EPA-650/4-74-005-0
March 1976 Environmental Monitoring Series
GUIDELINES FOR DEVELOPMENT
OF A QUALITY ASSURANCE PROGRAM:
VOLUME XV - DETERMINATION
OF SULFUR DIOXIDE EMISSIONS
FROM STATIONARY SOURCES
BY CONTINUOUS MONITORS
\
Ul
o
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environ-
mental Protection Agency, have been grouped into series . These broad
categories were established'to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:
1. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2 . ENVIRONMENTAL PROTECTION TECHNOLOGY
3. ECOLOGICAL RESEARCH
4. ENVIRONMENTAL MONITORING
5. SOCIOECONOMIC ENVIRONMENTAL STUDIES
6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
9. MISCELLANEOUS
This report has been assigned to the ENVIRONMENTAL MONITORING
series. This series describes research conducted to develop new or
improved methods and instrumentation for the identification and quanti-
fication of environmental pollutants at the lowest conceivably significant
concentrations. It also includes studies to determine the ambient con-
centrations of pollutants in the environment and/or the variance of
pollutants as a function of time or meteorological factors.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
-------
EPA-650/4-74-005-0
GUIDELINES FOR DEVELOPMENT
OF A QUALITY ASSURANCE PROGRAM:
VOLUME XV - DETERMINATION
OF SULFUR DIOXIDE EMISSIONS
FROM STATIONARY SOURCES
BY CONTINUOUS MONITORS
by
Pamela Wohlschlegel
Research Triangle Institute
Research Triangle Park, North Carolina 27709
Contract No. 68-02-1234
ROAP No. 26BGC
Program Element No. 1HA327
EPA Project Officer: Steven M. Bromberg
Environmental Monitoring and Support Laboratory
Office of Monitoring and Technical Support
Research Triangle Park, North Carolina 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
March 1976
-------
EPA REVIEW NOTICE
This report has been reviewed by the Environmental Monitoring and
Support Laboratory, Research Triangle Park, North Carolina, of the
U. S. Environmental Protection Agency and approved for publication.
Approval does not signify that the contents necessarily reflect
the views and policies of the Environmental Protection Agency, nor
does mention of trade names or commercial products constitute en-
dorsement or recommendation for use.
-------
TABLE OF CONTENTS
Page
List of Figures iv
List of Tables v
Abstract vl
1.0 INTRODUCTION 1-1
2.0 DETERMINATION OF SULFUR DIOXIDE EMISSIONS 2-1
FROM STATIONARY SOURCES BY CONTINUOUS MON-
ITORS
5.11 Determination of Sulfur Dioxide Emissions 2-2
from Stationary Sources by Continuous
Monitors
5.11.1 Performance Specification 2 2-4
5.11.2 Plan Activity Matrix 2-15
5.11.3 Operational Procedures 2-25
5.11.4 Auditing Procedures 2-75
3.0 FUNCTIONAL ANALYSIS 3-1
3.1 Definition and Discussion of Error Sources 3-3
3.1.1 In-Situ Monitoring Systems 3-3
3.1.2 Extractive Monitoring Systems 3-12
LIST OF REFERENCES R-l
Appendixes
A PERFORMANCE SPECIFICATION 3 A-l
B GLOSSARY OF SYMBOLS B-l
C GLOSSARY OF TERMS C-l
Xll
-------
LIST OF FIGURES
Figure Description Page
5.11.1 Method procedures relationships 2-16
5.11.2 Operational flow chart 2-26
5.11.3 Example of an extractive sulfur dioxide 2-43
monitoring system
5.11.4 Sample data form for the analysis of the 2-53
calibration gas mixture
5.11.5 Sample log sheet for calibration error 2-54
test data
5.11.6 Sample data form for operational test 2-56
5.11.7 Sample data form for response time test 2-60
5.11.8 Sample form for data analysis report cal- 2-61
culations
5.11.9 Sample control chart for use in determin- 2-72
ing trends in zero drift
5.11.10 Sample operation and maintenance table 2-73
5.11.11 Sample audit data form 2-78
5.11.12 Block diagram of a typical in-situ monitor- 3-4
ing system
5.11.13 Block diagram of a typical extractive moni- 3-4
toring system
5.11.14 Major components of system errors in an in- 3-6
situ monitoring system
5.11.15 Major components of system errors in an ex- 3-13
tractive monitoring system
IV
-------
LIST OF TABLES
Table Description Page
5.11.1 Methods of monitoring variables 2-14
5.11.2 Values for t __,. 2-67
-------
ABSTRACT
Guidelines for the quality control of the continuous
measurement of sulfur dioxide emissions by Performance
Specification 2 are presented. These include:
1. Good operating practices;
2. Directions on how to assess performance and to
qualify data;
3. Directions on how to identify trouble and to im-
prove data quality;
4. Directions to permit design of auditing activities.
The document is not a research report. It is designed
for use by operating personnel.
This work was submitted in partial fulfillment of con-
tract Durham 68-02-1234 by Research Triangle Institute under
the Sponsorship of the Environmental Protection Agency.
Work was completed as of January, 1976.
Vl
-------
1.0 INTRODUCTION
1-1
-------
1.0 INTRODUCTION
This document presents guidelines for implementing a
quality assurance program for the measurement of sulfur
dioxide emissions from stationary sources by continuous
monitors.
The objectives of this quality assurance program are to:
1. 'Provide routine indications for operating pur-
poses of satisfactory performance of personnel
>
and/or equipment,
2. Provide for pirompt detection and correction of
conditions that contribute to the collection of
poor quality data,
3. Collect and supply information necessary to
describe the quality of the data.
To accomplish the above; objectives, a quality assurance program
must contain the following components:
1. Routine training and/or evaluation of operators,
2. Routine monitoring of the variables and/or
parameters which may have a significant effect
on data quality,
3. Development of techniques to detect defects,
4. Development of methods to qualify data,
5. Action strategies to increase the level of pre-
cision in the reported data and/or to detect
equipment defects or degradation and to correct
same.
Implementation of a quality assurance program will result in
data that are more preferable in terms of precision and accuracy
and will enable the monitoring network to continuously generate
data of acceptable quality.
1-2
-------
This document is divided into three sections, as follows:
1. 1.0 INTRODUCTION—This section lists the overall objec-
tives of a quality assurance program and delineates the
program components necessary to accomplish the given
objectives.
2. 2.0 DETERMINATION OF SULFUR DIOXIDE EMISSIONS FROM
STATIONARY SOURCES BY CONTINUOUS MONITORS—This
section includes the method description, plan
activity matrix, operational procedures, and
auditing procedures. Subsection numbering is con-
sistent with section 5.11 of a larger document, The
Quality Assurance Handbook for Air Pollution Measure-
ments , Volume III, Stationary Sources Specific Methods.
3. 3.0 FUNCTIONAL ANALYSIS—This section includes an
estimate of the precisions and biases of the various
measurement components which collectively comprise
the continuous monitoring method, and an estimate of
the total precision and bias of the method.
1-3
-------
2.0 DETERMINATION OF SULFUR DIOXIDE EMISSIONS
FROM STATIONARY SOURCES BY CONTINUOUS
MONITORS*
*
The numbering in this section is consistent with
section 5.11 of a larger document, The Quality
Assurance Handbook for Air Pollution Measurements,
Vol. Ill—Stationary Sources Specific Methods.
2-1
-------
5.11 DETERMINATION OF SULFUR DIOXIDE EMISSIONS FROM
STATIONARY SOURCES BY CONTINUOUS MONITORS
This section describes a quality control/quality assur-
ance program designed specifically to ensure that sulfur
dioxide data generated from continuous stack monitoring systems
is of acceptable quality. To present the program, the section
has been divided into the following subsections:
5.11.1 Performance Specification 2,
5.11.2 Plan Activity Matrix,
5.11.3 Operational Procedures, and
5.11.4 Auditing Procedures.
A reprint of "Performance Specification 2--Performance
Specifications and Specification Test Procedures for Monitors
of SO- and NO from Stationary Sources" is included in section
£• X.
5.11.1 for convenient reference. The equipment specifications
and procedural guidelines detailed in this quality assurance
document were based on statements made in the Performance
Specification 2. The scope of this document, however, does
not include discussions concerning oxides of nitrogen monitors
even though they are covered in Performance Specification 2.
Also contained in section 5.11.1 is a block diagram which
delineates the procedures for continuously measuring sulfur
dioxide emissions.
The Plan Activity Matrix in section 5.11.2 includes the
characteristic checks for each of the operating procedures.
The operating procedures themselves are described in some de-
tail in section 5.11.3. The audit procedures described in
section 5.11.4 provide a basis for assessing the performance
in each operational category.
The method described in this document for the measure-
ment of sulfur dioxide emissions can be broken down into the
2-2
-------
following series of operational categories:
1. Definition of monitoring requirements and system
hardware,
2. Installation and testing,
3. Preventive maintenance and routine testing.
Within each category both routine and nonroutine procedures
are required. The routine procedures form the basic criteria
for monitoring sulfur dioxide at a given stack and obtaining
data. The nonroutine procedures provide characteristic checks
for each of the routine procedures to insure that valid data
will be ultimately collected from the system and to properly
assess the quality of that data. Such nonroutine procedures
are elements of an internal quality assurance or quality con-
trol program. The audit procedures are elements of an exter-
nal quality assurance program (sometimes referred to simply
as quality assurance). The purpose of section 5.11 is to
provide the user with guidelines for developing quality con-
trol and quality assurance programs for a given source mon-
itoring system.
2-3
-------
5.11.1 Performance Specification
PERFORMANCE SPECIFICATION 2—PERFORMANCE
SPECIFICATIONS AND SPECIFICATION TEST PRO-
CEDURES FOR MONITORS OF SOO AND NO FROM
*
STATIONARY SOURCES
Reproduced from Appendix B, "Performance Specifications,"
of 40 CFR Part 60, Federal Register, Volume 40, Number 194;
Monday, October 6, 1975.
Only the determination of sulfur dioxide (SO2) emissions
by continuous monitors are discussed in this quality
assurance document.
2-4
-------
5.11.1.1 Principle and Applicability
1. Principle and Applicability.
1.1 Principle. The concentration or sulfur
dioxide or oxides of nitrogen pollutants In
stack emissions is measured by a continu-
ously operating emission measurement sys-
tem. Concurrent with operation of the con-
tinuous monitoring system, the pollutant
concentrations are also measured with refer-
ence methods (Appendix A). An average of
the continuous monitoring system data is
computed for each reference method testing
period and compared to determine the rela-
tive accuracy of the continuous monitoring
system. Other tests of the continuous mon-
itoring system are also performed to deter-
mine calibration- error, drift, and response
characteristics of the system.
1.2 Applicability. This performance spec-
ification is applicable to evaluation of con-
tinuous monitoring systems for measurement
of nitrogen oxides or sulfur dioxide pollu-
tants. These specifications contain test pro—
cedures, installation requirements, and data
computation procedures for evaluating the
acceptability of the continuous monitoring
systems.
5.11.1.2 Apparatus
2. Apparatus.
2.1 Calibration Gas Mixtures. Mixtures of
known concentrations of pollutant gas in a
diluent gas shall be prepared. The pollutant
gas shall be sulfur dioxide or the appropriate
oxide(s) of nitrogen specified by paragraph
6 and within subparts. For sulfur dioxide gas
mixtures, the diluent gas may be air or nitro-
gen. For nitric oxide (NO) gas mixtures, the
diluent gas shall be oxygen-free «10 ppm)
nitrogen, and for nitrogen dioxide (NO,) gas
mixtures the diluent gas shall be air. Concen-
trations of approximately 50 percent and 90
percent of span are required. The 90 percent
gas mixture Is used to set and to check the
span and Is referred to as the span gas.
2.2 Zero Gas. A gas certified by the manu-
facturer to contain less than 1 ppm of the
pollutant gas or ambient air may be-used.
2.3 Equipment for measurement of the pol-
lutant gas concentration using the reference
method specified in the applicable standard.
2.4 Data Recorder. Analog chart recorder
or other suitable device with Input voltage
range compatible with analyzer system out-
put. The resolution of the recorder's data
output shall be sufficient to allow completion
of the test procedures within this specifi-
cation.
2.5 Continuous monitoring system for SO,
or NO* pollutants as applicable.
5.11.1.3 Definitions
3. Definitions.
3.1 Continuous Monitoring System. The
total equipment required for the determina-
tion of a pollutant gas concentration ln_ a
source effluent. Continuous monitoring sys-
tems consist of major subsystems as follows:
3.1.1 Sampling Interface—That portion of
an extractive continuous monitoring system
that performs one or more of the following
operations: acquisition, transportation, and
conditioning of a sample of the source efflu-
ent or that portion of an in-situ continuous
monitoring system that protects the analyzer
from the effluent.
3.1.2 Analyzer—That portion of the con-
tinuous monitoring system which senses the
pollutant gas and generates a signal output
that is a function of the pollutant concen-
tration.
3.1.3 Data Recorder—That portion of the
continuous monitoring system that provides
a permanent record of the output signal In
terms of concentration units
3.2 Span. The value of pollutant concen-
tration at which the continuous monitor-
ing system Is set to produce the maximum
data display output. The span shall be set
at the concentration specified In each appli-
cable subpart.
3.3 Accuracy (Relative). The degree of
correctness with which the continuous
monitoring system yields the value of gas
concentration of a sample relative to the
value given by a defined reference method.
This accuracy is expressed in terms of error,
which is the difference between the paired
concentration measurements expressed as a
percentage of the mean reference value.
3.4 Calibration'Error. The difference be-
tween the pollutant concentration indi-
cated by the continuous monitoring system
and the known concentration of the test
gas mixture.
3.5 Zero Drift. The change in the continu-
ous monitoring system output over a stated
2-5
-------
5.11.1.3 Definitions (continued)
period of time of normal continuous opera-
tion when the pollutant concentration at
the time for the measurements Is zero.
3 6 Calibration Drift. The change in the
continuous monitoring system output over
a stated time period of normal continuous
operations when the pollutant concentra-
tion at the time of the measurements Is the
same known upscale value.
3 7 Response Time. The time interval
from a step change In pol'utant concentra-
tion at the input to the continuous moni-
toring system to the time at which 95 per-
cent of the corresponding final value Is
reached as displayed on the continuous
monitoring system data recorder.
3.8 Operational Period. A minimum period
of time over which a measurement system
is expected to operate within certain per-
formance specifications without unsched-
uled maintenance, repair, or adjustment.
3.9 Stratification. A condition identified
by a difference In excess of 10 percent be-
tween the average concentration In the duct
or stack and the concentration at any point
more than 1.0 meter from the duct or stack
wall.
5.11.1.4 Installation Specifications
4. Installation Specifications. Pollutant
continuous monitoring systems (SO, and
NOX) shall be installed at a sampling loca-
tion whare measurements can be made which
are directly representative (4.1), or which
can be corrected so as to be representative
(4.2) of the total emissions, from the affected
facility. Conformance with this requirement
shall be accomplished as follows:
4.1 Effluent gases may be assumed to be
nonstratifled if a sampling; location eight or
more stack diameters (equivalent diameters)
downstream of any air in-leakage is se-
lected. This assumption and data correction
procedures under paragraph 4.2.1 may not
be applied to sampling locations upstream
of an air preheater in a stream generating
facility under Subpart D of this part. For
sampling locations where effluent gases are
either demonstrated (43) or may be as-
sumed to be nonstratified (eight diameters),
a point (extractive systems) or path (in-sltu
systems) of average concentration may be
monitored.
4.2 For sampling locations where effluent
gases cannot be assumed to be nonstrati-
fied (less than eight diameters) or have been
shown under paragraph 4 3 to be stratified,
results obtained must be consistently repre-
sentative (e.g. a point of e.verage concentra-
tion may shift with load changes) or the
data generated by sampling at a point (ex-
tractive systems) or across a path (in-situ
systems) must be corrected (4.2.1 and 4 2.2)
so as to be representative of the total- emis-
sions from the affected facility. Conlorm-
ance with this requlremert may be accom-
plished in either of the following ways:
4.2.1 Installation of a diluent continuous
monitoring system (O» or CO, as applicable)
in accordance with the procedures under
paragraph 4.2 of Performance Specification
3 of this appendix. If the pollutant and
diluent monitoring systems are not of the
same tvpe (both extractive or both. In-sltu),
the extractive system must use a multipoint
probe.
4 2.2 Installation of extractive pollutant
monitoring systems using multipoint sam-
pling probes or in-situ pollutant monitoring
systems that sample or view emissions which
are consistently representative of the. total
emissions for the entire cross section. The
Administrator may require data to be sub-
mitted to demonstrate that the emissions
sampled or viewed are consistently repre-
sentative for several typical facility process
operating conditions.
4.3 The owner or operator may perform a
traverse to characterize any stratification of
effluent gases that might exist in a stack or
duct. If no stratification is present, sampling
procedures under paragraph 4.1 may be ap-
plied even though the eight diameter criteria
is not met.
4 4 When single point sampling probes for
extractive systems are Installed within the
stack or duct under paragraphs 4.1 and 4.2.1,
the sample may not be extracted at any point
less than 1.0 meter from the stack or duct
wall. Multipoint sampling probes installed
under paragraph 4.2.2 may be located at any
points necessary to obtain consistently rep-
resentative samples.
5.11.1.5 Continuous Monitoring System Performance Specifications
5. Continuous Monitoring System Perform-
ance Specifications.
The continuous monitoring system shall
meet the performance specifications in Table
2-1 to be .considered acceptable under this
method.
2-6
-------
5.11.1.5 Continuous Monitoring System Performance Specifications
TABLE 2-1.—Performance specification.*!
Parameter
Specification
1 Accuracy ' --- ~^0 pet of the mean value nf the reference method test
"" ~ " " daia
2. Calibration error '
... .. _ < spot of each (50 pet, U0 pet) calibration gas mixture
value.
3. Zero drift (2 h) i 2 pet of span
4. Zero drift (24 h) 1 - --- --- Do.
5. Calibration drift (2 h) 1.. - --- --- Do
6. Calibration drift (24 h) i - 2.5 pet of span
7. Kesponso time ---- 15 nun maximum
8 Operational period. - 168 h minimum.
i Expressed as sum of absolute mean value plus 95 pet confidence interval of a series of tests
5.11.1.6 Performance Specification Test Procedures
6. Performance Specification Test Proce-
dures. The following test procedures shall be
used to determine conformance with the
requirements of paragraph 5. For NOX an-
requirements of paragraph 5. For NCK an-
alyzers that oxidize nitric oxide (NO) to
nitrogen dioxide (NO2), the response time
test under paragraph 6.3 of this method shall
be performed using nitric oxide (NO) span
gas. Other tests for NO« continuous monitor-
Ing systems under paragraphs 6.1 and 6.2 and
all tests for sulfur dioxide systems shall be
performed using the pollutant span gas spe-
cified by each subpart.
6.1 Calibration Error Test Procedure. Set
up and calibrate the complete continuous
monitoring system according to the manu-
facturer's writen instructions. This may be
accomplished either In the laboratory or In
the field.
6 1.1 Calibration Gas Analyses, Triplicate
analyses of the gas mixtures shall be per-
formed within two weeks prior to use using
Reference Methods 6 for SO, and 7 for NO,.
Analyze each calibration gas mixture (50%,
90 %,) and record the results on the example
sheet shown in Figure 2-1 Each sample test
result must be within 20 percent of the aver-
aged result or the tests shall be repeated.
This step may be omitted for non-extractive
monitors where dynamic calibration gas mix-
tures are not used (6.1.2).
6.1.2 Calibration Error Test Procedure.
Make a total of 15 nonconsecutive measure-
ments by alternately using zero gas and each
caliberatlon gas mixture concentration (e.g.,
0%, 50%, 0';", 90%, 50%, 90%, 50%, 0%,
etc ). For nonextractive continuous monitor-
ing systems, this test procedure may be per-
formed by using two or more calibration gas
cells whose concentrations are certified by
the manufacturer to be functionally equiva-
lent to these gas concentrations. Convert the
continuous monitoring system output read-
ings to ppm and record the results on the
example sheet shown In Figure 2-2.
6.2 Field Test for Accuracy (Relative),
Zero Drift, and Calibration Drift. Install and
operate the continuous monitoring system In
accordance with the manufacturer's written
instructions and drawings as follows:
6.2.1 Conditioning Period Offset the zero
setting at least 10 percent of the span so
that negative zero drift can be quantified.
Operate the system for an initial 168-hour
conditioning period in Formal operating
manner.
6 2.2 Operational Test Period Operate the
continuous monitoring system for an addi-
tional 168-hour period retaining the zero
offset The system shall monitor the source
effluent at all times except when being
zeroed, calibrated, or backpurged.
6 2.2.1 Field Test for Accuracy (Relative).
For continuous monitoring systems employ-
ing extractive sampling, the probe tip for the
continuous monitoring system and the probe
tip for the Reference Method sampling train
should be placed at adjacent locations In the
duct. For NOX continuous monitoring sys-
tems, make 27 NOX concentration measure-
ments, divided Into nine sets, using the ap-
plicable reference method. No more than one
set of tests, consisting of three Individual
measurements, shall be performed In any
one hour. All Individual measurements of
each set shall be performed concurrently,
or within a three-minute interval and the
results averaged. For SO2 continuous moni-
toring systems, make nine SO2 concentration
measurements using the applicable reference
method. No more than one measurement
shall be performed In any one hour. Record
the reference method test data and the con-
tinuous monitoring system concentrations
on the example data sheet shown in Figure
2-3.
6.2 2.2 Field Test for Zero Drift and Cali-
bration Drift. For extractive systems, deter-
mine the values given by zero and span gas
pollutant concentrations at two-hour Inter-
vals until 15 sets of data are obtained. For
nonextractive measurement systems, the zero
value may be determined by mechanically
producing a zero condition that provides a
system check of the analyzer Internal mirrors
and all electronic circuitry including the
radiation source and detector assembly or
by Inserting three or more calibration gas
cells and computing the zero point from the
upscale measurements. If this latter tech-
nique is used, a graph(s) must be retained
by the owner or operator for each measure-
ment system that shows the relationship be-
tween the upscale measurements and the
zero point. The span of the system shall be
checked by using a calibration gas cell cer-
tified by the manufacturer to be function-
ally equivalent to 50 percent of span concen-
tration Record the zero and span measure-
ments (or the computed zero drift) on the
example data sheet shown in Figure 2-4.
The two-hour periods over which measure-
ments are conducted need not be consecutive
but may not overlap All measurements re-
quired under this paragraph may be con-
ducted concurrent with tests under para-
graph 6.2.2.1.
2-7
-------
5.11.1.6 Performance Specifications Test Procedures (continued)
Reference ''etriod Used
ti^h-gange (span| Calibration Gas Mixture
SiTple J pprc
Sample Z ppnt
Sarple 3 ppm
Average _ppn
Flgars 2-1. Analysis of CaHbratlon Gas
RULES AND REGULATIONS
Calibration Gas Mixture Data (From Figure 2-1)
Mid (50%) ppm High (90%) ppm
Calibration Gas
Run # Concentration,ppm
Measurement System
Reading, ppm
Differences, ppm
10
11
12
13
14
15
Mean difference
Confidence interval
Calibration error =
Mid High
Pean Difference + C.I. x 100
Average Calibration Gas Concentration I
Calibration gas concentration - measurement system reading
>
"Absolute value
Figure 2-2. Calibration Error Determination
-------
5.11.1.6 Performance Specifications Test Procedure (continued)
Test
Ho.
1
2
3
4
E
6
7
8
9
lean
test
lean
;bK C
\ccur
• Exp
" Mi
Date
and
Time
reference IT
value (S02
differences
onf idence
,,.r, - He
Reference Method Samples
so
Sample 1
(ppm)
ethod
** =
ntervals - «
Sampfe 1
(ppm)
+
NO
Sample 2
(ppii)
NO
Sanpfe 3
(ppm)
M3 Sa-npl e
Average
(ppm)
Mean reference method
test value (NO )
ppm (SO,), s
ppm
Analyzer 1-Hour
Average (ppm)*
SO, N0x
D1 f ference
(ppm)
S02 N0x
1
Average of
the differences
ppm (NO..).
(50,). " *
an difference {absolute va ue) * 95% confidence Interval 1nn
Jcc' Mean reference method value " '"" -
lain and report method used to determine integrated averages.
an differences = the average of the differences minus the mean reference meth
ppm (NO ).
X {S02), » « (NOX).
od test value.
Figure 2-3. Accuracy Determination (SO^ and NOX)
Zero Span Callo'-c'tlon
Time Ze-o Drift Sps-. Drift CrUt
Begin End Ca'.e Feadlng ('.Zero) RejO'r-g (tSpan) ( S?=n- Ze'o)
Zero Dr~iff~=~ {Mean Zero Drift*
Cal ib-'iition Crift - [.Xedn Span T5~nt't"'r~
x 100 =
-! [Spa"] x 100 = .
Figure 2-4. Zero and" Calibration Drift (2 Hourp
2-9
-------
5.11.1.6 Performance Specification Test Procedures (continued)
6.2.2.3 Adjustments. Zero and calibration
corrections and adjustments are allowed only
at 24-hour Intervals or at such shorter In-
tervals as the manufacturer's written In-
structions specify. Automatic corrections
made by the measurement system without
operator Intervention or initiation are allow-
able at any time. During the entire 168-hour
operational test period, record on the ex-
ample sheet shown In Figure 2-5 the values
given by zero and span gas pollutant con-
centrations before and after adjustment at
24-hour Intervals.
6.3 Field Test for Response Time.
6 3.1 Scope of Test. Use the entire continu-
ous monitoring system as Installed, including
•sample transport lines If used. Flow rates,
line diameters, pumping rates, pressures (do
not allow the pressurized calibration gas to
change the normal operating pressure In the
sample line), etc., shall be at the nominal
values for normal operation as specified in
the manufacturer's written instructions. If
the analyzer Is used to sample more than one
pollutant source (stack), repeat this test for
each sampling point.
6.3.2 Response Time Test Procedure. In-
troduce zero gas into the continuous moni-
toring system sampling interface or as close
to the sampling Interface as possible. When
the system output reading has stabilized,
switch quickly to a known concentration of
pollutant gas. Record the time from concen-
tration switching to 95 percent of final stable
response. For non-extractive monitors, the
highest available calibration gas concentra-
tion shall be switched into and out of the
sample path and response times recorded.
Perform this test sequence three (3) times.
Record the results of each test on the
example sheet shown In Figure 2-6.
Date Zero Span Calibration
and Zero Drift Reading Drift
Time Reading (AZero) (After zero adjustment) (ASpan)
Zero Drift = [Mean Zero Drift*
C.I. (Zero)
•i [Instrument Span] x TOO =
Calibration Drift = [Mean Span Drift*
+ C.I. (Span)
« [Instrument Span] x TOO =
Absolute value
Figure 2-5. Zero and Calibration Drift (24-hour)
2-10
-------
5.11.1.6 Performance Specification Test Procedures (continued)
Date of Test
Span Gas Concentration
Analyzer Span Setting
Upscale
Average
Downscale
1
2
3
upscale
1
2
3
ppir.
ppm
seconds
seconds
seconds
response seconds
seconds
seconds
seconds
Average downscale response seconds
System average response time (slowe
^deviation from slower _ (averaoe
system average response
r time) = seconds.
upscale minus average downscale x ^gg, _
slower time \
Figure 2-6. Response Time
5.11.1.7 Calculations, Data Analysis, and Reporting
7. Calculations, Data Analysis and Report-
ing.
7.1 Procedure for determination of mean
values and confidence intervals.
71.1 The mean value of a data set is
calculated according to equation 2-1.
1 X r
n i=1 Equation 2-1
where:
xt=: absolute value of the measurements,
S = sum of the individual values,
x = mean value, and
n = number of data points.
712 The 95 percent confidence interval
(two-sided) Is calculated according to equa-
tion 2-2-
f t ^
n\/n—T
Equation 2-2
where:
£x, = sum of all data points,
t975 = ti-a/2, and
C 1.95 = 95 percent confidence interval
estimate of the average mean
value.
Values for '.975
10
.
13 .........
14 ________
15 ........ _
16..
975
706
303
182
776
571
447
365
306
262
228
201
179
160
145
131
The values in this table are already cor-
rected for n-1 degrees of freedom Use n
equal to the number of samples as data
points.
7 2 D:ita Analysis and Reporting
721 Accur.icy (Relative) . For each of the
nine le'crence method test points, determine
tho average pollutant concentration reported
by the continuous monitoring system. These
average concentrations shall be determined
fioiu the continuous monitoring system data
recorded under 722 by integia'ing or aver-
amn^' the pollutant concentrations over each
ot the time intervals concurrent with each
reference method testing period Before pro-
ceeding to the next step, determine the basis
(wet or dry) or the continuous monitoring
sjstem data and icference method test data
concentrations. If the bases are not con-
sistent, apply a moisture correction to either
reference method concentrations, or the con-
tinuous monitoring sj stern concentrations
as appropriate Determine the correction
fat tor liy moisture tests concurrent with the
reference method testing periods Report the
moisture test method and the correction pro-
cedure emplojed. For each of the nine test
runs determine the difference for each test
run by subtracting the respective leference
method test concentrations (use average of
eich set of three measurements for NCK)
from the continuous monitoring system inte-
grated or averaged concentrations Using
these data, compute the mean difference and
the 95 percent confidence interval of the dif-
ferences (equations 2-1 and 2-2). Accuracy
is reported as the sum of the absolute value
of the mean difference and the 95 percent
confidence interval of the differences ex-
pressed as a percentage of the mean refer-
ence method value. Use the example sheet
shown in Figure 2-3.
722 Calibration Error. Using the data
from paragraph C.I, subtract the measured
pollutant concentration determined under
2-11
-------
5.11.1.7 Calculations, Data Analysis, and Reporting^(Continued)
paragraph 6.1.1 (Figure 2-1) from the value
shown by the continuous monitoring system
for each of the five readings at each con-
centration measured under 6.1.2 (Figure 2-2).
Calculate the mean of these difference values
and the- 95 percent confidence intervals ac-
cording to equations 2-1 and 2-2. Report the
calibration error (the sum of the absolute
value of the mean difference and the 95 per-
cent confidence 'interval | as a percentage of
each respective calibration gas concentra-
tion. Use example sheet sihown in Figure 2-2.
7.2.3 Zero Drift (2-hour). Using the zero
concentration values measured each two
hours during the field test, calculate the dif-
ferences between consecutive two-hour read-
Ings expressed in ppm. Calculate the mean
difference and the confidence Interval using,
equations 2-1 and 2-2. Report the zero drift
as the sum of the absolute mean value and
the confidence interval as a percentage of
span. Use example sheet shown in Figure
2-4.
72.4 Zero Drift (24-hour). Using the zero
concentration values measured every 24
hours during the field test, calculate the dif-
ferences between the zero point after zero
adjustment and the zero value 24 hours later
just prior to zero adjustment. Calculate the
mean value of these points and the confi-
dence interval using equations 2-1 and 2-2.
Report the zero drift (the sum of the abso-
lute mean and confidence1 interval) as a per-
centage of span. Use example sheet shown in
Figure 2-5.
7.2 5 Calibration Drift (2-hour). Using
the calibration values obtained at two-hour
intervals during the field test, calculate the
differences between consecutive two-hour
readings expressed as ppm. These values
should be corrected for the corresponding
zero drift during that two-hour period. Cal-
culate the mean and confidence interval of
these corrected difference values using equa-
tions 2-1 and 2-2. Do not use the differences
between non-consecutive1 readings. Report
the calibration drift as the sum of the abso-
lute mean and confidence Interval as a per-
centage of span. Use the example sheet shown
in Figure 2-4.
72.6 Calibration Drift (24-hour). Using
the calibration values measured every 24
hours during the field test, calculate the dif-
ferences between the calibration concentra-
tion reading after zero and calibration ad-
justment, and the calibration concentration
reading 24 hours later after zero adjustment
but before calibration adjustment. Calculate
the mean value of these differences and the
confidence interval using equations 2-1 and
2-2. Report the calibration drift (the sum of
the absolute mean and confidence interval)
as a percentage of span. Use the example
sheet shown in Figure 2-5.
7.2.7 Response Time. 'Using the charts
from paragraph 6.3, calculate the time inter-
val from concentration switching to 95 per-
cent to the final stable value for all upscale
and downscale tests. Report the mean of the
three upscale test times and the mean of the
three downscale test times. The two aver-
age times should not differ by more than 15
percent of the slower time. Report the slower
time as the system response time. Use the ex-
ample sheet shown in Figure 2-6.
7.2.8 Operational Test Period. During the
168-hour performance and operational test
period, the continuous monitoring system
shall not require any corrective maintenance,
repair, replacement, or adjustment other than
that clearly specified as required in the op-
eration and maintenance manuals as routine
and expected during a one-week period. If
the continuous monitoring system operates
within the specified performance parameters
and does not require corrective maintenance.
repair, replacement or adjustment other than
as specified above during the 168-hour test
period, the operational period will be success-
fully concluded. Failure of the continuous
monitoring system to meet this requirement
shall call for a repetition of the 168-hour test
period. Portions of the test which were satis-
factorily completed need not be repeated.
Failure to meet any performance specifica-
tions shall call for a repetition of the one-
week performance test period and that por-
tion of the testing which is related to the
failed specification. All maintenance and ad-
justments required shall be recorded. Out-
put readings shall be recorded before and
after all adjustments.
5.11.1.3 References
8. References.
8.1 "Monitoring Instrumentation for the
Measurement of Sulfur Dioxide in Stationary
Source Emissions," Environmental Protection
Agency, Research Triangle Park, N.C., Feb-
ruary 1973.
8.2 "Instrumentation for the Determina-
tion of Nitrogen Oxides Content of Station-
ary Source Emissions," Environmental Pro-
tection Agency, Research Triangle Park, N.C.,
Volume 1, APTD-0847, October 1971; Vol-
ume 2, APTD-0942, January 1972.
8.3 "Experimental Statistics," Department
of Commerce, Handbook 91, 1963, pp. 3-31,
paragraphs 3-3.1.4.
8 4 "Performance Specifications for Sta-
tionary-Source Monitoring Systems for Gases
and Visible Emissions," Environmental Pro-
tection Agency, Research Triangle Park, N.C.,
EPA-650/2-74-013, January 1974.
2-12
-------
5.11.1.9 Quality Control Program Specifications--
5.11.1.9.1 Quality control programs — purpose. Each
laboratory and field station shall maintain a quality control
program consistent with the procedures described in this
manual. Such programs shall be designed to assure the qual-
ity and scientific reliability of laboratory and field data
used in EPA activities and documents.
5.11.1.9.2 Quality control program—provisions. Each
quality control program shall contain provisions for the
management of data quality, which shall include the following:
1. Requirements for the production of quality con-
trol data and the use of quality control records;
2. Control of technical documents, testing methods,
and calibration instructions, and changes
thereto;
3. Control and calibration of test and analytical
instruments and equipment used in certification
testing;
4. Control of purchased equipment to include checks
for analyzer conformance to specifications and
continuing surveillance of that conformance;
5. Audit of routine calibration and preventive main-
tenance schedules;
6. Establishment of an organizational structure to
carry out these provisions.
The intralaboratory quality control program shall provide
for the establishment and maintenance of total quality control
systems to assure continued precision and accuracy of the
instrument data including, as appropriate, requirements for:
1. The checks and audits listed in table 5.11.1,
2. Routine use of an appropriate reference method (either
Method 6 or 8) for replicate test data,
2-13
-------
Table 5.11.1 Methods of monitoring variables
Variable
Method of monitoring
1. Calibration gas
2.. Zero drift
Span drift
4. Response time
5..
Dynamic calibra-
tion
6,
Replicate sam-
pling (audit)
7.
Data processing
check
Analyze calibration gases using
Method 6 as part of routine operat-
ing procedures or as the need is indi-
cated by other checks. See section
5.11.3.5 page 2-51.
Make zero check and adjustment as
part of daily operating procedure.
See section 5.11.3.7 page 2-69.
Make span check and adjustment as
part of daily operating procedure
cind plot on control chart. See sec-
tion 5.11.3.7 page 2-69.
Check response time as part of bi-
weekly operating procedure. See
section 5.11.3.7 page 2-69.
Perform a dynamic three-point cali-
bration as part of bi-weekly opera-
tion procedure or as needed due to
major system modifications. See
section 5.11.3.7 page 2-69.
Manually sample the stack gas, ana-
lyze the sample, and determine a
sulfur dioxide concentration as
directed by the appropriate reference
method. See section 5.11.4 page 2-75
Manually check a randomly chosen week
of data for proper use of calibration
data, reasonableness of the averaged
data, and correct transfer functions.
See section 5.11.4 page 2-75.
2-14
-------
3. The use of X (mean) or R (range) or other
control charts and tests for significance of
differences to gain more information about
troublesome or out-of-control tests or
methods.
Figure 5.11.1 illustrates in a block diagram the method
procedural relationships for Performance Specification 2 as
applies to the measurement of sulfur dioxide.
5.11.2* Plan Activity Matrix
5.11.2.1 Purpose—Quality control procedures and
checks are designed to identify invalid data and to assure
that the reported data are of acceptable quality. These pro-
cedures and checks should be integral parts of the normal
operational functions. Special quality control checks or pro-
cedures may be required of the field operator if deemed neces-
sary by the supervisor in charge of the field operations. The
need for such exercises would be a function of the type of
source and the adequacy of the monitoring system.
Instructions for the continuous measurement of sulfur
dioxide are directed primarily toward 2-week performance tests
during which the compliance of the monitored source is deter-
mined. The continuous data generated from the monitoring
system other than during the performance test period provide
a daily indication of the operation and maintenance of the
process and emission control equipment. An auditing level of
one audit check every six months and three audits per performance
test period is recommended in this document. All apparatus should
satisfy Performance Specification 2 (section 5.11.1), and manufac-
turer's recommendations should be followed whenever possible.
This section presents, in tabular matrix form, a synop-
sis of control procedures for important sources of variation
2-15
-------
r\i
Cfl
to
C
O
•H
-P
W
0)
M
3
13
QJ
U
O
M
a
T)
O
X!
4-)
-------
associated with the operating procedures. Each important
measurement or control procedure is identified in a block at
the left of the tabular matrix. The information in the blocks
across the page indicates for each procedure: the acceptance
limits, the frequency of the check or measurement, the action
to be taken if requirements are not met, and the disposition
of records. Stepwise instructions for performing each oper-
ation in the matrix appear in the corresponding subsection of
section 5.11.3, "Operation Procedures," or, in the case of
audit instructions, in section 5.11.4, "Auditing Procedures."
The operational categories have been subdivided into the
following operational areas, which more readily lend them-
selves to a tabular formation:
5.11.2.2 System Definition and Procurement
5.11.2.3 Installation and Instrument Check
5.11.2.4 Performance Testing
5.11.2.5 Data Reduction, Validation, and Reporting
5.11.2.6 Operating Procedures for Non-Compliance
Monitoring
5.11.2.7 Auditing Procedures.
2-17
-------
X
H
EH
H
EH
U
04
X
O
0)
u
M-l
0
TJ
o
o
01
OS
1
0)
"H 0)
oTs
Oi "*"*
J*
M-l 2
tn
M 4J
0 ^
•H £
4J
C
0)
E
J_j
3
0
0
ft
TJ
IB
C
O
•H
4J
•H
C
•H
M-l
0)
Q
£
4-1
01
^,
CO
(N
•
CN
.
rH
rH
.
LT
4J =
o
"^C
4J
M-l C
O 01
E
TJ 01
0 Vi
x; 3
4-1 tn
01 rd
E 0)
^g"
^
85
_c
o ^
^ 3
rj 01
C 5
0) v?*
3 a
CT |.
U n
r4 °
01
4-1
•H
E
j
01
o
c
IS
4-1
a
01
o
o
F-C
o
•H
4-1
01
•H
i_(
01
4-1
O
IB
^
18
Xi
O
4-1 4-1
C C
0) £Z OI AS
EO e o
a o o. o
•H XI -H XI
3 Cn 3 en
cr o tr o
W rH CU rH
>1 *>1
M-l M-l
•H -H
TJ 01 TJ
O 18 0 Cn
e EC
4-1 -H
Vl C Vl rH 0)
o o TJ o Q* o
SO) E 18
4-1 3 M 4-1 IB M-l
O IH -H O 01 Vl
0) 4J 3 0) 01
•r-i 01 CT -n 01 4J
01 c oi ai x: c
(j; -H Vl OS 4-> -H
1 4-1 01
•H 3 0) 1 0) rH
Q* Qjx: IB x: QJ
>, e 4J a4-i E
4-1 -H IB
C 01 01 4-1 01
Or- IB 0> Vi 4-i rd
-H • rH 3 18 M-l 01
4-lro 0) Qi M -H 01 0 t)
O • 4-1 E IB M U IB
OlrO IB IB 0> CU 0> 4-1 M-l
01- rH 0] E O 4J 3 Vl
rH 3 H 0> O. 01
OlrH E rH TJ Ol £ 4-> 4-1
oi- -H IB e a, is 3 e
COLO CO O IB 18 Vl 0 -H
^
1 rH1
IB TJ XI
M-l rH rH £
O rH 01 01
IB -H 01
D 4-1 M-l 01 I-I
01 01 IB 01
IB C 0) N
x: -p -H x: rH >-,
O C -P (8 rH
Vl 0) 0 3 IS
3 £ 4-1 C 4-1 C
a, 3 -H o is
1-1 I-I IB
C 4-1 OCX!
O 01 -H O TJ 4-1
a C Vl -H C -H
3 -H ft 4-> IB 3
1
•H C 0)
M-l 1 -H r-l
•HO) X!
O I-I TJ 0) -H
0) 01 01 en 4-i
0,4-1 c x: IB IB >-i
0101-H^J1 4J 4J Ot O)
01 rH • 01 £ N
01 E 4J rH C O >1
3 • O 0) 0 rH
I-I 4-> 0 rH -H X! IB
01 01 rH 4-1 4-1 01 C
M 3 01 • -H -H 18
3 £ 4-1 in TJ O
4-1 C C 4-1 4-1 01
u 01 01 c o -H x:
<8 C £ O O 01 4-1
M-l O 0) -H rH Q)
3-HI-I4-1 4J Q< Vl £
C 4-1 -H CJ OlEOl-P
18(8301 3 18 X! -H
S 0 cr en S0133
He
It *
It 0)
l-l Cn O
0) C IB
N -H Mj
>1 rH Vl
rH Oi 0)
IB £4-1
C 18 C
< CO -H
4J
c
E 0
a, o
•H X!
3 Cn
cr o
H rH
01
^J
3
4-1
X
•H
E
01
(8
en
4-1
o
01
'o?
OH
c
>i O
XI -H
VO 1-'
0) O
N TJ 01
>i O tn
rH X!
IB 4-1 01
c o) a)
< S co
M-l
o
•U
Q,
•rH
01
o tn
0) 0)
l-l Vl
3
C 4J
o x
s's
01
TJ 01 M-l
0) 3 O
Vl rH
3 IB 4-1
01 > C
IS 0)
0) C 0
E 0 14
•H 01
M-I 4-1 a
O <8
Vi in
0) 4-1
Cn C C
(8 0) -H
vi u x:
01 C 4-1
> O -H
< 0 3
01 V<
IS 3
CnM-4
rH
C 3
O 01 C
•H 1 -H
4-1 01
IB 0) OI
Vl Vl TJ
XI 3 -H
•H 4-1 X
rH X O
IB -H -H
u e TJ
[ —
•
ro
•
'"I
rH
rH
•
tn
01
01
3
rH
CO
>
TJ
01
4-1
IB
4-1
01
*
*
M
•1-t
(0
OJ H
(U 0
J_l
M-l C
1 01
c en
01 o
Cn l-i
>|4J
X -H
o c
4-1
c
0) J<
e o
a o
'3 "§>
cr o
W rH
0)
i-i
3
4-1
X
• H
E
01
IB
en
4J
o
0>
01
05
01
c
•H
rH
•H
c
IS
01
0 TJ
l-l O
18 £
V4 4-1
18 01
ft. S
M-l
o
4-1
a,
•H
01
o
OI 0)
Vl Vl
3
C 4-1
ax
• H
D E
c
IB
X!
4J
01
01
01
rH
c
•H
18
4-1 CN
c o
O CO
o
E
•p a
01 a
3
S rH
4t
01
2.
tn
0
l-l
w
4J
11 AS
E O
a, o
•H X)
3 en
cr o
W rH
M 1
•H tn
IB 11
a, o
01 0)
1-4 C
U 01
O IB
>! E
M-l 01
•H 4-1 >,
TJ 0) Iq
O >i IB
S 01 tn
c
-H
TJ
01
4-1
o
0)
l-l rH
•H
TJ •
M-l
tn 01
<: .4
4-1
01
D
4-1
o
4-1
S-l
o
• H
ft
1 .
rrj M-I
O 01
•H r4
•H C
O -H
0)
O.TJ
01 4)
4-1
4-1 18
0) 4-1
01 01
E
tn
4-1 C
tn o
3 -H
S 4J rH
1
a
•H
3
cr
11
^0
TJ
0*
x: 4-i
01 0)
E E
_JJ
£
(D ^
e o
Cu O
•H XI
3 en
cr o
W rH
l-l 1
•H 01
rrj D
a u
01 01
Vl C
M 01
O IB
>i £
M-l D
•H 4J >,
TJ 01 r4
0 >. (B
S 01 tn
c
•H
TJ
0)
4J
0)
rH CN
•H
TJ •
M-l
01 0>
< n
4J
tn
01
4-1
x:
u
IB
01
o
4J
I,
o
• H
ft
CN
1 .
IS M-l
U 01
•H VI
M-l
•H C
CJ -H
O,TJ
01 0)
4-1
4-1 (8
D 4-1
01 01
01
4-1 C
01 O
3 -H
S J-1
1
1 0
0,-H
•H rH
3 a
Er a,
0) IB
00 13
(3
TJ — *
o —
x: 4-1 01
4-1 C rH
D 0) -Q
S E «
Jj
M
O) J^
E O
a, o
•H XI
cr o
W rH
VI
01
TJ
Vl
o
o
0)
4J
CJ
OI
•o
11
a;
c
0 r-
•H •
4-1 CO
o •
01 ro
01 •
rH
01 rH
0) •
co in
11
01
rs
x:
o
3
a,
c
o
a.
1
1 3
IS C
O IS
•H £
M-l
O XI
0)
a, TJ
01 CD
4J
4-1 rrj
U -P Vl
D 01 OI
01 3
4-1 C 4-1
tn o u
3 -H is
S 4J M-l
K
^1
01
TJ
o
o
0)
4J
18
u
c
•H
IB
4J
XI
o
0
4-1
Vl
OI
TJ
0
c
•H
0)
CJ
c
IS
0)
rH
o
4-1
TJ
0)
•H
M-l
•H
o
0>
tn
D
-P
c
•H
4-1
•H
3
TJ
01
rH
o
4-1
c
0
0
D
X!
4-1
01
3
x:
0
•H
x;
3
01
c
o
h
.
0)
.
•H
^
0
OI
x:
o
rH
IB
CJ
•H
4-1
•H
VI
o
(B
O)
4-1
18
0
•H
TJ
c
M
*
•
IB
4-1
rrj
TJ
01
rH
XI
IB
4J
o.
01
0
o
18
^J
rH
4-1
c
u
•H
M-l
•H
e
Cn
•H
tn
4-1
3
0
.c
4-1
•H
3
tn
c
o
•H
-U
18
•H
Vl
18
01
o
c
18
D
rH
o
4-1
1
M-l
o
1
4-1
3
o
rH
rH
o
0)
01
3
o
rH
rH
x:
o
• H
r^
>J
o
o
^
.
D
-H
«fc
^
0
D
s:
CJ
rH
18 >H
O 4-1
•Hrrt
•vH
•H (0
V) 3
0 er
c
O RJ
C 4-1
IB TJ
tn Cn
D C
4-1 -H
IS 4-1
O CJ
-H D
TJ MJ
C M-'
M ItJ
2-18
-------
H
>
H
B
U
<
X
H
«
EH
I
y
O
0)
O
4-1
C
01
g
O
01
x;
u
MH
o
•o
SM
o
o
01
i
>-i j.
. 4-1
3 .5-1
O1
«S
o
e
IS
4-1
a
01
o
o
0
01 -rl
C 4J
M
m
•
CN
01
•rl
SH
0)
4->
o
IS
rH lM
rM 1 IS
• II x:
^*n*
rH 0
is o
C X!
O Cn
•H 0
4J rH
(0
ri rO
01 4->
a is
O 73
01
4J
•rl 01
IS C
Q. O
Ol Pi TH
M g 01
O rl
SH O -H
O 3
£ O1
C 01 01
IS 4-1 SM
01 in
O 01 IS
^
o
01
o
rH
IS
3
01
•rl
>
4->
1 3 C
0 O 01 C iS
4-1 -H Cn
X. -rl 01 rH 10
01 c -H a
Oi 0 S C
S g g IS -H
0) 01 en
M M 4-1 0)
01 0) 01 73 X!
a > >i c
0) 01 10 O
01 C 4J
O QJ en MM
C JS C MM 01
O 3 -H 0 -H
01
01 CP A;
IH 1 1 C 10 73
3SM01 IOC-H 01CMM
4-1 01 CL r^ 0) CP XI rH IB -rl
o! 4-1 is is C -ri 3 —
•H flV-l CJ C QJ ^-* r-< OJ
O6-P M O r-i -HCOT
£ rQ U O C ttj *H i IM -H .— i G
IW 01 C "H OJ U-( -H
O-HfU ows-icn OtDr- 1
Q) O O *H i— j
CJ i-( C Q)O fl (UCijD1
OO-H y.H^-H oec
C IH C WSM CfO-H^.
QJ OQ OJ - C QJ OJUlrH^J
OltDr-H 05QJO-P W OO
,D c n3 ^tr^'Hfd XJ c o w
• • »
<— 1 CN <*>
^"*~
OJ*
Jj*
c e
-H (D
-p
tT1 yj
c >1
-H Cfl
i~i
d< CD
e u
a is
O 73
0 I
•rl 0 C
g 4J O
10 g
e c a —
>, o •-» i
T3 .H J2 > rM
4J 10 -H C
£ IS CJ 4J O
SH tj -H O
O XI rH iS 01
14M -rl Q< r< >H
U rH Q<4-l 0
01 fl IS X 4-1
CL, O ^^ D -H
__^
IN
g 1 •
O O in
SH Oi •
MH 4-1 01 n
14 *" •
0) IS rH
c x: 01 rH
•H o en •
s cm
M O.-H
(!) -H T3 C
4J V< 10 O
01 -U 01 -H
Q 01 IH 4J
C 1
4-1 O M
c a o
01 3 OJ C
•rl 01 O
O 'O -C "H CP
•rl O 4J 3 C C
MH -r( 4-1 -r( 0
MM iH MM SH 'H
3 01 O T3 3 4-1
01 c< C 73 is
a 10 >M
10 a, 3 >, oi
3 i E is a
Vj 1 4J OJ 73 O
oi E >-> 4-1
4J IH (0 in V4 rH
MH (0 4-1 >i 0) 10
< 3 01 01 a E
MM CP
o e 01
•rl rM
C rM 1 IS
O O 4-) M O
•HO C 01 01
4-1 O 01 tt
10 E rH
i-l 73 4-1 in rH
Q) C 01 3
Q, IS 3 v |UM
O -o
O~* cfl ""O OJ *H
MCg IS X) O
0) -H 01
0,4-1 4-1 O 4-1 4-1
O 10 01 rM 01 C
Ij 0) >i Ol 3 0)
a< x: 01 NEW
t t
f rH
a*
(S*
ac
01 0
•H
73 4J
C (S
IS IH
XI
O -H
IH rH
Ol iS
N U
^
rH O
is o
e xi
O CP
•H O
4J rM
IS
IH i
t£ 01
rH
iS
3
01
•H
>
01
O M
c o)
MM O 4J
O MH
73 ifl
a c ai
a is IH
1 01
4J £ S
M 01 4-1
10 4-1
4-1 01 >i
01 >i 10
in 73
c
O 0) rM
CUXI 01
a -u a
1 £*1
O rH
C Q-t
3 Ou
MH 3
rH 01 .C
IS 1) 43
E >; g -H
C -H 3
MM -H 4J D
O 4J E
OJ -4 C -H
OJ 4-1 0) 0) 4J
O IS 73 g
C 01 3 14 01 J<
01 c cf o oi o
01 O 01 O V4 O
X) -H 73 01 CT>rH
< -P < OS IS O
• •
rH CN rn
4-1
rM
IB
x:*
O rl
01
CP 73
O lM
rM O
IS U
C 01
< SH
^
rH O
IS O
C X)
o CP
•H O
4-1 rH
IB
SM rs
o j->
Qt is
O 'O
Ol
rH 1
01 X) 1 13
IS -H 0) CP C
01 SM 3 iS
IS 01 01 g
4-1 O IH
IS Q<-H in >i
73 10 (S XI M
01 a oi
Ol O -H
^J
01
0)
3
SM
01
a
01
o
c
0
73
C
IB
73
IS
Ol
Q
„
*r
0)
a
IS
4-1 MM
CP-H
o c ~-
•H -H
4J 73 g*
01 SM 01 —
C O 4-1 73
Cn CJ 01 Ol
IS i 01
S rl 01 3
.
73
0>
rH
rM
IS
4-1
in
c
•H
CP
C
0)
XI
g
Ol
4J
01
01
01
x;
4J
O
4-1
0)
rM
XI
IS
o
•rl
rH
a*
IB
MM
M
4-
2-19
-------
EH
u
X
H
K
H
a
CL,
en
C
•H
4-1
01
/a
EH
01
o
c
a
V4
o
m
s-i
01
04
Tjl
•
CN
•
rH
i-H
•
.*
O
CD
CJ
MH
O
T3
i-i
O
O
0)
OS
1
01
i! -w
3 3
CT**
5 4J
05 0
UJ Z
MH
'H ui
-, .p
Is
c
jj fc
«
Method of
Measurement
X
0 ii
«4J
Frequency of Ch
or Measuremen
01
4-1
•rl
E
•H
1-5
01
O
c
ifl
4-1
a
01
o
o
<
o
•rt
4-1
01
•H
M
01
•P
O
10
^
10
X!
in || u
^
* 8
10 o
C X)
0 Cn
•H O
4-1 rH
(0
M Ifl
01 4-1
a io
o -a
c
0
•rl
4-1
Ifl
M
X!
•H
rH
Ifl
D
, 01
4J ^4
O 3
01 4-1
•r-i X
0) -H
« E
1 -IX!
Take three sam
pies of each
mixture and,
using Method 6
determine
average concen
tration of eac
mixture
4J
01
01
4-1
At performance
and once every
6 months
e £
o o
!-4 14
M-l M-l
01 01
Cn en
C (8
(0 IH
X! 01
o >
10
4J
C T3 A!
OI 11 O
O 4-1 01
to ifl X!
« 74 CJ
04 3
O 4J
in rH 01
10 10
V |CJ rH
01
10
en
SH-
o*
•H*
•P 01
10 01
V4 IH
XI 3
•H 4-1
rH X
10 -H
CJ E
.*
rH O
io o
C XI
O CP
•H 0
4-1 rH
10
U 10
OI 4J
a io
O T3
OI
4J
10
>-l
X!
•H E
rH 01
(0 -U
O 01
01 >i
K 01
Manufacturer' s
instruction.
See section
5.11.1.5
4-1 1
0] M
01 X 01
4J OI Qi
At performance
and once per we
during normal o
ation
01
>-i
3 0) 1 1
4J en I* c
X! X C 0) O
o -H (o a o
(0 E fc
01 0 M
01 01 rH (1)
MH 10 x: -a
o en4J v ic
•H
4J C MH 01 i-H
C O -rl en >,
0) -H CO
O 4J ^4 -rl
H ifl O t3 tH
Oi V4 ifl O
a xi oi o)
•H 3 M 4J
in rH rH C
«j 18 M-I a)
v |CJ > O 0
c
o
•H
4-1
10*
M*
XI >H
•H 0
rH IH
Ifl ^4
U OI
C
o
-H
4J
10
It
4J
C
01
o
X
rH 0
ifl O
C XI
o en
•H O
4J rH
Ifl
M 10
0) 4J
a 10
O T3
MH
O 1
IH - 0)
01 o c 4J
01 O 0 ifl
3 -rt V4
io oi -u xi
o .* o -H
10 Ifl rH
><4J 10
MH a) o
•H - > 01
4J 4J -H M
C MH 4J
0) -H O T3
•a u o) c
H "0 U Ifl
c
O CN
•H •
4-1 in
o •
O) n
01 •
rH
0) i-l
0) •
CO in
1
18
Make 15 2-hour
interval deter-
minations and 7
2 4 -hour determi
tions
i
1 lH
M 0)
o> a
04
ON
CM c c
§ vr5
V 1 Q, Q,
0] ^01
^-» M
>H MH • 3 MH
SO O O
O X!
X! 4-1 4-1
C •* G
CN 01 (NO)
— O ~ 0
rH CN
*
4-1
MH
•H
>H
tJ
0
M
0)
N
j<;
rH O
10 O
C XI
o en
•H 0
-I-l rH
10
V4 Ifl
a 4J
Ifl
O T3
<4H
O 1
V4 - 0)
oi o c -u
0) O O 10
3 -rt (H
Ifl 0) 4-> XI
O J4 O -rl
Ifl Ifl i-H
>14J Ifl
MH DO
•H » > (U
4J 4J -H W
C M-l 4-1
01 -H o -a
13 >H 01 C
H T3 >H 10
171
Calculate from
span drift and
zero drift dat<
See section
5.11.3.5.2
1
10
c
Make 15 2-hour
interval deter-
minations and 7
24-hour determi
tions
i
1 )H
lH OI
01 a
O4
m
(N e C
5 v|io
via &
M — m
-^ M
t4 MH 3 MH
3 O O O
0 X!
x; 4J 4-1
C f C
(N 01 CN 01
• — o -^ o
rH CN
c
O
• H
4J
rfl*
V4*
XI 4-1
•rl 4H
rH -H
Ifl Vj
u -a
J^
rH O
(fl O
C XI
o en
•rl O
4J rH
ifl
V4 irj
OI 4J
a IB
O T3
)H 4J
o c -a
» 0) 01
>H 01 C 14
01 O O -H
N io a, 3
>irH E CT
rH Qj O 01
ifl 0) O !H
C !-4
Ifl C 01
» 18 IB
-U ^4 01
O -rt r-H 01
0) 10 O 4-1
•r-i a tl
0) OI >H (0
a; SH o a
i t
Compare to Ref-
erence Method
results as de-
scribed in sec-
>,
0 1
Test for accura
for one week
during each per
formance test
01
o
c c
ifl (U
01 M
E 0)
M-t 10
MH (U 4-1
O M (0
•0
4J 01
C X! 4J
E
«
«
>,
o
(0
>H
3
o
o
<
Ji!
rH O
(0 O
CXI
O Cn
•H O
4-> rH
10
lH Ifl
0) 4-1
Oi IB
O T3
tH
14 O 1
o o
^ (n c
M 01 4-1 3
0) O C MH
N IB OlrH
>irH C 10
rH O, 0 E
IB (u a
C ti E 01
Ifl 0^4
- O ifl Cn
•U M C
0-H CX! -H
01 fl IB O C
•r-i Q< 01 -H 0
01 0) rH x! -rl
a i-i o 3 4-1
in
(T .
tion 5.11.1.5.
As described ii
section 5.11.1
D1
C 1
•H -M
Three times dur
performance tes
ing
E
•H
X
fl
E
tn
0)
4J
3
c
•H
e
m
rH
Q)
>r4
4J
o
01
c
o
a
i
01
o
-P
(U
r~2
4-3
(Tl
Hi
0
•H
rH
o
x-u
Qj
fl
c
QJ
?5
-<-
2-20
-------
X
O
a)
43
u
<4H
o
73
)H
o
O
,^
c ; o 2
H C "*
4J ! S S
Ml 3 «"
O 01 |.
a, CD 2
CD SH °
OS
"§
IB
C
O
•H
4->
IB
13
• H
rH
IB
>
^
c
o
-H
4->
o
3
73
CD
OS
(8
4J
IB
Q
un
CN
.
rH
rH
BH
en
4->
-H
E
•H
J
CD
O
C
(B
4J
04
CD
O
rj
<
O
•rH
4-1
01
•H
IH
01
4->
u
IB
M
M >1
0 H
CD
CD 73 CD 4J
01 C > IB
3 IB -H -H
rB 4-1 73
U C O CD
O *H 0) -H
4J rH X. 4-1
01 IB 18 O
Q E 4J IB
Jrf
o
0)
43
O
rH
fl
3
cn
• H
>
>i
(8
73
iH
CD
a
IH C
01 (B 3
JS tH
UH O rH
0 Qj (B
•H E
CD iH
O 4J E
C 01 CD
01 4-1
en CD en
J3 £ >i
< 4J en
*
4-1 C
iH 0
(B -H
£ 4-1
CJ IB
a 3
•H rH
M IB
4-> >
cn CD
T3
iH
o
o
CD
iH
(8
4-1
IB
Q
•O
O
•H
01 01 iH
rH 01 C CD
IB IB -H Q<
> IH
IH T3 3 01
CD 01 T3 C
4J -H .H
C IH 73 jj
• H -H -H O
4-1 rH 4->
73 C IB -H
•H CD > C
0 73 C 0
> -H -H g
4J
iH
X IB
O J3
D O
js Q-
O -H
iH Ol
rH 4J C
18 en -H
3 73
en 73 IB
•H C CD
> (8 in
MH
O
0) (H
a> oi
c a
IB fB
js a
o
4-1
C iH
O C8
CUJ3
D U
rH
rH
IB IH 73
4= 3 O
en o en SH -H
M-l 4J CD SH
01 C > 0)
CD E -H o a
Cn O 0
18 IH CU73 0)
SH MH
CD (B O 3
> T3 4J rd C
IB CD id a-H
4J 73 01 g
SH 3
3 a CD xo
O E SH rH vc
,C O O rH
0 E IB JS
01 3 O
C 01 SH O1 IB
O A O 01 Oi
C
0
• H 01
4J4J
0 C
3 0)
73 S
CD 01
SH in
•r-j
IB 3
4-> O1
fl 01
Q iH
SH
3
0
42
0)
43
4->
C
O
cn
c
• H
O
C
CD
g
O
0
01
•H
en
>i
rH
IB
C 01
(B 4-1
SH
(8 O
4-> a
IB D
Q IH
0)
C
73 -H
SH E
Oi O SH
COO)
•H 0) 4J
01 SH 0)
en 73
•H g
E O SH
IH O
C IH
•H en
IB IB CD
4-1 4J rH
J2 IB -r|
O 73 MH
X
o
CD
43
o
rH
18
3
01
•H
>
IB E
4J O
18 M
73 «
4->
En c M
C 0 IB
•H -H 43
3 4J O
o o a
rH 3 -H
rH 73 IH
0 0) 4->
fa SH 01
>1 1
•H -H >,
SH rH O
0) IB IB
a o IH
O 3
IH rH O
Q*i-i O
fd rO
CD
43 43 73
4-> C
4-1 -H IB
en 3
3 C
E 73 O
CD -H
IB O14J
4-> Cn IB
IB O SH
Q rH 43
cn
C
•H
4-1
SH
O
a.
CD
SH
IB
4-1
IB
Q
1
en
01 C
rt 0
g -H
4->
SH IB
o e
4H SH
O
en MH
C C
0 -H
en
IB tji
01 C
SH -n
43
4-1
•H
3
•a
CD
73
3
rH
O 0)
IB JS
73 4J
en
-H
en
>i
rH
IB
C 01
IB 4-1
SH
18 O
4J a,
IB 01
Q M
rH
O
g M
O M 4J
iH O C
UH 4J 0
IB O
IB iH
4J CD O 18
IB G.4-1 4-1
73 O IB
73 73
C E 73
•H CD IB 4J
IB 4-1 iH
4-1 01 73 IB
42 >i C 43
O en id O
x
0
CD
43
O
rH
IB
3
en
•H
>
>,
rH
• H
IB
Q
X
MH O
O CD
43
Cn O
C
•H C
73 IB
in a. en
O 01 4-1
o c
CD 73 O)
SH C E
IB 4->
>i en
rH 0 3
•H SH T-i
IB 01 73
Q N IB
en
4->
SH
iB
43
O
rH
O
IH
4J
C
0
t_)
2-21
-------
EH
H
>
H
EH
U
4H C
O 0)
e
13 01
O SH
r; 3
JJ in
dJ n3
2 0)
S
^
0 in
0 f"
-C
u S
0 <»
3
^S
c
QJ S
3 s
C^
QJ ~
SH
fo
in
jj
•rH
g
•rH
,J
01
O
c
10
JJ
a
0)
o
o
"^
0
•H
JJ
in
•rl
SH
U
JJ
0
10
SH
to
c~|
1 U
y
rH 0
to O
C 43
o cn
•H O
JJ rH
10
in to
QJ JJ
a in
o -a
i
rH
(0
y
0)
3 >
01 M
C 3
0
ffl
c
01 O
SH -H
rO JJ
CX <0
OJ SH
*H 43
CX -H
c
O rH
•H
JJ in
o •
Oi o">
in •
rH
QJ rH
QJ .
CO in
in
4^
01
QJ
3
IN
^
SH
01
*>
0)
01
o
c
o
i
IQ
HH QJ
O S-l
C Q)
O IH
• H
JJ g
nj o
•H IH
0) 0)
13 -U >
C IH
S 0) 3
D O O
•H OJ C
x a o
S m in
i
•H
i — l
tO
0*
He
O C
•H O
E'H
(0 J-'
C flJ
*>1 r-4
Q43
^
rH O
fT] O
C rO
O Cn
•H O
JJ rH
10
SH iO
QJ JJ
tJj (0
O T3
0
• H
g
to
C C
>i O
JJ
g 10
SH SH
0 43
SH rH
01 to
ft. O
C
O rH
• rH *
jj m
o •
in •
rH
01 rH
QJ .
CO in
^
tO
T3
SH
OJ
a.
01
o
c
o
jj
c
01 QJ
Cn £ C
C JJ tO QJ
to in 43 >
43 3 JJ -rl
On JJ
T3 Q) 3
JJ 10 SH O
C 00)
0) JJ g W
o in c
M tO O O
0) H C O
a
£43 >H
in o JJ 3
v I'tn 3 in
C •*
to •<
a c
in O
•H
13 U
C tO
tO IH
Q
0 ^H
Q) 10
ISJ O
y
rH 0
« O
C 43
0 Cn
•H 0
JJ rH
(0
SH (0
Q) JJ
a. 10
O 13
E
SH 0>
O JJ 1
>i tn IH
SH in >, 0) JJ
ia c in cu in
in O 0)
m -ri c jj jj
01 JJ O fl
O 10 OJ OJ
QJ o in a o
C -H SH 0) C
KH -rl SH 10
01 -rl 10 E
•^ T3 Qj'O in
10 O 0) C 0
S £ SH <0 "M
C • rH
•rl rH C 00
• o
'O ^H 'O CO
01 rH OJ O JJ
43 • 4* 43 S
•H in to jj QJ
M E OJ g
0 C SO)
in O 4J SH
QJ -H Qi OJ 3
T3 JJ Q) OJ U)
O O SH «3
in QJ x 43 0)
< in QJ -u g
in IH
43 O 0) >i
JJ T-l C O >4H S rH
C fl -H <0 • -H Q) rH >,
O E "H SH JJ tO SH
g 0)SHQJ43inOQ)
tO T3 0) N -U >i-H >
i^ tOJJ£MC*in5oi
SH g C rH O tO
i>iQ) •HfO£OJGrO
SH > tn C SH >i 0)
01 01-rl Cn«3 SH-rlTJ-P
> C C Q) JJ 10 tn
QJ flj QJ 'i*4 QJ pi C -jJ ^4 ^
rC O^ 1^ r^ QJ O rQ QJ
QJ i5 C QJ 4-J QJ C *H QJ
O 10 E O QJ rH 3
OOOtnOOJJ-HOoj
1
r-J
Q)
MH
0)
0)
43 4H QJ
JJ O 3
rH
43 JJ to
JJ S3 >
•H QJ
3 c on
Oi In (0
in -H QJ o)
oi in a E
C"
C 0) 0 01
tO g CN O
43 tO C
o in v |o)
H*
H*
^
o
10
r4
3
o
0
fy
rH 0
tO O
C 43
O Cn
•H 0
JJ r-H
tO
SH <0
QJ JJ
Cu to
O ^3
0)
0
10
i-H
a
oj :>i
in M
(0
SH cn
O tn
01
SH o
•H Q)
(0 C
a
0) m
OS ra
m
_
>H W
0) C
SH O
3 -H
JJ JJ
O 0
tO p
3 JJ
C W
tO C
1 IH
in M 0)
(0 O ^
Cb SH
-4 X O
0 O 0
01
iO 43
T3 rC
n:3 -i-1
M (]3 -H
OJ -P S
QJ (|J Q)
u o) in
0) -H O 3
O T3 C
C C 0) C
O -H —1 -H
^
rH
a
O< rl
3 01
in tj
X O
ceo
HH O -H Q)
0 -H M
JJ 0)
Q) O JJ JJ
0 C 10 0
e :.! 3 o)
01 14-1 D1 SH QJ
in r-i 0) SH E
43 (0 t) O -H
«s e < u jj
rH IN m
cn
c
-,H*
"O g
SH 0)
O JJ
O Ul
0) >i
oi tn
X
rH 0
10 o
C 43
0 Cn
•rl O
JJ rH
18
SH tfl
Q) JJ
a< to
o -o
01
0
rH
a tn
QJ (fl
IH
tn
V4 •!-> >1
O S3 SH
01 (0
M c tn
•H o tn
to a oi
a- e u
0) O 01
OS O C
in
~
SH in
QJ C
-l 0
3 -H
JJ JJ
O O
tO 3
MH SH
3 JJ
c in
iO C
S -n
y
0)
3
IH
(1)
Q.
0)
O
C
o
rH C
(0 O
O -H
•rl JJ |
e (0 o •
tfl O C O
43 -H 3 JJ
O IH MH QJ
OJ 43 rH
£ 3 (C *
rH E in
c A;
0 SH i4H to
tn QJ O OJ
IH -U Ol rH
(0 SH O OJ
(0 C
CO -rl
2-22
-------
!5H
EH
H
>
H
EH
U
<:
X
H
OJ
O
.c
U
M-4
0
fl
0
0
01
K
1
41
'H a
Jra
iS •*•*
M-l P3
'H n
§S
•Ui
o
4J
M-l C
O 01
g
fl 0)
0 M
S in
S 41
2
o m
r- +*
Ss
M-l S
EH 0-
f31
S
§
j
PM
•v, 3
0 W
r* fl
oi S
3 *
01 L,
01 Q
fa
in
-H
g
•H
J
1 01
en
01
M
3
T3
01
O
O
J-l
Cu
en
c
O
c
-p
a
41
o
0
(T^
-H i
•H ! 0
'O
3
•a
3
01 en i
in 4J in in
1 fl -H C -H
C 1 0) fl M-4 O -P
0) >-l O C 01 O iH
O ft M O rH ft £*
01 3 3 en O O>
•HOIJ-I • >1 01 C -P rH
fl O C O 01 -H
O M-l O C M i
4J 4J -HO) rH > Vl
l^ J_|-P3rHflOO
a 01 0 fl cT-r-i 3 >-i -P
0) to u cij 4-) en ft cj
01 'H -H -H tl Cj -H £ fO
rH
fl
3
in
•H
>
1
0 0) G 0
l-l C 01 S-i C
JJ Cl rH C
\-t at C o) ft-H
0) 0) O J3 E --I
ft4J -POO
-P C O -P
01 0) co O 1 -rH
C C 01 O O
•o
c
rrj rH
rH
C fl
0
-P O
fl
0) n to
ft 01 4J
O C C
•H 0)
r-l rH C
0) C O
QJ 03 pi
O -l rH O
PJ O O
e
01 T) 0)
X) C -P
O fl en
&-( >1
ft - en c
0) O
Cn O V-i -H
C fl 0) -M
•H M-l N O
•H U >i 0)
ft (U rH ft
g 4J fl en
fl c c c
W -H fl -H
o o
o o
ja -0
o> «^
o o
r-H '""'
£ fl
3 3
01 CO]
1 01 j3 en 1 O 1 01
O.Q rH-PflO] T)"1'?)
C4JC4j'Ij014J4JrH O01J3E
raflflC4j^^,(8cu -HOI
S-lST! 3-P'H Cl) S"0'*
u-ienc-HO)oiOifl -HOifflo
a-HT301M-P3'O>i t!atl?,
C-Hg3fl (UCPOIr-l OOICOI
OOl-jflO)4->O £O 0) (H 0) M
flSS+ioio'-afljjSo IH fl § "
S3^55S.S525 SS^S
rH
rH
•
in
(N
c s •
0 O ^J*
i '.PH
O O rH
0) 01 .
r/i CO in
in
M O 0) -P 'rH M-4 0)
O-p oiOOftCn X Oen
M-i(Utnoi EEC -H3£I3
2-p-Pg O-H cnoi-PO
O M 01 01ft>
S-MgHflSrHO OCM-lgO
0) 01CO3C COi + |T3OCCG
O-UwOfd^tC OJ • 'M QJ-HnJOJ
(UU1 tTiCtU-UO.^ C -H QJ fQ QJ QJ i— I
CPU tn>iCT3-H4JO rH c td £ M S m 3
CT3COC OO TJWCJ-.13O jJtDjD-HM-jO
--H -n 3 -H n3 TJ C >i £ > -H M-t -H £ -H 03 -H &
•~H T3 O f~t M *T3 O O tfl rH U -M O S -M t-H ^ T3 W
ft«3ft30)JSft rHOCO fl
g Cg oin-pcncnflMflO) inoi^rjoioiin
flOl-Hfl Ofl010)3^0)0)tn -PXIO (-(tT'J-J
eoj3-Pen <],Q2r-iOinftg *~ ' ""^ 13 3 ^ , — |
3 4J 4J CU 4J ^| 3
in in c c x cu en
4) 3 01 -H -H > 0)
rH IN rXgrjggrrjl^i
{
XI
-P I
0) X T3
E 1 ^ r-l
0) en o rH 0) ft
M-l fl fl M-l E
41 fl i-l C S a
PS O -P fl 05 in
1
rH
fd
O OJ
^1
M-l 3
X
-P -H
01
o in
i-i fl
0) en
Q,,
C
O O
rH -r-f
4J fl
'O fl —
QJ ^-4 r—
2-23
-------
H
>
H
EH
U
X
H
^
4-1
O
O
M
iH
3
01
U
O
PU
tn
G
•H
.p
•rH
'O
rJ3
^
X
o
(U
u
MH
O
13
0
O
K
1
1)
rH 4-1
3 CU
era
0)
PJ 4-1
o
MH Z
iH
W
C 4-1
O C
H CU
•p e
u
rt,
4-1
UH G
O 0)
E
13 0)
O SH
4-i
.£ G
U 0)
g
UH 0)
O SH
>i W
O (0
C (D
CD S
tr SH
CD o
fa
ra
4J
-H
g
•H
CD
O
G
rO
4->
O,
CD
O
O
O
• H
P
(fl
•rH
^
r-- | 0)
rs!
1 O
(0
rH M
rH fl3
,
in
i ^C
o
o
Cl
^l
tn
Cl
i-l
4-1
•r-l
1
1
>i0'0
C M (U
c a4J
d nj
= rH G
f-H -H
a, 4J E -o
V- G-H tt)
M- a i-H N
0) 'iH.
a-p e
V. -H U) -rt
(C T)-H G
13 -H
>- IB E g
C CU
C MH .H >H
r- Q & O
1 tn
(Tl G
r-l -H fd
3 C 4-1
0 C (0
r-l -rH T3 • — '
nl tn m
U 0) S C •
X) tl O M1
r-l r-l -rH •
«l U) 4-1 iH
3 C ,G O iH
C! O 4-1 01 •
ni -H -H a) in
s: -P s —
CU
C U) <1)
1 O fi O
V. 4-1 G
C 13 C rt
^ C O -H
!>-' (3 E rH
O DJ
C-i 4-i «3 E tn
tn O G
(-1 CD M O -H
Ci -P CD 1 S4
c.. a G o
CD O 4-1
ai o x c -ri
o G a> G
c: d cu MH o
o e s o e
i
41 fH
01 rH
-J O JH
j: u_, ^ o o MH
O O CD O
Ul MH SH v
4-' 4J 0] 4-1
r-l a (0 4J g G
3 CD P SH CD 0)
u) o fO rfl 4-i o
CD S-i T3 42 ra SH
S-l •• CD O >, CD
G &4J a tn &
r-( .rH -H -H
n! ,C n T3 SH tn rH
c: -P a -P c
•r-l -H + 1(0 «] -rH +
•H
S-l CU •
QX1 rH CN
tn
G
•H
0!
(J)
a)
o
o
d,
nl
4.1
r.l
Cl
1
tn i
n) 13
e SH
O
SH O
O cu
rt) cu £
4J a, cu
•a 4-i u)
4-i o tn
•H -rH
T3 -P tn
3 0) G
«) G -H
2-24
-------
5.11.3 Operational Procedures
5.11.3.1 General Considerations—In the subsections
immediately following, specific procedures for measuring and
reporting sulfur dioxide emissions from stationary sources us-
ing continuous monitors are described. The sequence of opera-
tions to be performed is given in figure 5.11.2. Two sets of
columns are used. The left-hand set, with items numbered 1 through
24, lists operating procedures in sequential order for a given
monitcuring system. Calibration, maintenance, and procurement
procedures which are performed periodically are listed A
through P in the right-hand set of columns. Quality check-
points in the measurement process for which appropriate quality
control limits are assigned are represented by blocks enclosed
by heavy lines. Other checkpoints involve subjective judgments
by the operator with proper guidance as indicated in the pro-
cedures .
Directions throughout this manual are written in terms of
two modes: first, the performance testing mode, and secondly,
the continuous operation mode. Audit or checking levels of
three checks per 1-week operational test period and one check
every six months during non-compliance continuous operation are
assumed*
Examples of various pages which might be included in opera-
tional records are given throughout this manual. The operational
records should include:
1. Operational Data Logbook
2. Sampling Recorder Stripchart
3. Hourly Averages Record Sheet
4. Calibration Logbook
5. Equipment Logbook
6. Maintenance Logbook
7. Performance Testing Logbook
8. Audit Logbook
2-25
-------
tn
c
o
-H
4->
(0
rH
CU
&,
O
4H
o
a)
u
c
0)
0)
U)
^1 £"1
reliminar
ite Surve
ft CO
G tn
O G
terminati
Monitor!
System
QJ 4H
Q O
>»_
"
M
System
rocuremen
ft
G
System
stallatic
c
*•
u_i C
lalysis o:
llibratio
Gases
< o
4-)
rH
(0
41
o
8
rH
4H
rH
rd
O
•H
4J
CO
rH
0)
a.
o
in
0)
rH
O H
(0
OJ U
QrH
>i tn
4J >i
i O
4-> O
co tn
G
""O 'H
G H
rd Qj
g
- (0
tn tn
M
QJ tn
• l r*
QJ -H
tn
•H t3
X CD
QJ g
QJ
C
•H QJ
4-1 rH
QJ 42 4H
T3 -H O
4-1
""O fO w
G .
rO g
O
41 U tn 4-i
u
13
g 4-> rd fO QJ
QJ tn
QJ
QJ -P
cn tn 4-J o
Q) >vH 43
OH tn £ to
QJ
c
•H
C
0)
Cn g
C D.
•H -H
QJ O
T3 C
•H
QJ
C
QJ tO
W
rd
rH
rO
Pu
g m N
Q) rH >
O 4-> QJ H
O tn
cn
4J
G
•H
(-1
O
o
QJ
I
tn
o o H
tO -H •
"4-1 4-) rH
3 O rH
QJ G 3 •
us
g
in
tn G
G rH G 0
OJ 0) -H -M
O-i 4-1
H tn o
g-
fO QJ rH
4-) 4-> CD
tn tn M
H tn 4-)
(U
tn
0)
N
rd tn
G QJ
to tn
O DI
V£5 4-1
G
T3 tn O
O QJ -H
X! rH -P
4-1 3 ro
QJ ^ rH
S QJ 43
O -H
QJ O H
tn VH ro
ID a O
(N
2-26
-------
Make a minimum
CN
O
C/3
CU
c
•H
G
UH
O
concentration
measurements
i
00
vo -P
rH en
cu
Cn-P
£
-H H
M 3
3 0
T3 .C
period
0)
C
O
-H
4->
(0
M
Dynamic
Calibration
•"
Operation of
System for
168 Hours
>^_
'reparation for
eference Method
Measurements
m oL
>^
en
4-1
C
CU
T3 £;
O CU
-C J-l
-P 3
CU W
S W
QJ
SJ
i
Operation of
System for
Additional
168 Hours
o
&
o
o
•H
-p
RJ
M
Q)
o
o
oo
in
cu
o
-H
4-> O
(d cu
moo
CU 4-1 CO
Cu rd CX
I
S-l
o
I
u
I 4-1
g 3 en
CO fi C
4-1 (d -H
en g
>t en
to ^ -
CO h
CU Cl, CO
4-1 M
n3 >i n3 en
M rH 4-1 fi
CU rH U O
fl) fd fO -H
O g 4H 4->
C cu en
rd O 4->
c c
en Cn co CO
4-1 G SH g
G -H CU CO
CO rH U-r 5-|
rn P| fll rj
" K en
rfl
5-1 CU
O g
CU rH <4H
M rH
fd fd G
PU4-I -H
CO
5-1
!H C
Q
CU 03 (0 -P
CO
CU
o
>i
rH
en
O
en
4-> O
O CU G
4-> 3 -H
•H rH 4-1
C MH G
o m o
a cu u
in
-•H
2-27
-------
S-l
CD
S-l -P
O
w
O
CN-H
O -P
ft cu w (d
g -P rH
O rd CD -P
U H Cn C
CD
I C
C-H
•H C
O
O J^
CO
C -P
(tj *H
s &
3
o
H
(d
u
T3 cn
o -a
re; o
-P -H
(U >-t
S a)
CD rd
I 1 l_j J-<
O CU iD
C M -P 13
O O G -H
U m -H U
1
S-l
o
u
CD
C
•H
g
S-l
CD
-P
CD
P
M
O
-P
0
(0
M-H
£j
o
•H
-p
u
CD
S-i
1
S-l
3
cn
'^d
CD
cn
•H
!>
CD
S-i
dures
CD
U
O
S-i
ft
^3*
T3
O
r;
-p
Q)
g
cn
C
O
•H
-p
(0
5-1
(P
a
o
(U
u
c
cu
a)
cn
G
O
•H
-P
Hm
CMO
O-H
cn -P
rd
CD S-i
Cn-P
3 O rd G
u
rH
(d
u
>
S-i 0)
CD O
> G
0
U-l ^
0 V
S cn
CD
^ -H a) ^
>i U &1
1 i G ^
rd ^ rn
C ^
CD
p5
>^
^
CD
4-1
rd .
4-1 ft"
O O
S-i
ft;
f
C >i rd
O S-i -P
•H P rd
-P P
rd S-i
G
• H
o m
n
g -P
SH
0) cn
CD IS -H
4-1
cn
CD m rd
p o m
1 (D
CD cn 1 Of
i
cn
•H
H in
G
O
H S-4
P O
•J 4-1
y
H fd
H P^I
\
_)
•^^^fc^^
^
CU O CD G G -H G
T3 -POr-lO (D-POrrj
S-iS-i fdS-iftO S-i G O 4->
CD fd
•P X!
0 U
G ft
•H -H
G rH
1 -P
in cn
rH
g
CD O
-P M
rd m
rH
3 'd
£3 rd
fd CD
E-H S-l
o
o
CD
S-i
•H ft g CD O fd
s-i rrj CM IH o CD ^
ftT3 cn O CD
O O W « T
43
5 (D
S-l rC CD G -P S-l
CD ft-PNCD S-irdcnrrj
ft
fd
-p
u
•H
-P g
CD 0)
G -P
Cn cn
rd >i
g cn
ft CD >i-P O
rd ^1 i — t td cn M-i n
0 ft
3 g F
rO rH G -P ' ' O
G O rrj O -H -H T
•H G rH 4-1 cn
13 CD *• rd fd fd T
3 rd r
3 4-1
>
rd O
j
O S-i cn U M X! O
rH CD CD -P rC
H cn H->
J 3 C
-4
OCD3GCD x: CD O CD
[LI rt rrj rrj rj EH S
H 3 4
J
-P
S-l
(C
X!
U
fd
G
O
-H
-P
rd
S-t
CD
ft
O
-P
G
O
U
in
CD
S-l
•H
Pn
CM
2-28
-------
U
&
O
•H
4-1
U!
G
O
-H
-P
(d
^
CD
ft
o
4-1
O
cu
u
G
(U
3
tr1
QJ
C/}
4^
O W
mputation
Difference
o
o
[•
^.
4-t
0 QJ
Iculation
Mean
Dif ferenc
(d
u
>__
4-1
0 m
mputation
Conf idenc
Interval
o
u
i
>^
m
04J
Iculation
and Repor
Accuracy
td
o
(d
G
O
•H
4->
(d
M
CU
a,
o
-P
a
o
u
t!
O
,£
-P
0)
u
(0
Q)
S o
CD g
U O
C M O CU
D IH m 4J
c
cu c
4-) O
CU -H
(U -H
-P S-l
O -P W
G 3
(1) O
u
S-l
-p
0)
X! G G
3 O -H m
Cfl U -P O
W
4-1 0)
O O
G CU
CD CU
g 4-1
4-1
CD -H
,G rO -P
I 1 l_i
0 O
Q) f^ £1|
-P -H CD
£3 CH M
g CD T>
O ,G C
O -P (d
cu
x: "">
4J O^
4-) 0) (U
S-l rG O
O -P G
CU 4-1 T)
M O -H
4-t
T3 -P G
G -H O
(d g o
(U
-P
g
O
O
>H -p rd
G >
S-i CU S-i
CU O CD
di S-I -P
O-i CU G
3 OrH
I
5-1
-P Q)
S-l £4 G
O <3
CX (C 0)
0)
rtJ C1J
0)
G £**i -P
(d u
rfl 4-1
CD
-P
O
U CD
U Cr
-------
oo
O
(N
-P
n
td
^
u
8
rH
UH
W
C
O
•H
-P
id
SH
cu
ft
O
u-i
O
CU
O
c
CD
CU
Calculation
and Reporting of
Calibration
Error
^_
^™
MH
0
& 4J
o tn£|
4J -H T!
id -P 5
r-H S-l
d o 0
o a. r,
<-* ® m
5"*
T3
fi
(0
~^
^
m
o
d c
O tn O
•rH fi -.H
4-) -H -P 4-J
Id 4-> fd 4H
rH J-l S-l -H
3 0 X! SH
U ft-H Q
r-H Q) rH
id tf td
o u
T3
fd
"•
Calculation
and Reporting of
Response Time
Id
fi
o
-H
-P
(fl
H
QJ
&
O
-P
c
o
o
CN
cu id
m 3
0)
c
o
-H
CU Ti 4-> (d
O C Id rH
c id SH 3
CU X! O
S-l rH -H rH
H (d
id o
i] i j) rj
•HO tl
t3 G
C id
cu c -H
4-> cu 4-i en
3 CU rH (U
ft 15 3 3
g -P W rH
o cu cu (d
U X! S-i >
o
S-l
cu
cu o cu
S-l X! ft
fi
(d
cu
4-)
(d
rd
en ft
(d en
m o
-H
s-i d)
3 3 id
O O O 4-)
rH XJ S-l C
id i cu cu
CJ £N N O
4-> W
SH id
O S-j C
ft 3 4-> td
CU O 4-1 ft
XI
I
tn
cu
(d
rH S-l S-l
3 3 XJ
O O -H
rH XJ rH
U rs
o
o cu
•H tn
cu
o
S-l
cu
O ft
cu -
rH en
cu id 4-i
wow
C W CU
O ft-P
ft 3
w cu
CU rH rH
S-l rH td
id o
w
SH
cu
4-)
m
G
O £
rH 4-1 O
3 T!
o cu
rH g T3
(d -H G
cj -P id
cu
e
-H
S-l
* cu
w £
cu o
3 rH
rH W
id
> cu
C -P
td
CU -P
S M
o
CU ft
x: cu
•P SH
in
0)
SH
3
tn
2-30
-------
>1
G 13
4-1
CD M en
> O 3
•H g
-P CD
U O
CD . .
SH rd CD CD
SH C g Cn
O CD 4-> Cn
O 4-> en O
G 3 rH
G fd 13 CD
SH
O
13
G
rd
U
-H
4-1
rO
g
O
4J
3
<=C
13
C
rd
O
j_i
CD
IS!
t-H
rrj
3
C
g
1
CD
^
O CD
CD X!
x:
O 13
H
C 3
rd O
QjX!
en en
13
G
>i
rH
•H
rd
13
r£j
CD
g
o
UH
en
4J
G
CD
g
-P
en
3
•i— i
T)
rd
logged
CD
X)
13
rH
3
O
X!
en
libra-
rd
O
U
•H
g
rd
C
t>i
Q
CD
X!
13
rH
3
O
en
en
C
O
•H
4-1
CD
0)
1
•H
13
CD
g
O
U-l
J_)
CD
>i
rd
JH
CD
4-1
UH
rrj
SH
O
CD
•H
4J
O
CD
S-i
O
U
CD
0
G
C
CD
4J
C
-•H
rd
g
Q
en
G
O
-H
4-1
(C
M
Q)
Q)
O
C
Q)
CD
cn
H en en
porting A
djustment
nd Repair
CD (=
1
UH
O
n
ermination
i _\
CD
P
U rd
0)
en
O rX
G >il3 U
rd H G CD
e -H rd X!
M rd U
O Q O
U-l S-t G
M UH 0 rd
CD O N CL,
ft
f >
, 13
Test Perio
•C £
UH 0 3
°*s
> ^ft s >
* '
UH
Continuous
D
5 g
H fl) N^
J 4-> ^
ti en
D en
2.1
i_>
en 'rt
CD en en SH G
> rd 4-> CD O fd
CD
Repeat of
Performanc
en
CD
1 -P
en
r-1
Dynamic
ilibratioi
"J
U
r
Test as
Required
i
•H > en HO 4-1G g
CD CD4-I CDfd CDfdO -rH4->
O CD O
O G M O • 4-> OC3 CGO OO
C-HfdCDO en G-nen OCD
fd
g en H
H4->trird rdbi g3
G
rd fd
H
HCDGS g-nrd H4-> gO'
SH G • O -H MM 13 UH
O O H O ^ M 4-) OO4-1 H UH r
UH -H H
CD 3 en UH O 3 CD
M 4-> • O O 'O CD M UH G O
(0 SH U
d O
UH r
CD SH
H
m
3 4J
D fO
CDrdmGG4-> CD-H Xl ,M Xlv>i CDSH13
ft O
rdi3 ficn cnO4->H &,-H
•H G 13 G CD - en Cnrd rd
en
3 CD
CDU-tOCCDSH'^en 4->C|5 g4-)T33 4->O~lO
& -H -r-t rd 4-) -H O CD fd -H CD en
4J 0 4J
G3-HO CD4J4-) 4->
CDO4-)-iH&1SHO Djcnen eno)
UHfXCDCDrdCDCD3 CDCUCD >iXi
MO rd CD G
O 3 CD SH rd
O G d<
•H
CD-H CD en H
HencnggSnD-ien p^4J4J r_n4JS-i4-) Pnfdci,
H
CN
CN ro
CN CM
•^
CN
4-1
M
rd
x:
o
o
rH
m
fd
G
O
-iH
CD
a
o
4-)
G
o
U
CN
H
H
•
in
CD
tn
•H
PL,
2-31
-------
5.11.3.2 Preliminary Site Study—The optimum continuous
monitoring system for sulfur dioxide in stack gases can only
be defined after an in-depth study of the source in question
has been accomplished. The more important parameters which
will influence the choice of a system include:
1. Moisture content of the stack gas;
2. Stack gas temperature, temperature fluctuations,
stack gas velocity, and static pressure;
3. Existence of other gaseous components in the stack
gas, e.g., oxides of nitrogen, sulfuric acid;
4. Possible physical location of the sampling system,
an optimum sampling point, and physical dimensions
of the stack;
5. Particulate loading evident in the effluent;
6. Range of sulfur dioxide emission rates;
7. Ambient temperature range;
8. Level of competence and experience of the personnel
who will be charged with the operation and mainte-
nance of the monitoring system; and
9. Availability of laboratory facilities.
Most of the information asked for in the above list can
be obtained in a single1 visit to the plant site if not attain-
able from previously accumulated data from the same or similar
plants. Portable instrumentation is available to measure some
of the stack gas characteristics, i.e., pressure gage and dial
thermometer. The moisture content measurement can be made
using Method 4, Determination of Moisture Content in Stack Gases
(ref. 3). The gaseous constituents of the stack gas can be
determined with an Orsat analysis or estimated from details of
the particular process in question. The other parameters can
be sufficiently covered in discussions with the plant operator(s)
2-32
-------
5.11.3.3 System Definition and Procurement—The
determination of the optimal monitoring system for a given
stack must be made only after the preliminary site survey and
a careful study of several other parameters have been exe-
cuted. These parameters include:
1. Overall objective of the monitoring system program,
2. Expected operation and maintenance requirements
which can be tolerated by the program,
3. Need for ease in calibration,
4. Auditing program requirements which may be enforced.
In defining the overall objective of the sampling system, the
simplicity or complexity desired in the monitoring system
should become evident. The second parameter is definitely
dependent on parameter number 8 in section 5.11.3.2, the level
of competence and experience of the personnel to be charged
with the operation and maintenance of the system. Along these
lines, too, is the number of hours per week or per month of
the personnel's time which will be alloted to this role. The
need for ease in calibration will also be a function of the
operator's time and skills. As a result of defining potential
audit requirements, only systems with the desired characteris-
tics for auditing (i.e., reference gas inlets, etc.) will be
considered.
As stated in section 5.11.1.2 the system apparatus will
consist, generally speaking, of equipment for the reference
method measurement of the sulfur dioxide concentration, cali-
bration media, a continuous monitoring system (several varie-
ties of which are discussed below), and a data recorder. The
data recorder can be either an analog chart recorder or any
other suitable device with an input voltage range compatible
with the output of the analyzer system. If a magnetic tape
recording system is selected, a backup strip chart recorder is
recommended to meet immediate operation, maintenance, and
2-33
-------
troubleshooting needs.
Several types of continuous sulfur dioxide monitoring
systems are commercially available utilizing a wide range of
detection techniques. There are two main classifications for
these systems: in-situ and extractive. The in-situ measure-
ment system has the monitor mounted directly on the stack
where it senses the pollutant in the effluent stream without
need for removing samples from the stack gas. The majority
of the in-situ monitors are based on optical methods for tak-
ing the measurement at a point in the stack or taking a spa-
tially integrated measurement across the diameter of the stack.
The extractive monitor is one which requires the sample to
be drawn from the stcick and transported to the analyzer.
Depending upon the distance from the sampling point to the
analyzer and the analyzer's sample conditioning requirements,
the interface between the sampling point and the analyzer can
vary in complexity. Several in-situ and extractive monitors
are discussed below with indications given as to the types
of interfaces which might be considered when appropriate.
5.11.3.3.1 Dispersive spectroscopy. Several instruments
are available which utilize dispersive spectroscopy to measure
the gases across the stack rather than drawing a sample out of
the stack. One such instrument transmits a beam of light from
one side of the stack to the other so that it passes through
the gas stream. The polychromatic light beam contains certain
wavelengths of light which are optically absorbed by various
constituents of the gas. Hence, the reduction in the intensity
of the beam at a given wavelength as it is received at the
opposite side of the stack is proportional to the number of
molecules of the constituent absorbing at that wavelength.
The analyzer section of the instrument isolates the wavelengths
o
of interest (for SC>2 the corresponding wavelength is 3024 A)
and measures their individual intensities.
2-34
-------
A monitoring system of this type, although theoretically
simple, must be designed to withstand and/or compensate for
changes in ambient temperature and severe vibrations which are
likely at various times of the year. Other practical problems
which must be overcome include:
1. Calibration of the entire system.
2. Keeping the windows which separate the analyzer and
the stack gas clean,
3. Jarticulate concentrations in the stack gas which
will reduce transmission of the subject wavelength by
blocking its path rather than by S02 absorption,
4. Maintaining constant intensity of the source lamps,
5. Guaranteeing that the detectors maintain the same
detection sensitivity.
Two common variations to the dispersive spectroscopy
"double-ended" instrumentation described above are the "folded
path" design and the "slotted pipe" design. Unlike the "double
ended" system, the "folded path" system places both the source
and receiver at the same location. The energy beam is trans-
mitted through a slotted probe and reflected back into the
instrument for analysis. The "slotted pipe" extends either
between the transmitter and receiver or between the transmitter/
receiver combination and the reflector. The purpose of the
slotted pipe is either to prevent misalignment or to restrict
the absorption pathlength to maintain linear detector response.
One example of the more sophisticated "folded path"/
"slotted pipe" instruments is one which utilizes a flat zero/
read mirror. When the mirror is in the read position, reflected
light is directed back to the spectrometer. If SO2 is present
in the slotted pipe, it will absorb energy in regularly spaced
o
bands of 3024 A. The light reflected off the mirror is
reflected onto a diffraction grating which disperses the light
2-35
-------
allowing for a spatial display of the absorption spectra of
S02. When the mirror is in the calibration position, the light
is directed into the spectrometer bypassing the sample probe
which serves as a zero check. A temperature-stabilized gas
cell containing a known concentration of SC^ can then be
inserted into this same path to provide a span check.
An alternative to changing the position of the mirror is
to utilize a second light source in the analyzer which does
not pass through the stack gas. By inserting either a gas
cell of known concentration or a calibrated optical filter in
the path of the second beam, a calibration can be performed.
A possible means of performing a dynamic calibration is
to insert standard verified sealed cells into the path of the
original beam which will incrementally add to the unknown
stack gas. In this way a system operating curve can be deter-
mined. The entire dynamic calibration procedure can be accom-
plished automatically using a timer actuated solenoid arm
(ref. 5). Care must be taken, however, that the certification
value for the cell is determined at temperatures equivalent
to existing stack gas conditions. An alternative is for the
manufacturer to supply a calibration curve and to apply the
appropriate correction factors at each calibration,
Another of the problems stated above which has been
examined by several manufacturers is that caused by the accumu-
lation of particles on the windows which would block transmis-
sion of the light beam. A system designed to alleviate this pro-
blem is one which uses a specially designed forced-air purge system
to avoid dirt buildup on the windows and adjoining hardware.
The air for the system is supplied by a self-contained blower
which works continually (ref. 6).
2-36
-------
The third problem stated above deals with the particulate
loading characteristic of the stack gas. A relatively simple
way to avoid this problem is simultaneous detection of two
wavelengths of the polychromatic light beam in a double-ended
system or of the reflected light in the folded path system.
One wavelength, of course, is that which is absorbed by SC^-
The second is a reference wavelength not absorbed by any
stack gas constituent. Any deterioration of the intensity at
the second wavelength is not due to absorption and must be
caused by interference from the particulates in the gas stream.
This deterioration at the second wavelength can be appropri-
ately assigned to the first wavelength as an addition to the
originally measured intensity.
5.11.3.3.2 Second derivative spectroscopy. Second deriv-
ative spectroscopy is another in-situ method for measuring SO_
in stack gas emissions. It is an attempt toward greater spec-
ificity and resolution by determining the derivatives of the
intensity with respect to wavelength of a light beam which has
passed through the stack gas. The optical technique used in
this type of instrument is wavelength modulation where the wave-
length radiation is modulated sinusoidally with time.
One instrument utilizing this technique has within it a
grating monochrometer which spectrally disperses the radiation
from an ultraviolet source. The wavelength of this monochro-
matic radiation is then varied with respect to time by laterally
oscillating the entrance slit position electromechanically at a
frequency of 45 hertz. The amplitude of the wavelength modu-
lation can vary depending on the peak-to-peak displacement
chosen between .25 and 1 mm. The signal then enters an absorp-
tion cell with a wavelength centered about a value set by the
grating position and slightly modulated in time at 45 hertz.
2-37
-------
In the absorption cell, the light is reflected back and forth
across the length of the cell several times until it is focused on
a detector which is generally a photomultiplier tube. The
90 hertz component of the output signal from the photomulti-
plier tube is electronically analyzed using a fundamental
phase-lock amplifier technique. The amplitude of the 90 hertz
signal is presented as an output dc voltage which is propor-
tional to the second derivative of intensity with respect to
wavelength. Electronically dividing this signal by the total
output voltage of the PM tube results in a signal proportional
to gas concentration (ref. 7) .
All of the problems stated in section 5.8.3.3.1 are
applicable to the second derivative spectroscopy method. Cali-
bration at selected concentrations can be niade manually or
automatically using standard, sealed samples of SO- which are
inserted into the light path internally to provide measure-
ments corresponding to known concentrations. The problem of
accumulation of particulates on the windows encasing the optics
can be handled in very much the same way as stated above.
Because the technique is sensitive to particulates in the
stack gas, a ceramic probe filter can be used so as not to
allow particulates to enter into the measurement. All of the
electronics are located separately from the optics which are
in a sealed encasement. With this situation then, preventive
maintenance on the electronic components can be performed
without disturbing the rest of the system.
Two other parameters which should be discussed in relation
to second derivative spectroscopy are: sensitivity to high
stack gas temperatures and interferences by other constituents
of the stack gas. Designs available on the market today can
withstand stack gas temperatures up to 540°C. Some specificity
can be lost in the ultraviolet absorption method when two
2-38
-------
different constituents have absorbed some amount of energy at a
given wavelength. In other words, at wavelength X , several
constituents may add to the reduction in intensity. However,
with this technique of optically determining the second deriv-
ative, the amount of loss of intensity due to SOj can readily
be differentiated from absorption by any other constituent.
5.11.3.3.3 Extractive ultraviolet absorption. One of
the more simple extractive methods is that which extracts a
sample of the stack gas, maintains the sample at the high
stack gas temperatures, analyzes the sample for the sulfur
dioxide content, and then discharges the sample back into the
flue gas. Either a split beam or a dual beam source can be
used in the analysis. Simply, the first beam, at a given
wavelength absorbed strongly by S02, passes through the sample
cell through which the sample of stack gases flows. The
intensity of the light is detected by a wavelength phototube
along with the intensity of a second beam. The second beam
has passed through an optical filter that transmits only the
reference wavelength which is not absorbed by sulfur dioxide
and then through the sample cell simultaneously with the
first beam. The phototubes generate electrical signals which
are proportional to the intensity of each signal. The loga-
rithms of the two signals can then be compared and a single
photometer signal outputted which is proportional to the
concentration.
Advantages of such a system include:
1. Long sample lines are not required when the photometric
system is mounted on the stack;
2. Systematic errors such as particulate buildup on
the windows of the sample cell are avoided because
of the use of a reference;
2-39
-------
3. System components are at a minimum so that trouble-
shooting and the subsequent repair of equipment
should require considerably less time than other
more sophisticated systems;
4. Response is generally linear over a wide range;
5. The optics, the electronics, and the sample cell can
be' isolated from each other which allows the operator
to perform maintenance on any one of the three without
affecting the others;
6. Zero calibration can be performed by simply back-
flushing the probe and sample cell which simulates
zero absorption;
7. Span calibration can be performed by dropping a
reference cell of known concentration into the path
of the light beam during the backflushing operation;
8. Due to the elevated temperatures of the sample at
all times, problems due to condensation and cor-
rosion of equipment may be avoided;
9. An air aspirator is used to extract the sample as
opposed to a sample pump which is generally much
less reliable for long-term monitoring;
10. A system such as this can be readily adapted to
several different sources (ref. 8) .
5.11.3.3.4 Nondispersive infrared. Most analyzers using
the nondispersive infrared (NDIR) technique consist of either one
light source with a light chopper or two identical sources
whose beams are directed through two different cells. One of
the cells contains a gas which does not absorb infrared energy
at the same wavelengths at which sulfur dioxide absorbs infra-
red energy. Passing through the other cell is the sampled
stack gas. The beams pass through both of these cells and into
different half sections of a reference chamber. Separating the
2-40
-------
two half sections of the reference chamber is a flexible metal
diaphragm. Both sections contain the same amount of SG>2 vapor
kept at the same atmospheric pressure. The degree of absorp-
tion of infrared energy by the sample gas is directly propor-
tional to the amount of S02 in the sample gas. The absorption
by the sample gas will proportionally reduce the absorption by
the S09 vapor in the corresponding half section of the refer-
^w
ence chamber. The difference between the energy absorptions
in the two halves on the reference chamber, then, is a measure-
ment of the concentration of S0~ in the sample gas.
The flexible metal diaphragm when used with a stationary
metal plate forms one side of two-plate variable capacitor.
The capacitance, which changes with the detention of the dia-
phragm, is measured. The detention of the diaphragm is
proportional to the difference between the energy absorption
of the SC>2 vapor in the two sides of the reference chamber.
Hence, an electrical signal is generated which is propor-
tional to the concentration of SO,., in the sampled stack gas.
The primary sources of error in the NDIR method are the
blocking of the transmission of the light beam by particulates
and the inadvertent,absorption of infrared energy by moisture
in the sample gas. Both of these sources of error can be mini-
mized by adequate interfacing equipment, varieties of which
will be discussed below.
The sampling interface used with an NDIR analyzer must be
capable of removing fly ash and particulate matter as well as
removing or determining the quantity of moisture in the sample.
Particulate matter will tend to collect on the windows of the
sample cell. Water vapor will interfere inasmuch as the SO2
absorption band is overlapped by a water system in the
1200 cm"1 to 1400 cm"1 region.
2-41
-------
One possible sampling interface is described in reference
6 and shown in figure 5.11.3. The 0.5 micron filter used in
this system is adequate for removal of the small amount of
particulate in the emission from a source having a bag house
in use. The filter is placed in the stack as a portion of the
sample probe. Heat from the stack gas will keep the filter
hot and keep moisture from condensing as the gas is drawn
through the filter. When stack gases are not filtered by a
bag house filter systeir. or equivalent, it is necessary to in-
stall a portable bag house-type filter ahead of the 0.5 micron
filter in the sample line.
The sample gas transmission line shown in figure 5.11.3
is made up of 100-foot long Teflon segments with an individual
heating system for each segment. The overall transmission
heating system should maintain a nominal 120°C (250°F) gas
temperature from the sample probe to the first heat exhanger.
The heated gas from the transmission line is rapidly
cooled from 120°C (250°F) to 1°C (34°F) to remove all moisture.
To do this, a heat exchanger is usually employed to reduce the
gas temperature between 55 and 85°C (100 to 150°F) depending on the
atmospheric temperature. The sample gas is then passed through
an electro-mechanical refrigeration unit to further reduce the
gas temperature. Approximately 90 percent of the sample gas
is then removed from the conditioning system via the stack gas
pump. The remaining sample gas is passed through a second
electro-mechanical refrigeration unit to reduce the gas temper-
ature to approximately 1°C (34°F). The reduction of tempera-
ture causes the moisture in the gas to condense into a liquid.
A collection device is employed in conjunction with the above
conditioners to remove the liquid from the system and holds the
liquid for later measurement as evidence of the moisture con-
tent of the stack gas.
2-42
-------
INSIDF. GAS STACK
PROBE
=L
FILTER
DEKARON TRANSMISSION LINE
VACUUM
GAGE
NSTRUMENT
SAMPLE PUMP
STACK GAS
SAMPLE PUMP
SPAN GAS
BOTTLE
ZERO GAS
BOTTLE
GAS CONTROL VALVE
NSTRUMENT ANALYZER
EXHAUST GAS MANIFOLD
EXHAUST
Figure 5.11.3.
Example of an extractive sulfur
dioxide monitoring system.
2-43
-------
The stack sampling probe, when inserted in the stack, may
be at either positive or negative pressure. A system for
extracting the gas can either employ a pressure pump at the
stack, which would require the pump to operate at high temper-
atures, or a vacuum pump located after the conditioning equip-
ment. In the example illustrated in figure 5.11.3, the latter
and more practical configuration is used. The sample gas
leaving the pump is at a positive pressure sufficient to
satisfy the flow rate and pressure needs of all analyzers.
The pumped sample gas is then piped to the input of the gas
analyzer.
5.11.3.3.5 Pulsed fluorescence. In the pulsed fluor-
escence method, the extracted stack gas sample is passed in
front of a source of pulsed ultraviolet illumination equipped
with a monochromatic filter. The molecules of sulfur dioxide
contained in the sample will be energized by the high intensity
light source and will emit a monochromatic illumination. This
emitted light is linearly proportional to the concentration of
sulfur dioxide molecules in the sample. It is usually sensed
by a photomultiplier tube through a narrow-band filter and a
proportional analog output is generated.
The most commonly used region of ultraviolet radiation is
the 2300A to 1900A region. In this region sulfur dioxide mole-
cules exhibit stronger absorption and their fluorescence is
less likely to be quenched by other constituents of a stack gas
(ref. 9).
The quenching effects of oxygen, carbon dioxide and nitrogen
in concentration levels different from the ambient air levels
are discussed in reference 10. Calibration of pulsed fluores-
cence analyzers is accomplished using SOp-in-air reference
mixtures which alters the analyzer accuracy for other constituent
levels in the stack gas. As shown in test results discussed in
2-44
-------
reference 9, a reduction in oxygen level below 21% (ambient
level) will linearly increase the analyzer reading. A rise in
the CO- level in the stack gas will decrease the analyzer reading
with respect to the true value. The quenching effects due to
N2 are negligible in comparison to effects due to C>2 and CO-.
The level of water vapor content of the stack gas must be
kept constant to avoid inconsistent quenching of the fluorescence.
As discussed in reference 9, studies have shown that the pulsed
fluorescent analyzer, which has included within its internal
components a dryer which reduces the moisture to a very low and
discernible level, will not require any external drying devices
to maintain insensitivity to the moisture in the stack gas. In
the case of pulse fluorescent analyzers without internal drying
mechanisms, an external dryer or refrigeration unit must be used.
As in the majority of the extractive continuous analyzers,
particulate removal is important so as not to clog the analyzer
plumbing or to block the detection of light. Heated sample
lines are also vital in preventing water condensation and the
loss of sulfur dioxide from the sample.
Calibration of the system can be accomplished by inputting
zero and span gases at the probe and/or at the analyzer.
5.11.3.3.6 Flame photometric detection. The flame photo-
metric principle utilizes the detection of light, centered at
394 nm, emitted by sulfur-containing compounds when burned in a
hydrogen-rich flame. The burning creates an excited state
species of sulfur and the release of light energy upon the
return to ground state sulfur is the quantity measurable with
a photomultiplier tube optically shielded from the flame. The
response of the detector, which generally views the light
through a narrow band-pass interference filter, is directly
related to the concentration of total sulfur entering the
detector per unit time.
2-45
-------
The advantages of ci flame photometric detection system
include: a high degree of sensitivity to sulfur compounds,
low maintenance and downtime, and the fact that the system
is readily calibrated using NBS traceable permeation tubes
(ref.ll). Disadvantages; include downtime due to "flame-out,"
excessive analyzer drift, and sensitivity to small variations
in gas pressures and flows. In addition, the analyzers avail-
able on the market for source monitoring purposes must be
coupled with interface systems which have heated sample lines,
remove particulates, dry the sample, and dilute the sample to
within the measurement range of the air quality analyzer. The
interface systems should also be equipped to remove all sulfur-
containing compounds other than SO-.
Several of the disadvantages listed above are taken care
of in standard monitoring packages. Ceramic filters installed
in the probe will remove the majority of the particulates and
portable bag-house type filters can be installed downstream
from the probe if more protection is necessary. Downtime form
"flame-out" can be minimized with the use of an automatic flame
reignition system which senses the flameout condition and cor-
rects it by turning off or averting the hydrogen flow and reig-
niting the flame within 15 to 20 seconds. The hydrogen is then
started up or redirected across the flame. The problem of drift
can be avoided with daily or twice-daily calibrations. Utiliz-
ing an automatic calibration system is the most feasible way
to meet this requirement.
There are several different types of dilution systems
available. The disadvantages of all of them are:
1. The depletion of moisture is not adequate for most
analyzers which have a characteristic sensitivity
to water vapor.
2-46
-------
2. A more complex gas control system is required to
keep the flow rates constant.
An example jDf a dilution system is one which utilizes a perme-
able membrane. The dilution ratio acquired from such a system
would depend on the volume flow of the carrier gas stream, the
surface area and temperature of the membrane, and the perme-
ability of the membrane to the component of interest.
5.11.3.3.7 Procurement guidelines. Upon purchase of the analy-
zer and associated support equipment, certain procedures should
be followed which permit the operator to determine the adequacy
of the equipment. The analyzer system must ultimately meet the
specifications outlined in section 5.11.1.4. To determine the
acceptability of the analyzer, the operator must do one of two
things upon purchase of the instrument: 1. Acquire and review
the manufacturer's bench test data. Even more preferable is
the viewing of the actual bench testing at the factory by the
operator prior to shipment of the equipment; or 2. Design and
perform a bench test which simulates actual operating conditions,
generate, and review output data. The latter suggestion requires
immense experience on the part of the operator with the specific
equipment. It is beyond the scope of this document to define
a specific bench testing process, the extent of which is depen-
dent on the time and economic factors specific to the system
at hand, e.g. the necessity for the system to successfully meet
the performance testing requirements within a given time sche-
dule would alter bench testing requirements.
Prior to installation of the sampling interface and record-
ing equipment, the operator should simulate a typical operating
sampling condition, make several measurements, and study the
validity of the output data. Whenever quantitative measurements
are not practical, the operator should make a visual inspection
of each component to locate obvious defects and/or malfunctions.
2-47
-------
The calibration gas mixtures must be checked upon receipt
using Method 6 measurements. Nominal concentrations of 0 per-
cent of span, 50 percent of span, and 90 percent of span
must be acquired. The average concentration value for each
mixture, based on at least three measurements taken at least
24 hours apart, should not differ from the stated concentra-
tion value by more than 5 percent.
Any of the components of the system which do not meet
acceptable quantitative or qualitative limits should be rejected
prior to installation. Report and correct any malfunctions,
making this information available to all operating personnel.
5.11.3.4 Installation and Instrument Check Guidelines--
Because each system is usually specifically designed to meet
the requirements at the stack in question, it is recommended
that the system design engineer or manufacturing representa-
tive take the responsibilities of installation and initial
system checkout at the site. Also due to the uniqueness of
each system, it is advisable for the operator(s) of the system
to be present at the installation and to actively participate
in the system testing. With this experience and a detailed
operations and maintenance manual, the operator should be able
to trouble shoot, define, and resolve most of the problems he
may encounter during continuous operation. If the operator is
not skilled in the areas of electronics or chemistry, skilled
personnel should be available at all times to aid in resolving
problems with the minimum of downtime.
Of utmost importance during the installation and initial
system testing is that each step and the results be well docu-
mented at the front of the Operation and Maintenance Logbook
which should be updated daily by the operator. Installation
specifications as stated in Performance Specification 2 are
2-48
-------
reprinted in section 5.11.1.4.
For sampling of sources where the effluent gases cannot
be assumed to be nonstratified, Performance Specification 2
(section 5.11.1.4) requires that one of the following mea-
sures be taken:
1. The procurement and installation of either a con-
tinuous oxygen or a continuous carbon dioxide mon-
itoring system which meets the requirements of Per-
formance Specification 3 (reproduced in appendix A
of this document).
2. The procurement and installation of extractive S02
monitoring systems using multipoint sampling probes
or in-situ pollutant monitoring systems that sample
or view emissions which are consistently represen-
tative of the total emissions for the entire cross
section.
The continuous measurement of oxygen can be accomplished
with analyzers utilizing either the paramagnetic technique or
a zirconium element. Both of these systems require a clean,
dry sample which indicates the need for a sample interface
system such as described below in section 5.11.3.3.4. There
are no major interferences when using either of these methods.
However, problems do exist if particulate matter is allowed to
build up in the lines or if there are leaks into the system.
The continuous measurement of carbon dioxide is most com-
monly achieved with a nondispersive infrared analyzer. The
problems and error sources associated with this technique are
discussed below in section 5.11.3.3.4.
Research and development of multipoint probes has not yet
reached a point where they are readily available to the user.
Reference 4 provides a detailed description of errors due to
non-stratification and the ways in which the use of multipoint
probes will alleviate many of the inaccuracies caused by point
source continuous monitoring.
2-49
-------
Some general statements which apply to most in-situ and
extractive systems unless specified otherwise are listed below.
In some cases the manufacturer will make recommendations as to
the correct procedures. If not, the experience of the operator
and/or installation personnel will play an important role in
the success of the system.
The statements are as follows:
1. Temperatures outside the ambient temperature limits
of the analyzer may necessitate the use of special
environmental control equipment. This equipment
should be set up and tested for an undisturbed oper-
ating period of at least 1 month before installation
as support equipment in the system. The environment
during the testing period should be similar to the
extreme environmental conditions to be experienced
at the analyzier location. During this test period,
the temperature at the analyzer location should be
measured and recorded continuously or at intervals
which suffice: in determining that the temperature
requirements of the analyzer will be maintained.
2. Mount the analyzer horizontally or vertically as
recommended by the manufacturer.
3. Depending on the specific location of the analyzer
and the analyzer's sensitivity to vibration, pre-
cautions may or may not need to be taken to dampen
vibrations to which the analyzer may be subjected.
One suggestion is to use foam rubber or the equiv-
alent to isolate the instrument from the mounting
platform.
4. Typically, sample handling systems for extractive
monitors will incorporate the following:
2-50
-------
a. Valves to permit selection of gas sample,
zero calibration gas or upscale calibration gas.
*-*
b. needle valve inlet in sample line for flow
adjustment,
c. flow meter for flow measurement and/or indica-
tion of flow stoppage,
d. filters and filter holders.
Each of the above should be checked for cleanliness,
proper operation, and no undue disruption of air flow.
5. The layout of the equipment can be very important to
the operator's efficiency. A suggestion in this area
is that the recorder be near enough to the amplifier/
control section so that the operator can easily
observe the response to adjustments of the controls.
In the majority of systems the recorder response
should be observed rather than the meter response on
the control panel.
6. The type of pump chosen for the gas sampling in an
extractive system will dictate the location of the
pump. A simple water or air aspirator can be placed
downstream of the analyzer and will exhibit distinct
advantages over diaphragm and bellows pumps since
aspirators have no internal working parts and are
less prone to failure (ref. 12).
5.11.3.5 Performance Testing—The performance specification
test procedures presented in section 5.11.1.5 are reproduced
from Appendix B as published in the Federal Register, Octo-
ber 6, 1975. The testing procedures include the following
independent tests:
• Calibration test
• Field test for accuracy (relative), zero drift,
and calibration drift
• Field test for response time
2-51
-------
5.11.3.5.1 Calibration test. This test consists of an
analysis of the calibration gas mixtures (if used rn the system
in question) and a calibration of the complete measurement
system either in the laboratory or in the field. The latter
is preferred^ Nominal concentrations of 50 percent and 90
percent of span are recommended. The method of analysis is
as directed by Method 6. The analysis must be accomplished
within 2 weeks prior to start-up of the system with the
objective being to "demonstrate (the calibration gases)
to be accurate and stable," (section 5.11.1.2). Alternate
methods may be acceptable if approved by the enforcement agency.
Record all results on a form as shown in figure 5.11.4.
The entire system is calibrated with a series of five non-
consecutive readings for each concentration of span gas, e.g.,
50 percent, 90 percent, 50 percent, 90 percent, etc. This
method is subject to the particular operation of the system in
question. Alternative procedures may be acceptable as recom-
mended by the manufacturer and approved by the enforcement
agency. Record all data on a form such as shown in figure
5.11.5.
5.11.3.5.2 Field test for accuracy (relative), zero
drift and calibration drift. This test extends over a 2-week
period after the measurement system has been installed and started
up per the manufacturer's directions. First the operator must
offset the zero setting by at least 10 percent so that the
negative zero drift characteristic, if any, of the analyzer in
question can be quantified. The system should be operated for
one week (168 hours) in a normal operational manner. This is
called the conditioning period.
The conditioning period is immediately followed by a
168-hour operational test period when the system is on line at
all times except during calibration or backflushing of the
2-52
-------
ANALYSIS OF CALIBRATION GAS MIXTURES
DATE
REFERENCE METHOD USED
INITIALS
Calibration Gas Mixture "A":
Cylinder or cell designation
Test No. 1 ppm
Test No. 2 ppm
Test No. 3 ppm
Total _ ppm
Average = -° a = ppm
Calibration Gas Mixture "B":
Cylinder or cell designation_
Test No. 1 __ ppm
Test No. 2 _ ppm
Test No. 3
Total _ _ ppm
Average = -°- a = ppm
Figure 5.11.4 Sample data form for the analysis of
the calibration gas mixture.
2-53
-------
CALIBRATION ERROR TEST DATA
DATE
OPERATOR
CALIBRATION GAS MIXTURE "A": (
Reading
ppm)
Time
Measurement Sys. Reading
Volts
ppm
**
"'A
CALIBRATION GAS MIXTURE "B": (
ppm)
Reading
Time
Measurement Sys. Reading
Volts
ppm
**
**
Average value determined in analysis of calibration gas
mixtures.
k
d = Difference = Average Value (ppm) - Measurement Syst.
Reading (ppm)
ngure 5.11.5 Sample log sheet for calibration
error test data.
2-54
-------
sample lines and other support equipment. If, at any time
during this second period, monitoring is interrupted for main-
tenance procedures, the operational testing time up to the
interruption must be noted in the records and repeated (sec-
tion 5.11.1.6).
The field test for accuracy requires that a minimum of
nine SO- concentration measurements be taken in accordance
with the procedures stated in either Method 6 or 8 depending
upon the regulations for the particular source in question.
The sampling period should be 1 hour. In all sampling cases
the Reference Method sampling probe should be placed as near
as possible to the sampling point of the system in question.
Record this test data on a form such as demonstrated in figure
5.11.6.
Note: It is recommended that a chemist be available on site
to analyze the sample immediately following the sample
recovery. This will greatly enhance the efficiency of
the testing procedures.
If the basis (wet or dry) for the continuous measurement
system is inconsistent with the Reference Method measurements,
simultaneous moisture tests should be performed and a correc-
tion factor determined. Report the moisture test procedures
and data along with the correction factor calculations in the
performance test logbook. The column labeled "Reference Method
ppm" in figure 5.11.6 is designed for recording the reference
method data after the correction factor has been applied.
The field test for the 2-hour zero drift and 2-hour span
drift consists of an accumulation of 15 sets of data at 2-hour
intervals each. The zero and span values can be produced by
pre-tested calibration gases, simulated using equivalent gas
cells, or simulated electrically or mechanically within the
analyzer if approved by the enforcement agency. The 2-hour
intervals must not overlap and do not have to be consecutive.
2-55
-------
OPERATIONAL TEST DATA
START:
DATE_
TIME
END:
DATE_
TIME
OPERATOR(S)
ACCURACY DETERMINATION:
Test
No.
Oper.
Init.
Date
Time
Ref. Method
(ppm)
Analyzer Average
Interval Volts
ppm
di
d.2
d = Difference = Analyzer Average Value (ppm) - Reference Method Value (ppm)
Figure 5.11.6 Sample data form for operational test.
2-56
-------
OPERATIONAL TEST DATA
2-HOUR ZERO AND CALIBRATION DRIFT DATA:
Oper.
Init.
Date
Time
Begin
End
Zero
Reading
AZ
AZ'
' Span
Reading
Span
Drift
10
11
12
13
14
15
Figure 5.11.6 (Cont.) Sample data form for operational test,
2-57
-------
-------
Record the readings (fig. 5.11.6). Definitions of symbols is
found in Appendix B of this document.
During the field testing, zero and span calibration cor-
rections and adjustments are permitted only once every 24 hours.
Intervals shorter than 24 hours may be approved if recommended
by the manufacturer. The values given by zero and span before
and after the 24-hour interval adjustments are made should
be measured and recorded as shown in figure 5.11.6 Automatic
calibration systems may be used and their corrections can be
made at any time.
5.11.3.5.3 Field test for response time. This test
should be accomplished with the entire system in normal oper-
ation as specified by the manufacturer's instructions. Zero
gas is introduced at the probe or as close to the probe as
possible. When the output reading has stabilized, a span gas
mixture of known concentration is injected at the same point
that the zero gas had been injected. The recorded response
time is the time between the switching to a span gas and a
stable output at the span gas value. The span gas used for
this test should be 70 to 90 percent of span. This sequence
of operation should be repeated three times and the results
recorded on a log sheet as shown in figure 5.11.7.
5.11.3.6 Data Reduction, Validation, and Reporting—Of
utmost importance during the performance testing is that all
test data are logged correctly and directly upon acquisition
of the data. The figures referred to in the previous sections
should all be included in a single performance testing logbook
in chronological order. Having the data contained in a single
notebook will expedite the next steps which are data reduction,
validation, and reporting. Figure 5.11.8 shows a simple layout
for a form on which the necessary calculations can be made for
reporting the test data. Section 5.11.1.6 should be referred
to in making those calculations.
2-59
-------
DATE
OPERATOR
RESPONSE TIME TEST DATA
Test No.
Concentration
(ppm)
Upscale
Response Time
(seconds)
SUM 16
AVERAGE
Test No.
Concentration
(ppm)
Downscale
Response Time
(seconds)
Figure 5.11.7 Sample data form for response time test.
2-60
-------
DATA ANALYSIS REPORT CALCULATIONS
DATE
OPERATOR
CALIBRATION ERROR DETERMINATION "A"
Mean difference = —— = ppm
95% Confidence interval = C.I-95 = .2776 \5(Sum 2)-(Sum I)2
(Mean difference + C.I.g5
Calibration error for "A" = Avg. calib. gas conc. - 100
percent
CALIBRATION ERROR DETERMINATION "B1
Mean difference = —— = ppm
95% Confidence interval = C.I.g5 = .2776 V5 (Sum 4)-(Sum 3)2
percent
Mean difference
Calibration error for "B" =
+ C.I.
95
Avg. calib. gas conc.
percent
x 100
Figure 5.11.8 Sample form for data analysis report
calculations.
2-61
-------
DATA ANALYSIS REPORT CALCULATIONS
ACCURACY DETERMINATION
Mean difference == —3— = ppm
95% Confidence interval = C.I-95 = .0906 ^9(Sum 6)-(Sum 5)
Mean reference method value =
Mean difference + C.I.gg
Accuracy = Mean ref. method value * 100
= percent
ZERO DRIFT (2-hour)
Mean zero drift = = ppm
C.I. (zero)95 = .0382 Vl5 (Sum 9)-(Sum Q)2
Emission standard =
Mean zero drift + C.I.(zero)Q5
Zero Drift =
Emission Standard
percent
100
Figure 5.11.8 (Cont.) Sample form for data analysis
report calculations.
2-62
-------
DATA ANALYSIS REPORT CALCULATIONS
CALIBRATION DRIFT (2-hour)
Mean calibration drift = S"" 10 = ppm
C.I.(span) = .0382 \15 (Sum 11)-(Sum 10)2 =
95
Emission standard =
Mean span drift + C.I.(span)„,.
Calibration drift = =—, = 5 3 x ion
Emission standard ±fjtj
= percent
ZERO DRIFT (24-hour)
Mean zero drif i = —£LJ±- = pprn
C.I.(zero)g5 = .1427 V 7(Sum 13 )-(Sum 12 )2 =
Emission standard = ppm
(Mean zero drift!+ C.I.
(zero) qt-
Zero drift = =—= = r—3 5 — x 100
Emission standard
percent
Figure 5.11.8 (Cont.) Sample form for data analysis
report calculations.
2-63
-------
DATA ANALYSIS REPORT CALCULATIONS
CALIBRATION DRIFT (24-hour)
Mean calibration drift = Sum 14 = ppm
C.I.(span)95 = .1427 V 7(Sum 15)-(Sum 14;2
Emission standard = PPm
Mean span drift + C.I.(span)g5
Calibration drift =
Emission standard
percent
100
RESPONSE TIME
System reponse time = slower response time =
Percent deviation from slower time
_ avg. upscale - avg. downscale
x
slowest time
percent
Figure 5.11.8 (Cont.) Sample form for data analysis
report caculations.
2-64
-------
The personnel responsible for doing the data analysis for
the performance test should date the data analysis form and
initial it. Preferably this personnel will have participated
in the field testing, hence will be familiar with the data and
its origin.
The first determination which must be made is that of
the calibration error which applies to the calibration gas
mixtures. In figure 5.11.8 the terms "A" and "B" refer to
the raid-range gas mixture and the span range gas mixtures,
respectively. For each error determination the mean dif-
ference between the actual value of the gas mixture (as
resulted from a Method 6 test on the mixture within at
least one week prior to the calibration test) and the value of
the gas mixture measured by the continuous analyzer system.
These are calculated using the following equation:
= n i "i
where X. = individual difference values
^ = sum of the individual values
X = mean difference value
and n = number of data points .
More specifically, the mean difference for each calibration
value is calculated in this way (i.e., d = X)
23 (Measurement system reading . -actual value of "A")
2-65
-------
and
5
2J (Measurement system reading.-actual value of "B")
The 95 percent confidence interval, as defined in section
5.11.1.6, is calculated using the following equation:
r- T - f '=> /„ T v ^ _ r Y \ (4)
C.I.g^ - -*/n|i X.. ; - ( i AJ ) ^;
where £X. = sum of all data points,
t.975 = tl-a/2
and C.I. q,- = 95 percent confidence interval
estimate of the mean value of the differences.
The value for t Q7[- is dependent on the number of data points
• _/ / o
which, in the case of the calibration test, is five (n=5) for
each mixture. A table from which t 9?5 can be drawn for
values of n up to 16 is given below (table 5.11.2 ).
The calibration error for each gas mixture is determined in
the following manner:
(Mean differencej+ C.I. gr /c\
Calibration error = . calib. gas concentration x 10°'
The value CIqr- defined in section 5.11.1.6 is actually the
value which, when added to the mean, defines the upper limit
of the 95 percent confidence interval.
2-66
-------
Table 5.11.2 Values for t
.975
n
2
3
4
5
6
7
8
9
t.975
12.706
4.303
3.182
2.776
2.571
2.447
2.365
2.306
n
10
11
12
13
14
15
16
t.975
2.262
2.228
2.201
2.179
2.160
2.145
2.131
Av/^^X^N^V^(XJ^X^^^^^N^)?
This error is reported in percent and is not to exceed 5 per-
cent of each calibration gas mixture value (sec. 5.11.1.4).
The accuracy determination, reported in percent, is based
on the values recorded on the Operational Test Data Form
(fig. 5.11.6) as a result of the Method 8 sampling and analysis,
The equation used for the determination is:
Mean difference + C.I.
.95
Accuracy Mean reference method value
x 100
(6)
where
and
Mean difference =
Mean reference
method value
the average difference between
the Reference Method measurement
values and the corresponding
interval averaged data from
the continuous measurement
system,
the value calculated using
equation (1).
2-67
-------
The accuracy should not exceed 20 percent as specified
in section 5.11.1.4.
The 2-hour zero drift, 2-hour calibration drift, 24-hour
zero drift, and 24-hour calibration drift values are each
calculated using the same equation:
Mean drift
Drift = -
+ C.I. qD-
x 100 (7)
Emission standard
where Drift is reported in percent of emission
standard,
C.I. Q[- is calculated as shown in equation (4) ,
• y D
and Emission standard is that which applies to the
source in question.
The limits of the drifts, as set forth under Performance
Specification 2 in section 5.11.1.5, are as follows:
Zero drift (2 hour) £ 2 percent of span
Zero drift (24 hour) <_ 2 percent of span
Calibration drift (2 hour) ... <_ 2 percent of span
Calibration drift (24 hour) . . . <_ 2.5 percent of span
The response time calculation is based on data documented on
the form such as shown in figure 5.11.7. The system response
time is determined by the slower average response time (either
the measured upscale response time or the downscale response
time) and should not exceed 15 minutes. The specifications
(section 5.11.1.4) require that the response time be reported
2-68
-------
as the percent deviation from the slower time which is calcu-
lated in this way:
Percent deviation _ Avg. upscale - avg. downscale x ,QQ ,
from slower time slower time
5.11.3.7 General Operation and Maintenance Procedures for
Non-Compliance Monitoring--The performance test as described
above serves three purposes. They are:
1. Determines the adequacy of the installed system
in meeting the performance specifications
detailed in section 5.11.1.4.
2. Provides a 2-week period of uninterrupted data
collection which, if the system meets the per-
formance specifications, can be used to deter-
mine the compliance of the stack being monitored
to existing regulations.
3. Provides a basis for determining the operation
and maintenance procedures which will be required
by the continuous system as it is installed at
the location in question.
This section of the quality assurance document will expound on
various aspects of purpose number 3. Because Performance
Specification 2 is applicable to such a wide variety of sul-
fur dioxide monitoring systems, only general guidelines on
operation and maintenance can be provided here for the moni-
toring system when it is not being subjected to the performance
test and is not being used for compliance monitoring.
Unless otherwise specified, the data from the continuous
monitoring system should be reduced, validated and reported on
a regular basis. Generally one hour average computations are
required. These computations should be made using four equally
2-69
-------
spaced points over a 60 minute period commencing on the hour.
The system must be automatically checked for zero and span
drift and adjusted accordingly at least every 24 hours. When
the system is not equipped with automatic calibration at zero
and span, then a manual check should be made at least twice
weekly or as recommended by the manufacturer. A dynamic
calibration, the input of at least two non-zero calibration
gases and appropriate adjustments made, and a response time
check should be performed routinely once every two weeks.
Depending upon the drift characteristics of the particular
system in question, each of these statements may or may not
have to be altered as the operator gains experience with the
analyzer, interface system, and source effluent.
In all cases, calibration gases should be inputted as
close to the sampling probe as possible so that the calibra-
tion will include the interface system. For those systems
where this is not possible, the use of Method 6 or 8 measure-
ments at least once per month is recommended. Calculations
and comparisons of the Reference Method data with the corres-
ponding continuous system data would follow the procedures
discussed above in sections 5.11.3.5 and 5.11.3.6.
When a dynamic calibration is performed at least twice
a month, the use of Method 6 or 8 (refs. 1 and 2) measurements
once per quarter is suggested. This would involve taking at
least four samples, analyzing them, and comparing the data with
the continuous measurement data for the corresponding time period,
The comparison calculations would again follow the procedures
set forth above.
Manufacturer's recommendations should be followed in per-
forming preventive maintenance on the system in the first
weeks of operation. If the system exhibits excessive drifts
2-70
-------
or if accuracy tests prove the system to be malfunctioning,
preventive maintenance procedures should be reevaluated and
rewritten in view of the operating experience as discussed
below.
5.11.3.7.1 Use of control charts. The use of control
charts is highly recommended for three reasons:
1. They provide an "at-a-glance" description of
the operation of the system.
2. They are a visual aid in maintaining consistency
in the operation and maintenance procedures.
3. Trends in the monitoring data are obvious with
the continued use of control charts and oncoming
problems can often be foreseen and avoided.
One example of a possible use of control charts is in
determining zero drift trends with a chart as shown in
figure 5.11.9. In this example, the UCL (upper control limit)
and LCL (lower control limit) used are +5.0 percent and
-5.0 percent, respectively. Recommended operating policies
could inlcude the following:
1. No single 24-hour zero drift value from the last
adjustment should exceed 5 percent without
the immediate performance of troubleshoot
ing and correction procedures; and
2. Trouble shooting and correction procedures
should be performed if four consecutive checks
result in drifts of the same sign.
In addition to determining zero drift trends, other pos-
sible applications of control charts include the visual depic-
tion of the mean, the range, or the standard deviation of the
following measurement values over several sets of tests;
• accuracy checks
• calibration gas checks
2-71
-------
4J
§ + 5.0%
4-1
W
•<->" + 3.3%
JJ
to
us
4J
>W
-H
4J
d
a)
u
- 3.3%
CHECK NO. j
ACTION LIMIT
•UCL
WARNING LIMIT
ACTION LIMIT
•LCL
10
DATE /TIME
OPERATOR
PROBLEM AND
CORRECTIVE
ACTION
Figure 5.11.9 Sample control chart for use in
determining trends in zero drift.
2-72
-------
• zero/calibration drift (2-hour)
• zero/calibration drift (24-hour)
5.11.3.7.2 Use of operation and maintenance tables. The
use of operation and maintenance tables is recommended to
increase the operator's efficiency, to provide a visual dis-
play of the requirements of the system, and to provide a
simplistic means of updating and adding to the routine pro-
cedures as the operator gains experience with the system in
question. Figure 5.11.10 gives an example of such a table.
The operation and maintenance table should be formu-
lated by the operator who will have the major responsibility
for the system. The basis for the table should be: the
maintenance manuals supplied by the various component manu-
facturers, interaction with the factory representatives at
the installation of the system and, most important, personal
experience gained during the performance testing and any
time thereafter. If possible, the operator should visit the
manufacturing facility prior to the shipment of the system
and witness the bench testing. Valuable information can be
acquiesced at this time which may otherwise be acquired only
after several weeks of operation and hours of troubleshooting.
After the original operation and maintenance table is
formulated, its usefulness can be extended by a continuous
update of the information contained in the table. For
instance, upon survey of a month's data, the operator may
determine that either more or less frequent calibration is
eminent for greater data validity. Or, another case might
be extreme drift in the calibration values which would indi-
cate that restandardization of the calibration gas is
necessary.
2-73
-------
Q
^7-
<
O
Q)
0
C
e
o
0)
a,
c
in
(1 10
O JJ
C
A; -H
U 'H
fl)
x!
U
^
g
01
-P
in
^i
in
G
O
•H
-U
(3
S-j
0)
tn
•rH
SH
in
0)
S-i
A;
u
0)
•-•
JO
^
w
QJ
£>
rH
ni
r*
d)
rH
rrj
a)
a;
c
*x
u
CJ
^1
u
^*x
g
0)
4-1
M
£"1
in
tn
G
•rH
4-1
fd
(1)
r;
G
O
•H
4-1
d
V-4
XI
•H
rH
IT)
0
0)
N
^
rH
IS
C
<
-------
5.11.4 Auditing Procedures
The proper planning and implementation of an auditing
program allows the enforcement agency to estimate data
quality in terms of precision and accuracy at a given level
of confidence. To realize maximum benefits from an auditing
program, the auditing procedures should be conducted inde-
pendently of the routine operation of the sampling network.
That* is, the audit checks should be made by individuals other
than the regular operator and using calibration standards
supplied and tested by the auditing agency. Furthermore, the
audit checks should be performed without any special prepar-
ation or adjustment of the system.
In conjunction with the special checks required during
performance testing and assuming that all routine calibration
and maintenance checks are performed judiciously, three
auditing checks will be sufficient to properly assess data
quality.
1. A visual check of the sampling interface,
analyzer, and data recording system.
2. The extraction and analysis of at least four
samples using the procedures outlined in
Method 6 or 8 (refs. 1 and 2) and a comparison of
this data with the corresponding data for the
same sampling period as recorded by the con-
tinuous analyzer system.
3. A data processing check to evaluate data
reduction and computation errors.
An auditing level of three checks (n=3) per performance
test period and one check (n=l) every 6 months during non-
compliance monitoring is used here for illustration purposes.
This assumes that sampling is carried out continuously for
non-conformance monitoring purposes and that the 2-week
2-75
-------
performance test period and compliance monitoring are
synonymous. This would result in a minimum auditing level
of three checks (n=3) for a lot size of 12 (N=12) for data
reported quarterly. The enforcement agency will specify the
actual auditing level required to meet specific monitoring
needs.
Direction for performing each of the checks are given
here.
5.11.4.1 Visual Checks—The auditor should prepare a
standard check sheet for each individual system which lists
all of the major components of the system and the parameters
which can and should be visually assessed. Such parameters
would include:
• Proper operation of the data recording
system.
• Proper location and position of the
sampling probe.
• Any major changes in the sample
interface system.
• Cleanliness of system components.
• Undue deterioration of system
components,.
The list would, of course, be more specific in the items
listed but these are general guidelines for what should be
included. Any changes made in the system or information
resulting from an increased familiarity with the system
should be reflected in revisions of the list.
5.11.4.2 Redundant Monitoring Using Appropriate Refer-
ence Method--The auditor or auditing crew, other than the
regular operator(s) of the system, must extract at least four
samples using the sampling techniques discussed in reference
1 for sampling periods of 1 hour each. The manual method
probe tip must be placed adjacent to the continuous system
probe tip. The samples should be analyzed by the auditor or
2-76
-------
an independent laboratory. The data are documented and sub-
mitted for assessment. Data from the continuous system which
was recorded simultaneously with the manual sampling should
be reduced- from the stripchart or magnetic tape recording
system.
Results from both the Method 8 sampling and the continu-
ous system measurements should then be logged on an Audit
Data Form, an example of which is shown in figure 5.11.11.
The calculations given on the Audit Data Form correspond to
those specified in section 5.11.1.6. If the accuracy calcu-
lated on the audit data form exceeds 20 percent, corrective
action should be taken by the regular operator(s) to deter-
mine the cause of the extreme inaccuracies and the auditing
schedule should be changed to allow for more frequent audits.
An alternate to the manual extraction and analysis of
samples is the measurement of calibration standards. These
calibration standards used in implementing an audit check
should be prepared independently of the normal operations
and analyzed by Method 6 within 2 weeks prior to use or
demonstrated to be accurate and stable in some equiv-
alent manner. The results of the analysis should be well
documented and available for use in the assessment of the
data from the continuous monitoring system.
The calibration standards should be inputted at the
sampling probe or as close to the sample probe as possible.
In the case of in-situ optical sensors, special calibration
cells can be tested prior to the audit and inserted in the
optical pathlength of the sensor as specified by the manu-
facturer. Allow the system to stabilize and record the
measurement.
The use of three different upscale values of calibration
standards is recommended for audit purposes. The following
2-77
-------
AUDIT DATA FORM
AUDITOR
REPORT DATA_
LABORATORY
Test
No.
Date
Time
Method 8
Samples (ppm)
Analyzer 1-hour
average (ppm)
di
where
d.
1
Difference = 1 hour average minus the
Method 8 average for the 1th extraction
S, = Sum of the differences
S- = Sum of the squares of the differences
S, = Sum of the Reference Method values
Mean Difference
ppm
95 Percent Confidence Interval
C.I.95 = .4593
Mean Reference Method Value
ppm
Accuracy
100
percent
Check if accuracy < 20%
n
Figure 5.11.11 Sample audit data>form.
2-7!
-------
control limits should be exercised in analyzing the audit
data: If the difference between any one continuous measure-
ment system value and the predetermined calibration standard
value is greater than 20 percent of the latter value or if
the average of the three differences exceed 10 percent, checks
should be made to determine the cause of the inaccuracies,
corrective action(s) taken, and the audit procedures repeated.
The actions which must be taken might include:
1. Perform a regular zero and span check with the
calibration standards used routinely.
2. If, after number 1 has been performed, the
audit value still proves inconsistent with
the known concentration value, a dynamic cali-
bration should be performed with the calibra-
tion standards used routinely.
3. Repair the system or perform maintenance, as
required, based on the calibration results.
4. Resume normal operation only after the results
of an audit check prove satisfactory.
5.11.4.3 Data Processing Check—Independent checks for
data processing errors should be incorporated into every
auditing program. The checks should be made by an individual
other than the operator responsible for the original data
reduction, calculations, and reporting. The procedures for
making the data processing checks are highly dependent upon
the sophistication of the data acquisition system. In this
section, two types of systems will be discussed and general
procedural guidelines set forth. They are stripchart
recording systems and magnetic tape recording systems.
5.11.4.3.1 Stripchart recording systems. The auditor
must check the following items if the continuous analyzer is
interfaced with a stripchart recording system:
2-79
-------
1. Accuracy with which the operator reads the
data from the stripchart and records the data
on a logsheet.
2. Equations used to transform the analog volt-
age data to meaningful parts per million
units.
3. Averaging process to obtain the required one-hour
averages.
4. Method of data validation using calibration data
as recorded on stripcharts and determinations
of system malfunctions.
Because most of the transfer equations which would be
required for a continuous SO,, monitoring system are relatively
simple, most errors will be the result of carelessness on the
part of the person reducing the data from the stripchart.
Systematic errors should be eliminated by careful training of
the operator as to the treatment of roundoffs and illegible
data.
The auditor should reduce data for a randomly chosen
1-week period. The check should be made starting with the
raw stripchart data and ending with the recording of the 1-hour
average SCU measurements in concentration units of parts per
million on a form similar to or the same as those forms used
routinely by system operator. A comparison of the auditor's
measurement values with the originally determined values should
show a difference of less than + 3 percent. If determined
necessary by the enforcement agency, all data since the last
audit may have to be checked and corrected if the difference
exceeds + 3 percent.
5.11.4.3.2 Magnetic tape recording system. If the data
acquisition system consists of a magnetic tape recorder, then
2-80
-------
the first parameter listed in section 5.11.4.3.1 is not applic-
able. It can generally be assumed that the error in reducing
data from magnetic tape by a computer is minimal. However,
the latter two parameters which have to do with transfer
questions and data validation are very much applicable. The
auditor must determine if the programming handles the raw data
such that the final output reflects the measured quantities.
Some of the more sophisticated systems will use the calibra-
tion data to generate new adjustment equations and will apply
these equations directly to the raw data. The auditor must
study the raw data (in voltages) and manually apply the appro-
priate equations to arrive at the average sulfur dioxide con-
centration over the recording interval in concentration units.
The auditor's data should not differ more than 1 percent from
the computerized data report. If the difference is greater
than 1 percent, the entire system must be analyzed and the dis-
crepancies found and rectified.
2-81
-------
3.0 FUNCTIONAL ANALYSIS
3-1
-------
3.0 FUNCTIONAL ANALYSIS
Performance Specification 2 defines the performance speci-
fications and specification test procedures for continuous
monitors of sulfur dioxide and oxides of nitrogen from sta-
tionary sources. As stated before, this quality assurance docu-
ment is treating only the determination of sulfur dioxide. The
performance specification states that SO2 monitoring systems
must meet certain requirements but does not specifically state
the components of the system. For this reason, the func-
tional analysis will be general and will exemplify a systems
approach which might be taken to formulate a functional analy-
sis of a particular system.
This test method has been subjected to no collaborative
tests and relatively few performance tests (refs. 13 through 17),
so quantitative components is not available. Engineering judg-
ments, based on considerable experience with both stack or air
quality sulfur dioxide analyzers and the associated sampling
and interface systems, were used in lieu of collaborative test
data in some of these areas of variable evaluation. The sub-
ject of error analysis is discussed in references 18 and 19.
Continuous systems for monitoring sulfur dioxide in
stack emissions can consist of several different component
combinations. Basically though, the systems can be divided
into two categories, in situ and extractive. Both types of
systems are shown graphically in figures 5.11.12 and 5.11.13,
respectively.
3-2
-------
3.1 DEFINITION AND DISCUSSION OF ERROR SOURCES
The various phases of the continuous measurement will
each contain numerous sources of error. Some errors can be
eliminated by a conscientious system operating staff, while
others are inherent within the system and can only be con-
trolled. The paragraphs below discuss the components of error
for both types of systems. The standard deviation stated in
this section are estimates given for illustrative purposes
only, and have not been substantiated by sufficient field data
to be stated as fact.
3.1.1 In-Situ Monitoring Systems
Based on the fact that in-situ monitoring systems avoid
the need for extracting a gas sample from the stream, such
systems do not manipulate the sample of gas being measured.
This statement can lead to two conclusions: 1. in-situ
systems are simpler than extractive systems, and 2. the sources
of error will be fewer. Both of these conclusions are valid
but must be discussed further for one to fully understand all
implications.
First, the in-situ systems are generally simpler in that
they contain fewer major components and are usually based on
straightforward electro-optical principles. The sophistica-
tion of the system comes into play in the engineering, design,
and fabrication of a system which can withstand the adverse
environment after installation and which will, indeed, provide
a measurement which is indicative of the concentration of SOp
in the total emissions.
Second, the sources of error will be fewer with an in-situ
system. However, in defining the error components, the
analyst will determine that the error due to each of these
3-3
-------
ZERO
CALIB.
SOURCE
SPAN
CALIB.
SOURCE
STACK
GAS
STREAM
IN-SCTU
SO i
ANALYZER
>to
DATA
RECORDING
SYSTEM
•~
REPORT
Figure 5.11.12
Block diagram of a typical in-situ
monitoring system.
ESP AN
JALIB.
GAS
STACK
GAS
STREAM
SAMPLE
EXTRACTION
SYSTEM
SAMPLE
CONDITIONING
SYSTEM
Preferred System Design
Alternate System Design
REPORT
: ZERO
! CALIB.
GAS
SPAN I
CALIB. !
...GAS...J
SO 2
ANALYZER
DATA
RECORDING
SYSTEM
Figure 5.11.13
Block diagram of a typical extractive
monitoring system.
3-4
-------
components is greater than with extractive systems, and that
the overall system error for both types of systems will
usually prove to be comparable.
Error sources for in-situ monitoring systems include:
1. Measurement method errors, a ;
2. Calibration inadequacies and the associated
drift problems, a ;
\-f
3. Data recording errors, a , and
4. Data processing errors, a .
These major components of the total system error are illus-
trated with a block diagram in figure 5.11.14 and are dis-
cussed below.
3.1.1.1 Measurement Method Errors, a --Assuming that
most in-situ monitors are of the electro-optical variety, the
following list of error sources is generally applicable.
1. Accumulation of particulate matter on windows.
A primary problem with in-situ monitors is the
accumulation of particulate matter on the windows
of the analyzer case. The particulate matter
will block the transmission of light through the
windows resulting in a negatively biased response
by the photodetector. Two different methods are
now used by commercially avilable instruments in
an attempt to overcome the problem. The use of
dedicated fans that are directed across the face
of the windows to avoid the accumulation of
particulates is one method. This method has
proven partially inaffective due to the unreli-
ability and inefficiency of the fanning system
causing a negative error in the final measure-
ment. The second method is the use of a detection
module which separates the beam into two sets of
3-5
-------
2
H CO
EH oil O C
< M 04 O
Q CJ1 P^
q w
a!
c
•H
w
n
o
n
M
Q)
e
i
-------
component wavelengths (ref. 20). One range of
wavelengths is absorbed solely by S02 and the
second range is not absorbed by any of the con-
stituents of the stack gas. The result of this
system is to totally discount the effect of the
particulate blocking of the transmitted and
reflected light. Errors due to this second
method would be realized if the chosen wave-
lengths were indeed absorbed to some degree by
a stack gas constituent and would produce a
negative bias in the response of the photo-
detector.
Misalignment of the optical components. The
misalignment of light sources, mirrors, diffrac-
tion gratings, and photodetectors is another major
source of error in in-situ systems. Design
objectives for any in-situ system must include
the fabrication of an instrument with as few
moving parts as possible and that is insensitive
to the constant vibrations and drastic tempera-
ture differences to which the instrument will be
subjected when installed on a stack. A mis-
alignment of the optical components will cause
a loss of light intensity which will in turn
result in a negative bias in the system measure-
ments .
Detector insensitivity. The detector must contain
filters so as to block detection of wavelengths
other than the wavelengths of interest. In the
event of faulty filters or a secondary problem,
loss of detector sensitivity, the output readings
from the system will be biased. Assuming that the
3-7
-------
deteriorated filters will allow other wave-
lengths of light to pass through the analyzer
secton and be included with the proper wave-
lengths in determining an intensity propor-
tional to the concentration of sulfur dioxide,
the existence of deteriorated filters will
cause a positive bias in the output signal.
In the event of a loss of sensitivity of the
detector to the intensity of the light passing
through the analyzer, the output signal will
be negatively biased.
4. Interferences by other constituents of the
stack gas. Water vapor and other constituents
of the stack gas can cause interference in the
electro-optical measurement of SG>2 with an in-
situ monitor. Careful selection of the absc-rp-
tion peak chosen for a particular application
can minimize interferences.
5. Change in the absorption characteristics of
the sulfur dioxide component of the stack gas.
According to one study, references 7 and 15,
temperature variations of the gas sample in the
slot can broaden the spectrum seen by the spec-
trometer. Reported in the study was that a 6°C
change in stack gas temperature alters the output
signal by 3 percent. Such a radical temperature
effect will most likely occur only at very high
stack gas temperatures and, if the stack gas tem-
perature range is known, appropriate correction
factors can be applied to the data.
The standard deviation of the measured SG>2 concentration
as a result of error due to each of the above-mentioned
3-i
-------
sources cannot be assigned in this document because of the
lack of sufficient field test data to support individual vari-
able analysis. However, a standard deviation of the measured
SOp concentration resulting from all measurement method errors
is estimated to be 10 percent of the true concentration
in parts per million. This estimate is given for illustrative
purposes only. The standard deviation for a specific system
must be determined from performance test data and field
operation data for that system.
3.1.1.2 Calibration Inadequacies and the Associated
Drift Problems, (^--Calibration techniques on an in-situ
O
monitor are generally complicated if performed manually and
are less reliable if performed automatically. Supporting this
statement are the following problems existing in the calibra-
tion of in-situ monitors:
1. The manufacture and testing of calibration cells
for use in calibrating in-situ monitors is more
complex than the use of cylinder gas in cali-
brating extractive monitors. The pathlength
from the cell to the photodetector and the
amount of SO- in the cell are both variables in
determining the span gas concentration as
measured by the analyzer.
2. The pathlength between the photodetector and the
light reflected back through the sample stack
gas must generally be blocked during calibration.
This procedure requires at least one moving part
which must be able to function properly regard-
less of ambient conditions, i.e., vibration,
extreme temperatures, etc.
3. If the reference gas cell is calibrated at a given
stack gas temperature and is used at another tem-
perature without correction, the data will be
erroneous.
3-9
-------
4. If the monitor must be calibrated manually, a
qualified^ technician must climb the stack to the
location of the monitor to perform the calibration,
"Qualified" technician infers that the operator
must be trained in the routine calibration and
maintenance of the system so that he is aware of
any mechanical or electrical problems should they
occur (ref. 21).
5. Zero and span drift due to the analyzer's sensi-
tivity to pressure and temperature changes between
manual or automatic checks will alter the data
accuracy. In most cases the data cannot be accu-
rately adjusted based on subsequent calibration
data.
For illustrative purposes only an estimated standard
deviation, a , of 8 percent is assumed for the true SO-
C_- ^-*
concentration as a result of error due to calibration
inefficiencies.
3.1.1.3 Data Recording Errors, o --In continuous moni-
toring, data recording errors can result from inadequate
response of the mechanical stripchart recording to varying
sulfur dioxide concentration. If magnetic tape recording
systems are used, errors in the recorded signal can result
from improper amplification of the analyzer output or mal-
functions in the recorder. Because of the advanced state-
of-art in data acquisition systems, a standard deviation,
a , of 2 percent is estimated (for illustrative purposes
only) for the effects of data recording errors on the mea-
surement of sulfur dioxide by an in-situ monitor.
3.1.1.4 Data Processing Errors, a —Data processing
errors can be avoided through extreme conscientiousness of
3-10
-------
the data processing personnel. Errors may occur, though, in
making accurate estimates of average values-on illegible
stripchart recordings. Other error sources are:
1. Misreading of calibration data from strip-
chart recording or raw magnetic tape recording
data. This can take the form of data being read
as span data when the system has -not actually
stabilized at a span reading.
2. Carelessness in reading discrete measurement
values with the utmost accuracy.
3. Mistakes such as manually reading the correct
data and transposing the numerical figures
upon recording the data.
4. Calculation errors in determining average con-
centrations over an extended sampling period.
5. Time shifting of magnetic tape data resulting
from improper input of start-time data into the
computer or from loss of data without battery
carryover for recorder time synchronization with
real time (the latter is also applicable to strip-
chart records).
An estimate of the standard deviation of the measurement
value as a result of data processing errors, a , is 3 per-
cent (for illustrative purposes only).
3.1.1.5 Total Standard Deviation, a—All of the error
terms discussed above can be considered independent for
analysis purposes, hence the total bias in the continuous
measurement of sulfur dioxide by in-situ monitors is the
algebraic sum of the biases of the individual terms. The
2
variance of the data, a , is the sum of the variances of
the individual error terms as follows:
3-11
-------
2 2 ^ 2 ^ 2 ^ 2
aT = °m + CTc + ar + ap
Using the estimated example values given in the previous para-
2
graphs, a = .0177 or a = 13.3 percent of the true concen-
tration of sulfur dioxide. An exact value of o for a particu-
lar system can only be determined after a sufficient quantity
of field data has been taken and analyzed.
3.1.2 Extractive Monitoring Systems
The variable analysis of extractive monitoring systems
can only be accomplished with accuracy after a comprehen-
sive definition of the system components has been made. As
depicted in figure 5.11.15, the major components of system
error can be divided into five categories:
1. Sampling interface errors, a, ;
e
2. Method measurement errors, a ;
m
e
3. Reference sample errors, a ;
e
4. Data recording errors, a, ; and
e
5. Data processing errors, a
pe
Each of the above categories are discussed separately in the
paragraphs below.
3.1.2.1 Sampling Interface Errors, af —The sampling
•"-e
interface consists of the components of a measurement system
which extract and in some way modify the gas sample to be
analyzed as a representative sample of the total stack gas.
In accomplishing the objective of the interface system,
which is to supply a sample with characteristics compatible
with the analyzer in use, the interface system may perform
any of the following operations in addition to the extrac-
tion: transportation, filtration, or conditioning of the
sample in terms of temperature or water content reduction.
3-12
-------
DATA
PROCESSING
ERRORS
DATA
RECORDING
ERRORS
0)
'd
t>
w
K
D
co
i >i
en w
m en
O d
•H
tn M
-P O
C 4J
0) -H
C C
0 O
O 0)
u >
•H
H -P
O O
•t— i (0
in
0)
Q)
14-1
IT)
Q)
co
g
•H
CM
3-13
-------
Obvious sources of error in most types of interface
system include:
1. Loss of sample due to leaks or adsorption in
the sampling probe or sample lines,
2. Malfunction of control valves,
3. Loss of sample in water vapor reduction
'technique,
>4. Incorrect sample flow due to plugging by
particulates which can occur through the use of
ineffective filtering, failure to change filters
at reasonable intervals, or incorrect position-
ing of the probe filter.
5. Inaccurately assumed dilution ratio if dilution
system is used,
6. Improper operation of condensation or water
vapor removal system which results in higher
actual concentration of water vapor in sample
entering analyzer than is assumed,
7. Improper maintenance of sample gas temperature
compatible with the analyzer, and
8. Chemical reaction with interface construction
materials.
Most of the above sources of error can be eliminated with
a comprehensive operation and maintenance program designed to
meet the requirements of the specific system in use (ref. 10).
An estimated standard deviation (for illustrative purposes
only) of the final measurement due to interface system errors,
af , is 10 percent of the actual sulfur dioxide concentration.
e
3.1.2.2 Measurement Method Errors, am —Just as with the
e
in-situ monitoring systems discussed in section 3.1.1.1, each
3-14
-------
of the analyzing techniques used in extractive monitoring sys-
tems have certain inherent error sources. Because of the wide
variety of techniques this document will list only a few of
these error sources which may or may not be applicable to all
analyzers. The common error sources are:
1. Interferences by other constituents of the stack
gas. Water vapor, nitric oxide, sulfuric acid,
and hydrogen sulfides are some constituents of
the stack gas which may interfere with the
measurement of sulfur dioxide thereby reducing
the specificity of the measurement method.
Although this error source may be considered
dependent upon the efficiency of the condi-
tioning system with respect to the interfering
species, it is assumed that the inefficiency
of the conditioning system below specified
limits contributes to a,, and that those same
Ci
specified limits are used in determining
errors due to the sensitivity of the analyzer
to the minimum amount of the interfering species
which comes under a . This assumption allows
both error sources to be considered independent
for purposes of variable analysis in this document.
2. Plugging of the internal sample lines. The
flow rate through the analyzer can be hampered
by undue plugging of the sample lines internal
to the analyzer resulting from accumulation of
particulate matter and water vapor beyond the
minimum efficiency specifications of the inter-
face components.
3. Improper operation of internal sample dryers.
The malfunction of internal sample dryers not
3-15
-------
compensated for with external components can
cause interference above the amount adjusted
for by the analyzer.
4. Initial setup and calibration of system. Dis-
cussions on the influence of setting up and
calibrating a system at a large full-scale
value relative to existing emission levels have
shown that the analyzer accuracy can be greatly
reduced (ref. 17).
5. Zero and calibration drift. Zero and span drift
due to the analyzer's sensitivity to pressure
and temperature changes between manual or auto-
matic checks will alter the data accuracy. In
most cases the drift is random and the data can-
not be accurately adjusted based on subsequent
calibration data.
An estimated (for illustrative purposes only) standard
deviation of 8 percent of the true measurement value is used
here for illustrative purposes as representative of the error
due to the analysis technique.
3.1.2.3 Reference Sample Error, a —The analysis of
sulfur dioxide by extractive continuous monitors is subject
to error from inaccuracies in the calibration gases used
for either manual or automatic calibration and in the calibra-
tion technique. Most extractive measurement systems are
calibrated with cylinder gas. Error in the assumed concen-
tration value of the reference gas will result in a bias of
the measurements for the lifetime of the cylinder. Error in
the calibration technique is a function of the similarity
between the pathway of the sample stack gas through the
system and the pathway of the calibration gas. Ideally,
3-16
-------
the calibration gas should be input at the sample probe and,
upon entering the analyzer, would go through the same inter-
nal sample lines to the final measurement cell.
The standard deviation of the true measurement value due
to calibration error, a , is estimated at 8 percent of the
ce
true measurement value.
3.1.2.4 Data Recording Errors, a --As is the case with
in-situ monitoring, data recording errors can result from
inadequate response of the mechanical stripchart recorder to
varying sulfur dioxide concentration. If magnetic tape
recording systems are used, error in the recorded signal can
result from improper amplification of the analyzer output
or malfunctions in the recorder. Because of the advanced
state-of-art in data acquisition systems, a standard
deviation, a , of 2 percent is estimated for the
re
effects of data recording errors on the measurement of
sulfur dioxide by an extractive monitor.
3.1.2.5 Data Processing Errors, a —As with in-situ
Pe
monitoring systems, data processing errors can be avoided
through extreme conscientiousness of the data processing
personnel. Errors may occur, though, in making accurate
estimates of average values on illegible stripchart record-
ings. Other error sources are:
1. Misreading of calibration data from strip-
chart recording or raw magnetic tape
recording data. This can take the form of
data being read as span data when the system
has not actually stabilized at a span
reading.
2. Carelessness in reading discrete measurement
values with the utmost accuracy.
3-17
-------
3. Mistakes such as manually reading the correct
data and transposing the numerical figures
upon recording the data.
4. Calculation errors in determining average
concentrations over an extended sampling
period.
5. Time shifting of magnetic tape data resulting
from improper input of start time data into
the computer or from loss of data without
battery carryover for recorder time synchroni-
zation with real time (the latter is also
applicable to stripchart records).
An estimate of the standard deviation of the measurement
value as a result of data processing errors, a , is 3 per-
Pe
cent.
3.1.2.6 Total Standard Deviation, qm —All of the error
ie
terms discussed above are independent, hence the total bias
in the continuous measurement of sulfur dioxide by extractive
monitors is the algebraic sum of the individual terms. The
variance of the data, a^ ^, is the sum of the variances of
le
the individual error terms as follows:
Using the estimated values given in the previous paragraphs,
a 2 = .0183 and a =13.5 percent of the true concentra-
ie -"-e
tion of sulfur dioxide. This value is comparable to results
from evaluations of several monitoring systems which range
between 10 and 15 percent (see refs. 13 through 17).
3-18
-------
LIST OF REFERENCES
R-l
-------
LIST OF REFERENCES
1. J. W. Buchanan and D. E. Wagoner, Guidelines for Develop-
ment of a Quality Assurance Program, "Determination of
Sulfur Dioxide Emissions from Stationary Sources," EPA-
650/4-74-005-e, Office of Research and Development, U.S.
Environmental Protection Agency, Washington, D. C.;
November 1975.
2. J. W. Buchanan and D. E. Wagoner, Guidelines for Develop-
ment of a Quality Assurance Program, "Determination of
Sulfuric Acid Mist and Sulfur Dioxide Emissions from Sta-
tionary Sources," EPA Contract Number 68-02-1234, Program
Element Number 1HA327, Office of Research and Develop-
ment, U.S. Environmental Protection Agency, Washington,
D. C.; November 1975.
3. Franklin Smith and Denny E. Wagoner, Guidelines for
Development of a Quality Assurance Program, "Determina-
tion of Moisture in Stack Gases," EPA Contract Number
68-02-1234, Program Element Number 1HA327, Quality Assur-
ance and Environmental Monitoring Laboratory, National
Environmental Research Center, Research Triangle Park,
North Carolina, August 1974.
4. E. F. Brooks, C. A. Flegal, L. N. Harnett, M. A. Kolpin,
D. J. Luciani, and R. L. Williams, "Continuous Measure-
ment of Gas Composition from Stationary Sources," EPA-
600/2-75-012, U.S. Environmental Protection Agency,
Office of Research and Development, Washington, D. C.;
July 1975.
5. Dr. Harry C. Lord, "Verification of In-Situ Source Emis-
sion Analyzer Delta, " paper presented at ASTM/NBS/EPA
Symposium, Calibration in Air Monitoring, Boulder,
Colorado, August 1975.
6. Dr. Harry C. Lord, Dale W. Egan, Paul E. Paules, Dr.
Geoffrey B. Holstrom, "Instantaneous, Continuous,
Directly On-Stream Boiler Flue Gas Analysis," paper
presented at Instrument Society of America, 24th Annual
Power Industry Symposium, New York City, May 1971.
7. Robert N. Hager, Jr., "Derivative Spectroscopy with
Emphasis on Trace Gas Analysis," Analytical Chemistry,
Volume 45, Number 13, pages 1131A-1138A, November 1973.
R-2
-------
8. James B. Homolya, "Coupling Continuous Gas Monitors to
Emission Sources," Chemical Technology, Volume 4,
Number 7, July 1974.
9. William Zolner, Edward Cieplinski, Denis Helm, "Source
Level SC>2 Analysis Via Pulsed Fluorescence," Thermo
Electron Corporation, Waltham, Massachussets 02154.
10. J. A. Jahnke, J. L. Cheney, and J. B. Homolya, "Inter-
ferences in SC>2 Fluorescence Monitoring Instruments,"
Environmental Protection Agency, Environmental Research
Center, Research Triangle Park, North Carolina; 1976.
11. Robert K. Stevens, Lewis F. Ballard, and Clifford E.
Decker, "Field Evaluation of Sulfur Dioxide Monitoring
Instruments," U.S. Environmental Protection Agency,
Air Pollution Control Office, Raleigh, North Carolina.
12. K. J. McNulty, J. F. McCoy, J. H. Becker, J. R.
Ehrenfeld, and R. L. Goldsmith, "Investigation of
Extractive Sampling Interface Parameters," EPA-750/2-
74-089, Environmental Protection Technology Series,
Environmental Protection Agency, Research Triangle
Park, North Carolina, October 1974.
13. James Buchanan, "A Quality Assurance Program for the
Environmental Protection Agency Wet Limestone Scrubber
Demonstration Project, Shawnee Steam-Electric Plant,
Paducah, Kentucky," Final Report for Contract Number
60-02-1398, Task Order No. 20, Industrial Environmental
Research Laboratory, Environmental Protection Agency,
Research Triangle Park, North Carolina; December 1975.
14. Fredric C. Jay, "Monitoring Instrumentation for the
Measurement of Sulfur Dioxides in Stationary Source
Emissions," EPA Study Number EPA-R2-73-163, Contract
Number EHSO 71-23, Environmental Protection Agency,
Research Triangle Park, North Carolina; February 1973.
15. Robert K. Stevens and William F. Herget, "Analytical
Methods Applied to Air Pollution Measurements,"
Chemistry and Physics Laboratory, National Environmental
Research Center, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina; Ann Arbor
Science Publishers, Inc., P.O. Box 1425, Ann Arbor,
Michigan; 1974.
R-3
-------
16. James B. Homolya, "Data Output for Monitoring SC>2
Emissions from a Stationary Source," International
Conference and Exhibit at Houston, Texas, Instrument
Society of America Reprint, Pittsburgh, Pennsylvania
15222.
17. W. L. Bonam and W. J. Fuller, "Certification Experience
Extractive Monitoring Systems," paper presented at the
ASTM Symposium on Calibration Source Monitors, Boulder,
Colorado; August 1975.
18. Philip R. Bevingtcn, Data Reduction and Error Analysis
for the Physical Sciences, McGraw-Hill, New York; 1969.
191 D. C. Baird, Experimentation; An Introduction to
Measurement Theory and Experiment Design, Prentice-Hall,
New Jersey; 1962.
20. Robert J. Bambeck and Hermann F. Huettemeyer, "Operating
Experience with In-Situ Power Plant Stack Monitors,"
paper presented at the 67th Annual Meeting of the Air
Pollution Control Association, Denver, Colorado; June
1974.
21. F. Dale Huillet, "The Monitoring of SO2, NO, CO, and
Opacity with an In-Stack Dispersive Spectrometer," Scott
Paper Company, Everett, Washington.
R-4
-------
APPENDIX A
A-l
-------
46268
RULES AND REGULATIONS
Performance Specification 3—Performance
specifications and specification test proce-
dures for monitors of CO2 and O2 from sta-
tionary sources.
1. Principle and Applicability.
1.1 Principle. Effluent gases are continu-
ously sampled and are analyzed for carbon
dioxide or oxygen by a continuous monitor-
ing system. Tests of the system are performed
during a minimum operating period to deter-
mine zero drift, calibration drift, and re-
sponse time characteristics.
1.2 Applicability, This performance speci-
fication is applicable to evaluation of con-
tinuous monitoring systems for measurement
of carbon dioxide or oxygen. These specifica-
tions contain test procedures, installation re-
quirements, and data computation proce-
dures for evaluating the acceptability of the
continuous monitoring systems subject to
approval by the Administrator. Sampling
may include either extractive or non-extrac-
tive (in-situ) procedures.
2. Apparatus.
2.1 Continuous Monitoring System for
Carbon Dioxide or Oxygen.
2.2 Calibration Gas Mixtures. Mixture of
known concentrations of carbon dioxide or
oxygen in nitrogen or air. Midrange and 90
percent of span carbon dioxide or oxygen
concentrations are required. The 90 percent
of span gas mixture is to be used to set and
check the analyzer span and is referred to
as span gas. For oxygen analyzers, if the
span is higher than 21 percent O2, ambient
air may be used in place of the 90 percent of
span calibration gas mixture. Triplicate
analyses of the gas mixture (except ambient
a)r) shall be performed within two weeks
prior to use using Reference Method 3 of
this part.
2.3 Zero Gas. A gas containing less than 100
ppm of carbon dioxide or oxygen.
2.4 Data Recorder. Analog chart recorder
or other suitable device with input voltage
range compatible with analyzer system out-
put. The resolution of the recorder's data
output shall be sufficient to allow completion
of the test procedures within this specifica-
tion.
3. Definitions.
3.1 Continuous Monitoring System. The
total equipment required for the determina-
tion of carbon dioxide or oxygen in a given
source effluent. The system consists of three
major subsystems:
3 : 1 Sampling Interface That portion of
the continuous monitoring system that per-
forms one or more of the following opera-
tions: delineation, acquisition, transporta-
tion, and conditioning of a sample of the
source effluent or protection of the analyzer
from the hostile aspects of the sample or
source environment.
312 Analyzer That portion of the con-
tinuous monitoring system which senses the
pollutant gas and generates a signal output
that is a function of the pollutant concen-
tration.
3.1.3 Data Recorder. That portion of the
continuous monitoring system that provides
a permanent record of the output signal in
terms of concentration units.
3.2 Span The value of oxygen or carbon di-
oxidt! concentration at which the continuous
monitoring system is set that produces the
maximum data display output. For the pur-
poses of this method, the span shall be set
no less than 1.5 to 2.5 times the normal car-
bon dioxide or normal oxygen concentration
in tha stack gas of the affected facility.
3.3 Midrange. The value of oxygen or car-
bon dioxide concentration that as representa-
tive yt the normal conditions in the stack
gas of the affected facility at typical operat-
ing rates.
3.4 Zero Drift. The change in the contin-
uous monitoring system output over a stated
periol of time of normal continuous opera-
tion .vhen the carbon dioxide or oxygen con-
centration at the time for the measurements
is zero.
3.5 Calibration Drift. The change in the
conti luous monitoring system output over a
stated time period of normal continuous op-
eration when the carbon dioxide or oxygen
continuous monitoring system is measuring
the concentration of span gas.
3 6 Operational Test Period. A minimum
period of time over which the continuous
monitoring system is expected to operate
withii certain performance specifications
without unscheduled maintenance, repair, or
adjustment.
3 7 Response time. The time interval from
a step change in concentration at the input
to th« continuous monitoring system to the
time at which 95 percent of the correspond-
ing final value Is displayed on the continuous
monitoring system data recorder.
4. Installation Specification.
Oxygen or carbon dioxide continuous mon-
itoring systems shall be installed at a loca-
tion where measurements are directly repre-
sentative of the total effluent from the
affected facility or representative of the same
effluent sampled by a SO2 or NO, continuous
monitoring system. This requirement shall
be complied with by use of applicable re-
quirements in Performance Specification 2 of
this appendix as follows:
4.1 Installation of Oxygen or Carbon Di-
oxide Continuous Monitoring Systems Not
Used to Convert Pollutant Data. A sampling
location shall be selected in accordance with
the procedures under paragraphs 4.2.1 or
4.2.2, or Performance Specification 2 ol this
appendix.
4.2 Installation of Oxygen or Carbon Di-
oxide Continuous Monitoring Systems Used
to Convert Pollutant Continuous Monitoring
System Data to Units of Applicable Stand-
ards. The diluent continuous monitoring sys-
tem (oxygen or carbon dioxide) shall be in-
stalled at a sampling location where measure-
ments that can be made are representative of
the effluent gases sampled by the pollutant
continuous monitoring system(s). Conform-
ance with this requirement may be accom-
plished in any of the following ways:
421 The sampling location for the diluent
system shall be near the sampling location for
the pollutant continuous monitoring system
such that the same approximate polnt(s)
(extractive systems) or path (in-situ sys-
tems) in the cross section is sampled or
viewed.
4.2.2 The diluent and pollutant continuous
monitoring systems may be installed at dif-
ferent locations if the effluent gases at both
sampling locations are nonstratified as deter-
mined under paragraphs 4.1 or 4.3, Perform-
ance Specification 2 of this appendix and
there is no in-leakage occurring between the
two sampling locations. If the effluent gases
are stratified at either location, the proce-
dures under paragraph 4.2.2, Performance
Specification 2 of this appendix shall be used
for installing continuous monitoring systems
at that location.
5. Continuous Monitoring System Perform-
ance Specifications.
The continuous monitoring system shall
meet the performance specifications In Table
3—1 to be considered acceptable under this
method.
6. Performance Specification Test Proce-
dures.
The following test procedures shall be used
to determine conformance with the require-
ments of paragraph 4. Due to the wide varia-
tion existing in analyzer designs and princi-
ples of operation, these procedures are not
applicable to all analyzers. Where this occurs,
alternative procedures, subject to the ap-
proval of the Administrator, may be em-
ployed. Any such alternative procedures must
fulfill the same purposes (verify response,
drift, and accuracy) as the following proce-
dures, and must clearly demonstrate con-
formance with specifications In Table 3-1.
6.1 Calibration Check. Establish a cali-
bration curve for the continuous moni-
toring system using zero, midrange, and
span concentration gas mixtures. Verify
that the resultant curve of analyzer read-
ing compared with the calibration gas
value is consistent with the expected re-
sponse curve as described by the analyzer
manufacturer. If the expected response
curve is not produced, additional cali-
bration gas measurements shall be made,
or additional steps undertaken to verify
FEDERAL REGISTER VOL. 40, NO. 194—MONDAY, OCTOBER 6, 1975
A-2
-------
RULES AND REGULATIONS
46269
the accuracy of the response curve of the
analyzer.
6.2 Field Test for Zero Drift and Cali-
bration Drift. Install and operate the
continuous monitoring system in accord-
ance with the manufacturer's written in-
structions and drawings as follows:
TABLE 3-1.—Performance specifications
Parameter
Specification
1. Zero drift (2 h) i <0.4 pet Ch or CO;
2. Zero drift (24 h) 1 <05pct Chor COj.
3. Calibration drift (2 h) ' _. <0 4 pet 02 or CCh.
4 Calibration drift (24 h) 1. <0.5 pet 02 or CO;
5. Operational period 168 h minimum
6. Response time - 10 nun
i Expressed as sum of absolute mean value plus 95 pet
confidence interval of a series of tests.
6 2.1 Conditioning Period. Offset the zero
setting at least 10 percent of span so that
negative zero drift may be quantified Oper-
ate the continuous monitoring system for
an initial 168-hour conditioning period in a
normal operational manner
6.2.2. Operational Test Period. Operate the
continuous monitoring system for an addi-
tional 168-hour period maintaining the zero
offset. The system shall monitor the source
effluent at all times except when being
zeroed, calibrated, or taackpurged.
6.2.3 Field Test for Zero Drift and Calibra-
tion Drift. Determine the values given by
zero and midrange gas concentrations at two-
hour intervals until 15 sets of data are ob-
tained. For non-extractive continuous moni-
toring systems, determine the zero value
given by a mechanically produced zero con-
dition or by computing the zero value from
upscale measurements using calibrated gas
cells certified by the manufacturer. The mid-
range checks shall be performed by using
certified calibration gas cells functionally
equivalent to less than 50 percent of span.
Record these readings on the example sheet
shown in Figvire 3-1. These two-hour periods
ne
-------
46270
RULES AND REGULATIONS
3ata Zero Span
Set/'" Time Zero Drift Span Drift
to. Begin End Dat
-------
RULES AND REGULATIONS
46271
Datt of Test
Span Gas Concentration
Analyzer Span Setting
1.
Upscale 2.
3.
_ppm
seconds
seconds
seconds
Average upscale response
Downscal;
1.
2.
3.
seconds
seconds
seconds
Average downscale response
seconds
seconds
seconds
System average response time (slower time) = _
, j.«v,<*t,cft' frora slower ... average upscale minus average downscale
"
system average response
slower time
Figure 3-3. Response
[PR Doc.75-26565 Filed lC-3-75;&45 am]
FEDERAL REGISTER, VOL. 40, NO. 194—MONDAY, OCTOBER 6, 1975
A-5
-------
APPENDIX B
B-l
-------
APPENDIX B
GLOSSARY OF SYMBOLS
This is a glossary of the symbols used in this document.
Symbols used and defined in the reference method (section
5.11.1) are not repeated here.
A Given wavelength of light.
d Difference between average measured concentration
value of calibration gas A and the continuous
measurement system reading during calibration error
test.
d Difference between average measured concentration
value of calibration gas B and the continuous
measurement system reading during calibration error
test.
SUM 1 Sum of the d^ values determined in a calibration
error test.
2
SUM 2 Sum of the sc[uares of the d^ values (d^ ) used in
the calibration error calculation.
SUM 3 Sum of the dB values determined in a calibration
error test.
2
SUM 4 Sum of the squares of the dg values (dB ) used in
the calibration error calculation.
d. Difference between the reference method value (ppm)
1 and the average analyzer measurement for the ±tn
interval in sin accuracy test.
d.2 Square of d^ value for ith interval.
SUM 5 Sum of the d^ values for nine accuracy test
measurements.
2
SUM 6 Sum of the sc[uares of the di values (di ) used in
a calculation of continuous system accuracy.
SUM 7 Sum of the reference method measurements determined
in an accuracy test and used in the calculation of
system accura.cy.
AZ Difference in zero reading from zero reading taken
two hours before.
2
AZ Square of AZ value.
B-2
-------
C Calibration drift calculated using the 2-hour span
drift (ppm) minus the zero drift (ppm) for the
corresponding time period.
2
C Square of 2-hour calibration drift value.
SUM 8 Sum of the AZ values taken in a series of 2-hour
zero drift tests.
2
SUM 9 Sum of the squares of the AZ values (AZ ) taken in
a series of 2-hour zero drift tests.
SUM 10 Sum of the calibration drift (C) values calculated
from a series of calibration drift tests.
2
SUM 11 Sum of the squares of calibration drift values (C )
calculated after a series -of 2-hour calibration
drift tests.
AZ?. Difference in zero reading from zero reading taken
24 hours before.
AZ242 Square of the 24-hour zero drift value.
C~, Calibration drift calculated using the 24-hour span
drift (ppm) minus the zero drift (ppm) for the
corresponding time period.
C,-,.2 Square of the 24-hour calibration drift value.
SUM 12 Sum of the 24-hour zero drift values taken from a
series of 24-hour drift tests.
SUM 13 Sum of the squares of the 24-hour zero drift values
taken from a series of 24-hour zero drift tests.
SUM 14 Sum of the 24-hour calibration drift values.
SUM 15 Sum of the squares of the 24-hour calibration drift
values (C242)•
SUM 16 Sum of the upscale response time measurement values
(seconds).
SUM 17 Sum of the downscale response time measurement
values (seconds).
n Sample size for the auditing.
UCL Upper quality limit determined from performance test
data and used in assessment of routine calibration
data.
LCL Lower quality limit determined from performance test
data and used in assessment of routine calibration
data.
B-3
-------
d. Difference between continuous system 1-hour average
1 data minus the Method 8 value for the ith measure-
ment.
S, Sum of the differences (d^) calculated from a series
of Method 8 measurements compared with continuous
system data.
2
S~ Sum of the squares of the differences (dj_ )
S-, Sum of the reference method values calculated from
a series of measurements.
CR Mean reference method value calculated from audit
data.
a Standard deviation of in-situ continuous method
measurement due to measurement method errors.
a Standard deviation of continuous method measurement
due to calibration inadequacies and the associated
drift problems in in-situ monitoring systems.
a Standard deviation in in-situ measurement due to
r data recording errors.
a Standard deviation of in-situ measurement due to
" data processing errors.
a_ Standard deviation of the in-situ measurement which
is a function of the component errors.
a^ Standard deviation of extractive system measurement
e due to sampling interface errors.
a Standard deviation of extractive continuous method
me measurement due to measurement method errors.
a Standard deviation of continuous method measurement
°e due to calibration inadequacies and the associated
drift problems in extractive monitoring systems.
a Standard deviation of extractive measurement due to
re data recording errors.
a Standard deviation of extractive measurement due to
^e data processing errors.
a™ Standard deviation of the extractive measurement
e which is a function of the component errors.
B-4
-------
APPENDIX C
C-l
-------
APPENDIX C
GLOSSARY OF TERMS
The following glossary lists and defines the statistical
terms as used in this document.
Accuracy
Bias
Lot
Measurement method
Measurement process
Population
Precision
Quality audit
Quality control
check
Sample
A measure of the error of a process
expressed as a comparison between the
average of the measured values and the
true or accepted value. It is a func-
tion of precision and bias.
The systematic or nonrandom component
of measurement error.
A specified number of objects to be
treated as a group, e.g., the number
of field tests to be conducted by an
organization during a specified period
of time (usually a calendar quarter).
A set of procedures for making a
measurement.
The process of making a measurement,
including method, personnel, equipment,
and environmental conditions.
The totality of the set of items, units,
or measurements, real or conceptual,
that is under consideration.
The degree of variation among succes-
sive, independent measurements (e.g.,
on a homogeneous material) under con-
trolled conditions, and usually
expressed as a standard deviation or
as a coefficient of variation.
A management tool for independently
assessing data quality.
Checks made by the field crew on certain
Items of equipment and procedures to
assure data of good quality.
Objects drawn, usually at random, from
the lot for checking or auditing
purposes.
C-2
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO
EPA-650/4-74-005 o
3. RECIPIENT'S ACCESSIOWNO
4. TITLE ANDSUBTITLE
GUIDELINES FOR DEVELOPMENT OF A QUALITY ASSURANCE
PROGRAM: DETERMINATION OF SULFUR DIOXIDE EMISSIONS
FROM STATIONARY SOURCES BY CONTINUOUS MONITORS
5. REPORT DATE
Marrh 1Q76
6. PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
P. S. Wohlschlegel
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
Research Triangle Park, NC 27709
10 PROGRAM ELEMENT NO.
1HA327
11 CONTRACT/GRANT NO.
68-02-1234
12 SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Environmental Monitoring and Support Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D. C. 20460
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Guidelines for the quality control of the continuous measurement of sulfur
dioxide emissions by Performance Specification 2 are presented. These include:
1. Good operating practices;
2. Directions on how to assess performance and to qualify data;
3. Directions on how to identify trouble and to improve data qualityi
4. Directions to permit design of auditing activities.
The document is not a research report. It is designed for use by operating
personnel.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Quality Assurance
Quality Control
Air Pollution
Gas Analysis
13H
14D
13B
07D
14B
18 DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
20. SECURITY CLASS (This page)
Unclassified
21. NO. OF PAGES
123
e
EPA Form 2220-1 (9-73)
C-3
-------
m
D
C
O
33
H
CD
C
-o"
QJ —h —'
3^2
(Q O o
m" <° o>
-o
00
o
a "i s TJ
-H a .3
"D => I <_
a
.ox
t
O>
CJ1
N)
Nj
CJI
N
Q. 03 -»
CD ^ 3-
CD
W CD
Q. ~
Q. —
W CL
3 °
£ z
CD ,-,
c^H
^S
CO
^~
D
Q. CD
— O
O CD
CD CD
H cn
O
00
O
W H Tl
ui m m
O m
H en
CO
CO
CD
m
0
------- |