United States       Office of Air Quality        "{K/
Environmental Protection   Planning and Standards      January
Agency         Research Triangle Park NC 27711
Stationary Source Compliance Series	
A Compilation
of Opacity
Monitor
Performance
Audit Results

-------
                                             EPA-340/1-83-011
A Compilation of Opacity Monitor
      Performance Audit Results
                        Prepared by:

                       Steve Plaisance
                  Entropy Environmentalists, Inc.
               Research Triangle Park, North Carolina
                        Prepared for:

                      Darryl von Lehmden

                           and

                       Thomas Logan

                   Quality Assurance Division
            United States Environmental Protection Agency
                  QAD Contract No. 68-02-3431

                           and

                       Louis R. Paley
               Stationary Source Compliance Division
            United States Environmental Protection Agency
                 SSCD Contract No. 68-01 -6317
            U.S. ENVIRONMENTAL PROTECTION AGENCY
             Office of Air Quality Planning and Standards
               Stationary Source Compliance Division
                   Washington, D.C. 20460
                       January 1983   ,t ^ s
-------
The Stationary Source Compliance series of reports  is  issued  by the
Office of Air Quality Planning and Standards,  U. S Environmental
Protection Agency, to assist Regional  Offices  in  activities related  to
compliance with implementation plans,  new source  emission  standards,
and hazardous emission standards to be developed  under the dean Air
Act.  Copies of Stationary Source Compliance  Reports are available -
as supplies permit - from Library Services, U.S.  Environmental
Protection Agency, MD-35, Research triangle Park, North Carolina
27711, or may be obtained, for a nominal cost, from the National
Technical Information Service, 5285 Port Royal Road, Springfield,
Virginia  22151.

This report has been reviewed by the Office of Air  Quality Planning
and Standards, U.S. Environmental Protection  Agency, and approved for
publication as received from Entropy Environmentalists, Inc.  Approval
does not signify that the contents necessarily reflect the views and
policies of the U.S. Environmental Protection  Agency,  nor  does mention
of trade names or commercial products  constitute  endorsement  or
recommendation for use.
                             ii

-------
                                ABSTRACT

     Opacity monitor performance  audit  procedures and devices have been
developed and field tested on 93  opacity monitors *  The results of this
test program indicate that opacity monitoring  systems achieve high levels
of availability,  and are capable  of providing  accurate emissions data.  The
results also show that problems impacting data quality are generally
limited to monitor miscalibration  and/or misadjustment, as well as improper
or inadequate operating and maintenance practices.  It is believed that
improved monitor  performance and  data reliability can be achieved with
additional training of monitor operators and more frequent performance
audits.

     This document describes the  audit  program for continuous emission
monitors (CEMs) of effluent opacity.  Detailed explanations of the audit
methodology, monitor analyses, and analytical  criteria are included, and
both criteria- and monitor-specific results of installed opacity monitor
audits are delineated.  Finally,  conclusions are drawn as to the adequacy
of monitor performance and data reliability, and recommendations are
offered that can  optimize opacity monitoring system performance.
                                  iii

-------
                       II. PERFORMANCE AUDIT PROGRAM

                         Audit Program Description

     Performance audits of 93 opacity monitors were conducted at  41 sources
in EPA Regions IV, V, and VII through July 1982.   All  monitors were located
at coal-fired steam generators, with the exception of  one that was located
at a petroleum refinery catalytic regenerator.  These  test  sites  were
selected according to the following audit criteria:

             The test site had at least one  of the four representative
             monitor types in current use (Lear Siegler, Contraves Goerz,
             Dynatron, Esterline Angus).

             The monitor had  been in operation for a minimum of one year.

             The monitor had  undergone a PST that  indicated initial
             compliance with  Performance Specification  1, Appendix B,
             40 CFR 60.

     The opacity monitor audit program was designed  to provide accurate,
reliable analyses of monitor  performance through a simple, quick  field test
procedure which can be performed by a single  technician with a basic
understanding of transmissometry.  Equipment  necessary for a typical audit
includes a specialized retroreflector for the  specific monitor being tested
to simulate clear stack conditions.  In addition,  three neutral density
filters, traceable to the National  Bureau of Standards (NBS) , are included
to evaluate both the linearity and  the calibration error of the monitor.
All of the necessary equipment can  be transported  in a small suitcase.

                             Audit Methodology

     Opacity monitor field audits consist of two sequential procedures: (1)
a general information survey  and (2) the site  monitor audit.  The general
information survey, typically conducted prior  to the actual site visit,
serves to gather  information  required to tailor the  audit procedures and
equipment to the  specific source and monitor.   This  information includes:

         1.  Source identification, location,  fuel  used, emission control
             device(s)  (before and  after  the monitor).

         2.  Opacity monitor  manufacturer, model and serial numbers, dates
             of installation  and certification  (PST), stack diameters at
             monitor  location and exit,  proximity  to upstream or downstream
             flow disturbances,  types of opacity data recording equipment
             and  recording intervals.

         3.  Date  of most recent opacity monitor calibration and type of
             calibration  (off-stack using neutral  density filters, or
             on-stack during  clear-stack conditions) , recalibration
             criteria (zero/span error  or scheduled  recalibration), time
             interval between check/change of  air  purge filters, between
             check/change of  optical  surfaces, or  between zero/span
             adjustment.

-------
     The field audit procedures  are used to determine whether the monitor
has been properly operated  and whether the monitor accuracy and calibration
are of sufficient quality to  provide useful opacity data.  Although these
procedures may differ slightly in  their order for each type of monitor,
they all include the following three basic analyses:

     1.  Monitor Component  Analysis,

         a.  The fault lamp indicators on the monitor's control panel are
             checked to determine  whether the monitor is operating within
             the manufacturer's  prescribed limits.

         b.  The stack exit diameter and monitor pathlength are determined
             to verify the  accuracy of the monitor's preset stack exit
             opacity correction  factor.

         c.  Various internal electronic checks are performed using the
             controls in the monitor's control unit to further verify the
             operational status  of the monitor.

         d.  The control panel meter and chart recorder responses are
             compared to the monitor's internal span value, in order to
             determine the  accuracy of the control panel meter and internal
             zero and span  functions.

     2.  Monitor Maintenance Analysis,

             The optical alignment and dust accumulation on optical
             surfaces are checked  to determine the adequacy of the monitor
             mounting and maintenance frequency.

     3.  Calibration B-ror  Analysis,

             The calibration error and linearity of the monitor are checked
             using neutral  density filters.

     The Lear Siegler, Esterline Angus,  and  Contraves  Goer z monitors
require that an audit device with  an adjustable retroreflector be mounted
on the transceiver to simulate  clear stack conditions.  Neutral density
filters are then inserted into  the audit device to determine the
calibration/accuracy of the monitor at three  different opacity levels.   For
the  Dynatron opacity monitor, the  audit  device neutral density filters must
be inserted in front of the monitor's retroreflector.  This methodology
provides an incremental calibration check, because the filter opacity is
combined with the effluent opacity.

-------
                      Audit Procedures  and Terminology

     Each opacity monitor  field  audit comprises up to 10 specific analyses
which encompass the monitor's accuracy, precision, and the quality of
monitor operation and  maintenance practices.  These procedures and their
associated terminology are detailed as  follows:

     Fault Lamp Analysis.   The transceiver unit of a typical opacity
monitor has several fault  monitors that warn of monitor system malfunctions
and/or impending conditions of excessive opacity.  Monitor fault lamps are
indicative of a variety of conditions,  depending on the manufacturer, but
most units use fault lamps to monitor the intensity of the optical beam,
the quantity of dust on monitor  optical surfaces, the status of internal
circuitry that maintains monitor calibration, and/or the magnitude and rate
of increase of opacity. In general, a  fault lamp error exists if, at the
time of a performance  audit, any of the fault lamps are illuminated.
However , the absence of illuminated fault lamps does not preclude
malfunctions in the fault  lamp circuits or the lamp bulbs.

     Automatic Gain Control (AGO) Circuit Analysis.  Lear Siegler opacity
monitors employ an AGC circuit to compensate electronically for reductions
in the optical beam intensity resulting from power supply fluctuations or
normal bulb deterioration.  This compensation prevents beam intensity
variations from being  interpreted as variations in measured opacity.  Thus,
a fault condition exists when a  Lear Siegler monitor's AGC circuit is not
functional, and is indicated when the AGC lamp is not lit.  However, an AGC
circuit fault does not necessarily diminish the accuracy of opacity
measurements, provided that the  reference signal value is within the
specified range.

     Monitor Alignment Analysis. The optical alignment of the
transceiver/retroreflector system is critical in maintaining accurate
opacity measurements.   Misalignment of  the beam can cause erroneously high
opacity readings, because  a significant portion of the measurement beam is
not returned to the transceiver. Most  opacity monitor manufacturers
include provisions for an  optical alignment check, as either a standard
feature or an option.   Monitor alignment errors are indicated by an
off-center beam image  on the transceiver and/or retrore flee tor.

     Stack Exit Correlation Analysis.   Typically, the cross-stack optical
pathlength of the installed opacity monitor is not equal to the diameter of
the stack exit.  To obtain a true stack exit opacity value, the measured
opacity at the monitor location  is corrected to stack exit conditions
through the use of a pathlength  correction factor.  The stack exit
correlation error is the percent error  of the pathlength correction factor,
as preset by the manufacturer, relative to a pathlength correction factor
calculated through the use of actual measurements, blueprints, etc.

     Control Panel Meter Analysis.  Most opacity monitors have a panel
meter located on their control or transceiver units to monitor opacity
readings or to adjust  an internal monitor parameter.  The control panel
meter correction factor is the meter readings compared to the specified
opacity value for the  internal span filter.

-------
     Reference Signal Analysis.   The  Lear  Siegler monitor reference signal
is an internal monitor electrical signal output that indicates the
electronic alignment  of the  transceiver circuitry (usually 20 ma).  The
reference signal  error is  the  percentage difference between the measured
current value of the  signal  and  the specified 20-ma value.

     Internal Zero and Span  Analysis.  The zero and span errors are the
percentage difference between  the internal reference filter opacities that
the monitoring system should record and those opacity values actually
displayed on the  chart recorder. These errors occur due to one or a
combination of reasons:  (a) the values for the internal zero and span
functions are wrong,  (b) the internal electronics responding to the zero
and span functions are not working properly, (c) the signal from the
transceiver to the chart recorder is  being altered, and (d) the chart
recorder is not displaying the incoming signal correctly.

     Zero Compensation Analysis. The zero compensation circuit of the Lear
Siegler monitor automatically  adjusts the monitor's zero to compensate for
dust accumulation on  the transceiver's optical surfaces.  The zero
compensation analysis is based on recording the zero compensation before
and after cleaning the transceiver and retro-reflector optical surfaces.

     Optical Surface  Dust  Accumulation Analysis.  The optical surface dust
accumulation analysis determines the  amount of dust (measured in terms of
percent opacity)  found on  the  optical surfaces, based on the reduction in
opacity before and after cleaning of  the optical surfaces.  To obtain a
reliable assessment,  this  audit  analysis should be performed when the stack
opacity is relatively constant.

     Calibration Error Analysis. This analysis involves comparison of the
monitor responses to  the known opacity values for a series of three
reference neutral density  filters (as modified in opacity value by the
optical pathlength correction  factor).  The calibration of the reference
filters used in this  analysis  is traceable to the NBS.

                              Audit  Criteria

     Specific criteria have  been developed for the determination of opacity
monitor performance based  on (1) Performance Specification 1, Appendix B,
40 CFR 60, (2) manufacturer's  recommendations, and (3) extensive practical
experience.  For  the  previous  analytical procedures, these criteria are:

     Fault Lamp Analysis.  A fault lamp error is indicated when one or more
of the control unit fault  lamps  are illuminated.  Under performance audit
conditions, these lamps provide  an indication of monitor systems and
parameters that may be out-of-specification prior to the audit, thereby
signifying the level  of monitor  operation and maintenance.

     Automatic Gain Control  (AGC) Circuit Analysis.  A Lear Siegler AGC
circuit error occurs  when  the  AGC circuit is not operating.  The AGC LED,
located on the transceiver head, is illuminated when the AGC circuit is
operating.

-------
     Monitor Alignment Analysis.   The optical beam path should be properly
aligned as indicated  by checking the centering of the beam image on the
transceiver and/or  r etrore flee tor .

     Stack Exit Correlation Analysis.  The percent error of the optical
pathlength correction factor, as preset by the monitor manufacturer,
relative to that calculated from actual measurements, should be no greater
than +2 percent.

     Control Panel  Meter Analysis.  The control panel meter correction
factor, based on a  comparison of opacity values between the meter and that
of the internal span  filter, should fall within the range of 0.98 to 1.02,
or +2 percent opacity (differences of more than 10 percent opacity may be
indicative of other monitor problems) .

     Reference Signal Analysis.  The Lear Siegler reference signal error,
as indicated by the percentage difference in the internal reference signal
value and the manufacturer's specified value of 20 ma, should not exceed
+ 10 percent of the  specified value.

     Internal Zero  and Span Analysis.  Internal zero and span errors, based
on comparing the monitor internal  filter opacity values with those
indicated by the opacity recorder, should not exceed +2 percent opacity.

     Zero Compensation Analysis.   The zero compensation analysis, based on
a comparison of zero  compensations before and after cleaning of monitor
optics, should result in an automatic zero adjustment range of +0.018
optical density (not  in excess of +4 percent opacity)  The zero
compensation should approach 0.000 optical density after cleaning.

     Optical Surface  Dust Accumulation Analysis.  The optical surface dust
accumulation, based on the  difference in opacity readings before and after
cleaning of monitor optics, should not exceed 4 percent opacity.

     Calibration Error Analysis.   Transmissometer calibration error, based
on the sum of the absolute  value of the mean difference and 95 percent
confidence interval observed for the differences in opacity indicated by
the monitor and that  of the given  reference neutral density filters, should
be no greater than  3  percent opacity.

-------
               III.   SUMMARY AND DISCUSSION OF AUDIT RESULTS

     Opacity monitor  audit results are summarized and discussed  in  this
section for the four  representative types of monitors evaluated.  Table  1
summarizes results  of the audit analyses of the four monitors.  Figure  1
illustrates the relationship between actual and measured opacity for  all
monitor types, and  Figures 2 through 5 illustrate the same relationship for
each type of monitor  tested.

                   Results of Criteria-Specific Analyses

     General Operation and Maintenance Survey.  At the initiation of  the
opacity monitor field audit  program, all of the monitors were believed to
be functioning properly.  However, of the 93 monitors audited, only 89 were
fully operational.  Monitor  availability, as determined by reviewing  the
previous month's data, was  100 percent for 69 of the 93 monitors, and the
average monitor availability was approximately 95 percent.

     The general audit survey includes a variety of factors affecting
monitor operation and maintenance (e.g., the accessibility for maintenance,
the supply of spare parts, etc.).  In general, these factors had little
correlation with monitor performance.  However, one factor of significant
importance was whether the monitor maintenance was performed by  plant
personnel or by a monitor-servicing contractor.  Nearly all (six of seven)
of the monitors serviced by  monitor-servicing contractors had calibration
errors of _<3 percent  opacity.  In contrast, monitors serviced by site
personnel often lacked necessary maintenance, and two sources were
identified at which monitor  manufacturer's specifications for maintenance
were not followed .

     During the audit program, 30 opacity monitors that were installed in
exhaust ducts transporting effluent streams to a common exhaust  stack were
audited.  None of these common stacks had an installed opacity monitor .  To
determine the stack exit opacity, the source combined the duct opacities
measured by the respective monitors without weighting them according  to
duct flowrate .  In  addition, a few of these sources simply arithmetically-
averaged the duct opacities  to obtain a stack exit opacity.  Thus,  in cases
where the duct flowrates were unequal , the total opacities measured for
combining duct systems were  determined incorrectly.

     Fault Lamp Analysis.  The fault lamps provide a good indication  of the
current status of the monitoring system operation and maintenance,
particularly with reference  to the optical beam intensity, optical  system
dust accumulation,  or a zero/span malfunction.  The monitor parameter
indicated by a fault  lamp is considered to be "out of specification"  if the
control unit lamp is  illuminated.  Of the 93 monitors audited,
approximately 4 percent of the total were found to have one or more
illuminated fault lamps.  In addition, these same monitors had calibration
errors in excess of 10 percent, thereby indicating a positive correlation
between the fault lamp status and the reliability of measured opacity.

     Automatic Gain Control  (AGO Circuit Analysis.  The status  of the  AGC
circuit (anployed on  the  Lear  Siegler monitor only) is indicative of  the
monitor's ability to  compensate electronically for reductions in the
optical beam intensity resulting from fluctuations in the lamp electrical

-------
                                      TABLE 1.  OPACITY MONITOR AUDIT  RESULTS
AUDIT
ANALYSES
Number of Monitors Audited
MONITOR COMPONENT ANALYSIS
Fault Lamps
AGC Circuit
Monitor Alignment
Stack Exit Correlation Error
Panel Meter Status
Reference Signal Error
Internal Span Error
Internal Zero Error
MONITOR MAINTENANCE ANALYSIS
Zero Compensation Factor
Optical Eust Accunulc^ion
Calibration Error Analysis :^
Low Range (8.5% Opacity)
Midrange (18.5% Opacity)
High Range (42.5% Opacity)
LEAR
Number
Within
Specs.
SIEGLER
Number
Out of
Specs.
58

56
57
55
51
0
54
44
53

47
41

50
49
46

2
1
0
7
55 4
1
11
2

8
12

5
7
9
CONTRA VES GOERZ DYNATRON
Number Number Number Number
Within Out of Within Out of
Specs. Specs. Specs. Specs.
27

27
NA3
26
26
22
NA
22
25

NA .
22

25
22
19

0
NA
0
1
5
NA
5
2

NA
3

2
5
8

3
NA
3
2
4
NA
3
3

NA
3

1
1
0
5

2
NA
2
3
0
NA
1
1

NA
1

3
3
4
ESTERLINE ANGUS
Number Number % Total Number
Within Out of Monitors Tested
Specs. Specs. Out of Specs.
3

3
NA
2
3
2
NA
3
3

NA
3

3
3
2

0
NA
1
0
1
NA
0
0

NA
0

0
0
1

4
2
3
12
69
2
19
6

15
19

11
17
25
See Section II of this  report for audit methodology, definition of terms,  and  audit criteria.

The calibration error analysis for the Dynatron monitors utilized different  neutral density  filters:
Low Range = 17.0$ Opacity, Midrange = 56.5% Opacity, and High Range =  81.0%  Opacity.

Not Applicable

Lear Siegler panel meter  readings of opacity were generally accurate,  but  values  for  optical density
and circuit current were  typically erroneous.

-------
70
                       20         30         40

                Actual Opacity  (% Opacity)
    Figure   1.    Observed Opacity Responses for All
                 Audited Monitors:  8.9 Monitors
                           12

-------
   70
   60
•H
O

a
o
W
c
o

CO
0)
o
(0
QJ
M
3
CO

(1)
                   Actual Opacity  (%  Opacity)



       Figure  2. Observed  Opacity Responses  for  58  Lear Siegler

                 Monitors
                            13

-------
•H
O
td
D.:
O
(U
CO
c
o
O4
CO
0)
•p
•H
0
n}
a.
o

T)
a>
en
<0
    TO-
    GO-
    40.
    10-
                      Range of Measured Responses
             Average'Measured Responses
                                                        /
                                              ±3% Opacity

                                                Error Band
                                   Expected Responses
                               i I I i | i I i  i
                                          I l i I |  i i i i
IMIIIIII
                  10         20         30         40

                    Actual Opacity (% Opacity)
       Figure  3,   Observed Opacity Responses for  27

                    Contraves Goerz Monitors
                                                              50
                                 14

-------
  100
   80
-P
•H
O
(0
ft
O

Oft
4J
•H
O
0)
M
d
en

-------
    70
    60
    50
4J
•H
O
(0
Ck
o

df>
•H
O
T)
(U
M
d
en
(0
OJ
40
    20
    10
                      Range of Measured Responses
            Average Measured Responses —

                                            +3% Opacity
                                             Error Band
                                    Expected Responses
         M T I I [ l t i  i i l i  | i i l i  11 i i  i | i M i | i i i i i i i  i i  i

       0          10         20         30         40

                     Actual  Opacity (% Opacity)
                                                            50
        Figure  5.   Observed Opacity Responses for  3 Esterline Ang\

                    Monitors
                              16

-------
supply and/or deterioration of the  lamp bulb.  The  AGC circuit is
considered to be "out of specification"  if  the circuit is not operating, as
indicated by an LED on the Lear Siegler  transceiver.  Of the 58 monitors
evaluated, only one monitor (approximately  2 percent of the total) was
found to have an inoperative AGC circuit.

     Monitor Alignment Analysis.  The monitor alignment analysis indicates
whether the optical beam path across the effluent stream is centered on the
measurement r etr ore flee tor , with "out of specification" conditions being
indicated by an off-center beam path.   Improper alignment was found in
three of 89 monitors audited, representing  about  3  percent of the total.
In addition, two of these monitors  (Dynatron) were  out-of-calibration, and
the source indicated that attempts  at proper alignment resulted in the loss
of monitor zeroing capability.  During  o ff-stack calibration , however, an
alternative alignment was devised,  but  this alignment failed to correct the
zero calibration problem when the monitor was installed on the stack,
thereby indicating that such problems may require the attention of a
monitor service specialist.  Finally, three Lear  Siegler monitors had
faulty alignment sights, and therefore,  were not included in the tabulated
audit results.

     Stack Exit Correlation Analysis.   The  stack exit correlation analysis
evaluates the pathlength correction factor  which has an exponential impact
on the monitor opacity reading.  The audit  results  indicate that the most
common error in computing the monitor optical pathlength was the use of the
flange-to-flange distance, rather than  the  stack/duct inside diameter.
Approximately 12 percent of the monitors audited had incorrect pathlength
correction factors.  However, the data obtained for these monitors can be
used, provided that the erroneous opacity readings  are mathematically
corrected by using the proper pachlength correction factor.

     Panel Meter Analysis.  The panel meter analysis serves to evaluate the
accuracy of the panel meter readings of  opacity, optical density,  and (on
Lear Siegler monitors only)  reference circuit current values.   The panel
meter is considered to be "out of specification" if any of the meter
readings vary by more than 2 percent of  the true value.  Although 61 of the
monitors audited (about 69 percent  of the total) exhibited faulty panel
meter readings, 55 of these units were  Lear  Siegler monitors which
accuractely indicated opacity,  but  were  inaccurate  in their readings of
optical density and/or monitor  circuit current.  In addition,  zero and span
function adjustments were made on five monitors using inaccurate panel
meter readings.  In general, the  audit results indicated that  chart
recorder readings of opacity,  optical density, and monitor circuit current
should be used in monitor calibration and adjustment.  However, seven
monitors (8 percent of the total) were found to have incorrect chart
recorder readings.

     Reference Signal Analysis.  The reference signal analysis for the Lear
Siegler monitor serves as an internal verification of the beam intensity as
well as an indication of the status of the  photo detector and  its
associated electronics.  The reference signal is considered to be  "out of
specification" when it varies by more than -MO percent beyond  the  value
specified by the monitor manufacturer.   Based on the audit data, all  of the
                                 17

-------
Lear  Siegler monitors had reference  signals within the recommended range,
with  the exception of one unit.   This monitor had an incorrectly calibrated
panel meter, and correction of this  miscalibration resulted in a reference
signal reading within the manufacturer's recommended range.

Internal Zero and Span Analysis.   The internal  zero and span analysis
evaluates the monitor's ability to maintain calibration by automatically
adjusting its internal electronics to compensate for dust accumulation on
monitor optics.  The zero and  span are considered to be "out of
specification" when either of  their  errors exceed +2 percent opacity.  The
audit results indicated no direct correlation between zero/span errors and
monitor miscalibration.  Although 72 monitors had accurate zero/span
responses,  12 of these had calibration errors outside the specifications.
Likewise, of  17 monitors having zero/span errors exceeding the
specifications, only 10 had excessive calibration errors.

      Procedural error was identified at two sources having seven monitors
in which the zero and span were adjusted without cleaning the monitor
optics.  Because the zero mode is designed to simulate the amount of light
returned to the transceiver under clear stack conditions, any dust
accumulation on the transceiver window during the zero adjustment causes a
lesser amount of light to be returned to the detector, and thus, an
improper zero adjustment.  However,  the Lear Siegler monitor's zero
compensation circuit is automatically reset during the zero adjustment, and
the circuit will respond to the dirty optics as if they were clean.  Thus,
the adverse impact of making zero adjustments without cleaning the monitor
optics depends on whether the  monitor has a zero compensation function and
on its effectiveness, and may  or  may not bias the recorded opacity.

     Zero Compensation Analysis.   The zero compensation analysis evaluates
the extent to which the zero compensation circuitry of the Lear Siegler
monitor electronically nullifies  the adverse effects of dust accumulation
on monitor optics.  The zero compensation is considered to be "out of
specification" if the indicated value exceeds +0.018 optical density (4
percent opacity).  The  audit results indicated that the zero compensation
circuit values are an accurate measure of the amount of dust on the
transceiver optics, but give no indication of the dust accumulation on the
retroreflector .  Eight  (15 percent)  of the 55 Lear Siegler monitors had
zero compensation values exceeding 4 percent opacity, and eleven monitors
incorrectly used the zero compensation value as an indication of dust
accumulation on the entire system optics.  Thus, the Lear Siegler zero
compensation circuitry generally  fulfills its intended purpose of
accounting for dirt on  the transceiver optics.

     Optical Dust Accumulation Analysis.  The optical dust accumulation
analysis is a quantitative determination of the dust deposition on a
monitor's exposed optical surfaces.  An "out of specification" dust
accumulation is indicated by a value in excess of 4 percent opacity.  Of 85
monitors audited, 16 (19 percent  of  the total) had excessive dust on the
optical surfaces.  The  audit results indicated that dust deposition on
monitor optics is both  site- and  monitor-specific, with deposition
occurring at dissimilar, non-linear  rates for transceiver and
retroreflector optics.
                                     18

-------
      Calibration Error Analysis.  The calibration error  analysis determines
 the  accuracy and linearity of the entire opacity monitoring  system,
 excluding  retrore flee tor optics and monitor alignment.   The  calibration
 error is considered to be "out of specification" if the  difference between
 the  indicated opacity and that of the NBS neutral density filter employed
 exceeds  3  percent opacity.  In order to optimize the accuracy of the
 calibration  error analysis, all exposed monitor  optics were  cleaned prior
 to calibration, and the optical pathlength correction factor preset by the
 manufacturer was used to correct the audit filters for stack exit
 conditions.

      For all four types of monitors evaluated, the high  range calibration
 error was  the most prevalent with 25 percent of  the monitors having opacity
 readings in  excess of the specifications.  Furthermore,  the mid and low
 range error  analyses indicated corresponding reductions  in calibration
 error, with  these values being determined as 17  percent  and  11 percent,
 respectively.  Thus, a linear relationship between the magnitude of the
 opacity and  that of the calibration error was indicated  wherein monitors
 with lower range errors would not only have high range errors, but these
 errors would be of a greater magnitude (see Figure 1).   In general, the
 monitors audited with calibration errors in excess of 3  percent opacity
 were biased  high.  Poorly performing monitors tended  to  give falsely high
 opacity readings, and, therefore, it is in the source's  best interest to
 calibrate  and maintain opacity monitors properly.

                    Results of Monitor-Specific  Analyses

      Lear  Siegler.  The component analysis of 58 Lear Siegler opacity
 monitors indicated significant problems with panel meter and internal  span
 errors.  Although none of the monitors'  panel meters were within
 specifications, the erroneous readings were  limited to values of optical
 density and circuit current, with opacity readings of panel meters being
 generally accurate.  Internal span errors were present in 11  monitors  (19
 percent of the total)  thereby,  indicating  either  the presence of
 significant quantities of dust  on the  transceiver  system optics,
 miscalibration of the  monitor,  or an  incorrectly labeled internal  span
 filter.

      An analysis  of monitor  maintenance  indicated  that dust accumulation  on
 system optics was a  significant  factor, with approximately 30 percent  of
 the  Lear Siegler  monitors having  excessive dust accumulations.   As  a
 further consequence  of this  dust  accumulation, 17 percent of the monitors
 had  zero compensation  values in  excess of the specified  value.   In
 addition, several  of the  sources  indicated that they were unaware of the
 specifications for  the  zero  compensation value, thereby indicating  their
 lack of understanding  of  the role of the zero compensation as an indicator
 of dust accumulation on the  transceiver optics.

     Monitor  calibration  was  found to be generally accurate and  linear,
with the  greatest number  of  "out of specification" monitors resulting  from
 the high range calibration error analysis (9 monitors, or 16  percent of the
total).  As Figure  2 illustrates, the  Lear Siegler monitors tended  to  give
elevated  opacity measurements, but typically not  in excess of the 3 percent
                                 19

-------
opacity error band based  on  average measured responses.  However,  the range
of measured opacities increased disproportionately with increases  in the
actual opacity, thereby indicating that monitor accuracy deteriorated with
higher actual opacity.

     Contraves Goerz.  An analysis of the performance of 27 Contraves Goerz
monitors indicated that panel meter and internal span errors occurred in
about 20 percent of the monitors audited.  Monitor maintenance,  as
indicated by the optical  dust accumulation, was generally satisfactory,
with only 3 units (about  12  percent of the total) having excessive dust
accumulations.  Monitor calibration, as illustrated in Figure 3, was
adequate, with average low,  medium, and high opacity values falling within
the +3 percent opacity error band.  Contraves Goerz monitor opacity
measurements were typically  higher than actual opacity values, and this
bias increased with higher actual opacity values.

     Dynatron.  Audits of 5  Dynatron opacity monitors yielded problems with
stack exit correlation factors in 60 percent of the monitors (3  units) .
Also, monitor alignment errors were found in 2 units, and these  alignment
problems were not corrected. Fault lamp errors were encountered in 2
monitors, and were not previously repaired by the source because the
monitor control unit partially obscured view of the lamps.

     Maintenance of Dynatron monitors, as indicated by dust accumulation on
monitor optics, was satisfactory, with only one monitor having excessive
dust on system optics. Figure 4 illustrates the very high bias  found in
the calibration analysis, indicating severe inaccuracies in measured
opacity, particularly in  the higher range of actual opacity values.  In
fact, only 25 percent of  the Eynatron units were found to be within the
acceptable range of accuracy for low and mid-range opacity values.
Moreover, monitor linearity  was found to be deficient, with average
mid-range opacity values  having less accuracy than either low or high range
opacity values.

     Esterline Angus.  Of 3  Esterline Angus monitors audited, one  unit had
problems with monitor alignment and panel meter errors.  Optical dust
accumulations were found  to  be minimal, thereby indicating adequate monitor
maintenance.  Furthermore, only one unit was found to have excessive
calibration error (high range).  Monitor accuracy and linearity, as
illustrated in Figure 5,  were satisfactory, with both low and mid-range
calibration errors falling well within the +3 percent error band.
                                    20

-------
                IV.  CONCLUSIONS AND RECOMMENDATIONS

                       General Conclusions

 1.   Opacity monitoring systems installed at stationary sources
     typically achieve high levels of availability (approximately 95
     percent)

 2.   Contemporary opacity monitoring instrumentation  is capable of
     providing accurate emissions monitoring data.

 3-   Problems impacting opacity monitoring data  quality are generally
     limited to monitor miscalibration and/or adjustment, as well as
     improper or inadequate operation and maintenance practices.

 4.   The opacity monitor performance audit procedures which have been
     evaluated provide a reliable indication of  the accuracy of opacity
     monitoring data and the adequacy of monitor operation and
     maintenance practices.

                       Specific Conclusions

 1.   Monitors which were properly operated and maintained demonstrated
     acceptable performance relative to  audit test criteria and
     provided accurate  opacity monitoring data.

 2.   Most installed opacity monitors are  not  affected by optical
     alignment problems.  However,  opacity monitoring systems provided
     by one manufacturer may be more susceptible to alignment problems
     than other  monitoring  systems.

 3.   Incorrect stack exit correction factors  for opacity monitoring
     systems are often  encountered.

 4.   Inappropriate  methods  are  frequently used for determining
     stack-exit  opacity where multiple opacity monitors are  installed
     in separate  ducts  which are exhausted through a common  stack.

 5.  Measurement  values displayed on monitor panel meters are often
    inaccurate  and  are generally less reliable than strip chart
    readings.

 6.  Monitor responses  to internal  zero and span  checks  are  in  excess
    of acceptable  limits in many cases; however, a direct correlation
    between zero and span check results and calibration error  test
    results has not been observed.

7.  Automatic zero compensation circuits generally provide a reliable
    indication of dust accumulation on the transceiver  optics;
    however, some monitor operators are not aware  of  the 4 percent
    opacity zero compensation limit and/or incorrectly  interpret the
    zero compensation value.
                                 21

-------
Dust accumulation  on  exposed optical surfaces is both site- and
monitor-specific and  may occur at different rates for the
transceiver  and reflector components.  Excess dust accumulation  on
optical surfaces was  observed for approximately 19 percent of the
audited monitors.

Calibration  error  test results in excess of the acceptable limits
occurred most  frequently at high opacity levels; the magnitude of
calibration  errors was found to increase as the measured opacity
increased.  Poorly performing monitors tended to provide
measurements which were biased high relative to the correct value.
                           22

-------
                              Recommendations

     The following  recommendations are offered for those source operators
where installed  opacity monitoring systems failed to demonstrate acceptable
performance during  audits.

     1.   The personnel  who  perform the required daily zero and span checks
         should  receive adequate training to allow them to perform routine
         adjustments, as necessary, to eliminate excessive zero drift,  span
         drift,  and discrepancies between the control panel display values
         and the permanent  data recording system.  This training should be
         monitor-specific and  should include adequate explanation, as
         applicable, of automatic zero compensation functions, stack exit
         correction factors, and procedures for determining equivalent
         calibration filter values corrected for stack exit conditions.

     2.   After all  calibration/adjustment problems observed during audits
         are corrected , monitor operators should institute a well  defined
         operation  and maintenance program for the opacity monitoring
         systems.   Such a program must address monitor-specific and
         site-specific considerations, and should at least include the
         following  activities:

         a.  daily  checks,  and as necessary, adjustment of: fault  lamp
             indications, zero drift, span drift, panel meter accuracy,
             data recorder  adjustment, and available monitor-specific
             operational indicators,

         b.  periodic evaluation and, as necessary,  corrective actions  for
             dust accumulation on exposed optical surfaces,  optical
             alignment, reference voltages/curf ents, etc.

         c.  routine maintenance as specified  by the manufacturer  and
             including:  replacement of measurement  lamps,  fault lamp
             bulbs,  purge air filters, dessicant, etc.

         The  appropriate frequency for performing periodic  evaluations and
         routine maintenance should  be established through a  trial  and
         error procedure.  Such a procedure  involves: (1)  selecting an
         initial frequency for these  activities,  (2) performing  periodic
         evaluations and routine maintenance and  documenting  results of
         these activities over a sufficient  period of time to  evaluate
         their impact on monitor performance, and (3) modifying  the
         initially selected  frequency in  view of the  results obtained in
         order to minimize  the expenditure of resources  while  consistently
         monitoring instrument performance within acceptable limits.

     3.   Monitor operators should institute a self auditing  program which
         includes a calibration error  determination  at three points over
         the measurement range of the  instrument. At a minimun, this
         procedure should  be performed  concurrently  with the required
         annual optical  and  zero alignments of  Performance  Specification  1,
         Appendix B, 40 CFR  60.   The  frequency  of self-audits should be
         adjusted as necessary,  to maintain instrument performance within
         acceptable limits.

-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2 '" 	 	 	
EPA- 340/1-83 /Oil
4. TITLE AND SUBTITLE
A Compilation of Opacity Monitor Performance Audit
Results
7. AUTHOR(S)
Steve Plaisance
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Entropy Environmentalists, Inc.
P.O. Box 12291
Research Triangle Park, NC 27709
12. SPONSORING AGENCY NAME AND ADDRESS
OAQPS
Stationary Source Compliance Division
Waterside Mall, 401 M Street, SW
Washington, DC 20460
3. RECIPIENT'S ACCESSION NO.'
5. REPORT DATE
January 1983
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT N
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-01-6317
13. TYPE OF REPORT AND PERIOD COVERE
FINAL - IN-HOUSE
14. SPONSORING AGENCY CODE
EPA/200/04
15. SUPPLEMENTARY NOTES 	
  Opacity monitor performance audit procedures and  devices  have  been developed and
  field tested on 93 opacity monitors.  The  results of  this test program indicate that
  opacity monitoring systems achieve high  levels of availability,  and are capable of
  providing accurate emissions data.  The  results also  show that problems impacting
  data quality are generally limited to monitor miscalibration and/or misadjustment,
  as well as improper or inadequate operating and maintenance practices.   It is be-
  lieved that improved  monitor performance  and data reliability can be  achieved with
  additional training of monitor operators and more frequent performance  audits.
  This document describes the audit program  for continuous  emission  monitors (CEMs)
  of effluent opacity.  Detailed explanations of the audit  methodology, monitor ana-
  lyses, and analytical criteria are included, and  both criteria-  and monitor-specific
  results of installed opacity monitor audits are delineated.  Finally,  conclusions
  are drawn as to the adequacy of monitor  performance and data reliability,  and recom-
  mendations are offered that can optimize opacity  monitoring system performance.-
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Air Pollution
Opacity Monitoring Systems
18. DISTRIBUTION STATEMENT
Release to Public
b. IDENTIFIERS/OPEN ENDED TERMS
Audit Results
19. SECURITY CLASS (This Report)
unclassified
20. SECURITY CLASS (This page)
unclassified
c. COSATI Field/Group

21. NO. OF PAGES
42
22. PRICE
EPA Form 2220-1 (R«v. 4-77)   PREVIOUS EDITION is OBSOLETE

-------

-------
ct-
u
c+
y
o

-------
 United States                              Office of Air Quality Planning and Standards
 Environmental Protection                    Stationary Source Compliance Division
 Agency                                    Washington, D C  20460
Official Business                            Publication No  EPA-340/1-83-011                                                                                              Postage and
Penalty for Private Use                                                                                                                                                     Fees paitj
8300                                                                                                                                                                     Environmental
                                                                                                                                                                         Protection
                                                                                                                                                                         Agency
                                                                                                                                                                         EPA 335
                                                                         If your address is incorrect, please change on the above label,
                                                                         tear off, and return to the above address
                                                                         If you do not desire to continue receiving this technical report
                                                                         series, CHECK HERE D , tear off label, and return it to the
                                                                         above address

-------