£EPA
United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
Research Triangle Park NC 27711
EPA/600/4-88-003
January 1988
Research and Development
A Summary of the
1986 EPA National
Performance Audit
Program on Source
Measurements
-------
EPA/600/4-88/003
January 1988
A SUMMARY OF THE 1986 EPA NATIONAL PERFORMANCE AUDIT PROGRAM
ON SOURCE MEASUREMENTS
by
E. W. Streib and W. J. Mitchell
Quality Assurance Division
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
Tl a JS^lroMieptal Protection
:,;Y, Library (5PL-16)
•'V. . ^arborn Street, Room 1670
•. L 60604
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental
Protection Agency policy and approved for publication. Mention of trade
names or commercial products does not constitute endorsement or recommenda-
tion for use.
-------
FOREWORD
Measurement and monitoring research efforts are designed to anticipate
potential environmental problems, to support regulatory actions by develop-
ing an in-depth understanding of the nature and processes that impact health
and the environment, to provide innovative means of monitoring compliance
with regulations, and to evaluate the effectiveness of health and environ-
mental protection regulations through the monitoring of long-term trends.
The Environmental Monitoring Systems Laboratory, Research Triangle Park,
North Carolina, has responsibility for: assessment of environmental moni-
toring technology and systems; implementation of Agency-wide quality assur-
ance programs for air pollution measurement systems; and supplying technical
support to other groups in the Environmental Protection Agency, including the
Office of Air and Radiation and the Office of Toxic Substances.
The major task of this study was to report the results of the national
quality assurance audit program for stationary source test methods. Audits
were designed to estimate the minimal analytical and computational accuracy
that can be expected with Method 3 (carbon dioxide and oxygen), Method 5 (dry
gas meter only), Method 6 (sulfur dioxide), Method 7 (nitrogen oxides), and
Method 19 (coal). Statistical analysis was used to characterize the data.
James W. Falco, Director
EnvironnentaI Monitoring Systems Laboratory
Research Triangle Park, North Carolina
i i i
-------
ABSTRACT
In 1986, the Quality Assurance Division conducted the National Audits
for Stationary Source Test Methods. The audit materials consisted of: a
disposable gas cylinder for Method 3 (Orsat analyzer), a calibrated orifice
for Method 5 (dry gas meter only), five simulated liquid samples each for
Method 6 (S02> and Method 7 (NOX), and two coal samples for Method 19. Par-
ticipating laboratories sent their data to the Source Branch and in return
received a written report comparing their results to EPA's.
In the Method 3 audit, each gas component had only one concentration.
The mean C02 value for all participants differed by 3.7 percent from the
expected (EPA) value, and the mean for 02 was 0.4 percent from the expected
vaIue.
In the Method 5 audit, the mean value for all participants differed by
3.9 percent from the expected value. For the Method 6 audit, the average
mean differed by 1.5 percent from the expected value. The average mean in
the Method 7 audit was 6.3 percent from the expected value.
In the two coal audits, the parameters measured were sulfur, moisture,
ash, and Btu content. On the average for the sulfur analysis, 91 percent of
the participants measured within 10 percent of the expected value; for Btu,
99 percent of the participants measured within 10 percent of the expected
vaIue.
This report includes the results of the performance audits done during
the period from January to December 1986.
i v
-------
CONTENTS
Page
Foreword •• iii
Abstract i v
F i gures v i
Tables vi i
Acknowledgment ix
1. Introduction 1
2. Summary » 2
3. Method 3 Audit 6
4. Method 5 Dry Gas Meter Audit 8
5. Method 6 Audit 13
6. Method 7 Audit '. 17
7. Method 19 Audit 21
References • • 25
Append i x • 26
-------
FIGURES
Number page
1 Cumulative accuracy for participants in
Method 5 audit, 0686 10
2 Previous results of Method 5 audit 11
3 Results of Method 5 audit, 0686 12
4 SC>2 expected values compared to participants'
values 15
5 Percent difference of mean values and confidence
intervals from expected values 16
6 NOX expected values compared to participants'
values 19
7 Percent difference of mean values and confidence
intervals from expected values. * 20
v i
-------
TABLES
Number
1 Participants' Results from Method 3 Audit
(All Data—No Out I iers Removed) 2
2 Method 5 Audit - Summary Statistics 3
3 Participants' Results from Methods 6 and 7 Audits
(All Data—No Out Iiers Removed) 3
4 Participants' Results from Methods 6 and 7 Audits
(Out Ii ers Removed) 4
5 Participants' Results from Method 19 Coal Audits
(A I I Data—No Out I iers Removed) 5
6 Method 3 Audit Participants 6
7 Source Method 3 Audit 7
8 Method 5 Audit Participants 8
9 Results of Method 5 Audit, 1982-1986 9
10 Method 6 Audit Participants 13
11 Summary of Source S02 Audit 14
12 Method 7 Audit Participants 17
13 Summary of Source NOX Audit 18
14 Coal Audit Participants 22
15 Source Coa I Aud i t—0386 23
16 Source Coal Audit—0986 24
17 National Orsat Audit Frequency Distribution of
Absolute Percent Differences of Expected and
Reported Values 27
18 DGM Frequency Distribution of Absolute Percent
Difference 28
v i i
-------
TABLES (Continued)
Number Page
19 SC>2 Frequency Distribution of Percent Difference—No
Out Ii ers Removed 29
20 NOX Frequency Distribution of Percent Difference—No
Outliers Removed 29
21 National Coal Audit Frequency Distribution of Absolute
Percent Differences of Expected and Reported Values—
Study 0386 30
22 National Coal Audit Frequency Distribution of Absolute
Percent Differences of Expected and Reported Values—
Study 0986 31
v i i i
-------
ACKNOWLEDGMENT
We express our appreciation to the laboratories that participated in
the National Performance Audit Program for Stationary Sources. Thanks also
to the staff of the Standards Laboratory/Performance Evaluation Branch/
Environmental Monitoring Systems Laboratory (EMSL), who did the acceptance
testing of the audit samples, and to the programmers of the Monitoring and
Assessment Division/EMSL for providing the data systems to store and evafuate
the data.
IX
-------
-------
SECTION 1
INTRODUCTION
The Environmental Protection Agency's (EPA's) Environmental Monitoring
Systems Laboratory (EMSL) at Research Triangle Park, North Carolina, estab-
lished an audit program in 1977 to evaluate the performance of companies that
conduct compliance testing using EPA reference methods. The audits check the
participants' analytical accuracy in applying the analytical phase of EPA
Reference Methods 3. 6, 7, and 19 and the. catibration accuracy of the Method
5 control console.('' Accuracy is defined as the percent difference between
a participant's analytical results and the EPA expected value. By partici-
pating in this free and voluntary program, testing companies can compare
their performance to the performance of other laboratories conducting similar
measurements.
Source Test Methods 3, 5, 6, and 7 were each audited once, and Method 19
was audited twice in 1986. Each participating laboratory received an audit
package consisting of the audit sample, a data card, instructions, and an
envelope for returning the data to EPA. A label for returning the audit de-
vice was included with the Method 5 audit package. Participants had 8 weeks
to use the audit material and return their data to EPA. At the end of this
period, all data received were statistically analyzed to determine the accu-
racy with respect to the EPA expected value and those obtained by the parti-
cipants (see Appendix).
The Quality Assurance Division of EMSL also maintains a limited reposi-
tory of samples for EPA Methods 3, 6, 7, and 19 that are available to source
testing laboratories for purposes such as training new personnel or conduct-
ing quality control checks. Because the expected values for these samples
are included with the analysis instructions, there is no requirement for the
data to be returned to EPA. We recommend that source testing laboratories use
this sample repository to help improve their overall analytical performance.
This report summarizes the results obtained in the 1986 source audits.
-------
SECTION 2
SUMMARY
In 1986, EPA's EMSL at Research Triangle Park, North Carolina, conducted
National Quality Assurance Audits for Stationary Source Test Methods 3 (Orsat
analyzer), 5 (dry gas meter only), 6 (S02), 7 (NOX), and 19 (coal). Industrial
laboratories, contractors, foreign laboratories, as well as local, state, and
federal agencies, participated.
The results of the 1986 audit of Method 3 are summarized in Table 1.
Participants analyzed the gas sample twice for percentages of carbon dioxide
(C02), oxygen (02), and carbon monoxide (CO). The mean values of C02 and 02
differed by 3.7 percent and 0.4 percent from the expected values, respectively.
The mean values for CO differed only by 4 percent from the expected values.
TABLE 1. PARTICIPANTS' RESULTS FROM METHOD 3 AUDIT
(ALL DATA—NO OUTLIERS REMOVED)
Type of
sarnp le
Smal i
cy 1 i nder
(gas)
Parameter
% C02
% 02
% CO
No. of
ana lyses
56
55
56
55
37
36
Repl i-
cate
1
2
1
2
1
2
EPA
true
va lue
6.00
6.00
10.00
10.00
10.00
10.00
Parti
Mean
5.80
5.78
10.03
10.04
9.61
9.60
cipant resu Its
Median
5.80
5.80
10.00
10.00
9.20
9.20
Std.
dev.
0.75
0.69
0.68
0.74
2.10
2.13
One audit of Method 5 was conducted in 1986. The overall results (no
outliers removed) are summarized in Table 2. The mean for all participants
was 3.9 percent from the expected value, and the standard deviation was 7.7
percent. After the removal of 5 percent of the data to eliminate statisti-
cal outliers, the mean was 2.8 percent from the expected values and the
standard deviation was 2.2 percent. The participants' performance based on
the standard deviation of all data was good and consistent with previous
years.(2,3,4,5)
-------
TABLE 2. METHOD 5 AUDIT - SUMMARY STATISTICS
Mean
Med i an
n
AlI data
Outliers removed
552
526
3.9
2.8
2.5
2.4
Std dev.
7.7
2.2
Table 3 represents the data (no outliers removed) from the Methods 6 and
7 audits. In the Method 6 audit, the procedure requires the participants to
determine the sulfate content in five aqueous solutions using the titration
procedure. For all five concentrations, the mean of the combined partici-
pants' results was less than 2 percent from the expected value, and the median
median differed by less than 1 percent. In the Method 6 Audit, 82 to 39 per-
cent of the participants achieved an accuracy within 5 percent of the expect-
ed value.
The Method 7 audit procedure requires that the participants determine
the nitrate content in five aqueous solutions. For four of the five concen-
trations, the mean of the combined participants' results was less than 5
percent from the expected value, and the median differed by 3 percent. In
this audit, 70 percent of the participants achieved an accuracy within 10
percent of the expected value for four out of five concentrations.
TABLE 3. PARTICIPANTS' RESULTS FROM METHODS 6 AND 7 AUDITS
(ALL DATA—NO OUTLIERS REMOVED)
Type of
samp le
Aqueous
su 1 fate
Aqueous
n itrate
No. of
ana lyses
90
92
91
93
91
70
69
70
70
67
EPA true
va 1 ue
(mg/DSCM)
350.8
846.5
907.5
1151.5
1227.8
130.0
350.0
390.0
555.0
620.0
Partic
Mean
356.1
853.4
903.3
1148.3
1225.9
124.9
327.9
374.5
529.7
596.9
ipant resu
Med ian
350.0
844.9
904.3
1142.4
1217.0
129.0
348.8
388.0
557.0
636.0
Its
Std.
dev.
52.5
85.1
95.5
85.1
190.6
32.5
72.0
72.4
120.4
169.4
-------
Table 4 represents the data from Methods 6 and 7 after the removal of
outliers. Four percent of the Method 6 data and 6 percent of the Method 7
data were classified as outliers. However, even before the statistical out-
liers were removed, a great improvement was seen in this year's Method 6 data
(based on the standard deviation) over last year's data.(5)
TABLE 4. PARTICIPANTS' RESULTS FROM METHODS 6 AND 7 AUDITS
(OUTLIERS REMOVED)
Type of
sarnp 1 e
Aqueous
su 1 fate
Aqueous
n itrate
No. of
ana lyses
88
90
89
89
88
67
67
67
69
63
EPA true value
(mq/DSCM)
350.8
846.5
907.5
1151.5
1227.8
130.0
350.0
390.0
555.0
620.0
Parti
Mean
349.6
842.7
901.6
1138.4
1217.9
126.1
335.9
380.3
537.3
615.5
cipant resu I
Med ian
349.9
844.5
904.3
1142.0
1216.7
129.2
348.9
389.0
557.0
638.6
Its
Std.
dev.
21.9
35.2
51.0
36.9
60.3
20.7
54.5
44.7
102.8
106.8
Table 5 summarizes the results of the two coal audits that were con-
ducted in 1986. Participants analyzed each coal sample in duplicate for
percentages of sulfur, moisture, and ash, and for gross calorific value
(Btu/lb). The means of the combined ash and combined sulfur contents were
within 10 percent of the expected value; whereas, moisture values differed by
as much as 37 percent. An accuracy of one percent was achieved on the mean
of combined Btu content. The Btu content as measured by the participants is
lower than the EPA value because coal oxidation is not prevented in the audit
sample. The Btu content results are a good example of the accuracy with
which a measurement can be made when that measurement receives good labora-
tory quality control.
-------
TABLE 5. PARTICIPANTS' RESULTS FROM METHOD 19 COAL AUDITS
(ALL DATA—NO OUTLIERS REMOVED)
Audit
date
0386
0986
0386
0986
0386
0986
0386
0986
0386
0986
0386
0986
0386
0986
0386
0986
No. of
Parameter analyses
% S 130
135
130
135
% H20 131
135
131
135
% Ash 130
134
130
134
Btu/lb 126
131
126
131
EPA
va I ue
1.72
1.26
3.54
3.02
1.04
1.89
3.93
2.11
11.65
9.73
17.17
12.29
12412.0
13271.0
13684.0
13578.0
Mean
1.78
1.39
3.51
3.00
1.06
2.59
3.54
2.80
11.69
9.63
17.03
12.20
12299.2
13243.2
13633.4
13553.3
Participant resu
Med ian
1 .70
1.30
3.53
2.99
1.08
1.84
3.58
2.08
11.65
9.61
17.06
12.27
12326.5
13266.0
13658.5
13576.0
Its
Std.
dev.
0.88
0.74
0.25
0.25
0.26
0.24
0.52
8.19
0.33
0.31
0.27
0.87
201.4
196.6
144.0
132.7
-------
SECTION 3
METHOD 3 AUDIT
The Method 3 audit checks the participants' abilities to analyze a gas
sample using an Orsat analyzer. The audit package consists of a disposable
cylinder that contains a 4-liter (L) sample of C02, 02, and CO. The analyst
expels the gas into the Orsat analyzer using the positive pressure of the
cylinder. The gas sample is quantitatively analyzed for percentage of C02,
02, and CO.
In the 1986 audit, 55 percent of the 93 laboratories receiving the audit
package returned data. Table 6 shows the total number of laboratories re-
questing participation and the number that returned data for the Method 3
audit.
TABLE 6. METHOD 3 AUDIT PARTICIPANTS
No. receiving No. returning
Category samp les data
Contractors
I ndustry
Foreign
Federal
State
Loca 1
50
25
2
2
11
3
24
15
2
1
6
3
Total 93 51
Table 7 summarizes the Method 3 audit results. Each laboratory was
asked to analyze the sample in duplicate. Five and ten percent accuracies
were chosen for the reporting criteria for each of the parameters. Each
parameter had only one concentration.
In the 1986 audit, 64 percent of the reporting laboratories achieved an
accuracy within 5 percent for the C02, which was an increase from the last
audit.(5) Eighty-five percent of the laboratories achieved an accuracy with-
in 5 percent of the expected value for the 02 analysis. For the CO analysis,
only 30 percent of the laboratories achieved an accuracy within 5 percent,
and 34 percent of the laboratories did not report a value for CO.
-------
TABLE 7. SOURCE METHOD 3 AUDIT
Expected
value
6.00
10.00
10.00
No.
anal
(1)
(2)
(1)
(2)
(1)
(2)
of
yses*
56
55
56
55
37
36
Laborator i es
accurate within 5«
(%}
61
64
82
85
32
28
Laboratories
accurate within 10$
(?)
89
87
96
91
70
69
*Numbers in parentheses indicate first and second analyses.
-------
SECTION 4
METHOD 5 DRY GAS METER AUDIT
In the Method 5 audit procedure, participants use a calibrated orifice
to check the calibration of the dry gas meter in their EPA Method 5 control
console (meter box). They insert the orifice in the Method 5 meter box, al-
low the box to warm up, and then make three 15-min volume measurements. The
participants convert each of the three volumes to cubic meters at standard
conditions using the formula specified in Eq. 5.1 of Method 5 (Appendix A, 40
CFR 60) and record them on the data card. Then they return the orifice and
the data card to EPA, where the data undergoes statistical analysis.
In the 1986 audit, 63 percent of the 141 laboratories that received the
audit package returned data. Table 8 shows the categories of participants,
the number of participants who requested participation in the Method 5 audit'
and the number who actually returned data.
TABLE 8. METHOD 5 AUDIT PARTICIPANTS
No. receiving No. returning
Category samples data
Contractors
Industry
Foreign
Federal
State
Local
74
41
6
2
13
5
38
30
5
2
9
4
Total 141 88
Figure 1, a cumulative histogram, shows the accuracy obtained by parti-
cipants in the Method 5 audit, expressed as the percent difference from the
expected (EPA) value at various levels of accuracy. The Code of Federal
Regulations(1) requires that the dry gas meter be calibrated within an accu-
racy of 2 percent. Figure 2 shows that 41 percent of the reporting labora-
tories attained this accuracy. Table 9 summarizes Method 5 audits from 1982
to 1986 and shows an improvement in the mean value and standard deviation.
8
-------
TABLE 9. RESULTS OF METHOD 5 AUDIT, 1982-1986
Aud it
Number of
anal yses
Mean Median
(% from EPA value)
Std.
dev.
0382
0982
0383
0983
0684
0685
0686
827
769
763
614
631
633
552
7.6
12.5
5.7
4.1
3.1
10.6
3.9
2.5
2.5
2.2
1.9
1.9
2.2
2.5
39.5
81.4
32
21
4.9
102.1
7.7
The histogram in Figure 3 depicts the individual results from all parti-
cipants of the 1986 audit and includes the mean and median values. Most
laboratories reported values higher than the EPA value.
The standard deviation of each laboratory's triplicate analysis (repeat-
ability) indicated that 76 percent of the standard deviations for each set
were within 0.3 percent. Four percent of the 1986 data were identified as
outliers using Chauvenet's Criterion.(&)• Before the outliers were removed,
the mean value (absolute) differed by 3.9 percent from the expected value.
After deletion of outliers, this value was reduced to 2.8 percent.
-------
flN
e
a:
100
90
80
70
60
50
40
30
20
10
_ n=552
A
X
/
'/
-------
IL
31
(O
c
N>
•
-o
-j
CD
^—•
CO
CO
c_
rt-
CO
OJ
Ol
U\BORATORIES, percent
• §
Ol
Ol
a
s
o>
CJ
g
00
CJ
CD
|l
O
0>
00
Ol
o
Q.
Q
C
Q.
CD
OJ
OJ
O
O>
00
O>
Ol
ro
o
s
I
-------
(••0
20
15
I
E 10
i
NUMBER OF VOLUMES-552
mecm=0.6
OUTLIERS
REMOVED
medfan«1.1
OUTLIERS
REMOVED
<-11-10 -8 -6-4-2 02 4 6 8
DIFFERENCE FROM EPA VALUE, percent
10 12 >13
Rgure 3. Results of Method 5 audit, 0686
-------
SECTION 5
METHOD 6 AUDIT
The Method 6 audit checks the participants' abilities to quantitatively
analyze Method 6 samples for sulfur dioxide content. The audit set consists
of five aqueous dilutions of 10 N sulfuric acid in 25-miI IiIiter (ml) sealed
glass ampoules. The analyst withdraws 5 mL from each ampoule, adds 30 ml of
3 percent hydrogen peroxide, and dilutes this sample to 100 ml with distilled
water. A 20-mL aliquot is then withdrawn from the diluted sample, 80 mL of
100 percent isopropanol and thorin indicator are added, and the sample is
titrated with barium perch I orate to a pink endpoint. In calculating the
results, the participants assume an original liquid sample volune of 100 nL
and a gas sample volume of 0.021 dry standard cubic meter (DSCM) of stack gas.
Table 10 shows the categories of the participants and compares the total
number of participants requesting participation with the number returning
data. In the 1986 audit, 68 percent of the 137 laboratories that received the
audit package returned data.
TABLE 10. METHOD 6 AUDIT PARTICIPANTS
No. receiving No. returning
Category samp I es data
Contractors
Industry
Foreign
Federal
State
Loca I
72
41
f
1
11
7
44
31
3
0
9
6
Total 137 93
Table 11 shows the percentage of laboratories that achieved 2 percent
and 5 percent accuracy for each of the five different concentrations in the
Method 6 audit. At least 81 percent of the reporting laboratories achieved
an accuracy within 5 percent for all five concentrations. Five percent is
used as the criterion because it was established as the criterion for source
S02 compliance audit samples (Section 4.4 of Method 6).
13
-------
TABLE 11. SUMMARY OF SOURCE S02 AUDITS
Concentration ±5 ±1
(mg/DSCM) (% of laboratories) (% of laboratories)
350.8
846.5
907.5
1151.5
1227.8
81.1
89.1
85.7
88.2
87.9
88.9
93.5
93.4
94.6
92.3
93
Figure 4 shows the mean values for each S02 concentration and the confi-
dence intervals (Cl). Each Cl is calculated using the 95 percent ^confidence
level. Figure 5 shows that the mean values and the Cl's are within the 5
percent criterion. Overall, the S02 audit with its smaller standard devia-
tion showed great improvement from last year's audit.(5)
14
-------
1500
o
(/)
o
1*1000
tf
i
CL
O
500
350.8 846.5 907.5 1151.5 1227.8
EXPECTED VALUES, mg/DSCM
Figure 4. SCJ2 expected values compared
to participants' values.
-------
Lj.-^T^'1EB't^*»j4w^l!-'Lll''^- ^"ff-
91
PERCENT DIFFERENCE, percent
o 31
O in C
CL CD
A
§01 W
o • cji
CO P
__. u 00
5 n
^ CD TQ 00
of rt 3 bi
^* ^» o
3 CD* <
33 > co
CD p. o
CD O m *
-g ° w
30-.
o -•» 3
2.3 \ en
"^ CD O r-
< Q (/) Cll
Q D g
C < _*
CD Q 10
s £
CO CJI
Q
Q.
1 1 1 1
» CJI -t^ O4 IO
(III
-». c
1
> -» NJ C*l -f^ CJI O
1 1 1 1
-------
SECTION 6
METHOD 7 AUDIT
The Method 7 audit checks the participants' ability to quantitatively
analyze Method 7 samples for nitrate content. The NOX audit set consists of
five aqueous dilutions of a potassium nitrate solution in 25-mL glass ampoules
that are autoclaved after sealing so that bacteria that might attack the ni-
trate are destroyed. The analyst withdraws 5 ml of solution from an ampoule,
adds this with 25 ml of Method 7 absorbing solution to a flask, adjusts the
pH with sodium hydroxide, and dilutes to 50 ml with distilled water. A 25-mL
aliquot is withdrawn from the diluted sample, placed in an evaporating dish,
and analyzed as described in Section 4.3 of Method 7.^) After this treatment
is completed, the absorbance is measured at 410 nanometers (nm) with a cali-
brated spectrophotometer. In calculating the concentrations present, the
participant assumes that 2000 ml of stack gas was sampled.
Table 12 shows the total number of laboratories requesting participation
and the number that returned data for the 1986 Method 7 audit. Sixty-four
percent of the 110 laboratories receiving the audit package returned data.
TABLE 12. METHOD 7 AUDIT PARTICIPANTS
No. receiving No. returning
Category samples data
Contractors 63 36
Industry 33 22
Foreign 2 2
Federal 1 0
State 5 5
LocaI 6 5
Total 110 70
17
-------
The percentage of laboratories that achieved 10 and 20 percent accuracy
for each of the five concentrations is shown in Table 13. Ten percent is
used as the criterion because it was established as the criterion for the
source NOX compliance audit samples (Section 4.4 of Method 7). Fifty-seven
percent of the reporting laboratories achieved an accuracy within 10 percent
on the lowest concentration, and 74 percent achieved an accuracy within 10
percent on the highest concentrations.
TABLE 13. SUMMARY OF SOURCE NOX AUDIT
Concentration
(mg/DSCM) (% of laboratories) (% of laboratories)
130.0
350.0
390.0
555.0
620.0
57.2
71.0
74.3
70.0
74.6
78.6
87.0
90.0
85.7
83.6
70
Figure 6 shows the means for each NOX concentration and the CIs. The
CIs are calculated in the same way as they were for S02- Figure 7 shows that
all mean values were lower than the true values. The standard deviations for
NOx were low which accounts for the narrow (±6.8$) Cl.
18
-------
/uu
^ cstn
O 60U
Q
<
O^
EC f\f\
500
«k
u
^ Aon
_J *rUU
^
^
tf\
P 300
Q_
^ 200
100
_ —
_ —
r
r
_ —
' i^™^^«^^^ L n
-
i
130.0 350.0 390.0 555.0 620.0
EXPECTED VALUES, mg/DSCM
Figure 6. N0^< expected values compared
to participants' values.
-------
130.0 350.0 390.0 555.0 620.0
EXPECTED VALUES, mg/DSCM
Rgure 7. Percent difference of mean values and
confidence intervals from expected values.
_k _
0 C
I
PERCEN
* I I
5 00 O) 4
T DIFFERENCE, percent
t CO O N3 -1*. O> 00 ON
I I 1
1 1 1 1 1
-------
SECTION 7
METHOD 19 AUDIT
Standards of performance for newer electric utility steam generators
(Subpart Da of 40 CFR 60) allow coal sampling and analysis to serve as an
acceptable method for determining sulfur concentration in the scrubber inlet
flue gas. The coa I audit checks participants' ability to analyze coal samples
for sulfur, ash, moisture, and Btu content.
The coal audit samples consisted of two samples each with 50 grams (g)
of 60-mesh coal but with different parameter levels. The following American
Society for Testing and Materials (ASTM) procedures^) were recommended, but
not required, for participants' use in analyzing the coal samples:
* ASTM D-3177 (Standard Test Method for Total Sulfur in the
Analysis of Coal and Coke)
» ASTM D-3174 (Standard Test Method for Ash in the Analysis
Sample of Coal and Coke)
* ASTM D-3173 (Test for Moisture in the Analysis Sample of
Coal)
* ASTM D-2015 (Standard Test Method for Gross Calorific Value
of Solid Fuel by the Adiabatic Bomb Method)
Participants measured the parameters and reported their results for
moisture ($) on an as-received basis, and their results for sulfur (.%), ash
($), and gross calorific value (Btu/lb) on a dry basis.
In both audits, 85 percent of the laboratories that received the audit
package returned data. Eighty-two of the same laboratories participated in
both audits and returned data. This was an increase from last year's audit.
Table 14 shows the total number of laboratories requesting participation and
the number that returned data for coal audits 0386 and 0986.
21
-------
TABLE 14. COAL AUDIT PARTICIPANTS
Category
Total
No. requesting samples
0386 0986
109
112
No. returning data
0386 0986
Contractors
1 ndustry
Federal
State
Local
50
42
0
11
6
51
46
0
11
4
39
38
0
11
5
43
42
0
11
4
93
100
Tables 15 and 16 summarize the coal audit results. The number of analy-
ses is greater than .the number of participants because some companies had
more than one laboratory participating. Each laboratory received its own set
of samples and was asked to analyze the samples in duplicate. Accuracies of
5 and 10 percent were chosen as the reporting criteria for each of the four
parameters.
In the 0386 audit, 90 percent of the laboratories analyzed both sulfur
samples and submitted results within 10 percent of the expected value. In
the 0986 audit, 84 percent of the laboratories analyzed the sulfur samples
within 10 percent of the expected values. Only 50 percent were within the
10 percent criterion for moisture in the 0386 audit, whereas 60 to 70 percent
in the 0986 audit achieved within the 10 percent criterion for moisture. For
the ash analysis and Btu content, 98 to 100 percent of the reporting labora-
tories submitted results within 10 percent of expected values for both sample
concentrations.
Comparison of the 0386 audit to the 0986 audit shows an improvement in
the latter audit for only the moisture parameters. In the 0986 audit, the
sulfur and ash levels were considerably lower, therefore allowing a greater
chance for error in compliance testing.
22
-------
TABLE 15. SOURCE COAL AUDIT—0386
Expected
value
1.72
3.54
1.04
3.93
11.65
17.17
12412
13684
Laboratories
accurate
No. of within 5%
analyses* (%)
(1)
(2)
(1)
(2)
(1)
(2)
(1)
(2)
(1)
(2)
(1)
(2)
(1)
C2)
(1)
(2)
Su 1 fur
130
126
130
126
Moisture
131
126
131
126
Ash
130
126
130
126
Gross Calorific
126
123
126
123
70.0
69.0
75.4
77.0
24.4
35.7
27.5
25.4
94.6
92.1
97.7
96.8
96.8
95.9
99.2
99.2
Laborator ies
accurate
within 10$
(%)
90.8
92.9
94.6
93.7
48.9
53.2
53.4
50.0
100.0
93.4
100.0
100.0
100.0
100.0
100.0
100.0
^Numbers in parentheses indicate first and second analyses.
23
-------
TABLE 16. SOURCE COAL AUDIT—0986
Laboratories Laboratories
accurate accurate
Expected No. of within 5% within 10$
val ue ana lyses* (%) (%)
Sulfur
1.26 (1) 135 63.0 88.9
(2) 128 60.2 84.3
3.02 (1) 135 79.3 93.3
(2) 128 81.3 90.6
Moi sture
1.89 (1) 135 44.4 68.9
(2) 127 44.9 69.3
2.11 (1) 135 28.1 62.2
(2) 127 28.3 61.4
Ash
9.73 (1) 134 95.5 99.3
(2) 126 96.8 99.2
12.29 (1) 134 94.0 98.5
(2) 126 96.0 98.4
Gross Calorific
13271 (1) 131 98.5 99.2
(2) 125 99.2 100.0
13578 (1) 131 98.5 100.0
(2) 125 99.2 100.0
*Nunibers in parentheses indicate first and second analyses.
24
-------
REFERENCES
1. U.S. Environmental Protection Agency. Standards of Performance for New
Stationary Sources—Appendix A. Title 40, Part 60, Code of Federal
ReguI at ions.
2. Streib, E.W., and M.R. Midgett. A Summary of the 1982 EPA National
Performance Audit Program on Source Measurements. EPA-600/4-83-049,
U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina, Apr!I 1984.
3. Streib, E.W., T.J. Logan, and M.R. Midgett. A Summary of the 1983 EPA
National Performance Audit Program on Source Measurements. EPA 600/4-
85-004, U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina, January 1985. 38 pp.
4. Streib, E.W., T.J. Logan, and M.R. Midgett. A Summary of the 1984 EPA
National Performance Audit Program on Source Measurements. EPA 600/4-
86-005, U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina, February 1986. 34 pp.
5. Streib, E.W., T.J. Logan, and M.R. Midgett. A Summary of the 1985 EPA
National Performance Audit Program on Source Measurements. EPA 600/4-
87-001, U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina, February 1986. 37 pp.
6. Chauvenet, W. Manual of Spherical and Practical Astronomy: Volume II —
Theory and Use of Astronomical Instruments (Method of Least Squares).
J.P. Lippincott and Co., Philadelphia, Pennsylvania, 1863.
7. American Society for Testing and Materials. Annual Book of ASTM Stan-
dards—1984. Volume 5.05. 01-050584-13, American Society for Testing
and Materials, Philadelphia, Pennsylvania.
25
-------
APPENDIX
FREQUENCY DISTRIBUTIONS
26
-------
TABLE 17. NATIONAL ORSAT AUDIT FREQUENCY DISTRIBUTION OF ABSOLUTE
PERCENT DIFFERENCES OF EXPECTED AND REPORTED VALUES
Sample Std-
no. No. Min. 10% __jQj...30ig 40% 50% 60% 10% 80% 90% Max. Mean dev.
C02
6000 111 0.00 0.00 1.67 1.67 3.33 3.33 5.00 6.67 8.33 11.67 70.00 6.68 10.53
02
6000 111 0.00 0.00 0.00 1.00 1.00 2.00 2.00 3.00 4.00 8.00 44.00 3.33 6.21
CO
6000 73 0.00 1.00 2.00 4.00 7.00 8.00 9.00 10.00 16.00 19.00 110.00 11.77 17.82
-------
TABLE 18. DGM FREQUENCY DISTRIBUTION OF ABSOLUTE PERCENT DIFFERENCE
Samp le
no.
552
549
539
co
533
530
528
526
Win.
0.0
0.0
0.0
0.0
0.0
0.0
0.0
10$
0.3
0.3
0.3
0.3
0.3
0.3
0.3
20$
0.7
0.7
0.7
0.7
0.7
0.7
0.7
30$
1.2
1.2
1.1
1.1
1.1
1.1
1.1
40$
2.0
2.0
1.9
1.8
1.8
1.8
1.8
50$
2.5
2.5
2.5
2.4
2.4
2.4
2.4
60$
3.0
3.0
2.9
2.9
2.9
2.9
2.8
70$
3.8
3.8
3.7
3.5
3.5
3.5
3.5
80$
5.0
4.8
4.6
4.5
4.4
4.4
4.4
90$
7.2
7.2
6.7
6.4
6.3
6.2
6.0
Max.
93.6
25.4
13.5
11.4
10.7
10.3
9.7
Mean
3.9
3.4
3.0
2.9
2.8
2.8
2.8
Std.
dev.
7.7
3.9
2.6
2.4
2.3
2.3
2.2
Skewness
9.05
-0.03
-0.18
-0.27
-0.36
-0.45
-0.53
Med i an
2.5
2.5
2.5
2.4
2.4
2.4
2.4
-------
TABLE 19. S02 FREQUENCY DISTRIBUTION OF PERCENT DIFFERENCE—NO OUTLIERS REMOVED
Samp le
no«
1
4
5
6
8
No.
91
91
93
92
90
Mi n.
0.00
0.07
0.10
0.00
0.00
10$
0.23
0.39
0.30
0.18
0.26
20$
0.44
0.64
0.46
0.30
0.63
30$
0.83
0.94
0.83
0.72
0.86
40$
1.15
1.13
1.00
0.82
1.20
50$
1.49
1.37
1 .38
1.02
1.57
60$
1.99
1.69
1.79
1.62
2.28
70$
2.59
2.40
2.20
2.07
3.36
80$
3.14
3.22
3.03
3.01
4.33
90$
5.65
5.88
6.75
5.38
10.09
Max.
67.91
91.08
43.78
80.38
121.27
Mean
4.17
5.21
3.37
3.51
5.10
Std.
dev.
9.67
14.61
6.57
9.44
14.13
TABLE 20. NOX FREQUENCY DISTRIBUTION OF PERCENT DIFFERENCE-NO OUTLIERS REMOVED
Sanp 1 e
no.
2
3
5
7
9
No.
67
70
70
69
70
Mi n.
0.00
0.09
0.00
0.03
0.00
10$
1.00
0.68
0.62
0.31
0.38
20$
2.42
1.26
2.31
0.86
1.33
30$
3.08
2.43
3.08
1 .69
2.31
40$
3.85
3.60
4.62
2.86
2.77
50$
5.23
5.05
8.69
4.43
3.85
60$
5.97
6.47
11.54
6.74
5.67
70$
7.74
9.50
15.23
8.94
8.18
80$
12.95
13.75
20.08
12.60
12.92
90$
40.97
40.36
29.69
42.29
9.74
Max.
99.52
99.44
103.31
99.51
99.49
Mean
14.30
12.03
14.57
11.43
9.87
Std.
dev.
23.52
18.56
20.63
18.19
16.18
-------
TABLE 21. NATIONAL COAL AUDIT FREQUENCY DISTRIBUTION OF ABSOLUTE PERCENT
DIFFERENCES OF EXPECTED AND REPORTED VALUES - STUDY 0386
Samp le
no.
5000
7000
g 5000
7000
5000
7000
5000
7000
No.
256
256
257
257
256
256
249
249
Min.
0.00
0.00
0.00
0.25
0.00
0.00
0.01
0.0
10$
0.58
0.28
1.92
1.53
0.34
0.12
0.05
0.16
20$
1.16
1.13
3.85
3.56
0.77
0.35
0.12
0.35
30$
1.74
1.41
4.81
5.85
1.12
0.52
0.21
0.49
40$
2.91
1.69
7.69
7.89
1.46
0.76
0.32
0.60
50$ 60$
Su I fur
4.07 4.07
2.54 3.11
Moisture
9.62 12.50
9.67 11.45
Ash
1.80 2.32
0.99 1.28
Gross Ca lor i f ic
0.45 0.57
0.68 0.82
70$
5.23
3.95
15.38
15.52
2.83
1.57
0.69
0.97
80$
6.40
5.37
21.15
20.87
3.61
1.98
0.87
1.32
90$
8.72
7.34
31.73
26.72
4.38
2.74
1.38
2.89
Max.
579.07
57.34
138.46
69.47
13.65
6.64
6.52
8.79
Mean
8.73
3.99
14.91
12.65
2.28
1.32
0.65
1.15
Std.
dev.
50.84
6.16
18.93
11.05
1.88
1.22
0.83
1.44
-------
TABLE 22. NATIONAL COAL AUDIT FREQUENCY DISTRIBUTION OF ABSOLUTE PERCENT
DIFFERENCES OF EXPECTED AND REPORTED VALUES - STUDY 0986
Samp 1 e
no.
6000
8000
No.
263
263
Mi n.
0.00
0.00
10$
0.33
0.79
20$
0.99
1.59
30$
1.32
2.38
40$
1.99
3.17
50$
2.65
3.97
60$
Sulfur
3.31
4.76
70$
3.97
6.35
80$
4.97
7.94
90$
8.28
10.32
Max.
56.95
671.43
Mean
4.33
11.64
Std.
dev.
7.02
59.26
Moisture
6000
8000
6000
8000
262
262
260
260
0.00
0.00
0.00
0.00
0.53
1.90
0.24
0.31
1.59
3.79
0.57
0.51
3.17
5.21
0.90
0.92
4.23
6.16
1.14
1.13
5.82
8.06
1.38
1.64
7.94
9.95
Ash
1.63
1.95
Gross Cat or
6000
8000
256
256
0.00
0.00
0.07
0.04
0.14
0.08
0.19
0.13
0.24
0.18
0.30
0.23
0.37
0.27
10.05
11.85
1.95
2.36
i f ic
0.44
0.35
13.76
15.64
2.52
2.77
0.59
0.47
21.16
24.64
3.42
3.49
1.11
0.85
5032.28
4501 .90
75.67
28.06
14.18
7.81
51.32
45.16
2.35
2.00
0.54
0.44
440.37
383.51
6.82
2.64
1.08
0.87
------- |