United States
             Environmental Protection
             Agency
             Environmental Sciences Research
             Laboratory
             Research Triangle Park NC 27711
EPA 600 2-80-001
January 1980
             Research and Development
&EFK
Comparative
Study  of Plume
Opacity
Measurement
Methods

-------
                  RESEARCH REPORTING SERIES


 Research reports of the Office of Research and Development, U.S. Environmental
 Protection Agency, have been grouped into nine series. These nine broad cate-
 gories were established to facilitate further development and application of en-
 vironmental technology. Elimination of traditional  grouping was consciously
 planned to foster technology transfer and a maximum interface in related fields.
 The nine series are:

     1. Environmental Health Effects Research

     2. Environmental Protection Technology

     3. Ecological Research

     4. Environmental Monitoring

     5. Socioeconomic Environmental Studies

     6. Scientific and Technical Assessment Reports (STAR)

     7. Interagency Energy-Environment Research and Development

     8. "Special" Reports

     9. Miscellaneous Reports

 This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
 NOLOGY series. This series describes research performed to develop and dem-
 onstrate instrumentation, equipment, 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 required 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-600/2-80-001
                                                   January  1980
COMPARATIVE STUDY OF PLUME OPACITY MEASUREMENT METHODS
                          by

          William D. Conner and Norman White
  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 publica-
tion.  Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
                                     n

-------
                                  ABSTRACT

     The opacities of smoke-stack emissions were measured by three methods at
thirteen different plants and the results compared.  The three opacity meas-
urement methods were trained observer, in-stack transmissometer, and laser
radar (lidar).  The sources consisted of five coal-fired power plants, four
oil-fired power plants, a Portland cement plant, a paper mill  kraft recovery
furnace, a phosphate fertilizer plant rock dryer, and a small  oil-fired
boiler at a chemical plant.  The instrumental methods of lidar and in-stack
transmissometer correlated better with each other than with the observer
method and were also more precise than the observer method.  The observer
measurements were generally lower than the instrumental measurements.   This
is evidently due to a variable negative bias and/or low sensitivities  asso-
ciated with the observer method when evaluating plumes under viewing condi-
tions where plume visibility is less than desired for the method.   The data
show that between 10 and 40% opacity (range of opacity emission standards),
the observer readings averaged 6 and 11% opacity less than the instrumental
readings of lidar and transmissometer respectively, and varied by as much as
25% opacity below to 8% opacity above the instrumental readings.
                                     iii

-------
                                  CONTENTS

Abstract 	iii
Figures	vi
Tables	vi
Acknowledgments	vii

     1.   Introduction 	  1
     2.   Opacity Measurement Methods  	  2
               Trained observer  	  2
               In-stack transmissometers 	  2
               Lidar	:	3
     3.   Source and Test Descriptions	10
               Coal-fired power plants 	 12
               Oil-fired power plants  	 13
               Oil-fired boiler 0	15
                                 t*
               Portland cement plant P 	 16
               Paper mill K	16
               Phosphate plant R	16
     4.   Results	18
               Lidar and transmissometer comparison  	 18
               Observer and lidar comparison 	 18
               Observer and transmissometer comparison 	 22
     5.   Discussion	23

References 	 26

-------
                                     FIGURES

Number                                                                    Page

  1    Mobile lidar system 	  5
  2    Comparison of transmissometer and lidar measured opacities
         of screen targets	7
  3    Plume opacity measurements by lidar versus plume opacity
         measurements by in-stack transmissometers 	19
  4    Plume opacity measurements by trained observers versus
         plume opacity measurements by lidar 	 20
  5    Plume opacity measurements by trained observers versus
         plume opacity measurements by in-stack transmissometers 	 21
  6    Linear regression relationships between plume opacity
         measurements by trained observers and in-stack trans-
         missometer at a coal-fired power plant for different
         environmental conditions  	 24
                                     TABLES
Number                                                                    Page

  1    Mobile Lidar System Characteristics 	  6
  2    Concurrent Opacity Measurements by Different Methods at
         Various Plants  	 11
                                     VI

-------
                               ACKNOWLEDGMENTS

     We wish to express our appreciation to the personnel  at EPA's Region IV
and National Enforcement Investigation Center (NEIC) for assistance in the
selection of test sites, and to the plant personnel  at the various test sites
for assistance with the tests.  We also wish to express our appreciation to
Evelyn Adams, Dr. Edward Mangold, and Pat Thompson of NEIC, to Wayne Aronson
of Region IV, and to Bruce McElhoe of Northrop Services for their assistance
with various parts of the study.
                                    vn

-------
                                    SECTION 1
                                  INTRODUCTION

     Participate emissions from stationary sources must generally meet opacity
as well as mass standards.  The opacity standards are established primarily
because they are much easier to measure than the mass standards and can be
monitored by control agencies and plant operators to determine whether the
emission controls required to meet the mass standard are operating properly (1),
For this application of dual standards, the opacity standards established by
the U. S. Environmental Protection Agency (EPA) are set less stringent than
the mass standards so that a violation of the opacity standard will be a
clear indication of a decline in the efficiency of the control equipment and
a violation of the mass standards (2).

     The opacity of an emission is defined in the Federal Register as a
measure of the degree to which the emission reduces the transmission of light
and obscures the view of an object in the background (3).  Consequently, the
percent opacity of a plume is determined directly by measuring its percent
transmittance and subtracting from 100%.  This paper describes a comparative
study of three methods of measuring the opacity of smoke-stack plumes:
trained observer, in-stack transmissometer, and laser radar (lidar).  The
comparisons are of concurrently obtained measurements that were made under
field conditions at a variety of emission sources.

-------
                                    SECTION 2
                           OPACITY MEASUREMENT METHODS
TRAINED OBSERVER
     The  trained observer is the method generally used by control agencies
 for  evaluating  the opacity of source emissions, and it is the opacity compli-
 ance test method used by EPA (2).  The observers are trained at smoke inspec-
 tor  training  schools to evaluate opacities of training plumes with prescribed
 accuracies  relative to transmissometer measurement of their opacities.  Upon
 passing the course, they become certified by the school as capable of evalu-
 ating  the opacities of smoke plumes by visual inspection.  When inspecting a
 plume  the method requires that the observer stand (a) at a distance from the
 plume  sufficient to provide a clear view of the emissions, (b) with his line
 of vision approximately perpendicular to the plume direction, and (c) with
 the  sun oriented in the quadrant to his back.  The method also requires that
 readings  be made by observing the plume momentarily at 15 sec intervals over
 a  6  min period  and at the point of greatest opacity.  The average of these 24
 readings  is the observer opacity evaluation of the plume.  All of the viewing
 constraints were used for this study except that the readings were always
 made on the plume approximately 0.5 stack diameter above the stack exit;
 alsa,  observer  evaluations are not averages of 24 readings.  They are usually
 averages  of more than 24 readings taken over periods longer than 6 min which
 coincided with  or overlapped the lidar measurement periods.  Four different
 observers were  used during the study.

 IN-STACK TRANSMISSOMETERS

     In-stack transmissometers are often used to monitor the opacity of stack
effluents; their installation is required on some sources by local control
agencies and EPA.   Design and performance specifications for such

-------
transmissometers have been promulgated by EPA (4).  The in-stack opacity data
for sampling ports when plant transmissometers were not available.  The plant
transmissometers were all Lear Siegler* model RM4 instruments.  The portable
transmissometer was a Lear Siegler model RM41P.  All transmissometer opacity
data are transmissometer was a Lear Siegler model RM41P.  All transmissometer
opacity data are for pathlengths equal to the stack exit diameter of the
source being measured.  When the transmissometer pathlength was different
than the stack exit diameter, their opacity measurement was adjusted to
represent the opacity of the effluent for a pathlength equal the stack exit
diameter as required for monitoring opacity by transmissometer (4).

     The RM4 transmissometer was designed and developed in Germany by the
Irwin Sick Company to meet the requirements of the German pollution regula-
tory agency for in-stack particulate monitoring.  With modification of the
spectral response to meet EPA specifications, the instrument was marketed in
the United States by Lear Siegler, Inc. for in-stack opacity monitoring.   A
detailed description of the RM4 has been published by Beutner (5).  The RM4
has subsequently been replaced by a new model RM41 transmissometer.

     The RM41P portable transmissometer is basically the RM41 equipped with a
probe that is inserted into a standard 10.16 cm (4 inch) stack sampling port
for the opacity measurement.  The probe is a stainless steel tube with a
diameter of 7.6 cm and an overall length of 1.5 m through which light is
projected to a small (1.3 cm) retroreflector at the end where it is reflected
back to the RM41 transceiver.  A slot (1 m by 5.1 cm) in the tube allows  the
effluent to flow through the beam when inserted into the stack.  The reflector
and transceiver are protected from the effluent by a clean air purge.   For
detailed descriptions of the RM41 and RM41P transmissometers, the manufac-
turer's literature is recommended (6).

LIDAR

     The lidar evaluation of plume opacity is made by beaming a short pulse
of laser light through the plume and measuring the amount of light back-
 Mention of trade names or commercial products does not constitute endorsement
or recommendation for use by the U.S. Environmental Protection Agency.

-------
scattered from the pulse by the atmosphere in back of the plume relative to
the amount scattered by the atmosphere in front of the plume.  This ratio is
                                         2
a measure of the two way transmittance (T ) of the laser light through the
plume.  The opacity of the plume is (1 - T).  The method was first proposed
and demonstrated by the Stanford Research Institute (7).  The van mounted
lidar used for this study was designed and developed for EPA by the General
Electric Company specifically for plume opacity measurements.  The lidar is
shown in Figure 1.  Its design and operational characteristics are shown in
Table 1.  For a complete description of the lidar system and its operation,
the reports on its development are recommended (8,9).

     The accuracy of the lidar has been evaluated at various times by making
remote measurements on a series of neutral density screen targets of known
opacities(9).  The screen targets are 1 m in diameter, and their opacities
are known from laboratory calibration measurements with conventional trans-
missometers.  The laboratory transmissometers met the EPA design requirements
for monitoring opacity (4) and were checked for accuracy with neutral density
optical filters.  For the remote measurements the targets are usually located
approximately 200 m from the lidar.  Figure 2 shows a series of measurements
made on targets of 24-, 32-, 43-, and 52-% opacity just prior to the study.
These measurements and the earlier tests indicate that the accuracy of the
lidar is within 3% opacity for opacities under 50%.

     Since the lidar uses a ruby (red light) laser and the opacity standards
are for green light obscuration, it is necessary to correct the red light
lidar measurements to values that would be obtained for green light when the
measurements are made on emissions with predominantly submicrometer particle
sizes.  The correction is not needed if the emissions are composed of parti-
cles with mean sizes several micrometers or greater in size; however, for
very small submicrometer size particulate emissions, the red light opacity
measurement may be as much as 50% lower.  No wavelength correction was
required for the other methods since the training of the visible emission
observer and the spectral response requirements for the in-stack transmisso-
meters are for photopic or green light.

-------
                                        •
Figure 1.    Mobile lidar  system.

-------
        TABLE 1.  MOBILE LIDAR SYSTEM CHARACTERISTICS
Component
              Characteristic
Transmitter
  Laser
  Wavelength
  Pulse Width (FWHH)
  Maximum Output
  Repetition Rate
  Cooling
  Objective lens
  Beam Divergence
Receiver
  Objective Lens
  Field-of-view
  Bandpass (FWHH)  .
  Photomultiplier
  Off-gating
  Response
Rotating prism, Q-switched ruby
           694.3 nm
            <30 ns
             1.0 j
     3 pulses per minute
       Deionized wager
        12.7 cm, f/5
    ^0.5 mrad full angle

       15.25 cm, f/b
     4 mrad full angle
          1.2 nm
IT&T F4084 (modified S-20)
          >60 dB
          ^100 ns

-------
   60
   50
I 40

£

2
o
a 30
C/5
K
<
20
   10
          (  ) NUMBER OF MEASUREMENTS

     _  I
          MEAN VALUE OF MEASUREMENTS
          AND 95% CONFIDENCE INTERVAL
              10        20        30        40        50

                   TRANSMISSOMETER MEASURED OPACITY, percent
                                                            60
            Figure 2. Comparison of transmissometer and
            lidar measured opacities of screen targets.

-------
     The opacity-wavelength characteristics of the submicrometer emissions
were measured with dual -wavelength sun photometers.  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 that the sun not be obstructed
by clouds and that the shadow of the plume be in an accessible location.  For
accurate opacity measurements, it also requires 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
wavelengths were coupled and operated together as a unit to obtain concurrent
opacity-wavelength measurements; consequently, a well-defined plume cross
section was not necessary.

     To correct the lidar measurements, it is first necessary to calculate
the extinction coefficient ratio Q(green)/Q(red) from green and red light sun
photometer opacity measurements.  It follows from Bouguer's law (10), that
the Q(green)/Q(red) ratio is equal to the ratio of log (1-green light opacity)
to log  (1-red light opacity).  The corrected lidar opacity (Og) is then
calculated from the measured lidar opacity (0R) and the Q( green )/Q( red) ratio
with the equation:
                           = i-
                             i
                                        Q ( green )/Q( red)
Additional information on dual wavelength sun photometer instrumentation, and
its application to extinction-wavelength ratio measurements of plumes can be
found  in reference (10).

     In this study, only the lidar measurements on the oil-fired power plants
and boiler emissions required wavelength corrections.  Sun photometer measure
ments were made at all of the oil-fired power plants.  The Q(green)/Q(red)
ratio of the emissions varied from 1.5 at two of the plants to 2.0 at the
other two plants.  At the oil-fired boiler, sun photometer measurements were
not obtained due to overcast sky conditions and a Q ( green )/Q( red) average
value estimate of 1.75 from the oil-fired power plant measurements was used
to correct the lidar measurements.  Sun photometer measurements at the other
                                      8

-------
sources or particle size data on similar sources showed that the particle
sizes were large and wavelength correction of the lidar measurements  was not
necessary.

-------
                                    SECTION 3
                          SOURCE AND TEST DESCRIPTIONS

     Plume opacity measurements were made at thirteen different plants:  five
coal-fired power plants, four oil-fired power plants, a portland cement
plant, a kraft recovery furnace at a paper mill, a rock dryer at a phosphate
fertilizer plant, and a small oil-fired boiler at a chemical plant.  The
plants were generally located in the Southeastern United States.  Plant
selection was based primarily on the availability of in-stack monitoring data
from plant-owned transmissometers or on the availability of sampling ports
for installation of the portable transmissometer to obtain the in-stack
opacity data.  At seven of the plants, concurrent opacity measurements were
obtained with all three methods.  Opacity measurements were obtained with the
observer method at all plants.  However, no opacity measurements were obtained
with the in-stack transmissometer method at four of the plants; no opacity
measurements were obtained with the lidar method at one of the plants;
and no concurrent opacity measurements with the observer and lidar methods
were obtained at one of the plants.  At four of the coal-fired power plants
and at the phosphate plant rock dryer, measurements were made at two different
opacity levels.  The higher opacity levels were obtained by turning off
sections of the electrostatic particulate emission control equipment.  Two
sets' of measurements were made on consecutive days at one of the oil-fired
power plants.  Table 2 is a list of each concurrent opacity measurement set,
the type of plant studied, and the types of opacity measurements made for
each measurement set.
                                     10

-------
                                    CONCURRENT OPACITY MEASUREMENTS BY DIFFERENT METHODS AT VARIOUS PLANTS
Symbol
Cl
Cl
c2
C3
C3
C4
C4
C5
C5

°1
°2
°3
°4
°c
P
K
R
R
R
Plant
Type Time3, m
Coal -fired power plant No. 1
Coal -fired power plant No. 1
Coal -fired power plant No. 2
Coal -fired power plant No. 3
Coal-fired power plant No. 3
Coal -fired power plant No. 4
Coal -fired power plant No. 4
Coal-fired power plant No. 5
Coal -fired power plant No. 5
Oil-fired power plant No. 1
Oil-fired power plant No. 1
Oil-fired power plant No. 2
Oil-fired power plant No. 3
Oil-fired power plant No. 4
Oil-fired boiler (chemical plant)
Portland cement plant
Paper mill (kraft recovery furnace)
Phosphate plant (rock dryer)
Phosphate plant (rock dryer)
Phosphate plant (rock dryer)
16
12
18
11
17
30
20
—
—
9
13
23
17
13
17
10
18
5
--
—
Lidar
Nb Opacity0, 1
10
14
15
15
25
19
22
..
—
11
10
29
16
16
9
13
11
6
—
--
9+4
11+5
53+2
9±2
22+1
28+3
58±3
—
—
23±4
31 ±4
10±4
8+3
10+2
11±4
10±4
5±2
46±2
--
—
Transmissometer
', Typed Opacity6, 5
—
—
' —
S
S
S
S
S
S
P
P
—
P
P
—
S
P
P
P
P
—
—
—
5±1
18±3
36+2
58±7
13±1
25±2
199±2
209+2
__
7±2
3±1
—
14±3
4±0
34±5
28±4
57±5
Observer
£ Ob.f Time3, m Nb
A/B
A/B
A/B
A/B
A/B
C
C
D
D
D •
0
C
C
D
A
C/D
C
—
C/D
C/D
16/16
12/12
18/18
11/11
17/17
15
20
30
30/30
34
22
28
23
21
17 '
31/10
22
—
17/17
18/16
14/14
14/14
14/14
15/15
25/25
60
84
121
121/121
136
88
110
90
85
10
124/38
88
-_
68/68
72/64
Opacity0, %
0/0
0/0
35+4/33+2h
0/0
0/0
30+2
82±3
6+0
h
25±l/33±ln
27±1
30±1
17±1
7±1
6±1
0
7±l/5+lh
4±1
__
10+l/9±lh
32±l/33±2h
 Test period
 Number of opacity measurements during test period
cMean ± 95% confidence interval of opacity measurements
 S = Lear Siegler RM4 across-stack transmissometer
 P = Lear Siegler RM41P portable transmissometer
 Mean opacity ± opacity range (estimated by visual inspection of  chart)
 Specific observers are designated A, B, C, and D
^Transmissometer data not used (see text)
H
 Average of the two evaluations used for data analyses

-------
COAL-FIRED POWER PLANTS

     The generating capacities of units tested at the five coal-fired power
plants ranged from 150 to 680 megawatts.  'They all burned pulverized coal
from southern Appalachia, northwestern Kentucky, or Ohio, and they all had
electrostatic particulate emission control equipment.

Coal-fired Power Plant C,

     At this power plant, emissions were measured from two 150-MW units.  The
emissions from each unit were emitted through separate 61 m high by 3.7 m
exit diameter stacks.  No in-stack transmissometer data were obtained at the
plant.  Lidar and observer plume opacity measurements of the emissions were
made from a common location 300 and 360 m from the respective plumes.  Atmos-
pheric conditions for the lidar measurements were fair.  Conditions for the
observer measurements were poor.  The sky was overcast and the plumes were
not visible.

Coal-fired Power Plant C0
     At this power plant, emissions were measured from a 680-MW unit.  The
 emissions were emitted through an 85 m high by 4.6 m exit diameter stack.  No
 in-stack tramsmissometer data were obtained at the plant.  Lidar and observer
 plume opacity measurements were made from a common location 350 m northeast
 of the plume.  Atmospheric conditions for the lidar measurements were good.
 Conditions for the observer measurements were poor.  The plume background was
 a bright hazy sky which reduced plume visibility and made the observer
 measurements difficult.

 Coal -fired Power Plant C
     At this power plant, emissions were measured from a 150-MW unit.  The
emissions were emitted through a 25 m high by 2.7 m diameter stack on the
roof of the plant.  A velocity cone at the top of the stack abruptly reduced
the stack to an exit diameter of 1.9 m.  The top of the stack was 66 m above
ground.  An LSI RM4 transmissometer was located in the stack approximately 2
m above the plant roof for monitoring opacity.  Lidar and observer plume

                                    12

-------
opacity measurements of emissions were made from a common  location 340 m west
of the .plume.  Measurements were made at  two emission  levels.  Atmospheric
conditions for the lidar measurements were very good.  Conditions for the
observer measurements were poor.  The plume background was a bright hazy sky
which made the plumes not visible to the  observer.

Coal -fired Power Plant C
     At this power plant, emissions were measured from a 154-MW unit.  The
emissions were emitted  through  a  91 m  high  by 4.4 m exit diameter stack.  An
LSI RM4 was located  in  the  stack  at approximately the 46 m level.  Lidar and
observer plume opacity  measurements were made from different positions 300 m
north and 200 m  southwest of  the  plume, respectively.  Measurements were made
at two emission  levels.  Atmospheric conditions for the lidar measurements
were good.  Conditions  for  the  observer measurements were fair.  The observer
was viewing the  plume against a partly cloudy background.

Coal -fired Power Plant  C
      At this power plant,  emissions  were measured from a 300-MW unit.  The
 emissions were emitted through a  178 m high  by 4.7 m exit diameter stack.  An
 LSI  RM4 transmissometer was located  in the stack at the 54 m level.  No lidar
 data were obtained at this plant  because the lidar equipment was not avail-
 able at the time.   Observer plume opacity measurements were made from a
 location 340 m east of the plume.  Measurement conditions for the observer
 were fair.  The background varied from partly cloudy blue sky at the low
 emission level to high scattered  clouds and  light haze at the high emission
 level .

 OIL-FIRED POWER PLANTS

      The generating capacities of the units  tested at the four oil-fired
 power plants ranged from 70 to 520 MW. They all burned residual fuel oil
 with an additive to inhibit corrosion. None had emission control equipment.
                                     13

-------
Oil-fired Power Plant 0-j

     At this power plant, emissions were measured from a large 520-MW capac-
ity boiler burning high sulfur ('v 2.5% by weight) residual fuel oil. _ The
emissions were emitted through a 153 m high by 5.2 m exit diameter stack
that was breeched with a long horizontal duct (4.3 m wide by 8.4 m high by 51
m long).  The LSI RM41P portable transmissometer was installed near the
middle of the duct 160 m from the stack exit.  Lidar and observer plume
opacity measurements were made from a common location 375 m south of the
plume.  Measurements were made on two consecutive days near noon with the
plant operating at normal full load conditions.  Atmospheric conditions for
the lidar measurements were good both days.  Conditions for the observer
measurements were fair to good.  The sky background varied from hazy with
broken clouds the first day to blue with broken clouds the second day.

     The in-stack opacity data obtained at this plant clearly did not cor-
relate with the plume opacity measurements and were not used.  The measure-
ments showed that the opacity of the plume was much higher than in-stack,
indicating that the plume opacity is largely due to an interaction between
the effluent and the atmosphere and cannot be monitored by a transmissometer
in the stack.  This condition was apparently due to the high sulfur content
of the fuel oil burned at the plant, which resulted in the condensation and
hydration of sulfuric acid.  More extensive measurements at this plant made
over a range of plant operating conditions support this observation and also
show an increase in opacity of the plume with distance from the stack exit
(11).  The condition was not observed at any of the other plants.  The lidar
and observer plume opacity measurement data are used for this correlation
study since they were made on the plume at the same point.

Oil-fired Power Plant 00
     At this power plant, emissions were measured from a 380-MW unit that was
burning a mixture of 20% gas and 80% medium sulfur  (^ 1.5% by weight) resi-
dual fuel oil.  The emissions were emitted through  a 106 m high by 4.6 m exit
diameter stack.  No in-stack transmissometer opacity data were obtained at
this plant.  Lidar and observer plume opacity measurements were made from a

                                     14

-------
common location 315 m southwest of the plume.  Atmospheric conditions for the
lidar measurements were good.  Conditions for  the observer measurements were
good.  The plume was viewed against a clear  blue sky.

Oil-fired Power Plant 0
     At this power  plant, emissions were measured from a small 70-MW unit
that was burning a  medium sulfur  (^ 1.5% by weight) residual fuel oil.  The
emissions were  emitted  to the  atmosphere through a 45.6 m high by 3.3 m exit
diameter stack.  The  in-stack  opacity  was measured with the LSI RM41P port-
able transmi ssometer  that was  installed in the  stack  3 m from the stack exit.
Lidar  plume  opacity measurements  were  made from a location 320 m east of the
plume  and observer  plume opacity  measurements were made from a location 152 m
southwest of the plume.  Atmospheric conditions for the lidar measurements
were fair.   Conditions  for  the observer measurements  were good.  The observer
evaluation of  the  plume was made  against a blue sky background.

Oil-fired  Power Plant 0
      At this power plant,  emissions were measured  from  a  190-MW unit burning
 a medium sulfur (^ 1.1% by weight)  residual  fuel oil.   The  emissions were
 emitted through a 60 m high by 3 m  exit diameter stack.   The  in-stack opacity
 was measured with the LSI  RM41P portable transmissometer, that was installed
 in the stack approximately 30 m from the stack exit.  Lidar and observer
 plume opacity measurements were made from a  common location 260 m southwest
 of the plume.  Atmospheric conditions for the lidar measurements were poor.
 Conditions for the observer measurements were good.   Observer evaluations
 were made against a blue sky background.

 OIL-FIRED BOILER Or
                   c
      This source was a small oil-fired industrial  boiler  at a chemical  plant.
 The emissions were emitted through  a 53 m high by  3.7 m exit  diameter stack.
 No in-stack transmissometer data were obtained at  the plant.  Lidar and
 observer measurements were made from a common location  280  m  from the plume.

-------
Atmospheric conditions for the lidar measurements were fair.  Conditions for
the observer measurements were poor.  The plume was viewed against an over-
cast sky, and was not visible except for occasional puffs that were not
measured.

PORTLAND CEMENT PLANT P

     At this source, emissions were measured from a wet process rotary cement
kiln.  The emissions were emitted through a 76 m high by 4.6 m exit diameter
stack.  Particulate emissions were controlled with electrostatic precipitators,
The LSI RM41P portable transmissometer was installed in the stack breeching
15 m from the bottom of the stack for in-stack opacity measurement.  Lidar
and observer plume opacity measurements were made from different positions
240 m southeast of the plume and 150 m west of the plume, respectively.
Atmospheric conditions for the lidar measurements were good.  Conditions for
the observer measurements were good.  Conditions for the observer measure-
ments were fair.  The observer evaluations were made while viewing the plume
against a clear blue sky; however, the data were taken late in the day near
sunset, which may have made the plume difficult to read.

PAPER MILL K

     At this source, emissions were measured from the kraft recovery furnace
at a paper mill.  The emissions were emitted through a 93 m high by 2.8 m
exit diameter stack.  Particulate emissions were controlled with electro-
static precipitators.  An LSI RM4 transmissometer was located in the stack
approximately 5 m from the stack exit.  Lidar and observer plume opacity
measurements were made from a common location 700 m southeast of the plume.
Conditions for the lidar and observer measurements were good.  Observer
evaluation of the plume was made against a clear blue sky background.

PHOSPHATE PLANT R

     At this source, emissions were measured from a phosphate fertilizer
plant (rock dryer).   The emissions were emitted through a 25 m high by 2 m
exit diameter stack.  Particulate emissions were controlled with
                                      16

-------
electrostatic precipitators.  The LSI RM41P portable transmissometer was
installed in the stack approximately 2 m from the stack exit for in-stack
opacity measurements.  Lidar and observer plume opacity measurements were
made from different positions 246 m southeast and 120 m west of the plume,
respectively.  Atmospheric conditions for the lidar measurements went from
fair to impossible.  After obtaining one series of lidar measurements, a
conveyor upwind of the stack was put into operation and interference from
particulate emissions from the conveyor contaminated the atmosphere around
the stack and made its scatter too irregular and variable to obtain addi-
tional lidar measurements.  The observer plume opacity measurements were not
made concurrent with the  lidar measurements but were made at two emission
levels.  Conditions for the observer measurements were good.  The plume was
evaluated against  a clear blue sky.
                                    .  17

-------
                                    SECTION 4
                                     RESULTS

     The comparisons between the three opacity measurement methods are shown
 in Figures 3, 4, and 5.  The correlation coefficients, linear regression lines
 and their 95% confidence intervals are calculated and shown in the figures.
 The comparisons are of concurrently obtained measurements on relatively
 steady state emissions.  Whenever the opacity of a plume was evaluated by
 more than one observer during a test, the evaluations were averaged to give a
 single data point.  To prevent measurements from one plant overly affecting a
 comparison, no more than two emission levels were studied or two data points
 generated from any one plant.  In analyzing the data, the results for opaci-
 ties between 10 and 40% are considered most important because they form the
 range of the opacity emission standards.

 LIDAR AND TRANSMISSOMETER COMPARISON
     The best correlation observed was between the lidar and in-stack trans-
 missometer opacity measurement methods (Figure 3).  The coefficient of corre-
 lation between the methods was 0.95, and the linear regression line was
 within 3% opacity of the ideal one-to-one relationship over the entire range
 of opacities studied.  Over the 10 to 40% opacity range, the measurement
 differences between the methods (lidar-transmissometer) ranged from -8 to
 +12% opacity, and the lidar measurements averaged 2.2% opacity higher than
 the in-stack transmissometer measurements.

 OBSERVER AND LIDAR COMPARISON
     The next best correlation was observed between the observer and lidar
opacity measurement methods (Figure 4).  The coefficient of correlation
between the methods was 0.88.  The linear regression line was within ±5%
opacity of the ideal one-to-one relationship over the entire range of

                                     18

-------
K
a
j
>
<
         C3,C4 - COAL-FIRED POWER PLANTS 3,4
         03,04 - OIL-FIRED POWER PLANTS 3,4
     	   R  - PHOSPHATE PLANT, ROCK DRYER
            K  - PAPER MILL, KRAFT RECOVERY FURNACE
            P  -PORTLAND CEMENT PLANT
          IN-STACK TRANSMISSOMETER
              • LSI RM4
              O LSI RM41P
          CORRELATION
          COEFFICIENT=0.95
              10        20        30        40        50        60
           PLUME OPACITY (IN-STACK TRANSMISSOMETER), percent
             Figure 3.  Plume opacity measurements by lidar
             versus plume opacity measurements by in-stack
             transm issometers.
                             .  19

-------
    90
cc
Ul

cc
LU
CO
>

<
Q.
o
    80
    70
    60
                   COAL-FIRED POWER PLANTS
                   1,2,3,4
   — Oj, 02, 03, 04 - OIL-FIRED POWER PLANTS
                   1,2,3,4
               Oc - OIL-FIRED BOILER,
                   CHEMICAL PLANT
               K - PAPER MILL, KRAFT
                   RECOVERY FURNACE
               P  - PORTLAND CEMENT
                   PLANT
      CORRELATION
     'COEFFICIENT=0.88
50
40
         Cl,C2,C3,C4
                       20       30       40

                    PLUME OPACITY (LIDAR), percent
            Figure 4.  Plume opacity measurements by
            trained observers versus plume opacity
            measurements by lidar.
                                20

-------
C3. C4, CB - COAL-FIRED POWER PLANTS
   03,04 - OIL-FIRED POWER PLANTS
      R - PHOSPHATE PLANT, ROCK DRYER
      K - PAPER MILL, KRAFT RECOVERY
          FURNACE
      P - PORTLAND CEMENT PLANT
IN-STACK TRANSMISSOMETERS
       •  LSIRM4
       O  LSI RM41P
CORRELATION
COEFFICIENT=0.84
                         95%
                      CONFIDENCE
                        LIMITS
     10        20       30        40       SO

   PLUME OPACITY (IN-STACK TRANSMISSOMETER), percent
   Figure 5. Plume opacity measurements by
   trained observers versus plume opacity
   measurements by in-stack transmissometers.
                   21

-------
opacities studied.  Over the 10 to 40% opacity range, the measurement differ-
ences between the methods (observer-!idar) ranged from -22 to 7% opacity, and
the observer measurements averaged 5.9% opacity lower than the lidar
measurements.

OBSERVER AND TRANSMISSOMETER COMPARISON
     The correlation observed between the observer and in-stack transmisso-
meter (Figure 5} was similar but slightly below the correlation observed
above between the observer and lidar.  The coefficient of correlation between
the methods was 0.84.  The linear regression line was offset 4 to 6% opacity
from the ideal one-to-one relationship over the entire range of opacities
studied.  Over the 10 to 40% 'opacity range, the measurement differences
between the methods (observer-transmissometer) ranged from -24 to +4% opacity,
and the observer measurements averaged 11% opacity lower than the in-stack
transmissometer measurements.
                                     22

-------
                                    SECTION 5
                                   DISCUSSION

     The results of the study show that of the three opacity measurement
methods studied (lidar, observer, and in-stack transmissometer), the instru-
mental methods of lidar and in-stack transmissometer correlated best (corre-
lation coefficient 0.95).  The observer and lidar correlation was second
(correlation coefficient 0.88), and the observer and in-stack transmissometer
correlation was worst  (correlation coefficient 0.84).  The data also show that
the observer opacity measurements were generally lower than either the lidar
or the in-stack transmissometer opacity measurements.

     The lower opacity measurements by the observer method and its relatively
poor correlation with  instrumental methods is to be expected whenever the com-
parison is of data obtained under a variety of environmental lighting and
plume background viewing conditions.  It is primarily due to the low measure-
ments associated with  the observer method when evaluating plumes under less
than desirable viewing conditions.  Low observer opacity measurements relative
to transmissometer opacity measurements has also been observed by Hamil  (12)
during a series of comparative tests between observers and in-stack transmisso-
meters that were conducted at a coal-fired power plant under different environ-
mental conditions.  Their data (Figure 6) show that the low observer measure-
ments were related to  the environmental lighting and background viewing
conditions of the plumes which caused a reduction in sensitivity and/or
negative bias of the observer method.  In Figure 6, a reduction in sensiti-
vity and/or negative bias of the observer method relative to the transmisso-
meter are represented  by low slopes (less than 1) and/or negative translations
of the curves, respectively.  Low observer opacity measurements were also
reported by Hood (13)  for a comparative study between the observer and in-
stack transmissometer  opacity measurement methods at a kraft recovery furnace.

                                     23

-------
    40
S
Q»

S
cc
UJ
2
o
Ul
          ENVIRONMENTAL CONDITIONS:
TEST1.
TEST 2.
TEST 3.
CLOUDY, LOW HAZE
SOLID OVERCAST
CLOUDLESS SKY, BRIGHT SUNSHINE
                   10            20            30

           PLUME OPACITY (IN-STACK TRANSMISSOMETER), percent
                                           40
            Figure 6. Linear regression relationships be-
            tween plume opacity measurements by
            trained observers and in-stack transmissometer
            at a icoa I-fired power plant for different en-
            vironmental conditions(12).
                             24

-------
A variable negative bias of  the  observer  opacity measurement method is recog-
nized by EPA, and  its  significance  with respect to  the enforcement of opacity
standards by the trained observer  (Compliance  Test  Method 9} is discussed in
the introduction to the method  (2).   From the  enforcement point of view, it
is important to note that  over  the  important 10 to  40% opacity range of the
opacity emission standards,  the  data  show that'no observer measurements were
greater than in-stack  transmissometer or  lidar measurements by more than 8%
opacity.

     Although the  data reported  here  have been analyzed  statistically to show
the degree of correlation  and linear  regression line  relationships observed
between the three  measurement methods, the results  are based on too little
data to show much  more than  trends.  More data are  needed to completely
evaluate  the methods.   It  is likely that  with  more  data  on the general appli-
cation of the methods, (a) the  degree of  correlation  and relative accuracy
indicated for the  instrumental  methods of lidar and in-stack transmissometer
will not  change significantly,  (b)  the correlations between the observer and
the  instrumental methods will become  similar but remain  below the correlation
observed  between the  instrumental  methods, and (c)  the observer measurements
will continue to be generally lower than  the instrumental measurements.
However,  comparisons  of the methods at specific sources  at specific times may
show better or  worse  correlations  and accuracies.   These differences will be
related to  the  specific atmospheric conditions and/or real in-stack and plume
opacity differences.
                                      25

-------
                                   REFERENCES


1.   Federal Register Vol.  39, No.  47:9808-98089, March 8, 1974.

2.   Federal Register Vol.  39, No.  219:39872-39875, November 12, 1974.

3.   Federal Register Vol.  36, No.  247:24877, December 23, 1971.

4.   Federal Register Vol.  4D, No.  194:46259-46263, October 6, 1975.

5.   Beutner, H. P.  Measurement of Opacity and Particulate Emissions with an
     On-Stack Transmissometer.  J.  Air Poll. Control  Assoc., 24(9):865-871,
     1974.

6.   Lear Siegler, Inc., Englewood, Colorado.

7.   Evans, W. E.  Development of Lidar Stack Effluent Opacity Measuring
     System, NTIS PB 233-135/AS, Springfield, Virginia, 1967.   96 pp.

8.   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.

9.   Bethke, G. W.  Development of Range Squared and Off-Gating Modifications
     for a Lidar System, NTIS PB 228-715, Springfield, Virginia, 1973.  47 pp.

10.  Conner, W. D. and J. R. Hodkinson.  Optical Properties and Visual Effects
     of Smoke-Stack Plumes, U.S. Public Health Service Report No. 999-AP-30.
     NTIS PB 174-705, Springfield,  Virginia, 1967.   89 pp.

11.  Conner, W. D.  A Comparison between In-Stack and Plume Opacity Measure-
     ments at Oil-Fired Power Plants, In:  Proceedings of the Fourth National
     Conference on Energy and the Environment, Dayton Sect. Am. Inst. of
     Chem. Eng., Dayton, Ohio, 1976.  pp. 478-83.

12.  Hamil, H. F., R. E. Thomas, and N. F. Swynnerton.  Evaluation and
     Collaborative Study of Method for Visual Determination of Opacity of
     Emissions from Stationary Sources, NTIS PB 257 948/OBA, Springfield,
     Virginia, 1975.  70 pp.

13.  Hood, K. T., and A. L. Caron.   The Relationship between Particulate Mass
     Emission Rate and Observed Plume Appearance from Kraft Recovery Furnaces,
     Paper 74-AP-08, presented at PNNIS-APCA Meeting, Boise, Idaho, November
     1974.
                                     26

-------
                                   TECHNICAS REPORT DATA
                            (flease read Instructions on the reverse before completing;
 REPORT NO.
  EPA-600/2-80-001
                                                            3. RECIPIENT'S ACCESSION NO.
 .TITLE ANDSUBTITLE

 ;OMPARATIVE  STUDY OF PLUME  OPACITY MEASUREMENT  METHODS
                                       5. REPORT DATE
                                           January 1980
                                                            6. PERFORMING ORGANIZATION CODE
 . AUTHOR(S)

W.  D. Conner and N. White
                                       8. PERFORMING ORGANIZATION REPORT NO.
i. PERFORMING ORGANIZATION NAME AND ADDRESS


(Same as  Block 12)
                                       10. PROGRAM ELEMENT NO.
                                        1AD712B   BA-010    (FY-79)
                                       11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
invironmental  Sciences Research Laboratory -  RTP,  NC
3ffice of  Research and Development
J.S. Environmental Protection Agency
Research Triangle Park,  N.  C. 27711
                                       13. TYPE OF REPORT AND PERIOD COVERED
                                        In-house
                                       14. SPONSORING AGENCY CODE
                                        EPA/600/09
15. SUPPLEMENTARY MOTES
16. ABSTRACT
The opacity of smoke-stack  emissions was measured  by three methods at thirteen
different  plants and the  results compared.  The  three opacity measurement methods  are
trained  observer, in-stack  transmissometer, and  laser radar (lidar).  The instrumental
methods, lidar and in-stack transmissometer, correlated better with each other than
with the observer method  and were also more precise  than the observer method.
Observer measurements were  generally lower than  instrumental measurements.  Data show
that for the range of opacity emission standards  (between 10 and 40% opacity), the
observer readings averaged  6 and 11% opacity less  than the instrumental readings of
lidar and  transmissometer,  respectively, and varied  by as much as 25% opacity below
to 8% opacity above the instrumental readings.
 17.
                                 KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                               b.lDENTIFIERS/OPEN ENDED TERMS
                                                     c. COSATl Field/Group
  Air pollution
  Chimneys
  Plumes
  Opacity
  Measurement
  Comparison
  Observation
Optical  radar
Transmissometers
13B
13M
21B
17H
14B
18. DISTRIBUTION STATEMENT


       RELEASE  TO PUBLIC
                          19. SECURITY CLASS (Tl
                           UNCLASSIFIED
35
                          20 SECURITY CLASS (This page)
                           UNCLASSIFIED
                                                     22. PRICE
 EPA Form 2220-1 (Rev. 4-77)   PREVIOUS EDITION '5 OBSOLETE
                                             27

-------