EPA-650/4-74-005-b
February 1974
Environmental Monitoring Series
-------
-------
EPA-650/4-74-005-b
GUIDELINES FOR DEVELOPMENT
OF A QUALITY ASSURANCE PROGRAM:
VOLUME II - GAS ANALYSIS
FOR CARBON DIOXIDE , EXCESS AIR,
AND DRY MOLECULAR WEIGHT
by
Franklin Smith and Denny E. Wagoner
Research Triangle Institute
Research Triangle Park, North Carolina 27709
Contract No. 68-02-1234
Program Element No. 1HA327
ROAP No. 26BGC
EPA Project Officer: Steven M. Bromberg
Quality Assurance and Environmental Monitoring Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
February 1974
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
Publication No. EPA-650/4-74-005-b
11
-------
TABLE OF CONTENTS
SECTION PAGE
I INTRODUCTION 1
ii OPERATIONS mm. i\
2.0 GENERAL 4
2.1 APPARATUS REQUIREMENTS 4
2.2 PRESAMPLING PREPARATION 10
2.3 ON-SITE MEASUREMENTS 13
2.4 POSTSAMPLING OPERATIONS 20
111 MANUAL FOR FIELD TEAf1 SUPERVISOR 23
3.0 GENERAL 23
3.1 ASSESSMENT OF DATA QUALITY (INTRATEAM) 24
3.2 SUGGESTED PERFORMANCE CRITERIA 28
3.3 COLLECTION AND ANALYSIS OF INFORMATION
TO IDENTIFY TROUBLE 30
IV MANUAL FOR MNAGER OF GROUPS OF FIELD TW1S 36
4.0 GENERAL 36
4.1 FUNCTIONAL ANALYSIS OF TEST METHOD 37
4.2 PROCEDURES FOR PERFORMING A QUALITY AUDIT 44
4.3 DATA QUALITY ASSESSMENT 47
APPENDIX A METHOD 3 (AS PRINTED IN THE FEDERAL REGISTER) 56
APPENDIX B GLOSSARY OF SVMBOLS 57
APPENDIX C GLOSSARY OF TERMS 59
iii
-------
LIST OF ILLUSTRATIONS
FIGURE NO. PAGE
1 Operational Flow Chart of the Measurement Process 5-6
2 Illustration of Key Components of an Orsat Analyzer
for Measuring C00, 09, and CO Contents of Stack Gases 9
iL £»
3 Sample Orsat Field Data Sheet 17
4 Sample Control Chart for the Range, R, of
Field Analyses 34
5 Sample Control Chart for Calibration Checks 34
6 Number of Replications r for Estimating %C09 to
/t
Within + 10 Percent with 99 Percent Confidence 42
7 Example Illustrating p < 0.10 and Satisfactory
Data Quality 52
8 Example Illustrating p > 0.10 and Unsatisfactory
Data Quality 52
IV
-------
LIST OF TABLES
TABLE NO. PAGE
1 Suggested Performance Criteria 29
2 Computation of Mean Difference, d, and
Standard Deviation of Differences, s, 51
3 Sample Plan Constants, k for P{not detecting a lot
with proportion p outside limits L and U} < 0.1 51
v
-------
ABSTRACT
Guidelines for the quality control of gas analysis for carbon dioxide,
excess air, and dry molecular weight by the Federal reference method are
presented. These include:
1. Good operating practices
2. Directions on how to assess performance and qualify data
3. Directions on how to identify trouble and improve 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 Contract Durham
68-02-1234 by Research Triangle Institute, Research Triangle Park, N.C.,
Under the sponsorship of the Environmental Protection Agency. Work was
completed as of June 1974.
VI
-------
SECTION I INTRODUCTION
This document presents guidelines for developing a quality assurance
program for Method 3Gas Analysis for Carbon Dioxide, Excess Air, and
Dry Molecular Weight. For convenience of reference, this method as
published by the Environmental Protection Agency in the Federal Register,
December 23, 1971, is reproduced as appendix A of this report.
The objectives of this quality assurance program for Method 3 are to:
(1) Minimize systematic and random variability in
the measurement process,
(2) Provide routine indications for operating purposes
of unsatisfactory performance of personnel and/or
equipment,
(3) Provide for prompt detection and correction of
conditions which contribute to the collection of
poor quality data, and
(4) 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) recommended operating procedures,
(2) routine training and evaluation of personnel and
evaluation of equipment,
(3) routine monitoring of the variables and parameters
which may have a significant effect on data quality,
(4) development of statements and evidence to qualify
data and detect defects, and
(5) action strategies to increase the level of precision/
accuracy in the reported data.
Component (2) above will be treated in the final report of this contract.
Component (5) will be treated in the Quality Assurance Documents (of this
series) for pollutant specific methods which utilize the results of
Method 2.
Implementation of a properly designed quality assurance program should
enable measurement teams to achieve and maintain an acceptable level of
precision in their stack gas composition measurements. It will also
allow a team to report an estimate of the precision of its measurements
for each source emissions test.
-------
Variability in emission data derived from multiple tests conducted at
different times includes components of variation from:
(1) process conditions,
(2) equipment and personnel variation in field procedures, and
(3) equipment and personnel variation in the laboratory.
In many instances time variations in source output may be the most signi-
ficant factor in the total variability. The error resulting from this
component of variation is minimized by knowing the time characteristics
of the source output and collecting the gas sample at a rate proportional to
the stack gas velocity. The sampling period should span at least one com-
plete output cycle when possible. If the cycle is too long, then either the
sample collection should be made during a portion of the cycle representative
of the cycle average or multiple samples should be collected and averaged.
Quality assurance guidelines for Method 3 as presented here are designed
to insure the collection of data of acceptable quality by prevention,
detection, and quantification of equipment and personnel variations in
both the field and the laboratory through
(1) recommended operating procedures as a preventive measure,
(2) quality control checks for rapid detection of undesirable
performance, and
(3) a quality audit to independently verify the quality of the data.
This document is divided into four sections:
Section I, Introduction - The Introduction lists the overall
objectives of a quality assurance program and delineates the
program components necessary to accomplish the given objectives.
Section II, Operations Manual - The Operations Manual sets forth
recommended operating procedures to insure the collection of data
of high quality, and instructions for performing quality control
checks designed to give an indication or warning that invalid
data or data of poor quality are being collected, allowing for
corrective action to be taken before future measurements are made.
Section III, Manual for Field Team Supervisor - The Manual for
Field Team Supervisor contains directions for assessing data
quality on an intrateam basis and for collecting the
information necessary to detect and/or identify trouble.
Section IV, Manual for Manager of Groups of Field Teams - The
Manual for Manager of Groups of Field Teams presents information
-------
relative to the test method (a functional analysis) to identify
the important operations, variables and factors, and statistical
properties of and procedures for carrying out a quality audit
for an independent assessment of data quality.
The scope of this document has been purposely limited to that of a field
and laboratory document. Additional background information will be
contained in the final report under this contract.
-------
SECTION II OPERATIONS MANUAL
2.0 GENERAL
This Operations Manual sets forth recommended procedures for performing
stack gas analysis for carbon dioxide, excess air, and dry molecular weight
according to Method 3. (Method 3 is reproduced from the Federal
Register, and is included as appendix A of this document.) Quality control
procedures and checks designed to give an indication or warning that
invalid or poor quality data are being collected are written as part of the
operating procedures and are to be performed by the operator on a routine
basis. Results from certain strategic quality control checks will be used
by the supervisor for the assessment of data quality.
The sequence of operations to be performed for each field test is given in
figure 1. Each operation or step in the method is identified by a block.
Quality checkpoints in the measurement process, for which appropriate quality
control limits are assigned, are represented by blocks enclosed by heavy
lines. Other quality checkpoints involve go/no-go checks and/or subjective
judgments by the test team members with proper guidelines for decision
making spelled out in the procedures.
The precision/accuracy of data obtained from this method depends upon
equipment performance and the proficiency and conscientiousness with
which the operator performs his various tasks. From equipment checks
through on-site measurements, calculations, and data reporting, this method
is susceptible to a variety of errors. Detailed instructions are given for
minimizing or controlling equipment error, and procedures are recommended to
minimize operator error. Before using this document, the operator should
study Method 3 as reproduced in appendix A in detail.
It is assumed that all apparatus satisfies the reference method specifi-
cations and that the manufacturer's recommendations will be followed when
using a particular item of equipment (e.g., Orsat analyzer or sampling train
components).
2.1 APPARATUS REQUIREMENTS
A general description of the required apparatus along with desirable design
qualities is given for grab sampling, integrated sampling, and the gas
analyzer.
2.1.1 Grab-Sampling Train
2.1.1.1 General Description of the Grab-Sampling Train - A drawing of a
grab-sample train is given in figure 3-1 of appendix A. Equipment
specifications are given in subsection 2.1 of appendix A.
-------
APPARATUS SELECTION
1. SELECT EQUIPMENT ACCORDING TO THE GUIDELINES
GIVEN IN SUBSECTION 2.1, PAGE 4.
PRES/MPLING PREPARATION
2. INSPECT, CLEAN AND CHECK THE EQUIPMENT FOR
PROPER OPERATION ACCORDING TO INSTRUCTIONS
GIVEN IN SUBSECTION 2.2.1, STARTING ON
PAGE 10.
BEFORE EVERY THIRD FIELD TEST OR EVERY THREE
MONTHS, WHICHEVER COMES FIRST, PERFORM AN
OPERATIONAL CHECK OF THE ANALYZER USING A
SAMPLE WITH KNOWN CONCENTRATIONS AS DIRECTED
IN SUBSECTION 2.2.2, PAGE 12.
PACK THE EQUIPMENT FOR SHIPMENT SO AS TO
PRECLUDE DAMAGE WHILE IN TRANSIT.
ON-SITE PEASUREMEHIS
APPARATUS
SELECTION
APPARATUS
CHECK
I
OPERATIONAL
CHECK OF
ORSAT ANALYZER
PACKAGE EQUIPMENT
FOR SHIPMENT
5.
6.
ASSEMBLE THE EQUIPMENT FOR SAMPLING. VISUALLY
CHECK EACH ITEM FOR POSSIBLE DAMAGE SUSTAINED
DURING TRANSIT.
COLLECT THE GAS SAMPLE, EITHER GRAB OR INTEGRATED
AS APPROPRIATE, AS DIRECTED IN SUBSECTION 2.3.1
OR 2.3.2, RESPECTIVELY.
ASSEMBLE AND
CHECK THE
SAMPLING TRAIN
i
COLLECT
SAMPLE
7.
PERFORM REPLICATE ANALYSES UNTIL THE PERFORMANCE
CRITERIA SUGGESTED IN STEP 11 OF SUBSECTION 2.3.2
IS SATISFIED.
I
ANALYZE
SAMPLE
Figure 1: Operational Flow Chart of the Measurement Process
-------
POSTS/WONG OPERATIONS
8. COMPARE THEORETICAL AND MEASURED VALUES OF
PERCENT C02 AND DRY MOLECULAR WEIGHT AS
DIRECTED IN SUBSECTION 2.4.1, PAGE 20.
DISASSEMBLE AND VISUALLY INSPECT THE
EQUIPMENT FOR POSSIBLE DAMAGE SUSTAINED
DURING THE FIELD TEST. DOCUMENT THE
DAMAGE AND ESTIMATE ITS EFFECT ON THE
MEASURED VALUES.
10.
11
12.
PACK THE EQUIPMENT FOR SHIPMENT TO THE HOME
LABORATORY.
PERFORM CALCULATIONS FOR THE COMPONENT GASES
AND FOR THE DRY MOLECULAR WEIGHT OF THE
STACK GAS.
FORWARD THE DATA FOR FURTHER INTERNAL REVIEW
OR TO THE USER.
8
COMPARE MEASURED
AND THEORETICAL
VALUES
I
DISASSEMBLE
AND INSPECT
EQUIPMENT
10
I
PACKAGE
EQUIPMENT FOR
SHIPMENT
PERFORM
CALCULATIONS
12
I
REPORT
DATA
Figure 1: Operational Flow Chart of the Measurement Process (Continued).
-------
2.1.1.2 Desirable Design Qualities - The probe tip should be designed so
as to prevent the glass wool filter from being drawn from the probe when
sampling a source that has substantial negative pressure.
Even though a one-way squeeze bulb is acceptable as a pump and in certain
constrained environments may have to be used, it is recommended that, if
possible, a leak-free diaphragm-type pump be used in the train.
2.1.2 Integrated Gas-Sampling Train
2.1.2.1 General Description of the Integrated Gas-Sampling Train - A
drawing of an integrated gas-sampling train is given in figure 3-2 of
appendix A. Equipment specifications are given in subsection 2.2 of
appendix A. Specifications of and directions for using the Type-S pitot
tube are given in the Quality Assurance Document of this series for
Method 2.
2.1.2.2 Desirable Design Qualities - No volume specifications are given
for the air-cooled condenser in the sampling train. The main consider-
ation is that the condenser volume be kept to the minimum necessary to
sufficiently cool the sample air. This is important because the larger
the volume the more difficult it is to completely purge the sampling train
before collecting a sample.
An alternate means of cooling the sample air requiring less volume would
be to use a midget impinger in an ice bath as a condenser. The impinger
tip should be well above water level for this purpose.
2.1.3 Gas Analyzer (Orsat)
The Orsat analyzer is used to determine the stack gas composition in terms
of C02» 02, and CO concentrations. A sample is analyzed by successively
passing it through absorbents that remove specific gaseous components.
The difference in gas volume before and after the absorption represents the
amount of the constituent gas in the sample. Constant pressure and temper-
ature must be maintained throughout the analysis. Results are reported on
a dry, volume percentage basis.
2.1.3.1 General Description - The Orsat analyzer is illustrated in
figure 2. The apparatus consists basically of a glass burette to measure
gas volume, a water jacket to maintain constant temperature, a glass mani-
fold to control the flow of gases, three absorption pipettes, and a leveling
bottle to move the gases. The apparatus is usually assembled inside a case
equipped with front and rear doors and a carrying handle.
In this document two Orsat analyzer designs taken to represent the extremes
in obtainable precision are referred to as a standard Orsat analyzer and a
modified Orsat analyzer. The standard Orsat analyzer, as used herein,
implies an analyzer containing a burette with 0.2 ml divisions with a
-------
spacing between divisions of about 1 mm. The modified Orsat analyzer, on
the other hand, implies an analyzer equipped with a burette having 0.1 ml
divisions with the spacing between divisions being 1 mm or more.
2.1.3.2 Desirable Design Qualities - There is a wide variety of Orsat
analyzer designs on the market. Some specific design features increase
the precision and accuracy attainable with an Orsat analyzer. Some of
these design features are:
(1) Precision and perhaps accuracy are improved with a glass
burette configuration as shown in figure 2. That is, the
burette is designed with a large diameter column having a
volume of approximately 75 ml joined onto a portion of a
25 ml burette graduated in 0.1 ml divisions. Such a
design should result in less error in leveling the liquids
and in interpolating readings or the burette than with
burettes not as finely graduated and with less spacing
between divisions (ref. 1). This design is referred to as
a modified Orsat analyzer throughout this document.
(2) The volume reference mark should be located on the
capillary tubing at the top of the glass burette (fig. 2)
as opposed to being on the larger diameter burette. Having
the reference mark on a small bore capillary tube increases
the precision with which the total sample volume can be
determined from test to test. Also, it should result in a
more accurate burette calibration and thus a more accurate
sample volume determination.
(3) The glass manifold should have as small a volume as possible
to reduce the possibility of diluting the sample due to
incomplete purging of the manifold. It also minimizes the
increase in sample volume (i.e., the volume of gas in the
manifold between the reference mark on the burette and the
pipette is small).
(4) The Orsat apparatus and case should be designed so that the
leveling bottle and glass burette can be viewed side by side
when leveling the liquid. The liquid levels in the burette
and the leveling bottle must be the same when reading
volumes; otherwise, the sample air will not be at atmospheric
pressure.
(5) The inlet sample valve (see figure 2) should be a three-way
valve to allow purging of the manifold without causing the
sample bag or inlet gas to be diluted by ambient air.
-------
THREE-WAY INLET VALVE TO MANIFOLD
INLET VALVE TO CO PIPETTE
-INLET VALVE TO 02 PIPETTE
INLET VALVE TO C02 PIPETTE
GLASS MANIFOLD
SAMPLE
INLET
REFERENCE
MARK
WATER
JACKET
Figure 2: Illustration of Key Components of an Orsat Analyzer
for Measuring C02> 0 , and CO Contents of Stack Gases
-------
2.2 PRESAMPLING PREPARATION
2.2.1 Apparatus Check
2.2.1.1 Grab Sample (figure 3-1, Appendix A) - The grab-sample train should
be checked as follows:
(1) The probe should be visually checked for signs of
leaks. A glass probe will be either broken or
cracked, and a stainless steel probe will be corroded
as an indication of leaking. The probe should be
cleaned and both ends capped.
(2) The pump, either a one-way squeeze bulb or leak-free
diaphragm type pump, should be checked to see if it is
operating properly. All connectors and tubing should
be checked for leaks. This can be done by slightly
pressurizing the system and checking for leaks by
applying soap bubbles to the connections and joints and
observing to see if the bubbles burst more rapidly than
normal.
2.2.1.2 Integrated Sample (figure 3-2, Appendix A) - The integrated-gas
sampling train should be checked as follows:
(1) The probe should be cleaned with soap and water, rinsed
with distilled water and allowed to dry. It should be
visually checked for leaks. A glass probe will be either
broken or cracked, and a stainless steel probe will show
signs of corroding if leaking. Both ends of the probe
should be capped to prevent contaminants from entering
the probe while not in use.
(2) The air-cooled condenser or equivalent should be cleaned
and leak checked.
(3) The needle valve and rotameter should be disassembled,
cleaned, and reassembled at any sign of foreign matter
in the rotameter or erratic behavior of the rotameter.
(4) Leak test the Tedlar bag by filling it with air (under a
slight pressure), sealing, and letting sit overnight.
If there is any visual indication of collapsing, do not
use it for the field test. It is recommended that the
bag be mounted in a rigid container to prevent puncture
when in the field.
(5) Service the pump according to the manufacturer's directions.
Check the motor for proper electrical grounding, i.e., use
a three-wire system.
10
-------
2.2.1.3 Gas Analyzer (Orsat) - An Orsat analyzer or its equivalent is
utilized in the determination of CO
2'
0-, and, if desired, CO.
The Orsat
should be checked and serviced before each field test in the following
manner:
(1) The solution used in the leveling should be distilled
water containing approximately 5 percent by volume of
concentrated sulfuric acid (2 or 3 ml of methyl orange
indicator is usually added to indicate that the solution
is acidic). The solution is then saturated with a salt,
usually sodium sulfite or sodium chloride, at the temper-
ature at which the Orsat is expected to operate. The
sulfuric acid acts as a drying agent to remove any
moisture from the sample and the saturated salt solution
prevents the absorption of sample gases by the leveling
solution. This should be prepared as a stock solution
and checked before each field test or prepared fresh for
each field test.
(2) Absorbing solutions should be changed if more than 10
passes are needed to obtain a constant reading for any
component gas.
Charge the Orsat analyzer when applicable (following
the manufacturer's instructions) with fresh absorbing
reagents. Any absorbing reagent can be used if it does
not react with other gases in the sample (i.e., not
subject to interferences) and is quantitative for the
component gas being analyzed. Some commonly used
absorbents are:
Absorbent
Potassium hydroxide
Alkaline pyrogallic acid, chromous chloride,
or equivalent
Acid cuprous chloride over copper strips,
cuprous sulfate 1-beta naphthol, or equivalent
Gas
co2
°2
CO
(3) The stopcocks should be removed and cleaned. Stopcock
grease should be carefully applied so as to insure a
leak-free system and to preclude plugging the air
passages. Stopcocks are generally not interchangeable.
Replace each one in the same port that it was originally
removed from.
11
-------
(4) Assemble the Orsat analyzer and check for leaks as
follows:
(a) According to the manufacturer's procedures appro-
priate to the Orsat being used, adjust the level of
each absorbing reagent to its reference mark (see
fig. 2); then close the stopcock to that pipette.
(b) With the sample inlet valve open, lower the liquid
level in the glass burette to near the 0.0 ml mark.
Close the sample inlet valve (stopcock) and set the
leveling bottle on top of the Orsat case. This
pressurizes the sample air in the burette and
manifold.
(c) Observe the analyzer in this position for 10 to 15
minutes after the liquid level in the burette has
stabilized. If there are no leaks, all absorbing
solution levels should remain at their respective
reference marks and the liquid level in the burette
should remain fixed once it has stabilized after
placing the leveling bottle on the top of the case.
If leaks are present, they must be eliminated before performing
a field test.
2.2.2 Calibration Check of the Orsat Analyzer
It is recommended that the following calibration check be performed before
every third field test or before any field test in which the Orsat analyzer
has not been checked during the previous 3 months.
To check the 0. absorbing reagent and the operating technique of the
operator, it is recommended that the percent of 0~ in air be determined.
The average of three replicates should be 20.8 +0.7 percent when using the
standard Orsat and 20.8 + 0.35 when using a modified Orsat. A measured
average value higher than 21.5 percent indicates poor operator technique,
while a value lower than 20.1 percent indicates leaking valves, spent
absorbing reagent (for 0? only), and/or poor operator technique. (See
subsection 4.1 for the derivation of the above limits.)
A more thorough check that could be made if the required equipment is
available is to sample from a manifold containing a known mixture of CO,
CO, and air. This is applicable to grab samples or by the integrated bag
technique. In both cases the sample is analyzed for CO, CO-, and 0_ using
the Orsat analyzer. The average of three replicates should be within
12
-------
approximately 0.7 percent (absolute) of the known concentration of each gas
(or 0.35 percent when using the modified Orsat). Again, high measured
values indicate poor operator technique, while low values indicate leaking
valves, spent absorbing reagent, and/or poor operator technique.
If the above limits are exceeded, corrective action in the form of
equipment maintenance and/or operator training should be taken.
2.2.3 Package Equipment for Shipment
This aspect of the test method in terms of logistics, time of sampling, and
quality of data is very dependent upon the packing of the equipment in
regards to (1) accessibility in the field, (2) ease of movement on site,
and (3) optimum functioning of sampling and analytical devices in the
field. Equipment should be packed under the assumption that it will receive
severe treatment during shipment and field operation. Each item should be
packaged as follows:
(1) Probes, pumps, and condenser should be packed in cases
or wooden boxes filled with packing material or lined
with styrofoam.
(2) Rotameters, needle valves, and all small glass parts
should be individually packed in a shipping container.
(3) For integrated samples, it is advantageous that the
rigid container for the sampling bag serve also as
its shipping container.
(4) The Orsat should be disassembled and each item indi-
vidually packed in suitable packing material and rigid
containers. It is recommended that spare parts and
absorbent solution be shipped in another shipping
container.
2.3 ON-SITE MEASUREMENTS
The equipment, i.e., the sampling train and Orsat analyzer, are unpacked
at the sampling site and visually inspected for any damage that might have
been sustained during shipment from the laboratory.
It is suggested that, if at all possible, a nearby laboratory or room should
be utilized for conducting the gas analysis. The Orsat analyzer should be
used in the stack area as a last resort only. Accuracy and precision will
nearly always be enhanced by moving to a laboratory for analysis.
Assemble the Orsat analyzer and check for leaks as directed in subsection
2.2.1.3, step 4.
13
-------
Special precautions to observe when using an Orsat analyzer for gas analysis
are as follows:
(1) No ambient air must be allowed to enter the Orsat
analyzer during testing.
(2) The gas must always be sampled in the following
sequence:
Absorber No. 1 - CO,.,
Absorber No. 2-0
Absorber No. 3 - CO.
This is necessary because Absorber No. 2 will also absorb
C09 and Absorber No. 3 will absorb 09 and possibly CO ,
£ £ 2.
resulting in erroneous data.
(3) Solution in the burette must be saturated with the salt
at the operating temperature, or absorption of sample
gases prior to analysis will occur. Also, the solution
should be acidic (as indicated by the methyl orange
indicator) to absorb any moisture in the sample gas.
(4) Absorber solution must be kept from entering the capil-
lary column manifold. If solution enters the manifold,
the test should be voided and the sample manifold
cleaned. Acetone can be utilized in this cleanup.
(5) Gas samples must be allowed to come to temperature
equilibrium with the water jacket before analysis. A
minimum of 5 minutes should be allowed for equilibration.
(6) The data yields the molecular weight of the gas on a dry
basis and should be treated as such in future calculations.
(7) In placing the probe in the stack in any sampling method,
precautions should be taken to prevent dilution of the
stack gas by an influx of ambient air, i.e., if a negative
pressure exists in the stack.
(8) An Orsat analyzer must operate under constant temperature
and pressure; therefore, it is necessary when a reading is
taken from the Orsat analyzer that the levels of solution
in the burette and leveling bottle be the same to insure
equal pressure. The water jacket acts as a buffer for
temperature changes.
14
-------
(9) If the source being tested is known to have or is
suspected of having a high S02 concentration, it should
be measured quantitatively as by Method 6 and subtracted
from the (XL determinations. Measurement and subsequent
corrections should be made when the S02 level is suspected
of being as much as 3 percent (relative) of the CC>2
concentration when the data are to be used to correct
particulate emission rates to 12 percent C0_. If the data
are to be used just for calculating the molecular weight,
then interferences as high as 0.5 percent (absolute) or
5,000 ppm from S0? are acceptable (this level of interference
will result in an error of only about 0.1 Ib/lb-mole in M,).
2.3.1 Grab Sampling
Set up the grab-sampling train as depicted in Figure 3-1 of Appendix A.
Visually check each connection for possible leaks.
(1) Place the probe in the stack with the probe tip at least
12 inches from the stack wall.
(2) Plug the sampling port as well as possible with a sponge
or rag to prevent dilution of the stack gas by ambient
air if the stack pressure is negative.
(3) Purge the sampling train several times if a one-way
squeeze bulb is used, or for a few seconds if a leak-
free diaphragm-type pump is used.
(4) It is recommended that, if at all possible, a flexible
bag be used to collect the sample and that this bag be
transported to a laboratory for analysis.
(5) Draw sample gas into the analyzer and flush (i.e., allow
to bubble through the burette solution) at least three
times to saturate the liquid in the burette with the gas
being analyzed and to insure that the air remaining in
the manifold is of the same composition as the sample to
be analyzed.
Caution: Once the flushing operation has begun,
ambient air must not be allowed to enter
the manifold.
(6) Draw in a fixed volume (usually 100 ml) of the sample air
following the manufacturer's instructions. Allow a
minimum of 5 minutes for the sample air to come to a
temperature equilibrium with the water jacket around the
burette (unless the sample and the analyzer have both been
at the same temperature for a longer period of time).
15
-------
(7) Proceed with the sequential determination of C0«, 0», and
CO as directed by the instructions supplied by the manu-
facturer of the gas analyzer.
Note: If more than 10 passes are required to reach a
constant reading for any one of the component
gases, the applicable absorbing reagent should
be replaced.
(8) The Reference Method specifies that steps 4 through 7 be
repeated until three consecutive analyses vary no more
than 0.5 percent (absolute) by volume for each component
gas being analyzed. Since results from collaborative
tests (refs. 1,2) indicate a standard deviation of approxi-
mately 0.4 percent (absolute), the above criteria could be
difficult to satisfy. Also, the molecular weight deter-
mination is shown to be relatively insensitive to gaseous
component .measurement errors (see subsection 4.1.1.1).
Therefore, it is recommended here that the average of the
first three consecutive analyseswhere the range, R,
(i.e., the difference in the largest and smallest values)
for each of the component gases is not greater than
1.74 percent (absolute) or 0.87 percent (absolute) when
using the modified Orsat analyzerbe used. See subsection
4.1 for a discussion of these limits.
(9) Fill in the information required and the test data on an
Orsat Field Data Sheet as shown in figure 3.
2.3.2 Integrated Sampling
Integrated sampling specifies sampling at a rate proportional to the stack
gas velocity. This requires the use of a type-S pitot tube or equivalent
to monitor the stack gas velocity. Directions for the care and use of the
type-S pitot are given in the Quality Assurance Document of this series
for Method 2.
The procedure for collecting an integrated sample is as follows:
(1) Evacuate the flexible bag. This can be accomplished by
connecting one end of a piece of flexible tubing equipped
with a quick disconnect to the flexible bag and the other
end to the intake side of the sampling train pump.
(2) Set up the integrated gas-sampling train as shown in
figure 3-2 of appendix A, except do not connect the
flexible bag.
16
-------
Location
Date
FIELD TEST IDENTIFICATION INFORMATION
Comments:
Time
Opera to r_
Date* (
Reagents Used
(CO,),
(CO)
Average: %C02
MEASURED RESULTS
10 \J r\
M
THEORETICAL RESULTS
%CO
M
Ib/lb-mole
RECORDED FIELD DATA
Replicate
Number
Original
Volume
Reading 1
(C02)
Reading 2
(ml)
(co2)
Vol ume
(2-1)
(ml)
(02)
Reading 3
(ml)
(02)
Vol ume
(3-2)
(ml)
(CO)
Reading 4
(ml)
(CO)
Vol ume
(4-3)
(ml)
Date that the absorbing reagents were replaced.
Figure 3. Sample Orsat Field Data Sheet
17
-------
(3) Visually and physically inspect each connection to insure
that they are tight and leak-free.
(4) Insert a plug of borosilicate (Pyrex) filtering fiber
(glass wool) into the inlet end of the probe to filter
out particulates.
(5) Place the probe in the stack with the probe end at least
12 inches from the stack wall, and plug the sampling port
as well as possible (if the stack gas has a positive
pressure and there are hot and/or noxious gases, a packing
gland should be used on the sampling port to prevent the
escape of these gases).
(6) With the flexible bag still disconnected, purge the
sampling train by running the pump for a sufficient time
to completely purge the system, especially the air-cooled
condenser. The minimum purging time should be such that
a volume of sample gas at least as large as three times
the volume of the sampling train including the condenser
will pass through the train.
(7) With the pump turned off, connect the sampling train to
the flexible bag.
(8) Sample at a rate proportional to the stack gas velocity
as monitored by a type-S pitot tube. The rate of
sampling is varied according to the variation of the
square root of the velocity pressure differential, i.e.,
sampling rate as indicated by the rate meter is set and
subsequently adjusted according to the values of /AP.
(9) Disconnect the flexible sampling bag and remove to a
suitable area for performing the Orsat analysis. The
collected sample should be allowed to sit for about
30 minutes to insure thorough mixing and temperature
equilibrium. It is recommended that the analysis be
performed as soon as practical after the 30-minute
waiting period. Although no undesirable effects have
been reported for stored samples, a delay in analysis
of more than 8 hours should be avoided if possible.
(10) Following the instructions supplied by the manufacturer,
connect the Orsat analyzer to the sample bag and purge
the manifold, saturate the liquid in the burette with
sampling gas, and draw in the desired volume of sample
as discussed in steps 4 through 7 of subsection 2.3.1,
"Grab Sampling".
18
-------
(11) The Reference Method specifies that the analysis be
repeated until three consecutive analyses vary no more
than 0.2 percent by volume for each component gas being
analyzed. This performance criteria may be difficult
to achieve (ref. 2). It is suggested here that the
following criteria be used:
(a) When the data are to be used for computing the
stack gas molecular weight, accept the average of
the first three consecutive analyses where the
range, R, (i.e., the difference in the largest and
smallest values) for each of the component gases
is not greater than 1.74 percent (0.87 percent for
modified Orsat). (Note that a CO measurement is
not used in calculating molecular weight (see
subsection 2.4.4.2).) This is consistent with the
repeatability obtained in a collaborative test
using standard Orsat analyzers (ref. 2). See
section 3.2 for a discussion of this point. This
range limit implies that the average of the three
analyses is within + 1 percent (absolute) of the
true value (z 0.5 percent for the modified Orsat)
about 98 percent of the time (see subsection 4.1)
and, subsequently, the calculated molecular weight
would be in error by no more than about 1.5 percent
(relative) or about 0.75 percent when using the
modified Orsat.
(b) When the data are to be used to correct the partic-
ulate emissions level of incinerators to 12 percent
CO- as well as to determine the dry molecular weight,
the following procedures are recommended:
(1) Perform three replicate analyses. Use the averages
of the three replicates for %02 and %CO if the
ranges are within the limits given in (a) above.
(2) For %CO? compute the average, %CO~, from the
three replicates. From figure 6, page 42,
determine the required number of replicates,
r, in accordance with the type of Orsat being
used. The solid curve is applicable when a
standard Orsat with 0.2 ml divisions is used,
and the dashed curve when an Orsat with 0.1 ml
divisions (modified Orsat) is used.
(3) Counting the original three replicates, perform
any additional replicates as determined above.
19
-------
(4) Accept the \average of the r replicates as the
true value if their range, R, satisfies the
criteria given in Table 1 of Section III under
l(b), page 29.
If the criterion of (4) cannot be met, corrective action in
the form of replacing the absorbing reagent or having the
analyses repeated by another operator should be taken.
2.4 POSTSAMPLING OPERATIONS
2.4.1 Compare Measured Values Against Theoretical Values
After the analyses have been performed and before the equipment is
disassembled, the measured and theoretical results should be compared as a
quick check for gross measurement error.
Combustion nomographs are available commercially (ref. 3) for estimating
the percent by volume of CCL, CO, and 0? when the fuel composition is
known. Also, the molecular weight of the stack gas can be calculated using
the nomographs.
Perform the calculations on the measured data as directed in subsection
2.4.4, "Calculations". The following comparisons are suggested:
(1) If the measured data are to be used for determining the
stack gas molecular weight only, compare the estimated
dry gas molecular weight, M, , and the measured M, by
D = M, - M,
M^ dm de
where D = Difference in measured and estimated
d values, lb/lb-mole,*
M = The measured value as calculated in
Subsection 2.4.4.2, lb/lb-mole, and
M, = The estimated value using the combus-
tion nomograph, lb/lb-mole.
If D.. is less than 0.6 lb/lb-mole, accept the measured
Md
value; otherwise, it is recommended that another sample
be drawn from the stack and analyzed. The average of
the two measurements is used in subsequent calculations.
Theoretical values should be recorded on the form in
figure 3.
Molecular weight is numerically the same in lb/lb-mole, g/g-mole, or
Kg/Kg-mole, etc.
20
-------
(2) If the measured data are to be used to correct
particulate emission levels to 12 percent CCL in
addition to the comparison given in (1) above, perform
the following comparison:
= %C°2m - %C02e
where D = Difference in measured and estimated
CO
2 values, percent,
C0? = Measured CO,., in percent (average of
r replicates) , and
CO.-, = Estimated or theoretical CO in percent.
Accept the measured value if D is less than 2 percent
UU ry
(absolute); otherwise, check the equipment and technique
then collect and analyze r more samples. (Note: Some
judgment has to be made by the supervisor as to how good
the process information is before repeating the analysis.)
Record the estimated or theoretical values on the form in figure 3 (these
theoretical calculations could have been made and recorded before the field
test if sufficient prior knowledge of the process were available).
2.4.2 Disassemble and Inspect Equipment
When disassembling the equipment, it is important to visually inspect the
sampling train components and the Orsat analyzer for any signs of damage
that could have adversely affected the measured values.
Any identified damage that was not detected during the test should be
documented on the field data sheet and thoroughly evaluated, by performing
the appropriate apparatus check as directed in subsection 2.2.1, when back
in the laboratory. If after checking it is concluded that the damage could
have biased the measurements, a description of the damage and an estimate
of direction and magnitude of potential bias in the data should be a part
of the field test report. If possible, repeat the field test.
2.4.3 Pack Equipment for Shipment to Laboratory
Pack the equipment for shipment to the laboratory in the same manner
described in subsection 2.2.3. Also, the data sheets, which have been
prepared in duplicate, are returned to the laboratory; one copy by mail and
one copy hand-carried.
21
-------
2.4.4 Calculations
2.4.4.1 Component Gases - Compute the average value for each of the
component gases from the r consecutive analyses satisfying the suggested
performance criteria given in subsection 2.4.1. Round each average to the
nearest 0.01 percent. (The Reference Method says to report to the nearest
0.1 percent; it is felt that averaging three or more values justifies
rounding to 0.01.percent through intermediate calculations, and the final
molecular weight will be rounded to the first decimal.) Record the averages
on the form in figure 3 in the spaces for Measured Values.
2.4.4.2 Dry Molecular Weight - Compute the average measured molecular
weight of the stack gas on a dry basis by
M, = 0.44 (%C00) + 0.32 (%00) + 0.28 (%N_ + %CO)
dm 2. L 2.
where M, = The average measured dry molecular weight,
Ib/lb-mole,
%CO? = Percent C02 taken as average of three
analyses, percent,
%02 = Percent 0,, taken as average of three
analyses, percent, and
(%N2 + %CO) = 100 - (%C02 + %0~2), percent.
Round M, to the nearest 0.1 Ib/lb-mole. Record the value as M, on the
dm d
form in figure 3, page 17. File a copy of the completed form in the labor-
atory log book and forward the original for further internal review or to
the user.
22
-------
SECTION III MANUAL FOR FIELD TEAP1 SUPERVISOR
3.0 GENERAL
The term "supervisor" as used in this document applies to the individual
in charge of a field team. He is directly responsible for the validity and
the quality of the field data collected by his team. He may be a member of
an organization that performs source sampling under contract to government
or industry, a government agency performing source sampling, or an industry
performing its own source sampling activities.
It is the responsibility of the supervisor to identify sources of uncertainty
or error in the measurement process for specified situations and. if possible,
to eliminate or minimize them by applying appropriate quality control proce-
dures to assure that the data collected are of acceptable quality. Specific
actions and operations required of the supervisor for a viable quality
assurance program are summarized in the following listing.
(1) Monitor/Control Data Quality
(a) Direct the field team in performing field tests
according to the procedures given in the
Operations Manual.
(b) Perform or qualify results of the quality control
checks (i.e., assure that checks are valid).
(c) Perform necessary calculations and compare quality
control checks to suggested performance criteria.
(d) Make corrections or alter operations when suggested
performance criteria are exceeded.
(e) Forward qualified data for additional internal
review or to user.
(2) Routine Operation
(a) Obtain from team members immediate reports of
suspicious data or malfunctions. Initiate correc-
tive action or, if necessary, specify special
checks to determine the trouble; then take correc-
tive action.
(b) Examine the team's log books periodically for
completeness and adherence to operating procedures.
(c) Approve data sheets, data from calibration checks,
etc., for filing.
23
-------
(3) Evaluation of Operations
(a) Evaluate available alternative(s) for accomplishing
a given objective in light of experience and needs,
(b) Evaluate operator training/instructional needs for
specific operations.
Consistent with the realization of the objectives of a quality assurance
program as given in section I, this section provides the supervisor with
brief guidelines and directions for:
(1) collection of information necessary for assessing data
quality on an intrateam basis;
(2) isolation, evaluation, and monitoring of major components
of system error;
(3) collection and analysis of information necessary for
controlling data qulaity.
3.1 ASSESSMENT OF DATA QUALITY (INTRATEAM)
Intrateam or within-team assessment of data quality as discussed herein
provides for an estimate of the precision of the measurements made by a
particular field team utilizing an Orsat analyzer. Precision in this case
refers to replicability, i.e., the variability among replicates and is
expressed as a standard deviation. This technique does not provide the
information necessary for estimating measurement bias (see subsection
4.1.2 for a discussion of bias) which might occur, for example, from
failure to collect a representative sample, sampling train leaks, or
inadvertent exposure of the sample to ambient air. However, if the
operating procedures given in the Operations Manual (section II) are
followed, the bias should be small in most cases. The performance of an
independent quality audit that would make possible an interteam assess-
ment of data quality is suggested and discussed in subsection 4.2 of the
Manual for Manager of Groups of Field Teams.
3.1.1 Treatment of Information
The field data are used to derive a confidence interval for the reported
data. The two measurements of interest here are the dry molecular weight
and the percent CO measurements. Both measurements are reported when
testing incinerators for particulate emissions; otherwise, only dry
molecular weight is reported.
24
-------
3.1.1.1 Calculating the Precision of Dry Molecular Weight Determinations -
Using the first three consecutive analyses that satisfy the range criteria
given in Table 1, page 29, compute three dry molecular weight values using
the relationship
MJ = 0.44 (%CO_) + 0.32 (%00) + 0.28 (100 - %C00 - %0.)
d f. f- if.
where M = Dry molecular weight of the sample gas,
Ib/lb-mole, and
%CO- and %0- = The measured values of CO- and 0 ,
respectively, for the same replicate,
percent (absolute).
Calculate the standard deviation of the three M, values by
1/2
(\ - *a) + (\ - «,) ^ (V «
s{Md} :
where s{M,} = Sample standard deviation calculated from
three replicates, Ib/lb-mole.
M, /M, \/M, \ = Dry molecular weight for replicate 1
dlV V\ V (2) (3), Ib/lb-mole.
M = Average dry molecular weight for three
replicates, i.e., 1/3/M, + M, + M, \,
Ib/lb-mole. \ 1 2 3/
The estimated standard deviation of the average dry molecular weight
becomes
S{Md} = s{Md}//3 =
based on three replicates.
3.1.1.2 Calculating the Precision of C02 Measurements - Precision of the
C0? measurements should be calculated and reported when the data are to be
used for correcting particulate emissions from an incinerator to 12 percent
C0_.
25
-------
/ " " \ £
/yrn ypn \
/ AL«U _ AL.U _\ i
V 2i V
- (%CO - %CO N" + \-h
\ 2 / \
£co2 - %co2\
r /
r - 1
An estimate of the standard deviation is calculated from the r replicates
1/2
-
where s^CC^} = Sample standard deviation calcu-
lated from r replicates, percent.
%C02 /%CO \ /%CO \ = Percent C02 measured by
replicate 1(2) (r), percent.
%C02 = Average percent C02 for r
replicates, i.e., %CO =
l/r/%CO + %CO + +%CO
\ 1 2 i
The estimated standard deviation of the average percent C0_ is given by
a{%co2> = s{%co2}//r
based on r replicates.
3.1.2 Reporting Data Quality
Stack gas molecular weight and, in some cases the percent C0? as measured
by this method, are used in conjunction with pollutant specific methods to
arrive at average emissions levels for those pollutants. When reported as
individual quantities, they should be accompanied with precision statements.
3.1.2.1 Reporting Dry Molecular Weights - It is recommended that the
average measured dry molecular weight be reported with 90 percent confidence
limits. Assuming that M, is normally distributed (this is usually a valid
assumption since sample means tend to be normally distributed even for
nonnormal parent distributions) and using 5{M, } as calculated above to
dm
estimate the standard deviation, exact confidence limits can be calculated
for the true M, value using the Student t-distribution with r - 1 = 2 degrees
of freedom. This assumes no bias in the average values. The average
measured value with 90 percent confidence limits would be
M, +2.92 &{M, }
dm dm
26
-------
where M, = The average of three replicates, Ib/lb-mole,
dm
} = The estimated standard deviation of M based
on three replicates, Ib/lb-mole, and
2.92 = 95th percentile of the Student t-distribution
with 2 degrees of freedom which yields a
90 percent confidence interval.
For example, if &{**.} was calculated to be 0.14 Ib/lb-mole, the reported
value with 90 percent confidence limits would be
M, 4- 0.41 Ib/lb-mole.
dm
The utility of the above statement follows from the fact that if this
procedure for computing confidence limits is followed, then 90 percent of
the time the true M, value will be contained within the given limits
(assuming that M, is not biased).
3.1.2.2 Reporting CO- Measurements - Precision of CO- measurements is
reported only when the measurements are to be used to correct particulate
emissions of an incinerator to 12 percent COj. The same procedure as that
used for calculating the precision for M, is used for calculating the
precision of C0~ measurements, except the number of replicates is not
fixed at three. The average measured value of r replicates with 90 percent
confidence limits would be
^ + tn nc &{%C00}, for r - 1 degrees of freedom
2 ~~ 0.05 2
where %CO = Average percent C0? from r replicates,
percent (absolute),
9{%CO } = The estimated standard deviation of %CO~
based on r replicates, percent (absolute),
and
t = 95th percentile of the Student t-distribution
with r - 1 degrees of freedom which yields a
90 percent confidence interval.
27
-------
Values of t for r - 1 degrees of freedom are given in t-tables in most
general statistics books, e.g., table 2, page 54, ref. 4.
If CO- values lower than about 4 percent (absolute) are to be measured, the
number of replicates required becomes large for the standard Orsat with an
assumed a{%CO } =0.4 percent as can be seen in figure 6, page 42.
3.2 SUGGESTED PERFORMANCE CRITERIA
Data assessment as discussed in the previous subsection was based on the
premise that all variables are controlled at a given level. These levels
of suggested performance criteria are the values given in the Operations
Manual for determining when equipment and/or personnel variability is out
of control. Criteria for judging performance are summarized in table 1.
The criteria for determining the number of analyses to be performed for a
field test were arrived at in the following manner. The functional analysis
of subsection 4.1 shows that if the standard deviations associated with
determining %CO and %0~ are as assumed in the functional analysis, the
resulting variability in the mean value determined from three replicates
for the dry molecular weight, M,, can be expressed in terms of a standard
deviation as a{M,} = 0.14 Ib/lb-mole. Three sigma limits for the average or
three replicates then are + 0.42 Ib/lb-mole and indicate that a maximum
error only slightly larger than 1.5 percent (relative) in M, should seldom
occur when the measurement process is in control (assuming a minimum, value
of M of 29 Ib/lb-mole). To control the variability in M , it is sufficient
to control the measurement variabilities of %CO~ and %0~. Hence, for a
sample size of three (assuming a{%CO,)} = a{%0_} = 0.4 percent) the range, R,
i.e., the difference in the largest and smallest values of %CO or %0_
should not exceed 1.74 percent (absolute) or 0.87 for a modified Orsat, more
than about three times per thousand tests when the process is in control
(based on the control chart approach).
When determining the percent by volume of C0~ in cases where the particulate
emissions level of incinerator is to be corrected to 12 percent C0~, it is
desirable to control the variability as a function of the CO,, level being
measured. Figure 6 shows the number of replicates required to be 98 percent
confident that the average of the measured values is within +_ 10 percent
(relative of the true value as a function of the C0? level being measured.
These data are given for the standard Orsat (assumed standard deviation of
0.4 percent) and for the modified Orsat (assumed standard deviation of
0.2 percent).
28
-------
Table 1: SUGGESTED PERFORMANCE CRITERIA
1. Suggested Criteria for Determining the Number of Analyses
per Field Test
a) When the data are to be used for determining M, only, use the
d
first three consecutive analyses for which the range,
Standard Orsat: R jc 1.74 percent (absolute)
Modified Orsat: R£ 0.87 percent (absolute)
for both CO, and 0? determinations.
b) For determining percent C0? for correcting the particulate
emissions level to 12 percent C0«, determine the number of
replicate analyses, r, from figure 6, page 42, according to the
type of Orsat_being used. Use the average of the first three
analyses as %C09 .
zt
Also, the range, R, of the r replicates must be no greater than
1)2° as shown in the table, i.e., R <_ D~c for r replicates. (See
table 9, page 70, of reference 4 for additional values of D_.)
Number of
Replicates, r
, -
Multiple of
D (for 3a Values)
, -
R <_ Da (Percent)
a = 0.4 percent a = 0.2 percent
,
3
4
5
6
7
8
9
10
11
^
4.358
4.698
4.918
5.078
5.203
5.307
5.394
5.469
5.534
1.74
1.88
1.97
2.03
2.08
2.12
2.16
2.19
2.21
0.87
0.94
0.99
1.02
1.04
1.06
1.08
1.10
1.11
2. Suggested Criteria for Performing a Calibration Check
Perform a calibration check every third field test or after three
months, whichever occurs first.
3. Suggested Criteria for Replacing Absorbing Reagent
Replace the applicable absorbing reagent when 10 or more passes are
required to reach a constant volume for a component gas.
29
-------
It must be emphasized that these limits were arrived at from a small
quantity of data. They are only suggested as a starting point. After
25 or 30 tests have been performed, the limits should be reevaluated and
tightened if possible. The limits should not be relaxed without the
approval of the Environmental Protection Agency.
3.3 COLLECTION AND ANALYSIS OF INFORMATION TO IDENTIFY TROUBLE
In a quality assurance program, one of the most effective means of
preventing trouble is to respond immediately to indications of suspicious
data or equipment malfunctions. There are certain visual and operational
checks that can be performed while the measurements are being made to help
assure the collection of data of good quality. These checks are written
as part of the routine operating procedures in section II. In order to
effectively apply preventive-type maintenance procedures to the measurement
process, the supervisor must know the important variables in the process,
know how to monitor the critical variables, and know how to interpret the
data obtained from monitoring operations. These subjects are discussed in
the following subsections.
3.3.1 Identification of Important Variables
Determination of stack gas composition requires a sequence of operations
and measurements that yields as an end result a number that serves to
represent the average moelcular weight or average percent of a component
gas for that field test. There is no way of knowing the accuracy, i.e.,
the agreement between the measured and the true value, for a given field
test. However, a knowledge of the important variables and their char-
acteristics allows for the application of quality control procedures to
control the effect of each variable at a given level during the field test,
thus providing a certain degree of confidence in the validity of the final
result.
A functional analysis of this method of measuring the molecular weight of
stack gases was made to try to identify important components of system
error. Also, collaborative tests have been performed to determine the
repeatability and reproducibility of the Method (refs. 1 and 2). Results
from the collaborative tests are used as overall system error while the
individual error components are estimated using engineering judgment in a
manner such that their combined variability is consistent with overall
system error.
Two of the most important error sources are 1) the inability to maintain
a constant pressure throughout the test, i.e., the inability to return
the sample to atmospheric pressure before making each volume reading, and
2) volumetric reading errors.
30
-------
These variables are assumed to be important because a modified Orsat
analyzer with a section of the burette marked in 0.1 ml divisions to
reduce reading error and a small cross section to increase the spacing
between divisions making it easier to attain equal levels of the liquid
in the burette and the leveling bottle, thereby maintaining a constant
pressure, showed a marked improvement in system precision (ref. 1). The
standard deviations associated with the measurements of both CO- and CL
were reduced by one-half or better (with a burette as shown in figure 2)
over that obtained with a standard Orsat analyzer. Both of these errors
are expected to be random, normal deviates with zero means.
Other components of measurement error are:
(1) failure to make sufficient passes for complete
absorption of a component gas,
(2) failure to saturate the leveling bottle solution
with the sample gases, or failure to maintain a
saturated salt solution in the leveling bottle,
(3) exposure of the sample to ambient air,
(4) spent absorbing reagent, and
(5) S0~ interference with CO determinations.
Since the results of the collaborative tests do not indicate a bias among
field teams, and a modified analyzer would probably not significantly reduce
their effect, the above five error sources are lumped together as a
normally distributed variable with a zero mean and treated as a third
component of system error for this analysis. This third component accounts
for about 20 percent of the total variability when using a standard Orsat
analyzer under these assumptions.
A brief description of the assumptions made and the techniques used in the
functional analysis is given in subsection 4.1 of this document. A more
comprehensive treatment will be given in the final report for this contract.
The source and magnitude of uncertainty for each of the above parameters
are discussed below.
3.3.1.1 Sample Pressure - A constant pressure is maintained throughout the
test by leveling the liquid in the burette to that in the leveling bottle.
Some Orsat analyzers are constructed in such a manner that it is difficult
to view the burette and the leveling bottle simultaneously. Some standard
analyzer burette scales have 1 ml occupying only 4 mm of the burette length
(ref. 1). Under field conditions significant errors can easily result with
the standard Orsat analyzer. Level differences of as much as +_ 1 to 1.5 mm
would be expected under adverse conditions. This difference is equivalent
to a volume error of about 0.4 percent (absolute). It is assumed for this
31
-------
analysis that this is one standard deviation; i.e., on the average
68 percent of the errors due to unequal liquid heights will be less than
0.3 percent by volume. It is further assumed that the modified analyzer
with a burette scale such that 1 ml occupies about 11 or 12 mm and the
scale has 0.1 ml divisions will reduce the standard deviation by a factor
of 3 or to 0.1 percent by volume.
3.3.1.2 Volumetric Reading Error - Burette scales on which the distance
between 0.2 ml divisions is of the order of 1 mm are subject to signifi-
cant reading errors under the best of laboratory conditions. Under
typical field conditions, it is felt that 0.2 percent (absolute) is a
reasonable standard deviation for reading error. Increasing the distance
between divisions and marking the scale in 0.1 ml divisions could possibly
reduce this variability by a factor of four resulting in a standard
deviation of 0.05 percent by volume.
3-. 3.1.3 Other Error Sources - To obtain an overall standard deviation of
about 0.40, it is necessary that the other sources combined have a standard
deviation of 0.173 percent and account for less than one-fourth of the
total variability. Control of this component of variability is largely
dependent upon the conscientiousness and ability of the operator. It
should not be particularly sensitive to the equipment being used.
3.3.2 How to Monitor Important Variables
In general, if the procedures outlined in the Operations Manual are
followed, the major sources of random variability will be in control. It
is felt, however, that as a means of verification of data quality, as well
as a technique for monitoring personnel and equipment variability, two
quality control charts should be constructed and maintained as part of the
quality assurance program. The quality control charts will provide a basis
for action with regard to the measurement process; namely, whether the
process is satisfactory and should be left alone, or the process is out of
control and action should be taken to find and eliminate the causes of
excess variability. In the case of this method in which documented
precision data are scarce, the quality control charts can be evaluated
after a period of time to determine the range of variation that can be
expected under normal operating conditions.
The two recommended quality control charts are:
(1) a range chart for the analyses performed in the field
which should serve as an effective monitor of operator
variability and, to a lesser extent, equipment
variability, and
(2) a chart for the differences in measured and known values
as obtained from calibration checks to monitor equipment
and/or operator variability as well as systematic
errors (biases).
32
-------
Discussions of control charts and instructions for constructing and
maintaining them are given in many books on statistics and quality control,
such as in refs. 4 and 5.
It is good practice to note directly on control charts the reason for out-
of-control conditions, if determined, and the corrective actions taken.
It is also good practice to maintain control charts in large size, e.g.,
8-1/2 x 11 (inches) or larger and to keep them posted on a wall for viewing
by all concerned, rather than have them filed in a notebook.
3.3.2.1 Range Chart - Figure 4 is a sample control chart for the range.
The chart was constructed for a sample size of three; i.e., only three
replicates per field test are used. It is recommended here that the range
be computed for the first three analyses performed for a given field test.
A standard deviation of 0.4 percent for the measurement error was assumed
in computing R and UCL. (For small sample sizes (r <_ 6) the lower control
limit (LCL) is effectively zero and is not given here.) It is suggested
that the same limits be used for both C02 and 0~ analyses until sufficient
field data are available to calculate R and UCL applicable to each component
gas and to individual field teams.
The R values are plotted sequentially as they are obtained and connected to
the previously plotted point with a straight line. Corrective action, such
as instruction in proper operating technique and/or equipment calibration
check, should be taken any time one of the following criteria is exceeded:
(1) One point falls outside the UCL.
(2) Two out of three consecutive points fall in the
warning zone (between 2a and 3cr limits).
(3) Seven consecutive points fall above the R line.
Exceeding any one of the criteria will usually indicate poor technique or
equipment malfunction between analyses.
3.3.2.2 Difference Chart - A sample quality control chart for the differ-
ence between measured and known values is shown in figure 5. The chart
was constructed using a standard deviation of 0.4 percent for the measure-
ment error and assuming that the test gas concentrations are accurately
enough known not to substantially increase this variability. Also, it was
assumed that there were no biases in the measurements; hence, CL = 0. It
is suggested that the chart as set up in figure 5 be used for both C02 and
0« checks until sufficient field data are available to compute new limits.
33
-------
1.8
1.6
1.4
1.2
1.0
£ UCL = ^74 l_^
-!->
C
O)
fe 0.8
0.6
0.4
0.2
0.0
UCL - 1.74
Warning Line = 1.39
UCL (R + 3aR)
1 2"3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2(
Field Test Number
Figure 4: Sample Control Chart for the Range, R, of Field Analyses
0)
u
s-
QJ
Q.
+ 1.
+ 1.
+0.
+0.
0.
-0.
-0.
-1.
-1.
UCL (+3o)
Naming Line (+2a)
I l_l 1 1_
1 1 1 1 1
1 1 1 I 1 J L_
1
Warning Line (-2a)
LCL (-3a)
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Calibration Check Number
Figure 5: Sample Control Chart for Calibration Checks
34
-------
For each calibration check, compute
D = X = X.
m t
where D = The difference in the measured and known
concentrations of CO- (CL) in percent,
X = The measured concentration of C00(00) in
m £ i
percent based on the average of three
determinations, and
X = The true or known concentration of C0~ (()) in
the calibration gas in percent.
Plot each D value on the quality control chart as it is obtained and connect
it to the previously plotted point with a straight line.
Corrective action such as replacing the absorbing reagent, performing other
equipment repair, and/or providing instructions on proper operating pro-
cedures should be taken any time one of the following criteria is
exceeded:
(1) One point falls outside the region between the
lower and upper control limits.
(2) Two out of three consecutive points fall in the
warning zone, i.e., between the 2a and 3a limits.
(3) Seven consecutive points fall on the same side
of the center line.
Exceeding the first and second criteria will usually indicate poor technique
or equipment malfunction. The third criterion when exceeded indicates a
system bias due either to a faulty analyzer or a consistent error in
performing operating procedures.
35
-------
SECTION IV IWJUAL FOR MANAGER OF GROUPS OF FIELD TEAMS
4.0 GENERAL
The guidelines for managing quality assurance programs for use with Test
Method 3 - Gas Analysis for Carbon Dioxide, Excess Air, and Dry Molecular
Weight are given in this part of the field document. This information is
written for the manager of several teams for measuring source emissions and
for the appropriate EPA, State, or Federal Administrators of these programs.
It is emphasized that if the analyst carefully adheres to the operational
procedures and checks of section II, then the errors and/or variations in
the measured values should be consistent with the performance criteria as
suggested. Consequently, the auditing routines given in this section
provide a means of determining whether the stack sampling test teams of
several organization, agencies, or companies are following the suggested
procedures. The audit function is primarily one of independently obtaining
measurements and performing calculations where this can be done. The
purpose of these guidelines is to:
(1) present information relative to the test method
(a functional analysis) to identify the iraportant
operations and factors,
(2) present a data quality audit procedure for use in
checking adherence to test methods and validating
that performance criteria are being satisfied, and
(3) present the statistical properties of the auditing
procedure in order that the appropriate plan of
action may be selected to yield an acceptable level
of risk to be associated with the reported results.
These three purposes will be discussed in the order stated in the sections
that follow. The first section will contain a functional analysis of the
test method with the objective of identifying the most important factors
that affect the quality of the reported data and of estimating the expected
variation and bias in the measurements resulting from equipment and
operator errors.
There are no absolute standards with which to compare the routinely derived
measurements. Furthermore, the taking of completely independent measure-
ments at the same time that the routine data are being collected (e.g., by
introducing two pitot tubes into the stack and collecting two samples
simultaneously) is not considered practical due to the constrained environ-
mental and space conditions under which the data are being collected.
Hence, a combination of an on-site system audit, including visual observa-
tion of adherence to operating procedures and a quantitative performance
quality audit check, is recommended as a dual means of independently checking
on the source emissions data.
36
-------
The second section contains a description of a data quality audit procedure.
The most important variables identified in section 4.1 are considered in
the audit. The procedure involves the random sampling of n stacks from a
ot size of N = 20 stacks (or from the stacks to be tested during a 3-
month period, if less than 20) for which one firm is conducting the source
emissions tests. For each of the stacks selected, independent measure-
ments will be made of the indicated variables. These measurements will be
used in conjunction with the routinely collected data to estimate the
quality of the data being collected by the field teams.
The data quality audit procedure is an independent check of data collection
and analysis techniques with respect to the important variables. It
provides a means of assessing data collected by several teams and/or firms
with the potential of identifying biases/excessive variation in the data
collection procedures. A quality audit should not only provide an indepen-
dent quality check, but also identify the weak points in the measurement
process. Thus the auditor, an individual chosen for his background
knowledge of the measurement process, will be able to guide field teams in
using improved techniques. In addition, the auditor is in a position to
identify those procedures employed by some field teams that are improvements
over the current suggested ones, either in terms of data quality and/or
time and cost of performance. The auditor's role will thus be one of aiding
the quality control function for all field teams for which he is responsible,
utilizing the cross-fertilization of good measurement techniques to improve
the quality of the collected and reported data.
The statistical sampling and test procedure recommended is sampling by
variables. This procedure is described in section 4.3. It makes maximum
use of the data collected, and it is particularly adaptable to the small
lot size and consequently the small sample size applications. The same
sampling plans can be employed in the quality checks performed by a team
or firm in its own operations. The objectives of the sampling and test
procedure are to characterize data quality for the user and to identify
potential sources of trouble in the data collection process for the
purpose of correcting the deficiencies in data quality.
4.1 FUNCTIONAL ANALYSIS OF TEST METHOD
Test Method 3 - Gas Analysis for Carbon Dioxide, Excess Air, and Dry
Molecular Weight is described in the Federal Register of December 23, 1971
and reproduced as appendix A of this document. Under standards of
performance for new stationary sources, Method 3 is used to determine the
dry molecular weight of the stack gas by measuring the percent by volume
of COp and 0« in the gas. Also, when testing for particulate emissions of
incinerators, the measured CO is used to correct the emissions level to
12 percent C0?. The functional analysis addresses itself only to the
determination of molecular weight and to the special case where particulate
37
-------
emissions are to be corrected to 12 percent CO,, in which case a higher
accuracy in the CO determination is desired.
The dry molecular weight is given by
Md = 0.44 (%C02) + 0.32 (%02> +0.28 (%N2 + %CO) (1)
where M, = Dry molecular weight, Ib/lb-mole,
%CO? = Average percent carbon dioxide by volume from
at least three analyses, dry basis,
%0~ = Average percent oxygen by volume from at
least three analyses, dry basis, and
(%N2 + %CO) = 100 - %C02 - %02, dry basis.
The percent carbon dioxide by volume is determined by the average of at
least three analyses. The effect of errors in the analyses of the
individual components to the errors in estimating true dry molecular
weight, Mj., and the true percent C00 by volume, %C00 , are discussed
at / /£
in the following subsections.
4.1.1 Variance Analysis
The standard deviations of the percent carbon dioxide and percent oxygen
by volume as obtained by the Orsat analysis are both assumed to be
0.4 percent by volume, i.e., cr{%CO_} = o{%0.} = 0.4 percent (absolute) when
using a standard Orsat analyzer, and a{%CO~} = a{%Qj} = 0.2 percent (abso-
lute) when using a modified Orsat analyzer as shown in figure 2 of section
II. The values used for the modified analyzer are of the order of those
obtained from a collaborative test using such an analyzer (ref. 1).
Standard deviations of 0.38 percent and 0.82 were obtained for determina-
tions of CO and 09, respectively, from a collaborative test using a
standard Orsat analyzer (ref. 2). Although the variability in measuring
0~ was much larger than that for measuring C0~ using the standard Orsat
analyzer, it was somewhat less than using the modified Orsat analyzer.
For this analysis, the standard deviation for determining CO,, as obtained
from the collaborative tests is used for both component gases.
Briefly, for the standard Orsat analyzer, the following assumptions were
made for single measurements of either C02 or 02,
38
-------
02{$G02} = a2{%02> = a2 + a2 + a2
and a2{%C02} = 0.16, a2 = 0.09, a2 = 0.04, and a2 = 0.03,
where a = Variability in returning the sample to and
maintaining atmospheric pressure when making
volume readings (ability to attain equal
levels of the liquid in the burette and the
leveling bottle),
cr = Reading error judged to be 0.2 percent for the
standard Orsat.
cr = All other error sources combined.
If these assumptions are reasonable, then reading and pressure errors
account for about 80 percent of the total variability. Furthermore, if
the modified analyzer as shown in figure 2, page 9, accomplishes the
following
a2{%0.} = 02{%CO_} = 0.04, a2 = 0.0075, a2 = 0.0025, and cr2 = 0.03,
£ / r K U
then the reading and pressure errors account for only 25 percent of the
total variability leaving other error sources to account for 75 percent
of the variability. For further improvement on precision, actions should
be taken to identify and control sources of variations within this group
called "other error sources."
Furhter background concerning the assumptions made in the analyses and the
methodology of this section will be contained in the final report of this
contract.
4.1.1.1 Analysis of M^, - Referring to equation (1) for dry molecular
£_..- Q
weight, the following equation is determined to relate the variances of the
measurements, under the assumption that the errors in measuring percent CO-
and 0- are uncorrelated.
£. £ e y, /~if~^ \ i f r\ o o \ ^ T W r\ ~] i ff\ O O \ ^ ^"T oy-vr
CT {Md> = (0.44r a {%C02> + (0.32)^ a {%02> + (0.28)^ a {%N2 + %CO}. (2)
39
-------
The above assumption is conservative since reading errors have a negative
correlation of -0.5 which would result in a smaller value of a{Mj}.
a
Errors due to incomplete absorption of CO , then subsequent absorption
when determining 0~, would also have a negative correlation. Therefore,
this is a worst case approach.
The variance of the average of r replications of percent C0~ by volume is
o _ a2{%CO } 2
_
0{%C00}
2
Assuming a minimum of three (3) replications and that all the variances
2
are equal to (0.4) , then equation (2) becomes
cr2{Mj = (0.4)2 0.3744/3 = 0.02
a
or
a{Md> * 0.14 Ib/lb-mole.
The coefficient of variation of Md is CV{Md> = 100 x a{Md>/Mdt, where Mdfc
is the true mean value. Thus, if M =30.0 Ib/lb-mole, the coefficient
variation is about 0.5 percent based on an average of three replicates.
Hence, it is expected that the M, based on the average of three determi-
nations on each component using a standard Orsat will be estimated with
precision within 0.42 Ib/lb-mole (30 limits).
It should also be noted that the true value of M, should fall between
a
29 and 31.3 Ib/lb-mole, and that if the average value 30.15 were used
(independently of any measured values), the maximum error would be
1.15 Ib/lb-mole, or the maximum relative error would be 1.15/29 = 0.04 or
4 percent compared to 0.5 percent for measured M, based on the average of
three replicates. Because M, appears as v^lT in equations for the deter-
mination of emissions, the relative error in the final determination would
be 2 percent, indicating that for practical consideration the errors in
estimating M_, are not critical unless the final answer is to be determined
to a finer precision than the 2 percent value.
40
-------
4.1.1.2 Analysis of %CO? by Volume - A more critical measurement is the
%CO« when it is used to correct the emissions level to 12 percent C0« for
estimating the particulate emissions of incinerators. In the special case
where a high degree of precision is desired for the CO determination, the
required number of replicate analyses is computed as follows on the basis
that the measurements of %CO_ are normally distributed about a true mean,
, and with a standard deviation of a{%C02> = 0.4 percent by volume.
Using the fact that the mean of r replications is also normally distributed
(this assumption is reasonable even if the origianl individual measurements
were not normally distributed) about the same true mean (assuming no bias)
with a standard deviation of
> = 0.4//r ,
then the probability that the absolute deviation between the measured mean
and the true mean is less than a positive number e is given by
< e} - 2<|> - 1 , (3)
P{
,where a = 0.4/Vr, and (~) is the area under the standard normal curve to
the left of the value x = e/o". This value is tabulated in the standard
tables, e.g., see ref. 6. Now suppose that it is desired that e be
0.10 /%CO_ \, 10 percent of the true mean, and that the probability be
0.98. That is, 98 percent of the time the average of r measured values
will be within +_ 10 percent of the true C0_ concentration.
Under these conditions, = 2.33, the 99th percentile of the standard
normal variable, in order that 2<|>(2.33) - 1 = 2(0.99) - 1 = 0.98. Thus,
0
o.io
e_ _
a 0.4//r
and
2.33
The required number of replicates r, for the desired precision is given in
figure 6 as a function of %C00 for both standard and modified Orsats with
2t
assumed standard deviations of 0.4 and 0.2 Ib/lb-mole, respectively. A
41
-------
22
17
12 -
11 -
IS IU
1
o
I "t
a! 9
1 1 |
cc
\ I
0 8
LU
CO
1 7
Q
1 1 i
~ 6
"
Or
LU
or.
4
3
O
2
t-
i
i
_ ^
Jh
i
w
*1
i
:_. i
J|i i&p£' £ ;
rYJj-K- ^!/^-^-'V' ':' '?.. ''
-' .".j'.! -V ^'''.- .^ ''."''. '*" '"'
\.' ' ?_ '''iiv"--','V' ,'* v*' "
"- "';.' .'''''!',*'' - ~ ' ''.
'. ''.'!! -.''V"1'-. ];''''.. ' '.% ':
"v'.i i". '."'."'. v? . .' 'j
". :;";:\ .'.'.'.',\.''l{-v"' ; ";;- :
:.'; :;.::. ^V-''''''*--1:*-"" ~*V
H^^^^, 98 PERCENT CONFIDENCE
(a{%C09> = 0.4 PERCENT)
*
h
ti
98 PERCENT CONFIDENCE
1 ^--^ (a{%C09) = 0.2 PERCENT)
_^^"
« 1 MINIMUM NUMBER OF
/ REPLICATES RECOMMENDED
^%; '9-K^^v v^;)' ':>':;;.'v:v-;^':v;:--, -....:;.':v: !'
'{.":':. .i/''.':'. :'. ::'^'-:'^' ..;"V- ''' '-.'.V;'.1'" '.'".'.': .-'!'."'. ' .' ";.'.
:''. - 'V .''-'."".V-'v|"-. >' 'v'.''-.' ' .T-"-''j' '' ..' ' '.'.-' : "/ ' .
'.'.'.'.'"' ".;-i"«r-v!-.J-v-r.!.> '- ^ r ".. V.'-.'-V ...:..:''. . :' . '--.^
^'V;^''v, ' '(':- "-''' V''^^ ' '.'*!. '; :'-' ' "' .;'':.''' .:'^""'^
%'..'.;'.'.;.'.'-,;'--.''.''!'''''"''- '." V,' '''''', .'" . i '."''!' .'!!"' 'r.'
V^Vs'^VoV"1'-' '^'.\v--;r:.'i:':^',"-^;;'V '-''''"' .'''"' ';-: :"\
'V;'V^'vi'-^''-:-:'..,tv^V. ".- ''' -J- V '' y. '''.-'; .''i '/!.-\'.- r. '"'' ' A
10
Figure 6: Number of Replications r for Estimating %C00 to
''t
Within + 10 Percent with 98 Percent Confidence.
42
-------
minimum of three replicates is recommended under all conditions. Also,
the average of three replicates is recommended to estimate %C02 to
determine r. t
From figure 6 it can be seen that three replicates are sufficient for any
%CO levels greater than 3 percent (absolute) when the replicability is
2t
0.2 percent (absolute) or less. Three replicates are adequate for
2t
levels greater than about 6 percent (absolute) when the replicability as
expressed by a standard deviation is no greater than 0.4 percent (absolute).
4.1.1.3 Control Chart for the Range, R, of Replicate Analyses - The above
results depend directly on the assumption that a{%C02> = 0.4% or 0.2% by
volume for any CCL level. This assumption can be checked on a continuous
basis using the range, R, of the replicate analyses and comparing against
the expected variation in the range as given by a standard control chart
as recommended in subsection 3.3.2.1. The range of r replicates is defined
as the largest value less the smallest value. The +3cr limit, i.e., the UCL
is given in table 1, page 29, for r = 3 through 11, and for assumed stan-
dard deviations of 0.4 percent and 0.2 percent.
For r larger than 6, it is suggested that the measurements be subdivided
into groups of nearly equal size between 3 and 6, as is possible. If a
constant r = 3 can be used, then the points (ranges) can be plotted on a
control chart with upper limit 1.74, which should be exceeded only about
3 times in 1000 for a = 0.4 percent by volume. Such a control chart is
illustrated in figure 4, page 34.
4.1.2 Bias Analysis
The bias analyses of the two measures of interest, M, and %CO?, are
straightforward. Suppose that there is a bias in the measurement of the
percent carbon dioxide by volume due to faulty equipment or a consistent
error in performing the opeating procedures. The average of r replica-
tions of the measurement will also be biased by the same amount. Denote
the biases in %C00 (or %C00) and %00 by TnA and Tn , respectively.
2 2
Then the bias in 100 - %C00 - %00 would be - T__ - T . Substituting the
z / i_»u f\ 7
biases for each component into equation (1) yields the corresponding bias
in M,, that is,
43
-------
Hence, the bias in M, is relatively small compared to the biases in the
individual component because of the offsetting or cancellation effect.
These biases are estimated as described in section 4.3.1.
4.2 PROCEDURES FOR PERFORMING A QUALITY AUDIT
"Quality audit" as used here implies a comprehensive system of planned and
periodic audits to verify compliance with all aspects of the quality
assurance program. Results from the quality audit provide an independent
assessment of data quality. "Independent" means that the individuals
performing, and as much as possible of the equipment used in the audit, are
different from the regular field crew and equipment. From these data both
bias and precision estimates can be made for the analysis phase of the
measurement process.
The auditor, i.e., the individual performing the audit, should have
extensive background experience in source sampling, specifically with the
characterization technique that he is auditing. He should be able to
establish and maintain good rapport with field crews.
The functions of the auditor are summarized in the following list:
(1) Observe procedures and techniques of the field team
during on-site measurements.
(2) Analyze a split sample on-site using own analyzer.
(3) Check/verify applicable records of equipment calibration
checks and quality control charts in the field team's
home laboratory.
(4) Perform calculations using data obtained from the audit..
(5) Compare the audit value with the field team's test value.
(6) Inform the field team of the comparison results specify-
ing any area(s) that need special attention or improvement.
(7) File the records and forward the comparison results with
appropriate comments to the manager.
4.2.1 Frequency of Audit
The optimum frequency of audit is a function of certain costs and desired
level of confidence in the data quality assessment. A methodology for
determining the optimum frequency using relevant costs is presented in
the Quality Assurance Documents for the methods requiring the results of
Method 3 and in the final report for this contract. Costs will vary among
44
-------
field teams and types of field tests. Therefore, the most cost effective
auditing level will have to be derived using relevant local cost data
according to the procedure given in the final report on this contract.
Since the potential error is much greater for particulate emissions of
incinerators, cost considerations may indicate that only incinerator tests
should be audited.
4.2.2 Collecting On-Site Information
While on-site, the auditor should observe the field team's overall
performance of the field test. Specific operations to observe should
include, but not be limited to:
(1) Setting up and leak testing the sampling train.
(2) Purging the sampling train with stack gas prior to
collecting the sample.
(3) Proportional sampling.
(4) Transfer of sample from the collapsible bag to the
Orsat analyzer.
The above observations, plus any others that the auditor feels are
important, can be used in combination to make an overall evaluation of the
team's proficiency in carrying out this portion of the field test.
In addition to the above on-site observations, it is recommended that the
auditor have his personal Orsat analyzer, preferably a modified one, and
perform analyses of the gas at the same time that the field team is
performing its analyses. The auditor should perform the analyses according
to the procedures given in section II.
4.2.2.1 Comparing Audit and Routine Values of M, - In field tests
requiring only the molecular weight, the audit and routine (field team's
results) values are compared by
d. = M, = M,,
where d. = The difference in the audit and field test results
J th
for the j audit, Ib/lb-mole,
M = Audit value of dry molecular weight, Ib/lb-mole, and
j
M, = Dry molecular weight obtained by the field team,
j Ib/lb-mole.
Record the value of d. in the quality audit log book.
45
-------
4.2.2.2 Comparing Audit and Routine Values of CO,, - When testing an
incinerator for particulate emissions (Method 5), in addition to calcu-
lating d. as described above, it is necessary to compare audited and
routinely derived values for C0~ by calculating
D. = %CO, - %C00
3 2j 2a.
where D = The difference in the audit and field test results
th
for the j audit, percent,
%C02 = Audit value of %CO as the average of r replicates,
j percent, and
%CO = Field team's value of %CO as the average of r
j replicates, percent.
Record the value of D. in the quality audit log book.
4.2.3 Collecting Laboratory Information
When visiting the field team's home laboratory, the auditor should verify
by checking the records that the performance criteria as given in 1
of section II have been met over the period since the last audit was
performed.
4.2.4 Overall Evaluation of Field Team Performance
In a summary-type statement the field team should be evaluated on its
overall performance. Reporting the d. value and, when applicable, the D.
value as previously computed, is an adequate representation of the objective
information collected for the audit. However, unmeasurable errors can
result from nonadherence to the prescribed operating procedures and/or
from poor technique in executing the procedures. These error sources have
to be estimated subjectively by the auditor. Using the notes taken in the
field, the team could be rated on a scale of 1 to 5 as follows:
5 - Excellent
4 - Above average
3 - Average
2 - Acceptable, but below average
1 - Unacceptable performance.
In conjunction with the numerical rating, the auditor should include
justification for the rating. This could be in the form of a list of
the team's strong/weak points.
46
-------
4.3 DATA QUALITY ASSESSMENT
Two aspects of data quality assessment are considered in this section. The
first considers a means of estimating the precision and accuracy of the
reported data, e.g., reporting the bias, if any, and the standard deviation
associated with the measurements. The second consideration is that of
testing the data quality against given standards using sampling by
variables. For example, lower and upper limits, L and U, may be selected
to include a large percentage of the measurements and outside of which it
is desired to control the percentage of measurements to, say, less than
10 percent. If the data quality is not consistent with these limits, L
and U, then action is taken to correct the possible deficiency before
future field tests are performed and to correct the previous data when
possible.
4.3.1 Estimating the Precision/Accuracy of the Reported Data
Methods for estimating the precision (standard deviation) and accuracy
(bias) of the dry molecular weight measurements and the percent C0~ by
volume were given in section 4.1. This section will indicate how the
audit data collected in accordance with the procedure described in
section 4.2 will be utilized to estimate the precision and accuracy of
the two measures of interest. Similar techniques can also be used by
a specific firm or team to assess their own measurements. However, in
this case no bias data among firms can be obtained. Two sets of audit
data will be collected as a result of following the procedures in the
previous section. They are:
d. = M, - M,
j j j
and
D. = %C00 - %C00
j 2. 2a.
In practice, it may be decided to collect data on D. only in pertinent
circumstances because the errors in the determinations of M, are small.
d
These are differences between the field team results and the audited results
for the respective measurements. Let the means and standard deviations of
the differences d. and D. j=l, . . ,n be denoted by d, s, and D, S ,
respectively.
Thus
3 - E Vn'
1=1 J
47
-------
and
,1/2
n _ 2
(d - dr/(
J
n - 1)
Identical formulas replacing D for d are obtained for D and s_, respec-
tively. Now d and D are estimates of the biases in the measurements (i.e.,
relative to the audited value). Assuming the audited data to be unbiased,
the existence of a bias in the field data can be checked by the appropriate
t-test, i.e.,
d - 0
See ref. 7 for a discussion of the t-test.
If t is significantly large, say greater than the tabulated value of t
with n - 1 degrees of freedom, which is exceeded by chance only 5 percent
of the time, then the bias is considered to be real and some check should
be made for a possible cause of the bias. If t is not significantly large,
then the bias should be considered zero and the accuracy of the data is
acceptable.
The standard deviation s, (or s ) is a function of both the standard
deviation of the field measurements and of the audit measurements.
Assuming both the field and audited measurements are obtained using the
same type of Orsat analyzer and hence that the standard deviations are
expected to be the same, then s , is an estimate of v2 cr{M,} /r, where r
is the number of replications. Table 2, page 51, contains an example calcu-
lation of d and s, starting with the differences for a sample size of n = 7.
See the final report on the contract for further information concerning
this result.
Similarly, SD is an estimate of
/2 a{%C07>
~ .
These standard deviations can then be utilized to check the reasonableness
of the assumptions made in section 4.1 concerning o{%CO?} = cr{%0?} = 0.4 per-
cent (or 0.2 percent for the modified Orsat analyzer), and o(Md} = 0.14 lb/
Ib-mole. For example, the estimated standard deviation, s, may be directly
48
-------
checked against the assumed value, 0, by using the statistical test
procedure
f 2 '
a
2
where x /f is the value of a random variable having the chi-square distri-
2
bution with f = n - 1 degrees of freedom. If x /f is larger than the
tabulated value exceeded only 5 percent of the time, then it would be
concluded that the test procedure is yielding more variable results due to
faulty equipment or operational procedure. The values of s, and sn can be
2 2
used directly in the test given above, if a is replaced by 2cr , on the
assumption that the variance of the field measurements is equal to that for
the audited data. Thus,
2 2
X2 Sd SD
_ or t
* 2a {%CO,}/r 2a {M,}/r
2. d
The measured values should be reported along with the estimated biases,
standard deviations, the number of audits, n, and the total_number of
field tests, N, sampled (n £ N) . Estimates, i.e., s,, s , d, or D, that
are significantly different from the assumed population parameters should
be identified on the data sheet. For example, for M, , based on the data of
table 2, the results would be reported as M, = 30.4 (assumed), t_, = -0.2 lb/
o M^
Ib-mole, sd//2 = 0.406//2 = 0.29 Ib/lb-mole, n = 7, and N = 20.
2
The t-test and X -test described above, and in further detail in the final
report on this contract, are used to check on the biases and standard devi-
ations separately. In order to check on the overall data quality as
measured by the percent of measurement deviations outside prescribed limits,
it is necessary to use the approach described in subsection 4.3.2 below.
4.3.2 Sampling by Variables
Because the lot size (i.e., the number of field tests performed by a team
or laboratory during a particular time period, normally a calendar quarter)
is small, N = 20, and, consequently, the sample size is small, of the order
of n = 3 to 8, it is important to consider a sampling by variables approach
to assess the data quality with respect to prescribed limits. That is, it
is desired to make as much use of the data as possible. In the variables
approach, the means and standard deviations of the sample of n audits are
used in making a decision concerning the data quality.
49
-------
Some background concerning the assumptions and the methodology is repeated
below for convenience. However, one is referred to one of a number of
publications having information on sampling by variables; e.g., see
refs. 8-13. The discussion below will be given in regard to the specific
problem herein which has some unique features as compared with the usual
variable sampling plans.
In the following discussion it is assumed that only %CC>2 and M, are
audited as directed in sections 4.2.2.1 and 4.2.2.2.
The difference between the team-measured and audited value of M, is
d
designated as d., and the mean difference over n audits by d, that is,
n
V /M, - M. \
£4. ( d. da.)
^j _ ]-l \ J .37
n
Theoretically, M_, and M, should be measures of the same molecular weight,
and their difference should have a mean of zero on the average. In
addition, this difference should have a standard deviation equal to v2
times that associated with measurements of ML or M, . Recall from the
d da
variance analysis that the difference of two such measurements would have
a standard deviation equal to V2 x 0.14 Ib/lb-mole (based on the average
of 3 replicates).
Assuming three standard deviation limits, the values -3(0.14/2) and
3(0.14/2) Ib/lb-mole define lower and upper limits, L and U, respectively,
outside of which it is desired to control the proportion of differences,
d.. Following the method given in ref. 11, a procedure for applying the
variables sampling plan is described below. Figures 7 and 8 illustrate
examples of satisfactory and unsatisfactory data quality with respect to
the prescribed limits L and U.
The variables sampling plan requires the sample mean difference, d; the
standard deviation of these differences, s,; and a constant, k, which is
d
determined by the value of p, the proportion of the differences outside
the limits of L and U. For example, if it is desired to control at 0.10,
the probability of not detecting lots with data quality p equal to 0.20
(or 20% of the individual differences outside L and U) and if the sample
size n = 7, then the value of k can be obtained from table II of ref. 11.
The values of d and s, are computed in the usual manner; see table 2
50
-------
Table 2. COMPUTATION OF MEAN DIFFERENCE, d, AND
STANDARD DEVIATION OF DIFFERENCES, s
General Formulas Specific Example
d = M - M
d da Data (Ib/lb-mole)
dl
d2
d3
d4
d5
d6
d7
0.4
-0.2
0.1
-0.8
-0.6
-0.3
-0.1
0.16
0.04
0.01
0.64
0.36
0.09
n.Ol
Ed. Ed2 -1.5, 1.31
J J
Ed.
d = J- d = - 0.214
n
2 (Zd,)2
Ed
2 j n 2 . -,,.
s, = >*p r^r s, = 0.165
d (n - 1) d
sd = Vsd sd = °-406
Table 3. SAMPLE PLAN CONSTANTS, k for P{not detecting a lot
with proportion p outside limits L and U} j£ 0.1
Sample Size n p = 0.2 p = 0.1
3 3.039 4.258
5 1.976 2.742
7 1.721 2.334
10 1.595 2.112
12 1.550 2.045
51
-------
Figure 7. Example Illustrating p < 0.10 and Satisfactory Data
Quality.
p (percent of measured
differences outside
limits L and U) > 0.10
Figure g> Example Illustrating p > 0.10 and Unsatisfactory Data Quality.
52
-------
for formulas and a specific example. Given the above information, the test
procedure is applied, and subsequent action is taken in accordance with the
following criteria:
(1) If both of the following conditions are satisfied:
d - k s, > L = -0.42/2"= -0.597 Ib/lb-mole
d
d + k s, < U = 0.42/2" = +0.597 Ib/lb-mole.
a
the individual differences are considered to be consis-
tent with the prescribed data quality limits and no
corrective action is required.
(2) If one or both of these inequalities is violated, possible
deficiencies exist in the measurement process as carried
out for that particular lot (group) of field tests. These
deficiencies should be identified and corrected before
future field tests are performed. Data corrections should
be made when possible, i.e., if a quantitative basis is
determined for correction.
Table 3 contains a few selected values of n, p, and k for convenient
reference.
Using the values of d and s, in table 2, k = 1.721 for a sample size n = 7,
and p = 0.20, the test criteria can be checked; i.e.,
d - k sd = -0.214 - 0.698 = -0.913 < L = -0.597
d + k sd = -0.214 + 0.698 = 0.484 < U = 0.597.
Therefore, both conditions are not satisfiedspecifically the lower limit
L was exceededand the lot of N = 20 measurements is not consistent with
the prescribed quality limits. The plan protects one from not detecting
lots with 20 percent or more defects (deviations falling outside the
designated limits L and U) with a risk of 0.10.
The procedure for auditing the differences, D-S is identical to the above
with D. substituted for d. throughout, the standard deviation of D.,
J J J
53
-------
/2 x 0.4 %C02 by volume, is substituted for /2 x 0.14 Ib/lb-mole, and
L and U become
3/2 x
%CO by volume,
where r is the number of replications. The above audit checks assume that
the audit and the field measurements are both determined with the standard
Orsat analyzer. If the audit is performed using the modified Orsat
analyzer, the audit data would be more precise and the results of the above
tests would also have to be modified. The final report on this contract
will consider these variations in the test procedure.
54
-------
1. William J. Mitchell. Qn-Site Collaborative Test of Method 3 of the
New Source Performance Standards Using Modified Orsat Apparatus.
Environmental Protection Agency, Research Triangle Park, N.C.,
November 1973.
2. William J. Mitchell. Qn-Site Collaborative Test of Method 3 of the
New Source Performance Standards at a Municipal Incinerator.
Environmental Protection Agency, Research Triangle Park, N.C.,
August 1973.
3. Walter S. Smith and D. James Grove. Stack Sampling Nomographs for
Field Estimations. Entropy Environmentalists, Inc., Research
Triangle Park, N.C., 1973.
4. Glossary and Tables for Statistical Quality Control. American Society
for Quality Control, Statistics Technical Committee. Milwaukee,
Wisconsin, 1973.
5. Eugene L. Grant and Richard S. Leavenworth. Statistical Quality
Control. 4th ed. St. Louis: McGraw-Hill, 1972.
6. A. Hald. Statistical Tables and Formulas. New York: John Wiley and
Sons, 1952.
7. H. Cramer. The Elements of Probability Theory. New York: John Wiley
and Sons, 1955.
8. C. Eisenhart, M. Hastay, and W. A. Wallis, eds. Techniques of
Statistical Analysis. Statistical Research Group, Columbia,
Univ. New York: McGraw-Hill, 1947.
9. A. H. Bowker and H. P. Goode. Sampling Inspection by Variables.
New York: McGraw-Hill, 1952.
10. A. Hald. Statistical Theory with Engineering Applications. New York:
John Wiley and Sons, 1952.
11. D. B. Owen. "Variables Sampling Plans Based on the Normal Distribution,.".
Technometrics 9, No. 3 (August 1967).
12. D. B. Owen. "Summary of Recent Work on Variables Acceptance Sampling
with Emphasis on Non-normality." Technometrics^ 11 (1969):631-37.
13. Kinji Takogi, "On Designing Unknown Sigma Sampling Plans Based on a
Wide Class on Non-Normal Distributions." Technometrics 14 (1972):
669-78.
55
-------
APPENDIX A
24886
METHOD 3GAS ANALYSIS FOR CARBON DIOXIDE,
EXCESS AIR, AND DHT MOLECULAR WEIGHT
1. Principle and applicability.
1 1 Principle. An Integrated or grab pis
sample Is extracted from a sampling point
find analyzed tor Its components using an
Orsat analyzer.
1 2 Applicability. This method should be
applied only when specified by the test pro-
cedures for determining compliance with the
New Source Performance Standards. The test
procedure will Indicate whether a grab sam-
ple or an Integrated sample Is to be used.
2 Apparatus.
21 Grab sample (Figure 3-1).
2.1.1 ProbeStainless eteel or Pyrex 1
glass, equipped with a filter to remove partlc-
ulate matter.
2.1.2 PumpOne-way squeeze bulb, or
equivalent, to transport gas sample to
analyzer.
1 Trade name.
RULES AND REGULATIONS
2 2 Integrated sample (Figure 3-2).
2 2.1 ProbeStainless steel or Pyrex l
glass, equipped with a filter to remove per-
tlculate matter.
222 Air-cooled condenser or equivalent
To remote any excess moisture.
2.2 3 Needle valveTo adjust flow rate.
224 PumpLeak-free, diaphragm type,
or equivalent, to pull gas.
2.2 5 Rate meterTo measure a flow
range from 0 to 0.035 cfm.
2 2.6 Flexible bagTedlar,' or equivalent,
with a capacity of 2 to 3 cu ft. Leak test the
bag In the laboratory before using.
2 2.7 Pltot tubeType S. or equivalent.
attached to the probe so that the sampling
now rale can be regulated proportional to
the stack gas velocity when velocity is vary-
ing with time or a sample traverse is
conducted.
2.3 Analysis.
2.3.1 Orsat analyzer, or equivalent.
PROBE
FLEXIBLE TUBING
TO ANALKZER
TEfilG
FILTER (GLASS WOOL)
SQUEEZE BULB
Rgure3-1. Grab-sampling train.
RATE METER
r\ ,- .
\lfiffl
VALVE
AIR-COOLED CONDENSER / PUMP
! WOBE
FILTEnlGLASSWOOU
QUICK DISCONNECT
RIGID CONTAINER'
Figure 3-2. Integrated gas sampling train.
3 Procedure.
3 1 Grab sampling
3 1.1 Set up the equipment as shown In
Figure 3-1, making sure all connections are
leak-free. Place the probe in the etack at a
sampling point and purge the sampling line.
312 Draw sample into the analyzer.
3.2 Integrated sampling.
3 2.1 Evacuate the flexible bag. Set up the
equipment as shown in Figure 3-2 with the
bag disconnected Place the probe in the
stack and purge the sampling line. Connect
the bag. making sure that all connections are
tight and that there arc no leaks.
322 Sample at a rate proportional to the
stack velocity.
3 3 Analysis
331 Determine the CO., O,. and CO con-
centrations as soon as possible Make as many
passes as are necessary to give constant read-
Ings If more than ten passes are necessary,
replace the absorbing solution.
332 For grab sampling, repeat the sam-
pling and analysis until three consecuthe
samples vary no more than 0 5 percent by
volume for each component being analyzed
3 3.3 Pox Integrated sampling, repeat the
analysis of the sample until three consecu-
tive analyses vary no more than 0.2 percent
by volume for each component being
analyzed.
4 Calculations
4.1 Carbon dioxide Average the three con-
secutive runs and report the result to the
nearest 0.1% COr
4.2 Excess air. Use Equation 3-1 to calcu-
late excess air, and average the runs. Report
the result to the nearest 0.1% excess air.
KA =
°\, 02)-0.->(%CO)
0.264("v N2)-(',i02)+0.5(%
equation 3-1
where:
e,?E\-- Percent excess air.
ToOar= Percent oxygen by volume, dry basis.
%Nj = Percent nitrogen by volume, dry
basis
%COPercent carbon monoxide by vol-
ume, dry basis.
0.264 = Ratio of oxygen to nitrogen In air
by volume.
4.3 Dry molecular weight. Use Equation
3-2 to calculate dry moVecular weight and
average the runs Report the result to th«
nearest tenth.
Md = 044(%COJ) +0.32(%0.)
4 028(%NJ+%CO)
equation 3-2
where
M«=Dry molecular weight, Ib./lb-mola.
%CO2=Percent carbon dioxide by volume,
dry basis.
%Oi=Percent oxygen by volume, dry
basis.
%Nj=Percent nitrogen by volume, dry
basis.
0.44=Molecular weight of carbon dloxld*
divided by 100'.
0.32=Molecular weight of oxygen divided
by 100.
0.28=Molecular weight of nitrogen and
CO divided by 100,
6. References.
Altshuller, A. P., et al , Storage of Gases
and Vapors In Plastic Bags, Int. J. Air &
Water Pollution, 6:75-81, 1963.
Conner, William D , and J. S Nader. Air
Sampling with Plastic Bags, Journal of the
American Industrial Hygiene Association,
25:291-297, May-June 1964
Devorkln, Howard, et al , Air Pollution
Source Testing Manual. Air Pollution Con-
trol District, Los Angeles, Calif , November
1963.
FEDERAL REGISTER, VOL. 36, NO. 247THURSDAY, DECEMBER 23, 1971
56
-------
APPENDIX B
GLOSSARY OF SYMBOLS
This is glossary of symbols as used in this document. Symbols used and
defined in the reference method (appendix A) are not repeated here.
SYMBOL
N
CV{X}
CV{X}
DEFINITION
Lot size , i.e., the number of field tests to be treated
as a group
Sample size for the quality audit (section IV)
Number of replicate analyses per field test
Assumed or known coefficient of variation (100 a /u )
.A. .A.
Computed coefficient of variation (100 S../X) from a
finite sample of measurements
Assumed standard deviation of the parameter X
(population standard deviation)
Computed standard deviation of a finite sample of
measurements (sample standard deviation)
Assumed mean value of the parameter X (population
mean)
X
Computed average of a finite sample of measurements
(sample mean)
Computed bias of the parameter X for a finite sample
(sample bias)
R
Range, i.e., the difference in the largest and smallest
values in r replicate analyses
x
Random error associated with the measurement of
parameter X
(V
The difference in the audit value and the value of
M, (%C02) arrived at by the field crew for the
. th , .
2 audit
d (D)
Mean difference between M (%CO ) and M, (%CO ) for
, . Q £, del 2.
n audits a
Sd (SD}
Computed standard deviation of difference between
Md (%C02) and Md (%C02 )
a. a
Percent of measurements outside specified limits L and U
57
-------
APPENDIX B
SYMBOL
P(Y}
t
*(e/a)
L
U
CL
LCL
UCL
GLOSSARY OF SYMBOLS (CONT'D)
DEFINITION
Constant used in sampling by variables (section IV)
A positive number used to calculate the required
number of replicate analyses such that the error in
estimating %C09 by the mean of r replicates will be
zt
less than e.
Probability of event Y occurring
Statistic used to determine if the sample bias, d, is
significantly different from zero (t-test)
Statistic used to determine if the sample variance,
2
s , is significantly different from the assumed
2
variance, a , of the parent distribution (chi-square
test)
Area under a standard normal curve to the left of
(or less than) the value e/o
Lower quality limit used in sampling by variables
Upper quality limit used in sampling by variables
Center line of a quality control chart
Lower control limit of a quality control chart
Upper control limit of a quality control chart
58
-------
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 measured value and the true
value.
The systematic or nonrandom component of system
error.
A specified number of objects to be treated as a
group.
A set of procedures for making a measurement.
The process of making a measurement including method,
personnel, equipment, and environmental conditions.
A very large number of like objects (i.e., measure-
ments, checks, etc.) from which the true mean and
standard deviation can be deduced with a high degree
of accuracy.
The degree of variation among measurements on a
homogeneous material under controlled conditions,
and usually expressed as a standard deviation or,
as is done here, 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.
59
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-650/4-74-005-b
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Guidelines for Development of A Quality Assurance
Program: Volume II - Gas Analysis for Carbon Dioxide,
Excess Air, and Dry Molecular Weight.
5. REPORT DATE
February 1974
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
Franklin Smith, Denny E, Wagoner, A. Carl Nelson, Jr.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
Research Triangle Park, N. C.
10. PROGRAM ELEMENT NO.
1HA327
27709
11. CONTRACT/GRANT NO.
68-02-1234
12. SPONSORING AGENCY NAME AND ADDRESS
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D. C. 20460
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Guidelines for the quality control of gas analysis for carbon dioxide, excess
air, and dry molecular weight by the Federal reference method are presented.
These include:
1. Good operating practices
2. Directions on how to assess performance and quality data
3. Directions on how to identify trouble and improve data quality.
4. Directions to permit design of auditing activities
The document is not a research report.
personnel.
It is designed for use by operating
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Quality Assurance
Quality Control
Air Pollution
Gas Sampling
Stack Gases
13H
14D
13B
14B
21B
13. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
68
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
60
-------
-------
-------
m
?
o
3)
H O O 3
2. » fl> ±
------- |