United States     Atmospheric Research and Exposure
            Environmental Protection Assessment Laboratory
            Agency       Research Triangle Park NC 27711
                        EPA/600/3-89/005
                        January 1989
            Research and Development
£EPA
A Summary of the
1987 EPA National
Performance Audit
Program on Source
Measurements

-------
                                                EPA/600/3-89/005
                                                January 1989
A SUMMARY OF THE 198? EPA NATIONAL PERFORMANCE AUDIT  PROGRAM
                   ON SOURCE MEASUREMENTS
              E. W. Streib and W. J. Mitchell
                 Quality Assurance Division
        Research Triangle Park, North Carolina  27711
  ATMOSPHERIC RESEARCH AND EXPOSURE ASSESSMENT  LABORATORY
             OFFICE OF RESEARCH AND DEVELOPMENT
            U.S. ENVIRONMENTAL PROTECTION AGENCY
       RESEARCH TRIANGLE PARK, NORTH  CAROLINA   27711
                                    U S.  Environmental Protection Agency
                                    Eegion 5,  Library  (5PL-16)
                                    230 S. Doarborn Street, FiOom J-o/J
                                    Chicago,  IL   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.
                                       i i

-------
                                    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 Atmospheric Research and  Exposure Assessment Laboratory, Research Triangle
Park, North  Carolina,  has responsibility  for:   assessment  of  environmental
monitoring technology  and  systems;   implementation  of   Agency-wide  quality
assurance programs for air pollution measurement systems; and supplying tech-
nical support to other groups  in the Environmental  Protection Agency, includ-
ing 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.
                        Gary J. Foley, Acting Director
           Atmospheric Research and Exposure Assessment Laboratory
                    Research Triangle Park, North Carolina
                                    i i i

-------
                                   ABSTRACT
     In 198T, 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 for Method 6
(302) and Method 7  (NOX), and two coal  samples  for Method  19.   Participating
laboratories sent their data to the  Research and Monitoring Evaluation 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 CC>2, Og, and CO values for  all  participants differed by 2.6 percent,
1.6 percent, and 33 percent from the expected value, respectively.

     In the Method  5  audit,  the mean value for  all participants  differed by
3.1 percent  from the expected  value.   In  the  Methods  6 and  7  audits,  the
average mean  differed  by  0.7   percent  and 15.6  percent  from the  expected
value, respectively.

     In the two  coal audits,  the parameters measured were  sulfur,  moisture,
ash, and Btu content. On  the average, 91 percent of the participants measured
sulfur within 10 percent of the expected value and 98 percent of the partici-
pants measured  Btu  content  within  10  percent  of  the  expected  value.   The
corresponding results for ash  and moisture were 100 percent and  50 percent,
respectively.

     This report includes the  results  of the performance audits  done during
the period from January to December 1987.
                                      iv

-------
                                  CONTENTS
Foreword	  ill




Abstract	   iv




Figures	   vi




Tables	  vii




Acknowledgment 	   ix




     1.  Introduction 	    1




     2.  Summary 	    2




     3.  Method 3 Audit 	    5




     k.  Method 5 Dry Gas Meter Audit 	    7




     5.  Method 6 Audit 	   10




     6.  Method 7 Audit 	   12




     7.  Method 19 Audit 	   Ik




References	   18




Appendix	   19

-------
                                  FIGURES
Number                                                               Page

  1        Cumulative accuracy for participants in
           Method 5 audit 	     8

  2        Results of Method 5 audit 	     9

-------
                                    TABLES
Number                                                                 Page

  1       Participants'  Results from Method 3 Audit
          (All Data—No Outliers Removed) 	    2

  2       Participants'  Results from Methods 6 Audit
          (All Data—No Outliers Removed) 	    3

  3       Participants'  Results from Methods 7 Audit
          (All Data—No Outliers Removed) 	    3

  1*       Participants'  Results from Method 19 Coal Audits
          (All Data—No Outliers Removed) 	    U

  5       Method 3 Audit Participants	    5

  6       Source Method 3 Audit	    6

  1       Method 5 Audit Participants 	    7

  8       Method 6 Audit Participants 	   10

  9       Summary of Source S02 Audit 	   11

 10       Method 7 Audit Participants 	   12

 11       Summary of Source NO x Audit 	   13

 12       Coal Audit Participants 	   15

 13       Source Coal Audit—0387 	   l6

 Ik       Source Coal Audit—0987 	   17


Appendix

 A-l      National Orsat Audit Frequency Distribution of
          Absolute Percent Differences of Expected and
          Reported Values 	   20

 A-2      DGM Frequency Distribution of Absolute Percent
          Difference 	   22
                                    vii

-------
                             TABLES (Continued)
Number                                                                 Page

  A-3      S02 Frequency Distribution of Percent Difference—No
           Outliers Removed 	     21

  A-U      NOX Frequency Distribution of Percent Difference—No
           Outliers Removed 	     21

  A-5      National Coal Audit Frequency Distribution of Absolute
           Percent Differences of Expected and Reported Values—
           Study 038? 	     22

  A-6      National Coal Audit Frequency Distribution of Absolute
           Percent Differences of Expected and Reported Values—
           Study 0987 	     23
                                     Vlll

-------
                                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/Research  and Monitoring Evaluation
Branch/Atmospheric Research  and Exposure Assessment  Laboratory  (AREAL),  who
did the acceptance  testing of  the  audit  samples, and  to  the programmers  of
the Exposure  Assessment  Research  Division/AREAL  for  providing  the  data
systems to store and evaluate the data.
                                       IX

-------
                                  SECTION 1

                                 INTRODUCTION
     The Environmental Protection Agency's  (EPA's)  Atmospheric Research  and
Exposure Assessment  Laboratory  (AREAL)  at  Research  Triangle  Park,  North
Carolina, established 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 calibration accuracy
of the  Method   5  control  console.'1'   Accuracy  is  defined  as the  percent
difference between a  participant's  analytical results  and the EPA expected
value.  By participating 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  1987.  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.  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  accuracy achieved
with respect to the expected value and to provide the participants a means to
compare their performance  to  that of all the  laboratories  that participated
in the  audit.   This  latter  comparison  can be  done using  the  histograms in
Appendix A.

     The Quality Assurance Division of AREAL 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 _OL-iuct-
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  1987   audits  for
stationary sources.

-------
                                  SECTION 2

                                   SUMMARY
     In 198T, EPA's  AREAL 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 (302), 7  (NOX), and 19 (coal).  Industrial
laboratories, contractors, foreign laboratories, as well as  local,  state,  and
federal agencies, participated.

     The results  of the  1987 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  COp and  0%
differed by 2.6 percent and 1.6 percent from the  expected values, respectively.
The mean  values  for  CO  differed  by  33 percent   from  the  expected  values.
               TABLE 1.  PARTICIPANTS'  RESULTS FROM METHOD 3 AUDIT
                         (ALL DATA—NO OUTLIERS REMOVED)
Type of
sample Parameter
Small % C02
cylinder
(gas)
% 02

% CO

No. of
analyses
50
kQ

50
U8
37
35
Repli-
cate
1
2

1
2
1
2
EPA
true
value
7.00
7.00

11.00
11.00
1.00
1.00
Participant results
Mean
6.82
6.77

11.18
11.22
1.33
L.33
Median
6.80
6.80

11.00
11.10
1.00
1.00
Std.
dev.
0.65
0.65

1.02
1.07
2.71
2.80
     The mean for all participants  in the Method 5  flow audit was  3.1  percent
from the expected value, and  the  standard deviation was k.J percent.   After
the removal of 7  percent of  the data to  eliminate  statistical  outliers,  the
mean was 2.2 percent  from  the expected value and the  standard  deviation  was
1.9 percent.  The participants' performance  based  on the  standard  deviation
of all  data  was  good and   was   consistent  with   revious   ears. ^'->3,4 ,p J
                                                    previous  years.

-------
     Tables 2 and 3 present the data (no outliers removed)  from the Methods  6
and T audits,  respectively.   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 participants'  results  was  less than  2 percent  from the  expected
value, and the median differed "by less than 1 percent.

     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 was  greater  than 10 percent  from the expected value,  but
the median was less than 2 percent.
             TABLE 2.  PARTICIPANTS' RESULTS FROM METHOD 6 AUDIT
                       (ALL DATA—NO OUTLIERS REMOVED)

No. of
analyses
89
90
92
91
92
EPA true
value
(mg/DSCM)
259.3
289.8
617.7
678.7
1189.7
Participant results

Mean
260.3
295.6
621.0
67^.8
1190.3

Median
258.2
287.5
6l6.0
673.2
1182.5
Std.
. dev.
15.1
1*8.8
50.8
5U.U
89.8
             TABLE 3.  PARTICIPANTS' RESULTS FROM METHOD 7 AUDIT
                       (ALL DATA—NO OUTLIERS REMOVED)

No. of
analyses
62
62
6l
62
62
EPA true
value
(mg/DSCM)
99.5
125.1
329.9
590.0
61+9.5
Participant results

Mean
122.8
155-2
362.8
6l6.7
753.7

Median
97.6
122.9
331.3
590.2
636.9
Std.
dev.
228.0
262.5
279. U
397.6
1067.9

-------
     Table 4 summarizes the results  from the two coal audits that vere  con-
ducted in  1987.   Participants analyzed  each  coal  sample  in duplicate  for
percentages of  sulfur,  moisture, ash, and gross  calorific  value  (Btu/lb).
The means of the ash and sulfur results were within  6  percent of the  expected
value, and the mean  for  moisture was within 3 percent.  The  mean of the  Btu
results was within 1 percent  of  the expected value.   The Btu content as  mea-
sured by the participants is likely slightly 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 laboratory quality control.
          TABLE 4.  PARTICIPANTS' RESULTS FROM METHOD 19 COAL AUDITS
                    (ALL DATA—NO OUTLIERS REMOVED)
Participant results
Audit
date
038T
0987
0387
0987
0387
0987
0387
0987
0387
0987
0387
0987
0387
0987
0387
0987
No. of
Parameter analyses
% S 112
110
112
109
% H20 111
110
111
110
% Ash 112
110
112
110
Btu/lb 107
106
107
106
EPA
value
0.77
1.68
2.77
4.21
1.16
1.57
2.12
2.1+7
7.97
13.66
16.88
22.18
12392.00
11042.00
13464.00
12886.00
Mean
0.81
1.65
2.88
4.18
1.14
1.6l
2.18
2.1*0
7.94
13.66
16.93
22.10
12334.33
11129.27
13392.50
12861.51
Median
0.77
1.65
2.7^
4.14
1.21
1.67
2.27
2.1+5
7.96
13.70
16.93
22.11
12339.00
11052.00
13427.00
12847.00
Std.
dev.
0.1+0
0.10
1.45
0.26
0.21+
0.32
0.36
0.36
0.13
0.21
0.19
0.28
137.20
732.25
L51.34
470.82

-------
                                  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 U-liter (L) sample of  C02,  09, 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 1987 audit, 53  percent of the 9^ laboratories receiving the  audit
package returned data.  Table  5 shows the  total  number of laboratories  re-
questing participation and  the number that  returned data  for  the Method  3
audit.
                    TABLE 5.  METHOD 3 AUDIT PARTICIPANTS


                               No. receiving             No. returning
        Category	samples	data
Contractors
Industry
Foreign
Federal
State
Local
k6
31
2
2
10
3
20
19
2
1
6
?
        Total                       9^                         50
     Table 6 summarizes the  Method  3 audit results in terms  of the percent-
age of the  results within  5  and 10  percent  of  the  expected  value.   Each
laboratory was asked to analyze the sample in duplicate.   Fifty-eight percent
of the reporting laboratories achieved an accuracy within  5 percent  for C02,
and 80 percent  achieved an  accuracy within 5  percent for 02-  For the  CO
analysis, only 32 percent of the laboratories  achieved an accuracy  within 5
percent;  26  percent   of  the  laboratories  did  not report  a  value  for  CO.

-------
                      TABLE 6.  SOURCE METHOD 3 AUDIT


Parameter
-• ti / i • •
C02
02
CO


Expected
value
7.00
11.00
1.00


No. of
analyses*
(1) 50
(2) U8
(1) 50
(2) U8
(1) 37
(2) 35
Laboratories
accurate
within 5%
58
58
80
79
32
uo
Laboratories
accurate
within lOt
(*)
88
90
96
60
*Numbers in parentheses indicate first and} second analyses.

-------
                                  SECTION U

                         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, 1*0
CFR 60) a.nd record them on the data card.   Then they  return the orifice and
the data card to EPA.

     In the 1987 audit,  60  percent  of  the 1^9  laboratories that  received the
audit package returned  data.   Table  7 shows the  categories  of participants,
the number of participants who requested participation in the Method 5 audit,
and the number who actually returned data.  A total of 193 meter boxes  were
audited.
                    TABLE 7.  METHOD 5 AUDIT PARTICIPANTS


                                No. receiving        No. returning
            Category	orifice	data
Contractors
Industry
Foreign
Federal
State
Local
72
^7
5
3
16
6
39
33
5
2
7
3
            Total                   lU9                    89
     Figure 1  is a  cumulative  histogram that  shows  the percent  accuracy
obtained by participants in the Method 5 audit, expressed as the percent dif-
ference from the  expected  (EPA)  value.  The  Code  of Federal  Regulations^1'
requires that the dry  gas  meter be  calibrated within an accuracy of  2 per-
cent.  Figure 1  shows that  56  percent  of the  reporting laboratories  attained
this accuracy.  This is approximately the same as in 1986.

     The histogram in Figure 2 depicts the individual results from all parti-
cipants of the  1987 audit  and  includes  the  mean  and  median  values.   Most
laboratories reported values lower  than the EPA value  for  reasons  that are
unknown.

-------
   0)
   s-
   IO

    II
    c
O
O
O
0)
o
CO
                                          ,\\\\
o
N.
o
CO
o
10
o
CO
o
CM
                                                        00
                                                        CD
                                                        00
                                                            .c

                                                      CO     <2
                                                      I--     "c
                                                      5  c  §
                                                      y  co  a
                                                      r£  o  'n
                                                                  CO
                                                                  a
                                                      HI c£
                                                      O u


                                                      ^^
                                                      QC O
                                                      LU QC
                                                      U_ LL
                                                      LL

                                                      Q
                                                                  O
                                                                  CO

                                                                  D
                                                                  O
                                                                  O
                                                                  0
                                                                  CO "O
                                                                  •5 o
                                                                  D  0
                                                                  O  E
                                                            LU
                                                            QC
                                                            D

                                                            2
                                                            LJ_
                     XOVynOOV Q3ldlO3dS
               ONIA3IHOV S3iaoivaoavn
                                8

-------
2?
N 'll \X\\VV\\\\
• " J

co £
*w o
0
a
M UJ
D
_J
CM _*• •4—'
UJ 3
2 «
o O lO
nr
LL.
LL "D
LU r-
w 0 £
^ T rt^
I ^_ vU
UJ ^
LT ^
UJ H-
LL O
«t LL
1 Q «
CO
CD Q>
i tr
eg
oo C>
1 L-
v O)
Li.
D

-------
                                  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  mL  sealed  glass  am-
poules.  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 perchlorate to a pink endpoint.   In  calculating the  results, the par-
ticipants assume an original liquid sample volume of  100 mL and  a  gas sample
volume of 0.021 dry standard cubic  meter (DSCM) of stack  gas.

     Table 8 shows the categories of  the participants and compares the total
number of  participants  requesting participation with the  number  returning
data.  In the 1987 audit, 67 percent of the 137 laboratories  that received the
audit package returned data.


                    TABLE 8.  METHOD  6 AUDIT PARTICIPANTS
                              No. receiving        No. returning
              Category	samples	data
Contractors
Industry
Foreign
Federal
State
Local
69
k2
h
2
12
8
H3
32
3
0
8
6
              Total               137                   92
     Table 9  shows  the percentage  of laboratories  that achieved  5  and 10
percent accuracy for  each  of the  five different concentrations in the Method
6 audit.   At least  75 percent  of  the  reporting  laboratories achieved an
accuracy within 5 percent  for all  five concentrations.  This is the accuracy
criterion used  for  acceptable source  S02  compliance  audit  results (Section
h.k of Method 6).
                                      10

-------
               TABLE 9.  SUMMARY OF SOURCE SOg AUDITS
                        % Laboratories              % Laboratories
Concentration           accurate within             accurate within
  (mg/DSCM)                   ± 5%	        	± 10%
259.3
289.8
617.7
678.7
1189.7
75.3
81.1
93.5
86.8
90.2
95.5
93.3
98.9
95.6
97.8
            N                 92
                                  11

-------
                                  SECTION 6

                                METHOD 7 AUDIT
     The Method 7  audit  checks the  participants'  ability to  quantitatively
analyze Method T samples for nitrate content.  The NOX audit set  consists of
five aqueous dilutions of a potassium nitrate solution in glass ampoules  that
are autoclaved after sealing so that  bacteria that might attack  the  nitrate
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 U.3 of Method T.'^'   After  this treatment
is completed, the absorbance is measured at UlO nanometers  (nm) with  a cali-
brated spectrophotometer.  In calculating the concentrations, the  participant
assumes that 2000 mL of stack gas  was sampled.

     Table 10 shows the total number of laboratories  requesting participation
and the number that returned data  for the 1987 Method 7  audit.   Sixty  percent
of the 103 laboratories receiving  the audit package returned data.
                    TABLE 10.  METHOD 7 AUDIT PARTICIPANTS


                              No. receiving            No. returning
         Category	samples	data	

         Contractors               58                       30
         Industry                  30                       22
         Foreign                    3                        1
         Federal                    1                        0
         State                      5                        H
         Local                      6                        5

         Total                    103                       62
                                      12

-------
     The percentage of laboratories that achieved 10  and  20 percent accuracy
for each of the five  concentrations  is  shown in  Table 11.   Sixty-six percent
of the reporting laboratories achieved  an  accuracy  within 10 percent  on  the
lower concentrations, and 80 percent  achieved an accuracy within  10 percent
on the higher concentrations.  Ten percent is the minimum level of accuracy
for acceptable  NC>2  compliance  audit  samples  (Section k.k  of  Method  7).


                   TABLE 11.  SUMMARY OF SOURCE NOY  AUDIT
                             % Laboratories               % Laboratories
   Concentration             accurate within              accurate within
     (mg/DSCM)	± 5%	± 10%
99.5
125.1
329.9
590.0
6U9.5
66.1
66.1
78. T
79.0
82.3
77. U
82.3
88.5
90.3
90.3
                  N                              62
                                      13

-------
                                  SECTION 7

                               METHOD 19 AUDIT
     Standards of performance  for newer  electric utility  steam  generators
(Subpart Da of  UO  CFR 60)  allow coal sampling  and  analysis  to serve  as  an
acceptable method for determining sulfur concentration  in the  scrubber inlet
flue gas.  The coal 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-317^ (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,  8H percent  of  the laboratories that received  the audit
package returned data.  Seventy-nine  of the  same laboratories  participated in
both audits and  returned  data. This was  a  decrease  from last  year's audit.
Table 12 shows the total number  of laboratories  requesting  participation and
the number  that  returned  data  for  the  two  coal audits  (0387 and  0987).

-------
                      TABLE 12.  COAL AUDIT PARTICIPANTS
                        No. requesting samples        No. letarning data
     Category	038?     098?	038?     098?
Contractors
Industry
Federal
State
Local
55
38
0
12
5
52
kk
0
12
3
UU
35
0
9
h
1*5
39
0
9
2
     Total                  110      111                  92       95
     Tables 13 arid lU 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  target criteria  for  each of  the  four
parameters.

     Comparison of the  0387  audit  to  the 0987 audit shows  an  improvement in
the latter audit  for the moisture parameters and the  low  level sulfur  con-
centration.  In  the 0387  audit, the  sulfur level  was  considerably  lower,
therefore allowing a greater chance for error in compliance testing.
                                      15

-------
                      TABLE  13.   SOURCE  COAL  AUDIT—0387
Expected
value
0.77
2.11
1.16
2.12
1.91
16.88
12392
13U61*
No. of
analyses*
(l) 112
(2) 106
(1) 112
(2) 106
(1) 111
(2) 105
(1) 111
(2) 105
(1) 112
(2) 105
(1) 112
(2) 105
(1) 107
(2) 102
(1) 107
(2) 102
Laboratories
accurate
within 5%
(%}
Sulfur
55.3
56.6
7U.1
79.2
Moisture
21.6
19.0
18.0
20.0
Ash
99.1
99.0
100.0
99.0
Gross Calorific
99.1
99.0
99.1
99.0
Laboratories
accurate
within 10%
(*)
80. U
87.7
92.0
95.3
1+8.6
52.U
50.5
U3.8
99.1
100.0
100.0
100.0
100.0
100.0
100.0
100.0
*Numbers in parentheses indicate first and second analyses,
                                      16

-------
                      TABLE Ik.   SOURCE COAL AUDIT--0987
Expected
value
1.68
H.21
1.57
2.1+7
13.66
22.18
110U2
12886
No. of
analyses*
(1)
(2)
(1)
(2)
(1)
(2)
(1)
(2)
(1)
(2)
(1)
(2)
(1)
(2)
(1)
(2)
110
10U
109
103
110
101
110
101
110
103
110
103
106
101
106
101
Laboratories
accurate
within 5$
(*)
Sulfur
73
7U
79
77
Moisture
23
22
33
Hi
Ash
99
99
100
100
Gross Caloriiic
96
96
97
97
.6
.0
.8
.7
.6
.8
.6
.6
.1
.0
.0
.0
.2
.0
.2
.0
Laboratories
accurate
within 10f»
(%)
90
92
95
9H
ho
H7
66
69
99
100
100
100
98
98
98
98
.9
.3
,U
.2
.9
.5
.U
.3
.1
.0
.0
.0
.1
.0
.1
.0
*Numbers in parentheses indicate first and second analyses.
                                      17

-------
                                 REFERENCES
1.   U.S. Environmental Protection Agency.   Standards  of  Performance  for  New
     Stationary Sources—Appendix A.   Title ^0,  Part 60,  Code  of  Federal
     Regulations.

2.   Streib, E.W., T.J. Logan,  and M.R.  Midgett.   A Summary  of  the  1983 EPA
     National Performance Audit Program on  Source Measurements.  EPA 600A-
     85-00^, U.S.   Environmental  Protection  Agency,  Research  Triangle Park,
     North Carolina, January 1985.  38  pp.

3.   Streib, E.W. , T.J. Logan,  and M.R.  Midgett.   A Summary  of  the  1$>8U EPA
     National Performance Audit Program on  Source Measurements.  EPA 600/U-
     86-005, U.S.   Environmental  Protection  Agency,  Research  Triangle Park,
     North Carolina, February 1986.   3^  pp.

U.   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/i+-
     87-001, U.S.   Environmental  Protection  Agency,  Research  Triangle Park,
     North Carolina, February 1986.   37  pp.

5.   Streib, E.W.  and W.J. Mitchell.  A Summary of the 1986  EPA National
     Performance Audit  Program on  Source  Measurements.   EPA  600A-88/003,
     U.S. Environmental  Protection  Agency,  Research  Triangle  Park, North
     Carolina, January 1988.  31 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—-198H.   Volume  5.05.  01-05058^-13,  American  Society for  Testing
     and Materials, Philadelphia, Pennsylvania.
                                      18

-------
                                   APPENDIX

                           FREQUENCY DISTRIBUTIONS
     The following frequency distributions are provided to  allow  each parti-
cipating laboratory to  compare its performance  to the performance  of  other
laboratories that participated in the audit.  For  example,  Table  A-l (Orsat)
shows that  for  C02,  50 percent of  the  participants reported  results  within
U.29 percent of the true  value.   Similarly for C>2, 50 percent of the parti-
cipants reported results within 1.82 percent  of the true  value.
                                      19

-------



















H
EH

1—3
O CO
CO &H1
rO ^~|)
^3 t—\
<3J
p^i ^>
o
Q
S3 W

|— ) p^
E~* O
f^ pL(
pq w
M Pn
«
EH Q
CO S3
M  PL,
O? X
M EH"
OH
pr_^ [JH
o
EH
H CQ
Q W
& O

W
EH W
<: H
CQ fc
O M
O
j
<^ EH
S3 S3
0 W
M 0
EH K
< W
S3 OH


•
rH
I


w

j"O
2
EH
















• •
+3 0)
CO T3

C
Oj
0
s




•
X
s






"fe^-
o
ON





•65-
O
OO





"6^.
O





^^-
o
vo




•ftSS-
O
[f\




•*&-
0
J-



•&5.
0
00



"65.
o
OJ




"6^-
o
H



•
C
•H
*jr*

O
&


(D
H
t«
o
C!
CO
0 0 ON
00 t— 00
t— oo t-
vo
OJ
rH ON 00
t— ON OO
• • •
LA OO LA
VD



oo o o
J- 0 0
rH O 0
LA VD OO
VD
rH


oo oo o
-a- H o

iH OO O
rH t-



-H; LA O
rH LA O
• • •
t— -3- 0
-3-



H OO O
t- C— O
• • •
LA OJ O
00


H OJ O
t— OO O

LA rH O
OJ

OJ
O ON OJ OJ O O
O OJ O OO O O
• • •
-3- H 0
rH


VD oj o
CO CO O

OJ rH O
rH

VD H O
OO ON O
• • •
OJ O O


oo o o
J- O O
• • •
rH O O



O 0 0
o o o
• • •
0 O O



o o o
o o o
• • •
0 0 O

OO CO OJ
ON ON C—



O 0 O
o o o
o o o
t— t— t—
















w
o
S3

OH
w

ff^
1 — t
Q

EH
S3
W

OH
w
Pu

w
EH

[ "]
O
CQ
pq
H
0
S3
W
§
H
pcj
pr,

2
a


•
OJ
1

4J 0)
CO T3





C
oi
d)
s




•
X
Oi
s




"fe^-
o
o\



•55.
o
CO



•^5.
o
C" —



•a*.
O
VD


-gs.
O
LA



0
-=!•



"6^-
o
oo



S$J-
o
OJ



(§5.
0
H
B
C
•H
s


-------






@
0
2
te

CD
H
EH
t—~t
o
0
1

o
w

M
Q
EH

P3
O
w
fe
o
s
o
H
g
pq
1 — 1
EH
CD
Q

O
W
or
w
E
OJ
o
CD

.
on
i


d
EH







• .
73 |>
-P Q)
CD T3
C
03
2


cd
2


•6*.
O
ON


•feS.
O
00
^
O
t" —


"B^-
0
vo

•ss-
o
LA


•feS-
O
"*


•55.
o
on

•6^-
o
OJ


•^s.
O
1-1


•
c
•H
2
.
O
a
a;
rH
^ o
nj a
CD
CO VO rH rH OJ
-3- t— 0 O OJ
• • • • •
J" t— t— VO C—
i-H
rH LAOO t- On
t— t— C— LA LA
on cvi oj LA on
ON oo vo vo ON
oj o vo on oj
LA j- vo t— c~-
on t—vo oj LA
rH
J- O O ON H
j- o oj LA_=r
t— -d" -J- VO LA

-j- LA on o LA
OJ on oj ONVO
LA on on j- on
vo ONVO t— LA
on J- LA o o
j- CM oj -d- on


vo LACO j- t—
on co o oj on
on H oi on oj

OJ LAVO CO OJ
oo oj vo -=t r—
OJ H rH OJ rH


t— J- J- ON t—
t— O OJ LA_3-
rH rH H rH rH


t— LA OJ H ON
oj vo co on ON
rH O O rH O

ON o ON OJ on
oo -J- on t— vo
o o o o o


ON oo ONOO LA
on oj H OJ oj
o o o o o


O LA O O O
o o o o o
O 0 O 0 0

ON OJ OJ O rH
co ON ON ON ON


rH OJ On -3" LTN











Q
W
1
OH

CD
W
H
EH
O

O
1

u
S3
W

M
Q
EH

P3
O
cu
Cn
O
£q
0
H
EH

«
HH
EH
CD
O
K^j
?H
o
Of
w
f§
X
o
a

•
i
"^
w
9
EH








• ,
•o >
CD 13



C
cd
OJ
2

.
X
o5
2



%«-
o
ON



•feS-
o
OO

•6S.
t—




VO



IPv




•fcS-
o



•6*.
oo


OJ




H

•
c
J
•ri

o

0)
H
ft •
8 0
oj C
CD
ON ON OJ CO (•—
on t— -3- OJ ON
VO t— OJ LA OJ
OJ O VO VO CO
OJ OJ H
vo onoo OJ -*
O t— ON OJ IA
• • • • •
OJ t— ON t— ON
-^ on OJ H H

LA H VO O\ LA
J- VO O VO VO
ON OJ H rH OJ
oo J- oo o -3-
t— VO OJ LAVO
rH H rH

ON OJ vo vo t—
O\OO OO J- VO
OO LAOO t— ON
on oj H H H

O LA rH VO t—
VO H -4- OO VO
O -J" ON ON ON
OJ H
LA on -d- ONOO
OO OJ O LACO
O O VO LA J"
rH rH

J- ON LA OJ O
on ONOO -3- t—
co t— -=f on on

on oi oi t— vo
ON t— OO rH t—
LA j- on oj oj



OJ O VO VO ON
LA vo t— t— on
on on oj H oj


rH J- VO VO ON
on CM o on j-
OJ OJ OJ H rH
rH CO VO LACO
t— OJ J- t— CO
rH rH rH O O


O o ON t- OJ
LACO VO -3- -3-
o o o o o

O CO CO CO O
o o o o o
O O 0 0 0

OJ OJ OJ OJ rH
vo vo vo vo vo



-J LAVO CO ON


21

-------








1
o

H
IX C—
OO
W 00
EH 0
ID
. "I ^|
0 Q
CO D
W EH

PQ
M Q
« w
EH EH
CO K
M O
Q Pn
pgj]
>H K
O

w £23
^ ^
0?
pq Q
PH W
Pn EH
U
EH H
M 0,
£3 W
^*
J 0
O CO
o w
o
«! 5
£3 «
EH fti
< M
g^ f"A

IA
1


w
pq
3
EH










• «
Tj J>
-P 0)
CO T)

C
oJ
0)
s

•
X
cd
s




"fe^-
O
ON



"6^-
o
CO



VS.
o
t—




%^
o





'fe^-
o
IA



VI
o



v*.
o
00


"fe^-
o
CM
v*.
o
rH

.
C
•H
s
.
o

0)
rH •
to


CO
LA CO
-3" LA
• •
rH CM
LA LA
OO -=f
ONOO
• •
o oo
i-H
t— 0
OO \O
• •
ON LA
CM LA
LA LA


ON J-
ON O\

CM t-
H

ON LA
t- 0
• •
t— LA


ON t—
-3- ON
• •
M3 00



ON LA
H CM
• . 0)
J-i LA OO fn
3 2

H rH
OJ CM

O O
0 O
0 O
rH ON


CO
OJ
•
•H
LA
O
•
H
\£)
VQ
•
o
rH



OO
rH

OJ


rH
LA
•
H


O
O
•
rH



CO
CO
•
O

c^
CD

-------








EH
65
«
(X t-
oo
W ON
EH O
>-5 X
0 Q
CO P

ffl
M Q
K En1
CO K
M 0
Q a*
H
^H P*1
o

1"") ^
C?

OH H
En EH
O
SH W
M Cu
Q X
p E>5
En
r— 3 O
^J
O CO
0 W
o
j a
5 E3
Is; K
O E3
M En
EH En
< M
S£H Q


VO
1
^

w
«
1









, .
•0 >•
4J 0)
CO T3

C
03
0)
s


•
X
ctf
s




Vi
O
ON




^5.
O
oo



VS.
O
t—



v~-
o
vo


Vt
o
LA


Vi
O
-=>•



%^-
o
oo



•^s.
o
OJ


'S^-
o
rH

C
•rH
s
*
o
£n

0)
rH
So
oj C
CO
rH _=f
-=t OJ
• •
.rf LA

rH LA
rH C—
• •
-3 OO

-=t- OJ
t— j-
• •
OJ -*




OO LA
oo vo
• •
OO VO



LA OO
ON OJ

LA LA


VO -3-
t— 0

~f -^j*

^

-------