United States Environmental Protection Agency Environmental Monitoring Systems Laboratory Research Triangle Park NC 27711 EPA/600/4-87-001 December 1986 Research and Development vvEPA A Summary of the 1985 EPA National Performance Audit Program on Source Measurements ------- EPA/600/4-87/001 December 1986 A SUMMARY OF THE 1985 EPA NATIONAL PERFORMANCE AUDIT PROGRAM ON SOURCE MEASUREMENTS E. W. Streib, T. J. Logan, and M. R. Midgett 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 ------- 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 endorse- ment or recommendation 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), EPA 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. John C. Puzak, Deputy Director Environmental Monitoring Systems Laboratory Research Triangle Park, North Carolina iii ------- ABSTRACT In 1985, 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 (S02J and Method 7 (NOX), and two coal samples for Method 19A. Participating 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 4.0 percent from the expected (EPA) value, and the mean for 02 was 0.4 percent from the expected value. In the Method 5 audit, the mean value for all participants differed by 10.6 percent from the expected value. For the Method 6 audit, the average mean differed by 10 percent from the expected value. The average mean in the Method 7 audit was 4 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, 87 percent of the participants measured within 10 percent of the expected value; for Btu, 98 percent of the participants measured within 10 percent of the expected value. This report includes the results of the performance audits done during the period from January to December 1985. iv ------- CONTENTS Page Foreword Ill Abstract 1 v Figures vl Tabl es vll Acknowledgment vl 11 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 16 7. Method 19A Audit 19 References 23 Appendixes A. Frequency Distributions 24 B. Instructions for EPA Audit Materials 30 ------- FIGURES Number Page 1 Cumulative accuracy for participants in Method 5 audit, 0685 10 2 Previous results of Method 5 audit 11 3 Results of Method 5 audit, 0685 12 4 S02 expected values compared to participants' values 15 5 NOX expected values compared to participants' values 18 VI ------- TABLES Number Page 1 Participants' Results from Method 3 Audit (All Data—No Outliers Removed) 2 2 Method 5 Audit 0685 - Summary Statistics 3 3 Participants' Results from Methods 6 and 7 Audits (All Data—No Outliers Removed) 3 4 Participants' Results from Method 6 and 7 Audits (Outliers Removed) 4 5 Participants' Results from Method 19A Coal Audits (All Data—No Outliers Removed) 5 6 Method 3 Audit Participants 6 7 Source Method 3 Audit—0785 7 8 Method 5 Audit Participants 8 9 Previous Results of Method 5 Audit 9 10 Method 6 Audit Participants 13 11 Summary of Source SOg Audit—0585 14 12 Method 7 Audit Participants 16 13 Summary of Source NOX Audit—0485 17 14 Coal Audit Participants 19 15 Source Coal Audit—0385 21 16 Source Coal Audit—0985 22 17 National Orsat Audit Frequency Distribution of Absolute Percent Differences of Expected and Reported Values—Study 0785 25 18 DGM Frequency Distribution of Absolute Percent Di fference—0685 26 VI1 ------- TABLES Number Page 19 S0£ Frequency Distribution of Percent Difference—No Outliers Removed—Audit 0585 27 20 NOX Frequency Distribution of Percent Difference—No Outliers Removed—Audit 0485 27 21 National Coal Audit Frequency Distribution of Absolute Percent Differences of Expected and Reported Values- Study 0385 28 22 National Coal Audit Frequency Distribution of Absolute Percent Differences of Expected and Reported Values- Study 0985 29 ------- 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 evaluate the data. ------- 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 19A 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 other laboratories conducting similar measurements. Source Test Methods 3, 5, 6, and 7 were each audited once and Method 19A twice in 1985. 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 device 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 accuracy with respect to the EPA expected value and those obtained by the participants (see Appendix A). The Quality Assurance Division of EMSL also maintains a limited repository of samples for the EPA Methods 3, 6, 7, and 19A that are available to source testing laboratories for such purposes as training new personnel or conducting 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 1985 source audits. ------- SECTION 2 SUMMARY In 1985, 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 ($02), 7 (NOX), and 19A (coal). Industrial laboratories, contractors, foreign laboratories, as well as local, State, and Federal agencies participated. The results of the 1985 audit of Method 3 are summarized in Table 1. Participants analyzed the gas sample twice for percentages of carbon dioxide (C02), oxygen (03), and carbon monoxide (CO). The mean values of C02 and 62 differed by 4 percent and 0.4 percent from the expected values, respectively. In contrast, the mean values for CO differed as much as 63 percent from the expected values. TABLE 1. PARTICIPANTS1 RESULTS FROM METHOD 3 AUDIT (ALL DATA—NO OUTLIERS REMOVED) Type of Audit sample date Small 0785 cylinder (gas) Parameter % C02 % 02 % CO No. of analyses 46 45 47 46 37 36 Repli- cate 1 2 1 2 1 2 EPA true value 7.00 7.00 12.00 12.00 0.60 0.60 Participant results Mean 6.76 6.68 11.98 12.07 0.96 1.00 Median 6.60 6.60 12.00 12.00 0.50 0.50 Std.dev. 0.97 0.88 1.38 1.33 2.92 2.97 One audit of Method 5 was conducted in 1985. The overall results (no outliers removed) are summarized in Table 2. The mean for all participants was 10.6 percent from the expected value and the standard deviation was 102.1 percent. After the removal of 7 percent of the data statistical outliers, the mean was 2.3 percent from the expected values and the standard deviation was 1.7 percent. The participants' performance based on the standard deviation of all data was poor; however, when the 7 percent of the outliers was removed, the performance was consistent with previous years. ------- TABLE 2. METHOD 5 AUDIT 0685 - SUMMARY STATISTICS Mean Median Std. dev All data Outliers removed 633 588 10.6 2.3 2.2 2.0 102.1 1.7 Table 3 represents the data (no outliers removed) from the Methods 6 and 7 audits. In the Method 6 audit, the procedure requires the partici- pants to determine the sulfate content in five aqueous solutions using the titration procedure. For four out of five concentrations, the mean of the participants was 10 percent higher than the expected value. In contrast, the median differed by 2 percent. In the Method 6 audit, 60 to 85 percent of the participants achieved an accuracy within 5 percent of the expected value. The Method 7 audit procedure requires that the participants determine the nitrate content in five aqueous solutions. For each concentration, the mean of the participants was less than 4 percent from the expected value, and the median differed by less than 2 percent. In this audit, 70 percent of the participants achieved an accuracy within 10 percent of the expected value for all five concentrations. TABLE 3. PARTICIPANTS' RESULTS FROM METHODS 6 AND 7 AUDITS (ALL DATA—NO OUTLIERS REMOVED) Type of Audit sample date Aqueous 0585 sulfate Aqueous 0485 nitrate No. of analyses 89 91 90 91 89 72 73 73 72 73 EPA true value (mg/DSCM) 198.3 221.2 701.6 1250.7 1326.9 99.5 119.5 298.6 338.4 537.5 Participant results Mean 243.8 245.5 723.7 1332.6 1426.1 97.9 118.3 293.8 344.8 518.4 Median 201.5 225.0 698.3 1240.1 1323.1 101.1 119.0 298.0 339.4 534.0 Std. dev. 263.8 104.5 247.9 547.5 595.3 24.4 23.0 49.5 73.8 112.9 ------- Table 4 represents the data from Methods 6 and 7 after the removal of outliers. Two percent of the Method 6 data and 3 percent of the Method 7 data were classified as outliers. It is important to note that even with the statistical outliers removed, the performance of this year's Method 6 data (based on the standard deviation) is not as good as in previous years. TABLE 4. PARTICIPANTS' RESULTS FROM METHOD 6 AND 7 AUDITS (OUTLIERS REMOVED) Type of Sample Audit Date No. of Analyses EPA true value (mg/DSCM) Participant Mean Median results Std. dev Aqueous sulfate 0585 88 88 89 89 87 198 221 701 1250 1326 ,3 .2 ,6 .7 .9 217.6 227.6 699 1252 1338 201.3 224.6 697.8 1240.0 1322.2 91.9 25.2 95.9 103.6 117.0 Aqueous 0485 nitrate 69 71 71 70 71 99.5 119.5 298.6 338.4 537.5 101.5 120.8 299.4 342.9 530.9 101.9 119.1 298.5 339.4 534.0 17.5 17.7 36.3 49.5 84.8 Table 5 summarizes the results of the two coal audits that were con- ducted in 1985. 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 ash, moisture, and sulfur content were within 6 percent of the expected value except for a 20 percent difference on the lowest sulfur concentration. An accuracy of 1 percent was achieved on the 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 is a good example of the accuracy to which a measurement can be made when that measurement receives good laboratory quality control. ------- TABLE 5. PARTICIPANTS' RESULTS FROM METHOD 19A COAL AUDITS (ALL DATA—NO OUTLIERS REMOVED) Type of Audit sample date Coal 0385 0985 0385 0985 0385 0985 0385 0985 0385 0985 0385 0985 0385 0985 0385 0985 No. of Parameter analyses % S 111 108 110 106 % H20 111 106 112 108 % Ash 111 107 110 105 Btu/lb 108 11 102 12 109 13 104 12 EPA value 1.62 0.72 4.97 2.35 3.86 1.61 3.93 17.45 11.36 4.33 20.32 17.92 ,414.0 ,300.0 ,054.0 ,374.0 Participant results Mean 1.60 0.86 4.75 2.29 3.60 1.75 3.97 16.73 11.37 4.25 19.92 17.59 11,379.7 12,252.9 13,011.1 12,179.3 Median 1.59 0.70 4.76 2.29 3.70 1.53 3.95 17.41 11.33 4.37 20.07 17.83 11,383.5 12,273.0 13,025.0 12,319.0 Std.dev. 0.14 1.56 0.35 0.23 0.52 1.86 1.65 2.61 0.45 0.44 1.52 2.34 256.61 183.77 261.53 573.78 ------- SECTION 3 METHOD 3 AUDIT The Method 3 audit checks participants' ability 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 COg, Og, 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 1985 audit, 44 percent of the 103 laboratories receiving the audit package returned data. Table 6 shows the total number of laboratories requesting participation and the number that returned data for the Method 3 audit. TABLE 6. METHOD 3 AUDIT PARTICIPANTS No. receiving No. returning Category samples data Contractors Industry Foreign Federal State Local 57 28 1 2 12 3 16 15 1 2 8 3 Total 103 45 Table 7 summarizes the Method 3 audit results. Each laboratory was asked to analyze the sample in duplicate. Five and ten percent accuracy were chosen for the reporting criteria for each of the parameters. Each parameter had only one concentration. In the 1985 audit, 35 percent of the reporting laboratories achieved an accuracy within 5 percent for the C02, which was a decrease from the last audit.'4) Ninety percent of the laboratories achieved an accuracy within ------- 5 percent of the expected value for the 02 analysis. For the CO analysis, only 33 percent of the laboratories achieved an accuracy within 5 percent, and 21 percent of the laboratories did not report a value for CO. TABLE 7. SOURCE METHOD 3 AUDIT--0785 Expected value 7.00 12.00 0.60 No. of analyses* (1) 46 (2) 45 (1) 47 (2) 46 (1) 37 (2) 36 Laboratories accurate within 5% (%) 35 30 89 91 32 36 Laboratories accurate within 10% (%) 85 89 92 94 32 36 * 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, allow 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 1985 audit, 59 percent of the 160 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 81 51 4 2 18 4 37 35 2 2 14 4 Total 160 94 Figure 1, a cumulative histogram, shows the accuracy obtained by participants 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^) requires that the dry gas meter be calibrated within an accuracy of 2 percent. Figure 2 shows that 47 percent of the reporting laboratories attained this accuracy. These results are not as good as those from 1984. Table 9 is a summary of previous audits that, like Figure 2, shows a decline from 1984. ------- TABLE 9. PREVIOUS RESULTS OF METHOD 5 AUDIT Audit 0382 0982 0383 0983 0684 0685 Number of analyses 827 769 763 614 631 633 Mean Median (% from EPA 7.6 12.5 5.7 4.1 3.1 10.6 value) 2.5 2.5 2.2 1.9 1.9 2.2 Std. dev. 39.5 81.4 32.7 21.3 4.9 102.1 The histogram in Figure 3 depicts the individual results from all participants of the 1985 audit with the mean and median values. The majority of the laboratories reported values lower than the EPA value. The standard deviation of the triplicate analysis (repeatability) by each laboratory indicated that 71 percent of the standard deviations for each set were within 0.3 percent. Four percent of the 1985 data were identi- fied as outliers using Chauvenet's Criterion.(5) Before the outliers were removed, the mean value (absolute) differed by 10.6 percent from the expected value. After deletion of outliers, this value was reduced to 2.3 percent. ------- LABORATORIES ACHIEVING SPECIFIED ACCURACY, percent N> O U O (II O O) o Nj o 00 o (O o o o (Q C -^ CD 3 II O) w w o c 3 c_ D) 31 m Z O m O CD 0) O O C 3 o •a to o •6' Q> 3 W 3 CO m CO c 3) m m Z CO Tl m •o a> o a. 01 0) c a. rn •o CD 00 O en oo en (O ------- 55 50 CO ill I" oc o CO 40 35 DEGREE OF ACCURACY = 2% 0382/827 0982/769 0383/763 0983/614 0684/631 0685/633 AUDIT / NUMBER OF SAMPLES Figure 2. Previous results of Method 5 audit. 11 ------- 20 15 V UJ c 10 oc o CD NUMBER OF VOLUMES = 633 MEDIAN = -1.0 OUTLIERS REMOVED Mean = 0.7 OUTLIERS REMOVED <-12 -10 -8-6-4-20246 DIFFERENCE FROM EPA VALUE, percent 8 10 >10 Figure 3. Results of method 5 audit, 0685. ------- SECTION 5 METHOD 6 AUDIT The Method 6 audit checks participants' ability to quantitatively analyze Method 6 samples for sulfur content. The audit set consists of five aqueous dilutions of 10 N sulfuric acid in 25 milliliters (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 perchlorate to a pink endpoint. In calculating the results, the participants assume an original sample volume of 100 mL and a sample volume of 0.021 dry standard cubic meter of stack gas. Table 10 shows the categories of the participants and compares the total number of participants requesting participation with the number return- ing data. In the 1985 audit, 66 percent of the 137 laboratories that received the audit package returned data. TABLE 10. METHOD 6 AUDIT PARTICIPANTS Category No. receiving samples No. returning data Contractors Industry Foreign Federal State Local 74 40 3 1 13 6 42 30 2 1 11 5 Total 137 91 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 62 and 67 percent of the reporting laboratories achieved an accuracy within 5 percent for the two lower concentrations, but 83 percent of the laboratories showed an improvement on the higher concentra- tions. Five percent is used as the criterion because it was established as the criterion for source SOg compliance audit samples. 13 ------- TABLE 11. SUMMARY OF SOURCE S02 AUDITS—0585 Concentration ±2% ±5% (mg/DSCM) 198.3 221.2 701.6 1250.7 1326.9 28.1 34.1 58.9 39.6 64.0 61.8 67.0 83.3 87.9 86.5 91 Figure 4 shows the mean values for each S0£ concentration and the confidence intervals (CI). Each CI is calculated using the 95 percent confidence level. The CI for the lowest concentration is ± 22.5 percent; whereas the CI for the highest concentration is ± 8.7 percent. Overall, the CI's for the SOg were higher than expected for this method because of the high standard deviations. 14 ------- PARTICIPANTS'VALUES. mg/DSCM 198.3 221.2 701.6 1250.7 1326.9 EXPECTED VALUES, mg/DSCM Figure 4. SC>2 expected values compared to participants' values. * 00 N> 0 O O O C 0 0 O C II I I I I I I I I I II LLJ II I I I I I I I I I I ------- SECTION 6 METHOD 7 AUDIT The Method 7 audit checks 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 2b-mL 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 4.3 of Method 7.U) After this treatment is completed, the absorbance is measured at 410 nanometers (nm) with a calibrated spectro- photometer. 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 participa- tion and the number that returned data for the 1985 Method 7 audit. Sixty- two percent of the 117 laboratories receiving the audit package returned data. TABLE 12. METHOD 7 AUDIT PARTICIPANTS No. receiving No. returning Category samples data Contractors 67 36 Industry 36 25 Foreign 2 2 Federal 1 0 State 6 6 Local 5 4 Total 117 73 16 ------- The percentage of laboratories that achieved 5 and 10 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. Sixty percent of the reporting laboratories achieved an accuracy within 10 percent on the lowest concentra- tion, and 70 to 80 percent achieved an accuracy within 10 percent on the highest concentrations. TABLE 13. SUMMARY OF SOURCE NOX AUDIT—0485 Concentration (mg/DSCM) 99.5 119.5 298.6 338.4 537.5 ± 5% (*) 41.7 49.3 54.8 51.4 53.4 ± 10% (%} 58.3 69.9 80.8 72.2 71.2 73 Figure 5 shows the means for each NOX concentration and the CI's. The CI's are calculated the same tfs for SOg. The CI for the lowest concentration is ± 5.8 percent, whereas the CI for the highest is 5.0 percent. Overall, the CI's for the NOX were very low because of low standard deviations. 17 ------- 600 5 500 O (/> O N. O) Ul D (A O DC 400 300 200 100 - E 99.5 537.5 119.5 298.6 338.4 EXPECTED VALUES, mg/DSCM Figure 5. NOX expected values compared to participants' values 18 ------- SECTION 7 METHOD 19A COAL 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 scrubber inlet flue gas sulfur concentra- tion. 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^6) were recommended, but not required, for participants1 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, 80 percent of the laboratories that received the audit package returned data. Seventy-three of the same laboratories participated in both audits and returned data. Table 14 shows the total number of laboratories requesting participation and the number that returned data for coal audits 0385 and 0985. TABLE 14. COAL AUDIT PARTICIPANTS No. requesting samples No. returning data Category 0385 0985 0385 0985 Contractors Industry Federal State Local Total 57 42 1 11 4 115 52 37 1 12 5 107 42 37 0 10 3 92 42 33 0 10 5 90 19 ------- Tables 15 and 16 summarize the coal audit results. The number of analyses is greater than the number of participants because some companies had more than one laboratory participating. In this case, 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 (sulfur, moisture, ash, and gross calorific content). In the 0385, 90 percent of the laboratories were able to analyze both sulfur samples within 10 percent of the expected value. In the 0985 audit, 76 percent of the laboratories were able to analyze the sulfur content of the lower concentration sample within 10 percent of the expected value; however, 90 percent were able to analyze the high level sample within 10 percent of the expected value. Seventy percent were within the 10 percent criterion for both moisture concentrations in the 0385 audit, whereas 89 percent in the 0985 audit achieved 10 percent for the higher moisture concentration, but only 55 percent of the laboratories achieved 10 percent on the lower concentration. For the ash analysis and Btu content, 87 to 100 percent of the reporting laboratories were able to achieve an accuracy within 10 percent for both sample concentrations. Comparing the 0385 audit to the 0985 audit, there was improvement in the latter audit except for sulfur and moisture parameters. In the 0985, the sulfur and moisture levels were considerably lower, therefore causing a greater chance for error. 20 ------- TABLE 15. SOURCE COAL AUDIT—0385 Expected value 1. 4. 3. 3. 11. 20. 11, 13, 62 97 86 93 36 32 414 054 Laboratories No. of accurate analyses* within 5% (%) (1) (2) (1) (2) (1) (2) (1) (2) (1) (2) (1) (2) (1) (2) (1) (2) 111 107 110 106 111 104 112 105 111 105 110 104 108 103 109 104 Sulfur 76 78 53 60 Moisture 43 39 40 45 Asji 96 97 93 96 Gross Calorific 98 99 98 98 .6 .5 .6 .4 .2 .4 .2 .7 .4 .1 .6 .2 .1 .0 .2 .1 Laboratories accurate within 10% (%) 93. 95. 88. 85. 70. 68. 72. 74. 97. 97. 99. 99. 99. 99. 99. 99. 7 3 2 8 3 3 3 3 3 1 1 0 1 0 1 0 * Numbers in parentheses indicate first and second analyses. 21 ------- TABLE 16. SOURCE COAL AUDIT—0985 Expected value 0.72 2.35 1.61 17.45 4.33 17.92 12,300 12,374 No. of analyses (1) 108 (2) 102 (1) 106 (2) 100 (1) 106 (2) 100 (1) 108 (2) 102 (1) 107 (2) 101 (1) 105 (2) 99 (1) 102 (2) 97 (1) 104 (2) 99 Laboratories accurate within 5% Sulfur 46.3 50.0 62.3 65.0 Moisture 31.1 35.0 66.7 70.6 Ash 68.2 69.3 94.3 92.9 Gross Calorific 99.0 100 89.4 92.9 Laboratories accurate within 10% 77.8 75.5 87.7 93.0 51.9 57.0 88.0 90.2 86.9 86.1 95.2 96.0 100 100 93.3 92.9 Numbers in parentheses indicate first and second analyses, 22 ------- REFERENCES 1. U.S. Environmental Protection Agency. Standards of Performance for New Stationary Sources—Appendix A. Title 40, Part 60, Code of Federal Regulations. 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, April 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. 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. 6. American Society for Testing and Materials. Annual Book of ASTM Standards—1984. Volume 5.05. 01-050584-13, American Society for Testing and Materials, Philadelphia, Pennsylvania. 23 ------- APPENDIX A FREQUENCY DISTRIBUTIONS 24 ------- en TABLE 17. NATIONAL ORSAT AUDIT FREQUENCY DISTRIBUTION OF ABSOLUTE PERCENT DIFFERENCES OF EXPECTED AND REPORTED VALUES—STUDY 0785 C02 Sample no. No. 5000 91 Min. 10% 20% 30% 40% 50% 60% 70% 80% 90% Max. Mean Std.de v. 0.00 2.86 2.86 4.29 5.71 5.71 7.14 7.14 10.00 11.43 71.43 8.51 10.85 02 Sample no. No. 5000 93 Min. 10% 20% 30% 40% 50% 60% 70% 80% 90% Max. Mean Std.dev. 0.00 0.00 0.00 0.83 0.83 1.67 1.67 1.67 3.33 4.17 51.67 4.24 10.44 CO Sample no. No. Min. 10% 20% 30% 40% 50% 60% 70% 80% 90% Max. Mean Std.dev. 5000 73 0.00 0.00 0.00 0.00 16.67 16.67 16.67 33.33 50.00 66.67 2950.00 102.06 480.43 ------- TABLE 18. DGM FREQUENCY DISTRIBUTION OF ABSOLUTE PERCENT DIFFERENCE—0685 (S3 CT> Sample no. 633 630 617 609 600 591 588 Min. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 10% 0.6 0.6 0.6 0.6 0.6 0.6 0.5 20% 1.0 1.0 1.0 1.0 1.0 1.0 0.9 30% 1.3 1.3 1.3 1.2 1.2 1.2 1.2 40% 1.8 1.7 1.7 1.7 1.7 1.7 1.7 50% 2.2 2.2 2.2 2.1 2.1 2.0 2.0 60% 2.6 2.6 2.5 2.5 2.5 2.4 2.4 70% 3.2 3.2 3.1 3.1 3.0 2.9 2.9 80% 4.1 4.1 4.0 3.9 3.7 3.6 3.6 90% 6.3 6.1 5.7 5.2 5.0 4.8 4.7 Max. 1,562.9 75.2 20.1 11.4 9.6 8.6 7.7 Mean 10.6 3.6 2.8 2.6 2.5 2.4 2.3 Std.dev. 102.1 6.5 2.6 2.1 1.9 1.7 1.7 Skewness 14.34 -0.00 -0.00 -0.00 -0.00 -0.01 -0.01 Median 2.2 2.2 2.2 2.2 2.1 2.0 2.0 ------- TABLE 19. S02 FREQUENCY DISTRIBUTION OF PERCENT DIFFERENCE—NO OUTLIERS REMOVED—AUDIT 0585 Sample no. 2 4 7 8 9 No. 90 89 89 91 91 Min. 0.00 0.00 0.01 0.00 0.00 10% 0.40 0.35 0.35 0.38 0.54 20% 0.77 1.01 0.60 0.73 0.99 30% 1.08 2.12 0.88 0.99 1.81 40% 1.51 2.67 1.12 1.22 2.31 50% 1.62 3.68 1.40 1.66 3.35 60% 2.15 4.69 1.80 1.97 3.48 70% 2.94 5.65 2.55 2.44 5.06 80% 4.22 7.31 3.93 3.17 6.69 90% 11.05 15.33 6.17 5.83 14.69 Max. 309.06 1186.89 304.61 296.58 301.67 Mean 8.18 26.47 9.93 9.38 13.72 Std.dev. 34.50 132.35 44.38 43.25 46.52 ro TABLE 20. NOX FREQUENCY DISTRIBUTION OF PERCENT DIFFERENCE—NO OUTLIERS REMOVED—AUDIT 0485 Sample no. 1 3 5 6 8 No. 73 72 73 72 73 Min. 0.03 0.10 0.00 0.03 0.00 10% 0.37 1.51 0.65 0.62 0.50 20% 1.51 2.41 1.77 1.63 1.26 30% 2.31 2.61 2.75 2.63 2.09 40% 3.45 4.52 3.48 3.63 4.44 50% 4.39 6.63 4.52 4.43 4.60 60% 5.66 10.05 6.27 6.62 7.95 70% 7.23 14.87 9.08 9.04 9.62 80% 9.21 26.43 15.40 15.48 19.33 90% 20.73 33.97 32.76 27.45 32.13 Max. 79.60 95.18 98.81 117.55 79.33 Mean 9.23 15.01 11.69 11.61 11.58 Std.dev. 13.81 19.43 17.77 18.50 15.31 ------- TABLE 21. NATIONAL COAL AUDIT FREQUENCY DISTRIBUTION OF ABSOLUTE PERCENT DIFFERENCES OF EXPECTED AND REPORTED VALUES—STUDY 0385 ro oo Sample no. 1000 2000 Sample no. 1000 2000 Sample no. 1000 2000 Sample no. 1000 2000 No. 218 216 No. 217 215 No. 216 214 No. 213 211 Min. 0.00 0.00 Min. 0.00 0.00 Min. 0.00 0.00 Min. 0.00 0.00 10% 0.62 0.60 10% 0.76 1.04 10% 0.09 0.25 10% 0.08 0.09 20% 1.23 1.61 20% 1.27 2.07 20% 0.26 0.44 20% 0.18 0.16 30% 1.85 2.21 30% 3.05 3.37 30% 0.35 0.79 30% 0.25 0.25 40% 1.85 3.22 40% 4.58 4.66 40% 0.44 0.94 40% 0.36 0.33 Sulfur 50% 3.09 4.23 Moi 50% 6.62 6.22 50% 0.53 0 1.28 1 Gross 50% 0.47 0.46 60% 3.70 5.23 sture 60% 8.14 8.03 Ash 60% .79 .37 Calori 60% 0.59 0.63 70% 4.32 6.84 70% 9.41 10.10 70% 1.06 1.87 fie 70% 0.78 0.77 80% 5.56 7.85 80% 11.45 15.28 80% 1.41 2.26 80% 1.03 1.05 90% 8.02 11.47 90% 22.39 22.54 90% 2.64 3.54 90% 1.55 1.63 Max. 66.67 42.25 Max. 417.05 89.12 Max. 24.12 77.36 Max. 17.59 19.98 Mean 4.34 5.76 Mean 13.01 10.05 Mean 1.47 2.32 Mean 0.92 0.92 Std.dev. 6.98 6.22 Std.dev. 40.70 12.30 Std.dev. 3.51 7.46 Std.dev. 1.86 2.07 ------- TABLE 22. NATIONAL COAL AUDIT FREQUENCY DISTRIBUTION OF ABSOLUTE PERCENT DIFFERENCES OF EXPECTED AND REPORTED VALUES—STUDY 0985 ro vo Sulfur Sample no. 3000 4000 No. Min. 210 0.00 206 0.00 10% 20% 1.39 1.39 0.43 0.85 30% 2.78 1.70 40% 4.17 2.98 50% 5.56 3.83 60% 6.94 4.26 70% 8.33 5.96 80% 11.11 8.09 90% 18.06 9.79 Max 2256. 55. Mean 94 31.37 74 5.51 Std.dev. 219.04 7.60 Moisture Sample no. 3000 4000 No. Min. 210 0.11 206 0.00 10% 20% 0.63 1.43 1.24 2.48 30% 2.12 4.35 40% 2.58 6.21 50% 3.15 8.70 60% 4.01 11.18 70% 5.27 14.91 80% 6.76 19.88 90% 10.09 32.30 Max. Mean 99 1009 .08 6.64 .94 25.77 Std.dev. 14.28 106.36 Ash Sample no. 3000 4000 No. Min. 208 0.00 204 0.00 10% 20% 0.46 1.15 0.22 0.45 30% 1.62 0.61 40% 2.31 0.78 50% 3.00 1.17 60% 3.70 1.45 70% 5.08 2.01 80% 6.93 2.51 90% 14.09 4.13 Max. 83.83 87.67 Mean Std. 5.35 9. 3.82 12. dev. 10 53 Gross Calorific Sample no. 3000 4000 No. Min. 203 0.00 199 0.00 10% 20% 0.28 0.45 0.07 0.15 30% 0.61 0.24 40% 0.79 0.37 50% 0.99 0.50 60% 1.12 0.64 70% 1.48 0.89 80% 1.92 1.08 90% 3.51 1.63 Max. 18.52 9.34 Mean Std 2.30 4. 0.80 1. .dev. 19 04 ------- APPENDIX B INSTRUCTIONS FOR EPA AUDIT MATERIALS 30 ------- Page 1 of 2 INSTRUCTIONS FOR USING EPA METHOD 3 AUDIT MATERIALS Equipment Supplied with Audit Kit (1) Small gas cylinder containing four liters of gas (2) Nozzle for cylinder (taped on cylinder) (3) Tygon tubing Equipment to be Supplied by Participant (1) Orsat analyzer (2) Clamp (3) Extra Tygon or surgical tubing Procedure (1) Leak-check apparatus by clamping off tubing. SEE DIAGRAM. (2) Fill up burette with gas by using positive pressure. Vent this sample through the manifold to the atmosphere. (3) Repeat Step 2. (4) Fill up burette past fill mark with the gas and carefully vent out excess to the atmosphere, until the fill mark is reached. (5) Analyze for C09, 09 and CO as described in Sections 4.2.5, 4.2.6 and 4.2.7 of EPA Method 3. (6) Record the results on the data card enclosed with the sample. (7) Repeat Steps 4 through 6. CAUTION: If the tubing is punctured excess times, leakage can occur. Replace if necessary. Send the data card to the address below. (The cylinder gas can should not be returned.) Ms. Ellen Streib Quality Assurance Division (MD-77A) Environmental Monitoring Systems Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC 27711 31 ------- Page 2 of 2 Note: Site number will always be 001 except when other Orsat apparatus or participants are using the same gas sample. The extra apparatus or participants should be labeled 002, 003, etc. APPARATUS SET-UP inn o o NJ o 32 ------- Page 1 of 2 INSTRUCTIONS FOR USE OF ENVIRONMENTAL PROTECTION AGENCY METHOD 5 DRY GAS METER PERFORMANCE TEST DEVICE NOTE: All procedures referred to are from revised Method 5 published in the Federal Register, Volume 12, Number 160, Part II, Thursday, August 18, 1977, pp. 41776-41782 and references contained therein. This revised method should be adhered to in all details in the use of this quality assurance performance device. EQUIPMENT: The participant in this study should possess the following equipment, including the performance test device supplied by EPA. Quantity Item 1 Method 5/Source Sampling Meter Box 1 Stopwatch, preferably calibrated in decimal minutes 1 Thermometer, ambient range 1 Barometer. If unavailable, call nearest National Weather Service and request the ABSOLUTE barometric pressure. (Corrected for temperature and acceleration due to gravity, but not corrected for altitude.) 1 Performance Test Device. A calibrated flow orifice housed in a quick- connect coupling and identified with an engraved three-digit serial number. WARNING: THE DEVICE MUST NOT BE DISASSEMBLED UNDER ANY CIRCUMSTANCES, Use these devices at room temperature. PROCEDURE: Calibration of Vacuum Gauge — The vacuum pressure gauge on the meter box must be calibrated in the range of use (11-22" Hg) against a standard (Hg Manometer) to ensure accurate results. Remove the performance test device from its case and insert it into the gas inlet quick-connect coupling on the source sampling meter box. Turn the power to the meter box on and start the pump. Adjust the coarse flow rate control valve and the fine flow rate control valve to give a reading of 19" Hg (vacuum reading). CAUTION: The vacuum reading must be accurate and stable for the test period. Allow the orifice and source sampling meter box to warm up for 45 minutes with flow controls adjusted as described in Step 3 before starting quality assurance runs. 33 ------- Page 2 of 2 PROCEDURE: (continued) Make triplicate quality assurance runs. For each run, record initial and final dry gas meter volumes, dry gas meter inlet and outlet temperatures, internal orifice pressure drop (AH), ambient temperature, and barometric pressure. Run duration should be slightly greater than 15 minutes. The following procedure is recommended. Fifteen minutes after a run is started, the partici- pant watches the dry gas meter needle closely. As the needle reaches the zero (12 o'clock) position, the pump and stopwatch are stopped simultaneously. The dry gas meter volume and time are recorded. This complete run procedure is performed three times to provide the required triplicate quality assurance runs. Calculate the corrected dry gas volume for each run using equation 5.1 of the above-referenced Method 5. For each replicate, record the corrected dry gas volume in dry standard cubic meters, the sampling time in decimal minutes, the barometric pressure in mm Hg, and the ambient temperature in degrees Celcius on the enclosed data card. Be sure to record the performance test device serial number on the data card in the column headed "Orifice Number." NOTE 1: If you calculate dry gas volume in English Units, use the conversion factor of 0.02832 m3ft3 to obtain the volume in metric units. NOTE 2: If your stopwatch is not in decimal minutes, be sure to convert (e.g. 15 minutes 20 seconds is reported as 15.33 minutes). After recording the requested data on the enclosed data form, return the data form and the performance test device to: Quality Assurance Division (MD-77A) Environmental Monitoring Systems Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC 27711 ATTN: Ellen W. Streib A postpaid return envelope and label are enclosed for this purpose. 34 ------- INSTRUCTIONS FOR USE OF ENVIRONMENTAL PROTECTION AGENCY STATIONARY SOURCE QUALITY ASSURANCE S02 REFERENCE SAMPLES NOTE: All Method 6 procedures referred to are from the amended method pub- lished in the Federal Register Vol. j*2, No. 160, Part II, Thursday, August 18, 1977, PP 1*1782-U178U. This amended method should be adhered to in all details in the analysis of these reference stand- ards. 1. Prepare 3 percent hydrogen peroxide according to Section 3.1.3 of the method (30 ml is required for each sample and each blank). 2. Prepare each reference sample for analysis as follows: Wrap a paper towel around the ampoule and with the ampoule in an upright position break off the top at the prescored mark by exerting pressure sideways. From the ampoule pipette exactly 5 ml of the reference sample into a 100 ml volumetric flask. Add 30 ml of 3 percent hydrogen peroxide solution. Dilute exactly to the mark with deionized, distilled water. Analyze the sample in accordance with the procedure detailed in Section 1».3 of the method, beginning with "Pipette a 20 ml aliquot of this solution..." (Note: If more than 50 ml of barium perchlorate titrant is required for any sample analysis, a smaller aliquot should be selected to allow titration with less than 50 ml titrant.) 3. Calculate the concentration, Cso2 (concentration of sulfur dioxide, dry basis, corrected to standard conditions, mg/dscm), using Equation 6-2. A value of 21 x 10"^ dscm should be used for Vm(std), in the equation. A value of 100 ml should be used for Vsoinin the equation. U. Record the reference standard sample numbers and their corresponding SOg concentrations in mg/dscm on the enclosed data form. Return the form to: Quality Assurance Division (MD-77A) Environmental Monitoring Systems Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC 27711 ATTN: Ellen W. Streib If other than EPA Method 6 is used for your analyses, please explain in detail your analytical procedure on the back of the enclosed data form. 35 ------- INSTRUCTIONS FOR USE OF ENVIRONMENTAL PROTECTION AGENCY STATIONARY SOURCE QUALITY ASSURANCE NOX REFERENCE SAMPLES Note: All Method 7 procedures referred to are from the amended method published in the Federal Register Vol. 42, No. 160, Part 11, Thursday, August 18, 1977, pp 41784-41786. This amended method should be adhered to in all details in the analysis of these reference standards. 1. Prepare absorbing solution according to Section 3.1 of the method. 2. Prepare each reference sample for analysis as follows: Wrap a paper towel around the ampule and with the ampule in an upright position break off the top at the prescored mark by exerting pressure sideways. From the ampule pipette exactly 5 ml of the reference sample into a 100-ml beaker. Add 25 ml absorbing solution to the beaker; adjust the pH to 9-12 (using pH paper as indicated in Section 4.2 of the method) by dropwise addition of sodium hydroxide (1N). Quantitatively transfer the contents of the beaker to a 50-ml volumetric flask and dilute exactly to the mark with deionized, distilled water. Mix thoroughly and pipette a 25-ml aliquot of the diluted sample into a porcelain evaporating dish. Beginning with the evaporation step in Section 4.3, complete the sample analysis. 3. Calculate total jjg N02 per sample using Equation 7-3. Calculate the sample concentration, C (concentration of NOX as NC>2» dry basis, corrected to standard conditions, mg/dscm), using Equation 7-4. A value of 2000 ml should be used for Vsc in Equation 7-4. 4. Record the reference sample numbers and their corresponding concentrations, C, in mg/dscm on the enclosed data form. Return the form to: Quality Assurance Division (MD-77A) Environmental Monitoring Systems Laboratory Environmental Protection Agency Research Triangle Park, NC 27711 ATTN: Ellen W. Streib If other than EPA Method 7 is used for your analyses, please explain in detail your analytical procedure on the back of the enclosed data form. 3G ------- COAL AUDIT PROGRAM INFORMATION 1. There 1s approximately 50 grains of 60 mesh coal per bottle. 2. Analyze the coal samples for moisture and on a dry basis for ash, sulfur and gross calorific value. Report moisture, ash, and sulfur 1n weight percent with gross calorific value reported as BTU/lb. 3. All methods used In the analysis of these coal samples should follow American Society for Testing and Materials (ASTM) recomnended procedures or an accepted automatic analytical device. 4. Suggested procedures are: Moisture D-3173 Ash D-3174 Sulfur D-3177 Gross Calorific Value ... D-2015 Please note on the data card (columns 17-32) the ASTM method number. If an ASTM method was not used for analysis note that on the back of the data card. Be parameter specific. 5. If you cannot analyze the coal sample for all four parameters, analyze for what you can. Analysis of moisture 1s necessary to calculate on a dry basis any of the other three parameters. Analysis of sulfur 1s also necessary for the calculation of gross calorific value. 6. Analyze each sample In duplicate (If possible) and record results as analysis 1 and analysis 2 for each parameter. 7. Most laboratories will use site number 001. Multiple site numbers are used by laboratories that receive more than one set of samples. These central laboratories have requested auditing of their satellite laboratories. 8. After recording the requested data on the enclosed data card, return the data card to: Ms. Ellen W. Strelb Quality Assurance Division (MD-77) Environmental Monitoring Systems Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC 27711 A postpaid return envelope Is enclosed for this purpose. 9. If you have any questions concerning this or any source nethod audit, please call (919/541-7834). ?7 0' *U.S.Government Printing Office: 1987 - 748-121/40689 ------- |