&EPA
United States
Environmental Protection
Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
EPA-600/2-79-188
September 1979
Research and Development
Applicability of
Transmissometers to
Opacity
Measurement of
Emissions
i
Oil-Fired Power and
Portland Cement
Plants
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development U S Environmental
Protection Agency have been grcuced mtc nine series These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields
The nine series are
1 Environmental Health Effects Research
2 Environmental Protection Technology
3. Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8 "Special" Reports
9 Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series This series describes research performed to develop and dem-
onstrate instrumentation, eguipment. and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology reguired for the control and treatment
of pollution sources to meet environmental quality standards
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161
-------
EPA-600/2-79-188
September 1979
APPLICABILITY OF TRANSMISSOMETERS TO
OPACITY MEASUREMENT OF EMISSIONS
Oil-Fired Power and Portland Cement Plants
by
William D. Conner, Kenneth T. Knapp, and John S. Nader
Emissions Measurement and Characterization Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for publi-
cation. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
-------
ABSTRACT
In-stack transmissometers were evaluated for their capability to moni-
tor the opacity of smoke-stack plumes emitted by certain sources. In-stack
transmissometers were installed on three portland cement plants and three
oil-fired power plants. Tests were conducted to determine their performance
in four areas: the adequacy of the U.S. Environmental Protection Agency-
promulgated transmissometer design and performance specifications for the
sources, the correlation between the opacity of the emissions measured in the
plumes and in the stacks of the sources, as well as the existence of a func-
tional relationship between the transmissometer-measured opacity and mass
concentration of the particulate emission.
For portland cement plants, the results indicated that the promulgated
design and performance specifications for transmissometers are adequate. The
in-stack transmissometer measured opacity was found to be representative of
the plume opacity and to correlate to the mass concentration of the particu-
late emissions.
For oil-fired power plants, the results indicated that the promulgated
design and performance specifications for transmissometers are adequate
except for the spectral response design specification. For opacity monitoring
of submicrometer particulate emissions such as those from oil-fired power
plants, the allowable peak spectral response range of the transmissometer
should be reduced. In addition, the in-stack transmissometer-measured opaclcy
for oil-fired power plants was generally lower than the plume opacity, indi-
cating that much of the particulates in the plumes were forming in the
atmosphere out of the stack. Correlation was observed between the trans-
mic^ometer-measured opacity and mass concentration of the oil-fired power
plant particulate emissions.
m
-------
CONTENTS
Abstract iii
Figures vi
Tables vii
Acknowledgments viii
1. Introduction 1
2. Conclusions 3
3. Design Specifications 5
Spectral specifications 6
Collimation specifications 7
4. Performance Specifications 11
In-stack and plume opacity 11
In-stack opacity and mass concentration 12
5. Measurement Methods 14
Opacity measurement 14
Particulate mass measurement 15
Particle size measurement 15
Multiwavelength opacity measurement 18
6. Cement Plant Tests 19
Test sites 19
Results and discussion 23
7. Oil-Fired Power Plant Tests 33
Test sites 33
Results and discussion 34
References 47
-------
FIGURES
Number Page
1 Measured opacities as a function of true (theoretical)
opacities for a transmissometer with a 20° angle of view and
aerosols with different mean particle sizes ... 9
2 Measured opacities as a function of true (theoretical) opacities
for transmissometers with 2° and 20° angles of view and speci-
fied aerosol 10
3 Photograph of mobile lidar system 16
4 Schematic of cement plant test site I 20
5 Schematic of cement plant test site II 21
6 Schematic of cement plant test site III 22
7 Particle size distributions of portland cement plants 28
8 In-stack opacity of emissions from portland cement plants as a
function of the mass concentration of the particulates at
standard conditions and dry gas 31
9 In-stack opacity of emissions from portland cement plants as a
function of the mass concentration of the particulates at
actual conditions 32
10 Schematic of oil-fired power plant test site I .... 35
11 Schematic of oil-fired power plant test site II 36
12 Schematic of oil-fired power plant test site III 37
13 Particle size distributions of oil-fired power plants ....... 41
14 Particle size distribution of oil-fired power plant III . 42
15 In-stack opacity of emissions from oil-fired power plants as a
function of the mass concentration of the particulates at
standard conditions and dry gas 45
-------
FIGURES (Continued)
Number Page
16 In-stack opacity of emissions from oil-fired power plants as a
function of the mass concentration of the particulates at
actual conditions 46
TABLES
1 Specifications for Transmissometer Opacity Monitors 2
2 Mobile Lidar System Characteristic 17
3 Cement Plant Opacity Measurements 24
4 Cement Plant II Anderson Impactor Data 26
5 Cement Plant III Anderson Impactor Data 26
6 Cement Plant IV Pilat Impactor Data 27
7 Oil-Fired Power Plant Opacity Measurements 38
8 Oil-Fired Power Plant Pilat Impactor Data 40
-------
ACKNOWLEDGMENTS
We wish to express our appreciation to Mr. Norman White and Mr. Raymond C.
Steward of the U.S. Environmental Protection Agency, and Mr. Harold B. McElhoe
of Northrop Services, Inc., for assisting with the opacity and participate
size measurements. We also extend our thanks to the members of the Engineering
Science Division of McLean, Virginia, and Cottrell Environmental Sciences, Inc.,
of Bound Brook, New Jersey, for helping to measure the particulate mass con-
centrations.
vm
-------
SECTION 1
INTRODUCTION
In-stack transmissometer design and performance specifications for
monitoring the opacity of visible emissions were promulgated by the U.S.
Environmental Protection Agency (EPA) on October 6, 1975 (1). Although the
specifications (see Table 1) are intended to be geneially applicable to
monitoring the opacity of emissions from a wide range of industries, the
transmissometers should still be evaluated for each industry. The evaluation
should determine four performance areas:
the adequacy of the transmissometer design specifications for the
s'ource,
the adequacy of the transmissometer performance specifications for
the source,
the correlation between the opacity of the emissions measured in
the plume and in the stack of the source, and
the existence of a functional relationship between the transmis-
someter-measured opacity and mass concentration of the particulate
emissions.
The most extensive studies concerning the application of in-stack trans-
missometers to opacity measurements are reported for coal-fired power plants.
"Data are reported on the effects of transmissometer spectral response and
collimation characteristics on the measurement of opacity at a coal-fired
power plant (2), on correlation between in-stack and plume opacities at coal-
fired power plants (2,3,4), on correlations between opacity and mass concentra-
tion at coal-fired power plants (5,6), and on the performance stability of a
transmissometer at a coal-fired power plant (7). These studies have demon-
strated that the EPA-promulgated specification for transmissometers are adequate
for the conventional coal-fired power plant with electrostatic precipitator
-------
TABLE 1. SPECIFICATIONS FOR TRANSMISSOMETER
OPACITY MONITORS
Parameter
DESIGN SPECIFICATIONS
Specification
Spectral response
B Angle of view
C Angle of projection
Peak: within 500 to 600 nm
Mean: within 500 to 600 nm
Limit: less than 10% of peak
outside 400 to 700 nm
5° maximum (total angle)
5° maximum (total angle)
Parameter
PERFORMANCE SPECIFICATIONS
Specification
A Calibration error
B Zero drift (24 hours)
C Calibration drift (24 hours)
D Response time
E Operational test period
< 3% opacity
< 2% opacity
< 2% opacity
10 sec (maximum)
168 hr
particulate control equipment'. Additional studies have reported on the appli-
cation of transmissometers to the measurement of opacity at a steel mill (basic
oxygen furnace), at a cement plant, and at a sulfuric acid plant (8), at a
petroleum plant (catalytic cracker regenerator), an asphalt concrete plant, a
sludge incinerator, a secondary brass/lead smelter, and an oil-fired power
plant (9).
In this report, the specifications required for transmissometer opacity
monitors by EPA are reviewed. Also, the field tests are described that were
conducted at several portland cement and oil-fired power plants to evaluate
the adequacy of the specifications for monitoring the opacity of emissions
from these two sources.
-------
SECTION 2
CONCLUSIONS
The data for portland cement plants showed that the promulgated transmis-
someter opacity monitor design specifications for spectral response and
collimation are both adequate. The peak and mean spectral response specifica-
tion limits of 500 and 600 nm could be broadened to ^00 and 700 nm without
affecting the opacity measurement. Any broader expansion into the ultraviolet
and infrared spectral regions should be avoided, primarily because of water
vapor and other gas absorption bands. For similar reasons, the overall spectral
response limits of 400 and 700 nm should not be changed.
The performance (drift stability) specifications for the transmissometers
were found adequate for cement plants. Consequently, no need for changes were
indicated.
The in-stack opacity measurements when compared to out-of-stack plume
opacity measurements at cement plants indicated that the in-stack transmisso-
meter will monitor the plume opacity when installed and spanned according to
the promulgated specifications. However, there was a notable exception.
During cold weather the opacity of the plume can become higher than the in-stack
opacity due to condensation. When this occurred, the condensation and higher
opacities were not observed in the stack.
Good functional relationships were observed between the opacities and par-
ticulate mass concentrations of the portland cement plant emissions. Light
attenuation coefficient versus mass concentration curves for four plants were
linear with regression line slopes within ±18 percent when the mass measure-
ments were reduced to standard conditions and dry gas. When the mass measure-
ments were for actual conditions, the slopes were within ±33 percent.
The data for oil-fired power plants showed a potential problem with the
promulgated design specifications for spectral response when applied to this
-------
source. The opacities of the participate emissions from these plants were
observed to be very dependent upon the wavelength of the light within the visi-
ble part of the spectrum. The 500 to 600 nm wavelength range allowed in the
peak spectral response of transmissometers by the EPA specifications can result
in a variation of about ±25 percent (±5 percent opacity) relative to 20 percent
opacity green light (550 nm) measurement. This potential 10 percent opacity
difference between transmissometers with 500 and 600 nm peak spectral responses
is excessive and should be reduced. The potential error can be cut in half by
reducing the allowable peak response range of 100 nm to 50 nm. In reducing the
peak response range specification to 50 nm, the present range should be reduced
by subtracting 25 nm from both ends to allow a narrower range from 525 to 575 nm
wavelength. The new range would encompass and remain symmetrical with the peak
(phototopic) response of the human eye (550 nm). Such a reduction is recom-
mended when transmissometers are used to measure the opacity of emissions with
particulates as small as those emitted by oil-fired power plant sources.
The performance (drift stability) specifications for the transmissometers
were found adequate for oil-fired power plant in-stack opacity monitoring. No
need for change was indicated.
The data did show a problem when the in-stack transmissometer was used to
monitor the opacity of plumes emitted by an oil-fired power plant. The in-stack
opacity measurements were generally lower than the out-of-stack plume opacity
measurements. The higher plume opacities were attributed to condensing sul-
furic acid (H^SO^) as the gases exited from the stacks. The differences are
related to such variables as the sulfur content of the fuel, boiler excess air,
and probably atmospheric variables such as ambient temperature, humidity, etc.,
which would preclude a simple correction factor.
Functional relationships were also observed between the opacities and parti-
culate mass concentration of the oil-fired power plant emissions. Two distinct
light attenuation versus mass concentration curves were apparent. One curve
was for plants operating at excess oxygen firing levels of 1.5 to 3.0 percent.
The other curve was for a plant specially designed to normally operate at a low
excess oxygen level of 0.2 percent. Both curves were linear with slopes that
differed by factors of 3.0 and 2.2 when the mass measurements were for standard
conditions (dry gas) and actual conditions, respectively.
-------
SECTION 3
DESIGN SPECIFICATIONS
Design specifications to define opacity measurement accuracy are unique
to the opacity monitor. For gas monitors, accuracy is specified relative to
measurements obtained by a prescribed reference (compliance) method. For
opacity measurement, however, the compliance method is the trained observer,
while the reference opacity measurement method for the trained observer
certification is a transmissometer. Consequently, it is necessary to estab-
lish design specifications based on those optical characteristics of the
transmissometer that affect and define the opacity measurement obtained by
the transmissometer.
Two design requirements are specified for transmissometers used to moni-
tor opacity (1):
The peak spectral response of the transmissometer shall be between
500 and 600 nm wavelength and the responses at 400 and 700 nm wave
length shall be less than 10 percent of the peak response.
The angle of projection of the light beam and the angle of view of
the receiver shall both be less than 5° total angle.
Spectral response and collimation angles can have very significant effects on
the opacity measurement performance of transmissometers. They must be stand-
ardized to obtain valid opacity measurement. In general, the effects of
different spectral responses and collimation angles on opacity measurements
are related to the size of the particulates in the effluent: the smaller the
particle the greater the wavelength effect, the larger the particles the
greater the collimation effect.
-------
SPECTRAL SPECIFICATIONS
To determine the adequacy of the design specifications for spectral res-
ponse, the opacity-wavelength characteristic of the emissions near the center
(green) of the visible spectrum are first measured. This opacity-wavelength
characteristic is then used to calculate the differences between opacity
measurements by transmissometers with peak spectral responses at the specifi-
cation extremes of 500 and 600 nm with respect to a transmissometer operating
at 550 nm. The potential error is calculated relative to the opacity meas-
urement that would have been obtained at 550 nm because this is the peak
spectral response of the human eye and would be most closely related to the
visual effects of the emissions. The potential error is calculated for the
promulgated new source opacity standard of the source being examined.
For the calculation, Bouquer's law (often called the Lambert-Beer law)
for defining the transmission of light (T) through an aerosol is used. The
law may be written as
T = exp. (-N a" Q" £) (Eq. 1)
where N = the number concentration of the particles in the aerosol
IT = the mean particle area
Q = the mean particle extinction coefficient, and
£ = the length of the light path through the aerosol
If the opacity of the aerosol measured simultaneously at two different wave-
lengths in the visible spectrum is observed to be different, the mean diameter
of the particles can be expected to be less than 1 ym, and the extinction
coefficient can be expected to be varying in proportion to X~n where
0 < n < 4 (10). If the plume transmittances are TI and T2 when measured at
wavelengths Xi and X2, a value of n is obtained for an intermediate X from
the following relationship:
Jin T2/£n Tx = (\2/Xi)~" (Eq. 2)
The opacity (Ox) at 600 nm and 500 nm relative to the opacity standard (0 ) at
550 nm are calculated, from the relationship
Ox = 1 - (1 - 0S)R (Eq. 3)
with R = (.55/.60)n and R = (.55/.50)n, respectively.
-------
COLLIMATION SPECIFICATIONS
Particulate opacity is caused by a reduction in transmitted light due to
the scatter and absorption of light by the particles in the emissions. If the
correct opacity is to be measured, the transmissometer receiver must not detect
the light scattered by the particulates. To exclude the scattered light, the
light projection and receiver viewing angles of the transmissometer must be
limited to small angles, so that as much collimation is obtained as is
practical (7). In practice, complete collimation cannot be obtained; some
scattered light will be measured, resulting in an opacity measurement that is
lower than the opacity that would be measured by a theoretical, completely
collimated transmissometer.
The effect of scattered light measurement on the measurement accuracy of
transmissometers has received considerable study (2,11,12). These studies show
that the error is generally insignificant for measurements on aerosols with
particulate sizes less than 1/2 ym diameter, even when transmissometers with
relatively poor collimation are used. For larger particles, however, the
error can be significant if transmissometers with poor collimation are used.
The error is primarily due to the measurement of light scattered by the
particles in the forward direction by diffraction and has been shown to
increase with decreasing transmissometer collimation (increasing light pro-
jector and receiver viewing angles), increase with increasing particle size,
decrease with increasing width of the particle size distribution, and not
change significantly with changing composition or refractive index of the
particles. In addition, when the measurement pathlength is large compared to
the diameter of the beam and receiver aperture, all of the light scattered
within the receiver and projector collimation angles will not be measured.
As a result, the error will decrease with increasing pathlength. The theo-
retical calculations by Ensor and Pilat (12) illustrate the need for collimat-
ing transmissometers. Their data can be used to show the potential magnitude
of the error. Figure 1 depicts the measured opacities that would be obtained
with a transmissometer having a 20° receiver angle as a function of the true
opacity (the theoretical opacity that would be obtained with perfect collima-
tion) for aerosols with log-normal particle distributions with standard
deviations of three and different mean sizes. Figure 2 shows the opacities
-------
that can be expected with transmissometers have 2° and 20° receiver angles of
view for an aerosol with a log-normal particle size distribution with a
standard deviation of 3 and mean diameter of 6 ym.
Figures 1 and 2 are for transmissometers with complete collimation (no
divergency) of the projected light beam. A transmissometer that used a laser
for the light projector would have essentially no divergence of the projected
light beam. However, most commercial transmissometers use incandescent lamps
and collimating optics. Experimental measurements by Peterson and Tomaides
(2) of the effects of the transmissometer light beam projection and receiver
viewing angles on measuring the transmittance of a coal-fired power plant
emission showed that collimation of the light projection angle was as important
as collimation of the view of the light receiver.
The smallest collimation angles that appear feasible before alignment pro-
blems and costs become unreasonable are about 2° total angle. The. collimation
angles of most conventional transmissometers are between 2° and 3°, total angle,
Laser transmissometers have been proposed (and one is commercially available)
that would have projection angles near 0°; however, the receiver angle would
still have to be several degrees in order to obtain some alignment stability.
Long-term alignment may be a problem even then, as very stringent laser align-
ment must be maintained, particularly for large diameter stacks.
8
-------
PARTICLE REFRACTIVE INDEX = 1.5
LIGHT WAVELENGTH = 0.55 urn
DETECTOR ANGLE OF VIEW = 20°
LIGHT PROJECTION ANGLE = 0°
LOG-NORMAL PARTICLE SIZE
DISTRIBUTION WITH STANDARD
DEVIATION = 3
20 30
TRUE OPACITY, percent
Figure 1. Measured opacities as a function qfjtrue (theoretical) opacities for a transmissometer with
a 20° angle of view and aerosols with different mean particle sizes.
-------
PARTICLE REFRACTIVE INDEX
LIGHT WAVELENGTH = 0.55 urn
LIGHT PROJECTION ANGLE = 0°
LOG-NORMAL PARTICLE SIZE
DISTRIBUTION WITH:
STANDARD DEVIATION = 3
MEAN DIAMETER = 6 urn
20 30
TRUE OPACITY, percent
Fjgure^. Measured opacities as a function of true (theoretical) opacities for transmissometers with
2° and 20°^ angles of view and specified aerosol.
10
-------
SECTION 4
PERFORMANCE SPECIFICATIONS
The performance specifications for transmissometers are concerned with
calibration error (relative to neutral density calibration filters), response
time, and drift stability. These specifications are generally not related to
source effects, but are more related to the quality or the specific monitoring
systems and to the quality of the specific installations. For example, no
transmissometer can have the necessary drift stability if it is improperly
installed and becomes misaligned, or if the purge system selected for the
installation is not adequate to prevent the transmissometer optics from
becoming dirty. Furthermore, a particular monitoring system and installation
procedure that provides adequate drift stability at a particular location at
one plant may not necessarily provide adequate drift stability at another
plant or even at another location at the same plant.
Although the performance specifications are not generally related to
source effects, any unusual characteristics of the source are noted that may
prevent the transmissometer from monitoring the opacity of the emissions
according to the promulgated stability requirements of the performance
specifications. In addition, the capability of the in-stack transmissometer
to monitor the opacity of the plumes and mass concentration of the particu-
lates emitted by the sources are evaluated. Comparative evaluations of in-
stack and plume opacities and particulate concentrations are important for
proper interpretation of the transmissometer data.
IN-STACK AND PLUME OPACITY
Opacity emission standards apply to the opacities of the plumes emitted by
emission sources; consequently, the in-stack and plume opacities must be
equivalent, or at least a consistent relationship established, if the in-stack
11
-------
transmissometer is to be acceptable for monitoring the emission standard.
When the characteristics and/or concentration of the participates in the
stack of an emission source are not the same as in the plume, the in-stack
and plume opacities will most likely differ. These differences can be due to
reactions between particulates and gases, agglomeration of particulates,
formation of particulates by condensation, dilution of the effluent, etc. To
determine whether the in-stack transmissometer can monitor the plume opacities
of a source, concurrent measurements of the in-stack and plume opacities of
the source of interest are compared.
IN-STACK OPACITY AND MASS CONCENTRATION
The main purpose of opacity monitoring is to determine that the particu-
late emission control equipment is functioning properly (13). A functional
relationship between the opacity of the effluent and mass concentration of
particulates in the effluent must exist for changes in the performance of
particulate control equipment to be shown by changes in the opacity of the
particulate emissions. The simplest way to determine the opacity-mass rela-
tionship for an emission is to install a transmissometer in the source to
measure opacity, measure the mass concentration of the source by an acceptable
gravimetric sampling method, make the measurements at several emission levels
and finally plot a curve of opacity versus mass concentration. In practice,
the measurements should be made at a location in the stack (or duct) that is
acceptable for extracting the particulate samples for gravimetric analyses.
The transmissometer is preferably an across-stack (long path) transmissometer
that measures the opacity across the entire stack (or duct) and the particu-
late samples are obtained by traverse sampling along the axis and near the
light beam of the transmissometer. With proper precaution to avoid potential
problems due to stratification of the particulates, it is likely that valid
data can be obtained with portable (short path) transmissometers and point
samples. Several emission levels can usually be obtained at a plant by
varying the efficiency of the particulate emission control equipment.
The relationship between the opacity and mass concentration of a partial--
late emission can be defined by the following equation:
12
-------
0 = 1 - T = 1 - e-a£ = 1 - e-AC£ (Eq. 4)
where 0 = opacity,
T = light transmission,
£ - pathlength of opacity measurement,
a = attenuation of light per unit pathlength,
C = mass concentration, and
A = attenuation per unit path length per unit mass concentration
The attenuation coefficient a and the attenuation coefficient per unit
mass concentration A are both dependent upon the number concentration, and
the physical characteristics of the particles in the emission - size, size
distribution, shape, and refractive index. A (and C) is also dependent
upon the density of the particles. Equation 4 shows that the opacity-mass
concentration relationship of an emission depends on the chemical and physical
characteristics of the particulates, as well as on the pathlength of the
opacity measurement. Furthermore, the opacity and mass concentration are
logarithmically related if the chemical and physical characteristics of the
particulates are constant.
For easy comparison of data, the opacity versus mass concentration data
from the different plants are all shown for an opacity pathlength normalized
to 3 m (10 ft), even though the opacity measurements may have been made at
other pathlengths. In addition to the opacity scale, the attenuation coeffi-
cient (a) scale is shown since it is the optical characteristic of the
emission that can be calculated from the opacity measurement that is expected
to be proportional to the mass concentration of the emission. The slopes (A)
of the attenuation coefficient versus mass concentration curves determine the
consistency or lack of consistency between the transmissometer optical
measurement and mass concentration of the emission sources.
13
-------
SECTION 5
MEASUREMENT METHODS
A number of measurement methods were used to determine the physical
characteristics and opacities of the particulate emissions. The methods and
procedures used at a particular plant varied and will be noted when describing
the tests in Section 6 of the report. A brief description of all the measure-
ment methods will be given here with reference to more detailed descriptions.
OPACITY MEASUREMENT
The in-stack opacities were measured with either portable or "permanently"
installed Lear Siegler transmissometers. For the "permanent" installations
the Model RM4 was used. For the portable measurements either the Model RM41P
with the one meter probe or the RM4 equipped with the RM41P one meter probe
were used. Transmissometer calibrations were checked with neutral density
filters. The locations for the transmissometer measurements were either in
the stack or stack breeching and varied between plants. The locations were
usually dependent upon the availability of sampling ports and platforms. A
detailed description of the RM4 transmissometer has been reported by Beutner
(14). The manufacturer's literature is recommended for a description of the
RM41P (15).
The plume opacities were measured with one or more remote instrumental
methods. The trained observer (EPA reference Method 9) method was also used
at some plants. The instrumental methods include laser radar (lidar) and
contrasting target telephotometry.
The lidar plume opacity measurement is made by beaming a short pulse of
laser light through the plume and measuring the amount of light back-scattered
from the pulse by the atmosphere in back of the plume relative to the amount
back-scattered by the atmosphere in front of the plume. This ratio (R) is a
14
-------
measure of the two-way transmittance of the plume. The opacity of the plume
is (1- /R). The method was first proposed and demonstrated by the Stanford
Research Institute (16). The lidar is shown in Figure 3. Its design and
operational characteristics are shown in Table 2. A complete description of
the lidar system and its operation, are given in reports on its development
(17,18).
The contrasting target telephotometry method is used to measure the
opacity of a plume only when contrasting targets can be viewed through the
plume at the stack exit and when the ambient illumination of the plume and
targets are stable. The contrasting targets may be distant hills, tall
buildings, or towers and the sky adjacent to them. The plume opacity is
determined by using a narrow angle-view (less than 1/2°) telephotometer to
measure the ratio of the luminance difference between targets when viewed
through and beside the plume. Plume opacity measurement by telephotometry of
contrasting targets have been described in detail (19).
The trained observer method of measuring plume opacity is the method
generally used by control agencies; it is also the opacity compliance test
method used by EPA. The observer measurements were made by certified person-
nel using the procedures described in the Federal Register (20).
PARTICULATE MASS MEASUREMENT
The mass concentration of the particulates in the effluents was deter-
mined by sampling the effluents with an EPA Method 5 sampling train, which is
described in detail in the Federal Register (21). Any variations from the
Federal Register procedures will be noted when discussing the plant tests
below.
PARTICLE SIZE MEASUREMENT
The size of the emission particles was measured by size selective samp-
ling using cascade impactors. Two source test impactors were employed during
the study: the Anderson Model III and the Pilat Mark III. The impactors are
similar in size and are both inserted into the stack through standard 10-Cm
(4-in) sampling ports for in-situ sampling of effluent. They are usually
15
-------
en
Figure 3. Photograph of mobile lidar system.
-------
TABLE 2. MOBILE LIDAR SYSTEM CHARACTERISTIC
Component
TRANSMITTER
Characteristic
Laser
Wavelength
Pulse width (FWHH)
Maximum output
Repetition rate
Cooling
Objective lens
Beam divergence
Rotating prism, Q-switched ruby
694.3 nm
< 30 nsec
1.0 joule
3 pulses/ttiin
Deionized water
12.7 cm, f/5
^ 0.5 miliradian full angle
Component
RECEIVER
Characteristic
Objective lens
Field-of-view
Bandpass (FWHH)
Photomultiplier
Off-gating
Response
15.25 cm, f/5
4 miliradian full angle
1.2 nm
IT&T F4084 (modified S-20)
< 60 dB
'^ 100 nsec
operated within a flow rate range of 14 to 28 1/min; both accept an assortment
of straight or "gooseneck" nozzle sizes to permit isokinetic sampling. The
Pilat impactor consists of seven impactor stages (and "back-up" filter) with
size cuts ranging from 0.27 to 23 ym diameter and the impaction surfaces are
1 t^ 2 mill stainless steel shim stock coated with polyethylene glycol to
improve particle adhesion. The Anderson impactor consists of eight impactor
17
-------
stages (and "back-up" filter) with size cuts ranging from 0.52 to 13 urn
diameter. The impaction stages of the Anderson are designed to use glass
fiber filters as the impaction surfaces. In this case, adhesion is due to
penetration of the particles into the filters. A comparative evaluation of
impactors at a coal-fired power plant indicated similar performances for the
Anderson and Pilat impactors (22).
The procedures for collecting size selective samples with the impactor
are very similar to the EPA Method 5 procedures for determining mass loading.
The criteria for determining the preferred sampling locations, number of
traverse points, flow rate, and nozzle size selection for isokinetic sampling"
are the same. In the case of source test impactors, they must be allowed to
reside in the effluent for sufficient time prior to sampling in order to
obtain a temperature equilibrium (approximately 30 minutes for these impactors)
It is also necessary to maintain a constant sample flow rate through the
impactor during a test (to maintain constant impactor size separating charac-
teristics), and not collect too large a sample, overloading the stages.
Additional information on the impactors is given in the manufacturer's litera-
ture (23,24).
MULTIWAVELENGTH OPACITY MEASUREMENT
The opacity of the emissions was measured for different colors of light
within the visible spectrum with a dual wavelength sun photometer. The sun
photometer method of determining plume opacity requires measurement of the
relative intensity of the sun when viewed beside and through the plume with
the photometer. Clearly, this procedure requires the sun not be obstructed
by clouds and the shadow of the plume be in an accessible location. Accurate
opacity measurements also require that the sun be viewed through a defined
cross section of the plume. However, for this study, where only relative
measurements were required, two sun photometers operating at different wave-
lengths were coupled and operated together as a unit to obtain concurrent
opacity-wavelength measurements; consequently, a well-defined plume cross
section was not necessary. Detailed information on dual wavelength sun
photometer instrumentation, and its application to extinction-wavelength
ratio measurements of plumes can be found in Reference 19.
18
-------
SECTION 6
CEMENT PLANT TESTS
TEST SITES
Opacity measurements were conducted at four cement plants. Plants I,
II, and III used the wet-process rotary kiln production method and electro-
static precipitator dust collectors. Plant IV used the dry-process rotary
kiln production method and a baghouse dust collector. In-stack opacity
measurements were made at each plant by a transmissometer. In addition to
the in-stack opacity measurements, concurrent plume opacity measurements were
made at Plants I and II. The plume opacity measurements at Plant I were made
by lidar and trained observer, and by sun photometer at two wavelengths. The
plume opacity measurements at Plant II were made by contrasting target tele-
photometry; the in-stack opacity measurements by "permanent" transmissometer.
Concurrent measurements were also taken of the mass concentration and
size of the particulate emissions at Plants II, III, and IV. The mass con-
centration measurements were made by traverse sampling with the Method 5
train, the size measurements with the cascade impactors. At each plant, the
in-stack monitoring of particulate opacity and the extraction of particulate
samples were at adjacent locations in the stack.
At Plant II, the in-stack measurements were made at the 32-m (105-ft)
level of a 45.7-m (150-ft) high stack. The inside diameter of the stack at
the sampling location was 4.3-m (14 ft). The sampling location was 6.1-m (?0
ft) above the stack inlet and 13.7-m (45 ft) below the stack outlet (see
Figure 4).
To extract particulate samples, a set of four 10-cm (4-in) diameter
^-ts were located at 90° intervals around the stack. To monitor opacity, a
Lear Siegler RM4 transmissometer was located on the stack approximately 30-cm
(1 ft) above and at a 45° angle to the sampling ports.
19
-------
-3.5m
TEST PORTS (4)
.AND TRANSMISSOMETER
I I I i I
PLATFORM
& RAILING
ESP
PARTICULATE
COLLECTOR
Figure 4. Schematic of cement plant test site 11.
20
-------
10.1 m
TEST PORTS (4)
AND TRANSMISSOMETER
21.3m
PLATFORM
& RAILING
ESP
PARTICULATE
COLLECTOR
16.8m
Figure 5. Schematic of cement plant test site III.
21
-------
3 m
3.5 m I
12.8m
i
8.2m
I
TEST PORTS (4)
PLATFORM
& RAILING
BAGHOUSE
PARTICULATE
COLLECTOR
Figure 6. Schematic of cement plant test site IV.
22
-------
At Plant III, the measurements were made at the 38.1-m (125-ft) level of
a 48.2-m (158-ft) high stack. The inside diameter of the stack at the samp-
ling location was 3.7 m (12 ft). The sampling location was approximately
21.3-m (70 ft) above the stack inlet and 10.1-m (33-ft) below the stack
outlet (see Figure 5). To extract particulate samples, a set of four 10-cm
(4-in) diameter ports were located at 90° intervals around the stack. To
monitor opacity, a Lear Siegler RM4 transmissometer was located on the stack
approximately 30-cm (1-ft) above and essentially in line with two opposing
sampling ports.
At Plant IV, the measurements were made at the 21.0-m (69-ft) level of a
24.4-m (80-ft) high stack. The inside diameter of the stack at the sampling
locations was 3.0 m (10 ft). The sampling location was 12.8-m (42-ft) above
the stack inlet and 3.4-m (11-ft) below the stack outlet (see Figure 6). To
extract particulate samples, a set of four 10-cm (4-in) diameter ports were
located at 90° intervals around the stack. To monitor opacity, a Lear Siegler
RM41-P portable transmissometer was inserted in one of the sampling ports.
RESULTS AND DISCUSSION
Design Specifications
The sun photometer measurements of the blue and red light opacity of the
plume emitted by cement Plant I (see Table 3) showed that the opacity of the
plume did not vary as a function of the wavelength of the light in the visible
part of the spectrum. Consequently, the promulgated specification for spectral
response is more than adequate for the transmissometer used to monitor the
opacity of cement plant emissions. The cement plant particle size measure-
ments (see Tables 4, 5, 6, and Figure 7) indicate that the promulgated spec-
tral response specification is more than adequate for cement plants. These
data showed the mass mean particle size of the emissions to be near 2 um
diameter and the distribution of size to be relatively wide. Since the
spectral response specification limits the transmissometer response to wave-
lengths much smaller than 2 ym, a wavelength effect within the specification
response limits is not to be expected.
23
-------
TABLE 3. CEMENT PLANT OPACITY MEASUREMENTS
Measurement
Plant period, min.
I 60
II 40
42
48
47
55
60
42
Percent opacity
By in-stack By remote3
transmissometer Lidar observer
11-17 10±4 5±1
19-23
18-24
7-9
6-9
g_17
g_]2
13-17
Extinction ratio,
Telephotometer blue/red
]
23±4
16±3
3±1
11±1
10±1
16±2
20±2
Measurement ^ 1/2 stack diameter (2-3 m) above stack exit. Data are mean ± 95 percent confidence
interval of measurement set.
-------
The particle size data indicated as well that the coTlimation specifica-
tions for transmissometers are adequate for cement plants. For 2 ym diameter
monodisperse particles, a transmissometer with the maximum allowable five
degree collimation angles would be expected to measure opacity about 2 percent
opacity low at 20 percent opacity; a true (theoretical) opacity of twenty
percent would thus be measured as 18 percent opacity. This difference would
represent at most an upper limit for cement plants. The relatively wide size
distribution of the emissions (3 to 6 geometric standard deviation) would re-
duce the difference to within 1 percent opacity at a theoretical opacity of 20.
Performance Specifications
The in-stack transmissometer performance relative to zero and span drift
stability at the cement plants was good. No significant drift was observed
over monitoring periods up to 1 month. No problem related to the process was
observed that would prevent stable operation and preclude the application of
the promulgated drift specification to cement plants. The observed zero and
span drifts were less than the 2 percent opacity specifications and related
primarily to the stability of the instrument used for the measurement, to the
adequacy of its air purge system, and to the adequacy of the instrument
installation on the stack.
In-Stack and Plume Opacity
The data of in-stack and plume opacity measurements at cement Plants I and
II are shown in Table 3. At Plant I, the data showed that the average of a
set of 13 plume opacity measurements made by lidar was 4 percent opacity
lower than the average in-stack opacity measured by portable transmissometer,
and that the average of a set of 38 plume opacity measurements made by trained
observer was 9 percent opacity lower than the average in-stack opacity meas-
ured by the portable transmissometer. These measurements were made late in
the day (near dusk), during a 30-minute period of relatively steady plant
operation. The low observer readings may have resulted from poor plume
visibility.
Sun photometer opacity measurements, which were made at Plant I during
an earlier 6-minute time period before the in-stack transmissometer was
operating, averaged 14 percent opacity for both blue and red light. This
25
-------
TABLE 4. Portland Cement Plant II
Anderson Impactor Data
Stage
No.
1
2
3
4
5
6
7
8
F
D50 Cut,
ytn
13
8.2
5.4
3.8
2.4
1.25
0.75
0.52
Test 1
21
9
14
9
14
16
10
6
1
Percent
Test
8
7
10
9
10
16
16
11
13
of total samp!
2 Test 3
5
11
4 -
6
8
19
26
9
12
e weight
Test 4
8
8
7
7
5
21
24
12
8
by stages
Test 5
15
3
13
3
4
55
5
4
2
Average
11
8
10
7
8
25
16
8
7
TABLE
5. Portland Cement PI
Anderson Impactor Data
ant III
Stage
No.
1
2
3
4
5
6
7
8
F
050 Cut,
ym
13
8.2
5.3
3.8
2.4
1.25
0.75
0.52
Test 1
2
16
22
22
14
4
9
6
5
Percent
Test
8
6
8
16
12
5
12
13
20
of total sampl
2 Test 3
7
14
20
9
5
7
6
12
20
e weight
Test 4
4
9
7
6
8
5
26
17
18
by stages
Test 5
6
5
9
8
9
9
10
19
28
Average
5
10
13
12
10
7
13
13
17
26
-------
TABLE 6. Portland Cement Plant IV
Pilat Impactor Data
Stage
No.
1
2
3
4
5
6
7
F
D50 Cut,
23
10
4.7
1.9
1.0
0.52
0.27
Percent of total sample weijght by
Test 1
18
7
35
25
11
2
0
2
Test 2
27
15
29
20
7
1
0
1
Test 3
13
7
36
30
12
1
0
1
Test 4
15
7
34
29
10
3
0
2
Test 5
10
8
18
34
20
7
0
3
stages
Test 6
6
5
16
41
22
7
0
3
Average
15
8
28
30
14
3
0
2
27
-------
50
0.1
PLANT NO. II, WET PROCESS, ESP COLLECTORS
B PLANT NO. Ill, WET PROCESS, ESP COLLECTORS
A PLANT NO. IV, DRY PROCESS, BAGHOUSE COLLECTORS
'i 5 10 20 30 40 50 60 70 80
MASS LESS THAW INDICATED SIZE, percent
Figure 7. Particle size distributions of Portland cement plants.
90
95
98
28
-------
result showed that the promulgated spectral response specification for trans-
missometer opacity monitors is adequate, indicated that the size of the
particulate emission is larger than 1 or 2 ym diameter and agrees with the
particle size measurements described above.
At Plant II, the data showed that the averages of sets of plume opacity
measurements (100 to 150 measurements per set) made by the contrasting target
telephotometry method over short periods of time (less than 60 minutes)
during relatively steady plant operation were within ± 6 percent opacity of
the mean of the high and low continuous in-stack opacity measurements obtained
by transmissometer over the same time periods. The differences were random,
indicating no systematic variation between the plume and in-stack opacities
at the cement plants. The magnitude of the difference (± 6 percent) was in
part due to difficulty in making the remote telephotometer measurements. A
telephotometer opacity measurement requires about 4 seconds to execute and
should be made during steady plume conditions of opacity, lighting, and
background. The conditions were often not good for the measurement. With
good conditions agreement better than 2 to 3 percent opacity cannot be expected
On cold days condensation could occur in the plumes of the cement plants,
producing higher plume opacities. The presence of condensation was usually
obvious from the appearance plume. Depending on ambient temperature, the
condensation plume could be attached or detached from the stack. No condensa-
tion was observed in-stack during cold weather.
In-Stack Opacity and Mass Concentration
The comparisons of in-stack light attenuation and mass concentration
measurements at cement Plants II, III, and IV are shown in Figure 8 for the
mass concentration measurements at standard conditions and dry gas. Figure 9
depicts the same data with the mass concentration measurements at actual
stack conditions. Also shown in Figures 8 and 9 are linear regression lines
of data reported by Buhne (25) in 1972 for a dry process cement plant with
electrostatic precipitator emission controls. These data indicated that the
light attenuation coefficient of the particulate emissions at cement plants
is linearly related to the mass concentration of the particulate emissions
for both dry and wet process cement production. For mass concentration data
29
-------
at standard conditions and dry gas (see Figure 8), the slopes of the curves
(attenuation coefficient/mass concentration, m /g/m ) for standard conditions
are all within 0.57 ± 0.10.
When the mass concentration data were for actual conditions (see Figure
9), the slopes of the curves were distinctly different for the wet and dry
-1 3
processes. The slopes are represented by 1.55 ± 0.02 M /g/m averaged for
the wet processes and 0.92 ± 0.08 averaged for the dry processes. The slopes
of all data for actual conditions are within 1.2 ± 0.4.
30
-------
45
40
35
g 30
0)
Q.
I 25
u
a.
O
20
15
10
PLANT NO. II, WET PROCESS, ESP CONTROL
PLANT NO. Ill, DRY PROCESS, BAGHOUSE CONTROL
A PLANT NO. IV. DRY PROCESS, BAGHOUSE CONTROL
0.05 0.10 0.15 0.20 0.25
MASS CONCENTRATION, girT3
0.30
0.20
0.18
0.16
0.14
0.12 3
«
0.10
o
o
o
<
0.08
0.06
0.04
0.02
0.35
Figure 8. In stack opacity of emissions from Portland cement plants as a function of the mass
concentration of the particulates at standard conditions and dry gas.
31
-------
45
40
35
E
CO
Q.
o
30
25
20
15
0.20
« PLANT WO. II, WET PROCESS, ESP CONTROL
PLANT NO. Ill, WET PROCESS, ESP CONTROL
A PLANT NO. IV, DRY PROCESS, BAGHOUSE CONTROL
0.18
0.16
0.14
MASS CONCENTRATION, gm'3
Figure 9. In-stack opacity of emissions from Portland cement plants as a function of the mass
concentration of the particulates at actual conditions.
32
-------
SECTION 7
OIL-FIRED POWER PLANT TESTS
TEST SITES
Opacity measurements were conducted at three oil-fired power plants of
different sizes. They all burned residual fuel oils with an additive to
inhibit corrosion and none had emission control equipment. Plant I was a 70
megawatt unit burning medium sulfur (M.5% by weight) residual fuel oil.
Plant II was a 190 megawatt unit burning medium sulfur (a/l.0% by weight)
residual fuel oil. Plant III was a 520 megawatt unit burning high sulfur
(^2.5% by weight) residual fuel oil. Plants I and II normally operated at
excess oxygen boiler firing levels of 1.5 and 3 percent, respectively.
However, the boiler at Plant III was specially designed to normally operate
at a very low excess oxygen level of 0.2 percent. None of the plants utilized
emission control equipment.
In-stack opacity, plume opacity, particulate mass concentration, and
size measurements were made at each plant. The i.n-stack opacity measurements
were made by transmissometer. The plume opacity measurements were made by
lidar, trained observer, and sun photometer at two wavelengths. The particul-
ate mass concentration and size measurements were made by sampling the
emissions with a Method 5 train and a Pilat Mark III cascade impactor. At
each plant the in-stack monitoring of particulate opacity and the extraction
of particulate samples were at adjacent locations in the stack.
At Plant I the in-stack measurements were made at the 43-m (141-ft)
level of a 45.6-m (150 ft) high stack. The inside diameter of the stack at
the sampling location was 3.3-m (10.8-ft). The sampling location was 16.7-m
(55-ft) above the stack inlet and 2.7-m (9-ft) below the stack outlet (see
Figure 10). To extract particulate samples, a set of three 10-cm (4-in)
33
-------
diameter ports were located at 45° intervals around the stack. A Lear Siegler
RM41-P portable transmissometer was installed in the center port to monitor
opacity. The remaining 90° ports were used for particulate sampling.
At Plant II, the in-stack measurements were made at the 26-m (85-ft)
level of a 61-m (200-ft) high stack. The inside diameter of the stack at the
sampling location was 3-m (10-ft). The sampling location was 5-m (15-ft)
above the stack inlet and 35-m (113ft) below the stack outlet (see Figure
11). To extract particulate samples, two opposing 10-cm (4-in) diameter
ports were located in the stack. A third port was cut in the stack at an
angle of about 20° to one of the sampling ports to install a Lear Siegler
RM41-P portable transmissometer for opacity monitoring.
At Plant III, the in-stack measurements were in the duct breeching of
the stack. The stack was 153-m (500-ft) high with an exit diameter of 5.2-m
(17-ft). The breeching was a long horizontal duct (4.3-m wide by 8.4-m high
by 51-m long). The in duct sampling location was approximately 160-m from
the stack exit, 18-m from the stack inlet and 34-m past the duct inlet (see
Figure 12). To extract particulate samples and install a Lear Siegler RM41-P
opacity monitor, a vertical array of six 10-cm (4-in) diameter ports were
located at 1.4-m intervals on one side of the duct.
RESULTS AND DISCUSSION
Design Specifications
The sun photometer measurements of the blue and red light opacity of the
plumes emitted by the three plants (see Table 7) showed that the opacity was
a function of the wavelength of light in the visible part of the spectrum.
The blue (0.45 ym) to red 0.65 ytn) extinction ratio ranged from 1.9 to 2.8
for the oil-fired power plant emissions.
This difference showed that the extinction of light by the emissions
varies as a function of the wavelength of the light and indicated that the
green light extinction is proportional to X"n with n having a value between
1.7 and 2.8. This observed range in the value of n is likely due to small
differences in the size and composition of the particulates between plants.
34
-------
i
3.3m
^TEST PORTS (3)
T
2.7m
16.7m
/n
PLATFORM
& RAILING
Figure 10. Schematic of oil-fired power plant test site I.
35
-------
3m
TEST PORTS
(3)
~T
35 m
5 m
PLATFORM
& RAILING
Figure 11. Schematic of oil-fired power plant test site 11.
36
-------
5.2m
152m
33.5 ni-
TEST PORTS
(6)
BREECHING
SCAFFOLDING
STACK
Figure 12. Schematic of oil-fired power plant test site III.
37
-------
TABLE 7. OIL-FIRED POWER PLANT OPACITY MEASUREMENTS
CO
co
Percent opacity
Plant
I
II
III
Measurement
period, min.
15
20
35
20
15
Percent
excess 02
1.5
6.0
0.2
0.2
0.2
By in-stack
transmissometer
5-9
2-3
17-21
18-22
18-22
BY
Lidarb
8±3
10±2
23±4
31±4
43±3C
remote3
Observer
7±1
6±1
27±1
30±1
52±2C
Extinction ratio,
blue/red
2.8
2.4
1 .9
2.3C
'
aMeasurements ^ 1/2 stack diameter (2-3 m) above stack exit unless noted. Data are mean ± 95 percent
confidence interval of measurement set.
Lidar measurements corrected for wavelength.
Measurements a, 3 stack diameters (15 m) above stack exit.
-------
For example, a particle size variation of less than 0.2 ym diameter can
produce the observed range in n for these emissions.
For the worst case condition of n = 2.8, calculation of the opacity that
would be measured by transmissometers operating at the promulgated peak
response specification limits of 500 and 600 nm, showed a variation of ± 5
percent opacity relative to a 20 percent opacity measurement (the opacity
standard for new oil-fired power plants).
Except for Plant III, the particle size data obtained by sampling with
impactors (see Table 8 and Figure 13) indicated that most of the particulate
mass in oil-fired power plants emissions is associated with small particles
(less than 0.3 ym diameter). The impactor data from Plant III indicated a
larger mean size with a considerable amount of mass also associated with
large particles around 12 ym diameter. The bimodal size characteristic of
Plant III is shown in Figure 14. The reason for the large particle constitu-
ent at this plant is attributed to unburned carbon. The plant differed from
the others in that it was burning oil with a high sulfur content (^2.5 per-
cent), a high vanadium content (^500 ppm) and operated at a low excess oxygen
level of ^0.2 percent. The fine particle impactor observations support the
observation that the transmittance of the emissions varies with wavelength
(see Table 7).
Although the impactor data at Plant III did show a large particle component,
the multiwavelength measurements at Plant III showed that the fine particle com-
ponent in the effluent dominates its optical characteristics and opacity.
However, the variation in the transmittance of the emission with wavelength
is slightly less than that observed at the other plants and reflects the
presence of the larger particles. The collimation specification limits of 5°
are more than adequate for oil-fired power plants due to the small particle
size of the emissions.
Performance Specification
The in-stack transmissometer performance relative to zero and span drift
stability at oil-fired power plants was good. No significant drift was
observed over monitoring periods up to 1 month, and no problems related to
39
-------
TABLE 8. Oil-Fired Power Plant
Pilat Itnpactor Data
Percent
of total sample wei
on respective
Plant
I
Plant
II
Plant
III
Test
1
2
Average
Test
1
2
3
4
5
Average
Test
1
2
3
4
5
23
2
3
1.5
23
3
1
2
0
0
]
23
7
20
6
9
6
10
28
5
16.5
10
4
4
2
1
2
2
10
38
10
11
25
56
4.7
8
4
6
4.7
2
2
2
2
1
2
4.7
6
6
4
10
12
ght coll
ected
D50 (ym) stages of impactor
1.9
3
5
4
1.9
5
2
3
3
2
3
1.9
13
14
5
13
5
1.0
2
4
3
1.0
4
2
2
2
2
2
1.0
7
5
6
5
2
0.52
5
8
6.5
0.52
4
3
2
2
3
3
0.52
2
3
7
3
1
0.27
5
16
11.5
0.27
19
13
10
7
7
11
0.27
9
9
19
5
1
F
47
55
51
F
59
76
77
83
83
76
F
18
33
42
30
17
Average
10
28
10
28
40
-------
PLANT 1,1.5%SULFUR OIL-FIRED BOILER, EXCESS 02 =1.5%
PLANT II, 1.0%SULFUR OIL-FIRED BOILER, EXCESS 02 = 3.0%
A PLANT III, 2.5% SULFUR OIL-FIRED BOILER, EXCESS 02 0.16 to 0.25%
0.1
20 30 40 50 60 70 80
MASS LESS THAN INDICATED SIZE, percent
Figure 13. Particle size distributions of oil-fired power plants.
95
98
41
-------
ou
,_
2
UJ
LLJ
QC
O
UJ
M
CO
CD 20
O
J
a
UJ
i^
p
<
CJ
8 1R
iS * 3
z.
X
I-
g
CO
CO
< 10
s
u.
o
UJ
u
<
l-
z .
uj b
o
cc
UJ
a.
n
I
1
1
I
I
I
j
|
1
1
1
0.125 0.25 0.5 1.
2.5% SULFUR OIL-FIRED BOILER
BOILER EXCESS 02 = 0.16 TO 0.25%
NO EMISSION CONTROL
PILOT IMPACTOR, 5 RUN AVERAGE
MMD = 3.0 /urn
--ACTUAL UPPER AND LOWER SIZE
RANGE LIMITS ARE INDETERMINATE
1
0 2.0 4.0 8.0 16 32 64
PARTICLE SIZE,
Figure 14. Particle size distribution of oil-fired power plant III.
-------
the process were observed that would prevent stable operation and preclude
applying the promulgated drift specifications to oil-fired power plants for
monitoring the in-stack opacity by transmissometer.
Comparison of In-Stack and Plume Opacity
The comparison of in-stack and plume opacity measurements at the oil-
fired power plants (see Table 7) showed that the plume opacities were higher
than the in-stack opacities. The measurements were close enough only at
Plant I to indicate agreement, but even here, in-stack measurements appeared
slightly lower. At the other two plants, the plume opacity measurements were
clearly higher than in-stack, and at Plant III measurements several meters up
in the plume were higher than in the plume at the stack exit. These data
indicated that much of the particulates in the plumes form as the effluent
cools in the atmosphere out of the stack. The increase in opacity is attri-
buted to condensation of H2$04 (26). The plant~with the most opaque plume
and greatest in-stack to plume opacity difference was Plant III, which was
burning the high sulfur fuel (2.5 percent). In Table 7, it is interesting to
note that the blue/red light extinction measurement of the plume at Plant III
was slightly higher several meters up in the plume than at the stack exits.
This suggests that the increase in opacity is mostly due to an increase in
concentration, not an increase in particulate size.
Comparison of In-Stack Opacity and Mass Concentration
The comparison of in-stack light attenuation and mass concentration
measurements at oil-fired power plants I, II, and III are shown in Figure 15
for the mass concentration measurements at standard conditions and dry gas,
and in Figure 16 for the mass concentration measurements at actual stack
conditions. Also shown in Figures 15 and 16 are data reported by Reisman (9)
in 1974 for an oil-fired power plant.
Two distinct functional relationships are evident from the data. One
relationship concerns the data obtained at Plants I and II and the data
reported by Reisman. These plants all burned low-sulfur oil (less than 1.5
percent by weight) and boiler firings were at normal excess oxygen levels
between 1.5 and 3.0 percent. The data indicated that the correlation is
43
-------
linear. The light attenuation coefficient per mass concentration ratios were
-1 3
0.34 and 0.43 m /g/m with the mass concentration measurements at standard
conditions (dry gas) and actual conditions, respectively.
The second relationship is for data obtained at Plant III for two series
of tests. This plant burned high-sulfur oil (^2.5% by weight), and boiler
firing was at its normal excess oxygen level of 0.2 percent. The light
attenuation coefficient per mass concentration ratio at this plant was 0.11
-1 3
and 0.20 m /g/m with the mass concentration measurements at standard condi-
tions (dry gas) and actual conditions, respectively.
The low excess oxygen conditions for Plant III as compared to the higher
excess oxygen conditions for Plants I and II produced a great portion of
particles (unburned carbon) in the large size fractions. The size distribu-
tions in Figure 13 show a mass median diameter of about 3 ym for Plant III
emissions. The value is about an order of magnitude greater than the mass
median diameter for Plants I and II emissions. Consequently, the mass concen-
tration vs opacity relationships in Figure 15 are consistent with the physical
properties of the emissions. For the same mass concentration, the larger
number of submicron particles in Plants I and II emissions produced greater
light scatter and gave higher opacity readings.
44
-------
\L
11
in
IU
9
^
2 Q
w 0
X
J
t 7
CO
\ 6
a.
o
5
4
3
2
1
n
O 0 O OO-
~~ '' /
1 /
t *
/ '
/ /
1 /
/ /
1 /
I /
I /
§ f
//
, >
/ AA/i/
/ /
/ /
I /
/ /
/t
*
1 /
/ ^
/ /
/ /
/ / PLANT 1
T' / PLANT II
T2 DATA POINT AVERAGE * PLANT "' (m ^
1 / Q PLANT III (6 months later)
/ /' D REISMAN, 1974(9)
i / T
,/ \ t
^dATA POINT AVERAGE ' T?ANGE OF DATE FOR AVERAGES
1 '
0.04
0.03
"^
&
at
E
u
U-
LU
n n? o
U.U& * '
o
Z
O
1
z
LLJ
t
"*
0.01
0
0.05
0.10
MASS CONCENTRATION, gm'3
Figure 15. In-stack opacity of emissions from oil-fired power plants as a function of the mass
concentration of the particulates at standard conditions and dry gas.
45
-------
12
1 n
9
q
T
7
1
n
1 1 I T~
/
/
/
/ /
/ /
- ; /
/
/ /
/ /
h- >
/ /
* /
/
/ /
' /
' /
/ /
/ /
/ /
/ /
/ /
/ /
J - I
i /
HM PLANT 1
Jll DATA POINT AVERAGE .PLANT..
"T ' ' O PLANT III {6 months later)
/ / D REISMAN, 1974<9)
' ' T
/T ATA P°INT AVERAGE H^ANGE OF DATA FOR AVERAGES
/I
n nj
Om
s
0)
3
3
O
EFFICIENT, m
o
z
o
1
LU
t-
<
n m
n
0.05
0.10 0.15
MASS CONCENTRATION, gnV3
0.20
0.25
Figure 16. In-stack opacity of emissions from oil-fired power plants as a function of the mass
concentration of the particulates at actual conditions.
46
-------
REFERENCES
1. Federal Register Vol. 40, No. 194:46529-46262, October 6, 1975.
2. Peterson, C. M., and M. Tomaides. In-Stack Transmittance Techniques for
Measuring Opacities of Particulate Emissions from Stationary Sources.
NTIS PB 212-741, Springfield, Virginia, 1972. 87 pp.
3. Herget, W. F., and W. D. Conner. Instrument Sensing of Stationary Source
Emissions. Environ. Sci. Technol., 11(10):962-967, 1977.
4. Conner, W. D. Comparative Study of Plume Opacity Measurement Methods.
Submitted for publication to Environ. Sci. Technol.
5. Schneider, W. A. Opacity Monitoring of Stack Emissions: A Design Tool
with Promising Results. In: The 1974 Electric Utility-Generation Planbook,
McGraw-Hill, New York, New York, 1974. P. 73-75.
6. Conner, W. D. Measurement of the Opacity and Mass Concentration of Parti-
culate Emissions by Transmissometry. NTIS PB 241-251/AS, Springfield,
Virginia, 1974. 33 pp.
7. Nader, J. S., F. Jaye, and W. Conner. Performance Specifications for
Stationary Source Monitoring Systems for Gases and Visible Emissions.
NTIS PB 230-934, Springfield, Virginia, 1974. 74 pp.
8. Avetta, Edward D. In-Stack Transmissometer Evaluation and Application to
Particulate Opacity Measurement. NTIS PB 243 402/AS, Springfield, Virginia,
1975. 124 pp.
9. Reisman, E., W. D. Gerber, and N. D. Potter. In-Stack Transmissometer
Measurement of Particulate Opacity and Mass Concentration. NTIS PB 239-
264/AS, Springfield, Virginia, 1974. 107 pp.
10. Hodkinson, J. R. The Optical Measurement of Aerosols. In: Aerosol Sci.,
C. N. Davies, ed. Academic Press, New York, 1966. P. 287-357.
11. Franz, I., and J. Kraus. The Influence of Forward Scattering on Measure-
ments of the Degree of Transmission of Aerosols. Staub-Reinhalt Luft (in
English), 33(9):341-345, 1973.
12. Ensor, D. S., and M. J. Pilat. The Effect of Particle Size Distribution
on Light Transmittance Measurement. Am. Ind. Hyg. Assoc. J., 32(5):287-
292, 1971.
47
-------
13. Federal Register Vol. 39, No. 47:9308-9309, March 8, 1974.
14. Beutner, H. P. Measurement of Opacity and Participate Emissions with an
On-Stack Transmissometer. J. Air Poll lit. Control Assoc., 24(9):865-871,
1974.
15. Lear Siegler, Inc., Englewood, Colorado.
16. Evens, W. E. Development of Lidar Stack Effluent Opacity Measuring System
NTIS PB 223-135/AS, Springfield, Virginia, 1967. 96 pp.
17. Cook, C. S., G. W. Bethke, and W. D. Conner". Remote Measurement of Smoke
Plume Transmittance Using Lidar. Appl. Opt. 11(8):1742-1748, 1972.
18. Bethke, G. W. Development of Range Squared and Off-Gaiting Modifications
for a Lidar System. NTIS PB 228-715, Springfield, Virginia, 1973. 47 pp.
19. Conner, W. D., and J. R. Hodkinson. Optical Properties and Visual Effects
of Smoke-Stack Plumes. U.S. Public Health Service 99-AP-30. NTIS PB 174-
175, Springfield, Virginia, 1967. 89 pp.
20. Federal Register Vol. 39, No. 219:39874-39875, November 12, 1974.
21. Federal Register Vol. 36, No. 247:2488-24890, December 23, 1971.
22. McCain, J. D., K. M. Gushing, and A. N. Bird, Jr. Field Measurements of
Particle Size Distribution with Inertial Sizing Devices. NTIS PB 226 292,
Springfield, Virginia, 1973. 52 pp.
23. Pilat Impactor, Pollution Control Systems Corp., Seattle, Washington.
24. Andersen Impactor, Andersen Samplers, Inc., Atlanta, Georgia.
25. Buhne, K. W. and L. Duwel. Recording Dust Emission Measurements in the
Cement Industry with the RM 4 Smoke Density Meter Made by Messrs. Sick
and Staub, Staub-Reinhalt Luft (in English), 32(8):19-26, 1972.
26. Nader, J. S., and W. D. Conner. Impact of Sulfuric Acid Emissions on
Plume Opacity. Workshop Proceedings on Primary Sulfate Emissions from
Combustion Sources. Vol. 2, Characterization. EPA-600/9-78-020b.
P. 121, August 1978.
48
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/2-79-188
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
5. REPORT DATE
APPLICABILITY OF TRANSMISSOMETERS TO OPACITY MEASUREMENT
OF EMISSIONS
Oil-Fired Power and Portland Cement Plants
September 1979
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
W. D. Conner, K. T. Knapp, and J. S. Nader
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Same as Box 12
10. PROGRAM ELEMENT NO.
1AD605 DA-05 (FY-77)
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Science Research Laboratory - RTP, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N. C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 4/76-10/77
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT
In-stack transmissometers were evaluated for their capability to monitor the opacity of
smoke-stack plumes emitted by portland cement plants and oil-fired power plants. Tests
were conducted to determine their performance in four areas: the adequacy of the U.S.
Environmental Protection Agency-promulgated transmissometer design and performance spe-
cifications for the sources, the correlation between the opacity of the emissions
measured in the plume and in the stack of the sources, as well as the existence of a
functional relationship between the transmissometer-measured opacity and mass concen-
tration of the particulate emission.
The results indicated that the promulgated design and performance specifications for
transmissometers are adequate for both sources except for the spectral response design
specification; the allowable peak spectral response range may be too large for oil-fired
power plants. For opacity monitoring of submicrometer particulate emissions like those
from oil-fired power plants, the allowable peak spectral response range of the trans-
missometer should be reduced. In addition, the in-stack transmissometer-measured opa-
city for oil-fired power plants was generally lower than the plume opacity, indicating
that much of the particulates in the plumes were forming in the atmosphere out of the
stack. Correlations were observed between the transmissometer-measured opacity and mass
concentration of the particulate emissions from both sources.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
* Air pollution
* Transmissometers
Evaluation
Plumes
Opacity
Emission
Oil-fired power plants
Portland cement plants
13B
14B
21B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
!1 . NO. OF PAGES
57
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDI TION i s o BSOLET E
49
------- |