Handout
       Plant Inspection
  and Evaluation Workshop
     Opacity as an Indicator
of Control Equipment Performance
   U.S. ENVIRONMENTAL PROTECTION AGENCY
         OFFICE OF ENFORCEMENT
     OFFICE OF GENERAL ENFORCEMENT
         WASHINGTON, D.C. 20460
             April 1979

-------
                                           DRAFT
       OPACITY AS AN INDICATOR OF

      CONTROL EQUIPMENT PERFORMANCE
                   by

             Kirk E. Foster
  U. S. ENVIRONMENTAL PROTECTION AGENCY
          Office of Enforcement
Division of Stationary Source Enforcement
        Washington, D. C.  20460

                  and

           Gary L. Saunders
       PEDCo Environmental, Inc.
         505 South Duke Street
     Durham, North Carolina  27701
             November 1978

-------
     USE OF OPACITY IN EVALUATING AIR POLLUTION CONTROL
        EQUIPMENT OPERATING AND MAINTENANCE PRACTICES
     Measurement of plume or stack opacity can be a useful
tool in determining whether air pollution controls are
operating properly.  Air pollution sources are usually
subject to regulations limiting the plume opacity in addi-
tion to mass emission rates or concentration standards.  The
opacity standards established by the U.S. Environmental
Protection Agency, however, are generally set less stringent
than the mass emission standards so that a violation of the
opacity will be a clear indication of the decline in the
efficiency of the control equipment and presumably a viola-
tion of the mass emission standards.  Although agency field
inspectors are generally less concerned with opacity levels
which do not exceed the opacity limitations, these opacity
levels may be used as tools to indicate the performance of
the control equipment.  The presence of a noticeable plume
or unexplained changes in opacity even in the low opacity
ranges may signify deterioration in control equipment opera-
tion or changes in the process or raw material conditions
which adversely affect control equipment performance.
     Opacity of an effluent is an intrinsic property of the
plume and, depends upon the concentration of the particulates
in the gas stream as well as the particle size and other
optical properties.  Although the physical and chemical
characteristics of particulate emissions may differ appre-
ciably between source categories and even somewhat from
process to process within a given source category, for most
individual sources there is sufficient stability in these
optical properties to develop reliable opacity-mass emission
correlations.  Opacity-mass emission relationships have been

-------
developed and reported in the literature for a number of
source categories with a high degree of correlation.
     Opacity measurements may be made either instrumentally
 (transraissometer) or by a qualified observer using EPA
Method 9.  Unless there is physical or chemical secondary
formation of particulates in the plume as it leaves the
stack, the visual measurement has been shown to agree closely
with the in-stack transmissometer measurement.  The visual
measurements tend, in general, to be lower than the instru-
mental measurements.  This is due to the negative bias
and/or low sensitivities of the observer method when evalu-
ating plumes under viewing conditions where plume visibility
is less than desired for the method.  Also, the observer may
have difficulty in determining slight changes in opacity at
very low levels of less than 5% opacity.  The transmissome-
ter or opacity monitor with its higher measurement sensitiv-
ities would be able to detect slight changes in the low
opacity range and would, of course, provide a continuous
measurement of opacity when use of visual methods is not
possible.  Although the visual method is acceptable for
making gross estimates of the performance of a control
device, the transmissometer is more ideally suited for use
in determining proper operation and maintenance of air
pollution control devices.  The remainder of this discussion
will largely pertain to in-stack opacity measurements by
transmissometry and its use in determining the level of
performance achieved by a control device.
     Ensor and Pilat have developed a theoretical method for
calculating plume opacity from the properties and concentra-
tion of the particulate matter in the plume (Attachment 1}.
Use of constants (K)  that have been developed experimentally
from the effluent characteristics improves the reliability
of the theoretical method.  Once a suitable calibration
curve of adequate reliability is developed, it is possible
to monitor the process emissions on a continuing basis with
                              2

-------
opacity monitor measurements.
     A similar variation of the basic Beer-Bougert law for
the transmittance of light through a plume has been developed
by Conner* and is shown below.  The K factor used in the
equation developed by Conner is different in that it is the
reciprocal of the K values reported by Ensor and Pilat.

                                  _ Iff T
               % opacity = (1 - e  * m )  x 100

where:
                                   2
     „ _    attenuation       meter
     J\ —
         mass concentration    gram
    C  = mass concentration   gram/meter
     m
     L = path length of opacity measurements  meters
or:
                        I     %_0£\
               r  - -In U - 100  f
                m ~      (K)  (L)
     Figure 1 presents graphical  relationships that have
been developed for various source categories and the mass
concentration  (g/m ) of the particulate matter.  In this
figure, the value of K is simply  the slope of the  line
 m
(—).  If the plume path length is known as well as the
opacity, then an estimation of particulate mass-concentration
can be performed.
     Figure 2 presents an empirical method of determining a
K-value for emissions of an asphalt concrete plant con-
trolled by a baghouse.  The value of K is found by determin-
ing the slope of the line utilizing linear regression tech-
niques.  It should be noted  that  the term "extinction coeffi-
cient" is used here interchangeably with "attenuation coeffi-
cient" .
 *  ESRL, US  EPA,  Research  Triangle Park,  N.C.   27711.

-------
UJ
  0.22


  0.20


  0.18


  0.16


  0.14
o
u
*  0.12



 »| 0.10
 •
 L

 I < 0.08
 •
 r

   0.06



   0.04



   0.02



     0
                    ATTENUATION
                 IASS CONCENTRATIO

                     3
                                                                                        1.S
                          ( ) = NUMBER OF SOURCES STUDIED
                          GRAPH PREPARED BY WILLIAM CONNER.
                          ENVIRONMENTAL SCIENCE AND RESEARCH LAB.
                          U.S. ENVIRONMENTAL PROTECTION AGENCY
                          RESEARCH TRIANGLE PARK, N.C.

                                                                     0.5
                                                                                                   ci
                                                                                                   2
                                                                _

                                                                m
                                                                                                 0.2S
                0.02
0.04
0.06        0.08        0.10

 MASS CONCENTRATION, g/m3
                                                                        0.12
0.14
0.16

-------
                  OPACITY  -  MASS  CONCENTRATION   RELATIONSHIP

                  OF  THE  ASPHALTIC  CONCRETE   PLANT  EMISSIONS
   ,30
CSJ

cn  .20
>-
I—
t—I
o

o
    .10
                     20% Opacity
                                                       EPA Report No. 650/2-74-120
                                                       In-Stack Transmissometer Measurement
                                                       of Participate Opacity and Mass  Concentration
                                                       November 1974
                                                        K =  .33
                                                             I
                                                                    I
                      .16
                         .19
.32         >43 .48             .64


   MASS CONCENTRATION  (GRAMS  METER"

          FIGURE  2
                                                                                   .80
                                                                                                         .36
                                                                                                         .32
                                                                                                         .28
                                                                                                         .24
                                                                                                         .20
                                                                                                         .16
                                                                                                         .12
                                                                                                         .08
                                                                                                         .04
                                                                                                             oo
                                                                                                             O
                                                                                                             O
                                                                                                             o
                                                                                                  .96

-------
     Figure 2 also brings attention to another point.  In
the case of this asphalt plant, opacity levels above 10%
would indicate a violation of NSPS limitations if they were
applicable.  However, as mentioned previously, slight varia-
tions in opacity at these low-levels may not be detectable
by an observer.  Detecting levels below five percent opacity
would almost always entail the use of a transmissometer.
     The expected levels of particulate concentrations as
clear or near clear stack conditions are approached are
highly variable between source categories and processes
within categories, because of the optical properties of the
particulate matter.  For example, a pulverized coal-fired
boiler is predicted to approach 0.02 gr/ACF at clear stack
conditions whereas aluminum Soderburg processes are much
more sensitive and levels are predicted to be between  .001
to .003 gr/SCF.  A greater degree of control may be required
to achieve clear stack conditions.
     The use of an in-stack transmissometer is more ideally
suited to checking proper operation and maintenance of ESPs
and baghouses than it is to checking wet-scrubbers.  Scrub-
bers present a problem of entrained water droplets which
interfere with correct opacity readings.  Selection of a
monitor with narrow-band light filtration to avoid H.O
interferences is one approach for monitoring scrubbers.
Another approach is stack reheat to avoid water droplets.
An inexpensive single-pass transmissometer may be used to
indicate leaking bags as well as to optimize cleaning cycles
for baghouses  (Attachment 6).  Transmissometers may also be
used to "fine-tune" the rapper sequencing of electrostatic
precipitator to avoid excessive reentrainment of particulate
matter.  There are usually no problems of water vapor  in-
volved in these processes.
     A number of attachments are included which discuss in
further detail some of the topics covered in this discussion.
A bibliography of material to provide background information

-------
concerning optical properties of particulates,  measurement
techniques, and opacity-mass emission correlations is also
included.
     In summary, the use of opacity should be considered not
only as a tool to indicate the compliance of the source with
opacity regulations, but as an indication of operation and
maintenance procedures and for identifying potential problems
and deteriorating performance in the control equipment.  The
use of an in-stack transmissometer is more ideally suited to
this task than visual methods, particularly at low opacity
levels.  As long as the monitor is installed properly and
maintained it should provide both the source and the agency
with an indication of the control equipment's long-term
performance.

-------
Bibliography of Opacity Measurement Literature

-------
                       OPACITY AND OPTICAL PROPERTIES OF AERSOLS


Ensor, David S., and Michael J. Pilat.  The relationship between the visibil-
ity and aerosol properties of smoke-stack plumes.  Department of Civil Engineering,
University of Washington, Seattle, Wash.

Spaenkuch, D.  Atmospheric scattering functions in connection with the size
distribution of haze according to Jung's exponential law.  FTD-HT-23-340-69,
Foreign Technology Division, Wright Patterson Air Force Base, September 1969.

Pilat, M. J., and D. S. Ensor.  Plume opacity and particulate mass concen-
trations, Atmos. Environ. 4:163-173,  1970.

Ensor, D. S., and M. J. Pilat.  The effect of particle size distribution on
light transmittance measurement.   American Industrial Hygiene Association
Journal 32:287-292, May 1971.

Pilat, M. J., and D. S. Ensor.  Comparison between the light extinction aero-
sol mass concentration relationship of atmospheric and air pollutant emission
aerosols.  Atmosph. Environ. 5:209-215, May 1971.

Ensor, D. S., and M. J. Pilat.  Calculation of smoke plume opacity from par-
ticulate air pollutant properties.  JAPCA 21:496-501, August 1971.

Cornett, C. L.   Plume opacity.  M.S.  thesis.   Department of Civil Engineering,
Cincinnati University, Cincinnati, Ohio, 1972.

Ensor, D. S.  Smoke plume opacity related to the properties of air pollutant
aerosols.  Ph.D. thesis.  University of Washington, Seattle, Wash., 1972.

Peterson, C. M., and M. Tomaides.  In-stack transmissometer technique for
measuring opacities of particulate emissions from stationary sources.  Environ-
mental Research Corporation, St.  Paul, Minn.   EPA-R2-72-099, U.S. Environmental
Protection Agency, NTIS PB 212-741, April 1972.

Ensor, D. S., and M. J. Pilat.  Calculation of smoke plume opacity from par-
ticulate air pollutant properties.  JAPCA.   Reprint presented at 13th conference
on methods in air pollution and industrial hygiene studies, University of Cali-
fornia, Berkeley, Calif., October 30-31, 1972.

Franz, I., and J. Kraus.   The influence of forward scattering on measurement
of the degree of transmission of aerosols.   Staub Reinhaltung der Luft, vol.  33,
no. 9, September 1973.

Ensor, D. S., L. D. Bevan, and G. Markowski.   Application of nephelometry to
the monitoring of air pollution sources.  Presented at the 67th annual meeting
APCA, Denver, Colo., 1974.

Hermann, Jorg, and Hans Jorg Eiberweiser.  The influence of particle size
in extinction measurements.  Staub Reinhaltung der Luft. vol. 34, no. 5, May 1974.

Wolbach, C. Dean.  Approximations to discrepancies between visual and instru-
ment opacity readings for submicron particulate material.  Env. Sci. and Tech.
8:458-459, May 1974.


                                      11

-------
Hallow, J. S., and S. J. Zeek.  Prediction of Ringeltnann number and optical
characteristics of plumes.  JAPCA 23:676, 1974.

Ensor, D., and M. Pilat.  The correlation of plume opacity to particulate
mass  concentration.  Presented at 78th national meeting of American Institute
of  Chemical Engineers, Salt Lake City, Utah, August 18-21, 1974.

Conner, W. D.  Measurement of the opacity and mass concentration of particu-
late  emissions by transmissometry.  EPA-650/2-74-128, EPA National Environmental
Research  Center, U.S. Environmental Protection Agency, Research Triangle Park,
N.C., November 1974.

Pilat, Michael J.  Plume opacity related to particle properties.  Paper
presented at PNWIS-APCA meeting, Boise, Idaho, November 18-19, 1974.

Reisman,  £., W. D. Greber, and N. D. Potter.  In-stack transmissometer meas-
urement of particulate opacity and mass concentration.  EPA Contract Report 68-
02-1229,  EPA National Environmental Research Center, EPA-650/2-74-120, U.S.
Environmental Protection Agency, Research Triangle Park, N.C., November 1974.

Avetta, Edward D.  In-stack transmissometer evaluation and application to
particulate opacity measurement.  Owens-Illinois, Inc., Pittsburgh, Pa.  EPA-650/
2-75-008, U.S. Environmental Protection Agency, NTIS PB 243-402/AS, January 1975.

Weir Jr., Alexander, Dale G. Jones, and Lawrence T. Paypay.  Measurement of
particle  size and other factors influencing plume opacity.  Paper presented
at  the International Conference on Environmental Sensing and Assessment, Las
Vegas, Nev., September 14-19, 1975.

Fenn, R.  W.  Optical properties of aerosols.  In Handbook on Aerosols.  Edited
by  R. Dennis.  U.S. Energy Research and Development Administration TID-26608, 1976.

Harkowski, Gregory R., George J. Woffinden, and David S. Ensor.  Optical
method for measuring the mass concentration of particulate emissions.  Environ-
mental Protection Technology Series, EPA-600/2-76-062, U.S. Environmental Pro-
tection Agency, March 1976.

Weir Jr., A., et al.  Factors influencing plume opacity.  Env. Sci. and
Tech. 10:539-544, June 1976.

Pilat, M. J., and J. E.  Thielke.  Plume opacity related to particle size
and chemical composition (progress report).  EPA grant R8038 97-02, February 1977.

Weir Jr., Alexander.  Clearing the opacity issue.  Env. Sci. and Tech.. vol. 11,
no.  6, June 1977.
                                     12

-------
                     IN-STACK INSTRUMENTAL MEASUREMENT TECHNIQUES

Jahnke, J. A., and G. J. Aldina.  Continuous Monitoring for Source Emissions.
U.S. Environmental Protection Agency, Research Triangle Park, N.C.  EPA Contract 68-
02-2378.

James, J. E.  The use of a transmissometer as a process control tool to con-
trol emissions from a high efficiency electrostatic precipitator (undated report).

Wostradowski, R. A.  Continuous particulate monitoring of hog-fuel fired power
boilers and Kraft recovery boilers.  Macraillan Bloedel Ltd.  (in press).

General Federal procedures concerning control of emissions from stationary
sources as per paragraph 4.1 of the Federal Air Quality Act.   German Federal Republic

Sera, G. J. , et al.  Particulate mass:  A summary report.   State of the art:
1971, Instrumentation for Measurement of Particulate Emissions from Combustion
Sources, vol. I, NTIS PB 202-655, Springfield, Va.

Yocom, J. E.  Problems in judging plume opacity:  A simple device for
measuring opacity of wet plumes.  JAPCA 13:36, January 1963.

Buhne, Karl-Wilhelm.  Investigators into the directional dependence of photo-
electric smoke density measuring instruments.  Staub Reinhaltung der Luft,
vol. 31, no. 7, July 1971.

McKee, Herbert C.  Instrumental method substitutes for visual estimation of
equivalent opacity.  JAPCA 21:488-490, August 1971.

Cornett, C. L.  Plume opacity.  M.S. thesis.  Department of Civil Engineering,
Cincinnati University, Cincinnati, Ohio, 1972.

Ensor, D. S.  Smoke plume opacity related to the properties of air pollutant
aerosols.  Ph.D. thesis. University of Washington,  Seattle,  Wash., 1972.

Peterson, C. M., and M. Tomaides.  In-stack transmissometer technique for
measuring opacities of particulate emissions from stationary sources.  Environ-
mental Research Corporation, St. Paul, Minn.  EPA-R2-72-099,  U.S. Environmental
Protection Agency, NTIS PB 212-741, April 1972.
Gunther, Rainer.  An optical smoke density meter for direct indication of the
amount of solid:
September 1973.
                      2
amount of solids per m  flue gas.   Staub Reinhaltung der Luft.  vol.  33,  no.  9,
Cristello, J. C., and J. E. Walther.  An evaluation of an on-stack transmis-
someter as a continuous particulate monitor.  Presented at the 67th annual
meeting APCA, Denver, Colo., 1974.

Ensor, D. S., L. D. Sevan, and G. Markowski.  Application of nephelometry to
the monitoring of air pollution sources.  Presented at the 67th annual meeting
APCA, Denver, Colo., 1974.

Schneider, W. A.  Opacity monitoring of stack emissions:  A design tool with
promising results.  In Electric Utility Generation Planbook 1974, New York:
McGraw-Hill, 1974.

                                       13

-------
Nader, J. S., F. Jaye, and W. Conner.  Performance specifications for sta-
tionary source monitoring systems for gases and visible emissions.  EPA-650/2-74-
013, U.S. Environmental Protection Agency, January 1974.

Pfister, Ewald.  An integration instrument for timewise evaluation of emis-
sions.  Staub Reinhaltung der Luft, vol. 34, no. 2, February 1974.

Beutner, Heinz P.  Monitoring of particulate emissions from cement plants.
Rock Products. May 1974.

Wolbach, C. Dean.  Approximations to discrepancies between visual and instru-
ment opacity readings for submicron particulate material.  Env. Sci. and Tech.
8:458-459, May 1974.

McKee, Herbert C.  Texas regulation requires control of opacity using instru-
mental measurements.  JAPCA. vol. 24, no. 6, June 1974.

Schneider, W.  Evaluating particulate emission problems through opacity
monitoring.  Stack-Sampling News, July 1974.

Ensor, D., and M. Pilat.  The correlation of plume opacity to particulate
mass concentration.  Presented at 78th national meeting of American Institute
of Chemical Engineers, Salt Lake City, Utah, August 18-21, 1974.

Beutner, H. P.  Measurement of opacity and particulate emissions with an on-
stack transmissometer, JAPCA, vol. 24, no. 9, September 1974.

Conner, W.  D.  Measurement of the opacity and mass concentration of particu-
late emissions by transmissotnetry.  EPA-650/2-74-128, EPA National Environmental
Research Center, U.S.  Environmental Protection Agency, Research Triangle Park,
N.C., November 1974.

Cristello,  J. C., and J. E. Walther.  An evaluation of an on-stack transmis-
someter as a continuous particulate monitor.  Paper presented at PNWIS-APCA
meeting, Boise, Idaho, November 18-19, 1974.

Reisman, E., W. D. Greber, and N. D. Potter.  In-stack transmissometer meas-
urement of particulate opacity and mass concentration.  EPA Contract Report 68-
02-1229, EPA National  Environmental Research Center, EPA-650/2-74-120, U.S.
Environmental Protection Agency, Research Triangle Park, N.C., November 1974.

Walcher, Hans.  Digital intergrator to smooth the signals of instruments
for measuring smoke and dust density.   Staub Reinhaltung der Luft, vol. 34,
no. 12, December 1974.

Instrumentation for Environmental Monitoring, 1st ed.  Environmental Instru-
mentation Group, Lawrence Berkeley Laboratory, University of California, Berke-
ley, Calif., 1975, part 2.

Avetta, Edward D.  In-stack transmissometer evaluation and application to
particulate opacity measurement.  Owens-Illinois, Inc., Pittsburgh, Pa.  EPA-650/
2-75-008, U.S. Environmental Protection Agency, NTIS PB 243-402/AS, January 1975.

-------
John, Walter.  Investigations of particulate matter monitoring using contact
electrification.  Environmental Protection Technology Series, EPA-650/2-75-043,
U.S. Environmental Protection Agency, February 1975.

Cristello, J. C., and J. E. Walther.  Continuous onstack monitoring of par-
ticulate proves feasible.  Pulp and Paper, May 1975.

Shofner, F. M., G. Kreikebaum, and H. W. Schmitt.  In situ continuous measure-
ment of particulate mass concentration.  Paper presented at 68th annual meeting
APCA, Boston, Mass., June 5-20, 1975.

Nader, J. S.  Current technology for continuous monitoring of particulate
emissions.  JAPCA, vol. 25, no. 8, August 1975.

McKee, Herbert C.  Southwest Research Institute, Houston.   Letter to Texas
Air Control Board, re Owens-Illinois court proceedings,  August 25, 1975.

Winstanley, J. V., and M. J. Adams.  Point visibility meter:   A forward
scatter instrument for the measurement of aerosol extinction coefficient.
Applied Optics 14:2151-2157, September 1975.

Foster, Kirk E., and Norman White.  Use of the in-stack transmissometer in
manual source sampling for particulate mass concentrations (unpublished report).
Presented at East Central Section APCA/Source Evaluation Society Annual Meeting,
Dayton, Ohio, September 17-18, 1975.

Sticksel, P. R.  New directions in opacity measurements.  Battelle Memorial
Institute, Columbus, Ohio.  Paper presented at the Energy and the Environment Con-
ference, Oxford, Ohio, September 30, 1975.

Conner, W. D.  A comparison between in-stack and plume opacity measurements
at oil-fired power plants.  In Proceedings of the Fourth National Conference
on Energy and the Environment.  Am. Inst. of Chem. Eng., Dayton, Ohio, 1976.
pp. 478-483.

Markowski, Gregory R., George J. Woffinden, and David S. Ensor.  Optical
method for measuring the mass concentration of particulate emissions.  Environ-
mental Protection Technology Series, EPA-600/2-76-062, U.S. Environmental Pro-
tection Agency, March 1976.
Exton, R. J., and R. W. Gregory.  A four-channel solar radiometer for measuring
particulate and/or aerosol opacity of NO., and SO  in stack plumes.  N'ASA-TN'-D-
S182, NASA Langley Research Center, Langley, Va., June 1976.

Ensor, D. S., R. G. Hooper, R. C. Carr, and R. W. Scheck.  Evaluation of a
fabric filter on a spreader stoker utility boiler.  Report presented at the 69th
annual meeting APCA, Portland, Ore., June 27-July 1, 1976.

Hood, K. T., A. L. Caron, and R. 0. Blosser.  Variability observer and
instrument recorded plume opacity from Kraft recovery furnaces equipped with
high efficiency electrostratic precipitator.  National Council of the Paper
Industry for Air and Stream Improvement.  Report presented at the 69th annual
meeting APCA, Portland, Ore. June 27-July 1,  1976.
                                        15

-------
Hood, K. T., et al.  Measuring plume opacity.  TAPPI 60:141, 1977.

Buhne, Karl-Wilhelm, and Wolfgang Jockel.  Local and temporal distribution
of dust content in flue gas channels of larger steam boiler plants.  Staub Rein*
haltung der Luft. vol. 37, no. 5, May 1977.

Schmitt, Harold W. , Robert T. Muspliger, and Gerhard Kreikebaum.  Continuous
in situ particulate mass concentration measurements of industrial discharges.
Presented at 70th annual meeting APCA, Toronto, Canada, June 1977.

Chang, Daniel P. Y., and Bradford C. Greras.  Transmissometer measurement of
particulate emissions from a jet engine test facility.  JAPCA, vol. 27, no. 7,
July  1977.

Tomaides, M.  Instrumentation for monitoring the opacity of particulate emis-
sions containing condensed water.  Interpoll, Inc., St. Paul, Minn. EPA-600/2-
77-124, U.S. Environmental Protection Agency, August 1977.

Hood, K. T., and T. F. Briody.  Evaluation of the performance and applicabil-
ity of a laser light backscatter measurement instrument to the monitoring of a
particulate concentration prior to and exiting a Kraft recovery furnace electro-
static precipitator (special report).  National Council of the Paper Industry for
Air and Stream Improvement, December 1977.

Rudd, M. J.  Development of an optical convolution velocimeter for measuring
stack flow.  Environmental Protection Technology Series, EPA-600/2-78-049, U.S.
Environmental Protection Agency, March 1978.

Nader, John S.   Impact of sulfuric acid emissions on plume opacity.  U.S.
Environmental Protection Agency.  Presented at the Symposium on Transfer and
Utilization of Particulate Control Technology, Denver, Colo., July 23-28, 1978.

Sparks, L.   In-stack plume opacity from electrostatic precipitation scrubber
system at Harrington Unit 1 (draft).  U.S. Environmental Protection Agency,
Research Triangle Park,  N.C., October 1978.
                                       16

-------
            OPACITY/MASS EMISSIONS MEASUREMENTS:  SPECIFIC SOURCE CATEGORIES

James, J. E.  The use of a transmissometer as a process control tool to con-
trol emissions from a high efficiency electrostatic precipitator (undated report)

Stoecker, Wilbert F.  Smoke-density measurement.  Mech. Eng.,  October 1950.

Hurley, T. F., and D. L. R. Bailey.  The correlation of optical density with
the concentration and composition of smoke emitted from a Lancashire boiler.
J. Inst. Fuel 31: 534-540, 1958.

Edwards, L. V.  Smoke density measurement in municipal incinerators.  Pro-
ceedings 1966 National Incinerator Conference, American Society of Mechanical
Engineers, New York, May 1-4, 1966.

Duwel, L.  Comparative studies of different measuring principles for the
continuous monitoring of particulate emissions from lignite fired boilers.
Paper presented at the Second Int. Clean Air Congress.  Int,  Union of Air
Poll. Prevention Assoc., Washington, D.C., December 6-11, 1970.

Buhne, K. K., and L. Duwel.  Continuously recording surveillance of dust emis-
sions in the cement industry with the RM4 instrument.  Staub Reinhaltung der Luft
32:329-334, 1972.  (Summary of test results available from Lear Siegler, Inc.)

Larsen, S., et al.  Relationship of plume opacity to the properties of par-
ticulate emitted from Kraft recovery furnaces.  TAPPI 44:88-92, January 1972.

Buhne, Karl-Wilhelm, and Ludwig Duwel.  Recording dust emission measurements
in the cement industry with the RM 4 smoke density meter made by raessrs. Sick.
Staub Reinhaltung der Luft. vol. 32, no. 8, August 1972.

Ferguson, R. A.  Feasibility of a C. W. Lidar technique for measurement of
plume opacity.  Stanford Research Institute, EPA-650/2-73-037, U.S. Environ-
mental Protection Agency, NTIS PB 231-992/AS, 1973.

Feldman, P. L., and D. W. Coy.  Comparison of computed and measured opac-
ities:  Lignite fired boilers.  Paper presented at the 66th meeting APCA,
Chicago, 111., June 1973.

Rodriguez, D. G.  Measuring opacity of cat cracker flue gas.   Chem. Eng.
Profile 69:62-63, December 1973.

Schneider, W.  Evaluating particulate emission problems through opacity
monitoring.  Stack-Sampling News, July 1974.

Ensor, D., and M. Pilat.  The correlation of plume opacity to particulate
mass concentration.  Presented at 78th national meeting of American Institute
of Chemical Engineers, Salt Lake City, Utah, August 18-21, 1974.

Cristello, J. C., and J. E. Walther.  An evaluation of an on-stack transmis-
someter as a continuous particulate monitor.  Paper presented at PNWIS-APCA
meeting, Boise, Idaho, November 18-19, 1974.
                                        17

-------
 Hood, K.  T.,  and A.  L.  Caron.   The relationship between particulate  mass
 emission  rate and observed plume appearance from Kraft recovery furnaces.
 Paper presented at PNWIS-APCA  meeting,  Boise,  Idaho,  November 18-19,  1974.

 Reisman,  E.,  W.  D.  Greber,  and N.  D.  Potter.   In-stack  transmissometer meas-
 urement of particulate  opacity and  mass  concentration.   EPA Contract Report 68-
 02-1229,  EPA  National Environmental Research Center, EPA-650/2-74-120, U.S.
 Environmental Protection Agency, Research Triangle Park, N.C., November 1974.

 Fuirheller, W. R., and  T. L. Stewart.  Visible  emission  measurements at
 asphalt roofing  plants.  Monsanto Research  Corporation,  Dayton, Ohio.  Report
 76-ARM-ll.  U.S.  Environmental  Protection Agency Contract 68-02-1404.  (Two
 reports:  1975 and  1976).

 Cristello, J.  C., and J. E. Walther.  Continuous onstack monitoring of par-
 ticulate  proves  feasible.  Pulp and Paper,  May  1975.

 Shofner,  F. M.,  G. Kreikebaum, and  H. W. Schmitt.  In situ continuous measure-
 ment  of particulate mass concentration.  Paper  presented at 68th annual meeting
 APCA,  Boston,  Mass., June 5-20, 1975.

 Walton, John  M.,  Ernest C. Koontz,  and I. R. Mclnnis.  Air pollution in
 the woodworking  industry.  Tennessee Division of Air Pollution Control, Nash-
 ville, Tenn.   Report presented at the 68th  annual meeting APCA, Boston, Mass.,
 June  15-20, 1975.

 Visible emission  operations.  Kaiser Steel  Corporation.  EPA-330/2-75-010,
 U.S.  Environmental Protection Agency, November  1975.

 Knapp, K. T.,  W.  D. Conner, and R.  L. Bennett.  Physical characterization
 of particulate emissions from oil-fired power plants.  In Proceedings of the
 Fourth National Conference on Energy and the Environment.  Am. Inst.  of Chem.
 Eng., Dayton,  Ohio, 1976, pp. 495-500.

 Conner, W. D.  A  comparison between in-stack and plume opacity measurements
 at oil-fired  power plants.  In Proceedings  of the Fourth National Conference
 on Energy and  the Environment.   Am. Inst. of Chem. Eng., Dayton, Ohio, 1976.
 pp. 478-483.

 Ensor, D.  S., R. G. Hooper, R.  C. Carr, and R. W. Scheck.  Evaluation of a
 fabric filter on a spreader stoker  utility boiler.  Report presented at the 69th
 annual meeting APCA, Portland,  Ore., June 27-July 1, 1976.

 Hood, K.  T., A. L. Caron, and R. 0. Blosser.  Variability observer and
 instrument recorded plume opacity from Kraft recovery furnaces equipped with
 high efficiency electrostratic precipitator.  National Council of the Paper
 Industry for Air and Stream Improvement.  Report presented at the 69th annual
 meeting APCA, Portland,  Ore. June 27-July 1, 1976.

 Chang, Daniel P. Y., and Bradford C. Grems.   Transmissometer measurement of
particulate emissions from a jet engine test facility.  JAPCA, vol.  27, no. 7,
July 1977.

-------
Hood, K. T., and T. F. Briody.  Evaluation of the performance and applicabil-
ity of a laser light backscattcr measurement instrument to the monitoring of a
particulate concentration prior to and exiting a Kraft recovery furnace electro-
static precipitator (special  report).  National Council of the Paper Industry for
Air and Stream Improvement, December 1977.

Nader, John S.   Impact of sulfuric acid emissions on plume opacity.   U.S.
Environmental Protection Agency.   Presented at  the Symposium on Transfer and
Utilization of Particulate Control Technology,  Denver,  Colo.,  July 23-28,  1978.

-------
                     ATTACHMENTS

Calculation of Smoke Plume Opacity from Particulate Air
 Pollutant Properties, Ensor, D.S. and Pilat, M.J., APCA
 Journal, Vol. 21, No. 8, August 1971.

IGCI Consensus Report on Industrial Emission Levels
 Producing "Clear" or "Near Clear" Stacks, APCA Journal,
 Vol. 23, No. 7, July 1973.

Measurement of Opacity and Particulate Emissions with an
 On-Stack Transmissometer, Beutner, Heinz P., APCA
 Journal, Vol. 24, No. 9, September 1974.

Instrumental Method Substitutes for Visual Estimation of
 Equivalent Opacity, McKee, Herbert C., APCA Journal,
 Vol. 21, No. 8, August 1971.
Opacity  Monitoring  Technique Predicts Baghouse Mainten-
 ance, Saltz, J.  and  Cotler, L.,  Pollution Engineering,
 November  1978.

Continuous On-Stack Monitoring  of Particulate Proves
 Feasible, Cristello, J.C. and  Walther,  J.E., Pulp and
 Paper,  May  1975.

Impact of  Sulfuric  Acid  Emissions on  Plume Opacity,
 Nader,  J.S.  and  Conner, W.D.,  Symposium on  Transfer  and
 Utilization  of Particulate  Control Technology, Denver,
 CO, July  23  -  28,  1978.

Charts from  Evaluation and Collaborative Study of  Method
 for Visual  Determination  of Opacity  of  Emissions  from
 Stationary  Sources,  Hamil,  H., EPA 650/4-75-009,  January
 1975.
                          21

-------
                       Calculation  of  Smoke  Plume  Opacity from
                       Partfculate  Air  Pollutant  Properties
                       D. S. Ensor and M. J. Pllal
                       University of Washington
Calculation of $mok« plum* opacity from m« properties of tfw particulato emission
it facilitated with the  use of a parameter K (specific portieutete volume cm'/m1/
extinction coefficient m~') computed  from theory.   Graph* of K vs. me geometric
matt mean particle radius at geometric standard deviations  from  I (monodis-
perie) lo 10 are presented for particle refractive indices of  1.96-0.66i (carbon),
2.80-0.02i, 1.33 (water) and  1.50 at a wavelength of light of 550 nm. Experi-
mental data of K for various sources  are reported.   Application to the estimation
of the Ringelmann number is discussed and illustrated with an  example.
Nomenclature
*    extinction coefficient per volume of
        aerwol
(r)  particle number fraction frequency
        distribution
/    intensity of transmitted light
ft   intensity of incident light
I /It  light transmitUnce
I    <-!)»»
K   spn-ific particulate  volume /extinc-
        tion coefficient ratio
L   illumination  path  length  or di-
        trneler uf plume
m   refractive index of particle relative
        to air
•i   real part of refractive index
••   imaginary part ol refractive index
»{r)  particle Dumber freijiienry distribu-
        tion,  which,  multiplied  by the
        radius increment, f particles between r and r
        t Jr
Qt  extinction efficiency factor
r    particle riidius
Tf,  (scornclric numlwr mean radius
rr*  £'•', metric mass mean radius
Up  .-pocilic projected area or tlic specific
        extinction area
It"  total parti«'iil:ile  ina.is concentration
Grttk
      size parameter, 2rr/Jk
      wavelength uf light
      3.MI5!)
      average purtirle density
      particle  size  geometric  tUndard
        deviation
introduction
        Air Pollution Rofulationt
  Regulations to control particulate air
pollution from sources are usually of the
following types:

  1. visible  emissions   (Ringflmann
     No.);
  2. [articulate matter concentration
     grain loading (grains/scf);
  3. process weight code (Ib partitu'aie
     emission/It) process weipht);
  4. boundary line atmospheric aerosol
     coBicentrations  (micrograinM/'cubic
     meter).

  A detailed discussion of the various
air  pollution regulations and  the com-
munities imph'incnlinj; them  ran  be
found  ih  Stem.1  The  rvi;iil;itioii  of
sources by visible  evaluation of  the
cnu--ioH |j|ume o|xicity  (Kinjirlniann
No. limit)  U the lo:ist time i-un.-iiming
nnd loast exjionsive nH'th(MJ for a control
iim'iicy   to   itnpk'tnent.   For  these
reasons llini;elniuiin  numltrT ohservu-
tiuiii are  friH|iiently used  to control
Miurces.  In  Table  I the  Hitim-lnintm
number, plume opacity, uiul /ructioiiuj
light   lrut>suuUam-e  arc  ronijiared.
Ringelmann numbers are used by an ob-
server to describe black smoke plume*,
and equivalent opacity i* used  for white
or colored plumes.
  A  relationship between the  plume
opacity and the particle  )>liy$ii-al prop-
erties  (e*tx>ciiilly moss  concentration)
would uid in the design of control cqnijj-
ment ami indicate  cormspondoiicf be-
tween  the various types of regulations
for particulate air pollutants.   Often an
emission source will meet the paniculate
matter  concentration requircint'tit  but
not the visual standard (plume opacity;.

            Previous Work
  Correlations of particle nu-* concen-
tration to light trtinfmittitnce have been
ifjxirted for a few air pollution sources.
Stoecher1 and Hurley  and Bailey1 re-
ported  measurements  of  particulate
mas* concentration  and opacity from
eon I  combustion. Stern1 reported  the
ItinKclmann number ex|>eeted  for var-
ious  fly ush  concentrations from a rual
fired |>oivcr  plant.  sirniilt;nieo«i!« tneo-
suirrncnts of part irulate ma.-s con«-c.;tta-
tion  and plume transmittanrc for white
and black (fuel oil)  sinokp «nitrees weie
rc|>ortecl by  Conner and Hmlkiit.*on.'
Gun>ler* reported smoke meter measure-
imiils of stack tr:in«nuttaii(T ami nui««
concentration  for the cnii>M frntu a
Kraft reentry furnace.  TlH*>o >tuilie»
iinlii akil that in .1 siii|tk> -..iirci' thv light
extinction of the jiartivlo is diwtly rr-
latnl tu the ]i;irtlfli% lnu>s < iviurutl;tlioii.
Scatter in the i>\|H>rinieiit:il dnta ran U-
attiilinteil  in  part   to rhaiicis  in  tin*
pun irk site dittriliiition.  'i'lie inalttlity
to control the paiticli- >uo i(i>tri(mti,>n
was  ei(ed hy KiiKikihl' atnl l»y Mitclu-ll
                                                                             Journal of the Air Pollution Control Association
                                                       23

-------
 Table I.  Ufht extinction  relationships of
 plumes.   	
   Fractional                 Equivalent
   transmitted   Rincelmann    opacity
      light        number      Per cent
l.O
0.8
0.6
0.4
0.2
0
0
1
z
3
4
5
0
20
40
60
BO
100
 and Engdahl7 as a problem in developing
 an empirical relationship.
   The  two  general  theoretical  ap-
 proaches  have been  (1)  to assume a
 monodis|)erse  size  distribution  (or  a
 mean effective particle size) or (2) as-
 ciime a  mathematical relationship for a
 |>olydi-|>*rsc size distribution.  Huuks-
 ley et «/.,' re[>orted relationships between
 |iliune 1 115(1 1 extinction  unit  mass con-
 centrutimi for an average monodisperse
 -ize much smaller than the wavelength of
 light  :uul for an average  monodisperse
 size much larger than the wavelength of
 light.  (lux! agreement was reported
 with (he duta of Stoecher for the small
 l»artu-le case.'   Conner and Hodkinson4
 reported the use of a  mean effective
 |>artirle diameter to explain their duta
 and presented calculation* relating mass
 concentration  to  plume  transmittance
 for average particle sizes  in bluck and
 white nni>-iims.   Rohin.soti*  re|w>rted
 the u-e of size distribution data for iron
 foundry  emissions  to  calculate nu-
 merically the effect of control equipment
 on the plume opacity.   The Uuy Area
 Pollution Control Hoard Regulation 2*
 has a provision limiting the concentra-
 tion  from source* with a  relationship
 developed from a  Uincrlmjinn No.  2.
 Robin-Km reported that thi- regulation
 was devi>li)|H*d assuming an oil aerosol
 with  a geometric  mass mean radius  of
 O.J3M  niti\ a KPonwtrir  standard devia-
 tion of 3.4.  1'iliit and Kn-or" prex-ntrd
 general relationships of particubte mats
 concentration to plume triin.«mittance
 for log-normal particle size ilUtrihutKin.1
 and  Mack and while emissions.
        Obj«ctlv«« of Thl« P«p«r
  The objrc-tives of this pu|»er are:  (I)
 to extend the wiir-rul mathematical re-
 lationship* Itftwoen plume opacity and
 IwrticulatK air |x>llutunt  properties as
 reported by I'ilut and Kn-or to a wider
 range  of variables, (2)  to indicate the
effect and ran ire of values of particle re-
fractive index and density, and (3) to
        rrmilly rrconh-d siniiiltane«tu!«
           t* ul the plume o|iucity und
Optical »V«o«r«o« of Partkulate Matter
         Uf M Scattorhtf TMory
  Tin- attenuation of a rnllimated bt-um
tit ttKht thruiiKh  a  turbid  medium over

August 1971    Vokimt 21. No. I
 a path length L 'a given by the Bouguer
 law (Lambert-Deer law)

           ///• «• *xp| — bL\        (I)

 where I/It a the fraction of transmitted
 light and 6 is the extinction coefficient of
 a volume of aerosol (in'1).   Equation
 (1), in a slightly different form, has also
 been used to relate light transmit tunee
 of particle suspensions in liquids.  Skin-
 ner and Boas-Traube" reported an equa-
 tion in the form
              -exp|-Spiri|
(2)
 where Sp  is the specific projected par-
 ticle extinction area (m'/g) and H* is the
 paniculate mass concentration (g/m1).
 Skinner and Boas-Traub« verified equa-
 tion (2) for a wide variety of conditions
 for  a  transmittance range  of about
 SO-10%.   Equations (1) and (2) do not
 describe the effect  of  multifile tight
 scattering  which may be found in highly
 concentrated suspensions or over  long
 path lengths.   Hodkuison"  reported
 that multiple light  scattering  is rarely
 important  in extinction measurements
 of aerosols.
  The  light extinction coefficient  6 is
 well defined  theoretically and can be
 used  to develop relationships  for  the
 specific  projected   particle  extinction
 area Sp.  The extinction  coefficient for
 a volume of aerosol  coni|x>sed of spher-
 ical particles is  given by
      6 -  r I  n(r)dr    (4)
                                                               i
                                          where p is the average particle density.
                                          The particle size frequency distribution
                                          n(r) is related to the total particle num.
                                          ber concentration Ar and the panicle
                                          number fraction frequency /(r) by

                                                      »(r)  - tf/M          (5)
                                          The fraction of particles between r and
                                          r  + dr is obtained  by multiplying/(r)
                                          by the particle radius increment, dr.
                                            Horvuth and Charlson" retried that
                                          the ratio of paniculate mass concentra-
                                          tion to the extinction coefficient \V/b is
                                          useful for correlations of atmospheric
                                          visibility data.  The  theoretical
                                          ratio from equation (3) and (4) is
             , ,•
             /* -
                     4   r>
                     -fl
                     «»  J r,
                       Qt(«,m)r*n(T)dr
      The ratio  can be  made  concentration
      independent for single particle scattering
      by substituting equation  (5) into equa-
      tion (6)
                           r'/(r)dr
                   f'
                 LJr.
                                         CT)
      For  convenience,   Pilot  and  Eitsor
      (1970)"  assigned the name  parameter
      K  to the  integral ratio in  brackets
      (specific  particulatc  volume,  cm'/mV
      extinction  coefficient  m~').  The pa-
      rameter K is dimtrnsionully simitar to the
      volume Mirface characteristic  size de-
      scribed by Herdan (I960).'*   K  is pri-
      manly a  function  of  the particle size
      distribution, refractive index, and to a
      less degree, the wavelength of light.

                                        41?
                                                           24

-------
 Table IL  Selected *ir pollutant list distribution data.
Sou re*.
Uncontrolled
Wood smoke
Oil fired power
plant
Electric steel
furnace
Automobile tall
pipe
Cement dust
Pulverized coal
power plant
Hot mix asphalt
Spreader-stoker
coal furnace
Geometric mean
radius.
r«*iji
0.035

0.5

1.1

4.6
t.t

9.5
17

35
Geometric standard
deviation, r.
1.7

)

1.2

31
3

4
2

5
Reference
11

19

20

21
22

23
20

23
                                                                                   particle.  For   example,  eentospherea
                                                                                   (hollow  particles) with a density less
                                                                                   than I  g/cc  are sometimes found  in
                                                                                   power plant emiMtions."

                                                                                                 Calculations
                                                                                     The equation for parameter K it ob-
                                                                                   tained by substituting equation (11) for
                                                                                   the size frequency distribution /(r) into
                                                                                   the  integral ratio in  equation (7)  re-
                                                                                   sulting in
                                                                                   K -
  From equations (1) and  (2), the ex-
tinction coefficient b is related to  the
specific projected area Sp and particle
mass concentration by
              6 - SpW
(8)
By substituting equation (7) into equa-
tion (8), the speciSc projected area is
given by
                                   (9)
A working equation  can be developed
from equations (2) and (9)  to calculate
the expected mass  concentration  for
various values of  plume transmittance
(or opacity),  average particle density,
and plume diameter

         W  - -A' I in (///,)     (10)

      Range ef Important Variables
  Partick Sue Distribution.  The log-
normal particle size distribution model
can be used to describe a wide variety
of polydispersed systems such as  those
resulting from comminution of solids or
spraying  of  liquids.   A complete  de-
scription of the use  of the log-normal
size distribution  is given by Herdan"
and by Smith and Jordan."  The tog-
normal size distribution for a  number
frequency distribution is given by

               exp-
                       in' r/r.,
(11)
where rn is the geometric numlwr mean
particle  radius utid  f, is the geometric
standard deviation  (a  measure  of  the
polydispersity or breadth of the particle
sire distribution).   Lou-normal  distri-
butions  of  particle  tmmlxT,  area, and
mass arc related mathematically.  The
geometric  standard deviation remains
the same for these  tii>tril>utions.  The
relationship between geometric number
       and mass mean radii, r,.and r,^ respec-
       tively is

               lnrw  - lnr>.- 31i»V,   (12)

          Size distribution data can be reduced
       graphically by plotting ".smaller than"
       cumulative size  frequency versus size
       on logarithmic probability paper,  r,.,
       for mass distribution data, U the radius
       at the 50% size and t, is given by

                84.13%  sire     50% size
                 50% size     15.87% size

         Particle size distributions reported for
       air  pollutants  indicate that  both the
       geometric mass mean radius and the
       geometric standard deviation  can vary
       over wide ranges.  The  particle  size
       distribution in  the emissions from a
       given source depends on the nature of
       the source (combustion, metallurgical,
       etc.), and the degree  of collection.
       Hopefully, similar source*  should have
       similar  particle size  distributions.  A
       summary of typical size distributions is
       presented in Table II.  The  extremes
       in the geometric mass mean radius were
       from 0.04*. for wood  smoke to  35M for
       fly ash from an uncontrolled coal stoker.
       The extremes in the geometric standard
       deviation are  1.0 (monodisperse) to 31
       reported for automobile emissions.
         Refractiit Index.  The optical pro|>-
       erties of the particles are dcscritad by
       the  refractive  index.  The  refractive
       index is a complex number, m - n, - t'n:.
       The real part, n\, describes  the light
       scattering pro|x-rtic-s and the imaginary
       part, a,, describes  the light absorption
       of the particle material.  In Table III
       refractive  indices for pure crystalline
       solids und various liquids are presented.
       Air ixillut tnt particles may l«com)iou pre.«cntcd in Table
       III. For actual |*>liutanis the  particle
       der»ity may be much smaller than ex-
       (x-cted from the pun- aulistunce due to
       unhonioRt'ticou*  composition   of   the
                                                      I r.
                                                        I  rQ>
                                                       Jr>
                                                               (a,w)exp
The ]Miramctcr K wa* calculated using
equation (13) with equation (12) to ob-
tain r,. from r,,.  Tlie efficiency factor
Qi(a,m)  was  calculated  by  forward
recursion met hod* and was in agreement
with the results reported by Pcnndorf.*
Equation  (13)  was integrated with the
trapezoidal rule, reported by Deirmend-
jian* to be suitable for  integrations of
light scattering functions over particle
size.  The value of Ar was about 1% of
r over the wide range of particle sizes.
The theoretical limit Qt - 2.0 was used
for values of a greater than 75.  This is
a  good assumption if  the  acceptance
angle of the instrument used to measure
light extinction is very  small (for ex-
ample, the aci-cptuncc angle should be
less than 0.7° if the largest particle in the
duct is 5 jjm in diameter").

Results
    Calculation of A' versus r,r and »i
  The  results for refractive indices of
1.33, 1.50, 1.96-0.661, and 2.S-0.02tar»
presented  in Figures 1-4 respectively.
These results, as indicated in Table11*1,
cover the range of refractive index which
should  be important.  The  refractive
index 2.8-0.02i is a  value assumed to
represent  a  material such as iron oxide
with a weak light absorbing component.
A wavelength  of  light of 5oO nm  was
used a* an average for visible  light.   It
i*  approximately  the-  wavelength of
maximum sensitivity for the human eye.
The range of r,, and 9, is O.Ot  to 100^
nnd 1.0 {monodi^icT.-e) to 10.0, resix-c-
lively,  and  .-hould  incliule tin* values
ex|iertt%il for nio>t air |Milhit:tnt*.
   The  effect of refractive index i.< rel-
atively unimportant for particle railii
greater than ;il>mit O..V   However, for
jturticlc railii Jr., than abmit  ().;>*, the
nuiKiiitudcof tin-al>-i.il>inj: index (n>) is
very imimrtant.  The  light extinction
by pun-  lii;lil  Mallrniii: particle* (no
I III tit ubxirption  n,  -  0.0)  drcrc:i.c»
                                                         25
                                                                               Journal of the Air Pollution Control Association

-------
                                             Table III.  Physical properties ol pottlMe sir pollutanta,
 rnpiilly with u decreu'p wtth ti\e \\*t of e<|ua-
 liun <9) nnd the a»iiinmi> of an average
 particle dniMly.   A"  run uUo Iw drier-
 mtneil  with kriuwledm* of the  site dis-
 tribution parameter.* i geometric ma*s
 mron  rndiu* r,.  and  the  geometric
 standard  ili'viation -
 *rtw* «>f the aernsol.   OljvmiKly, ivhrn
 thfn« »rp !;iri»e t:-M'iiiiiiil  *in-  ilotritnition
 nui'li'l ivr ;i i-haiiKi* in  the  urnt-nl  fii  Ix'tM con
 tin*  '"itlM'iM »f purl nli-  -i
 nrnl  tln>  phine nparitv li
 ili'vi iii'ins (itiiii the tln'ou'tii al prc
 liwn- ma v rr«ult.   "i'lir niiMiiirml K
 in 'l.il'l'* l\* ni'lii'iile^  tlii> imiKiulnile of
 thr  n|i(i4-al  :u'tmty    (henry  -hiHilil lie
«>iiMida ZnO
Wood (mode
Sulfunc acid
90%
Banian*.
C,H,
n-Decane
Water


Density,
g/cc

1.5

1.8-2.1



5.2

2.7
2.J2-2.M
2.7
5.6
1.1

1.811

O.M
0.71
1.0
Wave-
Tempara- length of
tura. light.
'C nm

— 589.]

- 436
546
623

- 519.1

— 589.3
— 589.)
- 589.3
- 589.3
— 540

18. r 589.3

20* 589.3
20* 589.1
20* 589.3



Refractive Index.
r
n.

1.77

1.90
1.96
2.00

2.78-3

1.5-1.1
1.544
1.47
2.0
1.53

1.437

1.501
1.412
1.133
n « m — n,i
"i

to~*

0.68
O.M
0.66

.01 1

Mf
f
Vary small
r
r
0.00095

Vary small

Vary am«N
Vary smalt
Vary imill
Ref-
erance

24


29


M

M
2f
N
M
27

M

M
M
M
Table W. Measured optical propettls* of emissions.

Source
Orchard heater
(Black smoke)
Coal power plant
(Fly ash)
Coal itokar
(BUck «moka)
Coal itokar
(Black imoka)
Oil power plant
(BUck smoka)
Whita smoke
Generator
W - 0.!2 g/m'
W-0.«g/m»
W. 1.00 g/m'
Kraft mill recovery
furnaca W «
0.15 g/m'
Venaer dryer

Li.)

0.4J

1.14

0.30

0.15

0.2

0.2





1.S2
1

Instrument

Visual

Bolometer

Smoka mater

Smoke mater

Smoka matar

Smoke mater





Bolomatar
Visual
i
Average
Sp m'/g

20

0.78

6.1

4.6

8.7


2.5
1.8
5.75


1.7
2.8
(as- Avarag*
sumad) K cm'/m*

2.0 0.025

2.0 0.64

1.95 0.084

1.95 0.11

1.95 0.059

0.87
0.46
0.10
0.20


1.0 0.6
1.0 0.36
Haf-
erance

10

author

1

2

4

4





11
12.11
part tries" yr  cloud* of  spherical  |wr-
tirli".14   Liitht e.xtinrtirm by non-«plier-
iral partirU-t in randrun motion wn* re-
(Hirted by l|iNlkiti«in" to IN- nrnrly tin-
«uit\e  a*  thai fur  xphfriral   pui tii li>J
much lnrit'T and nitn-h minllrr  thun the
wavplencth of incident linht.  The effect
of irrri(iilnr particles nith sili'<  nr;ir (be
wnvclvintl\  of h|(ht  !•»  to  ^mU for nmnv yr»r<."   The li(ht
tnltering eorllirient of ft eloud of |wr-
ticli-s nas r«'|H>rtei| hy  Mnlland, flat.*
for tiilr dust and by ZUPV,  ft at ," for
ortifit ml  f 4.)   Rite distribution*  u
rtwim  in  FIKUIV*  l-t.   Aitiiitioually,
(he avernni1 (mrticle den!«it\  and the re.
Au«u*t
                       )\. No. I
                                                              26

-------
         \Geomatnc itandard deviation.
                             ft
             Rttrxtivt inde«-l 13
             w«v« length of I'jtit *J50 f»m
     10 T    10 '     10°      10'      10*
      Geometric mis* mean radiui,r,. (micron*)

       1. Parameter K at a function of the
log-normal «i»  dittrlbution parameter* for
liquid water.
 Tractive index  (the  magnitude of the
 imaginary part  is very important for
 small particles) may not be accurately
 known  for a given source.

           Simple Calculation
  The  estimation  of  the  maximum
 allowable  participate mass concentra-
 tion for a given plume opacity standard
 will be illustrated by an example cal-
 culation.   Let us u-vumc that the  regu-
 lation  s|>ecifies a  maximum allowable
 plume  opacity of Ringelmann No.  1
 (&09o light transmittance  through the
 plume).  It  is given  that the  stack
diameter L is 32.S ft (10 m). the exhaust
gases are  at  300°F, and  the  particle
 properties are a mass mean radius r,» of
 2>i, a geometric standard  deviation a,
of  3, a density p of  2 K 'cm1, :md a
 refractive index m of 1 .90-0. 6fii (carbon).
From Figure  3,  a  K of 0.6 cm1, m1  is
obtained.     With  equation  HO) the
particle mass concentration which cor-
responds  to the Ringelmann  No. 1  is
calculated
                                                                                       If/
                                                                                       10
                                                                                       to
                                                                                       10
       Geometne ttandard devwtion.
             R«ir*ctiveindnul.9« U.Mi!
             Wive lenfth ol Ii|hl*ii0 nm|
     10*10 l     10"      10'     1C1
     CtomilrK nwu mem radiut. f^ (micronu
Fl(ur* I. Parameter K aa a (unction of th*
loc-normal tit* distribution parameters for •
black aerosol.
    10''    10 '     10U      10'     10*
     Geometric mm mcin rtdiui, r,. (micrant)

Flfure t. Parameter Kit* (unction of the
lOaVnormil me distfibulion parameters for •
while aerosol.
                                                   -(0.6cmVm1)(2K/cmt) InO.S
                                                                10m

                                           ir - 0.027 g/cm' at stack conditions

                                           Converting to units of gr, ft1 at SOOT
                                           IP - 0.027 I'm-
                                          Converting to  standard conditions at
                                          60°F
Therefore  for this case the  maximum
allowable  particular  mats concentra-
tion which fonvsponds to a Riii^clmatui
No. 1 i> 0.012 cr, ACT or 0.01S vx SCF.
It shoulil Ix* notc*uin[>tiuu  of
constant  K.  Actually  there may  be
some  chaiiKi* in  K u]iou the in.siullation
of particle cullcction equipment >lue to
the decrease in  the particle sin-.  This
variation can be. taken into account if
information concerning the pnriii-le site
                                             JO*
                                                                                       10'
  10
                                                    Refractive mde«"iSO
                                                                HM-MO nm
10- •'     10-
 Gnomelric ni
10"
ID'
                                   io»
Ft«ura «, Parameter K at a (unction of the)
hypothetical refractive index for iron oaide.
too
                                                                                Journal of the Air Pollution Control Allocution
                                                          27

-------
dii.trilwt.ion of the controlled emission U
available (may be predicted either from
particle size  measurement*  at similar
source* or by calculations usin$ particle
collection  efficiencies of control equip-
ment as a function of particle size).

Conclusions

   A gfiieral  tlieoretic:i I relationship I*-
tween plume opacity and particle prop-
ertie< ha* Iwen developed and compared
\vitli ex|M>rimentul data. Numerical re-
sult« have l>ccn presented to include the
ramie*  of  physical  pro|>rrties expected
for air |M>ltutant*.   The parameter  K U
primarily a function of particle sue for
particle radii greater than about 0.5 urn
and primarily a  function of refractive
index   fur   smaller  particles'.  Plume
opacity can  be  estimated  for  given
source-. I'nmi the numerical re-ult*.


Acknowledgments

   The  it-search was supported by  the
United State* Public  Health  Service,
National   Air  Pollution  Control  Ad-
nuni'tratioii   Training  Grant  AP   29.
The assistance of  Professor August T.
Ro«ano, Rick A. Kestor, \Y'e$ Snowden,
John C. llox-h, and George Beckwith;
and dis.cussioti.i with Gene W. Smith,
Jon Peacy,  Dr.  Robert J. Charlson,
Steinar Lar.-M-n, and Leslie  E. Sparks
are (Hatefully acknowledged.  The  loan
of u Miioki* meter u. i* appreciated.

References

 I. Stern, A. C..  Air Pollution  Vol.  Ill,
    •JmJ oil.. A. C. Stern, lid.,  Academic
    Prev. New York. Chap 31. 196*.
 2. Mowker,   \V.  F..  "Smoke density
    measurement,"  Aleck. Eng.  72,  793
    (ItWOl.
 3. Hurley, T. F. and Bailey, P.  L., "The
    correlation of optical deiiMfy with the
    cimceniration  and composition of the
    sm.'ke  finitted  from  a Lancashire
    Holler," J. ln*t. purtil, W4 (19.58).
 4. Conner, U. U. and Hodkiruon, J. U.,
    Optical I'roprrtiei unit  I'l'iuaJ Elfectl
    of Smoke-Shirk riitmei, I'.H.rf. Publ.
    No. WJ-AP-:;mi!MJ7).
 5. <.::in.«ler, N. II., "The Use of a  Bo-
    lometer for Continuous Measurement
    of  Paniculate   Losses  from  Kraft
    Mill   Recovery  Furnaces,"  Pacific
    Northwest International Section.  Air
    Poll. Control Aisoc., Vancouver, B. C.
    'Nov.  I'.MW).
 6. EtiKdahl,   It.   B.,  "Correlation* of
    juiliiU  ronient  in ga«," Meek.  tng.
    73, -'4:1 (I»5I).

August 1971   Volum* 21. No. I
 8.




 9.



 10.


 11.



 12.



 13.



 14.


 15.



 16.

 17.




 18.




 19.


 20.

 21.




22.




23.



24.



25.
Mitchell,  R. I. and EngdmhJ, R. B.,
"A survey for improved method* for
the meuurement of particulate con-
centration in flowing  gin streams,"
/. Air Pill Control Attot. 13, 9 (1963).
Hawksley, P. U.,  Badzioch,  S. and
Blackett,  J.,  Mtasurtment of Solids
in Flue (jatet, British Coal Utilization
Re*.  Assoc.  Leatherhead,  England,
1901.
Robinson,  E.,  "The  effects  of air
pollution on visibility," Air Pollution
Vol. I  A. C. Stern", Ed., Academic
Pres*,  New York, Chap 7, 1962.
Bay Area Pollution Control  Board
Regulation 2, HMJ  Minion St., San
FrancL-co. Calif. (11(02).
Pilot. M. J. and Ensor, D. S., "Plume
opacity and  paniculate  ma&>  con-
centration,'1 Atmot.  Environ.  4,  163
(1970).
Skinner. O. G.  and Boo*-Trabe, 5..
"The  Ii(tht  extinction  method  of
particle sue estimation," Insl.  Ckem.
Engrt.23,57 (1947).
Hodkinson, J. U., "The optical mea-
surement of aero>ol.s," Arronol Seitnet,
C.  X.   Davies, Ed.  Academic Press,
New York. Chap  10. l'J(J45.
van de Hulst. H.C.  Light StaUt ring by
Small  f'ariiclet, John Wiley 4 SUM,
Inc., New York, I'J.57.
Hnrvath,  H.  and  CharL-ton,  R. J.,
"The direct optical measurement  of
atmospheric  air  pollution."  Amer.
Iml. Hi,g.  An. J., 30. 500 (19691.
H«rdan. (•.. Small  1'nrltcle Statittiei,
Biitterworths. London.  I'JOO.
.Smith.  J.  E. ami  Jordan,  M.  L.,
"MathematicaJ and graphical inter-
pretation  of  the  log-normal  law for
partirle gize  dUtrib>iti»n  analysis,"
J. Colloid  AVi. 19, A-l'J (1964).
Fo.-.(er, \V.  W.,  "The .->iie  of wood
amoke particles,"  Arrvlynnmic Cap-
ture af I'.irttclta, K. (.;.  IlicharU.-on,
Ed.  Periamon  PreMJ,  New  York,
     .  .   .
Smith, U'. S.. Atmo.iphrric EIHIMHUU
frUrn  Fuel  Oil  Cdmbtulion,  P.II.S.
Publ. No. fCjy-AP-2 (I9«i2).
Air  Pollution  Enijinrrring  .\fnniial.
P.H.S. Publ. No. 9«.»9-AP-«0 (I907).
Habit-i,
       ,   K.,   "Characterization  of
particulate  lead  in  vehicle exhaust-
exrwrimental   terhniepies,"   A.C:S.
Symp.  on  Air Quality  and  Lead.
Minneapolis, Minn. (April 19C9).
Kreichclf. T. E.. Kemnitz. D. A., and
Culle,  S.  T..  AtmoxpHeric  Emittioni
from  Uif  Ufamtifarture  of  Portland
Cemrnt. P.H.S. Publ. No. 999-AP-I7
/ I 'III** ^
Smith,  \V.  S.  and  Gruber, C. W.
Atmospheric  Emittiont  from   Coal
Comhuition.  An  Inventory  (luiJe,
P.H.rf. Publ. No. 9V.I-AP-24 (1966).
Pla»,  U. N.,  "Mie  scattering and
absorption cross  sections for alum-
inum oxide and  magnesium oxide,"
Appl. Optit» 3, S67 (1904).
StelTens, C., "Viability and air pc.l-
lulion," Air Pollution Handbook Magill,
P.  L..  Holden,  F.  R. and Ackley, C.,
!-M*    McGraw-Hill, New York, sec-
tion 6, 1956.
26. Hodgman,  C. D., Ed., Handbook of
    Chtmutru  and   Phyiict,  40th  ed.
    Chemical  Rubber  Co.,  Cleveland.
    1939.
27. Foster, W. W., "Attenuation of light
    by  wood  smoke,"  Brit.  J.  Appt.
    Pkyi. 10, 416 M959).
28. Penndorf, R. B., "New tables of total
    Mie scattering coefficients for spherical
    particles  of  real  refractive  indexes
    (1.33 to 1.50)," J. Opt. Soe. Amer. 47,
    1010 (1957).
29. Deirmendjian,   D.,   Electromagnetic
    Scattering on Spherical Polyditpenioiu
    American  Elsevier Publ.  Co.  New
    York, 1969, p. S9.
30. Snowden,  W. D., Valentine, Fisher
    and  Tomlinson,  Inc. Seattle,  Wash.,
    Personal  Communication (1970).
31. Bo-rh, J. C., "Size  Distributions of
    Aerosols  Emitted from  a Kraft Mill
    Recovery  Furnace,"   M.S.  Thesis,
    Chem. Eng. University  of  Wash..
    Seattle (1909).
32. Lars.sen,  S.,  Ensor,  D. S.,  Sparks.
    L. E. and Pilat,  M. J.,  "Size Dis-
    tribution of Aerosols Emitted from a
    Veneer Dryer,"   Presented at Pacific
    Northwest  International Section, Air
    Poll. Control Assoc., Spokane, Wash.

33. Bcckwith. G., Puget Sound Air Pollu-
    tion  Control  Agency, Seattle, Wash.
    Personal Communication (Jan., 1970)
34. Blau. H.  H., Jr.. McClee«!,.D. J. tod
    \Vut.son,  D. "Scattering by individual
    transparent spheres," Appl. Optict 9
    25£> (1970).
35. Holland,  A. C. and Draper, J. S., "An-
    alytical  and  experimental  investiga-
    tion  of  light  scattering from  poly-
    di.npersion*  of Mie particles,"  Appl
    Optic* o,  511 O'J67).
36. Zuev, V.  E., Kosh.'lev, B. P., Tvoro-
    gov, S.  I), and  Khmelcut), (Trans.
    bv l». A.  K.-eliii).
                                            Mr.  Ensor and  Dr. Pilat, Asso-
                                          ciate  Professor of Air  Resources)
                                          Engineering, are in the Water and
                                          Air Resources  Diy.,  Department
                                          of  Civil  Engineering,  Univer.titv
                                          of  Washington,  Seattle.    Thu
                                          paper  was  presented  as  Paper
                                          No. 70J%:i at the 63rd Annual Mwt-
                                          ing of the  Air Pollution  Control
                                          Association, June 1970.
                                                                           SOI
                                                            28

-------
                Reprinted  from AFCA  Journal,  Vol. 23,  No.  7,  July  1973
 IGCI  Reports Consensus  on  Industrial  Emission  Levels
 Producing  "C/ear"  or  "Near Clear" Stacks
  \  survey of the motiilirr companies
of tlic  Il\(lii>tri:il  (!;IH  ('It'Miiiiiv:  In-
htltuto (KiC'I) ha* developed tin' follow-
ing fon-cii5U-> as to pmi.-sioii levels which
ccnernlly produce or result in clear or
near clc.tr stacks for various industrial
applications.*
  Understandably, re|>orts  IGCI. there
are numerous <|ii:ilifications as to the
Mi-curacy of the data under all  eircum-
>tances.   The reader is  therefore cau-
tioned to  rounder any special factors
applicable to hi.i own operational con-
ditions in using these quideiine*.
  The data are for general informational
use to those concerned with the appli-
cation of  air |>ollutioii control equip-
ment.  Neither  the  IGCI  nor any
of its members make any representation
a* to the  Accuracy of the  data as A\t-
plied to any specific installation.
  Variables sui-h  a>  path length, angle
of incidence of liiiht, moisture  content
of effluent,  weather  condition!*, and
proi-e.-s changes may .-i^nincantly change
stack appearance irrespective  of  the
wtijjht of paniculate emivion.
  User*  and manufacturer* of air \m\\H-
tion  control  equipment  arc tilled  to
submit data  to IGCI relating to their
own  experience*  with emission levels
which result in clear or near clear stacks.
Confirming or contradictory data,  or
comment.-, in general will be welcomed
by  the  Industrial  Gas  Cleaning  In-
stitute,  1116 Summer Street, P.O. liox
1333, Stamford, Conn. 00004.
  * A paper wax prison led at  the May
IK,  I'.i'.'i, meeting in lialtimore  of the
South Atlaniic Suite* Action of  A PC A
 prepare
llii-  cim-cn-u-  data.  Tlic  paper  »'ua
presented by I'.nqcne 1*. SluMiiy, manager,
I'linlncl  riuiKiini;   nnd   Dcvcl'ipincnt,
l.nviii'iiiiinil:il  Sy-luiiiN  Uiv.,  Kopprr*
('ninpiiiiv, hie. and ctmriitly vice prrv
iilciii, IdC'I. Tin-«-ni in- PUJKT will iip|war
in the mcvlih;{ i>im-iitlim;«.
Industrial process emission levels generally cipeeted to produce visually clear
(or near clear) stacks, eiccpting condensed moisture plume.
Electrostatic precioitators.
fabric filters.
Industrial process Grams/ACF at stack txit
classification gas temperaturo. *F
Utilitiesaod industrial
power plant fuel-fired
boilers
Coal-pulverized
Coal-cyclone
Coal- stoker
Oil
Wood and bark
Bagasse
Fluid cok«
Pulp and paper
Kraft recovery
Boiler
Soda recovery
Boiler
Lime sludge kiln
Rock products (Kiln«)
Cement— Dry
Wat
Gypsum
AJumma
Lime
Bauxite
Mag. oxide
Steel
Basic oxygen
Furnace
Open hearth
Electric furnace
Sintering
Ore roasters
Cupola
Pyrites roaster
Taconite roaster
Hot scarfing
Mining and metallurgical
(Non-ferrous)
Zinc roaster (Ore)
Zinc smelter (Malting)
Copper roaster
Copper reverberatory
furnace
Coon«r converter
Aluminum— Hall process
Soderberg process
llmenite dryer
TiO: process
Molybdenum roaster
Ore beneficiation
Miscellaneous
Refinery catalyst
regenerator
Municipal incinerators
Apartment incinerators
Spray drying
Precious metal refininf


0.020  300
0.010 (A 300
0.010 <" 300
0.020 fe400


o.ois a 47;
0.015 ft 500
0.02 (•• 350
0.01 (.1 400
0.01 '" 400
Estimated
humidity*
Vol. %.
H:O


t
5
5
6
10
10
Dry


30

30
30

3
25
25
•20
1
3
20


20
1
S
10
10
20
10
10
25-40


9
%
S

S
S
S
25
2$
5
S
5


15-20
20
20
25
S
Wet scrubbers.
Grams/SCF Dry*
(SCF <& 70*F and
14.7 Psia)


0.029-0.031
0.014-O.Olt
0.030-0.090
0.0047-0.0053
0.090
0.072
O.C2Z-0.02*


0.040-0.043

0.040-0.043
0.046

0.026-0.030
0.034-0.040
0.048
0.040
0.039-0.050
0.035-0.037
0.024


0.021
O.C23-O.M7
0.026-0.032
0.040
0.036-0.040
0.029-0.03*
0.036-0.040
0.032
0.019-0.025


0.011
0.017
0.019

0.030
0.019
0.075
0.003
0.03*
0.015
0.015
0.034


0.032
0.034
0.038
0.022
0.0'. 7
                                        ' Based on etdmjted water vaoor content ol hot t[«ies

-------
                      Measurement  of  Opacity  and Particulate
                      Emissions with  an  On-Stack  Transmissometer
                      ttolnz P. Beutncr
                      LearSiegler, I no, Environmental Technology Division
An on-stock transmissometer system which is designed to
provide o precision meosurement of the opocity of visible
emissions is  described.   The sources of error in  opacity
measurements with regard to recent EPA emission monitor-
ing requirements and planned specifications are discussed.
Sources of error are voltage changes, temperature changes,
light source  and  detector aging and  effects of ambient
light.   Other  major  operational  errors  are caused by
alignment drift and soiling drift.  The methods employed to
minimize  these  errors  achieve  an accuracy  of  ±3% of
span and a maintenance free  operational  period of 3
months. The relationships between optical density, opacity
and transmittance are described. The instrument measure-
ment can  be correlated  with dust loading  provided the
particle size distribution is constant.  Examples are given
of correlations  obtained between optical density and por-
ticulate concentration in the gas on various types of emis-
sion sources and the  observed  error  margins are  sum-
marized.
     TrMKtlMr Unit
Flflura 1. Typical RM4 arrangement on a stack.
Two standard  methods for the continuous measurement of
pnrticulatc emissions are:
  Determination of the Ringelmann number or the equiv-
  alent opacity of visible emissions.
  Manual .sampling using particulutc sampling trains.
The problems of judging visual emission* by human observer*
are well known.  The original Kingclmann scale, comparing
the shade of gray of smoke with that of u chart, w useful only
for black smoke emission*.  TIM lay, smoke readers are trained
to judge 'the equivalent opacity of emissions of any color.
However, tlic result* are eiiorious manual procedure which, in large stacks,  may
require 2 to 4  man-weeks of effort involving u cost of from
$3000 to $10,000.  Typically, such sampling is performed
annually or hiamumlly, and as a result, it dix-s not provide in-
formation iilioiit the porforinam-e of control equipment during
the interval Ix'hvreii incMsiircmcnl.*.
                              ««7>">"'-d/romAPCA JOURNAL, Vol «, No 9, .SYptvmbcr IH74

                                                     31

-------
   Flfura 2.  RM4 trsn*ml«som«t»r optical system.
  Clearly, a better method for measuring and continuously
monitoring participate emissions is needed.  This paper dis-
cusses an on-stack optical  transmissometer system for pre-
cision measurement of the opacity of emissions and correlation
with  the particulate concentration in the gus.  The  R.M4
Visible  Emission Monitoring System was designed and  de-
velo|x.ld  in Germany and complies with the very strict long
term  operational i>crformance  s|x>cincations  established by
the German pollution regulatory agency for continuous partic-
ulate monitoring instruments.   It also meets the require-
ments of the United States Knvironmental  Protection Agency
(EPA) for opacity measuring instruments.  There arc over
1000 of these instruments now  o|x>rating in Euro)*1 and  the
United States.   The instrument* are  marketed and produced
in the United States,  under license, by the Environmental
Technology Division of LearSiegler, Inc.
Transmissometer System Description
  Figure 1 illustrates a generalised installation of the instru-
ment system, excluding the readout unit in the control  room.
This instrument consists of an optical transmitter, receiver
(transceiver) unit which is attached to one side of the duct or
stack, and a similar housing for the retroreflector on  the other
side.
  The transceiver contains a light source, a photocell/detec-
tor, and electronic signal-processing  circuitry.  The passive
reflector unit require* no electrical connections except for the
air blower provided for air (lushing on that side.
  Itoth  the transceiver ami the reflector are eqnip|M>d with
socially constructed air flow attachments  to keep the optical
windows free of dirt de|H>sits.   The purge air normally is
supplied by two high-volume nir blowers with filters.  They
provide up to NO cfm each, depending on pressure conditions
within the stack and pressure drop through the filter.  The
air-purging system typically provides 3 to  (i months of  main-
tenance-free o|>craticd, dual-beam, optical
system  (Figure 2).   This design automatically com|>ensatcs
fiir the HTi'ds of tcm|H-ruturi' changes, voltage  changes, and
component aging.  The instrument is non:nis|x»nsive to am-
bient light, since only pulsed light at cither W the two chop|ied
frequencies is measured.
  Light from a single source in the transceiver is divided into
a measuring Ix-am  and  a  reference  beam.  The measuring
beam is trmiHiiilted across tin- entire width  of  the smoke

NC                                                      32
channel to a corner-cube retrorefleetor which directs the beam
back through the smoke channel to the photocell detector, re-
gardless of small alignment variations between the two units.
The reference beam is projected directly onto the same photo-
cell detector.  A rotating disc modulates the two light beams
at different frequencies so that  both  beams can be measured
by  a common photocell, and electronic circuitry then .com-
pares the signals generated by the two beams.  The reference
beam provides automatic gain  control and, therefore, com-
|>cnsates for any chiingc in light output or photocell rcs|>onae
as a result of tomi>eraturc  variations, voltage fluctuations,
or component aging.  As  a  result, the instrument is free of
drift errors usually caused by these variables.
Tolerance for Misalignment
  Another common source of error  in smoke measuring in-
stallations is a variation in alignment as a result of buckling
and teni|>eraturc movements of  the stack or duct walla to
which the instrument is attached. As a result, some instru-
ments require the installation of a slotted pipe across the duct
or stick to maintain rigid alignment. Among the disadvan-
tages of this arrangement arc the ]>o*sibility of measuring a
non-representative sample of the smoke or dust, a reduced
instrument sensitivity  as a result  of measuring only  the
smoke passing through the slotted section of the pi|>e, and the
IKissibility of light vignetting, i.e., scattered light is reflected
off the walls ami  reaches the detector.  In large installations
and at  high  o|x>rating  tcm|>eratures, such a pijx? can nag,
causing a large measurement error, and the pipe usually is not
recommended for stuck widths greater than 20 ft.  The RM4
instrument system  does not utilize  a pi|>e across the smoke
channel, and it can be used over distances as great as 52 ft.
  The  instrument system tolerates  commonly encountered
alignment changes up to ±0.4°, without loss in accuracy, by
utilizing a simple patented concept.1'  The light beam  is
focused on the retrorellector for each  s|>ecinc distance and luw
a uniform intensity (±2%) throughout its cross section at the
focal distance.  For each x|H»cific distance, a reflector a|>cr-
ture size is chosen  that will maintain a large ratio of beam
cross sectional area to reflector  area.  Since the beam is uni-
form, the retrorefleetor always  returns a constant amount of
light, regardless of what |mrtion of the beam impinges U|*m
the  retrorefleetor.  Furthermore, the corner cube  retrore-
llector always returns light directly to the source, regardless of
the angl*' <>f incidence.  Within  its  accuracy specifications,
this instrument system tolerates alignment changes of ±0.4°'
which is equivalent to mcasuring-l>cam movements of 1 to 3
in. at distance* of 10 and .'W ft.   This is sufficient to account
for all alignment variations encountered in actual practice.
In addition, the instrument  is cquip|x-d with an optical bull's
eye which allows verification of  the alignment at any time
during o|>cration.  If necessary,  the alignment can  be  ad-
justed using two fine-adjustment screws on the instrument.
Automatic Calibration
  Temperature changes, voltage changes, accumulation of dirt
on the optical windows, and  alignment changes urc the major
sources  of drift in smoke or dust density measuring installa-
tions; the calibration of such installations must be  checked
frequently.  New source  |x>rformancc standards, published
by  the KPA,1 require o|«icity monitoring instrumentation in
all  new  steam electric power generating stations and further
.sjHH'ify that such instruments be calibrated at least once every
»lay.  None of the two-ended smoke-density monitoring  in-
strument* so far available can be calibrated during <>i>cration
on  the  stiflrk  since, in any two-ended system (light source
and detector on op|xisite sides of the smoke channel), it is
necessary to stop the exhaust gas flow in order to obtain a

               Journal of the Air Pollution Control Association

-------
 calibration for awro pereent opacity.  Tke angle-ended RM4
 system checks its iero and span calibration automatically at
 regular intervals without interrupting stack operation.  The
 calibration values obtained are recorded directly on the chart
 paper.  The calibration checks can also  be activated man-
 ually at any time from the control room.
  The calibration checks in the system are accomplished by
 inserting, in the light path outside of the optical window of the
 transceiver, a zero reflector which simulates the retroreflector
 on the  opposite side of the smoke channel.   A variable  dia-
 phragm on  the zero calibration reflector  is adjusted during
 instrument setup so that as much light  is returned  to the
 photocell  by the  calibration reflector as the retrorollector on
 the opposite side of the smoke would return under a "clear-
 stack"  condition.  Span calibration is accomplished by in-
 serting  a  neutral density filter of  known  opacity value into
 the reflected zero beam.  The value of the selected span cali-
 bration point depends upon the particular measurement range
 selected for  a specific  installation.  The  rero  calibration
 check indicates any drift in the instrument measuring circuit
and measures any loss of light transmission through the op-
tical window as a result of dust accumulation.  A deviation
from zero nlerts the operator in the control room when clean-
ing of the optical windows is required.  The loss of light as a
result of dirt accumulation on the reflector side of the instru-
ment system can normally be assumed to equal that on the in-
strument side, since both sides are protected by identical air
purging systems and arc subject to essentially the same stack-
gas flow and pressure conditions.
  An example of a smoke trace is shown  in Figure 3.  The
zero  and span calibration values are recorded automatically
at selectable, periodic intervals.  Figure 3  also shows periodic
emission spikes, resulting from the rapping of different precip-
itator sections.  Therefore, the instrument is a very effective
 tool for monitoring the ixrformancc of each  precipitator sec-
 tion and for detecting malfunctions in precipitator operation.
 Accuracy end Sensitivity
  As a result of the design features described, the instrument
system achieves an operational accuracy of better than ±3%
of span or  ±1.5% opacity,  whichever is greater, over an
operational period of  at least three months.  Test data ob-
tained in Germany as part of an acceptance test performed by
 the Technischer tjberwachungs-Vercin (TOY),' demonstrate
 that none of the following variables causes an error exceeding
 ±3% of span for a full scale range of 0 to 0.45 double-pass
 optical density* (0 to 0.225 single pass or 0 to 40.4% opacity),
 and most are well under this limit:
  Voltage variations within ± 10% of nominal.
  Temperature variations from -30°C to + S5'C (-22°F
  to+13l°F).
  Component aging and effects of ambient  light.
  Alignment variations within ±0.4° of the optical axis.
  Calibration and linearity errors in measurement scale.
 Operational tests on  stack emissions of  cement plants  and
 power plants confirm that none of the following errors exceeds
 ±3% of span  for the 0 to 0.45 double-pass optical density
 full-scale range:
  Zero  and .spun drift  over a maintenance-free operational
  period of 3 mo.
  Soiling and alignment drift over & maintenance-free O|xra-
   tinnal IKTKK] of 3 mo.
   • The iiutnimenfe light beam pae*ee through the emoke t*lee. therefore
 neacurM more than the opwtily eeea by aa obeerver (eee further below).
Figure 1.  Example of emission trace from cement plant showing
hourly zero and span calibration checks.
A choice of five full-scale measuring ranges which correspond
to stack opacities of 0 to 9.8%, 0 to 18.7%, 0 to 40.4%, 0 to
64.5%, and 0 to 87.4% is provided.  The instrument is sensi-
tive to  changes as low aa 0.05% opacity, and, on  its lowest
range, it can accurately measure opacities of 1 to 2%.  This
high sensitivity is a result both of the double pass of the light
through the entire stuck or duct  width and of the expanded
measuring scales.
  The absolute accuracy of the transmissometer is baaed  on
the calibration of the linear scale of optical density against
neutral  density  niters of known transmittance.   The mea-
sured values must be  within ±2% of the calibrated values of
the niters.
  The calibration against neutral density filters is acceptable
if the angles of projection and view are small.  A neutral den-
sity filter is not strictly equivalent to a smoke or dust aerosol,
since light attenuation in the filter is due primarily  to absorp-
tion, while attenuation in the aerosol is the result of both ab-
sorption and scattering.  Instruments with large angles  of
view measure a jwrtion of the forward (or backward) scat-
tered light and therefore measure a lower apparent opacity
for the aerosol than for a neutral  density filter of equal o]iao-
ity.*>'  Since,  for particle sizes under 2 microns,  nearly  an
order of magnitude less  tight is backscattcred than  forward
scattered, and since the angles of view arid projection are
very  small,  the folded beam system utilized in the trans-
missometer has an absolute accuracy equal to any equivalent
two-ended transmissometer.  The angle of view for this  in-
strument is  ± 1.8° and the angle  of projection is ±0.7° from
the optical axis.  Proposed E FA specifications for visible emis-
sion measuring instruments suggest limits of  ±2.5° for the
angles of view and  projection in order to  assure an accurate
absolute oimcity measurement by instruments from different
manufacturers.

Photopfc Light
  An additional  requirement for absolute accuracy of the
measurement is a specified spectral range, since light attenua-
tion by a polydisncne aerosol with mean  particle size of less
than I  it \s  a function of the wavelength of the light.  The
specified spectral distribution is pkotopic, the same as that of
the Unlit-adapted human eye, because one objective is  the
measurement of the opacity  of visible emissions as observed
by the human eye.
   The instrument  utilizes a photocell/filter combination to
achieve "photopie" H[X'etral sensitivity.   Figure 4 shows  the
8|>ectrul sensitivity  of the instrument compared to  the output
of  an incandescent light  source.  The  maximum light  in-
tensity  is at 0.55 micron wavelength compared to about  1.0
micron for the- incandescent source.  Instruments sensitive to
the entire output of the light source arc subject to two serious
 September 1974   Volume 24, No. 9
                                                         33
                                                       K7

-------
  100
                                                        2.f
 Retire 4.  Photopic spectral sensitivity of the RM4 compared
 to Incandescent light spectrum.
 errors.  First, water absorption bands in the near infrared
 portion of the s|>ectrum cause large measurement errors for
 •tack  gases containing high humidity.  Second, the instru-
 ment response to submicron participate matter is significantly
 lower  than the re.s|>onsc of the human eye, resulting in low
 readings when the stuck emissions contain substantial portions
 of submicron particulutc matter.
   Figure 5 shows the light attenuation of a |x>lydis|>erse aero-
 sol for two different spectral maxima as a function of the mean
 particle  size.7  The attenuation of the  1.0 M light falls off
 significantly at about 0.8 to 0.6 n particle size, while the 0.55
 M light is attenuated by particulate matter down to 0.3 to
 0.2^.size.
  In the absence of photopic light and es|>cciully when the
 major  instrument resjwnse is within the infrared region, as in
 instruments utilizing a bolometer detector cell, large errors
 result when the emissions contain large portions of submicron
 particulate mutter.  This generally occurs in  modern plants
                                       with strictly controlled emissions, since the control equipment
                                       primarily removes the larger particle sites, while submicron
                                       particles pass through.
                                       Instrument Output

                                         The instrument has a 0 to 20 milliampere output directly
                                       proportional to optical density units, with an elevated  in-
                                       strument zero at 10% full-scale (2 milliamps).
                                         Figure 6 shows the relationship between the optical density
                                       output of the instrument and the more familiar scales of opac-
                                       ity  or transmittance. Optical density (D) is defined as the
                                       logarithm (base 10) of one over transmittance (T):

                                                        D - log 1/T - -log T              (i)

                                       Since opacity (O)  is defined by 0  - 1 -  T, the relationship
                                       with optical density can be expressed as:

                                                   D - -log T - -log (1  - 0), or

                                                          T - 1 - 0 - 10~»                (2)

                                       If Tor (1 - 0) is plotted ona logarithmic scale, a linear correla-
                                       tion is obtained as shown in Figure 7.  The above relationship
                                       is known as Bouguer's  Law, equivalent to the well known
                                       Lambert-Beer's Law for the measurement of absorbance  by
                                       colored solutions.  It can be shown that the optical density
                                       for the case of light attenuation by a smoke or dust aerosol
                                       like the absorbance for  the case of light absorption in soh>
                                       tions, is proiwrtional to the jwthlcngth (d) and the concentra-
                                       tion (r) of the light attenuating material.
                                                             D- k
                                                                                             (3)
                                       This relationship is used to correct the measured optical den-
                                       sity values (DO for the measuring distance (rf,) at the instru-
                                       ment installation site to the values (D,) applicable for the
                                       stack-exit diameter (d,).  Since a folded-beam instrument
                                       measures through twice the smoke width at the installation
                                       site, the distance dt must be taken twice as follows:
                                                                     dt/2dt
                                                                                             (4)
                                                                                             \^/
                                      The corrected optical density values can then be converted to
                                      opacity units by means of the graphs in Figure 6 or 7, or equa-
                                      tion (1), or by using readily available tables  (Handbook of
                                      Chemistry  and  Physics)  or  nomographs.  An  electronic
                                      opacity converter and readout unit which is intended for in-
                                      stallation on the instrument panel in the plant control room
m
     (.t       (J    dJ  14 U        1J        MUM  M       10.1
                                •iillll DIMMH .MOMS
Wi.ilre I.  Llftrt attenuation of a polydlspef se aerosol as a function of particle size.


                               34                   Journal of trw Mr Pollution Central AsMctetfon

-------
 IB available.  It allows the selection of a meter readout in
 terms of either optical density at the measuring cite or opacity
 at the stack exit, and either or both of these outputs can be
 recorded on a strip chart recorder.  The available measure-
 ment ranees indicated in Figure 6 are single pass rrjui valents of
 the instrument's double pass measurements.
 Correlation with Dust Loading
   According to  llongucr's L-uv, the optical density instrument
 readout is directly proportional  to the dust concentration in
 the  gas.   The  constant of proportionality differs with  the
 density of the particulate matter, its size distribution, aiul its
 optical  properties.  The constant can be calculated if all data
 nre  known.'-'-*   In practice,  it  is  easier to c^jcrimentally
 determine the  proportionality  constant for  each  installa-
 tion.'•'<>-"
   Empirical calibration of the transmissometer in tonns of
 particulate concentration requires at least OIK- data jxjint  ob-
 tained by  stack sampling, using  an approved sampling train
 and  method (KPA method 5).  The average of  the  trans-
 missometer readings over the |>criod of the stack sampling
 test  equals the measured  dust  concentration in  mg/m* or
 gr ft' at actual conditions.   A  straight calibration line is
drawn from the origin through the ex|>erimental point.
   Calibration in terms of dust loading is valid only as long as
 the particle size distribution and other psirtii-iilntc pni|M>rties
% con-
fidence level are indicated in the  figure.   Figure Kb shows  the
name calibration curve after correction to standard conditions
(teni|ierature, pressure) and dry  gas.  The corrected calibra-
tion  curve can Ix1 useil when  gas temperature, pressure, and
humidity are constant.  Figure 9 shows a similar calibration
               1J
               O.f

               «.l
               e.$

               0.4

               0.9

               O.I

               0.1
           . %
                       10
                               I	I
     TrmminMC*, % 100 W M  TO  M  M  40  M  M 10  0

    "K»l» OMlwd Lhw* H*Brn*nl LItnlii »l Wnfte PCM
         M««tuf*m«nl
     Figure C.  Relationship between  optical density and
     opacity or trHnsriiittance (linear «cal«).

September 1974    Volume 24, No. 9
s
M 1






1 1
0 1
1
• 1





X
0 1
i i
i
« 7





X
« 1
1 1
r i
V 1






« «
3 •
1
0 1




X

D I
4 1
1 '
0 <




X

« 1
* 1
%fim
1
0 3



/*


0 T
( «
•»*•»•»
1
« 1


/
'


0 1
r 1
i
« i

^
^



0 •
* 1
t
I






0
*




.•f
"I
..«
1 J



                                                               Figure 7.  Relationship between optical density and opacity or trans-
                                                               mittance (logarithmic scale)
                                                               curve  obtained on  a lignite-fired boiler in Germany,  and
                                                               Figure 10 shows a .similar calibration curve for a bituminous
                                                               coal-fired jxiwer station.
                                                                 The system is being used in numerous installations in the
                                                               United States and  in Kiiro|H> for the continuous monitoring
                                                               of both opacity and  paniculate  concentration, based on the
                                                               correlations established Ix-tween the optical rending and the
                                                               dust loading.  In some states of the  German  Federal  Re-
                                                               public the instrument is approved for compliant* monitoring
                                                               on specific large emission  sources.  In this application the in-
                                                               strument calibration  in terms of particulate concentration is
                                                               cheeked by  manual testing  once it year.   In evaluating the
                                                               monitoring record, credit for the possible measurement error
                                                               is given.
                                                                 Table 1 summarizes the levels of error observed in a number
                                                               of correlation lines  for typical  emission sources.  Four dif-
                                                               ferent  definitions of  error have  been applied.   Values have
                                                               been calculated from the available raw  data using a computer
                                                               program.*   The error definitions are as follows:
                                                               A. Standard deviation is the root of the mean.
                                                               B. Ty |»e 1 error is the 9.r>% confidence level that the true mean
                                                                  of'all observed optical data for n given particulate concen-
                                                                  tration will lie within the limits.
                                                               C. Tyjx- 2 error is the 05% confidence level that the next one
                                                                  observation will fall within the limit.
                                                               D. Ty|K> 3 error is tlie 95% confidence level that 95% of all
                                                                  possible observation* (05% tolerance) will full within the
                                                                  limit.
                                                               In Europe type 3 error calculation is typically nppliett, while
                                                               in the 1T. S. the type 2 calculation is most  common. The
                                                               smallest error level is obtained by the  ty|x« 1 method,  As a
                                                               jH-rcentagv of the measured mean particulate concentration,
                                                               the ty|M-1 error level ningesfrom 5 to 21.1%.
                                                                 Itecuuse transinissometer data arc typically correlated with
                                                               1 hr stack sampling data, an individual duta |x>int in the con-
                                                               tinuous transmissometer  recording can be defined as a  1  hr
                                                        35
                                                                  •Th« author trmnki Gtrald McGowmn fur his uulitannln providing th«
                                                                cominitrr data.  The ormr mlrttlaliun in avaiUlile at a Mrviee from Lear
                                                                8i*«l«r. Inc. Knvinmm«nlal Trrlm»li>Ky Divinion.
                                                                                                                        tet

-------

         MO
         1*0
              •   Tn
                          (Without Onm)
            -•.11
            •.it
                  *.••     o.i
                    Optical
                                 ».«
                                         *.i
                                                 t.tl
        Mt
      w
      I.

              • —  Norm* Operating Mod*
              i —  Nornul Mod* Wllh Deviation*
              a—  Mroct Mod* (Without Oryw)
             •»—  TramMomlModt
                 0.01     (1      Oil     0.1
                   Oallcal Density - flitfta *M«
                                                 O.M
      Figure I. Relationship  between  optical density and
      participate mass concentration for a cement plant.
                   a.  Actual conditions.
                   b.  Standard conditions.
         average  reading.   The  continuously  recording  instrument
         easily allow* determining 8 hr or 24 hr average*.   The effect
         on the error probability is therefore equivalent to averaging
         8 or 24 individual data points.   In Europe the instrument is
         now used in many applications in conjunction with an inte-
         grator which provides simultaneous outputs of 1, 8 and 24 hr
         continuously ujxlaU'd averages.  If any of the averages ex-
         ceeds a  predetermined  value,  controls are  activated auto-
         matically.   The longer term averages improve the accuracy
         and reliability of the emission data recorded.
           The correlation  data  available from  different emission
         sources allow a comparison of stack opacities resulting from
         different emission sources at equal paniculate concentrations
         and  stack diameters.  Some calculated opacities for a con-
         centration of 150 ing/Nni1 (0.071  gr/scf) and a  stack  exit
         diameter of 3 m (10.12 ft) are summarized in Table II.
         Scop* of Applications
           Use  of the instrument for the continuous monitoring of
         particulatc mass  concentrations has been proven in the fol-
         lowing categories of emission sources:
           coal and oil-fired [rower stations
           residual fuel oil-fired industrial boilers
           Kraft recovery and hogged fuel fired boilers
           refuse incinerators anil teei>ce burners
           wet and dry process cement kilns
           catalytic cracker regenerators
           glass furnaces
           turbine engine test cells
           coke ovens
           electric arc furnaces

         When installed following electrostatic precipitators, the instru-
         ment precisely records the fluctuations resulting from precipi-
         tator rapping.  It indicates malfunctions and allows proper se-
         quencing of  rapping and adjustment of the precipitator for
         maximum ixrformance (fine tuning).1*
           The  instrument  system also monitors sodium losses on
         Kraft recovery boilers" and catalyst losses from catalytic
         cracking  regenerators.  The instrument's capability is with-
         out  equal for accurately measuring  low opacity values as
         required  for  cement and glass  plant emissions and turbine
         engine  tests.  In  monitoring the combustion process of oil
         fired  facilities or  waste incinerators, the  instrument can be
         reliably used for  the  automatic control of overfire air,  fuel
         additive, or other adjustments.
           Following  a fabric filter system, the instrument  clearly
         shows a surge of emission whenever a compartment comes
         back  on line after denning. A bag leak is signalled by a rising
         baseline of emissions but a sharp drop to normal levels when-
         ever  the compartment with the leaking bag is closed off for
         cleaning.
TaMe I. Error levels for correlations between optical density measurements and particulate mass concentration.
Error Level (mg/Nm') At Mean of Part. Cone.



Cement

Bituminous coal

Lignite

Municipal Incinerator

Kraft recovery"

Hogged fuel"

Bituminous Coal"


No. of
observ.
39

20

36

IS

12

13

11


Correl.
coeff.
0.985

0.760

0.915

0.909

0.957

0.750

0.974

Mean part.
cone.
(mg/Nm')
179.9

322.3

111.2

90.3

76.4

206.0

252.3


Std.
dev.
±13.9

±85.5

±31.4

±29.5

±12.6

±61.8

±56.4


95% Confid.
True mean
±9.0
(5.0%)
±40.2
(12.4%)
±10.6
(9.5%)
±16.4
(18.2%)
J16.1
(21.1%)
±39.0
(18.9%)
±38.4
(15.2%)

95% Confid.
Next single value
±57.2
(31.8%)
±184.0
(57.1%)
±64.6
(58.1%)
±65.7
(72.7%)
±57.9
(76.0%)
±145.6
(70.7%)
±133.2
(52.7%)
95% Confld.
95% of all
poss. values
±68.6
(38.1%)
±475.4
(147.5%)
±78.1
(70.2%)
±88.5
(98.0%)
±81.6
(106.8%)
±201.2
(97.6%)
±379.5
(150.4%)
170
36
                                                                             Journal of the Air Pollution Control Association

-------
                    LI|«M*
       I
       I
       8
0.10

                                                                  Table II.  Companion of stack opacities for different emission
                                                                  sources under equalized conditions.
                            0.10      (.is
                      Optlul DwMlty -

       Figure I.  Relationship between optical density and par-
       tlculat* mac* concentration  for a llgnit* fired boiler
       (standard conditions).
              •llu«4nou» Coal Find Boiler •nUtleno
400
200


«
•fl.os/y
71
i
M
'//
/

W
y













                   o.o*     0.1      e.tt     9.1
                    Oplteal D*n*lty - Slngl*
                                    •.II
       Figure 10.  Relationship between  optical density end
       participate  mass concentration  for a bituminous coal
       fired boHer (standard conditions).
   The instrument's continuous output signals for both opac-
ity and  optical  density, corre.spondinp; to particuluto  mass
concentration, can be used as inputs to a total emission moni-
toring system.  Other sensor inputs  required inohulc tem-
perature, humidity, pressure, and gas flow rate, as well its other
emission parameters such as sulfur dioxide and nitric oxide
content.  A small data processor  or controller can then cal-
culate mass emission data (Ib/hr or Ib/.MBtu) and applicable
averages (e.g., 10 min, 2 hr, and 8 hr averages) and provide
continuously updated printouts.  Such complete monitoring
systems will be used in the future.
   There are some limitations  to the applicability  of the on-
stack traiismissometcr.  Water  droplets in the gas are mea-
sured as if they  were paniculate matter.  Unless mist omis-
sions arc to be measured, efficient demisters must be employed
after wet scrubbers, or a portion of the gas must be reheated
prior to measurement in a by-pass duct.
   Similarly,  the  instrument cannot correctly measure the
opacity of emissions  which form visible plumes after the gas
exits from the stack (such  as sulfuric acid  mist  or vapors
which condense on contact with the cooler air).

 Soptwntoa* 1974    Voluitw 24,  No.  9


Emission
source
Cement
Bituminous
coal
Lignite
Municipal
incinerator
Kraft
recovery
Hogged fuel
Bituminous
coal
Mean
part.
cone.
(mgyNm1)
179.9

322.3
111. 2

90.3

76.4
206.0

252.3
Single
passO.D.
at mean
of part.
cone.
0.1029

0.0944
0.0580

0.0235

0.0746
0.1290

0.1432
Diam. at
measur-
ing site
(m)
1.95


3.25

4.00

3.66
1.52

4.88
Calcul. for
150 mg/Nm1 and
3 meter stack diam.


O.D.
0.1320


0.0722

0.0293

0.1200
0.1854

0.0523

%
Opacity
26.2


15.3

6.6

24.2
34.8

11.3
References

  1. K. Sick, (/. S. Palml \a. :{,fi17,7-VJ, issued November 2, 1071-
  2. K. W. Btthne, "Investigations into the directional dependence
    of photoelectric  smoke  density  measuring  instrument*."
    Staub 31: 22 (11171), (in Kngli.sh).
  3. "Stundfirds  of  j«?rfon, April 1072 (Knvironmental Research Corp.,  St
    Paul. Minn.).
  7. W. 1). Connor and J. II. Hodkinson, Optical Pro-prrtiet tmd
    Visual Effects of Smiikr Plumes, Knviroiunental  Protection
    Agency, Ollice of Air Programs Publication No.  AP30 (Co-
    operative Study Kdison ICIcctric Institute and Public Health
    Service, 1067).
 8. M. J. Pilat and  I). S. Kn.sor, "Plume opacity and  particulate
    mass concentration," At mm. Environ., 4: 103  (1970).
 9. IX S.  Knsor and M.  J. Pilat,  "Calculation of smoke plume
    opacity from paniculate air pollutant properties," J  Air Poll
    Control Asuoc., 21: 4!Nj (1971).
10. T. F. Hurley and 1). L. II. Bailey, "The correlation of optical
    density  with the  roncenl ration and  composition of  smoke
    emitted from a  Lancashire boiler,"  J. Intl.  Fuel. 31: 534
    (1058).
11. L. Diiwel,  "Coinpurative Sttidios of Different  .\fe*«uring
    Principles for the Continuous Monitoring of  ParticuUte
    Krnissions from Lipvitc Fired  Uoilers,"  Proceedings Second
    Int. Clean Air Congreti, Kdited by II,  M. Knglund and W. T.
    Berry.   Acaort,  1973.
1A. W. A. Schneider,  "Opacity monitoring of  stack  emiwioos:
    A design tool  with promising results/'  The 1974 etietric
    utility ... Generation  Plankook.  McGraw-Hill, New  York,
    1974, pg. 73.
                                                        Dr. Bcutner is a Vice President of the  Environmental
                                                      Technology Division of Ix;ar Siegler, Inc.,  One Inverness
                                                      Drive KasVKiiglewood, Colorado SOI 10.  Since 1971, Dr.
                                                      Ueulner has been responsible for the marketing and appli-
                                                      cations development, efforts for the division's pollution-
                                                      nicinitciring instrumentation.  His present responsibility is
                                                      business manager for Lear Siegler's air pollution control
                                                      equipment.  This is a revised version of Paper No. 73-169
                                                      presented at the With Annual Meeting of APCA at Chicago
                                                      in June 1073.  The author thanks Ernest E. Mau for his
                                                      assutance in the preparation of this poper.
                                                            37

-------
                      Instrumental  Method  Substitutes  for  Visual
                      Estimation  of  Equivalent  Opacity
                      Herbert C. McKee
                      Southwest Research Institute
Many air pollution control regulations limit the emission of visible effluents,  based
on the visual observation of "equivalent opacity."   Because of  difficulties en-
countered in wing visual observation, the Texas Air Control Board developed a
method of  calibration which made it possible  to use  an instrumental method for
measuring visible emissions.  A legal regulation based on this instrumental method
has been in effect for almost two years.   Despite 'minor difficulties  in calibra-
tion and  maintenance, results have been satisfactory.  The use of the instrumental
method avoids many of the difficulties inherent in using a regulation based on visual
observation, and continued use of the instrumental method is anticipated.
The regulation  of  emissions  of  dust
and other particular matter is a major
portion of most air pollution  control
programs.  This  is  usually  done  by
either or both of two different methods:
(1) regulations  based on  the weight of
material emitted, and (2) regulations
based on the optical pro|>crlies (opacity)
of the emission.
   The visual appearance of smoke emis-
sions  ha> bveu  regulated for  over 70
years by the  use  of the Kingclmann
chart to evaluate the density of gray and
Muck  smoke.1-*  More  recently,  the
Mime principle has been extended in the
control of emi-sions other than  gray or
lilark I'V Utilizing the concept of "equiv-
nlent opacity."1  It has been (Miniated
that between To and 101) governmental
control agencies now use this concept to
control visible emissions.  These  regu-
lations usually specify that the emission
should not obscure the view of an obser-
ver to a greater degree than the obseura-
tion which would be caused by gray or
black  smoke of  a specified density a«
measured by a Ringelmann smoke chart.
  To  aid  in   evaluating  equivalent
0|>acity, good correlation has been shown
between visual  estimates by a human
observer and   instrumental  readings
obtained  with   a   transmissomcter
installed in the staek of a social smoke
generator.4 This principle  has been
used to  train air pollution inspectors in
the estimation  of equivalent opacity.1
It has been shown that the errors inher-
ent in a subjective observation  can be
(•really  minimized in this way so that
visual estimates are sullieiently repro-
ducible to form a basi> fur lenal control.
This principle has also been upheld in
the courts.
.  Despite the use of smoke  generators
as training devices, however, there is
still some  uncertainty in making visual
observations  of equivalent opacity.
Furthermore, such observation* can be
made only with extreme difficulty at
night or during cloudy or rainy weather.
Therefore, the Texas Air Control Board
undertook  to   develop  an  alternate
method of measurement  which could
overcome at least some of these limita-
tions and could provide more accurate
and reproducible results.  This method
has now been in use for some time and
is specified by a legal regulation of the
Board  as  a substitute  for visual  e-ti-
mation.'
Principle of Method
  Since the reason for controlling emis-
sions on the basis of equivalent o|>ncity
is  to limit  the  emission of material*
which cause absorption or scattering of
light, the direct measurement of licht-
scattering pro|H>rties was selected »»tlie
basis  for  developing an  instrumental
method of  measurement.  Instruments
have l>een in u.-c for many years which
measure  light  transmission aciu.-«  a
stack a.s an aid in estimating the amount
of material emitted or the  density of
black  smoke.   Most of  these  in.-tru-
ment.s  include a liizlit source  and also a
photocell,  lioloinetcr, or  other optical
measuring  device: these are  located on
opposite sides of the stack in such a way
that the lidit U'uin pas«e.* through (he
stack  gases which are lienus emitted.
The amount  of  light  which  pa«fs
through the plume is measured, and the
result  is cxprc««cd  as percent transmit-
lance compared to a measurement with-
out stack gases present.
   Instruments of this nature are avail-
able from  several suppliers, and there-
fore it wn« decided to utilize this ty|x- of
4M
                                                                           Journal ol the Air Pollution Control Association

-------
 instrument as a Mubstitute f«w tlir vi-
 sual  estimation of equivalent opacity.
 By installing the instrument in a duct or
 stock, where the optical proiwrties of tlie
 fas  stream can be measured prior to
 leaving the stack, a continuous record of
 optical transmittance can IK* obtained.
 Since the  optical  propertie*  depend
 primarily on  the smaller particles pres-
 ent in the gad stream, considerations of
 isokinctic sampling do not ap|ky buck-
 ground,  emission condition*, or other
 factors which influence visuul observa-
 tions.

 Calibration
  In order to avoid error- due to < liaitur<
 in exit velocity or other emission  condi-
 tions, the |x»rmisstblc optical projicrties
 of the emission  were established on the
 basis  of  the total volume crnittvd.  In
 this way, I hi- total contribution  of the
 emission  to ihe optical propertii"* of the
community atmosphere was tin primary
controlling factor.
  20
 ho
i
        2      5         25
  10.000           100.000        l.OOO.OCO
            Volumt tmiiwd, elm

Flfur* 1.  light path required for 70% trint*
mitunce.

Augutt mi    Volume 21. No. 6
   The  inst-riiment re*>|>oit*e aUo varies
 de|>eiidingoit the light path across wliich
 the measiiremr-nt i.< mnilc.   I» eonven-
 ttoual in-tatlations, the  light path usu-
 ally is not extended the full width of the
 stack or duct in which the instrument is
 installed since  disturbances  in the flow
 pattern next to the  wall  might cause
 erroneous  readings.  To  calculate a
 calibration curve, an exit velocity of 40
 fps wns arbitrarily assumed so  that a
 direct calculation could  be made of the
 assumed  diameter of the  plume, and
 thus  e>tubli>h  the length of the light
 path  across  which the  required  trans-
 mittance  would  be  measured.    The
 required  minimum  transmittunce  was
 selected ns 70^, this twing somewhat
 more  restrictive than the iinual opacity
 limit equivalent to a \o. 2 Kingclniumi
 reading.  Figure  I  illustrates the  re-
 sults  of this  calculation, and >hows  the
 li^cht path fora minimum tr:m.»mittaiu'C
 of TOTi at any flow rate  between  10.000
 and   l.OOIl.LHJO  cfm.  This  volume  U
 biiM-d on tin-  total volume of gunc* emit-
 ted includins;  air,  combustion  ga>es,
 ga.seous impurities, and all gaseous mat-
 ter combined,  but excluding water or
 water va|»or, calculated  at the pitssure
 and temiK'ialure existing at tin' point of
 measurement.  This graph, thru. s|>eci-
 fiex the light  path across whii h the mini-
 mum  acceptable trati-mittuncc to deter-
 niiiie legal compliance a ~ta< k installatioos
 will not permit the UM- of this lenctli of
 tight |>ath. a  method wus mi'ilul to con-
 vprttlicTlX'i traiismittaucc loan eipuvn-
 teiit hgure over some other  li»bl path.
 This wa- ilniic by the u-«c of the follow-
 ing equation:
             -log T - U
                                          where
  T - trunsmittance
  I  - length of light  path
  k  — con>tant

In  this equation, k can be evulunU-inittunce for some  other
length of  light  path.
  Figure 2 is bu.Mt  pnth
olln-r than the length "(x-cified by Figure
I.  E\|N-nence  lia.s shown  thai  tlui
cc|iiutiossible that extrapola-
tion beyond these limits may introduce
some degree of error.
Precautions
  Certain precaution.-! should be kept in
mind  in the  use  of  thin  instrumental
method.   For  example, gases high  in
moisture content should  be measured at
a |>oint where the gas  temperature is
ulkive the dovv point to prevent eotideu-
sation; otherwise, loss of light truiismit-
tance mny occur due to  the presence of
condensed  water  droplets in  the gas
stream.  Certain  situations may  also
ari.se  in  which  the  method  cannot  be
used satisfactorily, such as the following;
  t.  Emissions may  change  in  char-
acter after emission to hvconu.1 more vis-
ible (so-called "detached  plume"),  so
that an omission which is invisible in the
stack may in fact cause deterioration of
visibility  in  the  ambient atmosphere
following emission.
  2.  Different omissions may mix and
react in the atmosphere to  form aerosols
which restrict visibility, although each
emission may be  essentially  invisible
when first emitted.  A familiar example
is the formation of a highly, visible aero-
sol  of ammonium  chloride by the reac-
tion of ammonia with either eliloriue or
hydrogen chloride.
  U'hore ilu-so or other coiulivum* make
it impossible to UM; this method. visual
iiis|M-ctiim or some other alternate must
lie found.  However, so far these excep-
tions have  occurred only infrwHirntly.
Illustrative Example
  Tluoimh the cooperation of the Dow
Uii-micnl Cuini'iiny, dutu lire prex-nu-d
lirtr in .|in\v a typical illu*tration of (hr

                                   Ot
                                                        40

-------
u*e of this method of measurement.  An
instrument supplied by the Bailey Meter
Company was installed  in  the stack
carrying flue gases from a series of lime
kilns  at the Freeport,  Texas plant of
Dow  Chemical Company.  Figure  3
shows a typical chart recording during a
24-hr JHTKM!, illustrating the minor var-
iations in transmittancc during this time
when the average  transmittance across
the 5-ft light path of the instrument was
around  90%.   Note that the recorder
wad set to record in the range of 70 to
100% traiismittanec, with 70% at the
periphery of the chart and 100% at the
center.
  Total volume of xases  emitted  is
approximately  130,000 cfm,  although
this naturally varies to some degree with
changes in  operating  conditions.   At
this flow rate. Figure 1  indicates that a
path length of 8.2 ft would he correct for
measuring the  required 70% transmit-
tance.  Since   the   instrument   was
installed with a 5-ft light path, Figure 2
was  used to determine the equivalent
reading with this installation; this indi-
cated that the transmittance should not
be less than approximately 80%, which
corresponds  to a 70%  transmittance
over a light path of 8.2 ft.   Thus, the
90%  transmittance shown in Figure 3
illustrates better precipitator perform-
ance  than  the minimum required  to
meet  the regulation.   This instrument
has been  o|>erated for several  months
and has provided a convenient means of
monitoring the o|>eration  of the  precipi-
tator to detect any minor changes indica-
ting a need  for maintenance or adjust-
ment.    The only instrument  mainte-
nance has been an occasional cleaning of
the windows on the liuht source  and  bo-
lometer, and  instrument stability and
reliability   have  been  re|x>rted  a*
satisfactory.
Advantages  of Instrumental Method
  In  the past two  years,  40 to 50
instruments have been installed  in vari-
ous industrial plants in Texas, for use in
meeting this now regulation.   Processes
include lime kiln*, cement plants, fertil-
izer  driers,  regenerators  in  catalytic
cracking units, and others.  liased on
the experience  which has  been accumu-
lated, the instrumental method  appears
to ixissess the  following advantages:
  1.  The instrumental mctln.nl is com-
pletely  objective and therefore is not
subject to the errors inherent in subjec-
tive evaluations by human obr-crvers or
to variations among duTctvnt  individ-
uals making vi-ual estimates.
  -.  The instrumental method  appear*
to IK- more accurate than visual nbiserva-
lions.   With adequate maintenance anil
calibration of the instrument, it  ap|x*ars
to be po-sible  to maintain ,tn accuracy
of a few  percent.   This  is  ln-tter than
the accuracy  generally obtained with
visual estimation, e*|>ecially with high

1*0
       I.  Twenty-four  hour chart  record,
showing typical MntmitUnc* m»i»uf«m«nt.
transmittance values above 60  percent
which are required by many regulations
now being written.  Previous  experi-
ence has indicated that an equivalent
opacity of Ringclmann No.  1 is more
difficult to estimate than No. 2, which
means  that transmittance  values  in
excess  of 60 to 70% may be subject to
considerable inaccuracy if measured by
visual  estimation.4
  3.  An automatic instrument can be
used at all times, and the results are not
affected by sunlight, cloud cover, dark-
ness, or other conditions affecting visual
observations.   Readings  can be ob-
tained on a 24-hr basis.
  4.  Continuous measurements can be
recorded automatically at less cost than
the cost  of frequent observations by
trained human  observers.
  5.  Since results arc available on a
continuous basis, they can be used as a
means of process control for continuous
processes.   Thus,  the  plant operator
can be aware of any malfunction of pol-
lution  control  equipment immediately
in  order  to take  necessary  corrective
action.  This is not possible with regu-
lations based  on  the  weight of  dust
emitted since stack sampling to measure
weight  if  tedious and  time-consuming
and subject to a time lag in  obtaining
measurements which may be as much as
several days.
  Thce.\|>erieiire todateliasnot revealed
any scriou- problems in the operation or
maintenance of the instruments.  The
major problem i» tiiat of rechecking cal-
ibration on production units  which are
Ojieratcd continuously for many months,
and various ways to accomplish this are
uniler study,  (ilass surfaces exposed to
flue gnse>  niu«t IT clcam-d oecasitiMully
although tin- can U1 mlm-ed by the use
of an air bleed which  mmiimio the con-
tact of dii»t laden ga«cs with I he window
surface!*.    Plant   operators  have re-
ported no uiiii-ual maintenance problems
winch  cannot  U-  bundled adequately
by the average plant instrument  depart-
ment.
  Because of the  advantages of  this
method, the Texa.s Air Control Board is
considering making this method manda-
tory for all industrial sources subject to
regulation on  the  basis  of opacity,
except for very small o|>erations where
the cost of installation ami maintenance
of an iiistnimpjit would be out of propor-
tion to the tola-fit  received.   A figure
of 10,000  cfm has been mentioned as a
minimum size for  mandatory installa-
tion, but this factor is still under consid-
eration.

Acknowledgment
  The author wishes to express appre-
ciation for the assistance of the Texas
Air Control board  staff in the develop-
ment  of   calibration procedures  and
other details in connection  with the new
method which is outlined.  In addition,
appreciation  is  also expres-ed  to the
Texas Division of Dow Chemical Com-
pany, for (icrmissiou to describe the lime
kiln installation which illustrates the u.*
of this method, ami for the illustration
furnished.

Reference*
1. Kmuelmann,  M.,  "Method of  esti-
   mating smoke  produced  by industrial
   installation*,"  Rtv.  Tirkniiiuf,  2fr»
   (June ls».«0.
2. U.  S.  Department of  the Interior,
   Bureau of Minrn Information Circular
   7718 (AiiBiist l!».V.J.
3. State of California. Health and Safety
   Code, Chanter 2,  D»vi>i«n 20, Section
   24-.M-.' fl'.il7>.
4. Conner, W. 11 and 1 lodkin-on,  J. H.
   Optical  Propirtim  anil  \'i3tial  Efftrt*
   of Smoki-Stark Pttimr*, PnMiration No.
   »99-AI'-:JO, f. ?s. I>«bli<: Health Ser-
   vice Uilfi").
5. Coons, J.  I)., Jame-, II. A., Johnwn,
   H. C.,  and \Vnlker, M  S.,  "Develop-
   ment, Calibration, and I'-r of :i Plume
   Evaluation Training t'nit."  J- Air
   Poll.  Canlntl  .4.««»-.,  15,   I!W  (May
   1%:»1.
0. HfCiilatii>n I.  "Control of Air Pol-
   lution from Srnukr. \'isihlc Knn—ion-
   and  lsii>pfii.
      Dr. Mi-Kee, »bi« i> A^Mnni Di-
    rvi-tur.  Departmeut  of  t'hein^try
    11 nd Clu-niiiul  KiitsiiiiTring nf the
    Soulhwesl   l!(-earrli   In-nniir —
    I IIHI-IMII. ;iNn .<>rr<— ;i- < 'K unii.in .if
    ilieTi'XU' Air Coiiirol |ln:iri|.  Tlii«
    |I:>|»T w:i- pre-entiil at ll«- 1'i.lnt An-
    iiuat MII-IIIIV  ..f Tl»- Air Polliiti.ni
    Coiiirol A->i'c. ii- Paper  NIL 7n->-4.
                 41
                                                                               Journal of th« Air Pollution Control Association

-------
           Reprinted from POLLUTION  ENGINEERING,  November  1978
   Opacity  monitoring  technique
   predicts baghouse  maintenance
   JULIAN SALTZ and LEE COTLER
            When  a  baghouse is  operating properly, typical
          collection efficiencies better than  99 percent can  be
          expected. However, if a bag becomes worn or if a break
          occurs, the emissions will increase, often above the legal
          allowable level. A major bag rupture can also allow a
          substantial quantity  of valuable product to  be emitted
          and lost  to the atmosphere. Therefore, it is important
          for several reasons to know the instant a break occurs.
            A time delay device can be installed in  an opacity
          monitor  and  can be set so that  a rupture can  be
          detected  from 0.5 sec to 2 min after occurrence. A relay
          in the opacity measuring control unit can then be used
          to disconnect power to the ID fan of the baghouse,
          • thereby preventing product loss and opacity violation.

          Opacity and mas* flow
            An opacity measurement is taken by projecting a
          well-collimated light beam across  the  diameter of a
    t'igure I. Common  duct of  a  5-compartmtni baghouse. Opacity
    instrument is located under the triangular weather cover.
    Julian Salt: is president of Datatest. Inc., designer} and manu-
    facturers (ij instruments for measuring air pollution. He has a
    H.S.t.h'. degree and lias an e\tensive background in instrumen-
    tal ion design.

    l.ce (.'oiler is a technical specialist with Datatest, Inc. He has
    many years of experience in manufacturing and engineering in
    the electronic.1; instrumentation field.
process outlet pipe or exhaust stack. Particles in the
effluent will block or absorb part of the light, thereby
reducing the amount of light that reaches a receiver
unit.
  In addition, it is possible to relate the mass  flow
(g/scf) to  opacity for a specific process.1 In a study
made on a municipal incinerator in Minneapolis,2 the
relationship of mass flow to opacity was found to be:
Opacity
where:
    ity= 1-L
 light with blockage
light with no blockage
100
           w =
40
                                                     43
       NV= paniculate mass concentration
           of the plume, g/m3
       Ki= plume parameter. cmj/mi
        P = average paniculate density, g/cm!
        L= plume diameter, m
       OP= plume opacity

  The  value  Ki for the incinerator operation varied
from 0.33 to 0.5. For this operation, the opacity for a
mass flow of 0.05 grains/scf varied from 8 to 12
percent.

Baghouse monitor
  The  instrument shown in Figure 1 is used to monitor
PVC emissions from a baghouse with a capacity of
50,000  cfm. The control unit is approximately 200 ft
from the receiver at ground level. All adjustments, such
as zero, span and set  point alarm, are made at the
control unit.
  Using this technique, schematically shown in Figure
2. the opacity of the effluent is measured to determine if
it is within acceptable limits. If a bag develops a  hole,
the opacity insirument will detect it. The output of the
opacity meter is connected to a recorder and a day-by-
day record is made of  the effluent opacity. Each time
the bag is cleaned, a small increase in the emission level
occurs. The peaks of these emissions may be observed
on the recorder. By monitoring the signature spikes of
an  individual compar'ment  and observing the
progressive increase each time the compartment is
cleaned, it is  possible  to see a hole enlarging or the
effects of increasing porosity of the bags.  Using this
technique, it  is possible to predict weeks ahead of time
when  baghouse maintenance is required.

Operation and control
  As shown  in Figure  2, the electrical signal from the
photocell in the receiver is divided  by the signal from

                             NOVEMBER 1978

-------
the light  source in the transmitter.  The resultant  is
subtracted from 1 and  this is the opacity signal which
goes to the meter and recorder. An alarm level may be
set for any value of opacity from 0  to  100  percent.

Low level measurements
  The baghouse has a unique characteristic that allows
for low-level  measurements of opacity.  If  there is a
small hole  or  opening in  a bag, or if  the  bags have
become thin or porous, the baghouse cleaning process
will release some of this material into the outlet pipe.
This emission  may be  very small, in the order  of 0.1
percent opacity, or above 0.0005 g/ft3. It is  impossible
to obtain such a low level measurement  using standard
opacity measuring methods.  However, a  high  mass
electrical  filter was connected to the  output  of the
opacity system shown in Figure 2, along with suitable
impedance matching circuits. Using this technique, drift
due to soiling of  the  optical  lenses exposed  to the
effluent  was eliminated and  low-level  measurements
were made as shown in Figure 3.
  In  this  particular  case,  five  baghouse  units  were
connected  to  a single  pipe which the  opacity  meter
spanned.  The  4-in. chart is calibrated for  10 percent
opacity full scale or 0.1  percent per small division. Note
the variations   from compartment  to  compartment,
from a low of 0.3 to 1.4  percent.
  As  the spikes on the  strip chart recorder increase in
amplitude, a bag  problem  is  being  reflected. In this
sequentially pulsed, multicompartment installation, the
amplitude  increase pinpoints  the compartment  with
problems.
  Therefore,  operators can   now  use  this  opacity
measurement technique 'to predict future problems such
as bag blowouts or excessive emissions. Equally impor-
tant, this  technique can be  used to measure extremely
low levels  of materials  which are hazardous or toxic.
Bibliography:
I.  Ensor and Pilot. "Calculations of Smoke Plume Opacity from
Paniculate Air Pollutant Properties." Journal of Air Pollution
Control Association, No. 3, 1971.
2.  Metropolitan  Waste  Control Commission, Minneapolis,
MN. Report No. 7065.
Figure  3. Low level measurement of effluent from a 5-compartment
50,000-cfm baghouse. Note increasing measurements in comparments I,
2 and 4. Steadily rising levels would indicate a bag beginning to break or
porosity increasing.
  10 i
               Baghouse Compartment Numbers
     Compartment Numbers Cleaning Cycle - 1 in./ hr Chart Speed
 Figure 2. Electro-optical diagram of an opacity instrument.
 POLLUTION ENGINEERING
                                                        44
                                                                                                                       41

-------
 AIR/WATER  QUALITY |
          Continuous  onstack   monitoring
          of  participate  proves  feasible
New  optical  instrument opens  way  to
continuously  and  accurately  monitor
particulate and  visual emissions from
kraft  recovery and hog-fuel boilers

•y J. C.CRISTELLO and J. E. WALTHER

D Continuous monitoring of particulate emissions from
boilers  is  becoming important for a number  of reasons.
First, a  continuous and  accurate record may be a necessity
in proving that boilers  are in compliance with visual and
mass emission standards. Secondly, new emission control
equipment can readily be evaluated to determine if it meets
the manufacturer's specifications. Lastly, continuous moni-
tors can aid boiler operators in optimizing the performance
of the boiler to reduce emissions.
   At one time, stack emissions were primarily graded by
visual measurements of smoke emissions by the Ringelmann
method. The Ringelmann text subjectively measures smoke
density  by visually comparing the darkness  of a plume
against  shades of gray, representing different-levels of
opacity. Most air pollution standards now specify the limits
of stack emissions in both visual and mass-volume concen-
tration  terms. Thus, there is a need to measure particulate
emissions with an instrument capable of accurately correla-
ting  particle  concentration and visual observation  with
smoke density.
   During 1974 a study  was  undertaken  to evaluate an
onsluck instrument capable of continuously  monitoring
stack emissions from  kraft  recovery  and wood  waste
boilers.  The test instrument evaluated was the Lear Siegler
Inc. (LSI) Model RM-4  transmissometer.1 Some of its most
significant features  are: (1) automatic builtin  calibration,
(2) double-puss measurement  of the gas over the  entire
stack, (3) and virtually maintenance-free operation.
   Since the LSI measures across the entire stack diameter,
it is necessary to focus  and zero the unit for each different
path length.  Mounting and adjusting  of the instrument
Mr.  Cmtello was  senior  process  engineer  with Crown
Zellerhach Knvironmental Services, Camas.  Wash., and is
now with Stevens,  Thompson A. Runyan Kngineers Inc.,
Portland, Ore. Dr.  Walther is supervisor of air programs.
Crown Zellcrbach  Knvironmental  Services,  Camas.  The
following is  bused on  a  paper presented at the  1974
PNWIS-AK'A show, Boise, Idaho.
                                                45
require about four hours. All of the initial work reported
was done with  the  transmissometer located at  the site
normally occupied by  the mill's  bolometers or opacity
meters which utilize  a  slotted tube. This means that the
actual double-pass measurement distance was limited to 10
ft even though the flange-to-flange distance was greater. All
data reported in this paper are based on a total measure-
ment distance of 10ft.
  Particulate sampling  was done  using either an instack
Hlter,  a  commercial EPA train or wet  impingers. The
instack  filter utilized a 70  mm, 0.3  micron, Type A,
glass-fiber filter. The National Council for Air & Stream
Improvement (NCASI)  has demonstrated  that results ob-
tained with the instack filter and the EPA train2 and those
with the  wet impinger, 600°C residue, and the EPA train3
are equivalent. Sampling procedures used with the instack
filter and the wet impingers were similar to those described
in ASME Test Method PTC 27, while those used with the
EPA  trains were similar  to  EPA Method 5. The major
variation from these methods involved sampling  at the
point of  average velocity rather than at several points. The
LSI RM-4 instrument  reports measurements in optical
density units.

Boiler tests

  The first test  site was choien at mill A with an 800-tpd
B&W kraft recovery boiler. A  three-field precipitator rated
at  98% efficiency preceded the transmissometer location.
Particulate samples were obtained  at the main stack, just
prior to  the bolometer  ports  using the  instack filter with
wet impingers. A plot of optical density versus particulate
concentration is shown in Figure 1 along with the 95%
confidence limits.
  The second evaluation was  at mill B for a B&W hog-fuel
boiler rated  at 300,000 Ib steam/hr at 600 psi. Particulate
control is provided by a cyclonic dust collecting unit. The
duct  out of this unit is approximately 20 ft wide at the
point where  the transmissometer was installed  and  the
particulate samples were taken. Three sampling ports were
available, each of which had a port at a quarter point. The
ports are perpendicular to  the bolometer tube and  the
middle port corresponds to the slot in the tube. Separate
particulate tests were made at  each port.
   A statistical analysis of the particulate measurements is
presented in Table  1.  This  analysis indicates that  the
particulate concentration is not uniform across the duct.
Therefore, to continuously, accurately monitor the emis-

                             PULP 4 PAPER MAY 1975

-------
      0.1
      0.7
   _0.6
   o
   g
      0.5
      0.3
      a?
      0.1



00 -O.WX -CM
!»*•«. r*

. *
//
•W
//
t
t
t

v/
•' / •'
///
'/s
y*


i


t
*
t
-
y,
*
t
f


legend
' ^
S









O InstKk filter
& Wei Impinqer.
600 C Rtiiduc
-~95*
Confidence limits
1

              01    0?   0.3   0.4     OS    06    07
              Front half particulate (X), grains/ltd cu ft
   Figure 1. Mill A recovery furnace. LSI optical density
   vs front half paniculate.
                                                                                               Middle Port
                                                                                             00 • 0.64 X -f 0.004
                                                                                                Rz • W. 5%
                                    o West Port
                                    O Middle  Port
                                    A East Port
                   o            0.10           a 20            a 30
                       Front half paniculate (X), grains/ltd cu ft
            Figure 2. Mill B hog-fuel boiler. LSI optical density vs
            front half particulate.
        0,25.
        0.15
        0.10
        0.05
                       	1	1	/ 7
                       00-1.ZKX0*.039    ,' /   -
                                              /   '


* / '
// /
,r


/












                                   Double Ptjs
                                   Path Length
                               UGtNO
                               95%  tonlidence
                                 Interval
           0       a 05      010      0.15       0.20
               F'om hall (MMiculatl   ui l"(i I  v-dito (In1 v,ct hulh lompera-
tine w.i\  .I)VMII  I >'• I   I nli.niirJ wjl.'i JiJ not appear  to
           be  a  problem because the measured  water content was
           approximately that predicted  by wet  bulb-dry bulb  tem-
           perature measurements. The particulate sampling train used
           for  this study was a commercial EPA-type unit. A plot of
           LSI optical density versus paniculate concentration for mill
           C is shown in Figure 3.
              A  highly  significant correlation between  LSI optical
           density and  particuhitc exists. Therefore, the LSI is suitable
                                                            46

-------
for use PS a continuous mass cnm\inn monitor. This implies
that the  need lor traditional  particulate  tests  would  be
reduced to an occasional but periodic check to verify the
instrument calibration curve.
   As shown  in  the appendix,  the optical  density and
opacity  are  merely  mathematical transformations of the
same  parameters.  Therefore, a good correlation betwe  n
optical density and participate assures an equal correlation
between  opacity  and participate  concentration (Table 2).
However,  a significant correlation between the LSI  trans-
missomcter opacity and purliculatc concentration docs not
yield  any  information  concerning  the  accuracy of  the
opacity  value. Accuracy of the instrument opacity would
be determined by correlation with the opacity determined
by a trained smoke observer.
   Based on the data presented and  casual  observations of
the plumes for mills A and C, we  were quite concerned
about the accuracy of opacity measured by  the instrument.
After  having  the  instrument checked out by  a factory
service  representative,  it was  decided  to remount  the
transmissometer at mill C in a location higher up on  the
stack. This alleviated any unrecognized problems associated
with using the  bolometer.
Appendix: Optical characteristics defined
The  transmittance, T, is defined as the  ratio of the  light
flux  out of the plume. I. to the light flux into the plume,
                             .
                             r
   The  transmittance  is  a  function  of the  particulate
concentration, C, and attenuation coefficient, K, and of the
measurement path length. L. For poly-disperse particle size
distributions, K tends to be insensitive to small changes in
the size distribution.
                      T..-KLC              (2)
   The opacity, 0, plus the transmittance is equal to unity.
                       9 + T - 1                (3)
   The  optical .density,  OD, which is a measure of the
decrease in visibility over  some distance is inversely related
to the transmittance.
                      00 -..g^               (4)
   By combining Equations  2  , id 4, it can be seen that the
optical  density  is  also  a 'function  of the  particulate
concentration and the measurement path length.
                    OD-KLCIqge            (5)
   It  has been our experience that this definition of optical
density is  not  a density  in the usual  sense. Since optical
density is dependent on the  total measurement path length,
this definition  tends to create confusion. It may  be more
appropriate and less confusing to redefine optical density in
terms  of  a  true  density   that  is independent of  the
measurement path  length and has the units of length per
unit length.
                    OD0 - K C log e            (6)
   If  the optical density at the stack exit, 00, , is desired,
the measured optical density is adjusted to account for the
difference  in measurement path lengths between the point
of measurement, L, and the stack exit, L,.
OD
                       .
                           L
(ODI
                                               (7)
   Equations 3. 4, and 7 may be used to find the in -stack
opacity at the stack exit.
                 9. - 1- (loo0 (t) (ODr1        (8)
                              !••
   It  is important to  remember  that the  opacity as
measured within the stack may  be significantly different
than that observed out of the stack.
Observation numburi

Number of observations
Mean visual opacity
Mean LSI opacity
Mean difference
Standard error of mean difference
t-test
1-62
60
41.4
47.9
6.5
0.73
8.89
65-141
74
34.7
24.1
11.6
0.56
19.14
Table 3: Hog  fuel boiler: comparison of opacity measure-
ments made visually and with LSI transmissometer.


Determining accuracy of instrument measurements

   Thereafter  data  were  collected  to determine  if the
instrumental opacity measurements were accurate by com-
paring instrument and visual opacity  readings as made by a
trained  smoke observer (Table 3).  The EPA states that
acceptable  errors for certification  of smoke observers is a
maximum of 15% and an average of 7.5%.4
   Opacity measurements made visually are dependent on
the plume dimensions and color, the relative position of the
sun  and  the   observer,  and  the background  that the
observations  are being  made  against. At  the  time the
measurements were first taken, the sky was clear, the sun
was to the observer's back and up in the sky; while at the
end of the observation period  the sun was on the horizon
and at a horizontal  angle of about 120° to the observer. It
is  known  that visual opacity measurements increase as the
angle  between  the sun and the  observer increases. This
phenomenon was apparent in our observation. Considering
the inaccuracies of making visual readings for opacity and
using the data  presented for mill C, it can be concluded that
the transmissometer accurately measures opacity.
   The EPA has recently published proposed requirements
for continuous monitoring  of stack  gas  opacity to be
recorded "and submitted in terms of  the magnitude and the
number of one-minute integrated averages which exceed the
applicable  standard."4  Needless to say, this kind of data
reduction would require  a computer for data processing.
Furthermore, continuous monitoring of opacity/paniculate
will  in effect serve  to  make current regulations  more
restrictive due  to a shorter time interval for measurement.
Consideration  must be given to modifying standards based
on traditional  particulate  tests  of one  to two  hours
duration.
   The  accurate and  rapid response of the LSI unit also
makes it well-suited for special studies.  As an example, we
installed the RM-4 on a recovery boiler, after a precipitator
but before a wet scrubber. This particular precipitator had a
history  of subpar operation due  to what we believe was
nonoptimal adjustment  of precipitator  operating para-
meters. The  RM-4  made it  possible  to  optimize these
parameters in a short  period of time, without performing
standard particulate tests.                             Q
References
1. II. P. Beutner, "Measurement of opacity and paniculate emis-
sions with an onstack transmissometer. Air Pollution Control Assn.
Journal 24:865 (1974).
2. National Council for Air & Stream Improvement. "Comparison
of  source  particulate  emission measurement  methods at krjfl
recovery furnace stacks." NCASI  Atmospheric Quality Impnivt
ment Bulletin No. 67. October 1973.
3  National Council for Ait & Stream Improvement. "Companion
of source particulate emission measurement methods at  lime kilns
and  smelt  dissolving tank vents." NCASI  Atmospheric Quality
Improvement Bulletin No. 64, May  1973.
4. "Stationary sources: proposed emission monitonnn and perform-
ance testing requirements," hejeral Register 39:32H$2. September
                                                         47
                                                                                             PULP ft PAPER MAY 1978

-------
IMPACT OF SULFURIC ACID EMISSIONS ON PLUME OPACITY
                        by

        John S. Nader and William D. Conner

       U.S. Environmental Protection Agency
    Environmental Sciences Research Laboratory
        Research Triangle Park, N.C. 27711
                  (919) 541-3085
                   Presented at

       SYMPOSIUM ON TRANSFER AND UTILIZATION
         OF PARTICULATE CONTROL TECHNOLOGY
                 Denver, Colorado
                 July 23-28, 1978

-------
          IMPACT OF SULFDRIC ACID EMISSIONS OH PLUME OPACITY
                  John S. Nader «nd William D. Conner
              Environmental Sciences Research Laboratory
                 U.S. Environmental Protection Agency
                Research Triangle Park, North Carolina
                             INTRODUCTION


     Emission standards for opacity and for mass concentration of par-
ticulate matter have been established for new sources and are applicable
to fossil-fired combustion sources1.  Continuous monitors for opacity
(transmissometers) are required to be installed on these sources to
verify the maintenance and satisfactory operation of control systems
used to meet the emission standards2.  Reference method 9 is the observer
method of measuring the opacity of the plume which forms as the parti-
culate matter exits the stack1.  The Lidar J^ight detection and ranging)
technique is an electro-optical instrumental technique to remotely measure
the opacity of the plume3 "*.  The transmissometers are installed in the
stack or in ducts leading to the stack.  These monitors measure the
opacity of the gas stream in the stack or duct prior to its exiting the
stack.

     The Stationary Source Emissions Research Branch (SSERB) of the
Environmental Sciences Research Laboratory has had in its program during
the past three years tasks to generate a data base of concurrent measure-
ments of in-stack gas opacity (Os) and plume opacity (Op) for emissions
from various industries.  The purpose of these measurements was to
identify those industries wherein the plume and in-stack opacities do
not agree.  Measurements conducted to date on combustion sources burning
coal with sulfur 12% show that Os and Op are comparable5 6.  These
results would imply that no significant (observable effect on opacity)
physical or chemical transformation was occurring in the contents of the
gas stream as it was transported through the stack.
                                  51

-------
     Similar opacity measurements were made on a power plant burning oil
with 2.4% S (sulfur) and 200-600 ppm V (vanadium).  In contrast with
opacity measurements made at sources burning coal or oil of lower sulfur
content the plume opacity was found to be significantly higher than the
in-stack opacity.  At this oil-fired power plant a concurrent study was
being conducted on sulfuric acid emissions.  The results of this acid
study support the conclusion that a physical transformation occurs as
the gas stream exits the stack and enters the atmosphere.  The following
phenomenon is indicated:  The sulfuric acid is above its dew point at
stack temperatures in excess of 150°C and does not affect the in-stack
opacity. When the gaseous sulfuric acid leaves the stack and is cooled
to ambient air temperatures which are below its dew point, it condenses
and the sulfuric acid droplets increase the plume opacity.  Additional
studies have been conducted and are on-going to obtain more data and
understanding of the effect of sulfuric acid emissions on plume opacity
for various operating conditions, fuel composition, and control systems
for a number of fossil-fuel-fired utilities.  This paper presents and
discusses the results of the above work that SSERB has conducted thus
far.

     The sulfur oxides potentially present in the stack gas stream at
temperatures above the sulfuric acid dewpoint are as shown in Figure 1.
The free HjSOi, and SC>2 are not sensed in the measurements of the opacity
of the stack gas stream.  In the plume with the temperature of the gas
stream dropping below the acid dewpoint and approaching ambient air
temperature, the free ^SOit condenses to form acid droplets.  The con-
densed acid droplets and the acid adsorbed on the fly ash add to the
opacity of the plume.  In our studies 03 was measured by transmisso-
meters and Op was measured by human observers or by a Lidar system; in
some instances, Op measurements were made by both of these methods.
                                52

-------
                      COMBUSTION SOURCE FEATURES
     Plume opacity measurements were conducted In conjunction with emis-
sions characterization studies at 2 oil-fired and 3 coal-fired power
plants.  Table 1 summarizes the physical and operating features of the
plants.

     There is a marked difference in composition of the fuel utilized
in the oil-fired sources in contrast to the coal-fired sources.  The
ash content of oil was two orders of magnitude less than the ash content
of coal.  The sulfur content of the coal was from 2 to 4 times the
sulfur content of the oil.  In addition, very high vanadium concentra-
tion (590 ppm) was found in the Venezuelan oil.  Excess boiler oxygen
was typically in the 3 to 5% range except for Plant A which operated
at very low oxygen levels at about 0.2%.  Oil-fired sources had no emis-
sion controls; however, fuel additives were used to minimize corrosion
problems and did provide some reduction in sulfate emissions7.  Coal-
fired sources had either particulate emission controls (electrostatic
precipitators, ESP) or both particulate and gaseous emissions controls
(two-stage vet scrubbers).
                                    53

-------
                          SAMPLING LOCATIONS
     Sampling locations for all in-stack measurements were at a common
location between the emission controls and the stack except for Plants M
and LC.  At Plant M all in-stack measurements were at a common location
in the stack proper.  At Plant LC in-stack opacities were monitored in
breechings leading into the stack and all other in-stack measurements
were at the output of one of eight scrubber modules that operated in
parallel for the total boiler output of 820 MW.  Each module in effect
handled about 100 MW of power output.  Except as noted, plume opacities
were measured within one stack diameter of the stack exit.
                                  54

-------
                                RESULTS
     Emission data were obtained on particulate concentration, S02,
and sulfates under various operating conditions of the boiler (excess
02) and of the control systems (cutting back on electric fields of ESP).
Opacity data, however, were obtained concurrently for a limited number
of operating conditions.  The Lidar systea was inoperative and under-
going repairs during the studies on the coal-fired sources.  Plume
opacity data in these instances were limited to Method 9.   Poor weather
conditions (fog) also restricted plume observations during the study at
Plant LC.

     Data from the various plant studies on emission concentrations of
gross particulate matter, SC>2, total water soluble sulfates (SO"),
plume opacity, and in-stack opacity were reviewed.  As much as possible,
data were selected for those periods of time when these measurements
were made concurrently.  Tables 2, 3, and A are a consolidation of these
data.  Table 2 summarizes the emission data for oil-fired power plants
without any emission control systems, Table 3, for coal-fired power
plants with ESP controls, and Table 4, for a coal-fired power plant with
2 stage wet scrubbers (particulate control plus flue &as desulfuriza-
tion, FGD).  Various constraints such as number of available sampling
ports in a given location did not permit the desired measurements to be
executed concurrently at Plant LC and this is reflected in the data in
Table 4.  These data represent sampling visits to this source on three
different dates.

     Figure 2 graphically portrays the data of Table 2 showing the com-
parison of t!ie plurcc opacity data with the in-stack opacity data at the
oil-fired power plants.  Figure 2 also provides data on the effect of
additional condensation on plume opacity with cooling of the plume down-
stream of the stack exit location.
                                  55

-------
                              DISCUSSION
     At the oil-fired power plants M and A (Table 2) the most distin-
guishing features are the vanadium content of the fuel oil, levels of
excess 02, and their impact on plume opacity.  Plant M was burning
domestic fuel oil with about 15 ppm V and 1.2% S compared to Plant A
which was burning Venezuelan fuel oil with 590 ppm V and 2.4% S.   Both
plants utilized fuel additives.  A significant difference between plume
and stack opacity existed for both fuels and for different levels of
excess Q£.

     The plume opacity at Plant A at normal operation was 31% at the
stack exic and above the opacity emission standard of 20%.  The opacity
Increased markedly to 43% further downstream as more condensation of
sulfuric acid occurred with cooling of the plume.  The observers tended
to read higher opacity values than the Lidar.

     Both solid particulate matter and condensed sulfuric acid (liquid
droplets) affect opacity.  Unfortunately, concurrent data are not com-
plete in Table 2, but qualitative data from acid emission measurements
in other studies' 910 reinforce some observations which can be made
from Table 2.  The concentration of solid particulate matter increased
as excess Ofc  for combustion was decreased to reduce acid formation.
We attribute this to unburned carbon soot particles which have been
observed in a related study11.  The increase of in-stack opacity with
decrease in excess 02 demonstrates this.  Since the acid is in the gas
phase in the stack, a decrease in acid with a decrease in excess 02 does
not affect the in-stack opacity data or the above interpretation of the
data.  As a matter of fact, the possibility of reduced sulfate salts
exists with the decreased acid and this would tend to counteract the
effect of the unburned carbon.  This would imply a greater impact of
acid on stack opacity than was actually observed.

     In the plume, the acid (at temperatures below its dewpoint) appears
as liquid droplets after condensation.  With increased excess 02, the
increased acid and sulfate salts tend to increase the plume opacity but
a counteracting effect is the concurrent reduction in unburned carbon
resulting from more complete combustion of the oil.  Since the condensa-
tion of the acid is a function of the plume temperature, one can infer
                                   56

-------
on a semi-quantitative basis the contribution of the condensed acid in
the plume opacity relative to that of the solid particulate (both
unburned carbon and sulfate salts) by a comparison of the plume opacity
at the exit to that downstream of the exit.  At 0.22 02 the plume opa-
city increase was from 31 to 43%.  At 0.6% 02 the increase was from 23
to 54X, indicating the presence of more acid at higher excess 02 levels.
The impact of the acid may actually be more than indicated because
dilution of the plume downstream can reduce the opacity and counteract
the effect of the increase in condensed acid.

     It is of interest to note that there appears to be an increase
(56%) in particulate loading as determined by Method 5 with an increase
in excess 02 (Table 2, Plant A).  One might expect a decrease because of
a reduction in unburned carbon with more complete combustion.  There is
the possibility of an increase in measured particulate loading due to
the collection of the gaseous acid and sulfate salts by Method 5.
Related studies in our laboratory have shown the glass fiber filter to
be a good collector of the gaseous acid12.  One can postulate that two
overlapping functions (one an increase in sulfate salts and acid and
another a decrease in unburned carbon with increasing excess 02) contri-
bute to the particulate loading.  The former will be a curve with a
positive slope, the latter a curve with a negative slope.  The resulting
curve on particulate loading as a function of excess 02 would have a
positive or negative slope depending upon which function has the steeper
slope.  The resulting curve would approach a straight line (zero slope)
as the two functions tend to exactly counteract each other.  Conse-
quently, depending upon the amount of excess 02, the particulate loading
may be on either the rising or declining slope of the curve or it may be
more or less constant.

     The emission characterization data for high sulfur  (>22 S) coal-
fired power plants with ESP controls show no significant difference
between plume and in-stack opacity under normal ESP operation or with
reduced electric fields in the ESP's.  It is possible that the high ash
content of the coal and resulting high particulate loading in the emis-
sion have a predominant effect on the in-stack and plume opacity.  In-
stack opacity at normal operation of the ESP was close to the opacity
emission standard for new sources and higher than that for the oil-fired
power plants.  The ratio of SOi*"  to particulate matter for the coal-
fired emissions is <1 and for the oil-fired emissions, >1.

     The stack gas environment for the coal-fired power plant  (LC) with
particulate and gas controls (two-stage wet-scrubber) was unlike that  for
plants with ESP controls.  The stack gas temperatures were below the
sulfuric acid dewpoint and the water vapor content from  the wet scrubbers
was high.  The result was that sulfuric acid will appear  in  the gas
stream as condensed acid droplets and these directly affect  the in-stack
opacity.

     Concurrent emission data could not be obtained  for  Plant  LC
(Table 4) in the same manner that it was for Plants  P  and MC  (Table  3)
                                57

-------
because the required number of sampling ports was not available.  None-
theless, the data obtained from the three visits to Plant LC do permit
some qualitative observations.  The emission data on the particulate
loading appear to fall within a narrow range of values indicating con-
sistent plant operation.  The plume and in-stack opacity data show no
significant difference but the very high opacity values are not consis-
tent with the particulate loading, size, and composition normally
associated with fly ash emissions.  Optical transmittance measurements
conducted at different wavelengths during the study gave data that
varied with wavelength in the visible portion of the spectrum13.  This
variation with wavelengths is indicative of submicron size distribution.
The submicron size was also substantiated with in-stack impactor
measurements11*.  The acid composition of the gas stream was substantiated
by the controlled condensation measurement data7.  In this case, we
attribute the high in-stack opacity levels to fine particle concentra-
tion with a mass mean diameter in the submicron size range and with a
significant percentage of the composition consisting of condensed
sulfuric acid.

     There are a number of important questions raised by the data  .
obtained thus far.  More studies are needed to adequately address these
questions and to determine the variation of these pollutant emissions
with operating parameters.  The questions can be briefly stated as
follows:

     0   What is the quantitative distribution of H^SOi, in the gas
         stream between free acid and acid adsorbed on particulate
         matter?

     •   What is the distribution of the sulfate ion (SOt*") between
         acid and salts?

     0   What is the size distribution of acid and salts?

There is need for more data on the physical properties of H^SOi* in both
the stack and plume environments to support proper interpretation of
optical data.
                                   58

-------
                                SUMMARY
     Emissions from oil-fired power plants without emission controls
«nd coal-fired power plants with ESP's and with FGD systems were charac-
terized for plume and in-stack opacity, S02, S0i»", and mass concentration.
Sulfuric acid content of the emissions from the oil-fired power plant
had a significant effect on the plurae opacity but no effect -on the
in-stack opacity.  In the case of the coal-fired power plants with ESP's
the in-stack and plume opacities were essentially the same.  This led to
the conclusion that the concentration of acid was low relative to the
non-acid particulate such that the acid did not contribute to any
significant degree to the opacity of the plume beyond that normally
associated with the fly ash.  The in-stack and plume opacities of the
emissions for the high sulfur coal-fired power plant with the two-stage
wet scrubber system were comparable but significantly high (70 to 90%).
The high opacity was attributed mainly to the sulfuric acid content of
the emissions and to submicron size of the particulate matter.
                                 59

-------
                              REFERENCES
1.   U.S. Environmental Protection Agency, Standards of Performance for
     New Sources.  Fed Regist.  36, 24876-24895, 1971.

2.   U.S. Environmental Protection Agency, Standards of Performance for
     New Sources.  Fed Regist.  40, 43850-43854, 1975.

3.   Cook, C. S., G. W. Bethke, and W. D. Conner.  Remote Measurement of
     Smoke Plume Transmittance Using Lidar.  Appl Opt.  11, 1742-1748,
     1972.

4.   Johnson, W. B., R. J. Allen, and W.  E. Evans.  Lidar Studies of
     Stack Plume in Rural and Urban Environments.  U.S. Environmental
     Protection Agency, Research Triangle Park, N. C.  Publication
     Number EPA 650/4-73-002.  1973.  112 p.

5.   Peterson, C. M., and M. Tomaides.  In-Stack Transmissometer Tech-
     niques for Measuring Opacities of Particulate Emissions from Sta-
     tionary Sources.  NTIS.  Publication Number PB-212-741.

6.   Herget, W. F., and W. D. Conner.  Instrumental Sensing of Stationary
     Source Emissions.  Environ Sci and Technol.  11, 962-967, 1977.

7.   Homolya, J. B.  Unpublished data.  U.S. Environmental Protection
     Agency, Research Triangle Park, N. C.  1978.

8.   Homolya, J. B., and J. L. Cheney.  An Assessment of Sulfuric Acid
     and Sulfate Emissions from the Combustion of Fossil Fuels.   In:
     Proceedings of Workshop on Measurement Technology and Characteriza-
     tion of Primary Sulfate Emissions from Combustion Sources.  Southern
     Pines, N. C.  April 24-26, 1978.  J. S. Nader, Ed.  EPA document in
     press.

9.   Dietz, R. N., and R. F. Wieser.  Operating Parameters Affecting Sul-
     fate Emissions from an Oil-Fired Power Unit.  In:  Proceedings of
     Workshop on Measurement Technology and Characterization of Primary
     Sulfate Emissions from Combustion Sources, Southern Pines,  N. C.
     April 24-26, 1978.  J. S. Nader, Ed.  EPA document in press.
                                  60

-------
10.  Cheney, J. L., and J. B. Homolya.  Characterization of Combustion
     Source Sulfate Emissions with a Selective Condensation Sampling
     System.  In:  Proceedings of Workshop on Measurement Technology
     and Characterization of Primary Sulfate Emissions from Combustion
     Sources, Southern Pines, N. C.  April 24-26, 1978.   J. S.  Nader,
     Ed.  EPA document in press.

11.  Bennett, R. L., and K. T. Knapp.  Chemical Characterization of
     Particulate Emissions from Oil-Fired Power Plants.   In:  Energy
     and the Environment, Proceedings of the Fourth National Conference,
     Cincinnati, Ohio.  October 3-7, 1976.  P. 501-506.   AICHE, Dayton,
     Ohio, 1976.  594 p.

12.  Cheney, J. L.  Unpublished data.  U.S. Environmental Protection
     Agency, Research Triangle Park, N. C.  1978.

13.  Conner, W. D.  Unpublished data.  U.S. Environmental Protection
     Agency, Research Triangle Park, N. C.  1978.

14.  Knapp, K. T.  Unpublished data.  U.S. Environmental Protection
     Agency, Research Triangle Park, N. C.  1978.
                                  61

-------
                        Table 1.   SUMMARY  OF  PHYSICAL AND NORMAL OPERATING FEATURES
                                          OF  POWER  PLANTS STUDIED
INJ
Plant
A
M
P
MC
LC
Fuel
Burned
Oil
Oil
Coal
Coal
Coal

Height
(m)
150
60
88
178
213
Stack
Diameter
(m)
6.8
3.
4.1
4.7
7.
Fuel Content
Tempera-
ture (°C)
160
127
154
166
77
Ash
M
0.17
0.07
8.
14.
30.
S
M
2.4
1.2
3.3
3.9
5.4
V
(ppm)
590
15
99
35
50
Excess
Boiler
0? (%)
0.2
3
5.
4.
3.8
Emission
Controls
only fuel
additives
only fuel
additives
ESP
ESP
2-stage
Power
Output
(KM)
525
190
100
330
820
                                                                                     wet scrubber
                                                                                     (particulate
                                                                                     plus FGD)
        aAs measured at economizer outlet

-------
                              Table 2.  EMISSION CHARACTERIZATION DATA FOR OIL-FIRED
                                             POWER PLANTS WITHOUT ESP

                                                      PLANT M
CT>
CO
Date

8/10/76

8/19/76
8/19/76
8/17/76
8/19/76
8/19/76
Part.
2
Time (mg/Nm )

1115-1215 27

1125-1145
1145-1200
1145-1205 250
0945-1000 390
1000-1015 390
SOZ S04" S04"/SOX Plume In-Stacka
(mq/m3) (mg/m3) (%) Opac. (%) Opac. (%)
Lldar
1,600 68 4.1 10±2
PLANT A
3U4
43±3b
3,800 340 8.2 23±3
23±2
54±3b
Obs.
6±1

30±1
52±2b
42±1
37±2
61±2b

2-3

18-22
18-22
11-15
11-15
11-15
Remarks

15 ppm 1

590 ppm
590 ppm
590 ppm
590 ppm
590 ppm
       aTransmissometer measurements.

        Measurements about 3 stack diameters (15 meters)  downstream of stack exit.

-------
                      Table 3.  EMISSIONS CHARACTERIZATION DATA FOR COAL-FIRED
                                       POWER PLANTS WITH ESP's
                                              PLANT Pd
Date
7/19/77
7/20/77
7/22/77
7/26/77
7/27/77
7/28/77
7/28/77
Part.
Time jmg/Nm )
1551-1636
1020-1120
0939-1039
1430-1545
0915-1015
0900-1000
1110-1210
840
1,100
360
330
2,800
2,100
450
SO,
3
(mg/m )
7,200
6,000
7,300
5,900
5,900
6,500
6,500
                                          (mg/m )
                                                   so4-/so
a8% Ash, 3.3% S, 99 ppm V, 100 MM
b!42 Ash, 3.9% S, 35 ppm V, 330 MW
C0bserver measurements
 Transmissometer measurements
250
330
180
PLANT
210
190
240
240
3/4
5.2
2.4
MCb
3.4
3.1
3.6
3.6
   Plume**     In-Stack
Opacity (%)  Opacity (%)  Remarks
27
23
32
18
86
80
20
21
28
31
21
81
72
17
                                                                                        2 fields off
                                                                                        2 fields off
                                                                                        2 fields off
                                                                                        1 field off
                                                                                        3 fields off
                                                                                        3 fields off
                                                                                        Normal opera-
                                                                                            tion

-------
                             Table 4.  EMISSIONS CHARACTERIZATION  DATA  FOR COAL-FIRED
                                             POWER PLANT LC WITH FGDd
tn

Date
9/19/77
9/19/77
9/19/77
9/19/77
11/2/77
11/3/77
11/3/77
4/3/78
A/3/78
4/4/78
4/4/78
4/5/78
4/5/78

Time
1014-1121
1201-1309
1390-1500
1528-1635
1529-1629
1345-1445
1530-1630
1245-1345
1515-1615
1145-1245
1345-1445
0930-1030
1130-1230
Part.b S0~b
1 *J
(mg/NnT) (mg/nT)
2,000
3,100
4,200
5,500
240

270
350
320
220
280
260
280
SO/ SO//SO b Plumec
*\ *\ *r X
(mg/nn (%) Opacity (%)
190 8.7
260 7.7
170 3.9
170 3.0

75.

>90
>90
>90
>90
>90
>90
In-Stackd
Opacity (%)




59-67
62-71
62-71
>90
>90
>90
>90
>90
>90
       a30% Ash, 5.42 S, 50-ppm V. 820 MW
       Measurements made on one of 8 parallel FGD modules.
       C0bserver Measurements
       dTransmissometer measurements at stack breeching.

-------
               Free 1I2S04 (Gas Phase)
              Particles with  Adsorbed H,S04 rrv '•'.'
                                 Sulfate Particles
          ••/•-.   Free i>02
                             Particles with Adsorbed SO
         Stack Ducting
Figure 1.  Sulfur Oxides  Present in the Stack Gas Stream

-------
+*
c
o
u
(4
o
d


O


t*


•O
•H
J
3

i-H

A
    50
   40
   30
    20
   10
^
                           • 1%S,  15 ppm V, 4*02

                           A '2.5%S, 590 ppm V, 0.2302

                           • 2.5%S, 590 ppm V, 0.6%02


                             Lidar Measurement

                             Location above Stack Exit

                             Unprimed - 2 to 3 meters

                             Primed - 15 meters

                          I   	I	I	
                10         20         30        40         50

            In-Stack Transciissometer Opacity, percent
Figure 2.  Concurrent Plume and  In-Stack Opacity Data

           for Emissions  from Oil-Fired  Power Plants
                          67

-------
Charts Showing In«Stack Opacity Measurements and Individual
Opacity Readings of Participants fn Collaborative Testing of
EPA Method 9.  [Data taken from tables In EPA Report No. 650/4-
75-009, Evaluation and Collaborative Study of Method for Visual
Determination of Opacity of Emissions from Stationary Sources,
H. Hamll, January 1975)
                        69

-------
          AFTER HAMIL AND THOMAS
          EPA-650/4-75-009
          JANUARY 1975
                    10         15         20         25

                       TRANSMISSOMETER OPACITY, percent

BLACK SMOKE GENERATOR
VIEWING CONDITIONS: IDEAL
TOTAL NO. OF POINTS: 132
NO. OF DETERMINATIONS PER OBSERVER:  APPROXIMATELY 16

-------
  35
  30
  25
I
I20
<
>  15
cc
LLJ
v»
CO
o
  10
      X
AFTER HAMIL AND THOMAS
EPA-65Q/4-75-OQ9
JANUARY 1975

O MULTI-POINT
    0         5         10         15         20         25

                          TRANSMISSOMETER OPACITY, percent
    WHITE SMOKE GENERATOR
    VIEWING CONDITIONS: IDEAL, BLUE SKY, BRIGHT SUNSHINE
    TOTAL NO. OF POINTS: 170
    NO. OF DETERMINATIONS PER OBSERVER: APPROXIMATELY 20
                                                    30
35
                                  73

-------
   30
   25
   20
0.
o

QC
GC
UJ
CO
CO
0
   15
   10
            AFTER HAMIL AND THOMAS
            EPA-650/4-75-009
            JANUARY 1975
                            10           15          20

                          TRANSMISSOMETER OPACITY, percent
25
30
    RIVERBEND STEAM STATION TEST NO. 1
    VIEWING CONDITIONS: CLOUDY. LOW HAZE
    TOTAL NO. OF POINTS: 45
    NUMBER OF DETERMINATIONS PER OBSERVER:  19
                                    1C
                                    I w

-------
            AFTER HAMIL AND THOMAS
            EPA-650/4-75-009
            JANUARY 1975
                               v-:   x
                        10           15          20

                        TRANSMISSOMETER OPACITY, percent
RIVERBENO STEAM STATION TEST NO. 2
VIEWING CONDITIONS: MARGINAL, SOLID OVERCAST
TOTAL NO. OF POINTS: 85
NO. OF DETERMINATIONS PER OBSERVER: 17
                              77

-------
   so
   40
   30
a.
o
OC
   20
CO
o
   10
                 AFTER HAMIL AND THOMAS
                 EPA-650/4-75-009
                 JANUARY 1975

                 O MULTI-POINT
                  10
20
30
40
                             TRANSMISSOMETER OPACITY, ptrcent
     RIVERBENO STEAM STATION TEST NO. 3
     VIEWING CONDITIONS: IDEAL. CLOUDLESS. BRIGHT SUNSHINE
     TOTAL NO. OF POINTS: 96
     NO. OF DETERMINATIONS PER OBSERVER: 24
50
                                     79

-------
  30
  25
  20
<
a.
o
cc
cc
LU
v>
CO
o
  10
                1            I            I
         NUMBERS-;* DETERMINATIONS AT POINT
*    *^   /     *
•3  v»* •¥        •
                      •3,»12»6
                    18/t*2«27«IO.tO

                        •12*3
                        10          15           20

                       TRANSMISSOMETER OPACITY, percent
STAUFFER CHEMICAL (ACID MIST PLUME)
VIEWING CONDITIONS: GOOD. BLUE SKY. LIGHT HAZE
TOTAL NO. OF POINTS: 297
NO. OF DETERMINATIONS PER OBSERVER: APPROXIMATELY 30
                                                                  25
                                                        30
                                    81

-------