DRAFT - Do not cite or quote

DRAFT Review of Emissions Test Reports for
Emissions Factors Development for Flares and
Certain Refinery Operations


-------
DRAFT - Do not cite or quote

This page intentionally left blank.

11


-------
DRAFT - Do not cite or quote

DRAFT Review of Emissions Test Reports for Emissions Factors Development for Flares and

Certain Refinery Operations

Contract \o l-P-D-l I-OX4
W ork Assignment \o 2-12

I S I ji\ iron mental Protection Agency
Office of Air Quality Planning and Standards

Sector Policies and Programs Division
Research Triangle Park, North Carolina 27711

August 2014

in


-------
DRAFT - Do not cite or quote

Disclaimer

This report has been reviewed by the Sector Policies and Programs Division of the Office of Air
Quality Planning and Standards, Office of Air and Radiation, EPA, and approved for publication.
Mention of trade names or commercial products is not intended to constitute endorsement or
recommendation for use.

iv


-------
DRAFT - Do not cite or quote

Table of Contents

Page

Section 1 Summary	7

Section 2 Background	9

2.1	Overview of Emissions Test Data Review	10

2.2	Overview of Emissions Factor Analysis and De\ elopment 	12

Section 3 Emissions Factor Development from Test Data Collected I Inder the 2011

Refinery ICR	13

3.1	Catalytic Reforming Units - CO	13

3.2	Catalytic Reforming Units - THC	14

3.3	Fluid Catalytic Cracking Units - HCN	15

3.4	Sulfur Recovery Units - CO	18

3.5	Sulfur Recovery Units - NOx	21

3.6	Sulfur Recovery Units - Tl IC		24

3.7	Hydrogen Plants - CO		26

3.8	Hydrogen Plants - \()\ 			27

3.9	Hydrogen Plants - Tl IC		29

Section 4 Discussion of Proposed Re\ isions to SO: I Emissions Factors in AP-42

Section S 13. Sulfur Reco\ery		31

Section 5 I-missions 1'aclor De\ elopment from Test Data Collected During the

Development of Parameters lor Properly Designed and Operated Flares	33

5.1	Flares-CO..		33

5.2	Flares - NOx			37

5.3	Flares - VOC	41

Section 6 References	46

Appendix A Emissions Test Report Data Fields Included In Test Data Summary Files
Appendix B EPA's "Test Quality Rating Tool" Template (ITR Template)

v


-------
DRAFT - Do not cite or quote

List of Tables

Table S-l. Summary of New and Revised Emissions Factors Developed	8

Table 1. Emission Sources and Pollutants with Emissions Test Report Data Reviewed	9

Table 2. Flare Pollutants and Emissions Test Report Data Reviewed	10

Table 3. Analysis of Emission Test Reports for CO from CRUs	14

Table 4. Overview of the Draft Emissions Factor for THC from CRUs	15

Table 5. Analysis of Emission Test Reports for THC from CRl s 	15

Table 6. Overview of the Draft Emissions Factor HC\ from I CO s	16

Table 7. Analysis of Emissions Test Reports for IICN from l-'CC I s	17

Table 8. Overview of the Draft Emissions Factor lor CO from SRI s 	18

Table 9. Analysis of Emission Test Reports lor CO from SRUs			19

Table 10. Overview of the Draft Emissions Factor for XO\ from SRUs		22

Table 11. Analysis of Emission Test Reports for \()\ from SRUs	22

Table 12. Overview of the Draft I -missions Factor for Tl IC from SRUs	25

Table 13. Analysis of Emissions lest Reports for Tl IC from SIU 's	26

Table 14. Analysis of F.mission Test Reports for CO from 11: Munis	27

Table 15. Overview of the Draft Emissions Factor for NOx from 11ydrogen Plants	28

Table 16. Analysis of Emission Test Reports for NOx from Hydrogen Plants	29

Table 17 Analysis of Fmissions Test Reports for Tl IC from Hydrogen Plants	30

Table IS ()\ u\ iew of the Draft Emissions Factor for CO from Flares	36

Table N Analysis of I-missions Test Reports for CO from Flares	37

Table 20 ()\ er\ iew of the Draft Fmissions Factor for NOx from Flares	40

Table 21. Analysis of Emissions Test Reports for NOx from Flares	41

Table 22. Overview of the Draft I-missions Factor for VOC from Flares	44

Table 23. Analysis of I-missions Test Reports for VOC from Flares	45

vi


-------
DRAFT - Do not cite or quote

Section 1
Summary

The purpose of this report is to document the review and analysis of test reports and
assess the use of test report data for developing emissions factors for flares and certain refinery
operations. These emissions factors are being proposed as an update to the Compilation of Air
Pollutant Emission Factors, Volume 1: Stationary Point and Area Sources, AP-42 (EPA, 1995).

On May 1, 2013, Air Alliance Houston, Community In-Power and Development
Association, Inc. (CIDA), Louisiana Bucket Brigade, and Texas Environmental Justice
Advocacy Services (TEJAS), (collectively, "Plaintiffs") filed a lawsuit against the U.S.
Environmental Protection Agency (EPA) alleging that the EPA had failed to review and, if
necessary, revise emissions factors at least once e\ cry three years as required in Section 130 of
the Clean Air Act (CAA). Air Alliance Houston, ci al. v. McCarthy. \o 1:13-cv-00621-KBJ
(D.D.C.). In the complaint, the Plaintiffs sought to compel the EPA to expeditiously complete a
review of the volatile organic compounds (VOC) emissions factors for industrial flares
("flares"), liquid storage tanks ("tanks"), and wastewater collection, treatment and storage
systems ("wastewater treatment systems"), and, if necessary, revise these factors. EPA entered
into a consent decree with the Plaintiffs to settle the lawsuit I nder the terms of the consent
decree, by August 19, 2014, EPA will re\ iew and either propose revisions to the VOC emission
factors for flares, tanks and wastewater treatment systems under CAA section 130, or propose a
determination under C A A section I 3d that re\ ision of these emission factors is not necessary.
By December 19, 2d 14. I-PA will issue final re\ isions to the VOC emission factors for flares,
tanks and wastewater treatment systems, or issue a final determination that revision of these
emission factors for flares is not necessary. I-PA will post each proposed revision or
determination (or combination thereof), and each final revision or determination (or combination
thereof), on its AP-42 website on the dates indicated above.

As part of its efforts to comply with the consent decree, EPA reviewed emissions test
data submitted by refineries lor the 2d I I Petroleum Refinery Information Collection Request
(2011 Refinery ICR) and test data collected during the development of parameters for properly
designed and operated flares and developed new emissions factors, as shown in Table S-l. The
EPA proposes to add these emissions factors to AP-42 sections 5.1 Petroleum Refining, 8.13
Sulfur Recovery, and 13 5 Industrial Flares. Please submit your written comments on the
proposed factors by October I '¦>. 2d 14. Comments should be e-mailed to
refineryfactor@epa. gov.

7


-------
DRAFT - Do not cite or quote

Table S-l. Summary of Proposed New and Revised Draft" Emissions Factors Developed



Kmissions lost data used







Emissions I nil
illlll I'olllllillll

No. of lest
reports

\o. or
mi its '*

list methods

Draft AIM2
Kmissions I'actor

RcpiTSCIlljI-
liM'llOSS

Catalytic
Reforming Unit
(CRU), Total
Hydrocarbon
(THC)

4

4

EPA Method 25A

4.0 x 10"4 lb THC
(as propane)/bbl
feed

Poorly

Fluid Catalytic
Cracking Unit
(FCCU), Hydrogen
Cyanide (HCN)

9

8

EPA Other lest
Method-"2^

X 0 x 10"3 lb
1 ICV/bbl feed

Moderately

Sulfur Recovery
Unit (SRU),
Carbon Monoxide
(CO)

21

20

IPX Method in

1.4 lb ( O tun
sulfur

Moderately

Sulfur Recovery
Unit, Oxides of
Nitrogen (NOx)

22



IPX Method 71:

i) 19 lb NOx/ton
sulfur

Moderately

SRU, THC

5

6

IPX Method 25.\

0.047 lb THC (as
propane)/ ton
sulfur

Poorly

Hydrogen Muni

NOx

X

7

IPX Method 7E

0.081 lb
NOx/MMbtu

Poorly

Flare CO

h

8

PFTIRc

0.34 lb
CO/MMBtu

Moderately

Flare NO\

4

5

PFTIRc

2.91b
NOx/MMBtu

Moderately

Flare Volatile
Organic
Compounds
(VOC)

1

9

PFTIR;c
DIAL d

0.55 lb
VOC/MMBtu

Moderately

a Draft factors are factors that are being proposed. They are not final factors.

b Number of units used during emissions factor development process. This number includes outliers.
c PFTIR is passive Fourier Transform Infrared.

dDIAL is Differential infrared absorption LIDAR (light detection and ranging).

8


-------
DRAFT - Do not cite or quote

Section 2
Background

In April 2011, EPA sent an ICR under CAA section 114 authority to facilities in the
Petroleum Refining industry (EPA, 2011) ("2011 Refinery ICR"). The 2011 Refinery ICR
consisted of four components, and two of these components requested emissions testing data
from refineries. Component 1 of the 2011 Refinery ICR requested all refineries to submit reports
for emissions tests that had been conducted since 2005. Component 4 of the 2011 Refinery ICR
requested that certain refineries conduct testing for specific pollutants at specific emissions
sources in accordance with an EPA-approved protocol and submit the test reports to EPA.
Emissions testing reports were collected for catalytic reforming units (CRUs), fluid catalytic
cracking units (FCCUs), sulfur recovery unils (SIU s), and hydrogen plants, along with several
other emissions sources. Testing was conducted for a number of pollutants, including carbon
monoxide (CO), hydrogen cyanide (HCN), oxides of nitrogen (NOx), and total hydrocarbons
(THC). Emissions testing reports were analyzed for multiple emissions sources and pollutants,
as shown in Table 1, for the purpose of updating or de\ eloping new emissions factors in AP-42.
In general, this project focused on the pollutants required under section 130 of the CAA (CO,
NOx, and VOC1), and those emissions units and pollutants for which there are no current AP-42
emissions factors (EPA 1995). For hazardous air pollutants (HAPs), we focused on HCN from
catalytic cracking units because that emissions unit is often the largest emissions source at the
refinery and HCN is a risk clri\ cr lor the petroleum refinery source category (EPA 2014).

Test data for the operating parameters and emissions from flares at petroleum refineries
and chemical plants are a\ ailablc as a result of \ arious enforcement actions related to flare
performance issues. The N\\ collected additional flare data during development of an analysis
of proper llare operating conditions (EIW 2< > 12) Hare data are available for CO, NOx, and
VOC. as shown in Table 2

This report documents the review and analysis of the available source test reports from
the 2011 Refinery TCR for the emissions sources/pollutants identified in Table 1 and from flare
studies for the pollutants identified in Table 2.

The background files for the AP-42 sections being proposed for revision contain the
information discussed in this document, including the data summary worksheets, the emissions
factor creation worksheets, the individual Test Rating (ITR) score sheets, and test reports that
were reviewed but not used in the calculation of the draft2 emissions factor. A link to the
background files can be found under the section's heading on the AP-42 website
(http://www.epa.gov/ttn/chief/ap42/index.html see sections 5.1 Petroleum Refining. 8.13 Sulfur
Recovery, and 13.5 Industrial Flares). The test reports that were used in the development of the

1	We also focused on THC as a surrogate for VOC.

2	Draft factors are factors that are being proposed. They are not final factors.

9


-------
DRAFT - Do not cite or quote

draft emissions factors are listed as references in the AP-42 sections being proposed for revision.
These references can be accessed by clicking the reference's name in the draft AP-42 section.

Table 1. Emissions Sources and Pollutants with Emissions Test Report Data Reviewed a

Kmissions source

I'olliiliinl

No. C omponent
1 omissions tost
reports

No. C omponent
4 emissions test
reports

Total
n ii in her of
emissions
lest
reports

Catalytic Reforming Units (CRUs)

CO

5

3

8



THC

13

1

14

Fluid Catalytic Cracking Units
(FCCUs)

HCN

i:

10

22

Sulfur Recovery Units (SRUs)

CO

45

5

50



NOx

4D

i

41



THC

17

(i

23

Hydrogen Plants

CO

5

3

8



\()\

11

3

14



TIIC

13

2

15

Total emissions test reports re\ ie\\ed







195

a This table provides llie loial number of test reports (and nol necessarily the number of emissions units).
Each test report ma\ Iki\ e lesl data for 1 or more emissions unit(s). and in some instances, an emissions
unit may have more lluin I lesl report

Table 2. Hare Pollutants and Kmissions l esl Ueporl Data Reviewed 11

Kmissions source

1'olliilnnl

No. emissions lest reports

1'lares

CO

6



NOx

4



VOC

7

Total emissions test reports re\ ie\\ed



7

a This table provides the lolal number of test reports (and not necessarily the number
of emissions units). Each test reporl may have test data for 1 or more emissions
unit(s).

2.1 Overview of Emissions Test Data Review

The facility and emissions information for each test report was compiled in a test data
summary worksheet called "Test_Data_Sum_(pollutant)_(emissionssource)". The data fields
included in the Test Data Summary file are provided in Appendix A. The Test Data Summary
file includes the field "QA Notes" in column DA that summarizes what data are available in the

10


-------
DRAFT - Do not cite or quote

test report and any potential issues with the data. The field "Looked at for EF?" identifies which
emissions factor the test report was reviewed for and the field "Used for EF?" identifies whether
the test report was included in emissions factor development.

To develop an emissions factor, two basic test data requirements need to be included in
the report: (1) pounds per hour (lb/hr) emissions rate, or enough data to calculate the lb/hr
emissions rate, and (2) process hourly production or process rate (process activity/hr), e.g., feed
rate in barrels per hour (bbl/hr), coke burn rate in lb/hr, or production rate in tons per hour
(ton/hr) or standard cubic feet per hour (scf/hr). Each test report was reviewed to confirm
whether the critical fields were available, and the calculations in the test report were reviewed for
accuracy.

For each emissions test report used in developing the emissions factor (i.e., "Yes"
response for field "Use in EF?"), an individual tesl rating (1TR) score was given to the test report
by completing the "Test Quality Rating Tool" Uih in the EPA's WebFIRE Template and Test
Quality Rating Tool (including instructions) spreadsheet (available on the ERT website at:
http://www.epa.gov/ttnchiel/ert/). The "Tesl Quality Rating Tool" template for the ITR is
provided in Appendix B. The ITR is a quantitati\ e measure of llie quality of the data contained
within a test report. The ITR score may range from iMo I on and gives a general indication of the
level and quality of documentation a\ ailuhle in the test report and the level of conformance with
the test method requirements. The '"Tesl Quality Rating Tool" includes a series of questions
related to "Supporting Documentation Provided" (columns A and B) and related to "Regulatory
Agency Review" (columns (i and 11) (ienerally. the "Supporting Documentation Provided"
columns are an indication of the completeness of the test report while the Regulatory Agency
Review" columns pro\ ide an indication of w hether the test was conducted according to the
requirements of the test method Columns A and 1} of the template worksheet were completed in
this analysis. Columns (i and 11, which are specific to State/Local agency reviewers, were not
completed

Ikvause only the "Supporting Documentation Provided" portion of the worksheet was
completed. ITR scores for the test reports in the analysis range from approximately 31 to 65. For
the "Supporting Documentation Provided" portion, the ITR includes 8 general questions, 8
questions for manual test methods, and 10 questions for instrumental test methods. Examples of
the general questions include \\ hether the testing firm described deviations from the test method
or provided a statement that de\ iations were not required; whether a full description of the
process and unit tested w as pro\ ided; and whether an assessment of the validity,
representativeness, achievement of data quality objectives and usability of the data was provided.
For manual test methods, examples of questions include whether the Method 1 sample point
evaluation was included in the test report; whether cyclonic flow checks were included in the
report; and whether a complete laboratory report and flow diagram of sample analysis was
included. For instrumental test methods, example questions include whether a complete
description of the sampling system was provided; whether the response time tests were provided;
whether the calibration error tests were included; and whether the drift tests were included. The
ITR scores for the test reports reviewed are provided in a spreadsheet called "Webfire-
template_(pollutant)_(emissionssource)".

11


-------
DRAFT - Do not cite or quote

2.2 Overview of Emissions Factor Analysis and Development

The emissions factor development approach followed EPA's Recommended Procedures
for Development of Emissions Factors and Use of the WebFIRE Database (EPA, 2013). The
emissions factor analysis for each draft emissions factor is provided in the spreadsheet

"EFCreation	(pollutant)_(emissionssource).xlsm". The recommended procedures in the 2013

guidelines were followed implicitly, including the handling of below detection limit (BDL) test
data, assigning an ITR score for those test reports that are used in the emissions factor analysis,
recommended statistical procedures for determining whether data sets are part of the same data
population, statistical procedures for determining whether tiny data points are outliers (i.e.,
outlier checks), and determining whether data for a particular emissions unit should be included
in the emissions factor. This last step, determining whether lo include data from each unit,
involves comparison of the Factor Quality Index (FQ1) for different emissions units. The FQI is
an indicator of the emissions factor's ability to estimate emissions lor the entire national
population, and it is related to both the ITR score and the number of units in the data set. Once
the statistical procedures are complete, the data set is ranked by ITR score (high to low), and a
FQI is developed for each unit in the candidate set The FQI should decrease with each
emissions unit. When the FQI increases, only a\ eraue test \ allies above the point where the FQI
increases are considered in factor development.

EPA's Emissions Factor Creation spreadsheet combines the emissions data from multiple
test reports conducted on a single emissions unit, so that each emissions unit is equally weighted
with other units. Because the F.P.Vs recommended emissions factor development procedures are
based on the premise that more test data values arc preferred over fewer test data values, the
scope of this project was limited to data sets containing test averages from at least 3 different
emissions units. Additionally, there are times u hen it is necessary to subcategorize the
emissions factor data from particular units because the emissions are dissimilar. The
recommended emissions factor de\ elopment procedures include a statistical procedure for
determining u hether emissions data are from the same data population, to indicate whether
emissions data should be subcategorized based on a characteristic of the emissions unit (e.g.,
type of APCD). This analysis requires 3 more emissions units from each potential subcategory.

Some of the data from instrumental test methods (e.g. Method 7E, Method 10, etc.)
included test run a\ erages reported as a negative value. The 2013 recommended procedures for
emissions factor development do not specify how this data should be handled. Because the
procedures are silent and it is not possible for emissions rates to be negative, this data has been
excluded from emissions factor development in this project.

12


-------
DRAFT - Do not cite or quote

Section 3

Emissions Factor Development from Test Data Collected Under the

2011 Refinery ICR

EPA has reviewed emissions test data submitted by refineries for the 2011 Refinery ICR.
The emissions data review and the draft emissions factor development for each emissions unit
and pollutant are described below.

3.1 Catalytic Reforming Units - CO

The available emissions test data from the 2<> I I Refinery ICR included multiple test
reports for CO from catalytic reforming units (CRl ) l.acli of the a\ ailable test reports was
reviewed, analyzed, and summarized, and for those lest reports included in the draft emissions
factor analysis, given an ITR score.

Based on the emissions test report review and analvsis. 2 emissions test reports for 2
emissions units had useable data and were a\ ailable for inclusion in development of an emissions
factor; these units had reformer charge rale data as the available production data. These useable
emissions test reports are provided in Table 3 In addition, another 2 emissions test reports for 2
emissions units had useable data, with coke burn rale data as the available production data.

These useable test reports are also provided in Table 3 A complete list of the available test
report information is pro\ ided in worksheet "Tesl_Dala_Sum_CO_CRU_2014Aug.xlsm". For
more detail on the analysis and OA conducted, see the field "QA Notes" for each test report.
The emissions data (lb CO hi ) in these test reports are based on measurements taken with EPA
Method 11 M \| |0). and the lest reports included production rate data for the CRU in bbl/hr feed
rate or lb hr coke burn rate

Certain test reports were excluded from the emissions factor analysis because production
rate data are not available

Overall, 4 test reports ha\ e useable data. Two emissions test reports include data on a
reformer charge rate basis while the other 2 emissions test reports include data on a coke burn
rate basis. These production data bases are not in comparable units, and there is no way to
calculate the production rate data on the same basis, so these test reports could not be combined
for emissions factor development. Because the scope of this project is limited to data sets
containing test averages from at least 3 emissions units and because there are only 2 emissions
units with useable test reports in each of the different production rate categories, an emissions
factor was not developed for CRU CO.

13


-------
DRAFT - Do not cite or quote

Table 3. Analysis of Emissions Test Reports for CO from CRUs

l-acililv II)
No.

l iicililv mime

Kmissions
unit

A PC 1)

Tesl
mot hod

A\ersi«»e lesl
result

Ilk

Production Data as Reformer Charge Rate, bbl/hr

MS3C0740

Chevron Refinery,
Pascagoula, Mississippi

EPN CH-

004

Chlorsorb

M10

4.5 x 10"6 lb
CO/bbl feed

46

OK2C0990

TPI Refining Company
Ardmore Petroleum
Refinery Ardmore.
Oklahoma

CRU400B

Venturi
Scrubber

M10

9.8 x 10"5 lb
CO/bbl feed

48

Production Data as Coke Burn Rate, lb/hr

OK2C0990

TPI Refining Company
Ardmore Petroleum
Refinery Ardmore,
Oklahoma

CRU40OB

Venturi
Scrubber

\IH)

2.9 x 10 : lb
CO/lb Coke
burn

48

TX3B1170

Exxonmobil Beaumont
Refinery, Beaumont,
Texas

PTR-4
Reactor
Regenerator
\enl

( auslic
Scrubber

Mlo

2 5x10"3 lb
CO/lb Coke
burn

38

3.2 Catalytic Reforming Units - I IIC

The available emissions lesl data from the 2" 11 Refinery ICR included multiple test
reports for THC from CRl units l-ach of the a\ ailable test reports was reviewed, analyzed, and
summarized, and for Ihose lesl reports included in llie draft emissions factor analysis, given an
ITR score. An o\ er\ iew of llie draft emissions factor is provided in Table 4.

Based on the emissions test report review and analysis, 4 emissions test reports for 4
emissions units had useable data and were included in the development of the draft emissions
factor. These emissions tests reports are provided in Table 5. A complete list of the available
test report information is provided in worksheet "Test_Data_Sum_THC_CRU_2014Aug.xlsm".
For more detail on the analysis and QA conducted, see the field "QA Notes" for each test report.
The ITR scores for these 4 test reports ranged from 33 to 41. The emissions data (lb THC, as
propane/hr) in these test reports are based on measurements taken with EPA Method 25A
(M25 A), and the test reports included production rate data for the CRU in bbl/hr feed rate. In
instances where both M25A and EPA Method 18 (Ml 8) were conducted in the same test report,
the THC data for M25 A alone were extracted from the raw data in the test report appendices, so
that the data from all tests was measured on the same basis.

Certain test reports were excluded from the draft emissions factor analysis for the
following reasons: production rate data are not available, the test method was not compatible
with THC (i.e, Ml8 test reports were excluded because Ml8 measures specific compounds
where M25 A counts total carbon) or the test method was not clearly identified.

14


-------
DRAFT - Do not cite or quote

EPA's recommended emissions factor development procedures were followed for the
CRU THC data. All 4 emissions units were combined for the draft emissions factor
development. The statistical analysis for determining outliers in the data set was conducted, and
no data were found to be outliers. The draft emissions factor is based on the emissions test data
for 4 units and is characterized as Poorly Representative. The draft emissions factor analysis for
CRU THC is provided in worksheet "EF Creation_THC_CRU_2014Aug.xlsm".

Table 4. Overview of the Draft Emissions Factor for THC from CRUs

Emissions lest d:it:i used

Test methods

Draft AP-42
Emissions l-'iiclor

Represenliilix eness

No. of tost
reports

\o. or
units

4

4

IPX Method 25.\

4"\l" 111 I I l( (LIS
pixipane) hhl leed

Poorlx

Table 5. Analysis of Emissions Test Reports for THC from CUl's

1 utility II)
No.

l-'iicilitv n:ime

Kniissions
unit

A PC 1)

Test
method

A\er:i»e test
result . Ih
THC. :is
propiine/hhl
leed

Ilk

MS3( 1)740

('he\ roil Relinerx.

Puscugoulu,

Mississippi

CRl 7w

( hlorsorli

M25A

1 4X \ Id

41

OK2C0990

TPI Refilling ( onipan>
Ardmore Petroleum
Refinery Ardmore.
Oklahoma

CRl 4oo|}

Venluri
Scrubber

M25A

1.4 x 10"5

37

TX3BI25"

The Premeor Refining
(iroup. Inc . Port
Arthur. Texas

CRl 1344

( hlorsorb

M25A

9.0 x 10"5

33

TX3B13 In

Valero Refining
Te\as. L.P., Corpus
Chrisli. Texas

CRU

Scrubber

M25A

1.5 x 10"5

34

3.3 Fluid Catalytic C racking Units - HCN

The available emissions test data from the 2011 Refinery ICR included multiple test
reports for HCN from FCCU units. Each of the available test reports was reviewed, analyzed,
and summarized, and for those test reports included in the draft emissions factor analysis, given
an ITR score. An overview of the draft emissions factor is provided in Table 6.

Based on the emissions test report review and analysis, 9 emissions test reports for 8
emissions units had useable data and were included in the development of the draft emissions
factor. These emissions tests reports are provided in Table 7. A complete list of the available
test report information is provided in worksheet

15


-------
DRAFT - Do not cite or quote

"Test_Data_Sum_HCN_FCCU_2014Aug.xlsm". For more detail on the analysis and QA
conducted, see the field "QA Notes" for each test report. The ITR scores for these 9 test reports
ranged from 46 to 65. The emissions data (lb HCN/hr) in these test reports are based on
measurements taken with EPA Other Test Method-029 (OTM-029) and in some instances with
EPA Conditional Test Method-033 (CTM-033). Test data using CTM-033 were considered
acceptable when the concentration of sodium hydroxide (NaOH) was high (6.0 N NaOH) and the
pH was maintained above 12 for the duration of the test. One test report based on CTM-033 did
not clearly indicate the NaOH concentration, and although the concentration used was unknown,
this test was included in the analysis in order to not exclude potentially compatible data. The test
reports included production rate data for the FCCU in bbl 'hr Iced rale.

Certain test reports were excluded from the emissions factor analysis for the following
reasons: production rate data were not available or the test method was not compatible with
OTM-029 (i.e., CARB Method 426 test reports and some CTM-033 test reports were excluded
because the tests did not involve the use of the higher concentration NaOH solution required in
OTM-029). Methods that use lower strength caustic solutions are not likely to measure the full
HCN emissions.

EPA's recommended emissions factor development procedures were followed for the
HCN FCCU data. For this draft emissions factor, there were two test reports with test runs that
were detection level limited (DLL), meaning that the laboratory result for at least one fraction of
the sample analysis was BDL. In the draft emissions factor calculations, the detection limit was
substituted for sample fractions that were BDL. Although the complete burn and partial burn
regenerators potentially emit different amounts of I ICY subcategories could not be formed for
complete and partial burn regenerators because all of the useable data was for complete burn
regenerators. Because 5 I¦'('('I s are controlled with scrubbers and 3 FCCUs are controlled with
electrostatic precipitators (l-SPs) and it is uncertain what effect each type of control device has
on the HCN emissions, a statistical analysis was performed to determine if these data belong to
the same population. The statistical analysis showed that all of the data belong to the same data
set. Also, w hi le 2 of the 1 CO s ha\ e CO boilers and 6 of the units do not have CO boilers, the
purpose of the CO boiler is to convert CO to CO2, not to control HCN. There is no data
indicating that the CO2 boiler has a significant impact on the HCN emissions. Therefore, all 8
FCCUs were combined for the draft emissions factor development. The statistical analysis for
determining oulliers in the data set was conducted, and no outliers were found. The draft
emissions factor is based 011 the emissions test data for the 8 units and is characterized as
Moderately Representati \ e The draft emissions factor analysis for FCCU HCN is provided in
worksheet "EF Creation_l lCN_FCCU_2014Aug.xlsm".

Table 6. Overview of the Draft Emissions Factor HCN from Complete Burn FCCUs

Emissions tost data (0 use

lost methods

Draft AIM2
Emissions I'aclor

kepiTsenliiliMMiess

No. of losl
reports

No. of units

9

8

EPA OTM-029;
CTM-033

0.0080 lb HCN/bbl
feed

Moderately

16


-------
DRAFT - Do not cite or quote

Table 7. Analysis of Emissions Test Reports for HCN from FCCUs

1 iicilitv II) No.

l-'iicililv 11:1111c

Emissions
unit ¦'

A PC "D

Test
method

A\crsi«»c
lost result

.lb
IK'N/hhl
I'oeil

Ilk

CA5A0190

ExxonMobil Torrance
Refinery, in Torrance,
CA

FCC b

ESP

EPA OTM-

029

0.0031

65

LA3C0560

Citgo Petroleum
Corporation, Lake
Charles Manufacturing
Complex, Lake
Charles, LA

FCCU317

Scrubber

i:pa otm-

(po

0.015

60

MN2B0720

Flint Hills Resources
Pine Bend, LLC Pine
Bend Refinery in
Rosemount, MN

fc( i ::x

i:sp

IPX ()T\I-

029

0.0010

56

MS3C0740
(2008 test)

Chevron Product
Company Refinen. in
Pascagoula, MS

1 ( ( IJ1NI3

i:sp

EPA CTM-

033

() 00014

57

MS3C0740
(2007 test)

Chevron Product
Coni|tan\ Refiners, in
Pascagoula. \IS

1 ( ( I |N)3

i:sp

I ;pa CTM-
033

0.00011

35

NJ1A0820

Hess ( or^oiaHon. Purl
Reading Reliner\. in
Purl Readme. \.l

FCCl -
PT1-A

Scrubber

EPA CTM-

033

0.0047

57

NJlAOXwi

Valero Relining
( oni|xin\. in
Paul shorn. Yl

l( ( I 1

Scrubber

EPA CTM-

033

0.0038

61

TX3BI25"

Valero Porl Arthur
Refinery, in Porl
Arthur, TX

I-CCU1241

Scrubber

EPA OTM-

029

0.014

65

VI6A1530

Hnensa I.I.C. in
Chrisliansled. 1 S
Virgin Islands

FCCU

Scrubber

EPA OTM-

029

0.022

64

a All of the FCCUs with useable data arc complete regeneration units.
b These FCCUs have CO boilers.

17


-------
DRAFT - Do not cite or quote

3.4 Sulfur Recovery Units - CO

The available emissions test data from the 2011 Refinery ICR included multiple test
reports for CO from SRU units. Each of the available test reports was reviewed, analyzed, and
summarized, and for those test reports included in the draft emissions factor analysis, given an
ITR score. An overview of the draft emissions factor is provided in Table 8.

Based on the emissions test report review and analysis, 21 emissions test reports for 20
emissions units had useable data and were included in the development of the draft emissions
factor. Several test reports provide emissions test data for SRU that share a common stack.

When emissions testing is conducted on more than one SIU lhat share a common stack, the
emissions units are counted as one "unit"; the total emissions rate is divided by the total
production rate of all SRU venting to the stack when de\ doping the units' average test results.

The emissions test reports used in the draft factor analysis are provided in Table 9. A
complete list of the available test report information is provided in worksheet
"Test_Data_Sum_CO_SRU_2014Aug.xlsm'\ I or more detail on the analysis and QA
conducted, see the field "QANotes" for each test report. The ITR scores lor these 21 test reports
ranged from 38 to 53. The emissions data (lb CO/hr) in these test reports are based on
measurements taken with EPA Method I'UMIO), and the test reports included production rate
data for the SRU in ton/hr sulfur production

Certain test reports were excluded from the draft emissions factor analysis because
production rate data are not a\ aiIahle. the concentration data lor the test run average in the test
report is a negative nine, or the SIU did not ha\e controls consistent with the other units (e.g., 2
SRU had no controls)

El\Vs recommended emissions factor de\ elopment procedures were followed for the
SRU CO data All 2<> SIU s ha\ e either an incinerator or a thermal oxidizer as the control
device Both incinerators and thermal oxidizers work on the same principles of combustion, and
these terms are often used inlcrchanucaMy by field staff. As such, there is no reason to believe
that these control devices would ha\ e differing levels of CO emissions. Therefore, all of these
units were combined for emissions factor development. The statistical analysis for determining
outliers in the data set was conducted, and no data were found to be outliers. The draft emissions
factor is based on the emissions test data for 20 units and is characterized as Moderately
Representative. The draft emissions factor analysis for SRU CO is provided in spreadsheet "EF
Creation_CO_SRU_2014.\nu \ I sin"'

Table 8. Overview of (lie Drsifl Kinissions Factor for CO from SRUs

Emissions lost data to use

lest methods

Draft AP-42
Emissions I'actor

Represent ill nen ess

No. of test
reports

No. of units

:i

"Ml

LIW Method 1U

1.4 lb CO. Ion sulfur

Nlodcralch

18


-------
DRAFT - Do not cite or quote

Table 9. Analysis of Emissions Test Reports for CO from SRUs

1 iu-ilily II)
No.

l iicililv iiiiine

Emissions unit

AI'C'I)

lost
method

A\ersi«»e
lesl results,
Ih (O/lon
sul I'u r

Ilk

LA3CU01U

laralhon Petioleuiii
Company LLC,
Garyville, Louisiana

srl::u

Lhermal

oxidizer

MlU

u.lu

50

LA3C0610

Marathon Petroleum
Company LLC,
Garyville, Louisiana

SRU234

Thermal

oxidizer

M10

0.21

50

LA3C0650

Valero Refining -
New Orleans, LLC.
St. Charles Refinery,
Norco, Louisiana

SRU1600

Thermal
o\idi/.er

M10

0.47

45

LA3C0650

Valero Refining -
New Orleans, LLC.
St. Charles Refinery,
Norco, Louisiana

SRI 3d

Thermal
oxidizer

\ll<)

0.35

41

OK2C0990

Total Petroleum, Inc.
Ardmore Refinery -
Ardmore, Oklahoma

SRI 1 (500A)

Incinerator

M10

0.038

46

OK2C0990

Total Petroleum. Inc
Ardmore Reliner\ -
Ardmore. Oklahoma

SRI 2 (5wi.\)

Incinerator

M10

0.0061

44

TX3A1190

Delek Relnihil:. I.TI)
Tyler Re1iner\. T\ ler.
'Levis

SRULSRl 2
TC1I2

Incinerator

M10

0.36

38

TX3A 12 3d

( onocoPh1111 ps
IJorger Petroleum
Refinery, Border.
1 e\as

SRI 43

Incinerator

M10

0.38

46

TX3A13()(i

Valem McKee
ReJinei'\. Sunra\.

Texas

EPNV-16 [Unit
830]

Incinerator

M10

8.2

51

TX3A1300 a

Valero McKee
Refinery, Sunrav.
Texas

EPNV-16 [Unit
830]

Incinerator

M10

7.1

51

TX3A1300

Valero McKee
Refinery, Sunray,
Texas

EPN V-5 [Unit
820]

Incinerator

M10

0.065

51

TX3B1090

Total Petrochemicals
USA, Inc., Port
Arthur, Texas

SRU1&2

Thermal
Oxidizer

M10

2.0

46

19


-------
DRAFT - Do not cite or quote

1 iu-ilily II)
No.

l iicililv iiiiino

Kinissions unit

A PC 1)

Test
modi oil

A\orsi«»o
losl results,
Ih ( ()/lon
sul I'u r

Ilk

TX3B111U

HP Products North
America Inc., Texas
City, Texas

SRI

Incinerator

MlU

1.7

44

TX3B1140

Valero Refining -
Texas, L P. East
Plant of Bill Greehey
Refinery, Corpus
Christi, Texas

SRU2

Incinerator

M10

0.061

49

TX3B1220

Motiva Enterprises,
LLC, Port Arthur,
Texas

SRU2&3
combined

Incincralor

\110

0.032

48

TX3B1240

ConocoPhillips
Company, Sweeny
Refinery, Old Ocean,
Texas

epn :x:

Incinerator

\ll<)

0.057

48

TX3B1250

Valero Port Arthur
Refinery, Port Arthur.
Texas

SRI 543

Incincralor

M10

7.7

49

TX3B1250
(2009 test)

Valero Purl Arlluir
Refincr\. Purl Arlluir.

Texas

SRI 544

Incincralor

M10

1.4

49

TX3B1250

Valero Port Arlluir
Refinery. Porl Arlluir.
Tc\as

SRI 545

Incinerator

M10

0.42

49

TX3B1250
(2011 tcsl)

Valero Porl Arlluir
Rclinciy. Porl Arlluir.

Texas

SRI 544

Incinerator

M10

5.3

46

TX3B13K)

Valero Refining. IJill
(ircchcy Refiner} -
Wcsl Plant. Corpus
Chnsli. Ic\as

SRI 1 &2Sulften

Incinerator

M10

2.6

41

TX3B1310

Valero Refining. liill
Grcchcx Re linen -
West Plant, Corpus
Christi, Texas

SRU3

Incinerator

M10

1.3

53

a Data is for same unit from same test report. Separate sets of test runs occurred on multiple days and
were reported separately.

20


-------
DRAFT - Do not cite or quote

3.5 Sulfur Recovery Units - NOx

The available emissions test data from the 2011 Refinery ICR included multiple test
reports for NOx from SRU units. Each of the available test reports was reviewed, analyzed, and
summarized, and for those test reports included in the draft emissions factor analysis, given an
ITR score. An overview of the draft emissions factor is provided in Table 10.

Based on the emissions test report review and analysis, 22 emissions test reports for 23
emissions units had useable data and were included in the de\ elopment of the draft emissions
factor. Two test reports provide emissions test data for SIU that share a common stack. When
emissions testing is conducted on more than one SRU tluil share a common stack, the emissions
units are counted as one "unit"; the total emissions rate is divided In the total production rate of
all SRU venting to the stack when developing the units' average test results. The majority of the
testing was conducted since 2005, although one lest report is from I

The emissions test reports used in the draft factor analysis are pro\ ided in Table 11. A
complete list of the available test report information is pio\ ided in worksheet
"Test_Data_Sum_NOx_SRU_2014.\uu \lsm" I or more detail on the analysis and QA
conducted, see the field "QA Notes" I'or each lest report. The ITR scores for these 22 test reports
ranged from 38 to 56. The emissions data (Ih NOx'hr) in these test reports are based on
measurements taken with EPA Method 71- (M7I-). and the test reports included production rate
data for the SRU in ton hr sulfur production.

Certain test reports were excluded from the draft emissions factor analysis because
production rate data are not a\ ailahle or the SIU did not have controls consistent with the other
units (e.g., 2 SRU had no controls)

I-I'A's recommended emissions factor development procedures were followed for the
SRU M)\ data. All 23 SRI1 units have cither an incinerator or a thermal oxidizer as the control
device. Both incinerators and thermal oxidizers work on the same principles of combustion, and
these terms are often used interchangeably by field staff. As such, there is no reason to believe
that these control devices would have differing levels of NOx emissions. Therefore, all of these
units were combined for emissions factor development. The statistical analysis for determining
outliers in the data set was conducted, and one data value was found to be an outlier and was
removed from the analysis The emissions test that was an outlier had the highest average test
result in the data set. The outlier test conducted on the remaining data set showed no additional
outliers. The draft emissions factor was based on the emissions test data for 22 units and is
characterized as Moderately Representative. The draft emissions factor analysis for SRU NOx is
provided in spreadsheet "EF Creation_NOx_SRU_2014Aug.xlsm".

21


-------
DRAFT - Do not cite or quote

Table 10. Overview of the Draft Emissions Factor for NOx from SRUs

Kmissions lost to use

l osl hum hods

DiiiK AP-42
r.iiiissions l';ic(or

Rep rcson l;ili\ ciioss

No. »l' k'sl
reports

No. ol° iiiii 1 s

22

23 a

EPA Method 7E

0.19 lb NOx/ton
sulfur

Moderately

a One SRU was shown to be an outlier for the data set and was removed from the draft emissions factor
analysis; the draft emissions factor is based on 22 SRUs.

Table 11. Analysis of Emissions Test Reports for NOx from SUUs

l-'iicility II)
No.

liicililv inline

Kinissions unit

A I'd)

Tesl
modioli

A\ersi«»e

(osl
rcsulls. lb
N()\/lon
sulfur

ITK

LA3CU01U

A laralhon IV' Hole urn
Company LLC,
Garyville, Louisiana

sri::u

Thermal
Oxidizer

\i7i:

() 32

5()

LA3C0610

Marathon Petroleum
Company LLC,
Garyville. Louisiana

SRI 234

Thermal
()\idi/.er

M7E

0.24

50

LA3C06503

Valero Refining - New
Orleans. I.I.C. St.
Charles Refinery.
Norco, Louisiana

SRI INK)

Thermal
Oxidizer

\I7E

0.86

48

LA3C0(on

Valero Refining - New
Orleans. I.I.C. Si
( liarles Relineiy.
Norco. Louisiana

SRI 3d

Thermal
Oxidizer

M7E

0.13

44

OK2C099O

Total Petroleum. Inc
Ardmore Refiner\ -
Ardmore. Oklahoma

SRI 1 (500A)

Incinerator

M7E

0.13

49

OK2C0990

Total IVlroleum. Inc
Ardmore Re line r\ -
Ardmore, Oklahoma

SRU2 (560A)

Incinerator

M7E

0.30

48

TX3A1190

Delek Refining, LTD.
Tyler Refinery, Tyler,
Texas

SRU1/SRU2
TGI2

Incinerator

M7E

0.27

38

TX3A1230

ConocoPhillips Borger
Petroleum Refinery,
Borger, Hutchinson
County, Texas

SRU34

Incinerator

M7E

0.32

50

22


-------
DRAFT - Do not cite or quote

l iu-ililv II)
No.

liicililv 11:1111c

Emissions unit

AI'CI)

lost
modioli

A\orsi«»o

lesi
results, Ih
.N()\/lon
sul I'u r

ITK

TX3A1230

ConocoPhillips Borger
Petroleum Refinery,
Borger, Hutchinson
County, Texas

SRU43

Incinerator

M7E

0.12

50

TX3A1300

Valero McKee
Refinery, Sunray,
Texas

EPN V-5 [Unit
820]

Incineralor

M7E

0.27

54

TX3A1300

Valero McKee
Refinery, Sunray,
Texas

EPNV-16 [1 nil
830]

Incineralor

\I7E

0.21

54

TX3A1300b

Valero McKee
Refinery, Sunray,
Texas

EPNV-I^ |l nil

830]

Incinerator

\l7i:

0.17

54

TX3B1090

Total Petrochemicals
USA, Inc., Port
Arthur, Texas

SRI l£2

Thermal
Oxn.li/.er

M7I!

0.21

49

TX3B1110

BP Products North
America Tuc . Texas
City, Texas

SRI

Incineralor

M7E

0.21

48

TX3B1140

Valero Refining -
Texas. L P LaslPlanl
of lilll (il'eelle\
Reliner\. ( oralis
( hrisli. Texas

SRIJ1

Incinerator

M7E

0.25

52

TX3BI 140

Valero Refining -
Texas, L.P. Easl Plain
of Bill Greehe\
Refinery, Corpus
Chnsli. Texas

SRU2

Incinerator

M7E

0.062

52

TX3B1220

Motiva Tnlei'|nises.
LLC, P01L Ailluir.

Texas

SRU2&3

Incinerator

M7E

0.13

52

TX3B1220

Motiva Enterprises,
LLC, Port Arthur,
Texas

SRU4

Incinerator

M7E

0.14

52

TX3B1240

ConocoPhillips
Company, Sweeny
Refinery, Old Ocean,
Texas

EPN 28.2

Incinerator

M7E

0.20

45

23


-------
DRAFT - Do not cite or quote

l iu-ililv II)
No.

liicililv 11:1111c

Emissions unit

AI'C'I)

lost
modioli

A\ersi«»e

lesl
results. Ih
.N()\/lon
sul I'u r

ITR

TX3B1250

Valero Port Arthur
Refinery, Port Arthur,
Texas

SRU543

Incinerator

M7E

0.085

52

TX3B1250

Valero Port Arthur
Refinery, Port Arthur,
Texas

SRU544

Incinerator

M7E

0.12

52

TX3B1250

Valero Port Arthur
Refinery, Port Arthur,
Texas

SRU545

Incineralor

M7E

0.086

52

TX3B1310

Valero Refining, Bill
Greehey Refinery -
West Plant, Corpus
Christi, Texas

SRUl&:Sullien

Incinerator

\l7i:

0.093

44

TX3B1310

Valero Refining, Bill
Greehey Refinery -
West Plant, Corpus
Christi, Texas

SRI 3

Incineralor

M7i:

0.22

56

a This emissions unit was show 11 to be an oullier lor I lie dala sel and was removed from the draft
emissions factor anal\ sis

b Data is for same unil I rum same lesl report. Separale sels ol' lesl runs occurred on multiple days and
were reported separalcK

3.6 Sulfur Recovery I nits - I II(

The a\ ailable emissions lest data from the 2011 Refinery ICR included multiple test
reports for Tl IC from SRI units Each of the available test reports was reviewed, analyzed, and
summarized, and for those lesl reports included in the draft emissions factor analysis, given an
ITR score. An o\ erview of the draft emissions factor is provided in Table 12.

Based on the emissions lesl report review and analysis, 5 emissions test reports for 6
emissions units had useable data and were included in the development of the draft emissions
factor. Two test reports provide emissions test data for SRU that share a common stack. When
emissions testing is conducted on more than one SRU that share a common stack, the emissions
units are counted as one "unit"; the total emissions rate is divided by the total production rate of
all SRU venting to the stack when developing the units' average test results. The majority of the
testing was conducted since 2005, although one test report is from 1996.

The emissions test reports used in the draft factor analysis are provided in Table 13. A
complete list of the available test report information is provided in worksheet
"Test_Data_Sum_THC_SRU_2014Aug.xlsm". For more detail on the analysis and QA

24


-------
DRAFT - Do not cite or quote

conducted, see the field "QA Notes" for each test report. The ITR scores for these 5 test reports
ranged from 33 to 44. The emissions data (lb THC [as propane]/hr) in these test reports are
based on measurements taken with EPA Method 25A (M25A), and the test reports included
production rate data for the SRU in ton/hr sulfur production.

Certain test reports were excluded from the draft emissions factor analysis because
production rate data are not available or the concentration data for the test run average in the test
report is a negative value.

EPA's recommended emissions factor development procedures were followed for the
SRU THC data. All 6 SRU units have either an incinerator or a thermal oxidizer as the control
device. Both incinerators and thermal oxidizers work on the same principles of combustion, and
these terms are often used interchangeably by field staff. As such, there is no reason to believe
that these control devices would have differing 1e\ els of THC emissions. Therefore, all of these
units were combined for emissions factor de\ elopment. The statistical analysis for determining
outliers in the data set was conducted, and no data were found to be outliers. The draft emissions
factor is based on the emissions test data for 0 units and is characterized as Poorly
Representative. The draft emissions factor analysis lor SIU Tl IC is pro\ ided in spreadsheet
"EF Creation_THC_SRU_2014Aug.\lsm"

Table 12. Overview of the Draft Emissions Inetor for Tl IC from SRUs

Emissions tost data to use







No. of test
reports

No. of units

lest methods

Draft AIM2
Emissions I'aclor

Rep resell latn en ess

5

(i

IPX Method 25A

0.U47 lb THC [as
propane |/ton sulfur

Puui'ly

25


-------
DRAFT - Do not cite or quote

Table 13. Analysis of Emissions Test Reports for THC from SRUs

l iu-ililv II)
No.

liicililv 11:11110

Emissions
unit

AI'CI)

Test
mot hod

A\erase lesl
results. Ih
1 IK |:is
|)r()|)2iiie|/ton
sul I'll r

Ilk

LA3C0650

Valero Refining - New
Orleans, LLC at St.
Charles Refinery in
Norco, LA

SRU1600

Thermal
Oxidizer

M25A

5.9 x 10"3

34

OK2C0990

Total Petroleum, Inc.
Ardmore Refinery -
Ardmore, Oklahoma

SRU500A

Incinerator

V125A

1.8 x 10"3

37

TX3B1090

Total Petrochemicals
USA, Inc. in Port
Arthur, TX

SRU1&2

Thermal
()\idizer

\ 125 A

8.2 x 10"2

39

TX3B1110

BP Products North
America Inc. in Texas
City, TX

SRL

Incinerator

M25A

IS \ I01

33

TX3B1220

Motiva Enterprises,
LLC in Port Arthur,
TX

SRI 4

Incinerator

\ 12 5 A

1.2 x 10"3

44

TX3B1250

Valero Purl Arthur
Refinei'\ 111 Port
Arthur. T.\

SRI 544

Incinerator

\ 12 5 A

7.4 x 10"3

37

3.7 Hydrogen Plants-CO

The available emissions lesl data from the 2<) 11 Refinery ICR included multiple test
reports for CO from Hydro uui Plants, l-ach of the available test reports was reviewed, analyzed,
and summarized, and for those lesl reports included in the draft emissions factor analysis, given
an ITR score

Based on the emissions lesl report review and analysis, 3 emissions test reports for 3
emissions units had useable data and were available for inclusion in development of an emissions
factor. The emissions units lor which emissions data are available include 2 condensate stripper
vents (prior to returning water to the site feed water system) and 1 reformer furnace. The
production data for each of these emissions units are not on the same basis. Hydrogen
production data in scf/hr is available for 1 of the condensate stripper vents, and production data
in the form of Methane Feed Rate in scf/hr are available for the other condensate striper vent.
For the reformer furnace, heat input rate is available as the process activity rate. Because these
production data are not in comparable units and there is no way to calculate the production rate
data on the same basis, these test reports could not be combined for emissions factor
development. These useable emissions test reports are provided in Table 14. A complete list of
the available test report information is provided in worksheet

26


-------
DRAFT - Do not cite or quote

"Test_Data_Sum_CO_H2P_2014Aug.xlsm". For more detail on the analysis and QA
conducted, see the field "QA Notes" for each test report. The emissions data (lb CO/hr) in these
test reports are based on measurements taken with EPA Method 10 (M10).

Certain test reports were excluded from the draft emissions factor analysis because
production rate data are not available or the concentration data for the test run average in the test
report is a negative value.

Because the scope of this project is limited to data sets containing test averages from at
least 3 emissions units and there are 2 emissions units with useable test reports for the
condensate stripper vent and 1 reformer furnace with useable lesl data, but none of these units
have production rate data on the same basis, an emissions factor was not developed for CO for
Hydrogen Plants.

Table 14. Analysis of Emissions Test Reports lor CO from Ih Plants

l-'iicility
II) No.

lacilitv iiiiine

Emissions
unit

AP( 1)

list
mot hod

A\itsi»c tost
results

Ilk

Condensate stripper vent

AR3D0110

Lion Oil Company in El
Dorado, AR

( oiideiisate
stripper \ cm
i prior in
hoilcr water
I'eed system)

None

\110

0.48 lb
CO/MMscf H2
Production

22

NJ1A0850

ConocoPlullips ( ompaiis
Bayway Refiners.
ConocoPhillips (ompaii\
in Linden. N.l

Condensate
shipper \ em
(prior lo
boiler water
feed svstemi

None

M10

0.0011 lb
CO/scf methane
process feed

36

Reformer

CO4A0340



Plant 1

Hydrogen
Furnace stack

None

M10

0.00077 lb
CO/MMBtu

31

3.8 Hydrogen Plants - NOx

The available emissions test data from the 2011 Refinery ICR included multiple test
reports for NOx from Hydrogen Plant units. Each of the available test reports was reviewed,
analyzed, and summarized, and for those test reports included in the draft emissions factor
analysis, given an ITR score. An overview of the draft emissions factor is provided in Table 15.

Based on the emissions test report review and analysis, 8 emissions test reports for 7
emissions units had useable data and were included in the development of the draft emissions
factor. The emissions test reports used in the draft factor analysis are provided in Table 16. A
complete list of the available test report information is provided in worksheet

27


-------
DRAFT - Do not cite or quote

"Test_Data_Sum_NOx_H2P_2014Aug.xlsm". For more detail on the analysis and QA
conducted, see the field "QA Notes" for each test report. The ITR scores for these 8 test reports
ranged from 23 to 52. The emissions data (lb NOx/hr) in these test reports are based on
measurements taken with EPA Method 7E (M7E), and the test reports included activity rate data
for the Hydrogen Plant in MMBtu/hr heat input.

Certain test reports were excluded from the draft emissions factor analysis because heat
input data are not available or the emissions unit did not have controls consistent with the other
units (e.g. 1 emissions units had ultra-low NOx burners, and 1 emissions unit had selective
catalytic reductions controls).

EPA's recommended emissions factor development procedures were followed for the
Hydrogen Plant NOx data. None of the 7 units have controls lor NOx, and all were combined
for emissions factor development. The statistical analysis lor determining outliers in the data set
was conducted, and no data were found to be outliers

One of the last steps in developing an emissions factor is a comparison of the Factor
Quality Index (FQI) for different units. The FQI is an indicator of the emissions factor's ability
to estimate emissions for the entire national population, and it is related to both the ITR score
and the number of units in the data set Once the statistical procedures are complete, the data set
is ranked by ITR score (high to low), and a I Ql is developed for each unit in the candidate set.
The FQI should decrease with each emissions unit. When the FQI increases, only average test
values above the point where the FQI increases should be considered in the factor development.
In the development of the draft emissions factor for NOx from Hydrogen Plants, the FQI
evaluation excluded one unit from the data set. so the draft emissions factor is based on the
emissions test data for (•> units and is characterized as Moderately Representative. The draft
emissions factor analysis for NOx from I lydrouen Plants is |iro\ided in spreadsheet "IT
Creation NOx 112P 2<)|4.\uu xlsm"

Tsihle 15. Overview of I lie Drsil'l Emissions Inelor lor NOx from Hydrogen Phi ills

Emissions (est data to use







No. of lest
reports

No. ol' units

lest methods

Draft AP-42
Emissions I'actor

Representation ess

S

7"

1 PA Method 7E

0.081 lb
NOx/MMBtu

Moderately

a One Hydrogen Plant was excluded from the data set during the emissions factor calculation due to a low
ITR score and was removed from the draft emissions factor analysis; the draft emissions factor is based
on 6 Hydrogen Plants.

28


-------
DRAFT - Do not cite or quote

Table 16. Analysis of Emissions Test Reports for NOx from Hydrogen Plants

l-';icilil> II)
No.

l'acili(\ iiiiiiic

(-'.missions miii

\P< 1)

losl
humhod

A\era tic (est

results. II)
NOx/MMIilu

1 IK

AL3D0020
(2007 test)a

Hunt Refining,
Tuscaloosa, Alabama

Reformers A, B,
and C

None

M7E

0.016

23

AL3D0020
(2010 test)

Hunt Refining,
Tuscaloosa, Alabama

No. 2 Hydrogen
Plant Reformer -
indirect heaters

None

M7E

0.016

38

IL2A0430

ConocoPhillips Company,
Wood River Refinery
Hydrogen Plant in
Roxana, Illinois

Hydrogen Plant 1

None

\n:

0.041

45

MT4A0790

ExxonMobil Billings
Refinery, Billings,
Montana

F-551 ] lulrnuen
Plant Process
Heater/Fu rii; ice

Nunc

\r i:

0.17

45

OH2A0910

BP Husky Refining LLC,
Toledo, OH

Hydrogen Furnace

None

M7E

() ()90

52

MT4A0800
(2008 test)

Montana Refining
Company, Great Falls,
Montana

11\ drouen Plani
kcfurnicr 1 leater
IIIXIO

None

M7E

0.11

51

C04A0340

Suncor Enerus Tnc
Commerce ( il> Refiners.
Commerce ( il\. ( nlnradu

ll-:iu|

Nunc

\I7E

0.052

31

a This facility was excluded from l he data set during I he draft emissions factor analysis.

3.9 Hydrogen Plants - TIIC

The available emissions lest data from the 2
-------
DRAFT - Do not cite or quote

the available test report information is provided in worksheet

"Test_Data_Sum_THC_H2Plants_2014Aug.xlsm". For more detail on the analysis and QA
conducted, see the field "QA Notes" for each test report. The emissions data (lb THC [as
propane]/hr) in these test reports are based on measurements taken with EPA Method 25A
(M25A).

Certain test reports were excluded from the emissions factor analysis for the following
reasons: production rate data were not available or the test method was not compatible with
THC measurements taken with M25A (i.e., M18 test reports and SCAQMD M25.3 test reports
were excluded because these methods measure specific com pounds where M25A counts total
carbon).

Because the scope of this project is limited to data sets containing test averages from at
least 3 emissions units and because there are 2 emissions units with useable test reports for the
condensate stripper vent and 1 reformer furnace with useable test data, but none of these units
have production rate data on the same basis, ;in emissions factor was not developed for THC
from Hydrogen Plants.

Table 17. Analysis of Emissions l ost Reports for I IIC from Hydrogen Plants

r~:icilit>
II) No.

l iicililv Miiinc

Emissions
unit

A PC 1)

lost
mot hod

A\ersi»e lost
results

Ilk

Condensate stripper vent

AR3D0110

Lion Oil Conip;in>. 1 !l
Diirculii. \R

( ninlcnsalc
stripper \ cm
(priii r in
hoiler water
I'eed system)

None

M25A

1.1 lb THC |as

propane]/MMscf
H2 product

13

NJ1A0850

( nnocoPhillips (nmpam
1 »;i\ w :iv Refinery.
ConocoPhillips Compam.
Linden. NJ

Condensate
stripper vent
(prior to
boiler water
feed system)

None

M25A

0.0035 lb THC
[as propane]/scf
methane process
feed

36

Reformer

AL3D0020

Hunt Refining in
Tuscaloosa, AL

Reformer

None

M25A

0.00046 lb
THC/MMBtu

15

30


-------
DRAFT - Do not cite or quote

Section 4

Discussion of Proposed Revisions to SO2 Emissions Factors in AP-42

Section 8.13, Sulfur Recovery

In addition to adding new emissions factors for sulfur recovery plants, as described in
sections 3.4, 3.5, and 3.6 for CO, NOx, and THC, respectively, revisions are being proposed to
the SO2 emissions factors presented in the 1993 version of Table 8.13-1 in Section 8.13 of AP-
42. The current emissions factors were based on assumed a\ erage sulfur recovery efficiencies
instead of on a statistical evaluation of measured emissions dala. While this approach is
technically sound, the current emissions factors do not appear lo be consistent with current sulfur
recovery plant performance data because mid-range values were used rather than developing a
more statistically-based approach. The background document lor AP-42 section 8.133 presents
test data for 16 sulfur recovery plants. Nine of the lo plants had SO: emissions of approximately
2 kg/Mg sulfur produced, but the smallest emissions factor in the 1993 \crsion of Table 8.13-1
was 29 kg/Mg. The footnotes to Table 8.13-1 indicated that test data for 2-staged "controlled"
units showed 98.3 to 98.8 percent sulfur recovery and that 3-staged "controlled" units showed 95
to 99.9 percent sulfur recovery; using the mid-range value, the 2-staged controlled units have the
lowest emissions factor (29 kg/Mg versus 65 kg/Mg). From review of the background
document, it is unclear how these ranges were determined unless incineration was considered an
SO2 control (in which case all units tested had "controls"). The data presented in the background
document show thai the highest a\ erage run dala lor a sulfur recovery plant with a tailgas
cleanup units was 7 S kg Mg. so thai the lowest "controlled" emissions factor in Table 8.13-1 is
roughly 4 times the highest emissions results from a Claus unit with tailgas cleanup. Thus, the
"controlled" emissions factors in TaMe S 13-1 do not appear to be representative of the Claus
sulfur recovery plants with tail gas clean-up

Due to the issues identified with the current \ersion of Table 8.13-1, revisions are being
proposed to the table to more accurately present emissions factors for different types of sulfur
recovery plants based on specific SCC codes, which include the expected sulfur recovery
efficiencies for those sulfur reco\ cry plants. Revisions are also being proposed for the
discussion of tailgas "controls" to more clearly distinguish between tailgas cleanup units, which
enhance sulfur reco\ cry efficiencies, and incineration, which merely converts reduced sulfur
compounds to SO2.

The proposed rev isions to the emissions factors in Table 8.13-1 are still based on a mass
balance approach assuming that all sulfur not recovered is emitted as SO2. The emissions factors
in Table 8.13-1 are applicable to sulfur recovery plants that are followed by a thermal oxidizer,
incinerator, or other oxidative control system in which hydrogen sulfide or other reduced sulfur
compounds in the tailgas can be converted to SO2 prior to atmospheric release. Revisions are
being proposed to the Title of Table 8.13-1 to clarify this applicability. The proposed title for

3 The 1993 background document for sulfur recovery is entitled "Background Report, AP-42 Section 5.18, Sulfur
Recovery." With the publication of the Fifth Edition of AP-42, the Chapter and Section number for Sulfur Recovery
changed to 8.13.

31


-------
DRAFT - Do not cite or quote

Table 8.13-1 is "S02 EMISSION FACTORS FOR CLAUS SULFUR RECOVERY PLANTS
WITH OXIDATIVE CONTROL SYSTEMS "

Additionally, Table 8.13-1 does not currently provide applicable SCC codes for the sulfur
recovery plants described in the table, and the footnote showing the calculation of the emissions
factor is incorrectly presented. Therefore, the proposed version of Section 8.13 is updated to
specify applicable SCC codes and to correct the footnote equations in Table 8.13-1.

32


-------
DRAFT - Do not cite or quote

Section 5

Emissions Factor Development from Test Data Collected During the
Development of Parameters for Properly Designed and Operated

Flares

EPA has reviewed the emissions test data in multiple recent flare studies. Several of
these test reports are based on studies that resulted from various enforcement actions related to
flare performance issues. The EPA collected additional flare data during development of an
analysis of proper flare operating conditions (EPA 2d 12) The emissions data review and the
draft emissions factor development for each pollutant is described below.

The available emissions test data included multiple test reports lor CO from flares.
[Additional discussion of these test reports is included in EPA's Review of A\ ailable Documents
Report (EPA, 2014b).] Each of the available test reports was reviewed, analyzed, and
summarized, and given an ITR score An overview of the draft emissions factor is provided in

Based on the emissions test report re\ iew and analysis. (¦» emissions test reports for 8
flares had useable data and were included in the de\ eloquent of the draft emissions factor. The
flares tested include 7 steam-assisted llares and one air-assisted flare. The test data are based on
the measurement principle of passive Fourier Transform infrared (PFTIR). The emissions data
for flares consisted of I -minute CO concentralion-pathlength data for approximately 10 to 15 test
runs for each flare, l-ach lest run was approximately 15 to 20 minutes in duration.

The mass emissions of CO were calculated using a carbon balance as follows:

5.1 Flares - CO

Table 18.

r.,,,=C inlet

Where

Eco

emissions rate of carbon monoxide (lbs/hr).

C_inlet= mass flow of carbon in the flare vent gas sent to the flare (lb/hr).

[CO]

PFTIR measured CO concentration (ppm-m).

[CO2] = PFTIR measured CO2 concentration (ppm-m).

CE

Measured flare combustion efficiency

28

molecular weight of carbon monoxide (lb/lb-mole).

33


-------
DRAFT - Do not cite or quote

12 = molecular weight of carbon (lb/lb-mole).

Cinlet was determined based on the standard flow rate of the vent gas and the carbon
constituents of the vent gas. C inlet was calculated as follows:

C	inlet = Q, x ^ x V (MF x CMN )

4 MYC tr	7

Where:

C_inlet=	mass flow of carbon in the flare \ cut uas sent lo the flare (lb/hr).

Qfg =	volumetric flow rate of flare uas (standard cubic feet per hour; scf/hr).

12 =	molecular weight of carbon (lb lb-mole).

MVC =	molar volume correction factor (scf/lb-mole) = 385 5 scf/lb-mole.

MFX =	mole fraction of compound "\" in the ll are vent gas (unit I ess)

CMNx = Carbon mole number of compound "\" in the flare vent gas (mole carbon
atoms per mole compound) I- u. CMN for ethane (C2H6) is 2; CMN for
propane (C;I In) is 3

12 = molecular weight of carbon (lb lb-mole)

Because the flare testing was conducted to help identify conditions where flare
performance deteriorates, there were main test runs conducted at operating conditions that
resulted in poor llare combustion efficiencies. These operating conditions are not representative
of normal llare performance Properly operated llares achieve at least 98 percent destruction
efficiency in the flare plume The one-minute data were reviewed to determine if the combustion
efficiency was less than 96 5 percent (considered to be equivalent to a destruction efficiency of
98 percent) (I -1\\, 2014b). Any data that did not meet this combustion efficiency was excluded
from the analysis Additionally, any data from one-minute periods where the CO2 concentration
could not be measured or the CO: concentration was reported as zero were excluded from the
analysis because the CO mass emissions could not be calculated for that minute. For steam-
assisted flares, periods of time when there was no steam flow to the flare were eliminated
because this would not be representative of normal operations. All data for a given flare with
measurable one-minute CO2 concentrations, steam flow (for steam-assisted flares) and
acceptable combustion efficiencies were used to calculate an average emissions value (in CO
mass per heat input of vent gas) for the flare.

Some test reports included multiple values for CO2 measurements. These measurements
represent the CO2 values determined by the PFTIR operator at up to three different wavelengths.
Conversations with the PFTIR operator indicated that one of the CO2 wavelength measurements
(generally labeled lk) is not as reliable as the other two wavelength measurements (generally
labeled 765 and 2k). If data were available for either 765 or 2k, the lk CO2 measurements were

34


-------
DRAFT - Do not cite or quote

discarded from consideration. If data were not available for either 765 or 2k, the lk CO2
measurements were used in the emissions calculations. Because the 765 and 2k measurement
values should be fairly close to each other, if data were available for both 765 and 2k these two
measurements were generally averaged. But at times, the measurement for either 765 or 2k (but
not always the same one) would drop to zero or near to zero. To remove these readings in order
not to artificially decrease the value of CO2 used in the emissions calculations, the CO2 measured
value at 765 was compared to the FTIR's calculated CO2 error at 765 and the CO2 measured
value at 2k was compared to the FTIR's calculated CO2 error at 2k. If the measured value for
765 or 2k dropped below the corresponding instrument error, the value at that wavelength was
removed from the average CO2 value for that minute of data and only the value that remained
above the FTIR's calculated CO2 error was used in the emissions calculations.

The emissions test reports used in the factor analysis are provided in Table 19. The
available data from each test report included in the draft emissions factor analysis is provided in
worksheet "Flare Calculation.xlsx". The ITR scores for these 7 test reports ranged from 38 to
52. The emissions data (ppm-m CO) in these lesl reports are based on measurements taken with
passive FTIR and the activity rate data in the test reports included flare vent gas flow rates and
compositions, from which C inlet (lb C/hr) and the net heal input (MMBtu/hr) 10 the flare could
be calculated.

EPA's recommended emissions factor development procedures include guidelines for the
inclusion of previous emissions data when existing emissions factors are revised. The existing
data should be included alongside the new data prior to running any statistical tests. The ITR
score for the existing data is hased on the letter-rating of the data. There is a current AP-42
emissions factor for CO emissions from flares (see AP-42 section 13.5), and so the draft
emissions factor analysis includes the existing CO emissions data. Per the EPA's recommended
emissions factor development procedures, since the current factor is B-rated, an ITR score of 80
was assigned to the existing data

I- P.Vs recommended emissions factor de\ elopment procedures were followed for the
flare CO data. Potential subcategories w ere considered for the flare emissions data based on the
type of flare and based on the heat input \ alue to the flare. With respect to flare type, because
there are 7 steam-assisted flares and only 1 air-assisted flare and the statistical analysis for
determining whether the data are part of the same population requires at least 3 emissions units
in each category, the statistical analysis for subcategorization could not be performed. However,
since the current AP-42 emissions factors are based on emissions from both air-assisted and
steam-assisted flares, it is appropriate to combine the emissions from both types of flares for this
draft analysis as well. Subcategorization based on heat input was considered because some
states recommend separate emissions factors for flares with net heat input values above 1,000
Btu/scf. However, there were less than 3 flares with test data that included heat input values
above 1,000 Btu/scf, so the analysis could not be performed. Furthermore, because the current
AP-42 emissions factors do not distinguish between flares with different heat input values, the
data from all available flares was combined, regardless of vent gas heating value, for this draft
analysis. All 8 units from flare test reports under the current analysis were combined for
emissions factor development, along with the existing flare emissions data in AP-42. The
statistical analysis for determining outliers in the data set was conducted, and one emissions unit

35


-------
DRAFT - Do not cite or quote

was shown to be an outlier. The CO emissions were significantly higher for the outlier, by two
orders of magnitude, than the other test values in the data set. After removing the outlier
emissions unit from the data set, the outlier statistical analysis conducted on the remaining data
showed no additional outliers. The draft emissions factor is based on the current flare CO
emissions factor in AP-42 and the emissions test data for 7 additional units and is characterized
as Moderately Representative. The spreadsheet "EF Creation_CO_flare_2014Aug.xlsm"
provides the analysis for the draft emissions factor for CO emissions from flares.

Table 18. Overview of the Draft Emissions Factor for CO from Flares

Emissions lost data to use







No. of test
reports

No. of units

lest methods

Draft AP-42
Emissions I'actor

Represent ill neness

6

8 a

(Measurement
technique is
Passive FTIR)

() 34 Ih CO \l\IUiu

ModeraleK

a One flare was an outlier for the data set and was remo\ed I'rom I he draft emissions lac lor analysis. The
draft flare CO emissions factor is based on 7 emissions mills plus llie current flare CO emissions factor in
AP-42.

36


-------
DRAFT - Do not cite or quote

Table 19. Analysis of Emissions Test Reports for CO from Flares

l-';ii'ilil> II)
No.

l-'acilil\ iiiiiiu*

Emissions unii

losl
mclhori

A\CTilliO (csl

IVSIlllS. II)

( O/MMIilu

1 IK

FHR

FHRAU

Flint Hills Resources Port
Arthur, LLC in Port Arthur, TX

Flai'e AU
(steam-assisted)

PFTIR

0.23

38

FHR

FHRLOU

Flint Hills Resources Port
Arthur, LLC in Port Arthur, TX

Flare LOU
(steam-assisicd)

PFTIR

0.15

38

MI2A0710

MPCDET

Marathon Petroleum Company,
LLC, Detroit, MI

Flare CI'
(steam-assisicd)

PI Ilk

0.27

51

TX3B1210 a

MPCTX

Marathon Petroleum Company,
LLC, Texas Refining Division
in Texas City, TX

I'larc Mam
(stcani-assisied;

PI Ilk

88

51

INEOS

INEOS

INEOS ABS Corporation in
Addyston, OH

I'larc Pool
(steam-assisicd)

I'FTIR

0 :x

38

TX3B1260

SHELL

Shell Deer Park Refiners in
Deer Park, TX

I'lare LP
(sicam-assisicd)

HTIR

0.58

41

NA

TCEQ icsiiuu conducted at John
Zinkfacililv

I'larc
(sieani-assisied)

HTIR

0.31

52

NA

TCEQ test inn conducted al John
/.ink lacililv

I'larc
(air-assisied)

PFTIR

0.37

52

NA

Iaisiiuu \IM2 CO emissions
laclor lor Hares i( )l.l)i

I'larc

PFTIR

0.37

80

a This I laic linn was show n in he ;iii mil her lor the data set and was removed from the emissions factor analysis.

5.2 Flares - NOx

The available emissions lest data included multiple test reports for NOx from flares.
[Additional discussion of these test reports is included in EPA's Review of Available Documents
Report (EPA, 2014b).] I-ach of the available test reports was reviewed, analyzed, and
summarized, and for those test reports that are to be included in the emissions factor analysis,
given an ITR score. An overview of the emissions factor is provided in Table 20.

Based on the emissions test report review and analysis, 4 emissions test reports for 5
flares had useable data and were included in the development of the emissions factor. The flares
tested include 4 steam-assisted flares and one air-assisted flare. The emissions data for flares
consisted of 1-minute NOx concentration-pathlength data for approximately 10 to 15 test runs
for each flare. Each test run was approximately 15 to 20 minutes in duration.

37


-------
DRAFT - Do not cite or quote

The mass emissions of NOx were calculated as follows:

We.xMl^±Ml^xcE

[C02]x12

Where:

Enox	=	emissions rate of nitrogen oxides (lbs/hr).

Cinlet	=	mass flow of carbon in the flare \ cut gas sent to the flare (lb/hr).

[NO]	=	PFTIR measured NO concentration (ppni-ni)

[NO2]	=	PFTIR measured NO2 concentration (ppm-m)

30	=	molecular weight of N()(ll"> lb-mole).

46	=	molecular weight of NO: (lb I b-molc)

[CO2]	=	1'TIR measured CO: concentration 
-------
DRAFT - Do not cite or quote

12 = molecular weight of carbon (lb/lb-mole).

Because the flare testing was conducted to help identify conditions where flare
performance deteriorated, many of the test runs were conducted at operating conditions that
resulted in poor flare combustion efficiencies. These operating conditions are not representative
of normal flare performance. Properly operated flares achieve at least 98 percent destruction
efficiency in the flare plume. The one-minute data were reviewed to determine if the combustion
efficiency was less than of 96.5 percent (considered to be equivalent to a destruction efficiency
of 98 percent). Any data that did not meet this combustion efficiency was excluded from the
analysis. Additionally, any data from one-minute periods w here CO2 concentration could not be
measured or the CO2 concentration was reported as zero were excluded from the analysis
because the NOx mass emissions could not be calculated for lhat minute. For steam-assisted
flares, periods of time when there was no steam flow 10 the flare was eliminated because this
would not be representative of normal operations All data for a ui\ en flare with measurable
one-minute CO2 concentrations, steam flow (for si cam-assisted flares) and acceptable
combustion efficiencies were used to calculate an a\ erage emissions \ alue (in NOx mass per
heat input of vent gas) for the flare.

Some test reports included multiple values for CO: measurements. These measurements
represent the CO2 values determined In the PFTIR operator at up to three different wavelengths.
Conversations with the PFTIR operator indicated that one of the CO2 wavelength measurements
(generally labeled lk) is not as reliable as the other two wavelength measurements (generally
labeled 765 and 2k) Tf data were available for either 765 or 2k, the lk CO2 measurements were
discarded from consideration If data were not available for either 765 or 2k, the lk CO2
measurements were used in the emissions calculations. Because the 765 and 2k measurement
values should be fairly close to each other, if data were available for both 765 and 2k these two
measurements were generally a\ eraued Ikil at times, the measurement for either 765 or 2k (but
not always the same one) would drop to zero or near to zero. To remove these readings in order
not to artificially decrease the \ alue of CO: used in the emissions calculations, the CO2 measured
value at 7(o was compared to the FTIR's calculated CO2 error at 765 and the CO2 measured
value at 2k was compared to the FTIR's calculated CO2 error at 2k. If the measured value for
765 or 2k dropped below the corresponding instrument error, the value at that wavelength was
removed from the a\ erage CO: \ alue for that minute of data and only the value that remained
above the FTIR's calculated CO: error was used in the emissions calculations.

The emissions test reports used in the factor analysis are provided in Table 21. The
available data from each test report included in the draft emissions factor analysis is provided in
worksheet "Flare Calculation.xlsx". The ITR ratings for these 4 test reports ranged from 38 to
52. The emissions data (ppm-m NOx) in these test reports are based on measurements taken
with passive FTIR, and the activity rate data in the test reports included flare vent gas flow rates
and compositions, from which C inlet (lb C/hr) and the net heat input (MMBtu/hr) to the flare
could be calculated.

EPA's recommended emissions factor development procedures include guidelines for the
inclusion of previous emissions data when existing emissions factors are revised. The existing
data should be included alongside the new data prior to running any statistical tests. The ITR

39


-------
DRAFT - Do not cite or quote

score for the existing data is based on the letter-rating of the data. There is a current AP-42
emissions factor for NOx emissions from flares (see AP-42 section 13.5), and so the draft
emissions factor analysis includes the existing CO emissions data. Per the EPA's recommended
emissions factor development procedures, since the current factor is B-rated, an ITR score of 80
was assigned to the existing data.

EPA's recommended emissions factor development procedures were followed for the
flare NOx data. Potential subcategories were considered for the flare emissions data based on
the type of flare and based on the heat input value to the flare. With respect to flare type,
because there are 4 steam-assisted flares and only 1 air-assisted flare and the statistical analysis
for determining whether the data are part of the same population requires at least 3 emissions
units in each category, the statistical analysis for subcalcuoii/alion could not be performed.
However, since the current AP-42 emissions factors are based on emissions from both air-
assisted and steam-assisted flares, it is appropriate to combine the emissions from both types of
flares for this draft analysis as well. Subcategorization based on heat input was considered
because some states recommend separate emissions I actors for flares with net heat input values
above 1,000 Btu/scf. However, there were less than 3 flares with test data that included heat
input values above 1,000 Btu/scf, so the analysis could not be performed. Furthermore, because
the current AP-42 emissions factors do not distinguish between flares with different heat input
values, the data from all available flares u as combined, regardless of vent gas heating value, for
this draft analysis. All 5 units from llare test reports under the current analysis were combined
for emissions factor development, along with the existing llare emissions data in AP-42. The
statistical analysis for determining outliers in the data set determined that no outliers existed.
The draft emissions factor is bused on the current llare NOx emissions factor in AP-42 and the
emissions test data for 5 additional units and is characterized as Moderately Representative. The
spreadsheet "EF Creation NOx llare 2'> 14Aug xIsm" provides the analysis for the draft
emissions factor for N< )x emissions from flares

Table 20. Overview of (lie Drsil'l Emissions Factor for NOx from Flares

Emissions (est ilaln lo use







No. of (est
reports

No. of units

lest methods

Draft AP-42
Emissions l-':iclor

Kep resen 1 sit i\ en ess

4

5

(\ leasurement
technique is
Passive FTIR)

2.9 lb NOx/MMBtu

Moderately

40


-------
DRAFT - Do not cite or quote

Table 21. Analysis of Emissions Test Reports for NOx from Flares

l;icili(\ II)
No.

l-'iicilil\ iiiiiiK*

I'lmissioiis unit

Tcsl

HUMhod

\\er;i!ie losl

IVSIlllS. II)

N()\/MMUlu

1 IK

FHR

FHRAU

Flint Hills Resources Port Arthur,
LLC in Port Arthur, TX

Flare AU
(steam-assisted)

PFTIR

16

38

MI2A0710

MPCDET

Marathon Petroleum Company,
LLC, Detroit, MI

Flare CP
(steam-assisted)

PFTIR

0.011

51

INEOS

INEOS

INEOS ABS Corporation in
Addyston, OH

Flare I'oul
(steam-assisied)

HTIR

0.47

38

NA

TCEQ tests conducted at John
Zink facility

1'lai'c
(slcani-asMSlcd)

PI 1 IK

0.13

52

NA

TCEQ tests conducted at John
Zink facility

llare
(air-assisted)

PFTIR

U58

52

NA

Existing AP-42 NOx emissions
factor for lares (OLD)

Hare

I'PTIR

() ()(>S

80

5.3 Flares - VOC

The available emissions lest data included multiple lest reports for VOC related data from
flares. [Additional discussion of these test reports is included in liPA's Review of Available
Documents Report (I ¦ l\\. 2< > 14b).] Each of the a\ ai lable test reports was reviewed, analyzed, and
summarized, and for those test reports that are to he included in the emissions factor analysis,
given an TTR score An o\ er\ iew of the emissions factor is provided in Table 22.

IJased on the emissions test report review and analysis, 7 emissions test reports for 9
flares had useable data and were included in the development of the emissions factor. The flares
tested include S steam-assisted flares and one air-assisted flare. The PFTIR emissions data for
flares consisted of I-minute Tl IC and individual hydrocarbon concentration-pathlength data for
approximately I <> to 15 test runs for each flare. Each test run was approximately 15 to 20
minutes in duration The l)l.\l. data for flares consisted of multiple scans directly measuring the
mass emissions of C.i hydrocarbons. As the mass emissions of "C3+ hydrocarbons" was
directly reported in the l)l.\l. study, only the heat input to the flare had to be calculated. Data on
vent gas composition and flow rate were available to perform this calculation.

41


-------
DRAFT - Do not cite or quote

The mass emissions of VOC from the PFTIR tests were calculated as follows. Any
measurement data for methane and ethane were excluded from the VOC calculation:

£[HCx]xMWhc,

E = C inlet x —	x CE

voc "	[C02]x12

Where:

Evoc = emissions rate of volatile organic compounds (lbs/hr),

Cinlet = mass flow of carbon in the flare \ enl uas sent to the flare (lb/hr).

[HCx] = PFTIR measured hydrocarbon constituent "\" concentration (other than
methane or ethane) (ppm-111)

MWhcx = molecular weight of hydrocarbon constituent "\" (lb lb-mole).

[CO2] = PFTIR measured CO2 concentration (ppm-m).

12 = molecular weight of carbon (lb lb-mole)

CE = Measured flare combustion efficiency

C inlet was determined based on the standard flow rate of the vent gas and the carbon
constituents of the vent uas C inlet was calculated as follows

C

Where:
C inlet

Qfg =

12

MVC =
MFX =
CMNx =

12

inlet (V —— y(\ll- ' CM\ )

nut	*	*'

MVC

mass llow of carbon in the flare vent gas sent to the flare (lb/hr).
\ olumetric llow rate of flare gas (standard cubic feet per hour; scf/hr).
molecular weight of carbon (lb/lb-mole).

molar volume correction factor (scf/lb-mole) = 385.5 scf/lb-mole.

mole fraction of compound "x" in the flare vent gas (unitless)

Carbon mole number of compound "x" in the flare vent gas (mole carbon
atoms per mole compound). E.g., CMN for ethane (C2H6) is 2; CMN for
propane (C3H8) is 3.

molecular weight of carbon (lb/lb-mole).

42


-------
DRAFT - Do not cite or quote

Because the flare testing was conducted to help identify conditions where flare
performance deteriorated, there were many test runs conducted at operating conditions that
resulted in poor flare combustion efficiencies. These operating conditions are not representative
of normal flare performance. Properly operated flares achieve at least 98 percent destruction
efficiency in the flare plume. The one minute data were reviewed to determine if the combustion
efficiency was less than of 96.5 percent (considered to be equivalent to a destruction efficiency
of 98 percent). Any data that did not meet this combustion efficiency was excluded from the
analysis. Additionally, any data from one-minute periods where CO2 concentration could not be
measured or the CO2 concentration was reported as zero were excluded from the analysis
because the VOC mass emissions could not be calculated for that minute. For steam-assisted
flares, periods of time when there was no steam flow to the llare were eliminated because this
would not be representative of normal operations. All data lor a given flare with measurable
one-minute CO2 concentrations, steam flow (for steam-assisted llai es) and acceptable
combustion efficiencies were used to calculate an a\ erage emissions \ alue (in VOC mass per
heat input of vent gas) for the flare.

Some test reports included multiple values lor CO: measurements These measurements
represent the CO2 values determined by the PFTIR operator al up to three different wavelengths.
Conversations with the PFTTR operator indicated thai one of the CO2 wavelength measurements
(generally labeled lk) is not as reliable as the other two \\a\ elength measurements (generally
labeled 765 and 2k). If data were available for either 765 or 2k. the lk CO2 measurements were
discarded from consideration. If data were not a\ ailaMe lor either 765 or 2k, the lk CO2
measurements were used in the emissions calculations. Ikvanse the 765 and 2k measurement
values should be fairly close to each other, if data were a\ ailaMe for both 765 and 2k these two
measurements were generally averaged. But at times, the measurement for either 765 or 2k (but
not always the same one) would drop to zero or near to zero. To remove these readings in order
not to artificially decrease the \ alue of CO2 used in the emissions calculations, the CO2 measured
value at 7(o was compared to the 1 Tills calculated CO2 error at 765 and the CO2 measured
value at 2k was compared to the FTIR's calculated CO2 error at 2k. If the measured value for
765 or 2k dropped below the corresponding instrument error, the value at that wavelength was
removed from the average CO: value for that minute of data and only the value that remained
above the FTIR's calculated CO: error was used in the emissions calculations.

The emissions test reports used in the factor analysis are provided in Table 23. The
available data from each test report included in the draft emissions factor analysis is provided in
worksheet "Flare Calculation \ls\"\ The ITR scores for these 7 test reports ranged from 38 to
52. The emissions data (ppni-111 or lb/hr) in these test reports were based on measurements taken
with passive FTIR and DIAL, and the activity rate data in the test reports which included flare
vent gas flow rates and compositions, from which C inlet (lb C/hr) and the net heat input
(MMBtu/hr) to the flare could be calculated.

In the existing AP-42 section for Industrial Flares, there is an emissions factor for THC
(measured as methane equivalent), but there is no current emissions factor for VOC. Even
though THC is often used as a surrogate for VOC, the measurement methods for the two
compounds vary. In this case, the measurements for THC and VOC are not directly comparable.

43


-------
DRAFT - Do not cite or quote

As such, there is no existing emissions factor from AP-42 included in the current emissions
factor analysis.

EPA's recommended emissions factor development procedures were followed for the
flare VOC data. Potential subcategories were considered for the flare emissions data based on
the type of flare and based on the heat input value to the flare. With respect to flare type,
because there are 8 steam-assisted flares and only 1 air-assisted flare and the statistical analysis
for determining whether the data are part of the same population requires at least 3 emissions
units in each category, the statistical analysis for subcategorization could not be performed.
However, since the current AP-42 emissions factors are based on emissions from both air-
assisted and steam-assisted flares, it is appropriate to combine ilie emissions from both types of
flares for this draft analysis as well. Subcategorization bused on heat input was considered
because some states recommend separate emissions factors for lltires with net heat input values
above 1,000 Btu/scf. However, there were less limn 3 flares with lest data that included heat
input values above 1,000 Btu/scf, so the analysis could not be performed Furthermore, because
the current AP-42 emissions factors do not distinguish between flares with different heat input
values, the data from all available flares was combined, regardless of vent gtis heating value, for
this draft analysis. All 9 units from flare test reports under the current analysis were combined
for emissions factor development. The statistical analysis for determining outliers in the data set
was conducted, and one emissions unit was shown to be an outlier. The VOC emissions were
significantly lower for the outlier, by one order of magnitude, than the other test values in the
data set. After removing the outlier emissions unit for the data set, the outlier statistical analysis
conducted on the remaining data showed no additional outliers. The draft emissions factor is
based on the emissions test data for 8 units and is characterized as Moderately Representative.
The spreadsheet "El¦' Creation V()C_flare_2<)|4.\ug xlsm." provides the analysis for the draft
emissions factor for VOC emissions from flares

Tsihle 22. Overview of I lie Drsil'l Emissions Inetor lor VOC from l-'hircs

Emissions test to use







No. of lest
reports

No. of units

lest methods

Proposed AI>-42
Emissions I'iU'tor

Represent 
-------
DRAFT - Do not cite or quote

Table 23. Analysis of Emissions Test Reports for VOC from Flares

l-'iicilil> II)
No.

l ;icili(\ iiiiinc

I'lmissioiis unit

li-sl
iiuMliod

A\emtio lost

IVSIlllS. II)

\()( VMM Bin

1 IK

FHR

FHRAU

Flint Hills Resources Port
Arthur, LLC in Port Arthur, TX

Flare AU
(steam-assisted)

PFTIR

0.50

38

FHR

FHRLOU

Flint Hills Resources Port
Arthur, LLC in Port Arthur, TX

Flare LOU
(steam-assisted)

PFTIR

0.95

38

MI2A0710

MPCDET

Marathon Petroleum Company,
LLC, Detroit, MI

Flare CI'
(steam-assisiedi

1'1'HR

0.42

51

TX3B1210 a

MPCTX

Marathon Petroleum Company,
LLC, Texas Refining Division in
Texas City, TX

I'lare Mam
(sieam-asbisled)

1'irik

0.016

51

INEOS

INEOS

INEOS ABS Corporation in
Addyston, OH

I'lare Pool
fsteani-assisiedi

I'FTIk

u "0

38

TX3B1260

SHELL

Shell Deer Park Refiners in Deer
Park, TX

1 hue LP
(sieam-assisied)

PFTIR

0.53

41

NA

TCEO lesiiim cmiducled al John
Zink I'aalilv

I'lare
(sieam-abbisled)

PFTIR

0.59

52

NA

TCEQ Icsiiim cmiducled al John
Zinkfacilil>

I'lare
(air-assisted)

PFTIR

0.47

52

TX3B1I lu

131'

l evis ( n\. 1 \

I 'lare No. 6
(steam-assisted)

DIAL

0.25

40

a This I'lare was ;in outlier fur I lie dala sol and \\;is removed from the draft emissions factor analysis.

45


-------
DRAFT - Do not cite or quote

Section 6
References

Air Alliance Houston, et al. v. McCarthy, No. l:13-cv-00621-KBJ (D.D.C.).

EPA (U.S. Environmental Protection Agency). 1995. Compilation of Air Pollutant Emission
Factors, Volume 1: Stationary Point and Area Sources, AP-42, Fifth Edition, U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards. January
1995. Available at: http://www.epa.gov/ttn/chief/ap42 index.html

EPA (U.S. Environmental Protection Agency). 2011. Information Collection Request for
Petroleum Refinery Sector New Source Performance Standards (NSPS) and National
Emissions Standards for Hazardous Air Pol I nuints (NESHAP) Risk and Technology
Review. April 2011.

EPA (U.S. Environmental Protection Agency). 2<)|2 I'arameiersfor Properly Designed and
Operated Flares. Prepared for U.S. Environ menial Protection Agency, Office of Air Quality
Planning and Standards, Research Triangle Park. \C April 2012. Available at:
http://www.epa.gov/airtoxics/flaie 2" 12llaretechienoil ixlf

EPA (U.S. Environmental Protection Agency). 2d I 3 Recommended Procedures for

Development of Emissions Factors ami I se <>J the We hi IRE Database. U.S. Environmental
Protection Agency. Office of Air Quality Planning and Standards. August 2013 (Draft
Final). Available at http:/ www cna.go\ tin chief efpac/procedures/index.html

EPA (U.S. Environmental Protection Agency) 2<)|3a Worksheet
"EF Creation (pollutant) (emissionssoiiive)\lsm".

EPA(l S. I ji\ironmental Protection Agency) 2<) I .>h Worksheet "Web fire-
template (pollutant) (cmissionssource)_(lacilily lD)_(unitID)",

EPA (U.S. En\ iron menial Protection Agency). 2014a. Draft Residual Risk Analysis for the
Petroleum Refinery Source Sector. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards. May 2014. (Draft). Available at:
http://www.regulalions uo\ 'documentDetail:D=EPA-HO-OAR-2010-0682-0225

EPA (U.S. Environmental Protection Agency). 2014b. Review of Available Documents for
Developing Proposed Emissions Factors for Flares, Tanks, and Wastewater Treatment
Systems. U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards. August 2014. Available at:

http://www.epa.gov/ttn/chief/consentdecree/draft report review.pdf

46


-------
DRAFT - Do not cite or quote

RTI (RTI International). 2011. Emission Estimation Protocol for Petroleum Refineries.
Version 2.1.1, Final ICR Version Corrected. Prepared for U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC. May
2011. Docket Item No. EPA-HQ-OAR-2010-0682-0060. Most current version also
available at http://www.epa.gov/ttn/chief/efpac/protocol/.

47


-------
DRAFT - Do not cite or quote

Appendix A

EMISSIONS TEST REPORT DATA FIELDS INCLUDED IN TEST

DATA SUMMARY FILES


-------
Appendix A. Data Fields in the Test Data Summary Files

Table column

I'ield name

A

Column

B

Facility ID Number

C

Unit ID Number

D

APCD ID(s)

E

Combustion controls used to reduce air pollution (from combustion sources)

F

General Description

G

Code for Process Unit Type

H

Test Report ID

I

Test Date (mm/dd/yyyy)

J

Pollutant Name

K

Pollutant CAS No.

L

Pollutant Class

M

Test Method

N

Run 1 Hourly Production Rate (value)

0

Run 2 Hourly Production Rate (value)

P

Run 3 Hourly Production Rate (value)

Q

Average Hourly Production Rate (value)

R

Hourly Production Rate (units)

S

Production comment

T

Run 1 Hourly Production Rate (value)

U

Run 2 Hourly Production Rate (value)

V

Run 3 Hourly Production Rate (value)

w

Average Hourly Production Rate (value)

X

Hourly Production Rate (units)

Y

Production comment

Z

Run 1 Airflow Rate Outlet (acfim)

AA

Run 1 Airflow Rate Outlet (scfm)

AB

Run 1 Airflow Rate Outlet (dscfm)

AC

Run 1 Gas Moisture Outlet (%)

AD

Run 1 Gas Temp Outlet (F)

AE

Run 1 Gas Pressure Outlet (in. Hg)

AF

Run 1 Gas Oxygen Outlet (%)

AG

Run 1 Gas C02 Outlet (%)

AH

Run 2 Airflow Rate Outlet (acfim)

AI

Run 2 Airflow Rate Outlet (scfm)

AJ

Run 2 Airflow Rate Outlet (dscfm)

AK

Run 2 Gas Moisture Outlet (%)

AL

Run 2 Gas Temp Outlet (F)

AM

Run 2 Gas Pressure Outlet (in. Hg)

AN

Run 2 Gas Oxygen Outlet (%)


-------
Tsihlc col ii in 11

l-lcld nsiiiic

AO

Run 2 Gas C02 Outlet (%)

AP

Run 3 Airflow Rate Outlet (acfm)

AQ

Run 3 Airflow Rate Outlet (scfm)

AR

Run 3 Airflow Rate Outlet (dscfm)

AS

Run 3 Gas Moisture Outlet (%)

AT

Run 3 Gas Temp Outlet (F)

AU

Run 3 Gas Pressure Outlet (in. Hg)

AV

Run 3 Gas Oxygen Outlet (%)

AW

Run 3 Gas C02 Outlet (%)

AX

Average Airflow Rate Outlet (acfm)

AY

Average Airflow Rate Outlet (scfm)

AZ

Average Airflow Rate Outlet (dscfm)

BA

Average Gas Moisture Outlet (%)

BB

Average Gas Temp Outlet (F)

BC

Average Gas Pressure Outlet (in. Hg)

BD

Average Gas Oxygen Outlet (%)

BE

Average Gas C02 Outlet (%)

BF

Run 1 Outlet concentration

BG

Run 1 Outlet concentration units

BH

Run 1 Outlet Detect Flag

BI

Run 1 Outlet (lb/hr)

BJ

Run 2 Outlet concentration

BK

Run 2 Outlet concentration units

BL

Run 2 Outlet Detect Flag

BM

Run 2 Outlet (lb/hr)

BN

Run 3 Outlet concentration

BO

Run 3 Outlet concentration units

BP

Run 3 Outlet Detect Flag

BQ

Run 3 Outlet (lb/hr)

BR

Average Outlet concentration

BS

Average Outlet concentration units

BT

Count Outlet Non-Detect Runs

BU

Average Outlet (lb/hr)

BV

Sampling comments

BW

Analytical comments

BX

QA Comments

BY

Other comments

DA

QA Notes

DB

RTI Reviewer initials

DC

Looked at for EF?

DD

Used in EF?

DE

see


-------
Tsihlc col ii in 11

l-lcld nsiiiic

DF

NEI POLLUTANT CODE

DG

PROCESS DESCRIPTION

DH

CONTROL CODE1

DI

CONTROL CODE2

DJ

MDL

DK

FACTOR

DL

UNIT

DM

MEASURE

DN

MATERIAL

DO

ACTION

DP

FLAG

DQ

TEST REPORT RATING

DR

REF ID

DS

REFERENCE TEXT

DT

No. pages


-------
DRAFT - Do not cite or quote

Appendix B

EPA'S "TEST QUALITY RATING TOOL" TEMPLATE

(ITR TEMPLATE)

August 2013


-------


A

B I G

H I N

1

Name of Facility where the test was performed





2

Name of Company performing stationary source test



3

SCC of tested unit or units



4

Name of assessor and name of employer.



5

Name of regulatory assessor and regulatory agency name.

NA

6







7

8

Emissions Factor Development Quality Indicator Value Rating 0

10

/ $/ /

Supporting Documentation Provided / & / Regulatory Agency Review /

/ <5* / / o

/«?/ A*

V/

^5 / Justification
/

11

General



12

As described in ASTM D7036-12 Standard Practice for
Competence of Air Emission Testing Bodies, does the testing
firm meet the criteria as an AETB or is the person in charge of
the field team a Ql for the type of testing conducted? A
certificate from an independent organization (e.g., Stack
Testing Accreditation Council (STAC), California Air
Resources Board (CARB), National Environmental Laboratory
Accreditation Program (NELAP)) or self declaration provides
documentation of competence as an AETB.



As described in ASTM D7036-12 Standard Practice for
Competence of Air Emission Testing Bodies, does the testing
firm meet the criteria as an AETB or is the person in charge of
the field team a Ql for the type of testing conducted? A
certificate from an independent organization (e.g., STAC,
CARB, NELAP) or self declaration provides documentation of
competence as an AETB.





13





Was a representative of the regulatory agency on site during
the test?





14

Is a description and drawinq of test location provided?



Is a description and drawinq of test location provided?





15

Has a description of deviations from published test methods
been provided, or is there a statement that deviations were not
required to obtain data representative of typical facility
operation?



Is there documentation that the source or the test company
sought and obtained approval for deviations from the
published test method prior to conducting the test or that the
tester's assertion that deviations were not required to obtain
data representative of operations that are typical for the
facility?





16





Were all test method deviations acceptable?





17

Is a full description of the process and the unit being tested
(including installed controls) provided?



Is a full description of the process and the unit being tested
(including installed controls) provided?





18

Has a detailed discussion of source operating conditions, air
pollution control device operations and the representativeness
of measurements made during the test been provided?



Has a detailed discussion of source operating conditions, air
pollution control device operations and the representativeness
of measurements made during the test been provided?





19

Were the operating parameters for the tested process unit and
associated controls described and reported?



Is there documentation that the required process monitors
have been calibrated and that the calibration is acceptable?





20





Was the process capacity documented?





21





Was the process operating within an appropriate range for the
test program objectives?





22





Were process data concurrent with testing?





23





Were data included in the report for all parameters for which
limits will be set?





24

Is there an assessment of the validity, representativeness,
achievement of DQO's and usability of the data?



Did the report discuss the representativeness of the facility
operations, control device operation, and the measurements of
the target pollutants, and were any changes from published
test methods or process and control device monitoring
protocols identified?





25

Have field notes addressing issues that may influence data
quality been provided?



Were all sampling issues handled such that data quality was
not adversely affected?





26

Manual Test Methods



27

Have the following been included in the report:









28

Dry gas meter (DGM) calibrations, pitottube and nozzle
inspections?



Was the DGM pre-test calibration within the criteria specified
by the test method?





29





Was the DGM post-test calibration within the criteria specified
by the test method?





30





Were thermocouple calibrations within method criteria?





31





Was the pitottube inspection acceptable?





32





Were nozzle inspections acceptable?





33





Were flow meter calibrations acceptable?





34

Was the Method 1 sample point evaluation included in the
report?



Were the appropriate number and location of sampling points
used?





35

Were the cyclonic flow checks included in the report?



Did the cyclonic flow evaluation show the presence of an
acceptable average gas flow angle?





36

Were the raw sampling data and test sheets included in the
report?



Were all data required by the method recorded?





37





Were required leak checks performed and did the checks meet
method requirements?





38





Was the required minimum sample volume collected?





39





Did probe, filter, and impinger exit temperatures meet method
criteria (as applicable)?





AppB_WF_ITR_POL_UNIT_Fac_template.xlsx
Test Quality Rating Tool


-------


A

B

G

H

N

40



¦flH

Did isokinetic sampling rates meet method criteria?





41





Was the sampling time at each point greater than 2 minutes
and the same for each point?





42

Did the report include a description and flow diagram of the
recovery procedures?



Was the recovery process consistent with the method?





43





Were all required blanks collected in the field?





44





Where performed, were blank corrections handled per method
requirements?





45





Were sample volumes clearly marked on the jar or measured
and recorded?





46

Was the laboratory certified/accredited to perform these
analyses?



Was the laboratory certified/accredited to perform these
analyses?





47

Did the report include a complete laboratory report and flow
diagram of sample analysis?



Did the laboratory note the sample volume upon receipt?





48





If sample loss occurred, was the compensation method used
documented and approved for the method?





49





Were the physical characteristics of the samples (e.g., color,
volume, integrity, pH, temperature) recorded and consistent
with the method?





50





Were sample hold times within method requirements?





51





Does the laboratory report document the analytical procedures
and techniques?





52





Were all laboratory QA requirements documented?





53





Were analytical standards required by the method
documented?





54



lllM

Were required laboratory duplicates within acceptable limits?





55





Were required spike recoveries within method requirements?





56



I1I1B

Were method-specified analytical blanks analyzed?





57



If problems occurred during analysis, is there sufficient

umentation to conclude that the problems did not adversely
affect the sample results?





58





Was the analytical detection limit specified in the test report?





59





Is the reported detection limit adequate for the purposes of the
test proqram?





60

Were the chain-of-custody forms included in the report?



Do the chain-of-custody forms indicate acceptable
management of collected samples between collection and
analysis?





61

Instrumental Test Methods



62

Have the following been included in the report:

I	

63

Did the report include a complete description of the
instrumental method sampling system?



Was a complete description of the sampling system provided?





64

Did the report include calibration gas certifications?



Were calibration standards used prior to the end of the
expiration date?





66

Did report include interference tests?



Did interference checks meet method requirements?





67

Were the response time tests included in the report?



Was a response time test performed?





68

Were the calibration error tests included in the report?



Did calibration error tests meet method requirements?





69

Did the report include drift tests?



Were drift tests performed after each run and did they meet
method requirements?





70

Did the report include system bias tests?



Did system bias checks meet method requirements?





71

Were the converter efficiency tests included in the report?



Was the NOX converter test acceptable?





72

Did the report include stratification checks?



Was a stratification assessment performed?





73

Did the report include the raw data for the instrumental
method?



Was the duration of each sample run within method criteria?





74



Ml

Was an appropriate traverse performed during sample
collection, or was the probe placed at an appropriate center
point ("if allowed by the method)?





75





Were sample times at each point uniform and did they meet
the method requirements?





76





Were sample lines heated sufficiently to prevent potential
adverse data quality issues?





77





Was all data required by the method recorded?





88

89

90

91

92

93

Total
Manual Test 0
Instrumental Test 0

AppB_WF_ITR_POL_UNIT_Fac_template.xlsx
Test Quality Rating Tool


-------