative distributions. Note that the inlet data are categorized by samp-
ling location because port-to-port traverses were not possible in each day's
test. Agreement between- the Brink and the Soviet Model II is reasonably good
in most cases, but the Soviet Model I deviates considerably. Cumulative and
differential distribution plots of all of the inlet data are categorized by
sample location and presented in Appendix J.
Figures 16, 17, and 18 are differential distribution plots of the inlet
data by device. These plots give an indication of the variability of separate
measurements, which is a function of both variability of true size distribu-
tion in the ducts with time and position and variability of individual tests
with the same device.
Table 7 also presents a comparison of the results obtained at the outlet
with the Andersen impactor and the Soviet Model I. With both devices the last
four tests are reasonably consistent, but the Soviet device indicates a higher
mass median diameter. This is probably a result of differences in the methods
of calibrating the impactors and may be resolved at a future date when com-
parable calibrations for the two sets of equipment become available. Differ-
ential distribution plots of the outlet data by device are presented in
Figures 19 and 20. Cumulative and differential distribution plots of all of
the outlet data are categorized by date and presented in Appendix J.
50
-------
TABLE 7. COMPARATIVE RESULTS OF PARTICLE SIZING DEVICES
Inlet tests
Brink Soviet Model I
Date Location
March 13 Bl, Port 7
March 15 B2, Port 4
March 15 Bl , Port 3
March 16 Bl , Port 5
March 17 B2, Port 2
March 19 B2, Port 6
MMD,ym ag
17 3.4
28 3.7
28 3.3
17 4.1
18 3.8
26 3.6
MMD,ym
70
-
35
-
18
45
£2
11
-
2.9
-
2.8
5.6
Soviet Model II
MMD,ym ag
54 3.2
24 2.4
31 2.8
23 2.3
22 2.4
23 3.1
Outlet tests
Date
March 12
March 13
March 15
March 16
March 17
March 18
March 19
Andersen
MMD,ym ag
4.1 4.0
6.4 2.6
30 20
11 3.1
10 2.5
11 3.7
9.4 3.9
Soviet
Model I
MMD ,ym ag
4.0
17
9.4
22
17
18
14
4.0
5.5
7.8
2.9
2.8
4.2
2.7
51
-------
10'
O
r-i
1.0(
10
Aerodynamic Particle Diameter, Micrometers
100
Figure 16. Inlet size distribution on a differential mass
distribution basis - Brink data.
52
-------
10'
Q
&>
0
101
10'
1 1 10
Aerodynamic Particle Diameter, Micrometers
1QQ
Figure 17. Inlet size distribution on a differential mass
distribution basis - Soviet Model I data.
53
-------
10'
Q
cn
O
10'
10
10'
.1
1 10 100
Aerodynamic Particle Diameter, Micrometers
Figure 18. Inlet size distribution on a differential mass
distribution basis - Soviet Model II data.
54
-------
Q
O
iH
<
10
10"2
.1
1 10
Aerodynamic Particle Diameter, Micrometers
100
Figure 19- Outlet size distribution on a differential mass
distribution basis - Andersen data. (Open
symbols - 120 minute samples, solid symbols -
288 and 360 minute samples).
55
-------
tn
S
Q
tn
0
10~2
Aerodynamic Particle Diameter, Micrometers
100
Figure 20.
56
-------
The sample entry nozzles for all the Soviet samplers and for the Brink
impactors were straight while "buttonhook" nozzles were used with the Andersen
impactors. The particulate which was recovered from the sampling nozzles was
added to the cyclone or first-stage collection in all cases. Filter catches
and cyclone or first-stage catches were omitted in all the differential size-
distribution data plots because of the difficulty in assigning a representa-
tive particle size to them.
Sampling rates and gas volumes sampled were determined with orifice meters
and stopwatches for the Soviet equipment and the Brink impactors, while orifice
meters and dry gas meters were used with the Andersen impactors. Calibrations
of the Soviet orifice meters with a dry gas meter at the test site subsequent
to the actual testing determined that the Soviet orifice meter used at the
inlet indicated 9 percent less than the true flow, and the outlet orifice
meter, as used, indicated 19 percent less than the true flow. The flow rates
and all concentrations and size cuts reported herein have been adjusted to
account for this discrepancy.
In addition to the complete set of plotted data, Appendix J contains a
tabulation of data and results by run after all corrections for flow rate,
temperature, and other physical parameters were made. Corrections were also
made for gas phase interferences by subtracting the appropriate blank or con-
trol correction from each stage weight. Appendix J reports the average weight
changes and standard deviations for the blank, or control, runs. Unfortunate-
ly, a single homogeneous batch of impaction substrates was not available from
the manufacturer for use with the Andersen impactors. As a result, the data
for the last 2 days of testing with these impactors were somewhat confused by
intermixing of two different substrate sets, one of which was much more
reactive with the flue gas than the other. This can readily be seen in the
data from the final Andersen blank runs in Appendix J. Correction of the
Andersen data for the final 2 days of testing was difficult because of this
mixing of the materials, and the proper corrections may not have been applied
for all stages for these runs although best judgment was used to correct for
gas phase interference as accurately as possible.
The stage cut sizes (diameters for 50 percent collection) were based on
Southern Research Institute calibrations for the Andersen and Brink impactors
and on Soviet calibrations for the Soviet devices. All cut sizes were calcu-
57
-------
lated on the basis of unit'density spheres, thus the reported values are aero-
dynamic diameters. Calibration of the Andersen has shown that cut sizes of
the first and second stages are much closer than theory predicts. This
feature results in an anomalous spike in the differential mass distribution
plot if normal calculation methods are used. To avoid this anomaly, the
second and third stages were lumped together to calculate the differential
mass distribution. In Appendix J, the lumped geometric mean diameter and
resulting differential mass distribution value are shown adjacent to the
uncorrected result. The lumped values were used in all differential mass
distribution plots.
Data obtained with the Soviet series cyclones are included in Appendix J.
However, because the data are not amenable to the same treatment as impactor
data and no data reduction techniques were provided, the cyclone results are
not shown on the figures in this section. Preliminary analysis of some of the
cyclone data indicated general agreement between that data and the data ob-
tained from the impactors.
A summary of total particulate mass concentrations as determined with the
various particle sizing devices and with the total mass devices is given in
Table 8. Fractional efficiencies derived from the Brink and Andersen impactor
data (excluding anomalous extreme values) are shown in Figure 21. The outlet
data, especially the high concentrations of large particles obtained with both
the Soviet and U.S. devices, indicate that rapping reentrainment losses in this
precipitator contribute significantly to the overall emissions.
FLY ASH RESISTIVITY DETERMINATION
The results of the fly ash resistivity tests conducted with the Southern
Research Institute point-to-plane probe are presented in Table 9. The average
value of the resistivity during the test period was 1.9 x 10 ohm-centimeters.
SULFUR DIOXIDE AND SULFUR TRIOXIDE DETERMINATIONS
Sulfur oxide tests were performed at the precipitator inlet ducts using
the U.S. EPA Method 8 test apparatus. Six tests were performed, coinciding
with the first six particulate mass tests performed on the precipitator. On
a dry basis the sulfur trioxide results averaged 2.38 ppm by volume with a
standard deviation of 1.91 ppm, and the sulfur dioxide concentration averaged
58
-------
TABLE 8. AVERAGE PARTICIPATE MASS LOADINGS BY SAMPLING DEVICE
en
UD
Inlet
Device
Grand average
(mg/ACM)
Standard devi-
ation (mg/ACM)
Average omitting
extremes (mg/
ACM)*
U.S. mass train
All runs Runs 6,7,8
4,942 4,931
319 473
4,918
i
U.S.S.R. mass train
All runs Runs 6,7,8
3,815 5,348
1,350 742
3,660
Soviet
Brink Model I
3,244 10,491
828 11,336
3,254 5,832
Soviet Soviet
Model II cyclone
4,794 4,596
1,635 1,613
4,424 4,136
Outlet
Device
Grand average
(mg/ACM)
Standard devi-
ation (mg/ACM)
Average omitting
extremes (mg/
ACM)*
U.S. mass train
All runs Runs 6,7,8
15.06 12.22
3.85 1.26
14.61
U.S.S.R. mass train
All runs Runs 6,7,8
30.86 27.25
11.27 7.09
28.39
Ander- Soviet
sen Model I
5.47 6.90
1.88 5.02
5.49 6.09
*The single highest and single lowest values omitted in each case.
-------
99.99
cn
O
40.
Aerodynamic Particle Diameter, urn
Figure 21. Precipitator fractional efficiency from Brink
and Andersen data.
60. 80. 100.0
-------
TABLE 9. FLY ASH RESISTIVITY RESULTS
Date
'^•••••Mm*
3/13
3/16
3/17
3/18
Gas
Time Temperature ( C)
0900-1000
1000-1100
1445-1545
1030-1130
1030-1130
1230-1330
1430-1530
1645-1745
347
349
342
344
346
345
343
343
Resistivity (n-cm)
3.5 x 1010
1.2 x 1010
1.5 x 1010
3.6 x 1010
1.5 x 1010
1.4 x 1010
1.3 x 1010
1.2 x 1010
818.2 ppm with a standard deviation of 124.0 ppm. Table 10 reports the
results of each test.
RESULTS OF BASIC CORONA TESTING
The Soviet method of evaluating the effects of back corona by studying
the voltage/current relationship of the precipitator produced no evidence of
back corona at Allen Unit 3. The V-I data are shown for reference in
Appendix Q.
GAS AND COAL ANALYSES - THEORETICAL GAS VOLUME
Theoretical gas volumes (TGV) were calculated from the ultimate and
proximal coal analysis, gas composition (Orsat) data, and coal flow rate. The
coal analyses are presented in Table 11. The moisture of the coal is reported
as received by York Research Corporation after preparation at Plant Allen.
TGV calculations assumed that the moisture of the coal entering the burners of
Unit 3 on a given day was the same as the moisture of the corresponding sample.
Results of the TGV calculations are presented in Table 12 and compared
with measured gas volumes. The tests were performed on ducts Bl and B2, which
comprise 50 percent of the total duct cross-sectional area at the test loca-
tion. The calculated gas volume is 50 percent of the total volume of flue gas
entering the Unit 3 precipitators. The TGV's are also based on an assumed
coal flow rate which was calculated using an average heat rate supplied by
Duke Power. The assumption was necessary because there were no coal feed
61
-------
TABLE 10. SULFUR OXIDES TEST RESULTS
Test no. Date
1 3/12
2 3/13
3 3/15
4 3/16
5 3/17
6 3/18
Time
start
1405
1053
1027
1018
0950
1137
' Gas
Standard
DNCMM
6,074
5,093
4,741
5,052
4,883
4,782
DSCFM
214,494
179,849
167,428
178,413
172,451
168,887
Flow
Actual
ACMM
13,927
11,488
10,896
11,698
11,324
10,931
TABLE 11
ACFM
491 ,831
405,710
384,779
413,097
399,915
386,029
. COAL
S03 Emission Si
ppm
4.16
1.77
5.37
.933
.942
1.10
kg/hr
5.06
1.80
5.09
.94
.92
1.06
Ib/hr
11.15
3.98
11.23
2.08
2.03
2.33
Ppm
819.8
627.3
991.5
906.4
780.5
783.6
0? Emission
kg/hr
796.65
511.11
752.09
732.58
609.78
599.53
Ib/hr
1,756.28
1,126.79
1,658.05
1,615.04
1,344.32
1,321.72
ANALYSIS
Btu/lb
% Ash
IX.
Test no.
1
2
3
4
5
6
7 & 8
Dry
12,535
12,627
12,559
12,540
12,429
12,650
12,453
As
rec'd
12,344
12,473
12,338
12,397
12,296
12,502
12,315
Dry
14.67
13.64
13.41
14.05
14.09
13.56
13.17
As
rec'd
14.45
13.47
13.23
13.89
13.94
13.40
73.02
Dry
1.04
1.01
0.98
0.99
0.98
0.97
0.94
As
rec'd
1.02
1.00
0.97
0.98
0.97
0.96
0.93
Dry
1.39
1.32
1.30
1.42
1.37
1.38
1.37
As
rec'd
1.37
1.30
1.28
1.40
1.36
1.36
1.35
Dry
72.61
72.68
73.13
72.02
72.45
72.78
73.35
As
rec'd
71.51
71.79
72.14
71.20
71.67
71.93
72.53
Dry
4.83
4.85
4.85
4.76
4.85
4.93
4.93
As
rec'd
4.76
4.77
4.78
4.71
4.80
4.87
4.88
%H20
1.52
1.22
1.36
1.14
1.07
1.17
1.11
TABLE 12. MEASURED AND THEORETICAL GAS VOLUMES
Test no.
1
2
3
4
5
6
7 & 8
Date
3/12
3/13
3/15
3/16
3/17
3/18
3/19
Time
start
1405
1030
1000
0948
0918
1046
1009
Measured
Soviet
DNCMH
5,859
5,376
5,200
5,292
5,257
5,335
4,991
ACMM
13,089
12,548
11,683
12,226
12,147
12,214
11,443
gas flow*
EPA Method 5
DNCMM
5,061
4,950
4,982
4,749
4,905
5,084
5,277
ACMM
11,658
11,548
11,355
11,133
11,375
11,572
12,232
TGV
calculation's
DNCMM
6,947
6,940
6,846
7,185
7,054
6,668
7,031
ACMM
•16,034
16,038
15,635
16,785
16,277
15,310
16,126
*Measured flow through inlet ducts Bl and 82 which comprise 50 percent of the total cross-sectional
duct area entering the precipitator.
Fifty percent of total calculated gas flow entering the precipitator.
62
-------
scales located at the Allen Steam Plant. The TGV calculations indicate that
the total gas flow might not be equally distributed to the four ducts entering
the precipitator. Complete computer summary sheets of each test can be found
in Appendix 0.
FLY ASH ANALYSIS
Results from the chemical analysis of the combined fly ash samples are
presented in Table 13. As stated in Section 3, each test sample was made up
of a composite of fly ash collected from the two boilers associated with
Unit 3. There is no assurance that the collected fly ash was identical in
size distribution-to the ash entering the precipitator. Since chemical
analysis of fly ash is known to depend to some extent on the particle size,
the results may not be precisely indicative of the composition of ash collected
by the precipitator or of the small quantity of ash contained in the stack gas.
TABLE 13. FLY ASH CHEMICAL ANALYSES (all results in percent of total mass)
Test no.
Loss on Ignition
Si02
A1203
Fe203
Ti02
CaO
MgO
Na20
K20
L120
soa
P205
Total
1
2.54
55.91
27.20
8.41
1.51
1.16
0.92
0.46
1.09
0.032
0.40
0.34
100.07
2
2.18
55.89
28.95
7.90
1.05
1.05
0.77
0.49
1.28
0.033
0.15
0.33
100.07
3
5.11
54.30
28.35
7.21
0.92
0.92
0.74
0.45
1.08
0.034
0.16
0.30
100.07
4
2.80
55.56
28.63
7.60
1.34
1.17
0.74
0.50
1.06
0.032
0.20
0.37
100.00
5
2.84
55.92
28.63
7.77
1.13
1.08
0.73
0.47
0.78
0.029
0.25
0.36
100.04
6
2.25
55.63
27.69
9.17
1.05
1.26
0.92
0.52
1.06
0.030
0.14
0.38
100.10
7
2.48
56.37
28.53
7.67
0.78
1.11
0.95
0.51
1.07
0.028
0.20
0.35
100.05
8
5.21
54.77
29.87
6.23
0.94
1.16
0.87
0.47
0.84
0.026
0.26
0.33
100.98
63
-------
TECHNICAL REPORT DATA
(Please read Inuructiom on the reverse before completing)
1. REPORT NO.
EPA-600/2-77-002
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
v* • » • ^ mf r—ii v v ***»•!••* • ^vm
Comparative U.S. /USSR Tests of a Hot-Side
Electrostatic Precipitator
5. REPORT DATE
January 1977
6. PERFORMING ORGANIZATION CODE
7. AUTMOR(S)
Charles H. Gooding, Joseph D. McCain, and
Diane K. Sommerer
3. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC 27709
10. PROGRAM ELEMENT NO.
1AB012: ROAP 21ADL-029
11. CONTRACT/GRANT NO.
68-02-1398, Task 33
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
Task Final: 1/75-9/76
14. SPONSORING AGENCY CODE
EPA-ORD
is. SUPPLEMENTARY NOTES jERL-RTP project officers for this report are D. C. Drehmel and
D. B. Harris, 919/549-8411 Exts 2925 and 2557, Mail Drops 61 and 62.
16. ABSTRACT
The report describes a U.S./USSR cooperative test program to quantify and
characterize particulate emissions from a U.S. coal-burning power plant boiler,
equipped with a hot-side electrostatic precipitator, at Duke Power Co. 's Allen Steam
Station in March 1976. U.S. and Soviet equipment and procedures were used to deter-
mine flue gas composition and velocity, total particulate mass concentration of the
gas stream, particle size distribution, electrical resistivity of the particulate enter-
ing the precipitator, evidence of back corona in the precipitator, SO2 and SOS concen-
trations in the flue gas, and chemical composition of the fuel and fly ash. The test
site and test procedures are described. Results of the comparative tests are presen-
ted and discussed. In 1972, the U.S. and the USSR signed a bilateral agreement
pledging cooperation on environmental protection. As part of this agreement, the
Working Group on Stationary Source Air Pollution Control was subsequently formed
by the U.S. EPA and the USSR Research Institute of Industrial and Sanitary Gas Clea-
ning to conduct cooperative programs in several areas of air pollution control tech-
nology, including particulate emission control. This is one of those cooperative
programs.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lOSNTIFISHS/OPSN ENOSO TERMS
COSATI Field/Group
Air Pollution
Electrostatic
Precipitators
Flue Gases
Dust
Coal
Coronas
Boilers
Electric Power
Plants
Sulfur Oxides
Electrical Resis-
tivity
FlvAsh
Air Pollution Control
Stationary Sources
Particulates
Back Corona
13B
21B
11G
21D
13A
10B
07B
20C
3. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report}
Unclassified
21. NO. OF PAGSS
76
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
SPA Form 2220-1 (3-73)
64
------- |