1、 STD*API/PETRO PUBL 348-ENGL 1998 0732290 ObLOLO3 TOO Air Toxics Emission Factors for Combustion Sources Using Petroleum-Based Fuels Volume 1 Development of Emission Factors Using APWSPA Approach Health and Environmental Affairs Department API PUBLICATION NUMBER 348 PREPARED UNDER CONTRACT BY: DAVID
2、 HANSELL GLENN ENGLAND 18 MASON IRVINE, CALIFORNIA 9271 8 ENERGY AND ENVIRONMENTAL RESEARCH CORPORATION AUGUST 1998 American Petroleum Institute STDmAPIIPETRO PUBL 348-ENGL 1998 W 0732290 Ob10104 947 = FOREWORD API PUBLICATIONS NECESSARILY ADDRESS PROBLEMS OF A GENERAL NATURE. WITH RESPECT TO PARTIC
3、ULAFt CIRCUMSTANCES, LOCAL, STATE, AND FEDERAL LAWS AND REGULATTONS SHOULD BE REVIEWED. API IS NOT UNDERTmG TO MEET THE DUTIES OF EMPLOYERS, -AC- TURERS, OR SUPPLIERS TO WARN AND PROPERLY TRAIN AND EQUIP THEIR EMPLOYEES, AND OTHERS EXPOSED, CONCERNING HEALTH AND SAFETY RISKS AND PRECAUTIONS, NOR UND
4、ERTmG THEIR OBLIGATIONS UNDER LOCAL, STATE, OR FEDEFUIT., LAWS. NOTHING CONTAINED IN ANY API PUBLICATION IS TO BE CONSTRUED AS GRANTING ANY RIGHT, BY IMPLICATION OR OTHERWISE, FOR THE MANU- FACTURE, SALE, OR USE OF ANY METHOD, APPARATUS, OR PRODUCT COV- THE PUBLICATION BE CONSTRUED AS INSURING ANYON
5、E AGAINST LIABIL- ERED BY LETTERS PATENT. NEITHER SHOULD ANYTHING CO“ED IN ITY FOR INFRINGEMENT OF LETERS PAmNT. All rights reserved No part of this work may be reproduced, stored in a retrieval system, or transmitted by any means, electronic, mechanical, photocopying, recording, or otherwise, witho
6、ut prior written pennikswn from the publishex Contact the publisher, API Publishing Services. 1220 L Street, N.W., Wmhington, D.C. 20005. Copyright Q 1998 Amencan Petroleum Instiiute iii Previous page is blank STD.API/PETRO PUBL 3YB-ENGL I1998 m 0732290 Ob10105 883 m ACKNOWLEDGMENTS THE FOLLOWING PE
7、OPLE ARE RECOGNEED FOR THEIR CONTRIBUTIONS OF TIME AND EXPERTISE DURING THIS STUDY AND IN THE PREPARATION OF THIS REPORT API STAFF C ONTACTS Tim Hunt, Health and Environmental Affairs Department Karin Ritter, Health and Environmental Affairs Department WSPA STAFF CONTACT Scott Folwarkow, Concord CA
8、MEMBERS OF THE REVIEW WORKGROUP Tom Brooks, Marathon Oil Company Bill Freemen, Shell Exploration and Production Lee Gilmer, Temo Research Arthur Lee, Texaco Miriam Lev- polychlorinated dibenzodioxins and dibenzofurans; polycyclic aromatic hydrocarbons and other semivolatile organic compounds; benzen
9、e, toluene and other volatile organic compounds; fomaldehyde and other aldehydes; and hydrochloric acid. Procedures developed in a separate CARB-sponsored program were used to screen and validate data, eliminating those data points or sets with significant problems andor reporting deficiencies. Thro
10、ugh this process, the best data sets were selected for emission factor development. Emission factors were developed encompassing all industries in California, and petroleum-industry- specific emission factors. As a result of this study, air toxics emission factors for combustion devices have been de
11、veloped using the best available source testing information. These emission factors can be used by environmental health and safety engineers to develop more accurate and complete emission inventories without additional source testing. This report consists of three volumes: Volume 1 presents emission
12、 factors derived specifically for petroleum industry combustion devices. Volume 2, which presents emission factors derived for inclusion in the CARI3 database, and Volume 3, providing detailed results of data validation and statistical comparisons are available from MIS web site: http:/iuww.apiorg/e
13、hs/ Publicationd348. hhn. STD-APIIPETRO PUBL 3YB-ENGL 1998 0732290 ObLOL07 656 TABLE OF CONTENTS VOLUME 1 Section PaJz+e EXECUTIVE SUMMARY . e5-1 1 . INTRODUCTION 1 . 1 2 . DATA VALIDATION 2-1 2.1 SCREENING 2-1 2.2 DETAILED VALIDATION . 2. 1 2.3 OUTLIER ANALYSIS . 2-4 EMISSION FACTORS . 3-1 3 . 3.1
14、3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 DESIGN AND OPERATING PARAMETERS . 3-2 MAJOR GROUPS 3-4 AIR TOXCS 3-4 NORMALIZING UNITS 3-6 DETAILED DATA LISTING 3.6 TREATMENT OF NONDETECTS . 3-8 SUBGROUP EVALUATION 3-11 3.7.1 External Combustion 3-12 Reciprocating Internal Combustion Engines 3-19 EMISSION FACTOR TA
15、BULATION 3-25 COMPARISON OF CARB AND APVWSPA EMISSION FACTORS 3-31 3.7.2 3.7.3 Gas Turbines . 3-25 REFERENCES R- 1 APPENDIX A . EXTERNAL COMBUSTION EMISSION FACTORS . A-1 APPENDIX B . INTERNAL COMBUSTION ENGINE EMISSION FACTORS B-1 APPENDIX C . DIRECT-FIRE COMBUSTION EMISSION FACTORS . C-1 VOLUMES 2
16、 and 3 . ht:/www.apiorg/ehsPublicatons/348. htm STD.API/PETRO PUBL 348-ENGL 1998 = 0732290 Ob10108 592 LIST OF FIGURES Figure 3- 1 3-2 3-3 3-4 3-5 3-6 3-7 3-8 Nomializing Unit Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7 External Combustion System Type
17、 Comparison . . . . . . . . . . . . . . . . . . . . . . . . . 3-14 External Combustion Fuel Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17 External Combustion NOx Control Comparison . . . . . . . . . . . . . . . . . . . . . . . . . 3-18 External Combustion Emission Fac
18、tor Group Comparison . . . . . . . . . . . . . . . . . 3-21 Reciprocating Internal Combustion Engine Comparison . . . . . . . . . . . . . . . . . . . 3-23 Reciprocating Internal Combustion Engine Emission Factor Group Comparison . 3-26 Gas Turbine Comparison . . . . . . . . . . . . . . . . . . . . .
19、 . . . . . . . . . . . . . . . . . . . . . . 3-27 STD-API/PETRO PUBL 348-ENGL 1998 m 0732290 ObLOLO9 429 Table ES- 1 ES-2 ES-3 ES-4 2- 1 3- 1 3-2 3-3 3-4 3-5 3-6 3-7 3-8 3-9 3- 10 UST OF TABLES . In general, the emissions data for each substance in each group were observed to vary over several order
20、s of magnitude. This variability is due to a combination of measurement uncertainty and dierences in the design and operation of devices tested. The variability was reduced, if pclssible, by identifying design and operating parameters responsible for the variation and further divukg the group into s
21、ubgroups, if warranted. Engineering judgment and statistical analysis were used to determine whether the design or operating parameters had a significant impact on emissions. Pooling of different system and fuel types within each of the five groups was allowed to jnuease the quality of the resultant
22、 emissions factors. Low sensitivity data also were eliminated from the subgroup evaluation process and emission factor calculations. FINDINGS Key observations and findings from the subgroup analysis are listed below. External Combust ion Devices (Boilers an d Process Heaters) 0 Hazardous air poliuta
23、nt (HAP) emission factors for boilers and process heaim are similar; HAP emission factors for external combustion sources fked by natural gas and process gas are similar; ES-2 STDOAPIIPETRO PUBL 348-ENGL 1998 0732290 Ob10114 876 Polycyclic aromatic hydrocarbon (PAH) emission factors for liquid-fired
24、 sources are higher than for gas-fired sources; e Benzene, toluene, and xylene (BTX) and formaldehyde emission factors for liquid- and gas-fired sources are similar; and e HAP emission factors for boilers and process heaters with and without NOx emission controls are similar. Recimocatine Internal C
25、ombustion Eneines e HAP emission factors for gas- and diesel oil-red sources are similar, except that formaldehyde emissions from diesel sources are lower; e HAP emission factors for 2 and 4 stroke sources are similar, except that total BTX emissions fi-om 2 stroke sources are higher; and e Rich bur
26、n reciprocating internal combustion engines (RICE) have higher total BTX and lower formaldehyde emissions than lean burn RICE. Based on these observations, external combustion devices and reciprocating internal combustion engines were divided into the subgroups shown in Tables ES-1 and ES-2, respect
27、ively. The arithmetic mean emission factor and EPA quality rating are provided in these tables for each substance. Gas turbines were divided into two subgroups-with and without duct burners (Table ES-3). Emission factors for direct-fired devices (asphalt blowing and coke calcining) are provided in T
28、able ES-4. The gas turbine and asphalt blower subgroups were developed from small data sets and could not be evaluated statistically for signifkance. The coke calcining emission factors were developed from source testing results on a single unit, therefore, no subgroup development was possible. The
29、EPA ratings listed in Tables ES-1 through ES-4 are assigned on a scale of A to E: EPA Oualitv Ratinq Emission Factors A - Excellent 28 B - Above Average 2 C - Average 12 D - Below Average 260 E - Poor 107 NR - Not Reported 25 ES-3 _ STD*API/PETRO PUBL 3YB-ENGL 1998 m 0732270 Ob10115 722 m As shown,
30、most of the emission factors developed in this project have EPA ratings of D or E. The predominant reason for low EPA ratings is that emission factors were developed from tests of only one or two sources. Under the EPA rating system, these emission factors can only receive a D or E raiing even thoug
31、h the measurement data may be of high quality as a result of the screening and validation procedures. With additional future testing, most of the D-rated emission factors would move to the A through C categories because these factors were collected using CARB or EPA source test methodologies. Emissi
32、on factors with E ratings also were developed using sound source test methodologies and data points with significant problems have been identified and eliminated. The E rating was assigned because the source test methodologies used (mainly CARB Method 428, CARI3 Method 430, and EPA Method 29/CARl3 M
33、ethod 436 results) have been significantly revised since tests were performed. CARB has not reqked any retesting for data collected using previous versions of test methods and still considers data collected using these methods to be valid. Some discussion of the gas turbine and oil-fired extemai com
34、busfion metals data is warranted because these data were collected using versions of EPA Method 29/CARB Method 436 which have been significantly revised to reduce contaminaton. Specifically, Manganese emission factors for the oil-fired external combustion sources and gas turbines are suspect because
35、 of possible cross-contamination between the HNO,/H,O, and KMnO,/H,SO, imphgers. Arsenic, Chromium, Copper, Lead and Zinc for oil-fired external combustion sources were detected in levels of significance in field blanks compared to the sample. A significant level is defined as one where the ratio of
36、 the blank to sample is greater than 0.10. The gas turbine trip blanks indicated significant levels of Mercury and Zinc for the sources with duct burners. A complete assessment of contamination was not possible for the gas turbine data sets because field blanks were not collected. REPORT ORGANIZATIO
37、N Only Volume 1 is presented here; Volumes 2 and 3 are available fiom MIS web site: http:/www.apiorg/ehflublic and 0 Asphalt blowers and coke calciners. The database includes sources fired with a variety of gaseous and liquid fuels, including natural gas and petroleum industry process gases (e.g., r
38、efinery gas, field gas, casing vapor recovery gas). The database includes information on sources located in California only. The data may not be representative of emission factors for sources outside of California. Since California generally has more stringent air pollutant emission regulations than
39、 other states, California sources probably have more controls for criteria pollutants (e.g., N Statistical comparisons were conducted assuming the data are log-normally distributed; and Undetected test results collected using low sensitivity analytical techniques were excluded. 1-3 Section 2 DATA VA
40、LIDATION Results from the validation process are summarized below. The validation process included screening and detailed validation steps, and an outlier analysis. A complete discussion of the validation process is provided in Volume 2. 2.1 SCREEMNG The objective of the screening analysis was to el
41、iminate test reports with insufficient process information to develop emission factors or documentation to evaluate the accuracy of test results. Supporting documentation included device and method descriptions, sample, lab, and blank data. Some tests were accepted without al of the supporting docum
42、entation to increase the size of the source pools. Test results for non-combustion sources and tests without air toxic emission results were excluded. of the 161 tests screened, 68 were selected for detailed data validation and extraction. The 68 tests were comprised of 33 “old” tests evaluated in a
43、 previous CARB project to develop emission factors and 35 “new” tests. 2.2 DETAILED VALIDATION Data validation procedures unique to each test method were used to evaluate the quality of emissions data. Data validation included checking to ensure that the correct sampling and analysis procedures were
44、 used, identifymg significant problems such as high field blanhs, checking calculations, and evaluating the accuracy of the test results. Major problems encountered during the method validation are documented in Table 2-1 for each test. The most comon problem was the lack of a fuu set of internal st
45、andards used during Method 429 (PAH) analyses. PAH data from 15 tests in document 2599 were eliminated from the emission factor development process because none of the required internai standards were used in the analytical procedures, high levels of contaminants were found in many of the samples, a
46、nd LRMS was used yielding high detection limits. Five Method 430 (HCHO) calculation check failures were documented; these data were either corrected or eliminated. 2-1 !i M .% e 2-2 $* p II 2-3 STD-APIIPETRO PUBL 348-ENGL 1778 0732290 Ob10127 217 m 2.3 OUTLIER ANALYSIS Before developing emission fac
47、tors, outliers were identined and evaluated. For this study, the Dixon test was used to identify outliers per substance per test and per substance per major group. To use the Dixon test, a group of data was selected and sorted from lowest to highest emissions. Then the high and low points were exami
48、ned statistically in relation to the other points in the data set. The test identified if the high and low points were outliers at a 95 percent level of confidence. Outliers were evaluated to determine if sampling problems, calculation errors, or process upsets occurred. Outliers with calcuiation er
49、rors were corrected; those with samphg problems were rejected. Twenty-two outliers were identifiled in the analysis by device and substance, and four in the analysis by major group and substance. These results were excluded from the emission factor calculations. 2-4 STD.API/PETRO PUBL 348-ENGL 1998 0732290 Ob10130 T39 Section 3 EMISSION FACTORS CARI3 developed a procedure.to provide emission factors of known quality for a wide range of air toxics and source types (Hanse& 1996). This procedure considers the design and operation of the sources,
