EPA453/R-93-026
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-453/R-93-026
JUNE 1993
AIR
Protocol for
Equipment Leak
Emission Estimates
-------
EPA-453/R-93-026
Protocol for Equipment Leak
Emission Estimates
Emission Standards Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
June 1993
-------
This report has been reviewed by the Emission Standards
Division of the Office of Air Quality Planning and Standards, EPA
and approved for publication. Mention of trade names or
commercial products is not intended to constitute endorsement or
recommendation for use. Copies of this report are available
through the Library Services Office (MD-35), U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina 27711;
from the Office of Air Quality Planning and Standards Technology
Transfer Network, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina 27711; or, for a fee, from the
National Technical Information Services, 5285 Port Royal Road,
Springfield, Virginia 22161.
Publication No. EPA-453/R-93-026
-------
FOREWORD
The EPA's protocol for estimating equipment leak emissions
is the result of detailed information gathering and data
analysis. The protocol was written to provide a thorough
understanding of acceptable approaches to generating process
unit-specific emission estimates. In preparing this document,
EPA has encouraged knowledgeable individuals in industry and the
regulatory community to provide comments.
The EPA has put forth considerable effort to make this
document as comprehensive as possible. However, it should be
understood that not all details and topics pertaining to
equipment leaks could feasibly be included in this document.
Additionally, it should be understood that the protocols
presented in this document are not necessarily suitable for all
applications. There will be cases where it will be necessary for
the user of the document to make a professional judgement as to
the appropriate technical approach for collecting and analyzing
data used to estimate equipment leak emissions.
Additional data on equipment leak emissions continues to be
collected. It is the intent of the EPA to periodically update
this document after analysis of the data warrants such an update.
For example, data is presently being collected in refineries, and
this data may be used to revise the existing refinery factors and
correlations, which are based on data collected in the late
1970s. Furthermore, as new techniques for collecting and
analyzing data are developed, they will be included in updated
versions of this document.
Mention of any manufacturer or company name within this
document does not represent endorsement by the EPA.
-------
TABLE OF CONTENTS
Section Page
FOREWORD i
1.0 INTRODUCTION 1-1
2.0 DEVELOPMENT OF EQUIPMENT LEAK EMISSION ESTIMATES .... 2-1
2.1 Introduction 2-1
2.2 General Information on the Approaches for
Estimating Equipment Leak Emissions 2-2
2.2.1 Equipment Leak Emission Estimation
Approaches 2-2
2.2.2 Overview of Equipment Leak Data
Collection 2-5
2.3 Approaches for Estimating Equipment Leak
Emissions 2-8
2.3.1 Average Emission Factor Approach 2-9
2.3.2 Screening Ranges Approach 2-15
2.3.3 EPA Correlation Approach 2-20
2.3.4 Unit-Specific Correlation Approach .... 2-30
2.4 Special Topics 2-37
2.4.1 Speciating Emissions 2-37
2.4.2 Using Response Factors 2-39
2.4.3 Monitoring Instrument Type and Calibration
Gas 2-43
2.4.4 Estimating Emissions for Equipment Not
Screened 2-44
2.4.5 Using Screening Data Collected at Several
Different Times 2-44
2.4.6 Estimating VOC Emission Rates from Equipment
Containing Non-VOC's 2-45
2.4.7 Estimating Equipment Leak Emissions of
Inorganic Compounds 2-46
2.5 References 2-48
3.0 SOURCE SCREENING 3-1
3.1 Introduction 3-1
3.2 Monitoring Instruments 3-1
3.2.1 Operating Principles and Limitations of
Portable VOC Detection Devices 3-2
3.2.2 Specifications and Performance Criteria of
Portable VOC Detection Devices 3-4
3.2.3 Use of Monitoring Devices That Do Not Meet
Method 21 Requirements 3-10
3.3 The Screening Program 3-12
3.3.1 Identification of Equipment to be Screened 3-12
3.3.2 Procedure for Screening 3-14
3.3.3 Data Handling 3-23
3.4 References 3-27
ii
-------
TABLE OF CONTENTS (Continued)
Section page
4.0 MASS EMISSIONS SAMPLING 4-1
4.1 Introduction 4-1
4.2 Sampling Methods 4-1
4.2.1 Vacuum Method 4-4
4.2.2 Blow-Through Method 4-8
4.3 Source Enclosure 4-12
4.3.1 Valves 4-14
4.3.2 Pumps and Agitators 4-14
4.3.3 Compressors 4-16
4.3.4 Connectors 4-16
4.3.5 Relief Valves 4-17
4.4 Analytical Techniques 4-17
4.4.1 Analytical Instrumentation 4-17
4.4.2 Calibration of Analytical Instruments . . 4-18
4.4.3 Analytical Techniques for Condensate . . . 4-18
4.4.4 Calibration Procedures for the Portable
Monitoring Instrument 4-20
4.5 Quality Control and Quality Assurance
Guidelines 4-20
4.5.1 Quality Control Procedures 4-20
4.5.2 Quality Assurance Procedures 4-24
4.6 References 4-28
5.0 ESTIMATION OF CONTROL EFFICIENCIES FOR EQUIPMENT LEAK
CONTROL TECHNIQUES 5-1
5.1 Introduction 5-1
5.2 Equipment Modification Control Efficiency 5-1
5.2.1 Closed-Vent Systems 5-3
5.2.2 Pumps 5-3
5.2.3 Compressors 5-4
5.2.4 Pressure Relief Valves 5-4
5.2.5 Valves 5-5
5.2.6 Connectors 5-5
5.2.7 Open-Ended Lines 5-5
5.2.8 Sampling Connections 5-6
5.3 Leak Detection and Repair Control Effectiveness . . 5-6
5.3.1 Approach for Estimating LDAR Control
Effectiveness 5-10
5.3.2 Example Application of Approach 5-25
5.4 References 5-29
iii
-------
TABLE OF CONTENTS (Continued)
Section Page
APPENDICES
A Example Calculations A-l
B Leak Rate Screening Value Correlation Development and
Revision of SOCMI Correlations and Emission Factors . . B-l
C Response Factors C-l
D Selection of Sample Size for Screening Connectors . . . D-l
E Reference Method 21 E-l
F Development of Leak Rate Versus Fraction Leaking
Equations and Determination of LDAR Control
Effectiveness F-l
IV
-------
LIST OF TABLES
Table Page
2-1 SOCMI Average Emission Factors 2-10
2-2 Refinery Average Emission Factors 2-11
2-3 Gas Plant Average Emission Factors 2-12
2-4 SOCMI Screening Value Range Emission Factors . . 2-16
2-5 Refinery Screening Value Range Emission Factors . 2-17
2-6 Gas Plant Screening Value Range Emission Factors 2-18
2-7 SOCMI Leak Rate/Screening Value Correlations . . 2-21
2-8 Refinery Leak Rate/Screening Value Correlations . 2-22
2-9 Default-Zero Values 2-29
2-10 Greater than 100,000 ppmv Screening Value Emission
Factors for SOCMI and Refinery Process Units . . 2-34
3-1 Performance Criteria for Porvoc Detectors 3-6
3-2 Porvoc Detection Instruments 3-11
3-3 Equipment Leak Emission Sources 3-13
3-4 Example Field Sheets for Equipment Screening Data 3-26
4-1 Calculation Procedures for Leak Rate when Using
the Vacuum Method 4-9
4-2 Calculation Procedures for Leak Rate when Using
the Blow-Through Method 4-13
4-3 Example GC Calibration Data Sheet 4-19
4-4 Example Data Collection Form for Fugitive
Emissions Bagging Test (Blow-Through Method) . . 4-21
4-5 Example Data Collection Form for Fugitive
Emissions Bagging Test (Vacuum Method) 4-22
4-6 Example Drift Test Report Form 4-27
5-1 Summary of Equipment Modifications 5-2
5-2 Control Effectiveness for an Ldar Program at a
SOCMI Process Unit 5-8
-------
LIST OF TABLES (Continued)
Table Page
5-3 Control Effectiveness for an Ldar Program at a
Refinery Process Unit 5-9
5-4 Equations Relating Average Leak Rate to Fraction
Leaking at SOCMI Units 5-19
5-5 Equations Relating Average Leak Rate to Fraction
Leaking at Refinery Units 5-20
5-6 Values Used in Example Calculation 5-26
5-7 Example Calculation to Determine the Final Leak
Frequency of SOCMI Gas Valves in a Monthly
Monitoring LDAR Program with a Leak Definition
of 10,000 ppmv 5-28
VI
-------
LIST OF FIGURES
Figure Page
2-1 Overview of Data Collection and Analysis
Approaches for Developing Equipment Leak
Emissions Inventory 2-4
2-2 SOCMI Correlations Relating Total Organic Compound
(TOC) Leak Rate to Screening Value:
0 - 1,000 ppmv 2-23
2-3 SOCMI Correlations Relating Total Organic Compound
(TOC) Leak Rate to Screening Value:
1,000 - 1,000,000 ppmv 2-24
2-4 Refinery Correlations Relating Non-Methane Organic
Compound (NMOC) Leak Rate to Screening Value:
0 - 1,000 ppmv 2-25
2-5 Refinery Correlations Relating Non-Methane Organic
Compound (NMOC) Leak Rate to Screening Value:
1,000 - 1,000,000 ppmv 2-26
3-1 Gate Valves 3-17
3-2 Globe Valves 3-18
3-3 Lubricated Plug Valve 3-19
3-4 Ball Valve and Butterfly Valve 3-20
3-5 Weir-Type Diaphragm Valve and Check Valves ... 3-21
3-6 Centrifugal Pumps 3-22
3-7 Spring-Loaded Relief Valve 3-24
4-1 Sampling Train for Bagging a Source Using the
Vacuum Method 4-5
4-2 Equipment Required for the Blow-Through Sampling
Technique 4-10
5-1 SOCMI Gas Valve Average Leak Rate Versus Fraction
Leaking at Several Leak Definitions 5-11
5-2 SOCMI Light Liquid Valve Average Leak Rate Versus
Fraction Leaking at Several Leak Definitions . . 5-12
5-3 SOCMI Light Liquid Pump Average Leak Rate Versus
Fraction Leaking at Several Leak Definitions . . 5-13
vii
-------
LIST OF FIGURES (Continued)
Figure page
5-4 SOCMI Connector Average Leak Rate Versus Fraction
Leaking at Several Leak Definitions 5-14
5-5 Refinery Gas Valve Average Leak Rate Versus
Fraction Leaking at Several Leak Definitions . . 5-15
5-6 Refinery Light Liquid Valve Average Leak Rate
Versus Fraction Leaking at Several Leak
Definitions 5-16
5-7 Refinery Light Liquid Pump Average Leak Rate
Versus Fraction Leaking at Several Leak
Definitions 5-17
5-8 Refinery Connector Average Leak Rate Versus
Fraction Leaking at Several Leak Definitions . . 5-18
5-9 Simplified Graphical Presentation of Changes in
Leak Frequency After Implementation of an
LDAR Program 5-22
viii
-------
1.0 INTRODUCTION
This document is an update to the original EPA equipment
leaks protocol document ("Protocols for Generating Unit-Specific
Emission Estimates for Equipment Leaks of VOC and VHAP,"
EPA-450/3-88-010, October, 1988). The purpose of this document
is the same as the original protocol document - to present
standard protocols for estimating mass emissions from equipment
leaks. However, this document publishes the results of
additional data collection and analysis that has occurred since
the original protocol was published, and also expands on the
topics that were covered in the original protocol.
Four particular items to note regarding this updated
protocol are:
(1) New correlations applicable to the synthetic organic
chemical manufacturing industry (SOCMI) that relate
screening values obtained using a portable monitoring
instrument to mass emissions have been developed.
These new SOCMI correlations have been used to revise
the SOCMI emission factors.
(2) The document has been expanded to include source
category-specific information on estimating emissions
from petroleum refineries and natural gas liquid
plants.
(3) The use of response factors when estimating equipment
leak emissions has been addressed.
(4) A chapter has been added that provides information on
estimating the control efficiency of equipment leak
control techniques.
As with the original protocol document, this document
presents standard protocols for general use in generating unit-
specific emission estimates for permitting and inventories. The
document describes methodologies the EPA considers appropriate
for development of equipment leak emission estimates. These
methodologies are intended to assist States and industry in their
efforts to estimate equipment leak emissions.
The updated protocol is divided into five chapters and
several appendices. Chapter 2.0 describes how to estimate
1-1
-------
equipment leak emissions. Chapter 3.0 describes collecting
screening data that can be used in the emission estimates.
Chapter 4.0 describes collecting unit-specific equipment leak
mass emissions data. Chapter 5.0 describes how to estimate the
control efficiencies of equipment leak control techniques. The
appendices support information contained in the chapters. Each
of these chapters and the appendices are briefly described below.
In Chapter 2.0, four different approaches for estimating
equipment leak emissions are described. These approaches are:
• Average Emission Factor Approach;
• Screening Ranges Approach;
• EPA Correlation Approach; and
• Unit-Specific Correlation Approach.
Additionally, several topics that are relevant to estimating
equipment leak emissions are addressed. These topics include
speciating equipment leak emissions of individual compounds from
an equipment piece containing a mixture, using response factors,
estimating emissions of inorganic compounds, and other topics not
specifically related to any one of the four approaches in
particular.
In Chapter 3.0, information is provided on how to perform a
screening survey at a process unit. Requirements for the use of
a portable monitoring instrument are described. These
requirements are based on EPA Method 21. Additionally, in
Chapter 3.0 guidance on how to set up a screening program and how
to screen individual pieces of equipment is provided.
In Chapter 4.0, information on how a process unit can
collect equipment leak rate data by enclosing individual
equipment and measuring mass emissions is provided. These data
can be used to develop unit-specific leak rate/screening value
correlations. Chapter 4.0 details the rigorous steps that need
to be followed when collecting the data to generate unit-specific
correlations. These steps are intended to ensure that the data
are of high quality.
In Chapter 5.0, information is provided that can be used to
estimate the control efficiency of equipment leak control
1-2
-------
techniques. The two primary control techniques for reducing
equipment leak emissions are (1) equipment modifications (such as
replacing a standard valve with a sealless type) and (2)
implementing a leak detection and repair (LDAR) program. Control
efficiencies for different equipment leak modifications are
summarized, and an approach for estimating the control efficiency
of any LDAR program is provided.
Appendices A through F provide additional information
supporting the material in the chapters. Appendix A contains
detailed example calculations using the approaches described in
Chapter 2.0. Appendix B documents how the SOCMI correlations and
emission factors were revised. Appendix B also serves as a
demonstration of how data can be analyzed to develop unit-
specific correlations. Appendix C summarizes available data on
response factors. Appendix D provides guidance on how to collect
representative screening data for connectors. Appendix E
contains a copy of Method 21. Finally, Appendix F demonstrates
how LDAR control efficiencies presented in Chapter 5.0 were
calculated.
1-3
-------
2.0. DEVELOPMENT OF EQUIPMENT LEAK EMISSION ESTIMATES
2.1 INTRODUCTION
The purpose of this chapter is to describe how to estimate
emissions from equipment leaks in a chemical processing unit.
Four approaches for estimating equipment leak emissions are
presented:
Approach 1: Average Emission Factor Approach;
Approach 2: Screening Ranges Approach;
Approach 3: EPA Correlation Approach; and
Approach 4: Unit-Specific Correlation Approach.
General information on these approaches is presented in
Section 2.2, and detailed information on applying each of the
approaches is presented in Section 2.3. Included in Section 2.3
are emission factors and leak rate/screening value correlations
for use in estimating emissions from equipment leaks in
refineries, natural gas/gasoline processing plants (gas plants),
and the synthetic organic chemical manufacturing industry ( /
(SOCMI). The SOCMI emission factors and correlations have been
recently revised and are introduced in this document. The focus
of this document is estimating emissions of volatile organic
compounds (VOC's), and for the purpose of this document, VOC's
include all organic compounds except those excluded by the EPA
due to negligible photochemical activity.
After the four approaches have been discussed, topics that
are not specifically related to any particular approach, but are
relevant to how equipment leak emissions are estimated, are
addressed in Section 2.4. These topics include:
• Estimating emissions of individual compounds within a
mixture;
• Using response factors when estimating emissions;
• Considerations regarding the monitoring instrument
used;
• Estimating emissions of equipment not screened when
other equipment have been screened;
2-1
-------
• Using screening data collected at different times;
• Estimating VOC emissions from equipment containing
organic compounds excluded from EPA's classification of
VOC's.
• Estimating emissions from equipment containing
inorganic compounds.
Appendices A through D contain supporting documentation for
the material presented in this chapter. Appendix A contains
detailed example calculations that demonstrate the four
approaches for estimating equipment leak emissions, as well as
the topics discussed in Section 2.4. Appendix B presents details
on how unit-specific correlations can be developed, and also
presents background information on the revision of the SOCMI
correlations and emission factors. Appendix C offers a detailed
listing of available response factors. Appendix D contains
information on the minimum number of connectors in a process unit
that must be screened in order to obtain a representative sample.
2.2 GENERAL INFORMATION ON THE APPROACHES FOR ESTIMATING
EQUIPMENT LEAK EMISSIONS
This section presents general information on the four
approaches for estimating equipment leak emissions. Each
approach is briefly described, and data requirements for each are
summarized. Additionally, background information is presented to
provide an historical overview of data collection and analysis on
emissions of VOC's from equipment leaks.
2.2.1 Equipment Leak Emission Estimation Approaches
The four approaches described here can be used by any
chemical-handling facility to develop an inventory of VOC
emissions from equipment leaks. The approaches, in order of
increasing refinement, are: Average Emission Factor Approach,
Screening Ranges Approach, EPA Correlation Approach, and Unit-
Specific Correlation Approach.
In general, the more refined approaches require more data
and provide more accurate emission estimates for a process unit.
In the Average Emission Factor Approach and the Screening Ranges
Approach, emission factors are combined with equipment counts to
2-2
-------
estimate emissions. To estimate emissions with the EPA
Correlation Approach, measured concentrations (screening values)
for all equipment are individually entered into general
correlations developed by the EPA. In the Unit-Specific
Correlation Approach, screening and leak rate data are measured
for a select set of individual equipment components and then used
to develop unit-specific correlations. Screening values for all
components are then entered into these unit-specific correlations
to estimate emissions.
Figure 2-1 is an overview of the data collection and
analysis required to apply each of the approaches. As can be
seen from this figure, all of the approaches require an accurate
count of equipment components by type of equipment (i.e., valves,
pumps, connectors, etc.). Additionally, for some of the
equipment types, the count must be further described by service
(i.e., heavy liquid, light liquid, and gas).
Except for the Average Emission Factor Approach, all of the
approaches require screening data. Screening data are collected
by using a portable monitoring instrument to sample air from
potential leak interfaces on individual pieces of equipment. A
screening value is a measure of the concentration of leaking
compounds in the ambient air that provides an indication of the
leak rate from an equipment piece, and is measured in units of
parts per million by volume (ppmv). The protocols for collecting
screening data are presented in Chapter 3.0.
In addition to equipment counts and screening data, the
Unit-Specific Correlation Approach requires bagging data.
Bagging data consist of screening values and their associated
measured leak rates. A leak rate is measured by enclosing an
equipment piece in a bag to determine the actual mass emission
rate of the leak. The screening values and measured leak rates
from several pieces of equipment are used to develop a unit-
specific correlation. The resulting leak rate/screening value
correlation predicts the mass emission rate as a function of the
screening value. Protocols for collecting bagging data are
described in detail in Chapter 4.0.
2-3
-------
Count Equipment Components
(by type and service)
Approach 1
Conduct Complete
Screening Survey
Approach 2
Approach 3
Bag Components for Each
Equipment Type and Service
Develop Unit-Specific
Correlations
Apply Average Emission Factors
and Composite Total Emissions
Apply>10,000/
<10,000 ppmv Emission Factors
and Composite Total Emissions
Apply EPA Correlations and
Composite Total Emissions
Apply New Correlations
and Composite Total Emissions
Inventory Section 2.3.1
Inventory Section 2.3.2
Inventory Section 2.3.3
Inventory Section 2.3.4
Figure 2-1. Overview of Data Collection and Analysis Approaches
for Developing Equipment Leak Emissions Inventory
2-4
-------
Each of the approaches are applicable to any
chemical-handling facility. However, the EPA has developed more
than one set of emission factors and correlations, and the type
of process unit being considered governs which set must be used
to estimate emissions. Historical data collection on emissions
from equipment leaks in refineries, SOCMI, and natural gas plants
has yielded emission factors and correlations for these source
categories. Emission factors and correlations specific to other
source categories have not been developed.
For process units in source categories for which emission
factors and/or correlations have not been developed, the factors
and/or correlations already developed can be utilized. However,
appropriate evidence should indicate that the existing emission
factors and correlations are applicable to the source category in
question. Criteria for determining the appropriateness of
applying existing emission factors and correlations to another
source category may include one or more of the following:
(1) process design, (2) process operation parameters
(i.e., pressure and temperature), (3) types of equipment used,
and (4) types of material handled. For example, in most cases,
SOCMI emission factors and correlations are applicable for
estimating equipment leak emissions from the polymer and resin
manufacturing industry. This is because, in general, these two
industries have comparable process design and comparable process
operation, they use the same types of equipment, and they tend to
use similar feedstock.
2.2.2 Overview of Equipment Leak Data Collection
As mentioned above, data on equipment leak emissions of
organic compounds have been collected from refineries, gas
plants, and SOCMI process units. Emission factors and
correlations have been developed for the following equipment
types: valves, pumps, compressors, pressure relief valves,
connectors, and open-ended lines. For sampling connections, an
average emission factor has been developed that is an estimate of
the typical amount of material purged when a sample is collected.
2-5
-------
A brief history of the development of these factors and
correlations is presented below.
2.2.2.1 Refinery Assessment Study.^'2 In the late 1970s,
EPA initiated the Petroleum Refinery Assessment Study, and
equipment leak data from 13 refineries were collected. In this
study, equipment was screened and the majority of sources that
had screening values over 200 ppmv were bagged. Bagged equipment
emission rates were reported as non-methane organic compound
emission rates. Average emission factors and correlations for
each equipment type were developed based on the screening and
bagging data collected in this study.
The Refinery Assessment Study included an investigation of
possible correlations between equipment leaks and process
variables. The only process variables found to correlate with
mass emission rates in a statistically significant manner were
(1) the phase of the process stream, and (2) the relative
volatility of liquid streams. This finding led to the separation
of data for valves, pumps, and pressure relief valves by type of
service. Three service categories were defined:
• Gas/vapor - material in a gaseous state at operating
conditions;
• Light liquid - material in a liquid state in which the
sum of the concentration of individual constituents
with a vapor pressure over 0.3 kilopascals (kPa) at
20°C is greater than or equal to 20 weight percent; and
• Heavy liquid - not in gas/vapor service or light liquid
service.
2.2.2.2 Gas Plant Studies.3 A total of six gas plants were
screened in two studies: Four were screened by the EPA and two
by the American Petroleum Institute. Average emission factors
were developed, and information on the percentage of equipment
with screening values equal to or greater than 10,000 ppmv was
presented. The average factors include emissions of ethane and
methane, which are hydrocarbons but are not classified as VOC's.
2.2.2.3 Original SOCMI Average Emission Factors and
Correlations. In 1980, two studies were coordinated by the EPA
2-6
-------
to collect data from SOCMI process units. These studies were the
24-Unit Study,4 and the Six-Unit Maintenance Study.5 In the
24-Unit Study, screening data were obtained from equipment
containing organic compounds at 24 individual chemical process
units representing a cross-section of the SOCMI. In the Six-Unit
Maintenance Study, bagging data were collected from six of the
process units within the 24-Unit Study to determine the effect of
maintenance on equipment leak emissions. Most of the bagging
data were collected from equipment with screening values above
1,000 ppmv. As part of the Six-Unit Maintenance Study,
correlations were developed for light liquid pumps, gas valves,
and light liquid valves.
The original SOCMI average emission factors were first
presented in the document "Fugitive Emission Sources of Organic
Compounds—Additional Information on Emissions, Emission
Reductions, and Costs."6 This document is referred to as the
Fugitive Emissions Additional Information Document (AID). In the
Fugitive Emissions AID, the data from the Refinery Assessment
Study were further analyzed to develop "leak/no leak" emission
factors. (A "leak" was defined as a screening value greater than
or equal to 10,000 ppmv.) With the exception of the factor for
gas valves, the original SOCMI average emission factors were
developed using (1) the leak/no-leak emission factors developed
from the Refinery Assessment Study data, and (2) the leak
frequencies from the SOCMI 24-Unit Study screening value data
set. This approach was based on statistical comparisons that
indicated that the most significant characteristic that
distinguished equipment in SOCMI facilities from that in
refineries was not the leak rate for a given screening value, but
rather the fraction of equipment that had screening values
greater than or equal to 10,000 ppmv.
Thus, the following equation was used to calculate the
original SOCMI average emission factors:
SOCMI Average Factor = x * RLF + (1 - x) * RNLF
2-7
-------
where:
x = Fraction of sources from the 24-Unit Study that
screened greater than or equal to 10,000 ppmv;
RLF = Refinery leaking emission factor; and
RNLF = Refinery non-leaking emission factor.
For gas valves, the previously collected data suggested that
for a given screening value the leak rate at a SOCMI facility was
not statistically equivalent to the leak rate at a refinery.
Therefore, data from the Six-Unit Maintenance Study were used to
develop the gas valve average emission factor.
2.2.2.4 Revised SOCMI Emission Factors and Correlations.
In 1987 and 1988, screening data were obtained from 19 ethylene
oxide and butadiene producers, and, in 1990, bagging data were
collected from 16 of these process units. Screening and bagging
data were collected from light liquid pumps, gas valves, light
liquid valves, and connectors. A specific goal of the program
was to bag equipment that had screening values less than 1,000
ppmv. The bagging data were combined with bagging data
previously collected in the Six-Unit Maintenance Study, and this
combined bagging data set was used to revise the SOCMI
correlations. Likewise, the new screening data were combined
with screening data previously collected in the 24-Unit Study,
and this combined screening data set was used with the revised
correlations to generate new SOCMI emission factors.
Appendix B-2 contains more detailed information on how the
revised SOCMI correlations and emission factors were developed.
2.3 APPROACHES FOR ESTIMATING EQUIPMENT LEAK EMISSIONS
In this section, each of the approaches for estimating
equipment leak emissions are discussed. The description of each
approach focuses on the basic method to be used to estimate total
VOC emissions. Each of the approaches are demonstrated in
example calculations contained in Appendix A.
2-8
-------
2.3.1 Average Emission Factor Approach
One accepted approach for estimating emissions allows use of
average emission factors developed by the EPA in combination with
unit-specific data that are relatively simple to obtain. These
data include: (1) the number of each type of component in a unit
(valve, connector, etc.), (2) the service each component is in
(gas, light liquid, or heavy liquid), (3) the VOC concentration
of the stream, and (4) the time period each component was in that
service. The average emission factors for SOCMI process units,
refineries, and natural gas plants are presented in Tables 2-1,
2-2, and 2-3, respectively. The SOCMI and gas plants average
emission factors predict total organic compound emission rates,
whereas the refinery average factors predict non-methane organic
compound emission rates. Note that limited data has been
collected on the leak rate of agitators, and, until additional
data are collected for emissions from agitator seals, the average
factor for light liquid pump seals can be used to estimate
emissions from agitators.
Although the average emission factors are in units of
kilogram per hour per individual source, it is important to note
that these factors are most valid for estimating emissions from a
population of equipment. The average factors are not intended to
be used for estimating emissions from an individual piece of
equipment over a short time period (i.e., 1 hour).
To estimate emissions using the Average Emission Factor
Approach, the concentration of VOC in weight percent within the
equipment is needed. The VOC concentration should not include
inorganic compounds. Also, some organic compounds (such as
methane and ethane) are not classified as VOC's, and these
compounds should not be included in the VOC concentration. (One
exception to this is when methane is encountered in refineries.
This exception is further discussed later in this section.) The
VOC concentration in the equipment is important because equipment
with higher VOC concentrations tend to have higher VOC leak
rates. When using the Average Emission Factor Approach,
equipment should be grouped into "streams" where all the
2-9
-------
TABLE 2-1. SOCMI AVERAGE EMISSION FACTORS3
Equipment type
Valves
Pump seals0
Compressor seals
Pressure relief valves
Connectors
Open-ended lines
Sampling connections
Emission factor*3
Service (kg/hr/ source)
Gas
Light liquid
Heavy liquid
Light liquid
Heavy liquid
Gas
Gas
All
All
All
0.00597
0.00403
0.00023
0.0199
0.00862
0.228
0.104
0.00183
0.0017
0.0150
aThe emission factors presented in this table for gas
valves, light liquid valves, light liquid pumps, and
connectors are revised SOCMI average emission factors.
bThese factors are for total organic compound emission
rates.
cThe light liquid pump seal factor can be used to estimate the
leak rate from agitator seals.
2-10
-------
TABLE 2-2. REFINERY AVERAGE EMISSION FACTORS3
Equipment type
Valves
Pump seals0
Compressor seals
Pressure relief valves
Connectors
Open-ended lines
Sampling connections
Service
Gas
Light liquid
Heavy liquid
Light liquid
Heavy liquid
Gas
Gas
All
All
All
Emission factor*3
(kg/hr/ source)
0.0268
0.0109
0.00023
0.114
0.021
0.636
0.16
0.00025
0.0023
0.0150
aSource: Reference 2.
^These factors are for non-methane organic compound
emission rates.
°The light liquid pump seal factor can be used to estimate the
leak rate from acntator seals.
2-11
-------
TABLE 2-3. GAS PLANT AVERAGE EMISSION FACTORS3
Equipment type
Valves
Pump sealsc
Compressor seals
Pressure relief valves
Connectors
Open-ended lines
Service
All
Liquid
All
All
All
All
Emission factor*5
(kg/hr/ source)
0.020
0.063
0.204
0.188
0.0011
0.022
aSource: Reference 3.
^These factors are for total organic compound emission rates.
cThe pump seal factor can be used to estimate the leak rate from
agitator seals.
2-12
-------
equipment within the stream have approximately the same VOC
weight percent.
To apply the average emission factors, use the following
equation to estimate VOC emissions from all of the equipment in a
stream of a given equipment type:
VOCS = AEF * WFVOC * N
where:
VOCS = VOC emission rate from all equipment in the
stream of a given equipment type (kg/hr);
AEF = Applicable average emission factor for the
equipment type (kg/hr/source);
WFvoc = FOR SOCMI AND GAS PLANTS: Average weight
fraction of VOC in the stream;
FOR REFINERIES: Average weight fraction of
VOC in the stream (assuming that methane is
not included as part of the VOC weight
fraction) plus the average weight fraction of
methane within the stream (up to a maximum of
10 percent by weight methane); and
N = Number of pieces of equipment of the
applicable equipment type in the stream.
Note that the term "WFVOC" is defined differently for refineries
than for SOCMI and gas plants. It is necessary to add the
methane weight fraction back into the "WFVOC" term when applied
to refineries because when the refinery factors were developed,
the methane was subtracted out. Including the methane in the
"WFVOC" term for refineries is a way to correct for this. Two
guidelines when correcting the "WFVOC" term when applied to
refineries are as follows:
• The correction should only be applied to equipment
containing a mixture of VOC's and methane; and
• The maximum correction for the methane weight fraction
should not exceed 0.10, even if the equipment contains
greater than 10 weight percent methane. (This reflects
that equipment in the Refinery Assessment Study1/2
typically contained 10 weight percent or less methane).
2-13
-------
Thus, at a SOCMI process unit, if there were 100 gas valves
in a stream containing, on average, 90 weight percent VOC and
10 weight percent water vapor, emissions would be calculated as
follows:
VOCS = AEF * WFVoc * N
= 0.00597 kg/hr/gas valve * 0.9 * 100 gas valves
= 0.54 kg/hr of VOC from gas valves in the stream
At a refinery, if there were 100 gas valves in a stream that, on
average, contained 80 weight percent VOC, 10 weight percent water
vapor, and 10 weight percent methane, emissions would be
calculated using the above equation as follows:
VOCS = AEF * WFVOC * N
= 0.0268 kg/hr/gas valve * (0.8+0.1) * 100 gas
valves
= 2.41 kg/hr of VOC from gas valves in the stream
If there are several streams at a process unit, the total
VOC emission rate for an equipment type is the sum of emissions
from each of the streams. The total emission rates for all of
the equipment types are summed to generate the process unit total
VOC emission rate from leaking equipment.
As mentioned earlier, the average emission factors are not
intended to provide an accurate estimate of the emission rate
from a single piece of equipment. Rather, the average factors
are more appropriately applied to the estimation of emissions
from populations of equipment. Data indicate that the range of
possible leak rates from individual pieces of equipment spans
several orders of magnitude. As a result, the majority of total
emissions from a population of equipment at any given time will
normally occur from a small percentage of the total equipment.
The average emission factors account for the span of possible
2-14
-------
leak rates, but, as a result, they are not necessarily an
accurate indication of the mass emission rate from an individual
piece of equipment.
Furthermore, the average emission factors do not reflect
different site-specific conditions among process units within a
source category. Site-specific factors can have considerable
influence on leak rates from equipment. Nevertheless, in the
absence of screening data, the average emission factors do
provide an indication of equipment leak emission rates from
equipment in a process unit.
2.3.2 Screening Ranges Approach
The Screening Ranges Approach (formerly known as the
leak/no-leak approach) offers some refinement over the Average
Emission Factor Approach, thereby allowing some adjustment for
individual unit conditions and operation. This approach and the
other two remaining approaches require that screening data be
collected for the equipment in the process unit. The screening
data are an indication of leak rates. When applying this
approach, it is assumed that components having screening values
greater than 10,000 ppmv have a different average emission rate
than components with screening values less than 10,000 ppmv.
This approach may be applied when screening data are
available as either "greater than or equal to 10,000 ppmv" or as
"less than 10,000 ppmv." Emission factors for SOCMI, refinery,
and gas plants for these two ranges of screening values are
presented in Tables 2-4, 2-5, and 2-6, respectively. As with
the average factors, the SOCMI and gas plant screening range
factors predict total organic compound emissions, whereas the
refinery screening range factors predict non-methane organic
compound emissions. Note that there are not screening range
factors for sampling connections because emissions from sampling
connections occur when the line is purged, and, thus, are
independent of any screening value. Also, as with the average
factors, the screening range factors for light liquid pumps can
be applied to agitators.
2-15
-------
TABLE 2-4. SOCMI SCREENING VALUE RANGE EMISSION FACTORS3
Equipment type
Service
>10,000 ppmv
Emission
factor*3
(kg/hr/source)
<10,000 ppmv
Emission
factorb
(kg/hr/source)
Valves
Pump seals0
Gas
Light liquid
Heavy liquid
Light liquid
Heavy liquid
0.0782
0.0892
0.00023
0.243
0.216
0.000131
0.000165
0.00023
0.00187
0.00210
Compressor
seals
Pressure relief
valves
Connectors
Open-ended
lines
Gas
Gas
All
All
1.608
1.691
0.113
0.01195
0.0894
0.0447
0.0000810
0.00150
aThe emission factors presented in this table for gas valves,
light liquid valves, light liquid pumps, and connectors are
revised SOCMI > 10,000/< 10,000 ppmv emission factors.
^These factors are for total organic compound emission rates.
GThe light liquid pump seal factors can be applied to estimate
the leak rate from agitator seals.
~0
2-16
-------
TABLE 2-5. REFINERY SCREENING VALUE RANGE
EMISSION FACTORS*
Equipment type
Service
>10,000 ppmv
Emission
factor*3
(kg/hr/source)
<10,000 ppmv
Emission
factorb
(kg/hr/source)
Valves
Pump seals0
Compressor seals
Pressure relief
valves
Connectors
Open-ended lines
aSource: Reference
Gas
Light liquid
Heavy liquid
Light liquid
Heavy liquid
Gas
Gas
All
All
6.
0.2626
0.0852
0.00023
0.437
0.3885
1.608
1.691
0.0375
0.01195
0.0006
0.0017
0.00023
0.0120
0.0135
0.0894
0.0447
0.00006
0.00150
bThese factors are for non-methane organic compound emission
rates.
GThe light liquid pump seal factors can be applied to estimate
the leak rate from agitator seals.
2-17
-------
TABLE 2-6. GAS PLANT SCREENING VALUE RANGE EMISSION FACTORS
Equipment type
Service
>10,000 ppmv
Emission
factor3
(kg/hr/source)
<10,000 ppmv
Emission
factora
(kg/hr/source)
Valvesb
Pump sealsb/c
Compressor seals*3
Pressure relief
valves^
Connectors^
Open-ended lines^
All
All
All
Gas
All
All
0.098
0.150
0.442
0.863
0.0336
0.174
0.0029
0.020
0.025
0.0447
0.00006
0.0015
aThese factors are for total organic compound emission rates.
bThese factors were calculated based on information in
Reference 3 on the fraction of sources screening >10,000 ppmv
and the percent of total emissions from sources screening
>10,000 ppmv.
°The light liquid pump seal factors can be applied to estimate
the leak rate from agitators.
AThese factors were calculated based on information in
Reference 3 on the fraction of sources screening >10,000 ppmv
and assuming that sources screening <10,000 ppmv had on average
emissions equal to the refinery <10,000 ppmv factors.
2-18
-------
The Screening Ranges Approach is applied in a similar manner
as the average emission factor approach in that equipment counts
are multiplied by the applicable emission factor. However, in
the Screening Range Approach, no adjustment is made for inorganic
compounds in the equipment because the screening value on which
emissions are based is a measurement of only organic compound
leakage.
An equation for applying the Screening Ranges Approach is
presented below. This equation is intended for cases when all
organic compounds in the equipment are classified as VOC's. In
cases where equipment pieces contain organic compounds (such as
methane and ethane), which are not classified as VOC's, the
equation can be corrected to subtract out the non-VOC portion.
This is discussed in more detail in Section 2.4.6.
To calculate VOC emissions using the Screening Ranges
Approach, the following equation is used:
VOC = GEF * Nge + LEF * N].e
; vere:
VOC = Total VOC emission rate for an equipment type
(kg/hr);
GEF = Applicable emission factor for sources with
screening values greater than or equal to
10,000 ppmv (kg/hr/source);
Nge = Equipment count (specific equipment type) for
sources with screening values greater than or
equal to 10,000 ppmv;
LEF = Applicable emission factor for sources with
screening values less than 10,000 ppmv
(kg/hr/source); and
NIS = Equipment count (specific equipment type) for
sources with screening values less than
10,000 ppmv.
The screening range emission factors are a better indication
of the actual leak rate from individual equipment than the
average emission factors. The greater than or equal to
2-19
-------
10,000 ppmv emission factor is particularly useful, because the
maximum readout is 10,000 ppmv on many of the screening
instruments; thus, the actual screening value can only be
determined by adding a dilution probe to the instrument. To
avoid having to use a dilution probe, the greater than or equal
to 10,000 ppmv factor can be applied. Nevertheless, available
data indicate that measured mass emission rates can vary
considerably from the rates predicted by use of these factors.
2.3.3 EPA Correlation Approach
This approach offers an additional refinement to estimating
emissions from equipment leaks by providing an equation to
predict mass emission rate as a function of screening value for a
particular equipment type. Correlations developed by EPA
relating screening values to mass emission rates for SOCMI
process units and refineries are presented in Tables 2-7 and 2-8,
respectively. Figures 2-2 through 2-5 plot the correlations.
The SOCMI correlations predict total organic compound emission
rates, whereas the refinery correlations predict non-methane
organic compound emission rates. Appendix B contains additional
information on the development of the correlation equations.
The EPA Correlation Approach is preferred when actual
screening values are available. This approach involves entering
the screening value into the correlation equation, which predicts
the mass emission rate based on the screening value.
Correlations for SOCMI are available for (1) gas valves;
(2) light liquid valves; (3) connectors; and (4) light liquid
pump seals. Correlations for refineries are available for
(1) gas valves; (2) light liquid valves; (3) connectors;
(4) a single equation for light liquid pump seals, compressor
seals, and pressure relief valves; and (5) heavy liquid pump
seals.
There is a single refinery correlation for liquid pump
seals, compressor seals, and pressure relief valves because
statistical tests performed on the bagging data collected from
these equipment types during the Refinery Assessment Study2
indicated that one correlation could represent these component
2-20
-------
TABLE 2-7. SOCMI LEAK RATE/SCREENING VALUE CORRELATIONS3
Equipment type _ Correlation*3/ c _
Gas valves Leak rate (kg/hr) = 1.87 * 10~6 (SV)O-873
Light liquid valves Leak rate (kg/hr) = 6.41 * 10~6 (SV) 0.797
Light liquid pumpsd Leak rate (kg/hr) = 1.90 * 10~5 (SV)0-824
Connectors Leak rate (kg/hr) = 3.05 * 10"6 (SV) 0.885
aThe correlations presented in this table are revised SOCMI
correlations.
&SV = Screening value.
cThese correlations predict total organic compound emission
rates .
correlation for light liquid pumps can be applied to
compressor seals, pressure relief valves, agitator seals, and
heavy liquid pumps.
2-21
-------
TABLE 2-8. REFINERY LEAK RATE/SCREENING VALUE CORRELATIONS9
Equipment type Correlation**/c
Gas valves Leak rate (kg/hr) - 2.18 * 10~7 (SV)!-23
Light liquid valves Leak rate (kg/hr) = 1.44 * 10~5 (SV)0.80
Light liquid pumps, Leak rate (kg/hr) = 8.27 * 10"5 (SV)°-83
compressors, pressure
relief valves3
Connectors Leak rate (kg/hr) = 5.78 * 10~6 (SV)0-88
Heavy liquid pumps Leak rate (kg/hr) = 8.79 * 10~6 (SV)1.04
aSource: Reference 2.
= Screening value.
cThese correlations predict non-methane organic compound
emission rates.
correlation can be applied to agitators.
2-22
-------
SOCM Correlation Curves
Screening Values 0-1,000 ppmv
0.006
r\
L
£
\
0)
.V
\J
If)
flj
+J
-------
SOCMI Correlation Curves
Screening Values 1,000-1,000,000 ppmv
n
L
£
\J
in
(0
DC
.
id
V
J
u
0
h
400
600
800
1000
D Gas Valves
Screening Value (ppmv) in thousands
+ Light Liquid Valves o Light Liquid Pumps
A Connectors
Figure 2-3.
SOCMI Correlations relating total organic compound
(TOC) leak rate to screening value:
1,000 - 1,000,000 ppmv
2-24
-------
Refinery Corre at I on Curves
Screening Values 0-1,000 ppmv
n
L
r
v
0
X
u
(0
CL
J
U
0
200
400
600
800
1000
Screening Value
D Gas Valves + Light Liquid Valves 0 Light Liquid Pumps
X Heavy Liquid Pumps
A Connectors
Figure 2-4.
Refinery correlations relating non-methane organic
compound (NMOC) leak rate to screening value:
0 - 1,000 ppmv
2-25
-------
Refinery Corre at ion Curves
Screening Values 1,000-1,000,000 ppmv
n
L
£
u
10
a>
cr
(0
a>
j
o
0
z
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
400
600
800
1000
D Gas Valves
Screening Value (ppmv) in thousands
Light Liquid Valves 0 Light Liquid Pumps
X Heavy Liquid Pumps
A Connectors
Figure 2-5.
Refinery correlations relating non-methane organic
compound (NMOC) leak rate to screening value:
1,000 - 1,000,000 ppmv
2-26
-------
types. Limited bagging data for compressors and pressure relief
devices have been obtained at SOCMI plants. However, because
statistical tests performed as part of the Refinery Assessment
Study2 indicated that emissions from light liquid pumps,
compressors, and pressure relief valves could be expressed with a
single correlation, until additional data are collected, the
SOCMI equation for light liquid pump seals can be applied to
estimate emissions for compressor seals and pressure relief
valves in SOCMI process units.
Bagging data for agitator seals at refineries and SOCMI
process units are unavailable at this time. Compared to those
equipment types that have correlations, agitators most closely
resemble light liquid pumps, and, for this reason, the applicable
light liquid pump correlation can be used to estimate agitator
emissions. Similarly, the SOCMI light liquid pump correlation
can be used to estimate emissions from SOCMI heavy liquid pumps.
Correlations can be used to estimate emissions for the
entire range of non-zero screening values, from the highest
potential screening value to the screening value that represents
the minimum detection limit of the monitoring device. All non-
zero screening values can be entered directly into the
correlation to predict emissions associated with the screening
value.
The "default-zero" leak rate is the mass emission rate
associated with a screening value of zero. (Note that any
screening value that is less than or equal to ambient
[background] concentration is considered a screening value of
zero.) The correlations mathematically predict zero emissions
for zero screening values. However, data collected by the EPA
show this prediction to be incorrect. Mass emissions have been
measured from equipment having a screening value of zero.
A specific goal when revising the SOCMI correlations was to
collect mass emissions data from equipment that had a screening
value of zero. These data were used to determine a default-zero
leak rate associated with equipment with zero screening values.
2-27
-------
Table 2-9 lists the default-zero leak rate for each of the
equipment types. These default-zero leak rates are applicable
only when the minimum detection limit of the portable monitoring
instrument is 1 ppmv or less above background. The default-zero
leak rates in Table 2-9 were collected from SOCMI facilities.
However, these leak rates are based on the best available data
and are considered applicable for all source categories.
The portable monitoring device used to collect the default-
zero data was sufficiently sensitive to indicate a screening
value of 1 ppmv or less. In cases where a monitoring instrument
has a minimum detection limit greater than 1 ppmv, the default-
zero leak rates presented in Table 2-9 are not applicable. For
these cases, an alternative approach for determining a default-
zero leak rate is to (1) determine one-half the minimum screening
value of the monitoring instrument, and (2) enter this screening
value into the applicable correlation to determine the associated
default-zero leak rate.
Assuming all of the organic compounds in the equipment are
classified as VOC's, total VOC emissions for each equipment type
are calculated as the sum of emissions associated with each of
the screening values. Section 2.4.6 discusses a correction that
can be made to the predicted VOC emissions rate if some of the
organic compounds in the equipment are not classified as VOC's
(such as methane and ethane). Each equipment piece with a
screening value of zero is assigned the default-zero leak rate.
For all equipment with a non-zero screening value, the screening
value associated with each individual equipment piece is entered
into the applicable correlation to predict emissions. Jt should
be noted that each individual screening value must be entered
into the correlation to predict emissions for an equipment piece.
Do not average screening values and then enter the average value
into the correlation to estimate emissions.
2-28
-------
TABLE 2-9. DEFAULT-ZERO VALUES3
Equipment type
Default-zero emission rates
(kg/hr/source)
Gas valve
Light liquid valve
Light liquid pumpb
Connectors
6.56 * 10~7
4.85 * ID"7
7.49 * 10~6
6.12 * 10~7
aThe default-zero mass emission rates presented in this table
are recently revised default-zero values. These values
predict total organic compound emission rates and are
applicable to all source categories.
light liquid pump default zero value can be applied to
compressors, pressure relief valves, agitators, and heavy
liquid pumps.
2-29
-------
2.3.4 Unit-Specific Correlation Approach
To develop unit-specific correlations screening value and
corresponding mass emissions data (i.e., bagging data) must be
collected from process unit equipment. (See Chapter 4.0 for a
detailed discussion on the protocols for bagging equipment.) The
equipment selected for bagging should be screened at the time of
bagging. The mass emissions rate determined by bagging, and the
associated screening value, can then be used to develop a leak
rate/screening value relationship (i.e., correlation) for that
specific equipment type in that process unit. The correlations
must be developed on a process unit basis to minimize the error
associated with differing leak rate characteristics between
units.
If a unit-specific correlation is developed, as long as the
protocols for bagging discussed in Chapter 4.0 are followed, it
is not necessary to demonstrate that the correlation is
statistically different from the EPA correlation for it to be
applied. However, before developing unit-specific correlations,
it may be desirable to evaluate the validity of the EPA
correlations to a particular process unit. As few as four leak
rate measurements of a particular equipment type in a particular
service can be adequate for this purpose. The measured emission
rates can be compared with the rates that would be predicted by
the EPA correlations to evaluate whether or not the EPA
correlations provide reasonable mass emission estimates. A
simple method of comparison is to determine if measured emission
rates are consistently less than or greater than what would be
predicted by the EPA correlation. If there is a consistent
trend, such as all of the measured leak rates being lower than
the rate predicted by the EPA correlation, the EPA correlation
may not provide reasonable emission estimates for the process
unit.
A more formal comparison is the Wilcoxon signed-rank test.
This test can be performed by comparing the logarithm of the
measured mass emission rates to the logarithm of the
corresponding rates predicted by the EPA correlation. The
2-30
-------
absolute magnitude of the differences are then ranked (e.g., the
pair with the smallest difference is assigned a rank of 1, the
pair with the next smallest difference a rank of 2, etc.), and
the sum of the ranks associated with positive differences is
computed. For example, if four bags were measured and they each
predicted higher mass emission rates than the EPA correlation,
the value of the sum of the ranks associated with those pairs
with positive differences would equal:
1+2+3+4= 10
On the other hand, if four bags were measured and three predicted
higher mass emission rates than the EPA correlation, but the one
with the greatest absolute difference predicted a lower rate than
the EPA correlation, then the sum of the positive ranks would
equal:
1+2+3=6. (Note: The sum of the negative ranks would
equal 4).
The value of the sum of the positive ranks can be compared to
given values on statistical tables to evaluate if there are
statistically significant differences between the measured rates
and the rates predicted by the EPA correlation.
However the comparison is performed, in cases where the EPA
correlations provide an adequate estimate of emissions, then the
potential increase in accuracy obtained by developing unit-
specific correlations may not be worth the effort. Consideration
should also be given to the typical screening value measured at a
process unit. If a process unit normally has very low screening
values, then the difference between the sum of unit equipment
leak emissions predicted by a unit-specific correlation and the
EPA correlation will likely be relatively small.
In developing new correlations, a minimum number of leak
rate measurements and screening value pairs must be obtained
according to the following methodology. First, equipment at the
2-31
-------
process unit is screened so that the distribution of screening
values at the unit is known. Then, mass emissions data must be
collected from individual sources that have screening values
distributed over the entire range. The criteria for choosing
these sources is as follows. For each equipment type
(i.e., valves, pumps, etc.) and service (i.e., gas, light liquid,
etc.), a random sample of a minimum of six components should be
chosen for bagging from each of the following screening value
ranges:
Screening Value Range (ppmv)
1 - 100
101 - 1,000
1,001 - 10,000
10,001 - 100,000
> 100,000
The requirement of six bags per screening value range is based on
EPA experience with bagging components. There are two primary
reasons for the above requirement: (1) to be confident in the
representativeness of the data, and (2) to accurately reflect the
range of possible mass emission rates associated with a given
screening value. The importance of the first reason is
self-evident: The more data collected the better the
representativeness. The importance of the second reason is that
a given screening value does not necessarily have a "true"
emissions rate. For a single screening value, the mass emissions
may range over several orders of magnitude depending upon several
factors, including the equipment type (i.e., gate valve versus
ball valve versus plug valve, etc.) and operating parameters
(i.e., chemical handled, temperature, pressure, etc.). This
range of possible mass emission rates is accounted for when the
correlation is developed (see discussion on the scale bias
correction factor), and it is important to obtain enough data to
accurately reflect the range. If six sources are not available
in a particular screening value range, additional sources from
2-32
-------
the nearest range should be tested so that a minimum of
30 emission rate/screening value pairs are obtained for each
source type. If 30 or more bags are collected, the process unit-
specific correlation can be used to estimate emissions across the
entire range of screening values (1 to 1,000,000 ppmv).
In some cases, it may be desirable to develop a correlation
with fewer than 30 bags. This can be accomplished by developing
a correlation that is not valid across the entire range of
screening values. Two alternatives are available: (1) to develop
a correlation valid for screening values ranging from 1 to
100,000 ppmv, or (2) to develop a correlation valid for screening
values ranging from 1 to 10,000 ppmv. These alternatives may be
preferable for process units with equipment that do not normally
have high screening values. An example of this type of process
unit is one that already has a leak detection and repair program
in place to prevent the release of odor-causing chemicals. At
this type of process unit, leaks may be quickly detected and
repaired.
For the first alternative, a minimum of 24 bags are
required, rather than 30, because sources with screening values
greater than 100,000 ppmv do not need to be bagged. Thus, a
minimum of six sources each should be chosen for bagging from
each of the screening ranges presented above except for the
greater than 100,000 ppmv range. In the event that a source
screens at 100,000 ppmv or greater, emissions can be estimated
using the greater than or equal to 100,000 ppmv emission factors
presented in Table 2-10.
For the second alternative, a minimum of 18 bags are
required, because sources screening greater than 10,000 ppmv do
not need to be bagged. Thus, a minimum of six sources should be
chosen for bagging from the 1 to 100 ppmv range, the 100 to
1000 ppmv range, and the 1,000 to 10,000 ppmv range. In the
event that a source screens at 10,000 ppmv or greater, emissions
can be estimated using the applicable greater than or equal to
10,000 ppmv emission factor presented in either Table 2-4, 2-5,
or 2-6. An advantage of using the greater than or equal to
2-33
-------
TABLE 2-10. GREATER THAN 100,000 PPMV SCREENING VALUE EMISSION
FACTORS FOR SOCMI AND REFINERY PROCESS UNITS
Equipment type
SOCMI >100,000 ppmv
factor
(kg/hr/source)a
Refinery >100,000 ppmv
Emission factor
(kg/hr/source)
Gas valves
Light liquid
valves
Light liquid pump
seals0
Connectors
0.114
0.150
0.623
0.216
1.20
0.349
2.93
0.384
aThe SOCMI factors predict total organic compound emission rates,
refinery factors predict non-methane organic compound
emission rates.
GThe light liquid pump seal >100,000 ppmv emission factor can be
applied to compressors, pressure relief valves, and agitators.
2-34
-------
10,000 ppmv emission factors is that several of the available
portable monitoring instruments have a maximum readout of
10,000 ppmv, and to obtain a screening value from a source
screening at 10,000 ppmv, it is necessary to install a dilution
probe. However, if the greater than or equal to 10,000 ppmv
factor is used, installing a dilution probe is not necessary.
The above groupings and recommended number of sources are
given as guidelines. They are based on experience in measuring
leak rates and developing leak rate/screening value correlations.
Other source selection strategies can be used if an appropriate
rationale is given.
With mass emissions data and screening values, leak
rate/screening value correlations can be generated using the
following methodology. Least-squares regression analyses are
completed for each equipment type/service, regressing the log of
the leak rate on the log of the screening concentration,
according to:
Log10 (leak rate [in kg/hr]) = /3g + 0i * Logic (SV)
where:
00* 01 ~ Regression constants; and
SV = Screening value.
Note that the results are the same whether the base 10 or natural
logarithm are used (see Appendix B). The equations presented
here are written assuming the Base 10 logarithm is used. All
analyses should be conducted using logarithms of both the leak
rate and screening value because this type of data has been shown
to be log-normally distributed. A scale bias correction factor
(SBCF) is required in transforming the equation in the log-scale
2-35
-------
back to the original units. The transformed equation is the
unit-specific correlation, and is expressed as:
*° *
Leak rate = SBCF * 10
where :
Leak rate = Emission rate from the individual equipment
piece (kg/hr) ;
SBCF = Scale bias correction factor;
00 »01 = Regression constants; and
SV = Screening value.
The SBCF is a function of the mean square error of the
correlation in log space. The greater the range of possible
emission rates for a given screening value, the greater the SBCF
will be. The purpose of the SBCF is to reflect this range when
transforming the correlation out of log space. When regressed in
log space, in general, approximately half of the data points will
lie above the correlation line and half will lie below it, and,
for a given screening value, the correlation will pass through
the mean log leak rate (i.e., the geometric mean). Thus, one way
of thinking of the correlation in log space is that it predicts
the geometric mean emissions rate across the range of screening
values. However, the geometric mean always underestimates the
arithmetic mean.
A simplified hypothetical example will help demonstrate this
point: For a screening value of 500,000 ppmv, three bagging data
points were obtained with mass emission rates of 0.1 kg/hr,
1 kg/hr, and 10 kg/hr. In log space, these emission rates
correspond to log^o (0-1) = -I/ l°9io (1) = O/ and logio (10) =
1, respectively. Thus, the geometric mean of these three points
is (-1 + 0 + l)/3 = 0. Directly transforming this geometric mean
to normal space predicts an emission rate for a screening value
of 500,000 ppmv of 10° = 1 kg/hr, whereas the arithmetic mean of
the emission rates is (0.1 + 1 + 10)/3 =3.7 kg/hr. From this
2-36
-------
example, it can be seen that the geometric mean underestimates
the arithmetic mean.
Thus, if the correlation was directly transformed, it would
underestimate the true average emission rate associated with a
given screening value, and, for this reason, the SBCF is
necessary to transform the correlation out of log space.
In Appendix B, additional details on developing a
process-unit specific correlation are presented. Appendix B also
contains information on development of the revised SOCMI
correlations.
2.4 SPECIAL TOPICS
There are several special topics relevant to estimating
equipment leak emissions that are not specific to any one of the
four approaches that have been described. These special topics
are discussed in this section:
• Speciating emissions;
• Using response factors;
• Monitoring instrument type and calibration gas;
• Estimating emissions for equipment not screened (when
other screening data are available);
• Using screening data collected at several different
times;
• Estimating VOC emission rates from equipment containing
organic compounds not classified as VOC's (such as
methane and ethane); and
• Estimating equipment leak emissions of inorganic
compounds.
Each of these topics above are addressed in the following
sections.
2.4.1 Speciating Emissions
For each of the approaches, the equations presented are used
to estimate total VOC emissions. Often, in a chemical-handling
facility, material in equipment is a mixture of several
chemicals, and, in some cases, it may be necessary to estimate
emissions of a specific VOC in the mixture. The following
2-37
-------
equation is used to speciate emissions from a single equipment
piece:
CHEMX = VOC * WPx/WPvoc
where:
CHEMX = The mass emissions of VOC "x" from the
equipment (kg/hr);
VOC = The VOC mass emissions from the equipment
(kg/hr);
WPX = The concentration of VOC "x" in the equipment
in weight percent; and
WPVOC = The total VOC concentration in the equipment
in weight percent.
An assumption in the above equation is that the weight percent of
the chemicals in the mixture contained in the equipment will
equal the weight percent of the chemicals in the leaking
material. In general, this assumption should be accurate for
single-phase streams containing (1) any gas/vapor material, or
(2) liquid mixtures containing constituents of similar
volatilities.
If the material in the equipment piece is a liquid mixture
of constituents with varying volatilities, in certain cases this
assumption may not be correct. Whether or not the assumption is
valid for a liquid mixture of varying volatilities depends on the
physical mechanism of how the leakage occurs from the equipment.
If the physical mechanism is one in which the liquid "flashes"
before it leaks from the equipment, the leaking vapor may contain
a higher concentration of the more volatile constituents than is
contained in the liquid mixture, on the other hand, if the
mechanism is one in which the liquid material leaks from the
equipment and then evaporates, the assumption that the weight
percent of each constituent in the liquid will equal the weight
percent of each constituent in the vapor is valid. There are no
clear guidelines to determine what mechanism is taking place for
any given piece of equipment; for this reason, unless there is
2-38
-------
information to suggest otherwise, it should be assumed that the
leaking vapor has the same concentrations as the liquid.
For those cases where it is suspected the leaking vapor will
have different concentrations than the liquid, engineering
judgement should be used to estimate emissions of individual
chemical species. An example might be equipment containing
material in two phases. Another hypothetical example is a case
where equipment contain a liquid mixture of two constituents with
one of the constituents having a very low vapor pressure and the
other a much higher vapor pressure. Leaks may occur from the
equipment such that the constituent with higher vapor pressure
volatilizes to the atmosphere, but the constituent with lower
vapor pressure is washed to the waste water treatment system
prior to volatilization.
2.4.2 Using Response Factors
A correction factor that can be applied to a screening value
is a response factor (RF) that relates the actual concentration
to the measured concentration of a given compound, using a
specific reference gas. As stated earlier, screening values are
obtained by using a portable monitoring instrument to detect
VOC's at an equipment piece leak interface. An "ideal" screening
RF value is one that is equal to the actual concentration of
VOC's at the leak interface. However, portable monitoring
instruments used to detect VOC concentration do not respond to
different VOC's equally. (This is discussed in more detail in
Chapter 3.0). To demonstrate this point, consider a monitoring
instrument calibrated using a reference gas. If the instrument
is calibrated correctly and is used to measure the concentration
of the gas with which it has been calibrated, it will indicate
the actual concentration. However, when used to measure other
gases for which the monitoring instrument is more or less
sensitive than the calibration gas, it will not indicate the
actual concentration. To correct for this, RF's have been
developed. The RF is calculated using the equation:
RF = AC/SV
2-39
-------
where:
RF - Response factor;
AC = Actual concentration of the organic compound
(ppmv); and
SV = Screening value (ppmv).
The value of the RF is a function of several parameters.
These parameters include the monitoring instrument, the
calibration gas used to calibrate the instrument, the compound(s)
being screened, and the screening value.
The correlations presented in this chapter have been
developed primarily from screening value/mass emission data pairs
collected from equipment containing compounds that had RF's less
than three. Thus, for cases in which a calibrated instrument is
used to measure concentrations of a compound for which that
instrument has an RF of three or less, reasonably accurate
emission estimates can be obtained directly without adjusting the
screening value. However, for a case in which a compound has an
RF greater than three for the calibrated instrument, the
emissions estimated using the unadjusted screening value will
generally underestimate the actual emissions. The EPA recommends
that if a compound (or mixture) has an RF greater than three,
then the RF should be used to adjust the screening value before
it is used in estimating emissions.
A detailed listing of published RF's is contained in
Appendix C. These RF's were developed by injecting a known
concentration of a pure compound into a monitoring instrument and
comparing that actual concentration to the instrument readout
(i.e., screening value).
As an example of applying a RF, consider chloroform. From
Table C-2 in Appendix C, it can be seen that the RF for
chloroform at an actual concentration of 10,000 ppmv is equal to
4.48 for a Foxboro OVA-108 monitoring instrument calibrated with
methane. Thus, when the actual concentration of chloroform is
10,000 ppmv, the instrument will read 10,000 ppmv divided by
2-40
-------
4.48, which equals 2,230 ppmv. if the measured value for
chloroform was directly entered into the correlation, it would
tend to underestimate emissions. (Note that when the RF is less
than 1 the unadjusted screening value will tend to overestimate
actual emissions.)
The RF's in Appendix C are for pure compounds. Those RF's
can be used to estimate the RF for a mixture using the equation:
where:
RFm = Response factor of the mixture;
n = Number of components in the mixture;
xi = Mole fraction of constituent i in the mixture;
and
RFi = Response factor of constituent i in the mixture.
This equation is derived in Appendix A.
An alternative approach for determining the RF of a pure
compound or a mixture is to perform analysis in a laboratory to
generate the data used to calculate a RF. The approach for
generating these data in the laboratory is described in
Chapter 3.0. The approach involves injecting samples of a known
concentration of the material of interest into the actual
portable monitoring instrument used to obtain the screening
values and calculating the RF based on the instrument readout.
In general, calculating the RF by performing analysis on site
will give the most accurate RF information, since, among other
factors, RF's have been shown to be a function of the individual
monitoring instrument.
Ideally, when using screening values to estimate equipment
leak emissions, the RF would be equal to 1, and, in this way, the
screening value would be the actual concentration. However,
because RF's are a function of several parameters, this cannot
2-41
-------
normally be achieved. Response factors can be used to correct
all screening values, if so desired. To evaluate whether a RF
correction to a screening value should be made, the following
three steps can be carried out.
(1) For the combination of monitoring instrument and
calibration gas used, determine the RF's of a given
material at an actual concentration of 500 ppmv and
10,000 ppmv. (See Appendix C; in some cases, it may
not be possible to achieve an actual concentration of
10,000 ppmv for a given material. In these cases, the
RF at the highest concentration that can be safely
achieved should be determined.)
(2) If the RF's at both actual concentrations are below 3,
it is not necessary to adjust the screening values.
(3) If either of the RF's are greater than 3, then EPA
recommends an RF be applied for those screening values
for which the RF exceeds 3.
One of the following two approaches can be applied to
correct screening values:
(1) Use the higher of either the 500 ppmv RF or the
10,000 ppmv RF to adjust all screening values.
(2) Generate a response factor curve to adjust the
screening values.
A RF curve can be generated in one of two ways. The
simplest way is to assume that the RF value is a linear function
of the screening value. The first step to generate a line
relating screening value to RF is to convert the RF at the actual
concentration to the RF at the associated screening value. This
is done by dividing the RF by the actual concentration to get the
associated screening value. Thus, if, at an actual concentration
of 10,000 ppmv, an instrument has a RF of 5, this corresponds to
a screening value of 2,000 ppmv (i.e., 10,000 ppmv divided by 5).
This procedure is implemented at both actual concentrations of
10,000 ppmv and 500 ppmv, and a line is drawn between the RF's at
each associated screening value. This line can then be used to
estimate the RF at any given screening value. (See Appendix A
for a demonstration of this procedure.) The line should not be
2-42
-------
extrapolated for screening values beyond the endpoints. For
these screening values, the endpoint RF should be applied.
For some materials, the RF is nonlinear as the screening
value increases. For these materials, RF's at several screening
values can be estimated by collecting data in a laboratory, as
mentioned earlier. The RF/screening value relationship can then
be generated by fitting a curve through the data pairs.
When an RF is used, the screening value is multiplied by the
RF before mass emissions are estimated. Thus, if a screening
value is 3,000 ppmv and the associated RF is 4, then the
screening value must be adjusted to 12,000 ppmv (i.e., 3,000
multiplied by 4) before mass emissions are predicted.
It should be noted that if it is possible to calibrate the
monitoring instrument with the material contained in the
equipment that is being screened, the RF should equal 1. Thus,
theoretically, the screening values will equal the actual
concentration, and no RF adjustment will be necessary. If it is
necessary to apply RF's, site personnel should use engineering
judgement to group process equipment into streams containing
similar compounds. All components associated with a given stream
can then be assigned the same RF, as opposed to calculating an RF
for each individual equipment piece.
2.4.3 Monitoring Instrument Type and Calibration Gas
When the correlations presented in Section 2.3 were
developed, in general, for each of the source categories, the
data were collected using a specific type of monitoring
instrument calibrated with a specific calibration gas. The
correlations are intended to relate actual concentration to mass
emissions. For this reason, screening values obtained from any
combination of monitoring instrument and calibration gas can be
entered directly into the correlations as long as the screening
values are an indication of actual concentration. If the
screening values are not an indication of the actual
concentration, the guidelines set forth in the previous section
on RF's can be applied to correct the screening values
(i.e., screening values should be adjusted if the RF is greater
2-43
-------
than 3). Otherwise, it is not necessary to correct screening
values to account for the instrument type and calibration gas
that were used to develop the correlation curves developed by
the EPA.
2.4.4 Estimating Emissions for Equipment Not Screened
Often, screening data cannot be collected for all of the
equipment pieces in a process unit. In some cases, equipment are
difficult or unsafe to screen. Difficult or unsafe to screen
equipment must be included in the equipment counts. For these
equipment pieces, the average emission factors must be used to
estimate emissions.
In other cases, it is not possible to screen every equipment
piece due to cost considerations. This is particularly true for
connectors. Appendix D provides criteria for determining how
many connectors must be screened to constitute a large enough
sample size to identify the screening value distribution for
connectors. If the criteria in Appendix D are met, the average
emission rate for connectors that were screened can be applied to
connectors that were not screened. It should be noted that if
connectors must be included in a leak detection and repair
program as part of an equipment leaks standard, then all
connectors must be screened. For equipment types other than
connectors, if they are not monitored, the Average Emission
Factor approach should be used to estimate emissions.
2.4.5 Using Screening Data Collected at Several Different
Times
When screening data is collected and used to estimate
emissions, the emissions estimate represents a "snapshot" of
emissions at the time the screening data were obtained. Over
time, it is possible that more screening data will be collected,
and that for individual equipment pieces, several screening
values will have been obtained at different time periods. For
example, if quarterly monitoring is performed on a valve, in an
annual period four screening values will be obtained from the
valve. The annual emissions from the valve should be calculated
by determining the emissions for each quarter based on the
2-44
-------
operational hours for the quarter, and summing the quarterly
emissions together to arrive at emissions for the entire year.
See Appendix A for an example of estimating emissions from an
equipment piece for which more than one screening value has been
obtained.
2.4.6 Estimating VOC Emission Rates from Equipment Containing
Non-VOC's
Some organic compounds not classified as VOC's can be
detected by the screening instrument. Because the compounds are
detected, the emissions associated with the screening value will
include emissions of the "non-VOC's." The two key organic
compounds not classified as VOC's are methane and ethane, but
other organic compounds not classified as VOC's include methylene
chloride, 1,1,1-trichloroethane, and several chlorofluorocarbons.
An approach very similar to that outlined in Section 2.4.1
(Speciating Emissions) is used to estimate VOC emissions from
equipment containing these non-VOC's mixed with VOC's.
Once screening data have been used to estimate the
"uncorrected" VOC emissions (either by using the screening range
emission factors or the correlations), the corrected VOC
emissions from a group of equipment containing similar
composition can be calculated using the equation:
VOCcorr = VOCuncorr * WPvoc/WPorg
where:
VOCcorr = The corrected VOC mass emissions from the
equipment (kg/hr);
VOCuncorr = The previously calculated "uncorrected" VOC
mass emissions from the equipment (kg/hr);
WPVOC = FOR SOCMI AND GAS PLANTS: The concentration
of VOC in the equipment in weight percent;
= FOR REFINERIES: The concentration of VOC plus
the concentration of methane in the equipment
in weight percent (up to a maximum of
10 percent by weight methane); and
2-45
-------
wporg = Tne total concentration of organic compounds
in the equipment in weight percent.
Note that the term "WPVOCM is defined differently for refineries
than for SOCMI and natural gas plants. It is necessary to add
the methane weight percent back into the "WPVOC" term when
applied to refineries because the refinery screening range
factors and correlations predict non-methane organic compound
emissions, whereas the SOCMI and natural gas plant screening
range factors and the SOCMI correlations predict total organic
compound emissions.
2.4.7 Estimating Equipment Leak Emissions of Inorganic
Compounds
The majority of data collected for estimating equipment leak
emissions has been for VOC's and not for inorganic compounds.
Accordingly, the emission factors and correlations presented in
Section 2.3 are not intended to be applied for the used of
estimating emissions of inorganic compounds. However, in some
cases, there may be a need to estimate equipment leak emissions
of inorganic compounds—particularly for those that exist as a
gas/vapor or for those that are volatile. Some examples of
inorganic compounds include sulfur dioxide, ammonia, and
hydrochloric acid.
The best way to estimate equipment leak emissions of
inorganic compounds would be to develop unit-specific
correlations as described in Section 2.3.4. To do this, it would
be necessary to obtain a portable monitoring instrument that
could detect the inorganic compounds. If it is not possible to
develop a unit-specific correlation, but a portable monitoring
instrument (or some other approach) can be used to indicate the
actual concentration of the inorganic compound at the equipment
leak interface, then the "screening values" obtained with this
instrument can be entered into the applicable correlations
presented in Section 2.3.3 to estimate emissions. Alternatively,
the equal to or greater than 10,000 ppmv, or the less than 10,000
ppmv emission factors could be applied. In the event that there
2-46
-------
is no approach that can be used to estimate the concentration of
the inorganic compound at the leak interface, then in the absence
of any other data, the average emission factors can be used.
2-47
-------
2.5 REFERENCES
1. Radian Corporation. Assessment of Atmospheric Emissions
from Petroleum Refining: Volume 4. Appendices C, D, E.
Prepared for U.S. Environmental Protection Agency. Research
Triangle Park, NC. Publication No. EPA-600/ 2-80-075d.
July 1980.
2. Wetherhold, R.G., L.P. Provost, and C.D. Smith (Radian
Corporation). Assessment of Atmospheric Emissions from
Petroleum Refining: Volume 3. Appendix B. Prepared for
U.S. Environmental Protection Agency. Research Triangle
Park, NC. Publication No. EPA-600/2-80-075c. April 1980.
3. DuBose, D.A., J.I. Steinmetz, and G.E. Harris (Radian
Corporation). Frequency of Leak Occurrence and Emission
Factors for Natural Gas Liquid Plants. Final Report.
Prepared for U.S. Environmental Protection Agency. Research
Triangle Park, NC. EMB Report No. 80-FOL-l. July 1982.
4. Blacksmith, J.R., et al. (Radian Corporation.) Problem-
Oriented Report: Frequency of Leak Occurrence for Fittings
in Synthetic Organic Chemical Plant Process Units. Prepared
for U.S. Environmental Protection Agency. Research Triangle
Park, NC. Publication No. EPA-600/2-81-003.
September 1980.
5. Langley, G.J. and R.G. Wetherhold. (Radian Corporation.)
Evaluation of Maintenance for Fugitive VOC Emissions
Control. Prepared for U.S. Environmental Protection Agency.
Research Triangle Park, NC. Publication No. EPA-600/
S2-81-080. May 1981.
6. U.S. Environmental Protection Agency. Fugitive Emission
Sources of Organic Compounds - Additional Information of
Emissions, Emission Reductions, and Costs (Section 2).
Research Triangle Park, NC. Publication No. EPA-450/3-
82-010. April 1982.
2-48
-------
3.0 SOURCE SCREENING
3.1 INTRODUCTION
This chapter presents protocols for screening equipment
components with a portable volatile organic compound (VOC)
analyzer. When performing source screening, the portable
analyzer probe opening is placed at the leak interface of the
equipment component to obtain a "screening" value. The screening
value is an indication of the concentration level of any leaking
material at the leak interface. A screening value is not a
direct measure of mass emissions rate, but, as discussed in
Chapter 2.0, can be entered into a mass emissions/screening value
correlation equation to estimate mass emissions.
This chapter is divided into two sections. The first
section provides a description of the portable analyzers that can
be used when conducting screening surveys. Operating principles
of the analyzers and performance criteria and specifications in
EPA Reference Method 21 (the method describing the use of
portable VOC analyzers)1 are described, and the use of monitoring
devices that do not meet Method 21 requirements is discussed.
The second section presents the protocol for successfully
conducting a screening program. This section includes methods to
identify components to be included in the screening program, a
discussion on the development of a systematic approach for
performing the screening survey, the protocol for screening each
of the equipment types, and recommendations for collecting and
handling data.
3.2 MONITORING INSTRUMENTS
A number of portable VOC detection devices have the
potential to measure the concentration level at the leak
interface of equipment. Any analyzer can be used, provided it
meets the specifications and performance criteria set forth in
EPA Method 21, Section 3.O.1 Method 21 is included in this
document as Appendix E.
In general, portable VOC monitoring instruments are equipped
with a probe that is placed at the leak interface of a piece of
3-1
-------
equipment. A pump within the instrument draws a continuous
sample of gas from the leak interface area to the instrument
detector. The instrument response is a screening value—that is,
a relative measure of concentration level. The screening value
is in units of parts per million by volume (ppmv). However, the
screening value does not necessarily indicate the actual total
concentration at the leak interface of the compound(s) being
detected because the sensitivity of instruments vary for
different compounds. As discussed in Section 2.4.2, response
factors (RF's) relate actual concentration of a compound to the
observed concentration from the detector. Before a monitoring
instrument is used, it must first be calibrated using a reference
gas containing a known compound at a known concentration.
Methane and isobutylene are frequently used reference compounds.
3.2.1 Operating Principles and Limitations of Portable VOC
Detection Devices
Monitoring instruments operate on a variety of detection
principles, with the three most common being ionization, infrared
absorption, and combustion. Ionization detectors operate by
ionizing the sample and then measuring the charge (i.e., number
of ions) produced. Two methods of ionization currently used are
flame ionization and photoionization. Each of these detector
types are briefly described below.
A standard flame ionization detector (FID) theoretically
measures the total carbon content of the organic vapor sampled,
but many other factors influence the FID readout. Although
carbon monoxide and carbon dioxide (CO2) do not produce
interferences, FID's react to water vapor at a low sensitivity.
Furthermore, erratic readings may result if water condenses in
the sample tube. A filter is used to remove particulate matter
from the sample. Certain portable FID instruments are equipped
with gas chromatograph (GC) options making them capable of
measuring total gaseous nonmethane organics or individual organic
components. Certain organic compounds containing nitrogen,
oxygen, or halogen atoms give a reduced response when sampled
3-2
-------
with an FID, and the FID may not respond to some organic
compounds.
Photoionization detectors use ultraviolet light (instead of
a flame) to ionize organic vapors. As with FID's, the detector
response varies with the functional group in the organic
compounds. Photoionization detectors have been used to detect
equipment leaks in process units in the SOCMI, especially for
certain compounds, such as formaldehyde, aldehydes, and other
oxygenated compounds, which will not give a satisfactory response
on a FID or combustion-type detector.
Nondispersive infrared (NDIR) instruments operate on the
principle of light absorption characteristics of certain gases.
These instruments are usually subject to interference because
other gases, such as water vapor and C(>2, may also absorb light
at the same wavelength as the compound of interest. These
detectors are generally used only for the detection and
measurement of single components. For this type of detection,
the wavelength at which a certain compound absorbs infrared
radiation is predetermined and the device is preset for that
specific wavelength through the use of optical filters. For
example, if set to a wavelength of 3.4 micrometers, infrared
devices can detect and measure petroleum fractions, including
gasoline and naphtha.
Combustion analyzers are designed either to measure the
thermal conductivity of a gas or to measure the heat produced by
combustion of the gas. The most common method in which portable
VOC detection devices are used involves the measurement of the
heat of combustion. These detection devices are referred to as
hot wire detectors or catalytic oxidizers. Combustion analyzers,
like most other detectors, are nonspecific for gas mixtures. In
addition, combustion analyzers exhibit reduced response (and, in
some cases, no response) to gases that are not readily combusted,
such as formaldehyde and carbon tetrachloride.
3-3
-------
3.2.2 Specifications and Performance Criteria of Portable VOC
Detection Devices
As previously stated, any portable analyzer may be used as a
screening device, provided it meets the specifications and the
performance criteria called for in Method 21. (See Appendix E.)
Method 21 specifies the requirements that must be met when a
facility is collecting screening data to comply with a
regulation. The requirements of Method 21 are also applicable
when screening data are collected for the sole purpose of
estimating emissions. When the requirements of Method 21 refer
to a "leak definition," this is the screening value indicating
that a piece of equipment is leaking as defined in the applicable
regulation. If screening data are collected for the sole purpose
of estimating emissions, the equivalent to the "leak definition"
concentration in the text that follows is the highest screening
value (i.e., 10,000 ppmv) that the monitoring instrument can
readout.
Method 21 requires that the analyzer meet the following
specifications:!
The VOC detector should respond to those organic
compounds being processed (determined by the RF);
Both the linear response range and the measurable range
of the instrument for the VOC to be measured and the
calibration gas must encompass the leak definition
concentration specified in the regulation;
• The scale of the analyzer meter must be readable to
±2.5 percent of the specified leak definition
concentration;
• The analyzer must be equipped with an electrically
driven pump so that a continuous sample is provided at
a nominal flow rate of between 0.1 and 3.0 liters per
minute;
• The analyzer must be intrinsically safe for operation
in explosive atmospheres; and
• The analyzer must be equipped with a probe or probe
extension for sampling not to exceed .25 inch in
outside diameter, with a single end opening for
admission of sample.
3-4
-------
Note that the suction flow rate span allowed by Method 21 is
intended to accommodate a wide variety of instruments, and
manufacturers guidelines for appropriate suction flow rate should
be followed.
In addition to the above specifications, criteria for the
calibration gases to be used are specified. A minimum of two
calibration gases are required for analyzer performance
evaluation. One is a "zero" gas, which is defined as air with
less than 10 ppmv VOC; the other calibration gas, or reference
gas, uses a specified reference compound in an air mixture. The
concentration of the reference compound must approximately equal
the leak definition specified in the regulation. If cylinder
calibration gas mixtures are used, they must be analyzed and
certified by the manufacturer to be within ±2 percent accuracy.
The shelf life must also be specified. Calibration gases can
also be prepared by the user as long as they are accurate to
within ±2 percent.
The instrument performance criteria that each analyzer must
meet are presented in Table 3-1 and discussed in greater detail
in the following sections.
3.2.2.1 Response Factor. The sensitivity of an analyzer
varies, depending on the composition of the sample and
concentration of VOC detected. The RF quantifies the sensitivity
of the analyzer to each compound. The RF is defined by:
RF = Actual Concentration of Compound
Observed Concentration from Detector
An RF must be determined for each compound that is to be
measured. Response factors may be determined either by testing
or from referenced sources. (The RF's for many commonly screened
compounds are presented in Appendix C.) The RF tests are
required before placing the analyzer into service, but do not
need to be repeated. The RF for each compound to be measured
must be less than 10 for an analyzer to be acceptable for use in
a screening program. According to Method 21, the RF can either
3-5
-------
TABLE 3-1. PERFORMANCE CRITERIA FOR PORTABLE VOC DETECTORS3
Criteria
Requirement
Time interval
Instrument
response factor
Instrument
response time
Calibration
precision
Must be <10 unless
correction curve is
used
Must be <30 seconds
Must be <10 percent
of calibration gas
value
One time, before
detector is put in
service.
One time, before
detector is put in
service. If
modification to sample
pumping or flow
configuration is made,
a new test is required.
Before detector is put
in service and at
3-month intervals or
next use, whichever is
later.
aSource: Reference 1.
3-6
-------
be measured in the laboratory using a prepared gas concentration
at 80 percent of the applicable leak definition, or it can be
taken from values published in the literature. When no
instrument is available that meets this criteria when calibrated
with the reference compound specified in the applicable
regulation, the available instrument may be calibrated with one
of the VOC's to be measured. However, the analyzer RF must still
be less than 10 for each VOC to be measured.
As discussed in Section 2.4.2, RF's depend on several
parameters, including the compound, the screening value, the
monitoring instrument, and the calibration gas. In Chapter 2.0,
guidance was provided on when and how to apply RF's for
estimating emissions. Methods were presented on calculating an
RF for a given chemical at a screening value other than one for
which data were published. Methods were also presented for
calculating RF's for mixtures.
In this chapter, several additional issues pertaining to
RF's are discussed. These issues are (1) the consideration of
RF's when selecting a monitoring instrument, (2) how laborsitory
analysis can be performed to generate data to determine an RF for
a given compound, and (3) when laboratory analysis is
recommended.
Response factors contained in Appendix C can be used as a
guide for selecting an appropriate monitoring device. If at the
applicable leak definition, the RF of an instrument is greater
than 10, that instrument does not meet Method 21 requirements
unless a substitute reference gas is used to calibrate the
instrument. For example, at a concentration of 10,000 ppmv, it
can be seen that when screening equipment in a process unit that
contains cumene, an FID can be used (RF = 1.92 at an actual
concentration of 10,000 ppmv), while the catalytic oxidation
detector cannot (RF = 12.49). Similarly, at a concentration of
10,000 ppmv, neither of these devices respond to carbon
tetrachloride and, therefore, cannot be used unless calibrated
with a substitute VOC such that an RF of under 10 can be
calculated for this compound.
3-7
-------
Response factors can be determined by laboratory analysis
using the following method. First, the analyzer is calibrated
using the reference gas. Then, for each organic species that is
to be measured, a known standard in air is obtained or prepared.
The concentration of the organic species should be at
approximately the leak definition value. This mixture is then
injected into the analyzer and the observed meter reading is
recorded. The analyzer is then "zeroed" by injecting zero air
until a stable reading is obtained. The procedure is repeated by
alternating between the mixture and zero air until a total of
three measurements are obtained. An RF is calculated for each
repetition and then averaged over the three measurements. This
procedure can be repeated at several different concentration
values. The data can then be used to generate a curve that
relates RF to screening value. (See Appendix A.)
The most accurate method for estimating RF's is to perform
laboratory analysis. This is particularly true because RF's
vary, not just for the detector type, but also for each
individual instrument. However, in some cases, time and resource
constraints may require the use of published RF data.
Nevertheless, a limitation of the published data is that it is
typically specific to a pure compound for a single actual
concentration value, detector type, and calibration gas.
Additionally, although an RF for mixtures can be calculated as
described in Section 2.4.2 (i.e., if an RF is known for each
individual compound), the most accurate RF for a mixture is
calculated by preparing known standards of air for the mixture
and injecting the standard into the analyzer as described
earlier.
3.2.2.2 Response Time. The response time of an analyzer is
defined as the time interval from a step change in VOC
concentration at the input of a sampling system, to the time at
which the corresponding concentration value is reached as
displayed on the analyzer readout meter. Response time is
determined by introducing zero air into the instrument sample
probe. When the meter reading has stabilized, the specified
3-8
-------
calibration gas is injected. The response tine is measured as
the time lapsed between switching to the calibration gas and the
time when 90 percent of the final stable reading is obtained.
This test is performed three times and the response time is
calculated as the average of the three tests. The response time
must be equal to or less than 30 seconds for the analyzer to be
acceptable for screening purposes.
The response time test is required before placing an
analyzer in service. The response time must be determined for
the analyzer configuration that will be used during testing. If
a modification to the sample pumping system or flow configuration
is made that would change the response time (e.g., change in
analyzer probe or probe filter, or the instrument pump), a new
test is required before the screening survey is conducted.
3.2.2.3 Calibration Precision. Calibration precision is
the degree of agreement between measurements of the same known
value. To ensure that the readings obtained are repeatable, a
calibration precision test must be completed before placing the
analyzer in service, and at 3-month intervals, or at the next
use, whichever is later. The calibration precision must be equal
to or less than 10 percent of the calibration gas value.
To perform the calibration precision test, three
measurements are required for each non-zero concentration.
Measurements are made by first introducing zero gas and adjusting
the analyzer to zero. The specified calibration gas (reference)
is then introduced and the meter reading is recorded. This
procedure must be performed three times. The average algebraic
difference between the meter readings and the known value of the
calibration gas is then computed. This average difference is
then divided by the known calibration value and multiplied by 100
to express the resulting calibration precision as percent. The
calibration precision of the analyzer must be equal to or less
than 10 percent of the calibration gas value.
3.2.2.4 Safety. Portable instruments to detect VOC
emissions from equipment leak sources are required to be used in
potentially hazardous locations such as petroleum refineries and
3-9
-------
bulk gasoline terminals. The National Electrical Code requires
that instruments to be used in hazardous locations be certified
to be explosion-proof, intrinsically safe, or purged.
Hazardous locations are divided into three classes:
Class I, class II, and Class III. Each class is divided into two
divisions (Division 1 or 2) according to the probability that a
hazardous atmosphere will be present and also into seven groups,
depending on the type of hazardous material exposure: Groups A
through D are flammable gases or vapors, and Groups E, F, and G
apply to combustible or conductive gases. Class I, Division 1,
Groups A, B, C, and D locations are those in which hazardous
concentrations of flammable gases or vapors may exist under
normal operating conditions. Class I, Division 2, Groups A,B, C,
and D locations are those in which hazardous concentrations of
flammable gases may exist only under unlikely conditions of
operation.
Any instrument considered for use in potentially hazardous
environments must be classified as intrinsically safe for
Class I, Division 1 and Class II, Division l conditions at a
minimum. The instrument must not be operated with any safety
device, such as an exhaust flame arrestor, removed.
Table 3-2 lists several portable VOC detection instruments.
Table 3-2 includes manufacturer, model number, pollutants
detected, principle of operation, and range. Note that
additional instruments, not listed here, may be available.
3.2.3 Use of Monitoring Devices That Do Not Meet Method 21
Requirements
In some cases, a monitoring device may not be available that
meets all of the performance specifications of Method 21. For
example, there are several cases (e.g., phosgene) where the RF at
10,000 ppmv is greater than 10. An instrument may meet all other
requirements, but fail as a Method 21 instrument because it
cannot meet the RF requirement. If an instrument fails to meet
Method 21 requirements, it can still be used for the purpose of
estimating emissions if its reliability can be documented.
3-10
-------
CO
EH
CO
55
H
S5
O
H
S
H
EH
U
g
w
9
OJ
I
n
S
s
£u
c «j
52
i«
I
nt Co
ornia
Is.
f •
u —
a u
L.
§
o
u
u
I-
*•
8
Z - s
.o .o a
L.
tl
& S
= i
s
o o
o
o
§
5 8
• i
o o
»- •- in
ii i
oo o
g s
U U
i -
*• 0) — A
fep:
-8*;c
S
i
2
i
t^ i
o £ x
a
v
c
—
o — •
CO
»*§
Igl
5S
§ 5
3« jj«
SS -S
^ «S
5^ «-g
I1 f 1
4J .c *n »
«» 0»
it) .- .-
: X X a.
8*<
«J
X CL
3-11
-------
Two primary steps must be taken to document the reliability
of an analyzer that fails to meet the Method 21 requirements.
First, a laboratory program must be undertaken to demonstrate the
response of the monitoring instrument to the compounds being
measured; that is, an instrument response curve must be developed
for the entire screening value range and documented so that
screening values taken in the field can be adjusted to actual
concentrations if necessary. Second, the testing program must be
sufficiently well-documented to demonstrate how the instrument
will be used when screening equipment. For example, if the
response time of the candidate instrument exceeds the Method 21
performance specification, the test plan should reflect added
screening time at each potential leak point. Once this
laboratory demonstration has been completed and the screening
value correction curve has been established, the instrument may
be used in a screening program.
3.3 THE SCREENING PROGRAM
The goal of the screening program is to measure VOC
concentrations at seals, shafts, and other potential leak points.
All equipment to be included in the screening survey needs to be
identified before the screening program starts. A list of
equipment types that are potential sources of fugitive emissions
is provided in Table 3-3.
3.3.1 Identification of Equipment to be Screened
The first step in the screening survey is to precisely
define the process unit boundaries. This is usually
straightforward, but occasionally multiple units may be built on
the same pad and share some common facilities. A process unit
can be defined as the smallest set of process equipment that can
operate independently and includes all operations necessary to
achieve its process objective. The exact basis for the unit
definition should be documented. A plot plan of the unit should
be obtained and marked with the appropriate boundaries.
The next step is to obtain a simplified flow diagram of the
process and note the process streams. The actual screening and
data collection can be done efficiently by systematically
3-12
-------
TABLE 3-3. EQUIPMENT LEAK EMISSION SOURCES
Equipment types
Pump seals
Compressor seals
Valves
Pressure relief devices
Flanges, screwed connections, etc.
Open-ended lines
Agitator seals
Service
Gas/vapor
Light liquid
Heavy liquid
3-13
-------
following each stream. For example, a logical starting point
would be where one of the feed lines enters the process boundary.
The screening team would follow that line, screening all sources,
until the line terminates at the connectors of a reactor or
separation step. This approach offers the advantage of screening
groups of equipment with roughly the same composition of material
in the line. Screening would then continue on the outlet side of
the reactor or separation equipment. Minor loops, such as a
bypass around a control valve, pump, or heat exchanger, should be
screened on the initial pass. Larger loops of process equipment,
such as parallel passes and processing alternatives, are more
effectively treated as separate streams.
Each source should be uniquely identified to indicate that
it has been screened. For example, sources can be tagged. Tags
can consist of any form of weatherproof and readily visible
identification. Alternatively, a process unit can be considered
appropriately tagged if the unit has a system of identifying
markings with an associated diagram allowing easy location of
marked sources. Once all the equipment along the major streams
has been screened, the unit should be divided into a grid to
search for fittings missed on the initial survey. Consistent
with equipment leaks standards, equipment that is unsafe to
monitor or very difficult to access does not need to be included
in the survey. Documentation must be provided, however, to
substantiate the unsafe or confined nature of such equipment.
3.3.2 Procedure for Screening
Once the equipment to be screened has been identified, the
procedures outlined in Method 21 to screen each equipment type
are followed.1 The probe inlet is placed at the surface of the
potential leak interface where leakage could occur. (The
potential leak interface is the boundary between the process
fluid and the atmosphere.) For equipment with no moving parts at
the leak interface, the probe should be placed directly on the
leak interface; for equipment with moving parts (e.g., pumps,
compressors, and agitators), the probe should be placed
approximately 1 centimeter off from the leak interface. Care
3-14
-------
must be taken to ensure that the probe is held perpendicular, not
tangential, to the leak potential interface; otherwise,
inaccurate readings will result. The probe must then be moved
along the interface periphery while observing the instrument
readout. If an increased meter reading is observed, slowly move
the probe along the interface where concentrations register until
the maximum meter reading is obtained. The probe inlet should be
left at this maximum reading location for approximately two times
the instrument response time. The maximum reading is recorded as
the screening value.
The instrument measurement may exceed the scale of the
instrument. This is referred to as a "pegged" readout. For
example, for several instruments, the highest readout on the
scale is 10,000 ppmv. For the purposes of generating an
emissions estimate, a dilution probe should be employed to
measure concentrations greater than the instrument's normal range
unless average emission factors for greater than or equal to the
"pegged" readout are applied. It is important to note that
extending the measurement range necessitates the calibration of
the instrument to the higher concentrations.
Care should be taken to avoid fouling the probe with grease,
dust, or liquids. A short piece of Teflon* tubing can be used as
a probe tip extender and then can be snipped off as the tip
fouls. In areas with a noticeable particulate loading, this
tubing can be packed loosely with untreated fiberglass, which
acts as a filter. (Note that the instrument must also be
calibrated with this filter in place.) If a surface to be
screened is obviously dirty, hold the probe tip just over the
surface to avoid scooping up contaminants. Some fouling is
unavoidable, so it is recommended that the probe tip filter be
cleaned at least daily and any other filters on a weekly basis.
Normally, these filters can be cleaned by just tapping them
lightly on a table top, but if the deposits are wet and caked on,
they should be washed with an aqueous solution of soap and
alcohol. This solution also can be used to wash the probe and
3-15
-------
transfer line periodically. Care should be taken to blow the
equipment dry before reuse.
This general procedure can be used to screen equipment such
as valves, connectors, pumps and compressors, pressure relief
devices, and other potential sources of VOC leakage, such as
open-ended lines or valves. The following sections describe the
location on each type of equipment where screening efforts should
be concentrated.
3.3.2.1 Valves. For valves, the most common source of
leaks is at the seal between the stem and housing. To screen
this source, the probe opening is placed where the stem exits the
packing gland and is moved around the stem circumference. The
maximum reading is recorded as the screening valve. Also, the
probe opening is placed at the packing gland take-up connector
seat, and the probe is moved along the periphery. In addition,
valve housings of multipart assemblies should be screened at the
surface of all points where leaks could occur. Figures 3-1
through 3-5 illustrate screening points for several different
types of valves.
3.3.2.2 Connectors. For connectors, the probe opening is
placed at the outer edge of the connector - gasket interface and
the circumference of the connector is sampled. For screwed
connectors, the threaded connection interface must also be
screened. Other types of nonpermanent joints, such as threaded
connections, are sampled with a similar traverse.
3.3.2.3 Pumpsf Compressors, and Agitators. Pumps,
compressors, and agitators are screened with a circumferential
traverse at the outer surface shaft and seal interface where the
shaft exits the housing. If the source is a rotating shaft, the
probe inlet is positioned within 1 centimeter of the shaft - seal
interface. If the housing configuration prevents a complete
traverse of the shaft periphery, all accessible portions must be
sampled. All other joints on the pump or compressor housing
where leakage could occur should also be sampled. Figure 3-6
illustrates screening points for two types of centrifugal pumps.
3-16
-------
Screen
Here
Stem
Packing Gland
Packing
Packing Out
Screen
Here
Disk or Wedge
Packing Nut
Nonrising Stem Type
Packing Gland
Packing
Disk or Wedge
Rising Stem Type
Figure 3-1. Gate Valves
3-17
5
tD
s
-------
Stem
Packing
Gland
Packing-
Screen
J / "ere
./y Packing Nut
*** Screen
'Here
Body
Manual Globe Valve
Globe Type Control Valve
Figure 3-2. Globe Valves
3-18
-------
Screen Here
Figure 3-3. Lubricated Plug Valve
3-19
-------
Screen Here
Ball Valve
Screen Here
Disk
Butterfly Valve
Figure 3-4. Ball Valve and Butterfly Valve
3-20
-------
Stem
Screen Here
Weir-Type Diaphragm Valve
Screen Here
Screen
Here
Lift Check
Check Valves
Figure 3-5. Weir-Type Diaphragm Valve and Check Valves
3-21
I
-------
Moving Shan
Screen Here
Stationary Casing
Vertical Centrifugal Pump
Screen Here
Horizontal Centrifugal Pump
Figure 3-6. Centrifugal Pumps
3-22
-------
3.3.2.4 Pressure Relief Devices. The configuration of most
pressure relief devices prevents sampling at the sealing seat.
>
Because of their design and function, pressure relief devices
must be approached with extreme caution. These devices should
not be approached during periods of process upsets, or other
times when the device is likely to activate. Similarly, care
must be used in screening pressure relief devices to avoid
interfering with the working parts of the device (e.g., the seal
disk, the spring, etc.) For those devices equipped with an
enclosed extension, or horn, the probe inlet is placed at
approximately the center of the exhaust area to the atmosphere.
It should be noted that personnel conducting the screening should
be careful not to place hands, arms, or any parts of the body in
the horn. Figure 3-7 illustrates the screening points for a
spring-loaded relief valve.
3.3.2.5 Qpen-Ended Lines. Fugitive leaks from open-ended
lines are emitted through a regularly shaped opening. If that
opening is very small (as in sampling lines of less than 1 inch
in diameter), a single reading in the center is sufficient. For
larger openings it is necessary to traverse the perimeter of the
opening. The concentration at the center must also be read.
3.3.3 Data Handling
To ensure that data quality is maintained, it is recommended
that data be recorded on prepared data sheets. The data
collected should include the following:
1. Monitoring instrument type and model number.
2. Operator's name.
3. Date.
4. Component identification number (ID number). (If
permanent ID'S are not in place, assign ID's as each
source is screened.)
5. Component type (i.e., valve, connector, open-ended
line, etc.)
3-23
-------
Screen Here
Tension Adjustment
Thimble
Spring
Disk
Alternate Screening
Area if Horn
Inaccessible Nozzle
To Process
Figure 3-7. Spring-Loaded Relief Valve
3-24
-------
6. Location/stream. (Provide brief description of where
the screened component is located and the composition
of material in the equipment.)
7. Service (i.e., gas, light liquid, or heavy liquid).
8. Number of hours per year the component is in service.
9. Screening value.
10. Background concentration.
11. Comments. If any explanation is required, it should be
noted in a "comments" section.
In some cases, it may be necessary or desirable to adjust
the screening values for RF. In these cases, the data sheet
should be designed to accommodate extra columns for RF and
corrected screening values. Table 3-4 provides an example data
sheet that may be used to log measurements taken during a
screening program.
3-25
-------
5
a
I
*.-
HI
I*
8
£
1
L. L.
o o
V k w
*• II 4J
4i 5. a
o O a
3-26
-------
3.4 REFERENCES
1. Code of Federal Regulations, Title 40, Part 60, Appendix A.
Reference Method 21, Determination of Volatile Organic
Compound Leaks. Washington, DC. U.S. Government Printing
Office. Revised June 22, 1990.
3-27
-------
4.0 MASS EMISSIONS SAMPLING
4.1 INTRODUCTION
This chapter describes the procedures for "bagging"
equipment to measure mass emissions of organic compounds. An
equipment component is bagged by enclosing the component to
collect leaking vapors. Measured emission rates from bagged
equipment coupled with screening values can be used to develop
unit-specific screening value/mass emission rate correlation
equations. Unit-specific correlations can provide precise
estimates of mass emissions from equipment leaks at the process
unit. However, it is recommended that unit-specific correlations
are only developed in cases where the existing EPA correlations
do not give reasonable mass emission estimates for the process
unit. The focus of the chapter is on bagging equipment
containing organic compounds, but similar procedures can be
applied to bag equipment containing inorganic compounds as long
as there are comparable analytical techniques for measuring the
concentration of the inorganic compound.
This chapter is divided into four sections. In Section 4.2,
the methods for bagging equipment are discussed. Considerations
for bagging each equipment type are discussed in Section 4.3. In
Section 4.4, techniques used in the laboratory analysis of bagged
samples are discussed. Section 4.4 also includes a description
of a rigorous calibration procedure for the portable monitoring
device that must be followed. Finally, in Section 4.5, quality
assurance and quality control (QA/QC) guidelines are provided.
4.2 SAMPLING METHODS
The emission rate from an equipment component is measured by
bagging the component—that is, isolating the component from
ambient air to collect any leaking compound(s). A tent
(i.e., bag) made of material impermeable to the compound(s) of
interest is constructed around the leak interface of the piece of
equipment. A known rate of carrier gas is induced through the
bag and a sample of the gas from the bag is collected and
analyzed to determine the concentration (in parts per million by
4-1
-------
volume [ppmv]) of leaking material. The concentration is
measured using laboratory instrumentation and procedures. Mass
emissions are calculated based on the measured concentration and
the flow rate of carrier gas through the bag.
In some cases, it may be necessary to collect liquid leaking
from a bagged equipment piece. Liquid can either be dripping
from the equipment piece prior to bagging, and/or be formed as
condensate within the bag. If liquid accumulates in the bag,
then the bag should be configured so that there is a low point to
collect the liquid. The time in which the liquid accumulates
should be recorded. The accumulated liquid should then be taken
to the laboratory and transferred to a graduated cylinder to
measure the volume of organic material. Based on the volume of
organic material in the cylinder (with the volume of water or
nonorganic material subtracted out), the density of the organic
material, and the time in which the liquid accumulated, the
organic liquid leak rate can be calculated. Note that the
density can be assumed to be equivalent to the density of organic
material in the equipment piece, or, if sufficient volume is
collected, can be measured using a hydrometer. It should be
noted that in some cases condensate may form a light coating on
the inside surface of the bag, but will not accumulate. In these
cases, it can be assumed that an equilibrium between condensation
and evaporation has been reached and that the vapor emissions are
equivalent to total emissions from the source.
When bagging an equipment piece, the enclosure should be
kept as small as practical. This has several beneficial effects:
• The time required to reach equilibrium is kept to a
minimum;
• The time required to construct the enclosure is
minimized;
• A more effective seal results from the reduced seal
area; and
4-2
-------
• Condensation of heavy organic compounds inside the
enclosure is minimized or prevented due to reduced
residence time and decreased surface area available for
heat transfer.
Two methods are generally employed in sampling source
enclosures: the vacuum method and the blow-through method. Both
methods involve enclosing individual equipment pieces with a bag
and setting up a sampling train to collect two samples of leaking
vapors to be taken to the laboratory for analysis. Both methods
require that a screening value be obtained from the equipment
piece prior to and after the equipment piece is enclosed. The
methods differ in the ways in which the carrier gas is conveyed
through the bag. In the vacuum method, a vacuum pump is used to
pull air through the bag. In the blow-through method, a carrier
gas such as nitrogen (or other inert gas) is blown into the bag.
In general, the blow-through method has advantages over the
vacuum method. These advantages are as follows.
(1) The blow-through method is more conducive to better
mixing in the bag.
(2) The blow-through method minimizes ambient air in the
bag and thus reduces potential error associated with
background organic compound concentrations. (For this
reason the blow-through method is especially preferable
when measuring the leak rate from components with zero
or very low screening values.)
(3) The blow-through method minimizes oxygen concentration
in the bag (assuming air is not used as the carrier
gas) and the risk of creating an explosive environment.
(4) In general, less equipment is required to set up the
blow-through method sampling train.
However, the blow-through method does require a carrier gas
source, and preferably the carrier gas should be inert and free
of any organic compounds and moisture. The vacuum method does
not require a special carrier gas.
Details of the sampling train of each of these bagging
methods are discussed in Sections 4.2.1 and 4.2.2, respectively.
These sections also contain summaries of the steps of the
sampling procedure for each method. For both methods, the
approach described above for collecting and measuring liquid leak
4-3
-------
rates can be utilized. In addition to the sampling descriptions
presented in the following sections, the quality control and
assurance guidelines presented in Section 4.5 must also be
followed when bagging equipment.
4.2.1 Vacuum Method
The sampling train used in the vacuum method is depicted in
Figure 4-1. The train can be mounted on a portable cart, which
can be moved around the process unit from component to component.
The major equipment items in the sampling train are the vacuum
pump used to draw air through the system, and the dry gas meter
used to measure the flow rate of gas through the train. In
previous studies that EPA conducted, a 4.8-cubic feet per minute
Teflon* ring piston-type vacuum pump equipped with a
3/4-horsepower, air-driven motor was used. Other equipment that
may be used in the train includes valves, copper and stainless
steel tubing, Teflon* tubing and tape, thermometer, pressure-
reading device, liquid collection device, and air-driven
diaphragm sampling pumps. It also may be necessary to use
desiccant preceding the dry gas meter to remove any moisture.
The bag is connected by means of a bulkhead fitting and
Teflon* tubing to the sampling train. A separate line is
connected from the bag to a pressure-reading device to allow
continuous monitoring of the pressure inside the bag. If a
significant vacuum exists inside the bag when air is being pulled
through, a hole is made in the opposite side of the bag from the
outlet to the sampling train. This allows air to enter the bag
more easily and, thus, reduces the vacuum in the enclosure.
However, it is important to maintain a vacuum in the bag, since
VOC could be lost through the hole if the bag became pressurized.
In practice, it has been found that only a very slight vacuum
(0.1 inches of water) is present in the bag during most of the
sampling, even in the absence of a hole through the bag wall.
Sufficient air enters around the seals to prevent the development
of a significant vacuum in the bag. A small diaphragm sampling
pump can be used to collect two samples into sample bags or
4-4
-------
•o
o
0)
g
O
(0
0)
x:
•P
c
•H
(0
0)
o
M
g
(fl
Cn
en
(0
e
•rH
(0
-P
tJl
(0
4-5
-------
canisters, which are then transported to the laboratory for
analysis.
The diaphragm pump can also be used to collect a background
sample of the ambient air near the bagged component. The
concentration in the background bag is subtracted from the
average concentration in the sample bags when calculating the
leak rate. Often this correction is insignificant (particularly
for components with high leak rates or in cases where there is no
detectable volatile organic compound (VOC) concentration measured
by the portable monitoring device), and collection of a
background bag is optional. However, in some cases collection of
a background bag is important so that emission rates are not
biased high.
Any liquid that accumulates in the bag should be collected
using the approach described in Section 4.2. Note that if there
is a concern that condensation will occur in equipment downstream
from the bag outlet, a cold trap can be placed as close to the
bag outlet as possible to remove water or heavy organic compounds
that may condense downstream. Any organic condensate that
collects in the cold trap must be measured to calculate the total
leak rate.
The flow rate through the system can be varied by throttling
the flow with a control valve immediately upstream of the vacuum
pump. Typical flow rates are approximately 60 liters per minute
(£/min) or less. A good flow rate to use is one in which a
balance can be found between reaching equilibrium conditions and
having a high enough concentration of organic compounds in the
bag outlet to accurately measure the concentration in the
laboratory. As the flow rate is decreased, the concentration of
organic compounds increases in the gas flowing through the
sampling system. The flow rate should be adjusted to avoid any
operations with an explosive mixture of organic compounds in air.
It may also be possible to increase the flow rate in order to
minimize liquid condensation in the bag.
The flow rate should be set to a constant rate and kept at
that rate long enough for the system to reach equilibrium. To
4-6
-------
determine if equilibrium conditions have been reached, a portable
monitoring device can be used to indicate if the outlet
concentration has stabilized.
It is not recommended that the vacuum method be used to
measure the leak rate from equipment that have low screening
values (approximately 10 ppmv or less), because considerable
error can be introduced due to the background organic
concentration in the ambient air that is pulled through the bag.
In summary, the vacuum sampling procedure consists of the
following steps.
(1) Determine the composition of material in the designated
equipment component, and the operating conditions of
the component.
(2) Obtain and record a screening value with the portable
monitoring instrument.
(3) Cut a bag from appropriate material (see Section 4.3)
that will easily fit over the equipment component.
(4) Connect the bag to the sampling train.
(5) If a cold trap is used, immerse the trap in an ice
bath.
(6) Note the initial reading of the dry gas meter.
(7) Start the vacuum pump and a stopwatch simultaneously.
Make sure a vacuum exists within the bag.
(8) Record the temperature and pressure at the dry gas
meter.
(9) Observe the VOC concentration at the vacuum pump
exhaust with the monitoring instrument. Make sure
concentration stays below the lower explosive limit.
(10) Record the temperature, pressure, dry gas meter
reading, outlet VOC concentration and elapsed time
every 2 to 5 minutes (min).
(11) Collect 2 gas samples from the discharge of the
diaphragm sampling pump when the outlet concentration
stabilizes (i.e., the system is at equilibrium).
(12) Collect a background bag (optional).
4-7
-------
(13) Collect any liquid that accumulated in the bag as well
as in the cold trap (if used) in a sealed container.
(14) Take a final set of readings and stop the vacuum pump.
(15) Transport all samples to the laboratory, along with the
data sheet.
(16) Remove the bag.
(17) Rescreen the source with the portable monitoring
instrument and record.
Based on the data collected in the steps described above, mass
emissions are calculated using the equation presented in
Table 4-1.
4.2.2 Blow-Through Method
The sampling train for the blow-through method is presented
in Figure 4-2. The temperature and oxygen concentrations are
measured inside the bag with a thermocouple (or thermometer) and
an oxygen/combustible gas monitor. The carrier gas is metered
into the bag through one or two tubes (two tubes provide for
better mixing) at a steady rate throughout the sampling period.
The flow rate of the carrier gas is monitored in a gas rotameter
calibrated to the gas. Typical flow rates are approximately
60 £/min or less. It is preferable to use an inert gas such as
nitrogen for the blow-through method so as to minimize the risk
of creating an explosive environment inside the bag. Also, the
carrier gas should be free of any organic compounds and moisture.
The pressure in the bag should never exceed 1 pounds per square
inch gauge (psig).
The flow rate through the bag can be varied by adjusting the
carrier gas regulator. As mentioned in Section 4.2.1, a good
flow rate to use is one in which a balance can be found between
reaching equilibrium conditions and having a high enough
concentration of organic compounds in the bag outlet to
accurately measure the concentration in the laboratory.
Adjustments to the flow rate may also help minimize liquid
condensation in the bag. Any liquid that does accumulate in the
4-8
-------
TABLE 4-1. CALCULATION PROCEDURES FOR LEAK RATE WHEN USING THE
VACUUM METHOD
Leak Rate
(kg/hr)
where:
Q
MWa
T
GCb
9.63 X 10~10
P
VL
t
16.67
9.63 X IP"10 (Q)(MW)(GC)(P) + (p)(VL)
T + 273.15 16.67(t)
Flow rate out of bag (£/min);
Molecular weight of organic compound(s) in
the sample bagc or alternatively in the
process stream contained within the equipment
piece being bagged (kg-mol);
Temperature at the dry gas meter (°C);
Sample bag organic compound concentration
(ppmv) minus background bag organic compound
concentration0 (ppmv);
Absolute pressure at the dry gas meter
(mmHg);
A conversion factor using the gas constant
(°K * 106 * kg-mol * min)/(£ * hour * mmHg);
Density of organic liquid collected (g/m£);
Volume of liquid collected (m£);
Time in which liquid is collected (min); and
A conversion factor to adjust term to units
of kilograms per hour (g * hr)/(kg * min)
aFor mixtures calculate MW as:
n
S
n
2
where :
n =
Molecular weight of organic compound i ;
Mole fraction of organic compound i; and
Number of organic compounds in mixture.
mixtures, the value of GC is the total concentration of all
the organic compounds in the mixture.
cCollection of a background bag is optional. If a background bag
is not collected, assume the background concentration is zero.
4-9
-------
0)
I
Q)
•P
w
.c
&>
o
T3
0)
^
•H
3
4J
0)
I
•H
W
CM
I
(U
»-
-------
bag should be collected using the approach described in
Section 4.2.
The carrier gas flow rate should be set to a constant rate
and kept at that rate long enough for the system to reach
equilibrium. In addition to the carrier gas flow through the
bag, some ambient air may enter the bag if it is not airtight.
The oxygen measurements are used to determine the flow of ambient
air through the bag. The oxygen measurements are also an
indication of the quality of the bagging procedure (the lower the
oxygen concentration the better). Once oxygen concentration
falls below 5 percent, the portable monitoring instrument is used
to check organic compound concentrations at several locations
within the bag to ensure that the bag contents are at steady
state.
Once the bag contents are at steady state, two gas samples
are drawn out of the bag for laboratory analysis using a portable
sampling pump. It may also be necessary to collect a background
bag sample, particularly if the source had screened at zero and
if there is still a detectable level of oxygen in the bag.
However, collection of a background bag is optional.
In summary, the blow-through method consists of the
following steps, which assume nitrogen is used as the carrier
gas.
(1) Determine the composition of the material in the
designated equipment component, and the operating
conditions of the component.
(2) Screen the component using the portable monitoring
instrument.
(3) Cut a bag that will easily fit over the equipment
component.
(4) Connect tubing from the nearest nitrogen source to a
rotameter stand.
(5) Run tubing from the rotameter outlet to a "Y" that
splits the nitrogen flow into two pieces of tubing and
insert the tubes into openings located on either side
of the bag.
4-11
-------
(6) Turn on the nitrogen flow and regulate it at the
rotameter to a constant rate and record the time.
(7) After the nitrogen is flowing, wrap aluminum foil
around those parts of the component where air could
enter the bag-enclosed volume.
(8) Use duct tape, wire, and/or rope to secure the bag to
the component.
(9) Put a third hole in the bag roughly equidistant from
the two carrier gas-fed holes.
(10) Measure the oxygen concentration in the bag by
inserting the lead from an oxygen meter into the third
hole. Adjust the bag (i.e., modify the seals at
potential leak points) until the oxygen concentration
is less than 5 percent.
(11) Measure the temperature in the bag.
(12) Check the organic compound concentration at several
points in the bag with the portable monitoring
instrument to ensure that carrier gas and VOC are well
mixed throughout the bag.
(13) Collect samples in sample bags or canisters by drawing
a sample out of the bag with a portable sampling pump.
(14) Collect a background bag (optional).
(15) Remove the bag and collect any liquid that accumulated
in the bag in a sealed container. Note the time over
which the liquid accumulated.
(16) Rescreen the source.
Table 4-2 gives equations used to calculate mass emission rates
when using the blow-through method.
4.3 SOURCE ENCLOSURE
In this section, choosing a bagging material and the
approach for bagging specific equipment types are discussed. An
important criteria when choosing the bagging material is that it
is impermeable to the specific compounds being emitted from the
equipment piece. This criteria is also applicable for sample gas
bags that are used to transport samples to the laboratory. A bag
stability test over time similar to the Flexible Bag Procedure
described in Section 5.3.2 of EPA Method 18 is one way to check
4-12
-------
TABLE 4-2. CALCULATION PROCEDURES FOR LEAK RATE
WHEN USING THE BLOW-THROUGH METHOD
Leak Rate = 1.219 x 10"5 (Q) (MW) (GC) + (p) (VL)
(kg/hr) T + 273.15 16.67(t)
where :
Q = flow rate out of tent (m3/hr) ;
= _ N2 Flow Rate (£/min) _ x [0.06 (m3/min) ]
1 - [Tent Oxygen Cone, (volume %)/21]
MWa = Molecular weight of organic compounds in the
sample bag or alternatively in the process
stream contained within the equipment piece
being bagged (kg/kg-mol) ;
T ~ Temperature in tent (°C);
GC*3 = Sample bag organic compound concentration
(ppmv) , corrected for background bag organic
compound concentration (ppmv) ;c and
1.219 x 10~5 = A conversion factor taking into account the
gas constant and assuming a pressure in the
tent of 1 atmosphere (°K * 106 * kg-mol/m3)
p = Density of organic liquid collected (g/m£) ;
VL = Volume of liquid collected (m£) ;
t = Time in which liquid is collected (urn); and
16.67 = A conversion factor to adjust term to units
of Kilograms per hour (g * hr)/(kg * min) .
aFor mixtures calculate MW as:
n n
X MWi Xi / S Xi
i=l i=l
where :
MWi = Molecular weight of organic compound i;
Xi = Mole fraction of organic compound i; and
n = Number of organic compounds in mixture.
bFor mixtures, the value of GC is the total concentration of all
the organic compounds in the mixture.
GCollection of a background bag is optional. If a background bag
is not collected, assume the background concentration is zero.
To correct for background concentration, use the following
equation:
GC = (Sample Bag Cone, [ppmv]) -
(Tent Oxygen Cone, [volume %])/21 *
(Background Bag Cone, [ppmv])
4-13
-------
the suitability of a bagging material.1 After a bag has been
used, it must be purged. Bags containing residual organic
compounds that cannot be purged should be discarded. Mylar*,
Tedlar®, Teflon®, aluminum foil, or aluminized Mylar* are
recommended potential bagging materials. The thickness of the
bagging material can range from 1.5 to 15 millimeters (mm),
depending on the bagging configuration needed for the type of
equipment being bagged, and the bagging material. Bag
construction for individual sources is discussed in
Sections 4.3.1 through 4.3.5. For convenience, Mylar* will be
used as an example of bagging material in the following
discussions.
4.3.1 Valves
When a valve is bagged, only the leak points on the valve
should be enclosed. Do not enclose surrounding flanges. The
most important property of the valve that affects the type of
enclosure selected for use is the metal skin temperature where
the bag will be sealed. At skin temperatures of approximately
200°C or less, the valve stem and/or stem support can be wrapped
with 1.5- to 2.0-mm Mylar* and sealed with duct tape at each end
and at the seam. The Mylar* bag must be constructed to enclose
the valve stem seal and the packing gland seal.
When skin temperatures are in excess of 200°C, a different
method of bagging the valve should be utilized. Metal bands,
wires, or foil can be wrapped around all hot points that would be
in contact with the Mylar® bag material. Seals are then made
against the insulation using duct tape or adjustable metal bands
of stainless steel. At extremely high temperatures, metal foil
can be used as the bagging material and metal bands used to form
seals. At points where the shape of the equipment prevent a
satisfactory seal with metal bands, the foil can be crimped to
make a seal.
4.3.2 Pumps and Agitators
As with valves, a property of concern when preparing to
sample a pump or agitator is the metal skin temperature at areas
4-14
-------
or points that are in contact with the bag material. At skin
temperatures below 200"C, Mylar® plastic and duct tape are
satisfactory materials for constructing a bag around a pump or
agitator seal. If the temperature is too high or the potential
points of contact are too numerous to insulate, an enclosure made
of aluminum foil can be constructed. This enclosure is sealed
around the pump and bearing housing using silicone fabric
insulting tape, adjustable metal bands, or wire.
The configuration of the bag will depend upon the type of
pump. Most centrifugal pumps have a housing or support that
connects the pump drive (or bearing housing) to the pump itself.
The support normally encloses about one-half of the area between
the pump and drive motor, leaving open areas on the sides. The
pump can be bagged by cutting panels to fit these open areas.
These panels can be made using thicker bagging material such as
14-mm Mylar®. In cases where supports are absent or quite
narrow, a cylindrical enclosure around the seal can be made so
that it extends from the pump housing to the motor or bearing
support. As with the panels, this enclosure should be made with
thicker bagging material to provide strength and rigidity.
Reciprocating pumps can present a somewhat more difficult
bagging problem. If supports are present, the same type of two-
panel Mylar® bag can be constructed as that for centrifugal
pumps. In many instances, however, sufficiently large supports
are not provided, or the distance between pump and driver is
relatively long. In these cases, a cylindrical enclosure as
discussed above can be constructed. If it is impractical to
extend the enclosure all the way from the pump seal to the pump
driver, a seal can be made around the reciprocating shaft. This
can usually be best completed by using heavy aluminum foil and
crimping it to fit closely around the shaft. The foil is
attached to the Mylar* plastic of the enclosure and sealed with
the duct tape.
In cases where liquid is leaking from a pump, the outlet
from the bag to the sampling train should be placed at the top of
the bag and as far away from spraying leaks as practical. A low
4-15
-------
point should be formed in the bag to collect the liquid so that
the volume of the liquid can be measured and converted to a mass
rate.
4.3.3 Compressors
In general, the same types of bags that are suitable for
pumps can be directly applied to compressors. However, in some
cases, compressor seals are enclosed and vented to the atmosphere
at a high-point vent. If the seals are vented to a high-point
vent, this vent line can be sampled. A Mylar* bag can be
constructed and sealed around the outlet of the vent and
connected to the sampling train. If the high-point vents are
inaccessible, the vent lines from the compressor seal enclosures
can be disconnected at some convenient point between the
compressor and the normal vent exit. Sampling is then conducted
at this intermediate point. In other cases, enclosed compressor
seals are vented by means of induced draft blowers or fans. In
these cases, if the air flow rate is know or can be determined,
the outlet from the blower/fan can be sampled to determine the
emission rate.
4.3.4 Connectors
In most cases, the physical configurations of connectors
lend themselves well to the determination of leak rates. The
same technique can be used for a connector whether it is a
flanged or a threaded fitting. To bag a connector with a skin
temperature below 200°C, a narrow section of Mylar® film is
constructed to span the distance between the two flange faces or
the threaded fitting of the leaking source. The Mylar® is
attached and sealed with duct tape. When testing connectors with
skin temperatures above 200°C, the outside perimeter of both
sides of the connector are wrapped with heat-resistant insulating
tape. Then, a narrow strip of aluminum foil can be used to span
the distance between the connection. This narrow strip of foil
can be sealed against the insulating tape using adjustable bands
of stainless steel.
4-16
-------
4.3.5 Relief Valves
Relief devices in gas/vapor service generally relieve to the
atmosphere through a large-diameter pipe that is normally located
at a high point on the process unit that it serves. The "horns"
can be easily bagged by placing a Mylar* plastic bag over the
opening and sealing it to the horn with duct tape. Because may
of these devices are above grade level, accessibility to the
sampling train may be limited or prevented. It is sometimes
possible to run a long piece of tubing from the outlet connection
on the bag to the sampling train located at grade level or on a
stable platform.
As discussed previously in Section 3.0, the purpose of
pressure relief devices makes them inherently dangerous to
sample, especially over a long period of time. If these
equipment are to be sampled for mass emissions, special care and
precautions should be taken to ensure the safety of the personnel
conducting the field sampling.
4.4 ANALYTICAL TECHNIQUES
The techniques used in the laboratory analysis of the bagged
samples will depend on the type of processes sampled. The
following sections describe the analytical instrumentation and
calibration, and analytical techniques for condensate. These are
guidelines and are not meant to be a detailed protocol for the
laboratory personnel. Laboratory personnel should be well-versed
in the analysis of organic compound mixtures and should design
their specific analyses to the samples being examined.
Also discussed is the calibration protocol for the portable
monitoring instrument. When bagging data are collected, it is
critical that the screening value associated with mass emission
rates is accurate. For this reason, a more rigorous calibration
of the portable monitoring instrument is required than if only
screening data are being collected.
4.4.1 Analytical Instrumentation
The use of analytical instrumentation in a laboratory is
critical to accurately estimate mass emissions. The analytical
instrument of choice depends on the type of sample being
4-17
-------
processed. Gas chromatographs (GC's) equipped with a flame
ionization detector or electron capture detector are commonly
used to identify individual constituents of a sample. Other
considerations besides instrument choice are the type of column
used, and the need for temperature programming to separate
individual constituents in the process stream with sufficient
resolution. For some process streams, total hydrocarbon analyses
may be satisfactory.
4.4.2 Calibration of Analytical Instruments
Gas chromatographs should be calibrated with either gas
standards generated from calibrated permeation tubes containing
individual VOC components, or bottled standards of common gases.
Standards must be in the range of the concentrations to be
measured. If cylinder calibration gas mixtures are used, they
must be analyzed and certified by the manufacturer to be within
±2 percent accuracy, and a shelf life must be specified.
Cylinder standards beyond the shelf life must either be
reanalyzed or replaced.
Field experience indicates that certified accuracies of
±2 percent are difficult to obtain for very low-parts per million
(ppm) calibration standards (<10 ppm). Users of low-parts per
million calibration standards should strive to obtain calibration
standards that are as accurate as possible. The accuracy must be
documented for each concentration standard.
The results of all calibrations should be recorded on
prepared data sheets. Table 4-3 provides an example of a data
collection form for calibrating a GC. If other analytical
instruments are used to detect the organic compounds from liquid
samples, they should be calibrated according to standard
calibration procedures for the instrument.
4.4.3 Analytical Techniques for Condensate
Any condensate collected should be brought to the laboratory
sealed in the cold trap flask. This material is transferred to a
graduated cylinder to measure the volume collected. If there is
enough volume to make it feasible, the organic layer should be
separated from the aqueous layer (if present) and weighed to
4-18
-------
TABLE 4-3. EXAMPLE GC CALIBRATION DATA SHEET
Plant ID
Instrument ID
Analyst Name
Certified Instrument
Gas Cone. Reading
Date Time (ppmv) (ppmv) Comments
4-19
-------
determine its density. If water-miscible organic compounds are
present, both the aqueous and organic phases should be analyzed
by GC to determine the total volume of organic material.
4.4.4. Calibration Procedures for the Portable Monitoring
Instrument
To generate precise screening values, a rigorous calibration
of the portable monitoring instrument is necessary. Calibrations
must be performed at the start and end of each working day, and
the instrument reading must be within 10 percent of each of the
calibration gas concentrations. A minimum of five calibration
gas standards must be prepared including a zero gas standard, a
standard approaching the maximum readout of the screening
instrument, and three standards between these values. If the
monitoring instrument range is from 0 to 10,000 ppmv, the
following calibration gases are required:
• A zero gas (0-0.2 ppm) organic in air standard;
• A 9.0 ppm (8-10 ppm) organic in air standard;
• A 90 ppm (80-100 ppm) organic in air standard;
• A 900 ppm (800-1,000 ppm) organic in air standard; and
• A 9,000 ppm (8,000-10,000 ppm) organic in air standard.
The same guidelines for the analysis and certification of the
calibration gases as described for calibrating laboratory
analytical instruments must be followed for calibrating the
portable monitoring instrument.
4.5 QUALITY CONTROL AND QUALITY ASSURANCE GUIDELINES
To ensure that the data collected during the bagging program
is of the highest quality, the following QC/QA procedures must be
followed. Quality control requirements include procedures to be
followed when performing equipment leak mass emissions sampling.
Quality assurance requirements include accuracy checks of the
instrumentation used to perform mass emissions sampling. Each of
these QC/QA requirements are discussed below.
4.5.1 Quality Control Procedures
A standard data collection form must be prepared and used
when collecting data in the field. Tables 4-4 and 4-5 are
4-20
-------
TABLE 4-4. EXAMPLE DATA COLLECTION FORM FOR FUGITIVE EMISSIONS
BAGGING TEST (BLOW-THROUGH METHOD)
Equipment Type Component ID
Equipment Category Plant ID
Line Size Date
Stream Phase (G/V, LL, HL) Analysis Team
Barometric Pressure
Ambient Temperature Instrument ID
Stream Temperature Stream Pressure
Stream Composition (Wt %) ,
Time Bagging Test Measurement Data
Initial Screening (ppmv) Equipment Piece Bkgd.
Background Bag Organic Compound Cone. (ppmv)a
Dilution Gas Flow Rate (£/min)
Sample Bag 1 Organic Compound Cone, (ppmv)
02 Concentration (volume %)
Bag Temperature (°C)
Dilution Gas Flow Rate (£/min)
Sample Bag 2 Organic Compound Cone, (ppmv)
02 Concentration (volume %)
Bag Temperature (°C)
Condensate Accumulation: Starting Time Final Time
Organic Condensate Collected (m£)
Density of Organic Condensate (g/m£)
Final Screening (ppmv) Equipment Piece Bkgd.
aCollection of a background bag is optional. However, it is
recommended in cases where the screening value is less than
10 ppmv and there is a detectable oxygen level in the bag.
4-21
-------
TABLE 4-5. EXAMPLE DATA COLLECTION FORM FOR FUGITIVE
EMISSIONS BAGGING TEST (VACUUM METHOD)
Equipment Type Component ID
Equipment Category Plant ID
Line Size Date
Stream Phase (G/V, LL, HL) Analysis Team
Barometric Pressure
Ambient Temperature Instrument ID
Stream Temperature Stream Pressure
Stream Composition (Wt %) ,
Time Bagging Test Measurement Data
Initial Screening (ppmv) Equipment Piecea Bkgd.
Background Bag Organic Compound Cone. (ppmv)b
Dry Gas Meter Reading (£/min)
Sample Bag 1 Organic Compound Cone, (ppmv)
Vacuum Check in Bag (Y/N) (Must be YES to collect sample.)
Dry Gas Meter Temperaturec (°C)
Dry Gas Meter Pressure0 (mmHg)
Dry Gas Meter Reading (£/min)
Sample Bag 2 Organic Compound Cone, (ppmv)
Vacuum Check in Bag (Y/N) (Must be YES to collect sample.)
Dry Gas Meter Temperature0 (°C)
Dry Gas Meter Pressure0 (mmHg)
Condensate Accumulation: Starting Time Final Time
Organic Condensate Collected (m£)
Density of Organic Condensate (g/m£)
Final Screening (ppmv) Equip. Piece3 Bkgd.
aThe vacuum method is not recommended if the screening value is
approximately 10 ppmv or less.
^Collection of a background bag is optional.
°Pressure and temperature are measured at the dry gas meter.
4-22
-------
examples of data collection forms for the blow-through and vacuum
methods of mass emissions sampling, respectively.
In addition to completing the data collection forms, the
following guidelines need to be adhered to when performing the
bagging analysis:
• Background levels near equipment that is selected for
bagging must not exceed 10 ppmv, as measured with the
portable monitoring device.
• Screening values for equipment that is selected for
bagging must be readable within the spanned range of
the monitoring instruments. If a screening value
exceeds the highest reading on the meter (i.e., "pegged
reading"), a dilution probe should be used, or the
reading should be identified as pegged.
• Only one piece of equipment can be enclosed per bag; a
separate bag must be constructed for each equipment
component.
• A separate sample bag must be used for each equipment
component that is bagged. Alternatively, bags should
be purged and checked for contamination prior to reuse.
• A GC must be used to measure the concentrations from
gas samples.
• Gas chromatography analyses of bagged samples must
follow the analytical procedures outlined in EPA
Method 18.
• To ensure adequate mixing within the bag when using the
blow-through method, the dilution gas must be directed
onto the equipment leak interface.
• To ensure that steady-state conditions exist within the
bag, wait at least five time constants (volume of bag
dilution/gas flow rate) before withdrawing a sample for
recording the analysis.
• The carrier gas used in the blow-through method of
bagging should be analyzed by GC before it is used, and
the concentration of organic compounds in the sample
should be documented. For cylinder purge gases, one
gas sample should be analyzed. For plant purge gas
systems, gas samples should be analyzed with each
bagged sample unless plant personnel can demonstrate
that the plant gas remains stable enough over time to
allow a one-time analysis.
4-23
-------
• The portable monitoring instrument calibration
procedure described in Section 4.4.4 should be
performed at the beginning and end of each day.
4.5.2 Quality Assurance Procedures
Accuracy checks on the laboratory instrumentation and
portable monitoring device must be performed to ensure data
quality. These checks include a leak rate check performed in the
laboratory, blind standards to be analyzed by the laboratory
instrumentation, and drift checks on the portable monitoring
device.
4.5.2.1 Leak Rate Check
A leak rate check is normally performed in the laboratory by
sampling an artificially induced leak rate of a known gas. This
can clarify the magnitude of any bias in the combination of
sampling/test method, and defines the variance in emissions
estimation due to the sampling. If the result is outside the
80 to 120 percent recovery range, the problem must be
investigated and corrected before sampling continues. The
problems and associated solutions should be noted in the test
report.
Leak rate checks should be performed at least two times per
week during the program. The leak rate checks should be
conducted at two concentrations: (1) within the range of 10
multiplied by the calculated lower limit of detection for the
laboratory analytical instrument; and (2) within 20 percent of
the maximum concentration that has been or is expected to be
detected in the field during the bagging program.
To perform a leak rate check, first induce a known flow rate
with one of the known gas concentrations into a sampling bag.
For example, this can be done using a gas permeation tube of a
known organic compound constituent. Next, determine the
concentration of the gas using a laboratory analytical instrument
and compare the results to the known gas concentration.
If the calculated leak rate is not within ±20 percent of the
induced leak rate, further analysis should be performed to
determine the reason.
4-24
-------
Areas that can potentially induce accuracy problems include:
• Condensation,
• Pluggage,
• Seal of bag not tight (leakage),
• Adsorption onto bag, and
• Permeation of bag.
The results of all accuracy checks should be recorded on prepared
data sheets.
4.5.2.2 Blind Standards Preparation and Performance
Blind standards are analyzed by the laboratory
instrumentation to ensure that the instrument is properly
calibrated. Blind standards must be prepared and submitted at
least two times per week during the program. The blind standards
are prepared by diluting or mixing known gas concentrations in a
prescribed fashion so that the resulting concentrations are
known. The analytical results should be within ±25 percent of
the blind standard gas concentration. If the results are not
within 25 percent of the blind standard concentration, further
analyses must be performed to determine the reason. Use of blind
standards not only defines the analytical variance component and
analytical accuracy, but it can serve to point out equipment
malfunctions and/or operator error before questionable data are
generated.
4.5.2.3 Drift Checks
Drift checks need to be performed to ensure that the
portable monitoring instrument remains calibrated. At a minimum,
drift checks must be performed before and after a small group of
components (i.e., two or three) are bagged. Preferably, drift
checks should be performed on the screening instrument
immediately before and after each component is bagged. These
checks should be performed by analyzing one of the calibration
gases used to calibrate the portable monitoring instrument. The
choice of calibration gas concentration should reflect the
anticipated screening value of the next component to be
monitored. For example, if a component had previously screened
4-25
-------
at 1,000 ppm and been identified for bagging, the calibration
standard should be approximately 900 ppmv.
Drift check data must be recorded on data sheets containing
the information shown in the example in Table 4-6. If the
observed instrument reading is different from the certified value
by greater than ±20 percent, then a full multipoint calibration
must be performed (see Section 3.2.4.1). Also, all those
components analyzed since the last drift check must be retested.
Drift checks should also be performed if flameout of the
portable monitoring instrument occurs. Using the lowest
calibration gas standard (i.e., approximately 9 ppmv standard),
determine the associated response on the portable monitoring
instrument. If the response is not within ±10 percent of the
calibration gas concentration, a full multipoint calibration is
required before testing resumes.
4-26
-------
TABLE 4-6. EXAMPLE DRIFT TEST REPORT FORM
Plant ID
Instrument ID
Analyst Name
Standard Measured ID Number of
Gas Cone. Cone. % Component Bagged
Date (ppmv) Time (ppmv) Errora Since Last Test
a% Error = Certified Cone. - Measured Cone. * 100
Certified Cone.
4-27
-------
4.6 REFERENCES
1. Code of Federal Regulations, Title 40, Part 60, Appendix A.
Reference Method 21, Determination of Volatile Organic
Compound Leaks. Washington, DC. U.S. Government Printing
Office. Revised June 22, 1990.
4-28
-------
5.0 ESTIMATION OF CONTROL EFFICIENCIES FOR
EQUIPMENT LEAK CONTROL TECHNIQUES
5.1 INTRODUCTION
In this chapter, control techniques for reducing equipment
leak emissions are described. There are two primary techniques
for reducing equipment leak emissions: (1) modifying or
replacing existing equipment, and (2) implementing a leak
detection and repair (LDAR) program.
Modifying or replacing existing equipment is referred to in
this chapter as an "equipment modification." Examples of
equipment modifications include installing a cap on an open-ended
line, replacing an existing pump with a sealless type, and
installing on a compressor a closed-vent system that collects
potential leaks and routes them to a control device. In
Section 5.2, possible equipment modifications for each of the
equipment types are briefly described. Also, the estimated
control efficiency is presented for each equipment modification.
An LDAR program is a structured program to detect and repair
equipment that is identified as leaking. The focus of this
chapter is LDAR programs for which a portable monitoring device
is used to identify equipment leaks from individual pieces of
equipment. In Section 5.3, an approach is presented for
estimating the control effectiveness of an LDAR program.
5.2 EQUIPMENT MODIFICATION CONTROL EFFICIENCY
Controlling emissions by modifying existing equipment is
achieved by either installing additional equipment that
eliminates or reduces emissions, or replacing existing equipment
with sealless types. Equipment modifications for each equipment
type are described in the following sections. A separate section
is included on closed-vent systems, which can be installed on
more than one type of equipment. Equipment modifications that
can be used for each equipment type are summarized in Table 5-1.
Table 5-1 also contains an approximate control efficiency for
each modification.
5-1
-------
TABLE 5-1. SUMMARY OF EQUIPMENT MODIFICATIONS
Equipment type
Modification
Approximate
control
efficiency
Pumps
Compressors
Sealless design
Closed-vent system
Dual mechanical seal with
barrier fluid maintained at a
higher pressure than the pumped
fluid
Closed-vent system
Dual mechanical seal with
barrier fluid maintained at a
higher pressure than the
compressed gas
100a
90b
100
90b
100
Pressure relief
devices
Valves
Connectors
Open-ended
lines
Sampling
connections
Closed-vent system
Rupture disk assembly
Sealless design
Weld together
Blind, cap, plug, or second
valve
Closed-loop sampling
c
100
100a
100
100
100
aSealless equipment can be a large source of emissions in the
event of equipment failure.
^Actual efficiency of a closed-vent system depends on percentage
of vapors collected and efficiency of control device to which
the vapors are routed.
cControl efficiency of closed vent-systems installed on a
pressure relief device may be lower than other closed-vent
systems, because they must be designed to handle both
potentially large and small volumes of vapor.
5-2
-------
5.2.1 Closed-Vent Systems
A closed-vent system captures leaking vapors and routes them
to a control device. The control efficiency of a closed-vent
system depends on the percentage of leaking vapors that are
routed to the control device and the efficiency of the control
device. A closed-vent system can be installed on a single piece
of equipment or on a group of equipment pieces. For use on
single pieces of equipment, closed-vent systems are primarily
applicable to equipment types with higher potential emission
rates, such as pumps, compressors, and pressure relief devices.
5.2.2 Pumps
Equipment modifications that are control options for pumps
include routing leaking vapors to a closed-vent system
(as discussed in Section 5.2.1), installing a dual mechanical
seal containing a barrier fluid, or replacing the existing pump
with a sealless type.
5.2.2.1 Dual Mechanical Seals. A dual mechanical seal
contains two seals between which a barrier fluid is circulated.
Depending on the design of the dual mechanical seal, the barrier
fluid can be maintained at a pressure that is higher than the
pumped fluid or at a pressure that is lower than the pumped
fluid. If the barrier fluid is maintained at a higher pressure
than the pumped fluid, the pumped fluid will not leak to the
atmosphere. The control efficiency of a dual mechanical seal
with a barrier fluid at a higher pressure than the pumped fluid
is essentially 100 percent, assuming both the inner and outer
seal do not fail simultaneously.
If the barrier fluid is maintained at a lower pressure than
the pumped fluid, a leak in the inner seal would result in the
pumped fluid entering the barrier fluid. To prevent emissions of
the pumped fluid to the atmosphere, a barrier fluid reservoir
system should be used. At the reservoir, the pumped fluid can
vaporize (i.e., de-gas) and then be collected by a closed-vent
system.
The actual emissions reduction achievable through use of
dual mechanical seals depends on the frequency of seal failure.
5-3
-------
Failure of both the inner and outer seals could result in
relatively large releases of the process fluid. Pressure
monitoring of the barrier fluid may be used to detect failure of
the seals, allowing for a quick response to a failure.
5.2.2.2 Sealless Pumps. When operating properly, a
sealless pump will not leak because the process fluid cannot
escape to the atmosphere. Sealless pumps are used primarily in
processes where the pumped fluid is hazardous, highly toxic, or
very expensive, and where every effort must be made to prevent
all possible leakage of the fluid. Under proper operating
conditions, the control efficiency of sealless pumps is
essentially 100 percent; however, if a catastrophic failure of a
sealless pump occurs, there is a potential for a large quantity
of emissions.
5.2.3 Compressors
Emissions from compressors may be reduced by collecting and
controlling the emissions from the seal or by improving seal
performance. Shaft seals for compressors are of several
different types—all of which restrict but do not eliminate
leakage. In some cases, compressors can be equipped with ports
in the seal area to evacuate collected gases using a closed-vent
system. Additionally, for some compressor seal types, emissions
can be controlled by using a barrier fluid in a similar manner as
described for pumps.
5.2.4 Pressure Relief Valves
Equipment leaks from pressure relief valves (PRV's) occur as
a result of improper reseating of the valve after a release, or
if the process is operating too close to the set pressure of the
PRV and the PRV does not maintain its seal. Emissions occurring
from PRV's as a result of an overpressure discharge are not
considered to be equipment leak emissions. There are two primary
alternatives for controlling equipment leaks from pressure relief
devices: use of a rupture disk (RD) in conjunction with the PRV,
or use of a closed-vent system.
5.2.4.1 Rupture Disk/Pressure Relief Valve Combination.
Although they are also pressure relief devices, RD's can be
5-4
-------
installed upstream of a PRV to prevent fugitive emissions through
the PRV seat. Rupture disk/pressure relief valve combinations
require certain design constraints and criteria to avoid
potential safety hazards, which are not covered in this document.
If the RD fails, it must be replaced. The control efficiency of
the RD/PRV combination is assumed to be 100 percent when operated
and maintained properly.
5.2.4.2 Closed-Vent System. A closed-vent system can be
used to transport equipment leaks from a pressure relief device
to a control device such as a flare. The equipment leak control
efficiency for a closed-vent system installed on a pressure
relief device may not be as high as the control efficiency that
can be achieved by installing a closed-vent system on other
equipment types. This is because emissions from pressure relief
devices can be either high flow emissions during an overpressure
incident or low flow emissions associated with equipment leaks,
and it may be difficult to design a control device to efficiently
handle both high and low flow emissions.
5.2.5 Valves
Emissions from process valves can be eliminated if the valve
stem can be isolated from the process fluid. Two types of
sealless valves are available: diaphragm valves and sealed
bellows valves. The control efficiency of both diaphragm and
sealed bellows valves is virtually 100 percent. However, a
failure of these types of valves has the potential to cause
temporary emissions much larger than those from other types of
valves.
5.2.6 Connectors
In cases where connectors are not required for safety,
maintenance, process modification, or periodic equipment removal,
emissions can be eliminated by welding the connectors together.
5.2.7 Open-Ended Lines
Emissions from open-ended lines can be controlled by
properly installing a cap, plug, or second valve to the open end.
If a second valve is installed, the upstream valve should always
be closed first after use of the valves to prevent the trapping
5-5
-------
of fluids between the valves. The control efficiency of these
measures is assumed to be essentially 100 percent.
5.2.8 Sampling Connections
Emissions from sampling connections occur as a result of
purging the sampling line to obtain a representative sample of
the process fluid. Emissions from sampling connections can be
reduced by using a closed-loop sampling system or by collecting
the purged process fluid and transferring it to a control device
or back to the process. The closed-loop sampling system is
designed to return the purged fluid to the process at a point of
lower pressure. A throttle valve or other device is used to
induce the pressure drop across the sample loop. The efficiency
of a closed-loop system is assumed to be 100 percent.
Alternatively, in some cases, sampling connections can be
designed to collect samples without purging the line. If such a
sampling connection is installed and no emissions to the
atmosphere occur when a sample is collected, then the control
efficiency can be assumed to be 100 percent.
5.3 LEAK DETECTION AND REPAIR CONTROL EFFECTIVENESS
An LDAR program is designed to identify pieces of equipment
that are emitting sufficient amounts of material to warrant
reduction of the emissions through repair. These programs are
best applied to equipment types that can be repaired on-line,
resulting in immediate emissions reduction, and/or to equipment
types for which equipment modifications are not feasible. An
LDAR program is best suited to valves and pumps, and can also be
implemented for connectors.
For other equipment types, an LDAR program is not as
applicable. Compressors are repaired in a manner similar to
pumps; however, because compressors ordinarily do not have a
spare for bypass, a process unit shutdown may be required for
repair. Open-ended lines are most easily controlled by equipment
modifications. Emissions from sampling connections can only be
reduced by changing the method of collecting the sample (since by
definition equipment leak emissions are the material purged from
the line), and cannot be reduced by an LDAR program. Safety
5-6
-------
considerations may preclude the use of an LDAR program on
pressure relief valves.
In this section, an approach is presented that can be used
to estimate the control effectiveness of any given LDAR program
for light liquid pumps, gas valves, light liquid valves, and
connectors. The approach is based on the relationship between
the percentage of equipment pieces that are leaking and the
corresponding average leak rate for all of the equipment. In
this approach, the three most important factors in determining
the control effectiveness are: (1) how a "leak" is defined,
(2) the initial leak frequency before the LDAR program is
implemented, and (3) the final leak frequency after the LDAR
program is implemented. The leak definition (or action level) is
the screening value at which a "leak" is indicated if a piece of
equipment screens equal to or greater than that value. The leak
frequency is the fraction of equipment with screening values
equal to or greater than the leak definition.
Once these three factors are determined, a graph that plots
leak frequency versus mass emission rate at several different
leak definitions is used to predict emissions preceding and
subsequent to implementing the LDAR program. In this way the
emissions reduction (i.e., control effectiveness) associated with
the LDAR program can be easily calculated.
A general description of the approach is provided in the
subsections below. This is followed by an example application of
the approach. The approach has been applied to determine the
control effectiveness at Synthetic Organic Chemical Manufacturing
Industry (SOCMI) and refinery process units for the following
LDAR programs: (l) monthly LDAR with a leak definition of 10,000
parts per million by volume (ppmv), (2) quarterly LDAR with a
leak definition of 10,000 ppmv, and (3) LDAR equivalent to that
specified in the proposed hazardous organic National Emission
Standard for Hazardous Air Pollutants (NESHAP) equipment leaks
negotiated regulation.1 Tables 5-2 and 5-3 summarize the
estimated control effectiveness for the three LDAR programs
mentioned above at SOCMI process units and refineries,
5-7
-------
EH
H
fe
CO
CO
W
CJ
o
«
Oi
H
§
O
CO
<
H
<
5
8
a
a<
i
Q
r3
S
0
fn
CO
CO
w
2
H
>
H
^ .
EH
CJ
W
Cn
In
W
i-3
O
K
O
CJ
•
fM
1
in
W
»J
i
^^
#>
0)
10
0)
c
0)
>
.,-1
4J
O
0>
H
2
2
BC
&
C
•HX
r4 <0
O -H
0 B-P
6 ft-H
ft C
•i^^ .^
?^ *rl
rH 0
tJ>
C X
•H (0
M 0)
O H C
•P O
•H > *H
C g -P
O ft-H
g ft C
•H
>iO
•P -
C o
siH
^<
Equipment type and
service
oi co in
o> co t^
t^ H in
VO VO M-
t"» ^ O\
co oo vo
•0 ^,
Valves - gas
alves - light liqui
Pumps - light liquic
> ~
o
a\
A
A
Connectors - all
i
4J
§
CO ft
S -H
i!f
10
N C
(0 O
TJ
•C Q)
0) (0
to
o» 5»o>
r4 0) H
h
Q) 0)
jd -o o
4J 0 C
4J <0
O (0 8
4J-H C
•P 0
0) O H rH
4J 10
4J^J C
(0 C O
a) -H
W B -P
ontrol effectivenes
rganic NESHAP equip
pecific leak defini
CJ O W
•H
•P
U
(U
<(H
rH
ata are not availab
Q
A
5-8
-------
EH
CO
CO
W
1
«
PH
^
6
F*<
2
tf
4
s
j§
O
«
"
1
1— 1
"
2
<
Qi
o
"
CO
CO
H
>
I—I
C_|
r^
CJ
w
fa
IV.
M-l
TABLE 5-3. CONTROL
^_,
<*>
CO
CO
0)
M/
0)
w
•H
•P
O
0)
-H
0 B4J
"as
>i -H
rH O 1-1
V4 0 0)
0) 0 TJ
•P ••
M 0
jrt •
(0 rn
s
&
c *^
•H id
O H C
•P O
•H > *H
5 St!
O ft'H
e a c
•H
H 0 Q)
«Cj ^3 *O
1 \ ^
C 0
O rH
s
Equipment type and
service
t^H
vo in co
en o^ co
o H in
(N. VO t
CO VO CO
CO (^ VO
Valves - gas
Ives - light liquid
imps - light liquid
£> ^
H
CO
A
A
Connectors - all
i
Q)
8
CO ft
3 -H
•§&
rl 0)
id
N C
id o
X!
T3 Q)
0) CO
co id
O *Q
ft
o -o
V-l Q)
ft-P
0) 6
-P -P
CO
*H Q)
O
0)
10 J-l
•P id
c
0) C
6 0
CD-H •
rl -P CO
•H Id rH
3 H 0)
trd >
Q) &> 0)
M Q) -H
0) 0)
J3 -O 0
4J 0) C
•P tO
o id e
•P-H R
•P 0
0) O ^
42 0) «
id c ft
4J
0 J^ ^
A id c
•H Q) Id
rl rH
4-> CO
4J -P C
id c o
ntrol effectiveness
ganic NESHAP equipme
ecific leak definiti
O M ft
cj o co
id
CO
CO
-------
respectively. It should be noted that, to calculate the control
effectiveness values presented in Tables 5-2 and 5-3, assumptions
were made that may not necessarily be applicable to specific
process units. For example, the control effectiveness values in
the tables are based on the assumption that the emission rate
prior to implementing the LDAR program is the emission rate that
would be predicted by the average emission factor. The best way
to calculate the effectiveness of an LDAR program is by
collecting and analyzing data at the specific process unit.
5.3.1 Approach for Estimating LDAR Control Effectiveness
As previously stated, the key parameters for estimating the
control effectiveness of an LDAR program are the leak definition,
the initial leak frequency, and the final leak frequency. The
leak definition is a given part of an LDAR program. It can
either be defined by the process unit implementing the program or
by an equipment standard to which the process unit must comply.
After the leak definition is established, the control
effectiveness of an LDAR program can be estimated based on the
average leak rate before the LDAR program is implemented, and the
average leak rate after the program is in place.
Figures 5-1 through 5-4 are graphs presenting mass emission
rate versus leak frequency for SOCMI-type process units at
several leak definitions for gas valves, light liquid valves,
light liquid pumps, and connectors, respectively. Figures 5-5
through 5-8 are graphs presenting mass emission rate versus leak
frequency for refinery process units at several leak definitions
for gas valves, light liquid valves, light liquid pumps, and
connectors, respectively. Using these figures, for a given leak
definition, the leak rate before and after the LDAR program is
implemented, along with the corresponding control effectiveness,
can be determined by plotting the initial and final leak
frequency on these graphs. Tables 5-4 and 5-5 present equations
for the lines in each of the SOCMI and refinery graphs,
respectively. Appendix F describes the approach that was used to
develop the equations.
5-10
-------
SOCMI Gas Valve Equations
A
01
u
L
D
0
CO
\
L
£
0>
.y
\j
id
DC
_
id
0)
j
OJ
0>
(d
L
(II
Fraction Leaking
D 500 ppmv Lk. Def. + 1,000 ppmv Lk. Def. 0 2,000 ppmv Lk. Def.
A 5,000 ppmv Lk. Def. X 10,000 ppmv Lk. Def V SOCMI Avg. Factor
Figure 5-1.
SOCMI Gas Valve Average Leak Rate Versus Fraction
Leaking at Several Leak Definitions.
5-11
-------
SOCMI Light Liquid Valve Equations
u
L
3
0
10
X
L
r
VJ
(P
fi
d
j
0.006 -
0.004
0.002 -
0.2
0.02
0.06
0.14
0.18
Fraction Leaking
D 500 ppmv Lk. Def. 4- 1,000 ppmv Lk. Def. o 2,000 ppmv Lk. Def.
A 5,000 ppmv Lk. Def. x 10,000 ppmv Lk. Def V SOCMI Avg. Factor
Figure 5-2. Light Liquid Valve Average Mass Emission Rate Versus
Fraction Leaking at Several Leak Definitions
5-12
-------
SOCMI Light Liquid Pump Equations
0.08
n
fl)
o
L
3
0
ID
x
L
£
10
cr
id
j
(D
0.07 -
0.06 -
0,05 -
0,04 -
0,03 -
0,02
0,01 -
0.05 0.1 0.15 0.2 0.25 0.3
Fraction Leaking
D 500 ppmv Lk. Def. + 1,000 ppmv Lk. Def. 0 2,000 ppmv Lk. Def.
A 5,000 ppmv Lk. Def. X 10,000 ppmv Lk. Def V SOCMI Avg. Factor
Figure 5-3.
SOCMI Light Liquid Pump Average Leak Rate Versus
Fraction Leaking at Several Leak Definitions
5-13
-------
SOCMI Connector Equations
0)
u
L
0
W
x
L
£
x
0>
Jsi
\j
fll
1-1
id
cr
.
(d
GJ
J
OJ
D
id
L
01
0.012
0.011
0.01
0,009
0.008
0.007
0.006
0.005
0,004
0.003
0.002
0.001
0
0 0.02 0.04 0.06 0.08 0.1
0.01 0.03 0.05 0.07 0.09
Fraction Leaking
D 500 ppmv Lk. Def. f 1,000 ppmv Lk. Def. 0 2,000 ppiw Lk. Def.
A 5,000 ppmv Lk. Def. X 10,000 ppmv Lk. Def V SOCMI Avg. Factor
Figure 5-4. SOCMI Connector Average Leak Rate Versus Fraction
Leaking at Several Leak Definitions.
5-14
-------
Refinery Gas Valve Equations
u
L
3
0
U)
x
L
x
\j
(0
10
L
0)
0.1
0.2
0.3
0.5
Fraction Leaking
D 500 ppmv Lk. Def. 4- 1,000 ppmv Lk. Def. 0 10,000 ppiw Lk. Def
V Refinery Avg Factor
Figure 5-5. Refinery Gas Valve Average Leak Rate Versus Fraction
Leaking at Several Leak Definitions
5-15
-------
Refinery Light Liquid Valve Equations
0.05
u
L
3
0
(D
x
L
£
id
o:
id
fl>
j
0.04 -
0.03 -
0.02 -
0.01 -
0.1
0.2
0.3
0.4
0.5
Fraction Leaking
D 500 ppmv Lk. Def. 4- 1,000 ppmv Lk. Def. 0 10,000 ppmv Lk. Def
V Refinery Avg Factor
Figure 5-6.
Refinery Light Liquid Valve Average Leak Rate Versus
Fraction Leaking at Several Leak Definitions
5-16
-------
Refinery Light Liquid Pump Equations
n
(P
u
L
D
0
L
1C
ff
J
0
01
(0
L
0)
0.35 -
0.25 -
0,15 -
0.05 -
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Fraction Leaking
D 500 ppmv Lk. Def. + 1,000 ppmv Lk. Def. 0 10,000 ppmv Lk. Def
V Refinery Avg Factor
Figure 5-7.
Refinery Light Liquid Pump Average Leak Rate Versus
Fraction Leaking at Several Leak Definitions
5-17
-------
Refinery Connector Equations
u
L
D
0
(D
x
L
£
0>
X
\J
-------
W
EH
H
*
H
S
Q
0
(A
EH
<
O
Jg
^•4
2
3
§
6
w
g
8
&
1
H
O
2
w
^>
<
o
55
H
EH
g
CO
o
H
EH
<
&
CX
H
•«*
1
in
W
EH
10
e
o
•H
4J
id
a
cr
W
C
O
•H
4J
•H
c c
•^ s
it 1 K
(1) ft
'O
(0
id
O
00 CO
CM ^*
0 0
o o
o o
o o
o o
"*" +
o o
id id
fa fa
33
CM CM
00 VO
9» VO
"* in
o o
0 0
II II
a a
«. «;
o o
o o
o o
H CM
H H
CO CO
0 H
0 0
O O
O O
O O
O 0
id id
tji Qyi
3 3
CO O
9> H
r« co
vo r-
o o
o o
II II
1 1
«! ««
O O
o o
o o
in o
H
t^ o\
CM CO
0 0
0 0
o o
o o
o o
+
O 0
id id
fa fa
3 3
H in
CM CM
r* co
M- in
o o
o o
II II
a a
«« «;
o o
o o
in o
H
Q)
iH
5
>
j
•rt
•H
4J
•H
a\ H in
in H vo
O H H
o o o
o o o
o o o
o o o
+ + +
0 O O
n> id id
14 14 14
fa fa fa
333
in r* H
CM O O
H r^ o>
vo r» co
o o o
o o o
II II II
191
«!«»«;
o o o
o o o
o o o
CM in o
H
VO CO
o in
CO ^
o o
o o
o o
o o
+ +
o o
id id
h b
fa fa
3 3
00 H
O> CM
^« CO
O^ iH
O H
0 0
II II
CM CM
«; «.
o o
o o
in o
r-l
a
e
g
a
•o
•H
p
•H
H
1
•H
VO CO CO
vo o vo
vo ^r co
O H H
o o o
o o o
o o o
+ + +
0 0 O
id id id
li tj l»
fa fa fa
333
H in CM
f« «* CO
CO p* H
CO O^ ^*
H H CM
O O O
II II II
s a a
«.«;«.
o o o
o o o
o o o
CM in o
H
r» in in
H fM CO
000
o o o
o o o
o o o
o o o
+ + +
0 0 O
id id id
tj l« li
fa fa fa
333
*t r- r>
co in o
vo a\ co
«* in r»
o o o
o o o
II II II
a s |
«««;«;
o o o
o o o
in o o
H CM
j^
0
4J
O
c
o
u
't H
in co
0 0
o o
o o
o o
o o
+ +
0 0
id id
33
91 O
r-. vo
H CM
9> H
O H
0 0
II II
9 1
«. «i
O O
o o
o o
in o
H
.
^T
o
^
o
w
t.
Q)
^1
X!
^.
0)
id
M
y<
id
0)
i-H
0)
id
0)
a
id
5-19
-------
to
H
>4
PS
W
25
H
fa
rt
H
<
o
H
M
^C
py
H
u
P§
fa
o
H
1
^
J
*~*
u
to
^^
»
w
2«
^
2
H
W
«
to
§
EQUATI
*
in
i
in
§
E-i
id
c
o
•H
1
W
d
o
•H
-P
C *~*
.,_! g
til 6
fl) ^
•«3
id
3
o
o
in
Q>
r-H
id
•o
•H
W
•H
H
•P
-J
J
O O
oo o
CN r»
O i-l
0 O
o o
o o
+ +
O 0
(d (d
^ M
fib Cb
33
co o
•* in
CM CO
••* CO
0 0
0 0
II II
PS Pa
< <
0 0
0 0
o o
H O
H
O
CM
CO
H
O
O
o
+
0
id
^
fa
3
0t
(•»
in
01
H
o
II
3
**<
o
o
tn
D.
a
•O
•H
&
H
4J
X3
•H
J
O
CO
01
H
O
0
o
+
o
id
^
fa
3
i^
CO
CO
CO
CM
O
II
a
^
o
o
o
H
O
o
o
CM
H
O
O
+
6
id
M
fa
3
o
0
in
CM
^
d
II
a
*•
o
o
o
o
H
CO CO O
H H VO
000
O O O
O O O
O O O
o o o
+ + +
u o o
id id id
M M M
fa fa fa
333
in co «*
in CM <•
co r^ r^
H H CO
000
o o o'
II II II
a a a
4
(u
0)
*
Oj
[^
id
5-20
-------
Figure 5-9 provides guidance on how to determine the initial
and final leak frequencies. This figure is a simplified
graphical presentation on how the leak frequency will change
after an LDAR program is implemented. When generating the
figure, it was assumed that all equipment pieces are monitored at
the same time. Each occurrence of equipment monitoring is
referred to as a "monitoring cycle," and it is assumed that equal
time periods lapse between monitoring cycles.
From Figure 5-9, it can be seen that there is an immediate
reduction in leak frequency after the LDAR program is
implemented, and then the leak frequency will oscillate over
monitoring cycles. This oscillation occurs because between
monitoring cycles a certain percentage of previously non-leaking
equipment will begin to leak. There are four key points on the
graph presented in Figure 5-9. These key points are:
• Point X - initial leak frequency;
• Point Y - leak frequency immediately after monitoring
for and repairing leaking equipment (i.e., immediately
after a monitoring cy^le);
• Point Z - leak frequency immediately preceding a
monitoring cycle; and
• Point F - average leak frequency between monitoring
cycles (final leak frequency).
The initial leak frequency is the fraction of sources
defined as leaking before the LDAR program is implemented. The
initial leak frequency is Point X on Figure 5-9. The lower the
leak definition, the higher the initial leak frequency. At a
process unit, the initial leak frequency can be determined based
on collected screening data. If no screening data are available,
the initial leak frequency can be assumed to be equivalent to the
leak frequency associated with the applicable average emission
factor. However, if a process unit already has some type of LDAR
program in place, the average emission factor may overestimate
emissions.
5-21
-------
(D
(1)
c
0
0
a)
11
0.08
0.07
0.06
0.05
0.04
0,03
0.02
0,01
% Point X Clnitlal leak frequency)
Point F Cflnal leak frequency)
>
Monitoring Cycle
Figure 5-9.
Simplified graphical presentation of changes in leak
frequency after implementation of an LDAR program.
5-22
-------
On Figures 5-1 through 5-8, the average emission factor for
each equipment type is plotted as a horizontal line. From this
line, an initial leak frequency can be determined for any of the
leak definitions. For example, on Figure 5-1, which is for gas
valves, the SOCMI average emission factor equals 0.00597
kilograms per hour (kg/hr). For a leak definition of 500 ppmv,
this average emission factor corresponds to a fraction leaking of
approximately 0.136. Similarly, for a leak definition of 10,000
ppmv, the average emission factor corresponds to a fraction
leaking of 0.075. These points are determined by finding the
intersection of the SOCMI average emission factor line and the
applicable leak definition line and reading off the corresponding
fraction leaking. Alternatively the fraction leaking associated
with the average factor can be calculated using the equations in
Tables 5-4 and 5-5.
The leak frequency immediately after a monitoring cycle is
Point Y on Figure 5-9. After an LDAR program is implemented for
a given time period, Point Y will reach a "steady-state" value.
As presented in Figure 5-9, Point Y depends on two key factors:
(1) the percentage of equipment successfully repaired after being
identified as leaking, and (2) the percentage of equipment that
was repaired for which leaks recurred. Two simplifying
assumptions when calculating Point Y are: (1) that leaking
equipment is instantaneously repaired, and (2) that the recurring
leaks will occur instantaneously after the equipment is repaired.
Based on these assumptions the value for Point Y is calculated
using the following equation:
Yi = zi ~ (FR * zi) + (FR * zi * R)
where:
Yi = Leak fraction immediately after monitoring cycle i;
Zi = Leak fraction immediately preceding monitoring cycle
i (note that Z^ equals Point X.);
5-23
-------
R = Fraction of repaired sources for which a leak
immediately recurs; and
FR = Fraction of leaking sources successfully repaired.
Point Z on Figure 5-9 is the leak frequency immediately
preceding equipment monitoring. After an LDAR program is
implemented for a given time period, Point Z will reach a
"steady-state" value. To go from Point Y to Point Z on
Figure 5-9, the occurrence rate is added to Point Y. The
occurrence rate equals the percentage of initially nonleaking
equipment that starts to leak between monitoring cycles. Use the
following equation to go from Point Y to Point Z:
Zi+i = Oc * (1 - YI) + Yi
where :
= Leak fraction immediately preceding monitoring
cycle i + 1;
Oc = Fraction of nonleaking sources which will leak in
the time period between monitoring cycles
(i.e, occurrence rate) ; and
Yi = Leak fraction immediately after monitoring cycle i.
After several monitoring cycles, the leak frequency will be
found to approximately oscillate between Points Y and Z. The
average value of these two "steady-state" values is the final
leak frequency. This is Point F on Figure 5-9. The final leak
frequency is the average percent of sources that are still
leaking after an LDAR program has been implemented.
Once the initial and final leak frequencies are determined,
they can be entered into the applicable equation from Table 5-4
or Table 5-5 to calculate the associated average leak rates at
these leak frequencies. Based on the initial leak rate and the
final leak rate, the control effectiveness for an LDAR program
can be calculated. The control effectiveness is calculated as:
Eff = (ILR-FLRJ/ILR * 100
5-24
-------
where:
Eff = Control effectiveness (percent);
ILR = Initial leak rate (kg/hr/source); and
FLR = Final leak rate (kg/hr/source).
5.3.2 Example Application of Approach
As previously mentioned, the approach described in
Section 5.3.1 was applied to estimate the control effectiveness
for three types of LDAR programs: (1) monthly inspection with a
leak definition of 10,000 ppmv, (2) quarterly inspection with a
leak definition of 10,000 ppmv, and (3) a program complying with
the requirements specified in the proposed hazardous organic
NESHAP equipment leaks negotiated regulation.1 Details of these
calculations are presented in Appendix F. As an example of
applying the approach, the control effectiveness for gas valves
at a SOCMI process unit implementing a monthly LDAR program with
a leak definition of 10,000 ppmv is presented in the following
paragraphs.
Table 5-6 presents the SOCMI gas valve occurrence rate,
recurrence rate, unsuccessful repair rate, and initial leak
frequency. (See Appendix F for details on how each of these
parameters were determined.) Using the values presented in
Table 5-6 and the approach presented in Section 5.3.1, the LDAR
control effectiveness can be calculated. Note that Figure 5-9 is
also based on monthly monitoring of gas valves in a SOCMI process
unit with a leak definition of 10,000 ppmv, and it is referred to
in this example demonstration.
For gas valves with a leak definition of 10,000 ppmv, the
initial leak frequency is 7.5 percent. This initial leak
frequency value is taken from Figure 5-1, by finding the value of
the fraction leaking at the intersection of the SOCMI average
factor line and the 10,000-ppmv leak definition line. The
initial leak rate for this leak frequency is the SOCMI gas valve
average emission factor, which equals 0.00597. After the LDAR
program is implemented and monitoring occurs on a monthly basis,
5-25
-------
TABLE 5-6. VALUES USED IN EXAMPLE CALCULATION3
Source Category: SOCMI
Equipment Type: Gas Valves
LDAR Program: Monthly Monitoring with a Leak Definition
of 10,000 ppmv
Occurrence Rate: 1.00%
Recurrence Rate: 14%
Unsuccessful Repair Rate: 10%
Initial Leak Frequency;b 7.5%
aSee appendix F for information on how the occurrence rate,
recurrence rate, and unsuccessful repair rate were determined,
on the SOCMI average emission factor for gas valves.
5-26
-------
the steady-state leak frequency immediately after monitoring (see
Point Y6 on Figure 5-9) equals 0.29 percent. The steady-state
leak frequency prior to monitoring (see Point Zg on Figure 5-9)
equals 1.29 percent. This gives an average of 0.79 percent as
the final leak frequency (see Point F on Figure 5-9). The
calculations performed to determine the final leak frequency are
shown in Table 5-7. Once the estimated gas valve final leak
frequency is determined, the associated leak rate can be found
using Figure 5-1 or the gas valve equation for a leak definition
of 10,000 ppmv listed on Table 5-4. The corresponding leak rate
associated with the final leak frequency of 0.79 percent at a
leak definition of 10,000 ppmv is 0.00075 kg/hr. Thus, the
control effectiveness of a monthly LDAR program with a leak
definition of 10,000 ppmv for gas valves is:
= (0.00597-0.00075)/0.00597 * 100
=87 percent.
5-27
-------
>
55 2
W PQ
»Z) U
oo
W p£
prf Q.
Si
r^ S3
< H
2 Od
H 0
fa EH Id
pq 55 £
W O CM
EH £ CM
H X 0
S3 i-3 0
H a o
i5 P^ ^
W O H
w Jjj
Q 53
EH 5 H
3 > W
p o
CJ CO
Hi rtl *£
•< o 2
O w
, H (J
H £
rl CJ rfj
£l| Q
§ W *
W O EC
.
r^
I
in
W
J
1
10
LJ
PI
fli
^v
.p
0)
id
M
id
a
J_
1 1
*•*
^
a)
w
m
(U
«v
1 1
^»
0)
g
id
id
a
tn
Start in<
L(
Q)
4-)
4H
id
>i
U
B
0)
Q^ ••
0) *-»
L| \0
o e
definit
10,000 p
id II
3
,
H
rH
0)
4J
id
•O
Q) ..
§ *~*
S vo
•H N
>i-P
O B
E-H
0) O
5 CM
O'N-'
0)
Ll
B
4J id L|
C 0) O
0) H p
O -H
Lt 0) C
a> -P o
aid 6
o\ u o
OJ 1 4J
• ^^
O Tl LI
id o
4J L|
CO CX
.
M
99
"cT ^-s
o c<<
^J
0) C 0)
O Q) O
CUB
occurre
1.00 per
recurre
*•*
Q) Q)
. ,
oj n
..
XI
*-*
fa
43
C
•H
O
U
C
0)
4J 5 -P
C O* C
0) 0) 0)
U Li O
Ll «H L|
a) a>
Qifi^ Oi
id
eft Q) w
fM rH t^
• •
H rH O
id
II C II
•H
fa
.
n
..
^^
n>
fa
3 o)
(0
*f W O
H Q) ON
U
II U II
3
CO
,
•"41
(Q
,_j
•n
id
g
w
rt
^_^
X
^j
c
•H
o
•^
^J
o
c
Q)
&
0)
M 4^
(Lj c
0)
^ u
id i-i
0) 0)
rH &
rH in
id •
•H f>
4J
•H ||
C
H
,
in
0
4J -H
N
O -P
1 f+
•H C
• . .._!
W 'H,
ft 0
"
C_^
••
Q) d^
M
0) LI
Li O
-H
*§
0) g
T
M
Ii -r1
fll ^*
IJ
T* IJ
dj C
^_ o
r\ CM
C
3 "
Ei1}?
i *H
iTl IH
** O
t, -P
irt *^
j O
B
ft
o
^J
C o r» o vo o o>
0) in vo vo n n oj
L| r- O>
L| H O O O O O
o
^J
0)
rH
U
5**
? 0)
.q H M n ^J1 in vo
LI
0
•H
C
s
0)
rH
id
•H
4J
•H
•o
0)
•H
HH
•H
4J
C
0)
•o
•H
to
^»
B
•H
o
rH
rH
id
O
1
rH TJ
id id
B Q)
•H .p
fa W
X)
^
pj
-K
•H
N
* -H
X
g -H
**•*
+ ^H
X
*^H 1
H
^ ^^^
g *
^ O
o
II
•H
N H
II +
•H -H
X N
O TJ
.
0)
4J
id
1)
U)
41
id
(D
4J
CO
U
id
0)
LI
•H
N
•a
B
id
•H
L|
O
(0
0)
3
rH
id
0)
4J
V
0)
rH
0
u
tp
B
•H
Ll
O
V
•H
B
O
g
•H
(0
0)
•P
L|
0)
4J
<4H
rtj
0)
5-28
-------
5.4 REFERENCES
1. National Emission Standards for Hazardous Air Pollutants for
Source Categories; Organic Hazardous Air Pollutants from the
Synthetic Organic Chemical Manufacturing Industry and Seven
Other Processes. Subpart H—Equipment Leaks. Federal
Register. Vol. 57, No. 252, pp 62765-62785. Washington,
DC. Office of the Federal Register. December 31, 1992.
5-29
-------
APPENDIX A: EXAMPLE CALCULATIONS
-------
A-l. INTRODUCTION
(
This appendix provides example calculations demonstrating
the approaches described in Chapter 2.0. A simple dataset from a
hypothetical process unit is expanded as needed to illustrate how
the data are is used in each approach. Table A-l summarizes
information used in the example calculations. This information
includes the equipment count, hours of operation, and composition
for each stream. The stream compositions presented in Table A-l
are completely hypothetical and were chosen for the sole purpose
of demonstrating the various approaches. Three streams are
presented in Table A-l. Note that the hours of operation are
based on the time in which the equipment contains material.
(Even if a process unit is shutdown, if the equipment contains
material, then the shutdown time must still be included in the
hours of operation.)
Two SOCMI equipment type/service categories are used in the
example calculations: pumps/light liquid and valves/gas. The
same technique used for these equipment type/service categories
can be followed for any equipment type/service. In each of the
calculations, emissions are estimated on an annual basis.
The following sections present the example calculations. In
Section A-2, the Average Emission Factor Approach is presented.
Section A-3 presents the Screening Value Ranges Approach. In
Section A-4, the EPA Correlation Equation Approach is presented,
and in Section A-5, the use of the Unit-Specific Correlation is
discussed. Section A-6 explains how to speciate emissions.
Section A-7 demonstrates three approaches for applying response
factors (RF's). Section A-8 demonstrates how to annualize
emissions when more than one screening value is collected from
individual equipment pieces over an annual time period.
Section A-9 shows how to estimate VOC emissions when screening
data are collected from equipment containing organic compounds
not classified as VOC's. Finally, Section A-10 addresses
estimating emissions from equipment containing inorganic compounds.
A-l
-------
SNOUT
g
5
0
M
^
J
O<
£
s
W
O
fe
Q
iH
w c ^
n ^j ^*
0 Q) ^"*
S3 &*
o
H
3
O
U
•P
C
I
H
01
c fa
Hi ®
"3 's^
HI 0)
ftl Ql
^1
•P
Q
H
i
0)
14
-P
U
o o o o in in o
00 (N i-t O\ \O CN iH
• • • • • • •
o o o o o o o
0) 0) Q)
*» 4> -P
id id id
r-t iH fH U
>» >i >i 0
W W H PI
u u o id
« id 0) « >
C ID
iH M rH 0) r-l C t4
>lfl) >,V4 >1« 0>
£*J JS >i 4543+J
4J Id 4J +J -P -P «
O > 0) CD 0) 0) 3
00 0
r* to r»
00 *fr 00
in
-------
A-2. AVERAGE EMISSION FACTOR APPROACH
The Average Emission Factor Approach is demonstrated for
Streams A and B, which contain light liquid pumps. The SOCMI
average emission factor for light liquid pumps is 0.0199 kg/hr.
Based on this emission factor and data contained in Table A-l,
total VOC emissions can be calculated. Note that Stream A
contains water, which is not a VOC. This is accounted for when
total VOC emissions are estimated from Stream A. Table A-2
summarizes the Average Emission Factor Approach calculations.
A-3. SCREENING VALUE RANGES APPROACH
The Screening Value Ranges Approach is demonstrated for
Streams A and B. The calculations for the Screening Value Ranges
Approach are similar to those used for the Average Emission
Factor Approach, except that an emission factor for each
screening value range is used, and no adjustment is made for the
weight fraction VOC in the stream. An adjustment is not
necessary because the screening value is based solely upon VOC
leaking from the equipment piece. In this example, the component
screening values are designated as either less than 10,000 ppmv
or equal to or greater than 10,000 ppmv. It is assumed that none
of the light liquid pumps in Stream A have a screening value
greater than or equal to 10,000 ppmv, and one of the light liquid
pumps in Stream B screens greater than 10,000 ppmv. It is also
assumed that one of the pumps in Stream B could not be screened.
Emissions from this pump are calculated using the average
emission factor. Table A-3 summarizes the calculations used in
the Screening Value Ranges Approach.
A-4. EPA CORRELATION EQUATION APPROACH
The EPA Correlation Equation Approach is demonstrated for
Streams A and B. The EPA Correlation Equation Approach involves
entering screening values into a correlation equation to generate
an emission rate for each equipment piece. In Table A-4, assumed
screening values and the resulting emissions for each individual
equipment piece are presented. Emissions from the pump that was
not screened are estimated using the average emission factor.
A-3
-------
Q
O
ffi
£
£
OS
FACTO
2
O
H
W
CO
H
§
H
O
>
<
•
rg
<
W
iJ
03
<
tH
id
o
c
o ^
o n
S >:
i ^^
i?
8
>
C
0
•^
V
«
Ij
H
8.^
0 >
^^
... n
*H ^
OC
n
• .
H
3
0
S3
^
C
o
L VOC
fractii
I!
0)
d
I ,
i factoi
source)
ST:
•H s
0^
?^
B™
W~
jj
C
0) «P
o» 3
•H 0
M | »
&°
H
Q
H
n
0)
b
-P
(Q
o o o
a\ in 'f
O O rH
^ % «
CS rH <*»
O O
\O 00
t^ ro
% fe
00 •*
o o
00 O
• •
O rH
rH rH
O O
0 0
in CN
rH rH
0)
C
o
•H
o
a
•ri
< « E
H
rH
«
+J
o
EH
C
O
~4
*J
id
14
8.
0
H-l
O
a
t.
H
f4
4^
*
^
y
^
c
0
~4
fract
•
£
*
^
0
it
+J
U
«
-------
Q
O
K
EH
H
*H
CO
H
O
H
3
3
>
o
2
H
55
H
H
«
CJ
CO
•
n
rf
^S
M
,q
EH
a
c
o
•^ ^^
a vi
a >
••H ^
IS
r
"H c —
°5^
•^ ^
0 4^^
Vl rt>
r* t, *
3 W r
O fl)
as £'-
0
V
V
*— s
0)
Emission
factor
g/hr/sourc
.i
*••"
U
fl) 4J
& 3
-H O
&°
H
Q
H
E
Q)
Vl
4J
W
£,
id
>
a
s
o
o
o
^
o
1-1
A|
o>
•H
c
0)
0)
M
U
Q
m
4J
c
01
c
3
Q
u
O M) 04
vo *r co
O C4
iH
o o o
00 VO 00
n r* n
* S *.
•* 00 «*
-J- t^ r-
r 03 00
n t-t in
* o o
(N O O
• * •
o o o
rH 1" 0
rH tH
id
>
I
A
O
o
o
o
rH
V
o>
c
-H
c
0)
0)
H
u
0)
m
.p
c
« fl) < M
1
e
o
u
r* o
00 00
•*
w
rH
O
00
n
*
*•
o»
en
rH
O
•
O
f-l
o"
•
rH
0
*J
rH
id
&
0)
c
JQ 0
•O -^
0) -P
C U
o> id
0) Vl
Vt IH
u
0) •
4J 0)
•P 3 C
0 0
C 0 -H
O n
m > to
4J — -^
C E
o) n at
c
ex id
e -p
0 O
O H
*
IH
C
fi 5
O 4>
•H U
-P «
id vi
vi IH
& .
o -y
•H C-
0
*
Vl —
3 Vi
O O
£ 4J
— O
id
* IH
-~ C
Vi O
0 ~H
4-> CD
u m
id -i
IU P
§
C
O 0)
^ o>
m id
m vi
~4 0)
0) id
* *
— —.
o n
U H
0) 0)
c c
ft. ft.
E E
U U
VH IH
O O
•
• • **•«.
0 0 C
C CO
V
II II id
Vl
0) ffl Q)
C C ft
0 00
•H .rl
co co IH
m o o
•H -H
E Em
0) 0) Vi
3
O O 0
O O J5
* Z~
A-5
-------
TABLE A-4. EPA CORRELATION EQUATION METHODa
Equipment ID
Screening value
(ppmv)
VOC mass emissions13
(kg/yr)
Total
A-l
A-2
A-3
A-4
A-5
A-6
A-7
A-8
A- 9
A-10
A-ll
A-12
A-13
A-14
A-15
Stream A Emissions:
B-l
B-2
B-3
B-4
B-5
B-6
B-7
B-8
B-9
B-10
B-ll
B-12 (100% VOC) Not
Total
Total
Stream B Emissions:
Emissions
0
0
0
0
0
20
50
50
100
100
200
400
1,000
2,000
5,000
0
0
0
10
30
250
500
2,000
5,000
8,000
25,000
screened
0.066
0.066
0.066
0.066
0.066
2.0
4.2
4.2
7.4
7.4
13
23
49
87
190
390
0.033
0.033
0.033
0.55
1.4
7.9
14
44
93
140
350
87
740
1,130
aEquipment type: Light liquid pumps.
Correlation equation: Leak rate (kg/hr) = 1.90*10~5 * (SV)°-824
Default-zero mass emission rate: 7.49 * 10~6 kg/hr
Hours of operation: Stream A = 8,760; Stream B = 4,380.
Emissions = (correlation equation or default-zero emission
rate) * (hours of operation)
CVOC Emissions = (average emission factor) * (wt. fraction
of VOC) * (hours of operation)
A-6
-------
A-5. UNIT-SPECIFIC CORRELATION APPROACH
Correlation equations may be developed for specific units
rather than using the more general EPA Correlation Equations.
Appendix B presents details on developing unit-specific
correlations. Once correlations are developed using the approach
outlined in Appendix B, they are applied in the same manner as
described for the EPA correlations.
A-6. SPECIATING EMISSIONS
The emission rate of specific compounds in a mixture can be
calculated if the concentration of the compound in the stream is
known. The equation for speciating emissions is
CHEMX = VOCi * CONCX/CONCVOC
where :
CHEMX = The mass emissions of VOC "x" from the
equipment piece (mass/time) ;
VOCi = Tne voc mass emissions from the individual
equipment piece (mass/time) ;
CONCX = The concentration of VOC "x" in the equipment
piece (wt. fraction) ;
CONCyoc = Tne total VOC concentration in the equipment
piece (wt. fraction) .
See Table A-5 for a demonstration of speciating emissions of
Stream B. Because all of the equipment in Stream B contains the
same composition, the emissions can be speciated on a stream-wide
basis.
A-7. RESPONSE FACTORS
Response factors are used to correct screening values to
compensate for variations in a monitor's response to different
compounds. Determination of whether an adjustment to the
screening value will provide more valid emission estimates can be
made by reviewing RF's at actual concentrations of 500 ppmv and
10,000 ppmv for the material in the equipment being screened.
The RF's can be taken from Table C-l in Appendix C, or may
be calculated based on analytical measurement performed in a
laboratory. For materials with RF's below three at both actual
A-7
-------
id
£
s
K
EH
CO
fa
0
CO
§
H
CO
CO
C
S
H
EH
H
U
W
O,
CO
in
i
^»
H
(Q
E2
jj
a
o
m _
* ^
C r*1
•"o?
41 "„
C
0)
(0
A
D
C
0
D
a
I-
•P-JJ.
»""t 44
J>t ^^
M
u
id
rH
^
A
4J
H
C
O
(0
(0
JvT
81
id
4J
o
H
C
0
jj
a)
rH
U
rH
id
u
«. vo
•"*
in «•> _,
S rH -
0 T3
O O 01
in n o
O CN ^
rH rH
M C
0 O
U i
id m id
"" £. &
^\ ^^
C C 0<
0 id
••H M C
m o
(Q Qt >r^
•rl C 4J
§-rl Cd
C rH
01 01
• 0) V4
O> M M
> 0 O
rt w o
c"
o
4J
u
•W
•
n
.
at
>
id
0
•P
g
rH
•o
•H
^H
•rl
*w
o
c
o
•rl
4J
U
id
h
^
^(
m
1
(0
CO
1
CD
c o
ft rH
•ad id
•rl 4J
3 C O
cro o -p
rH H 4J «T>
U • II
+J C «JO
J3 O V4 CO
O»-H M-I c a)
•rl 4J 0 -rl
r) 0 • -H O
,+j at 0 < <
4) 3 rH • ,M Q) 0)
4J U h ? O rH rH
C O U C JO JO
0) > n) 0) O id id
g C -rl H H
UlrH rH 01 CO
§J3
Q
&1 0 -P -P g rl M
H H H W H fe h
id £> O -0
1
0)
rH
•8
H
g
0
0)
A-8
-------
concentrations, the screening value does not need to be
corrected. If the RF at either concentration is above three, the
screening value obtained from the monitoring device should be
adjusted.
If it is necessary to adjust the screening value, one of two
approaches can be applied:
(1) Use the higher of either the 500 ppmv or 10,000 ppmv RF
to adjust all screening values, or
(2) Plot the RF versus screening value and determine the
applicable RF for each screening value.
Table C-l in Appendix C presents the RF's for chemical
compounds at actual concentrations of 500 ppmv and 10,000 ppmv
for several different monitoring devices. For the example
calculations presented here, data for the Foxboro OVA-108 is
utilized. Table A-6 presents the RF's for ethyl aerylate and
styrene. From Table A-6, it can be seen that at both
concentrations, the RF for ethyl acrylate is below three.
Therefore, it is not necessary to adjust any of the screening
values taken from the equipment in Stream A. (The only VOC
constituent in Stream A is ethyl acrylate.) Stream B contains
10 percent ethyl acrylate and 90 percent styrene. The RF's at
both concentration values for Stream B are calculated using the
following equation:
RFm
where:
RFm = Response factor of the mixture;
n = Number of constituents in the mixture;
Xi = Mole fraction of constituent i in the mixture; and
= Response factor of constituent i in the mixture;
A-9
-------
FACTORS3
W
CO
53
O
CO
w
K
_
0
[CATION
r i
I
VO
1
H
CQ
EH
O O |
(0 O a
_, ^3
QJ *J2 O
Sf 2 O
2 H *
0 ° H
ll i
« (0 °
>H .
O U
-P C >
O O B
*0 U ft
i •
4-> 0 4J 0
W ^ to ^-
O
^
O
•a
o
Pn
(0
C
o
•d
cu
(0
(0
Q)
(0
U
•H
e
Q)
ft
•d
(3 C
•H 3
O-H
H Q
1 BT3
U 0 C
U 3
0) O
•3 S &
A o e
(0 -HO
EH -P O
O
BO) (0 0)
•H (d
(US H
(0 -d o
Q) rH
W -P <0
r4 (0 U
•rl
ompound
nd i
O 3
o
C Pi
o e
4J 0
0
id O
Pi O S3
-------
The derivation of the above equation is presented in
Table A-7. Using the RF's and mole fraction information from
Table A-6, the RF for the mixture in Stream B is calculated as
follows:
RFm(6 500 ppmv) = (0.1036/2.49 + 0.8963/1.10)-1 = 1.17
and
RFm(e 10,000 ppmv) = (0.1036/0.72 + 0.8964/6.06)-1 = 3.43
From the above calculations, it can be seen that at an
actual concentration of 10,000 ppmv the RF is above three, which
means the screening values need to be adjusted. Table A-8
demonstrates the simplest approach for adjusting the screening
values. This approach involves multiplying all of the screening
values by whichever RF is higher.
Correcting the screening values by the approach described
above may be inaccurate in some cases. For example, if all or
most of the equipment have low screening values, using the RF
based on an actual concentration of 10,000 ppmv may cause an over
estimate in the calculated emission rate. A more precise
application of RF's is to plot the RF versus the screening value.
This can be done by fitting a straight line between the RF and
the corresponding screening values associated with the 500 and
10,000 ppmv actual concentrations. For the example case, this is
done as follows.
Screening value associated with actual concentration of
500 ppmv:
= (500 ppmv)/(RF at actual concentration of 500 ppmv)
= 500 ppmv/1.17
= 427 ppmv
Screening value associated with actual concentration of
10,000 ppmv:
= (10,000 ppmv)/(RF at actual concentration of
10,000 ppmv)
A-ll
-------
TABLE A-7. DERIVATION OF EQUATION USED TO ESTIMATE
RESPONSE FACTOR FOR A MIXTURE
(1) Response Factor (RF) Equation:
Actual Concentration (ppmv) _ A
Rr = . — —
Screening Value (ppmv) SV
(2) For a mixture, each compound will contribute to the actual concentration
and to the screening value, thus:
A « AI + A2 •*• A3 ..
SV = SV! + SV2 + SV3
Thus, the above equation converts to:
RF .
svx + sv2 + sv3 . . .
(3) The value for the screening value of each individual compound (SV^) is
calculated as:
SVi = —i-; substituting gives:
RFi
RF =
Al + A2 + A3
RF2 RF3
(4) The mole fraction of each individual compound (XjJ is calculated as
Thus, the actual concentration of compound i is calculated as:
/ substituting gives:
RF _ _ ATOT
X1ATOT + X2ATQT + X3ATQT ^ ^ ^ xl + X2 + X3
RFi RF2 RF3 ' ' * RFi RF2 RF3
(5) Thus, the response factor of a mixture is calculated as:
1
RF
n
E
i = 1
A-12
-------
fO
u
W
§
^
^4
o
p^
PM
CO
DC*
o
H
U
£
_ _
CO
o
CO
&
o
55
H
X
eu
CO
1
fr^
CQ
rt2
PJ
^^
u
0)
4>
id ^.^
V4 1.
B ^
j>"o>
m *"
•g
M
•°0)
Uf
3
i-l
•0 id
01 > — •
m C7> g
3 B ft
*o c ^^
0)
14
u
CO
o
ti 11
«( ^
uj 3
«X
m -H
§e
S*o
oj
Unadjusted
creening value
(ppmv)
m
o
H
4J
fi
i
Si
H
ro fO ro
ro ro ro
o O o in oo
ddd.Hrooiatoooo'O o
H (N CO Ot 00 00
i-H
0 O O^nooinoooo
0 « « n ° £ ^ » In 5 S
H \o r» r* in
4> 4J JM iH (M 00 I
«d *
•0 M id 03
§*id B 1
~4 01 0 0)
*H t^ *H ^4
Q jj
4> an
£ •• -H
.<
Hl-H fi
4J a o
« 01 -H
*• 3 itf ^J
Q) O* 3 (0
o, 0j ^
S O $
0 0) O
4> -H N
fi 4J 1 4 at M
&V4
X
B
o
I
>
•a
01
a
3
•O
id
B
3
^i^
||
U
*0
01
4)
0)
3
•o
•O
ja
B"
o
]p
rt
M
(11
a,
0
•w
O
n
3
5
^•^
4e
^^
01
4>
id
fj
-S
a
m
-H
al
0
a>
N
.p
•^
id
M-l
01
•o
on equation or
•H
^j
id
•H
0)
V4
O
u
"^
H
B
0
0)
01
•H
g
H
|
U
B
O
••H
0)
O
-------
10,000/3.43
= 2,915 ppmv
Figure A-l plots this screening value/RF relationship.
Table A-9 uses this plot to calculate emissions. Note that in
Table A-9, all of the screening values are adjusted. An
alternative would be to adjust only those screening values having
an associated RF greater than three. Note that for all screening
values less than 427 ppmv, the RF calculated at 427 ppmv is
applied, and, similarly, for all screening values above
2,915 ppmv, the RF at 2,915 ppmv is applied.
An alternative to using the RF's in Appendix C is to use the
analytical technique described in Chapter 3.0 to determine RF's
at several different actual concentrations. These RF's are then
related to the screening value. Once the RF's and associated
screening values are determined, a first-order or second-order
(if the relationship appears nonlinear) equation can be fitted to
the RF data. Table A-10 demonstrates how the collected data of
RF's at actual concentrations is converted to RF's for the
associated screening values. A hypothetical plot of the
RF/screening value relationship is shown in Figure A-2.
Table A-ll demonstrates how emissions can then be calculated by
applying the plot. Note that the line is not extrapolated beyond
the highest screening value for which data were obtained.
A-8. ANNUALIZING EMISSIONS
If more than one screening value is obtained from an
equipment piece, all of the screening values can be used to
estimate emissions, as long as the elapsed time between each
screening value obtained is known. This is demonstrated for pump
A-15 in Stream A. Table A-12 shows how emissions are calculated
for each period between the collection of screening values.
Notice that each screening value is used to estimate emissions
since the last screening value was obtained.
A-14
-------
3.5
L
0
*J
u
id
u.
Q)
U)
0
a
in
0)
tr
2.5
1.5
0.5
0.4
O.B
1.2 1.6
2.4 2.8
0.6 1 1.4 1.8 2.2 2.6
Screening Value Cppnw) in thousands
Figure A-l. Response Factor Curve Generated From Response Factor
Data in Table C-l
A-15
-------
(0
H
*C
H
g
O
H
£,
0
e
Q
f*
g
W
1
CO
o
CJ
r^
&4
H
CO
5
2
hH
W
3
g
CO
D
CO
r5
O
H
CO
CO
H
0
H
EH
g
H
EH
CO
W
•
T
H
»3
s
EH
U
01
4J
Vl
C ^^
-S ^
a *JJ;
•H rr.
8
W
8
r*
0<
1
•^
at
0*0 7
a 3 i
•a "id a.
4J
CO
3
s
m
o
Vl
o
,p
m vi
*• 5
ge
8,
m
2
at
3
•s?
4J *• ~
fjl
D J
u
CO
Q
l-l
g
s.
.
&
f*) f) ro c*)
O O O VO VO O
C3 ^3 ^3 C3 ^H ^^ ^^ ^^ C3 ^D ^3 *O ^D
v^ ^ft ^D 00 f*1* t*** C^
01 rn on oo ao
rH
O O O
at at at
N N NcsinroOOOOO
•P-P-P CMtOCMH^t^l
222 m t^ i^ in
Id Id « rH CN 00
at at at
o a a
r^t^c^^CNfnroro
rHrHiHCNU)^*^'^'
1 1 IrHrHtHrHOinrOnl
III 1
•o
g
8
at
ooooooooooo M
r-tromooooo o
CN in o o O o ca
M in ao in
M •iJ n
c
01
-p
CO
C 6
•H o at ••
•J-l C
4j m o
Id ID -H
•• 3 id jj
IcSl
o at o
4J-H N
a .P i M-I
01 Id 4J O
g rH rH
&ai 3 a
•H Vl Id Vl
P.H >M 3
&O 0) O
w o Q as
^^
rH
i
0)
3
&
-*-l
fo
g
0
«w
1
id
4J
0)
VI
s
"x
"g
M-l
o
^
>"
TJ
01
D
3
T?
id
c
^ff
II
CO
•o
0)
CO
3
•n
•o
A
"G
o
T}
id
Vl
8,
o
o
n
Vl
1
*•"*
^
"oT
Id
Vl
c
0
-H
m
n
o
Vl
01
4J
rH
3
id
*w
at
-o
Vl
0
o
•H
1
at
c
o
«
rH
0)
Vl
Vl
o
u
•V
II
c
o
•r)
a
a
•rl
1
g
u
*
,^
8
"*
id
^
nents
1
U
*w
o
• c
o o
C -rl
"™* -M
a)
H Vi
eg,
0 0
CO H-t
n o
e p
W M
3
U 0
^ **^
•o
A-16
-------
TABLE A-10.
GENERATION OF HYPOTHETICAL RESPONSE
FACTOR DATA FOR STREAM Ba
Actual
standard gas
concentration
(ppmv)
500
500
500
2,000
2,000
2,000
5,000
5,000
5,000
10,000
10,000
10,000
25,000
25,000
25,000
Sample number
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
Measured
screening
value
(ppmv)
375
390
390
Avg - 385
1,219
1,205
1P258
Avg = 1,227
1,865
1,930
1,872
Avg = 1,889
2,976
3,040
2.994
Avg - 3,003
6,361
6,394
6.476
Avg = 6,410
Response
factor
1.33
1.28
1.28
Avg =1.30
1.64
1.66
1.59
Avg = 1.63
2.68
2.59
2.67
Avg =2.65
3.36
3.29
3.34
Avg =3.33
3.93
3.91
3.86
Avg =3.90
aThis table is a demonstration of how analytical determination
of response factors can be used to generate a response
factor/screening value relationship.
A-17
-------
0
ed
n
-------
(0
CM
1
^*
W
g
5
H
fe
52
5
fe
Q
H
EH
r?
H
H
O
CO
Sf
o
£^
^^
.<
1*4
W
CO
Ai
CO
W
an
H
H
9
F^
9
P-I
u
JJ
* -^
** JH
^^
O ^
M
• „
m
^H
1
g>
•H
0)
01
o o) >"
m i3 §i
•O id Pi
«-
JJ
m
3
"I 1
*o
*•
o
Jj -»
*^ 01
•W ^j
• JJ
r* 6
o
I*. *M
BB ®
0)
***
0*
3
«?
JJ * — •
3§i
«!«
*g~
§2
u
m
p
M
ij
c
•rl
W
fO ro ro
ro ro ro lO
O o o r» on
OOOOrHrHrHf-OOOlJ O
H CM co r- w ^ r- r-
CM t rH CO O
•H* CM
O O Om^oinoooo
Vl IH ^rH^'Cn^'CMOOO
0)4)0) n oo vo o vo o
W W N fc » ^ ^
Z % S * ••» 2 !
333 * '
a) id id
<4_l 14.4 ^
o> ai a)
o o a
vo r^> vo o^ iH o o o
^t^inujnvocNcsj
Ii i , , . , _.. ___>. i
• i rH i~l rn rn (N <*> 'T *T '
s
0)
OOOOOOOOOOO 01
•nroinooooo *•
CM in o o o o u
> » % » B)
CM in co m
CM -P
1 «
i
0)
(0
MH
0
a
c
o
Hcsiro^'in»Or*eooAo'~'tv'"'"1
lllllllll^YV"
W
rH
a)
H
•*
CM
00
•
o
I
m
i
o
rH It
,rj
* *^
O tV
C/i
•VO
•H 1
O
H rH
— •*
rl
£ (Tt
^.^J- O
O* • CO
m X r* m
DO)**
O<4J 0) II
id JJ
•o vi id pa
•rl Ll
tn * c 1
•rl 0) 0 d)
rH >J-H rl
D JJ
J- .. -r|
O> C g
|J "rl fi
JJ 0) 0
« 0) -rl
.. 3 <0 JJ
0) D1 g Id
Q»
•a
id
c
3
N
CO
01
JJ
a
3
•n
T)
<
c"
0
5
^
&
o
u
id
rl
>
^^
*
mponents j
8
0
•
0
c
II
B
0 •
•rl *-»
D C
at o
•r>4 *^
^
rl
8S,
> 0
•a
A-19
-------
TABLE A-12. ANNUALIZING EMISSIONS
FOR LIGHT LIQUID PUMP A-15*
Hypothetical
date
Screening
value (ppmv)
Hours elapsed
since last
screening
value*3
VOC emissions
since last
screening
value0 (kg)
January 1
February 1
March 1
April 1
May 1
June 1
July 1
August 1
September 1
October 1
November 1
December 1
January 1
5,000
0
0
8,000
100
1,000
0
0
0
10,000
0
0
0
TOTALS:
—
744
672
744
720
744
720
744
744
720
744
720
744
8,760
—
0.006
0.005
23.3
0.6
4.2
0.005
0.006
0.006
27.0
0.006
0.005
0.006
55.1
aEquipment type: Light liquid pumps
Correlation equation: Leak rate (kg/hr) = 1.90 * 10~5 (SV)°-824
Default-zero mass emission rate: 7.49 * 10~6 kg/hr
bHours elapsed since the last screening value was obtained. For
example, the hours elapsed since the screening value obtained on
March 1 are the hours from February 1 to March 1, which equal
24 hr/day * 28 days, or 672 hours.
CVOC Emissions = (correlation equation or default-zero
emission rate) * (hours elapsed).
A-20
-------
A-9. ESTIMATING VOC EMISSIONS FROM EQUIPMENT CONTAINING ORGANIC
COMPOUNDS NOT CLASSIFIED AS VOC's.
Stream C contains ethane, which is an organic compound, but
is not classified as a VOC. When a monitoring instrument is used
to screen equipment in stream C, the resulting screening value
will include measurement of the ethane. However, the ethane
should not be included in the estimated VOC emission rate.
The following equation is applied to subtract out the ethane
contribution:
VOCcorr - VOCuncorr * WPvoc/WPorg
where:
VOCcorr = The corrected VOC mass emissions from the
equipment (kg/hr);
vocuncorr = Tne previously calculated "uncorrected" VOC
mass emissions from the equipment (kg/hr);
= Tne concentration of VOC in the equipment in
weight percent;
= The total concentration of organic compounds
in the equipment in weight percent.
Note that the above correction is only applied if screening data
are being used to estimate emissions.
The above calculation is demonstrated assuming screening
values have been obtained from equipment in Stream C as either
greater than or equal to 10,000 ppmv or less than 10,000 ppmv.
Assume 2 of the 40 gas valves in Stream C screened above
10,000 ppmv, and the remainder screened below 10,000 ppmv.
Uncorrected VOC emissions are calculated using the Screening
Ranges Approach:
VOC = GEF * Nge + LEF * Nle
where:
VOC = Total VOC emission rate for an equipment type
(kg/hr);
GEF = Applicable emission factor for sources with
screening values greater than or equal to
10,000 ppmv (kg/hr/source);
A-21
-------
Nge = Equipment count (specific equipment type) for
sources with screening values greater than or
equal to 10,000 ppmv;
LEF = Applicable emission factor for sources with
screening values less than 10,000 ppmv
(kg/hr/source); and
Nle ~ Equipment count (specific equipment type) for
sources with screening values less than
10,000 ppmv.
Thus,
VOC = 0.0782 kg/hr * 2 + 0.000131 kg/hr * 38
= 0.161 kg/hr
Converting to an annual emission rate gives:
= 0.161 kg/hr * 8,760 hr/yr
= 1,410 kg/yr
Using the weight fraction of the compounds in Stream C given
in Table A-l (65% ethyl aerylate, 25% ethane, and 10% water
vapor), the above emission rate is corrected as follows:
voccorr = vocuncorr * wpvoc/wporg
= 1,410 kg/yr * 65/(65 + 25)
= 1,020 kg/yr VOC emissions
A-10. ESTIMATING INORGANIC EQUIPMENT LEAKS
If the hypothetical process unit also had equipment that
contained a volatile inorganic compound, emissions could be
estimated using the following guidelines. If a monitoring device
is not available, the equipment emissions can be calculated using
the Average Emission Factor Approach. If a monitoring device is
available, the best approach for estimating the emissions is to
generate unit specific correlations, but the EPA Correlation
Equations could also be applied as in Section A-4. If the
monitoring device cannot accurately predict the screening value
A-22
-------
but can be used to predict concentrations greater than/less than
10,000 ppmv, the emissions may be estimated by applying the
Screening Value Ranges approach presented in Section A-3.
A-23
-------
APPENDIX B:
LEAK RATE/SCREENING VALUE CORRELATION DEVELOPMENT
AND REVISION OF SOCMI CORRELATIONS
AND EMISSION FACTORS
-------
APPENDIX B
The purpose of this appendix is to provide supplemental
information on the approach for developing site-specific
correlations as discussed in Chapter 2.0 of this document. Also,
this appendix contains background information on the data
collection and analysis performed to revise the SOCMI
correlations and emission factors, and presents summary
parameters associated with the SOCMI and refinery correlations.
Section B.I addresses the following:
• Analysis of bagging and screening data;
• Development of a correlation eguation; and
• Development of a default-zero leak rate.
Section B.2 addresses the following:
• Analysis of new SOCMI bagging data;
• Development of revised correlations and default-zero
leak rates;
• Development of revised SOCMI emission factors; and
• Summary of SOCMI and refinery correlation parameters.
B.I DEVELOPMENT OF SITE-SPECIFIC CORRELATION EQUATIONS
Development of site-specific correlations involves bagging
individual pieces of equipment. (Refer to Chapter 4.0 for
details on how equipment is bagged.) The emission rate and
associated screening value from several equipment pieces of the
same type (valve, pump, connector, etc.) and service (gas, light
liquid or heavy liquid) are used to develop a correlation. The
correlation predicts a leak rate based on a screening value. To
develop a correlation, "bagging data" must be collected. In this
appendix, "bagging data" refers to data used to estimate the mass
emission rate from an equipment piece, and the screening value
obtained with the portable monitoring instrument when the
equipment piece is bagged.
B.l.l Preliminary Analysis of Bagging Data.
For the purposes of this discussion, it is assumed the blow-
through method is used to bag the equipment piece. For each
B-l
-------
bagged (tented) equipment piece, two sample bags should be
collected. For each sample bag the following bagging data should
be recorded: (1) total organic compound concentration (ppmv)
measured in the sample bag at the laboratory using a GC or
similar instrument, (2) the mole percent and molecular weight of
each of the constituents in the sample bag (or alternatively in
the process stream contained within the enclosed equipment
piece), (3) the temperature in the tent when the sample bag is
collected, (4) the carrier gas flow rate out of the tent, (5) the
tent oxygen concentration (6) background bag organic compound
concentration measured at the laboratory (optional), and (7) the
density and volume of any organic liquid collected from the
bagged equipment piece and the time in which the liquid
accumulated.
In some cases, the sample bag total organic concentration
will be below the GC minimum detection limit. If this occurs,
one half the GC minimum detection limit should be used to
estimate emissions.
For each sample bag, the vapor leak rate is calculated using
the following equation:
(1.219*10-5)*(Q)*(MW)*(GC)
Vapor leak rate (kg/hr) =
T + 273.15
where:
1.219 x 10~5 = A conversion factor based on the gas constant
and assuming a pressure in the tent of
1 atmosphere (°K * 106 * kg-mol/m3)
Q = Flow rate out of tent (m3/hr)
_ N2 flow rate (£/min) 0.06 m3/min
• •"•' ' i • • -i-"— •• ..I.. ii - X • —''
1 - [tent oxygen cone, (volume %)/21] £/hr
T = Temperature in tent (°C)
B-2
-------
MW = Molecular weight of organic compounds in the sample bag or
alternatively in the process stream contained within the
equipment piece being bagged. For mixtures, MW is
calculated as follows:
n
MW =
n
where:
= Molecular weight of organic compound i;
Xi = Mole fraction of organic compound i; and
n = Number of organic compounds in the mixture.
GC = Sample bag organic compound concentration. If a background
sample bag is obtained, the value of GC can be corrected for
background organic compound concentration using the
following equation:
Oxy * BBC
GC = SBC - —
21
where:
SBC = Sample bag organic compound concentration
(ppmv);
Oxy = Tent oxygen concentration (volume %); and
BBC = Background sample bag organic compound
concentration.
The vapor leak rate calculated from the two sample bags is
averaged. Added to this average vapor leak rate is the leak rate
B-3
-------
of any liquid that is collected in the bag. The liquid leak rate
is calculated as follows:
Liquid leak rate (kg/hr) = p VL
16.67 t
where:
p = Density of organic liquid collected
(g/m£);
VL = Volume of organic liquid collected (m£);
t = Time in which liquid is collected
(minutes); and
16.67 = A conversion factor to adjust term to
units of kilograms per hour
[g * hr/(kg * min)]
Thus, the total emission rate for the bagged equipment piece is
as follows:
Leak rate (kg/hr) = Average vapor leak rate (kg/hr)
+ Liquid leak rate (kg/hr)
The screening value associated with each bagged equipment
piece is calculated by subtracting the background screening value
from the average of the initial and final screening values. In
cases where the background concentration was larger than the
average of the initial and final screening values, the screening
value should be recorded as 0 ppmv.
B.I.2 Correlation Equation Development.
After preliminary analysis of the bagging data is complete,
there will be a mass emission rate and corresponding screening
value associated with each individual equipment piece that was
bagged. All mass emission rate/screening value data pairs with
nonzero screening values are used to develop the site-specific
correlation. Data pairs with a screening value of zero can be
used to develop a default-zero leak rate, and this is discussed
in Section B.I.3.
B-4
-------
Two terms used in conjunction with developing the
correlation are defined as follows: "log space" — where the
logarithms of both the screening values and mass emission rates
are evaluated, and "arithmetic space" — where the actual screening
values and emission rates are evaluated. The data is first
analyzed in log space to develop an expression relating the
logarithm of the screening value to the logarithm of the mass
emission rate. This expression is then transformed to arithmetic
space to arrive at the correlation equation.
It is necessary to perform the initial analysis in log space
because both the screening value and mass emission rate data
typically span several orders of magnitude, and the data are not
normally distributed in arithmetic space. Normality of the data
is important for the validity of the statistical procedures being
used. Historically, the data have been shown to be approximately
log-normally distributed.
The first step in the development of the correlation
equation is to calculate the logarithm of each screening value
and mass emission rate. Note that the correlation developed will
be the same whether the natural logarithm or base 10 logarithm is
used. The next step is to perform simple linear (least squares)
regression in log space. The log of the mass emission rate
(dependent variable, Y) is regressed on the log of the screening
value (independent variable, X) . The resulting regression line
takes the following form:
vi = /30 + /*lxi
where :
Yi = Logarithm of the leak rate determined by
bagging equipment piece i;
Xj[ = Logarithm of the screening value for equipment
piece i ;
/3O = Intercept of regression line; and
jSjL = Slope of regression line.
B-5
-------
The value for the slope and intercept are calculated using the
following equations:
(XY) - (X) (Y)
PI - — — - zr~r —
X2 - (X) 2
and
where :
- E
X =
n
-_ EY
n
— E X,-
XY = ^
n
-•> E xi2
X2 = 1-
n
n = number of screening/bagging pairs.
Once these have been calculated, then the Mean Squared Error
(MSB) can be given by:
n
1
MSB =
n - 2
B-6
-------
where:
ri = Yi -/30 -/^Xi
The slope and intercept and a scale bias correction factor
(SCBF) are used in the final step to transform the regression
equation from log space to arithmetic space. The transformed
equation is the correlation equation and it is calculated as
follows:
Leak rate (kg/hr) = SBCF * (e or 10) ^o * (Screening value) "1
Note that if the natural logarithm of the leak rates and
screening values is used when developing the regression line,
then the "e" term should be raised to the power of the intercept
(£0) • On the other hand, if the base 10 logarithm of the leak
rates and screening values is used when developing the regression
line, then the "10" term should be raised to the power of the
intercept ((3$) .
The SBCF is a correction factor which accounts for the
variability of the data in the log space (see discussion in
Section 2.3.4). It is obtained by summing a sufficient number
(usually 10-15) of the terms from the infinite series given
below:
(m-l)*T (m-l)3*T2 (m-l)5*T3
SBCF = 1 + + + +
m m2*2!*(m+l) m3*3!*(m-fl)*(m+3)
where:
T (when regression performed using base 10 logarithms)
= (MSE/2)*((lnlO)2) ;
T (when regression performed using natural logarithms)
= (MSE/2);
MSB = mean square error from the regression;
InlO = natural logarithm of 10; and
m = number of data pairs.
B-7
-------
B.I.3 Determination of Default Zero Leak Rates
A default zero leak rate can be calculated based on the
emission rates measured from bagged equipment that have a
screening value of zero ppmv. The first step to determine the
default-zero leak rate is to take the logarithm of each of the
mass emission rates and then determine the average log leak rate.
The average log leak rate is used to calculate the default-zero
mass emission rate. Analysis is performed in log space rather
than just determining the arithmetic average because this gives
the most efficient estimator of the default-zero leak rate. The
average log leak rate and a scale bias correction factor, that
takes into account the variance of the log mass emission rates,
are then utilized in the following equation to calculate the
default zero leak rate:
Default Zero Leak Rate = SBCF * (10 or e)LOG:AVG
(kg/hr)
where:
SBCF = Scale bias correction factor for
the logs of the mass emission
rates; and
LOG:AVG = Average of the logs of the mass
emission rates.
The SBCF for the default zero determination is calculated using
the same equation for the SBCF as presented in Section B.I.2,
with the following exception: the variance of the log mass
emission rates is used in the "T" term for the default zero SBCF,
rather than the regression mean square error (MSB). The variance
(S2) is calculated as:
n
S2 = £ (LOG:LEAKi - LOG:AVG)2
11-1 i = l
B-8
-------
where:
LOGrLEAKi = Logarithm of leak rate from component i;
LOGiAVG = Average of the logs of the mass emission
rates; and
n = Number of data points.
B.2 DEVELOPMENT OF REVISED SOCMI CORRELATIONS AND FACTORS
In 1990 bagging data were obtained from several ethylene
oxide (EO) and butadiene (BD) producers. Bagging data were
collected from connectors, light liquid pumps, gas valves, and
light liquid valves. In 1987 and 1988 screening data had been
collected from the same EO/BD process units. These bagging and
screening data were used to revise the SOCMI correlations and
factors.
(Note that as used in the following discussion, "bagging
data" refers to the screening value/mass emission data pairs, and
"screening data" to the data set of screening values collected
independently of the bagging data. Normally, bagging data are
collected from a chosen set of equipment pieces to provide the
best data for developing a correlation. On the other hand,
screening data are collected from all equipment pieces to give a
representative distribution of screening values).
To revise the SOCMI correlations and factors, the data
collected from the EO/BD process units were compared with data
previously collected from SOCMI process units. In the following
discussion this previously collected data are referred to as
"old" data. The old SOCMI bagging data were collected in the
Six-Unit Maintenance Study (EPA-600/S2-81-080). The old SOCMI
screening data were collected in the 24-Unit Study
(EPA-600/2-81-003). The EO/BD data are referred to as "new."
When the data sets are joined, the resulting data set is referred
to as "combined."
B.2.1 Analysis of SOCMI Bagging Data
Following the approach described in Section B.I, the new
SOCMI bagging data were analyzed to develop new correlations.
B-9
-------
A comparison of the old and new bagging data was performed to
evaluate any differences. Note that for connectors, only new
bagging data were analyzed since connectors were not bagged as
part of the Six-Unit Maintenance Study. Attachment 1 includes
the complete list of each of the emission rate/screening value
datapoints and presents summary tables on the regression
statistics of the old, new, and combined data.
To evaluate the differences between the new and the old data
for light liquid pumps, light liquid valves, and gas valves, the
following statistical tests were applied:
• Wilcoxon test of paired differences, and
• F-test of statistical parameters.
The statistical tests did not have consistent results for the
three equipment types. For light liquid pumps, no statistically
significant differences were found, for light liquid valves, the
tests indicated significant differences, and for gas valves, the
tests were inconclusive.
A better comparison was a visual comparison of the data
plotted in log space. This comparison was made by developing
plots of the old and new bagging data with regression lines
superimposed. All of the regression equations are plotted in
Figures B-l through B-4. Figure B-l presents the new bagging
data and regression equation for connectors. Figures B-2 through
B-4 show old and new bagging data superimposed upon the old, new,
and combined regression equations for light liquid pumps, gas
valves, and light liquid valves, respectively. The regression
lines in these four figures are drawn to correspond only to the
data points from which they were derived.
Figures B-2 through B-4 suggest the old and new data points
appear to lie along a common axis with a similar amount of
scatter. Figures B-2 through B-4 also demonstrate that most of
the old data were from equipment which had screening values
exceeding 1,000 ppmv, whereas a significant portion of the new
data came from equipment screening less than 1,000 ppmv. The
B-10
-------
L
u
0)
p
<0
cr
c
0
fft
0)
UJ
01
0
o
r
(1)
CO
CO
03
-1
-2
-3
-4
-5
-6
-8
-1
Base 10 Log Screening Value (Log ppmv)
• New Data Point
Figure B-l. Connector Regression Equation
B-ll
-------
r\
L
r
cb
.y
Q>
0
0)
•H
id
cr
c
0
U)
in
UJ
o>
o
-------
r\
L
0
J
(0
o:
c
0
0)
(I)
m
o>
o
D
r
0)
U)
(0
m
-1
-2
-3
-4
-5
-6
-7
Conbined
New
•
0
1
2
Base 10 Log Screening Value [Log ppmv]
I New Data Point + Old Data Point
Figure B-3. Gas Valve Regression Equations
B-13
-------
n
L
\
D)
0)
0
J
U
(0
cr
c
0
in
en
i
HI
0)
o
(0
CD
-1
-2
-3
-4
-5
-B
-7
-8
• •. • ,••
• • -J^J I . • •
-1
Base 10 Log Screening Value (Log ppmv)
New Data Point + Old Data Point
Figure B-4. Light Liquid Valve Regression Equations
B-14
-------
correlation derived from combining the old and new bagging data
spans the greatest range of screening values. Additionally, for
each of the equipment types, the combined correlation equation
has the best fit. Since the combined regressions span the
greatest range of screening values and have the best fit, the
combined data set was used to develop the revised SOCMI
correlation equations.
B.2.2 Development of Revised SOCMI Correlations and
Development of Default-Zero Factors.
After the old and new bagging data were combined, an initial
regression analysis was performed on the logarithms of the
screening values and mass emission rates following the procedures
outlined in Section B.I on the development of correlation
equations. For the combined data sets outliers were removed.
The residuals (differences between measured log mass emission
rates and log mass emission rates predicted by the regression)
were used to flag outliers. A data pair was flagged as an
outlier whenever the absolute value of its studentized residual
(the residual divided by its standard error) was greater than or
equal to 3. These data pairs are indicated as outliers in the
table contained in Attachment 1, which lists the screening values
and mass emission rates for the combined bagging data set.
Attachment 2 contains a table listing all of the bagging
data used to develop the default zero mass emission rates. These
data were collected at the EO/BD process units, and were analyzed
using the approach outlined in Section B.I.3.
B.2.3 Revision of SOCMI Emission Factors
After the SOCMI correlations were revised, they were
utilized in conjunction with the "old", "new", and "combined"
screening value data sets to revise the SOCMI emission factors.
Recall that the "old" screening data were the data collected in
the SOCMI 24-Unit Study (EPA-600/2-81-003), the "new" screening
data were the data collected from the EO/BD process units in 1987
and 1988, and the combined data were the two data sets combined.
B-15
-------
Using screening data in conjunction with the applicable
correlation eguation, emission factors are calculated in the
following manner.
(1) Screening values with a value of zero are assigned the
default zero emission rate,
(2) All other screening values are entered into the
applicable correlation equation to determine the
associated mass emission rate, and
(3) The sum of all of the individual emission rates is
divided by the total number of screening values
(i.e., equipment pieces) to give the average factor.
These steps were followed to revise the SOCMI average emission
factors for connectors, light liquid pumps, gas valves, and light
liquid valves. The same approach was used to revise the SOCMI
Screening Range Emission factors (>10,000 ppmv / <10,000 ppmv),
except that the screening values were segregated into the two
ranges to calculate the average of each range.
Consistent with development of the revised SOCMI correlation
equations (which were developed from the combined bagging data
set), the combined screening data set was used to revise the
SOCMI factors. The combined data set has the advantage that it
reflects changes that have occurred in SOCMI process units since
the 24-Unit Study, and contains data from a representative
sampling of SOCMI process units.
To develop the emission factors it was necessary to make
adjustments to a small percentage of the screening values. These
adjustments were applied to large screening values that were
identified as "pegged data." The large screening value data are
important in the emission factor calculations and these
adjustments were made in an attempt to keep as many screening
values in the analysis as possible.
Examination of the frequency distributions of the screening
value data sets revealed spikes near 10,000 ppmv (between 9980
and 10,001 ppmv) and near 100,000 ppmv (between 99,980 and
B-16
-------
100,001 ppmv). These spikes indicate that the instrument was
"pegged" or unable to measure the concentration being sampled
because the concentration was beyond the measurement range of the
instrument. It was assumed that screening values pegged at
10,000 ppmv had actual values between 10,000 and 100,000 ppmv,
and that screening values pegged at 100,000 ppmv had actual
values greater than 100,000 ppmv. Because there were several
screening values greater than 10,000 ppmv and 100,000 ppmv that
were not pegged, an average from the two ranges
(10,000-100,000 ppmv and >100,000 ppmv) was calculated to
substitute for the pegged readings. For the 10,000-100,000 ppmv
range, the average was 33,620 ppm and for the greater than
100,000 ppmv range, the average was 302,367 ppm. These averages
were used in the emission factor analysis for pegged data from
the screening data sets. Thus, each pegged screening value was
assigned the applicable average screening value, which was
entered into the correlation to predict emissions.
Attachment 3 lists the average emission factors generated
from each of the screening data sets, using the revised SOCMI
correlations. There are thousands of screening values in the
data sets, and these data sets are not reproduced in this
appendix. Instead, figures plotting the distribution of the
screening values are presented in Attachment 3.
B.2.4 Summary of SOCMI and Refinery Correlation Parameters
Table B-l presents the regression line slope and intercept
and the SBCF associated with each of the SOCMI and refinery
correlations contained in Tables 2-7 and 2-8 of this document.
B-17
-------
TABLE B-l. SUMMARY OF SOCMI AND REFINERY CORRELATION PARAMETERS,
Equipment type
compressors,
pressure relief valves
Connectors
Heavy liquid pumps
Regression
intercept3
()8Q)
Regression
slope (/?n)
•5.543
-5.443
0.88
1.04
SBCF
SOCMI Correlations
Gas valves
Light liquid valves
Light liquid pumps
Connectors
Refinery Correlation
Gas valves
Light liquid valves
Light liquid pumps,
-6.529
-6.069
-5.273
-6.434
-7.343
-5.243
-4.743
0.873
0.797
0.824
0.885
1.23
0.80
0.83
6.315
7.520
3.563
8.298
4.81
2.53
4.58
2.02
2.44
aRegression intercepts are based on analysis in log space using
Base 10 logarithms of leak rates in kg/hr.
B-18
-------
APPENDIX B: ATTACHMENT 1
This attachment lists bagging data used to develop the
combined correlation equations for each of the equipment types in
Table B-l-1. Also included is a summary table (Table B-l-2) of
the regression statistics associated with the old, new, and
combined SOCMI bagging data sets. Note that the regression
statistics presented in Table B-l-2 are based on development of
the regression lines using natural log leak rates and natural log
screening values.
B-19
-------
Table B-1-1. Bagging data used to develop the combined correlation equations.
Plant
Type
EO
EO
EO
EO
EO
EO
BD
EO
EO
BD
EO
EO
EO
EO
EO
BD
BD
BD
BD
BD
BD
EO
BD
BD
EO
EO
EO
EO
EO
BD
EO
BD
BD
BD
BD
EO
EO
BD
EO
EO
BD
BD
BD
BD
EO
Data
Origin
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
...... cquipmer
Measured
Emission
Rate (kg/hr)
0.0000000728
0.0000000734
0.0000001004
0.0000001061
0.0000001101
0.0000001137
0.0000001265
0.0000001544
0.0000001613
0.0000001620
0.0000001644
0.0000001731
0.0000002953
0.0000002996
0.0000003195
0.0000003254
0.0000003346
0.0000003430
0.0000003442
0.0000003939
0.0000003994
0.0000004007
0.0000004288
0.0000004757
0.0000004798
0.0000005309
0.0000005812
0.0000005944
0.0000006075
0.0000006524
0.0000007355
0.0000007648
0.0000008560
0.0000008798
0.0000008869
0.0000008924
0.0000009888
0.0000010715
0.0000012661
0.0000016351
0.0000017995
0.0000018303
0.0000020777
0.0000022858
0.0000032725
11 i ype=bUNNCv. iv
Screening
Value (ppmv)
299.00
2.00
4.50
0.50
6.00
0.80
2.90
21.50
4.25
1.00
2.00
18.50
458.50
0.40
0.40
13.80
1.70
1.35
12.75
4.00
10.00
0.80
4.00
1.50
999.00
399.40
2.75
28.50
128.00
97.00
3.50
3.25
8.50
28.50
2.00
8.30
4.25
17.00
1.00
4.50
4.00
19.25
3.50
3.75
3.00
IK oervice=n
Natural
Log of
Emission
Rate
(kg/hr)
-16.4361
-16.4271
-16.1142
-16.0586
-16.0217
-15.9900
-15.8832
-15.6835
-15.6400
-15.6354
-15.6207
-15.5693
-15.0354
-15.0209
-14.9565
-14.9382
-14.9105
-14.8856
-14.8819
-14.7473
-14.7334
-14.7300
-14.6623
-14.5586
-14.5499
-14.4486
-14.3582
-14.3357
-14.3140
-14.2426
-14.1227
-14.0837
-13.9710
-13.9436
-13.9356
-13.9293
-13.8267
-13.7464
-13.5795
-13.3238
-13.2280
-13.2110
-13.0842
-12.9888
-12.6300
Natural
Log of
Screening
Value
(ppmv)
5.7004
0.6931
1 .5041
-0.6931
1.7918
-0.2231
1.0647
3.0681
1.4469
0.0000
0.6931
2.9178
6.1280
-0.9163
-0.9163
2.6247
0.5306
0.3001
2.5455
1.3863
2.3026
-0.2231
1.3863
0.4055
6.9068
5.9900
1.0116
3.3499
4.8520
4.5747
1.2528
1.1787
2.1401
3.3499
0.6931
2.1163
1.4469
2.8332
0.0000
.5041
.3863
2.9575
.2528
.3218
.0986
B-20
-------
Table B-1-1. Bagging data used to develop the correlation equations
... equipment i ype-uuNNei. 1 1
(continued)
Plant
Type
EO
BD
BD
EO
EO
BD
BD
BD
BD
EO
BD
BD
BD
EO
BD
BD
BD
EO
BD
EO
BD
BD
BD
EO
BD
BD
w_
BD
BD
BD
BD
BD
BD
EO
EO
BD
BD
BD
BD
BD
BD
BD
BD
BD
Data
Origin
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
Measured
Emission
Rate (kg/hr)
0.0000036190
0.0000036396
0.0000038387
0.0000041625
0.0000044784
0.0000046207
0.0000057784
0.0000080668
0.0000095125
0.0000100797
0.0000137255
0.0000140845
0.0000140911
0.0000142252
0.0000143958
0.0000151611
0.0000161064
0.0000166253
0.0000168916
0.0000178679
0.0000183124
0.0000191290
0.0000194650
0.0000197515
0.0000198244
0.0000227951
0.0000279813
0.0000348217
0.0000351763
0.0000359334
0.0000403480
0.0000423987
0.0000445724
0.0000509982
0.0000512445
0.0000595643
0.0000758688
0.0000860423
0.0000910990
0.0000947099
0.0001007398
0.0001051050
0.0001178839
0.0001397861
Screening
Value (ppmv)
1.60
0.80
8.50
6.50
48.00
7.80
41.50
12.00
100.00
297.00
19.75
4.50
14.00
63.50
195.50
16.00
13.50
18.50
195.00
0.95
123.50
4995.00
16.50
50.50
23.00
320.50
67.00
18.00
195.50
9.00
198.00
472.00
13.00
25.00
323.00
275.00
35.00
98.00
1049.00
94.40
197.50
38.80
94.80
371.00
JK service-*
Natural
Log of
Emission
Rate
(kg/hr)
-12.5293
-12.5236
-12.4704
-12.3894
-12.3162
-12.2850
-12.0614
-11.7278
-11.5629
-11.5050
-11.1963
-11.1704
-11.1700
-11.1605
-11.1486
-11.0968
-11.0363
-11.0046
-10.9887
-10.9325
-10.9079
-10.8643
-10.8469
-10.8323
-10.8286
-10.6890
-10.4840
-10.2653
•10.2551
-10.2338
-10.1180
-10.0684
-10.0184
-9.8837
-9.8789
-9.7285
-9.4865
-9.3607
-9.3036
-9.2647
-9.2030
-9.1606
-9.0458
-8.8754
LL
Natural
Log of
Screening
Value
(ppmv)
0.4700
-0.2231
2.1401
1.8718
3.8712
2.0541
3.7257
2.4849
4.6052
5.6937
2.9832
1.5041
2.6391
4.1510
5.2756
2.7726
2.6027
2.9178
5.2730
-0.0513
4.8162
8.5162
2.8034
3.9220
3.1355
5.7699
4.2047
2.8904
5.2756
2.1972
5.2883
6.1570
2.5649
3.2189
5.7777
5.6168
3.5553
4.5850
6.9556
4.5475
5.2857
3.6584
4.5518
5.9162
B-21
-------
Table B-1-1. Bagging data used to develop the correlation equations
(continued)
Plant
Type
BD
BD
BD
BD
BD
BD
BD
BD
BD
EO
BD
BD
BD
BD
BD
BD
BD
BD
Data
Origin
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
Measured
Emission
Rate (kg/hr)
0.0001721438
0.0001806903
0.0002038979
0.0002463283
0.0002731277
0.0002853205
0.0003727741
0.0004184529
0.0005627360
0.0008093015
0.0008566981
0.0013381945
0.0013408366
0.0017192076
0.0021650014
0.0085056085
0.0101785661
0.0587476684
Screening
Value (ppmv)
54.90
4747.00
895.00
97.00
549.00
345.00
198.50
199.00
195.00
997.00
99.00
1049.00
999.00
471.50
1997.00
2999.00
3996.00
99998.80
IK services
Natural
Log of
Emission
Rate
(kg/hr)
-8.6672
-8.6187
-8.4979
-8.3088
-8.2056
-8.1619
-7.8945
-7.7789
-7.4827
-7.1193
-7.0624
-6.6164
-6.6145
-6.3659
-6.1353
-4.7670
-4.5875
-2.8345
l_l_
Natural
Log of
Screening
Value
(ppmv)
4.0055
8.4653
6.7968
4.5747
6.3081
5.8435
5.2908
5.2933
5.2730
6.9048
4.5951
6.9556
6.9068
6.1559
7.5994
8.0060
8.2930
11.5129
107 (0 outliers)
B-22
-------
Table B-1-1. Bagging data used to develop the correlation equations
Plant
Type
BO
BD
BD
EO
BD
EO
EO
EO
RE
EO
EO
BO
RE
BD
BD
BD
BD
RE
BD
RE
EO
BD
BD
EO
BD
EO
BD
BD
EO
EO
RE
RE
RE
BD
RE
BD
RE
EO
EO
RE
EO
EO
RE
EO
BD
Data
Origin
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
OLD
NEW
NEW
NEW
OLD
NEW
NEW
NEW
NEW
OLD
NEW
OLD
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
OLD
OLD
OLD
NEW
OLD
NEW
OLD
NEW
NEW
OLD
NEW
NEW
OLD
NEW
NEW
— equipn
Measured
Emission
Rate (kg/hr)
0.0000003333
0.0000003346
0.0000004908
0.0000012091
0.0000021532
0.0000038359
0.0000055733
0.0000067016
0.0000068315
0.0000115240
0.0000137032
0.0000173708
0.0000182707
0.0000218470
0.0000234610
0.0000243023
0.0000262744
0.0000273344
0.0000287475
0.0000343297
0.0000385230
0.0000418537
0.0000474696
0.0000588925
0.0000715064
0.0000722114
0.0000978468
0.0001152858
0.0001232483
0.0001803724
0.0001957145
0.0001991513
0.0002209241
0.0002667811
0.0002999432
0.0003013546
0.0004782523
0.0005168934
0.0005477897
0.0005646821
0.0005681949
0.0005857415
0.0006402389
0.0006886734
0.0007364641
nent iype=KunK :
Screening
Value (ppcnv)
3.00
64.40
1.30
4.00
9.50
768.00
49.00
8.40
42.53
3.00
1.00
15.00
83.26
21.00
8.00
10.00
95.00
647.80
7.80
719.36
13.90
394.00
4.00
2.75
33.00
180.00
1.00
2.75
74.00
44.00
47.12
49.68
744.91
892.50
2388.28
65.00
49.86
105.00
499.00
16033.45
595.00
349.00
3102.49
199.00
598.00
>ervjce=LL -
Natural
Log of
Emission
Rate
(kg/hr)
-14.9141
-14.9104
-14.5272
-13.6256
-13.0486
-12.4711
-12.0975
-11.9132
-11.8940
-11.3711
-11.1979
-10.9607
-10.9102
-10.7314
-10.6602
-10.6249
-10.5469
-10.5074
-10.4570
-10.2795
-10.1643
-10.0813
-9.9554
-9.7398
-9.5457
-9.5359
-9.2321
-9.0681
-9.0013
-8.6205
-8.5389
-8.5214
-8.4177
-8.2291
-8.1119
-8.1072
-7.6454
-7.5677
-7.5096
-7.4792
-7.4730
-7.4426
-7.3537
-7.2807
-7.2137
Natural
Log of
Screening
Value
(ppmv)
1.0986
4.1651
0.2624
1.3863
2.2513
6.6438
3.8918
2.1282
3.7503
1.0986
0.0000
2.7081
4.4220
3.0445
2.0794
2.3026
4.5539
6.4736
2.0541
6.5784
2.6319
5.9764
1.3863
1.0116
3.4965
5.1930
0.0000
1.0116
4.3041
3.7842
3.8526
3.9057
6.6133
6.7940
7.7783
4.1744
3.9091
4.6540
6.2126
9.6824
6.3886
5.8551
8.0400
5.2933
6.3936
Outlier
Flag
OUTLIER
OUTLIER
B-23
-------
Table B-1-1. Bagging data used to develop the correlation equations
Equipment iype=runp s
(continued)
Plant
Type
RE
RE
RE
BD
RE
BD
RE
BD
EO
RE
BD
BD
EO
BD
BD
RE
RE
BD
BD
BD
BD
EO
BD
BD
RE
RE
RE
BD
RE
BD
RE
BD
BD
EO
BD
BD
RE
RE
BD
RE
RE
BD
RE
RE
Data
Origin
OLD
OLD
OLD
NEW
OLD
NEW
OLD
NEW
NEW
OLD
4JEW
NEW
NEW
NEW
NEW
OLD
OLD
NEW
NEW
NEW
NEW
NEW
NEW
NEW
OLD
OLD
OLD
NEW
OLD
NEW
OLD
NEW
NEW
NEW
NEW
NEW
OLD
OLD
NEW
OLD
OLD
NEW
OLD
OLD
Measured
Emission
Rate (kg/hr)
0.0007563452
0.0007987816
0.0009912542
0.0010889569
0.0011480956
0.0012930833
0.0013248663
0.0014886548
0.0016401471
0.0017660014
0.0018539657
0.0021087390
0.0022296212
0.0023007567
0.0025947420
0.0027435637
0.0029144932
0.0029456140
0.0033415187
0.0036014533
0.0036569429
0.0037009240
0.0037297151
0.0039913442
0.0041248489
0.0046220969
0.0046281246
0.0050222262
0.0054013839
0.0055450728
0.0070361493
0.0071307927
0.0081605157
0.0090139120
0.0098565101
0.0101206645
0.0108936908
0.0110475772
0.0115165376
0.0120415404
0.0120492786
0.0126046858
0.0135546418
0.0138366847
Screening
Value (ppmv)
1378.39
8095.43
289.26
471.00
521.79
348.00
2221.10
3197.00
101.20
24145.32
299.00
997.00
2000.00
5499.25
1993.80
2125.99
5870.47
5.75
125.00
1899.00
1393.90
3197.50
599.00
700.00
2775.53
16654.09
5538.83
1099.00
9501 .80
2998.00
1381.77
27.60
6498.00
7696.90
2548.00
2997.00
12820.53
14254.89
3194.50
20840.78
19187.09
5248.25
15011.05
10491.80
iervice=LL -
Natural
Log of
Emission
Rate
(kg/hr)
-7.1870
-7.1324
-6.9165
-6.8225
-6.7697
-6.6507
-6.6264
-6.5099
-6.4130
-6.3390
-6.2904
-6.1617
-6.1059
-6.0745
-5.9543
-5.8985
-5.8381
-5.8274
-5.7013
-5.6264
-5.6111
-5.5992
-5.5914
-5.5236
-5.4907
-5.3769
-5.3756
-5.2939
-5.2211
-5.1948
-4.9567
-4.9433
-4.8084
-4.7090
-4.6196
-4.5932
-4.5196
-4.5055
-4.4640
-4.4194
-4.4188
-4.3737
-4.3010
-4.2804
Natural
Log of
Screening
Value Outlier
(ppmv) Flag
7.2287
8.9991
5.6673
6.1549
6.2573
5.8522
7.7058
8.0700
4.6171
10.0918
5.7004
6.9048
7.6009
8.6124
7.5978
7.6620
8.6777
1.7492
4.8283
7.5491
7.2399
8.0701
6.3953
6.5511
7.9286
9.7204
8.7855
7.0022
9.1592
8.0057
7.2311
3.3178
8.7792
8.9486
7.8431
8.0054
9.4588
9.5649
8.0692
9.9447
9.8620
8.5656
9.6165
9.2583
B-24
-------
Table B-1-1. Bagging data used to develop the correlation equations
equipment iype=Kuw !
(continued)
Plant
Type
BO
BO
RE
RE
RE
BO
BO
RE
BD
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
BD
RE
RE
RE
BD
RE
BD
RE
BD
Data
Origin
NEW
NEW
OLD
OLD
OLD
NEW
NEW
OLD
NEW
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
NEW
OLD
OLD
OLD
NEW
OLD
NEW
OLD
NEW
Measured
Emission
Rate (kg/hr)
0.0154757686
0.0155724932
0.0156873305
0.0159032925
0.0196113751
0.0198424922
0.0219422932
0.0220953073
0.0221617288
0.0226278893
0.0232021936
0.0258831450
0.0263221310
0.0274280572
0.0300037851
0.0305561087
0.0361388265
0.0371630240
0.0409811410
0.0476567087
0.0480145702
0.0492542578
0.0556463965
0.0572488867
0.0586671574
0.0863688407
0.0977863072
0.1039387219
0.1074526291
0.2535689673
Screening
Value (ppmv)
3998.50
3998.00
300.60
51041.21
88270.79
2748.50
797.00
38632.61
6996.50
12H2.30
22078.88
10996.59
8527.17
193253.34
12130.06
16850.04
9472.44
37500.32
12196.61
130564.77
23101.38
5998.00
38446.34
3111.50
41504.10
99996.00
88269.36
5997.00
45285.17
99994.00
>erv?ce=LL -
Natural
Log of
Emission
Rate
(kg/hr)
-4.1685
-4.1622
-4.1549
-4.1412
-3.9316
•3.9199
-3.8193
-3.8124
-3.8094
-3.7886
-3.7635
-3.6542
-3.6373
-3.5962
-3.5064
-3.4882
-3.3204
-3.2924
-3.1946
-3.0437
-3.0363
-3.0108
-2.8887
-2.8603
-2.8359
-2.4491
-2.3250
-2.2640
-2.2307
-1.3721
Natural
Log of
Screening
Value Outlier
(ppmv) Flag
8.2937
8.2935
5.7058
10.8404
11.3882
7.9188
6.6809
10.5619
8.8532
9.4045
10.0024
9.3053
9.0510
12.1718
9.4034
9.7321
9.1561
10.5321
9.4089
11.7796
10.0476
8.6992
10.5570
8.0429
10.6335
11.5129
11.3881
8.6990
10.7207
11.5129
W = 119 (2 outliers)
B-25
-------
Table B-1-1. Bagging data used to develop the correlation equations
Plant
Type
EO
EO
EO
BO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
BD
BD
EO
BD
BO
EO
EO
BD
EO
EO
BD
EO
EO
BD
BD
BD
BD
BD
EO
EO
EO
BD
EO
BD
EO
EO
EO
Data
Origin
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
•- equip
Measured
Emission
Rate (kg/hr)
0.0000000717
0.0000000720
0.0000000737
0.0000001062
0.0000001082
0.0000001147
0.0000001167
0.0000001170
0.0000001172
0.0000001198
0.0000001251
0.0000001525
0.0000001579
0.0000001705
0.0000001964
0.0000002292
0.0000002537
0.0000002824
0.0000003468
0.0000003511
0.0000003724
0.0000004915
0.0000005202
0.0000005222
0.0000005551
0.0000006288
0.0000007041
0.0000007204
0.0000007597
0.0000008744
0.0000010541
0.0000013384
0.0000013799
0.0000013870
0.0000018645
0.0000018779
0.0000021100
0.0000022366
0.0000024148
0.0000025627
0.0000034003
0.0000036200
0.0000036375
0.0000038715
0.0000042396
nent I ype-vnLvc
Screening
Value Cpptnv)
37.50
35.00
2.00
1.00
4.00
4.00
0.10
9.00
5.00
4.00
21.50
1.20
1.00
2.00
98.25
3.00
224.30
9.00
6.20
1.75
0.40
1.00
1.50
108.00
4.00
1.25
0.20
1497.50
2.50
68.90
198.00
51.50
3499.30
15.70
6.00
1.50
99.00
0.20
598.00
28.00
678.00
6.00
19.00
118.25
38.40
aervice=u -
Natural
Log of
Emission
Rate
(kg/hr)
-16.4508
-16.4468
-16.4235
-16.0577
-16.0396
-15.9811
-15.9641
-15.9608
-15.9591
-15.9374
-15.8945
-15.6963
-15.6615
-15.5848
-15.4430
-15.2887
-15.1869
-15.0800
-14.8747
-14.8622
-14.8032
-14.5259
-14.4690
-14.4652
-14.4041
-14.2795
-14.1663
-14.1434
-14.0903
-13.9497
-13.7628
-13.5241
-13.4935
-13.4884
-13.1925
-13.1854
-13.0688
-13.0105
-12.9339
-12.8744
-12.5916
-12.5290
-12.5242
-12.4619
-12.3710
Natural
Log of
Screening
Value Outlier
-------
Table 8-1-1. Bagging data used to develop the correlation equations
equipment iype=VALVt
(continued)
Plant
Type
BD
EO
BD
EO
BD
BD
EO
RE
BD
BD
RE
RE
RE
EO
BD
EO
BD
BD
BD
BD
RE
RE
BD
BD
BD
RE
EO
BD
BD
RE
RE
RE
RE
RE
RE
RE
RE
BD
RE
RE
RE
RE
EO
RE
Data
Origin
NEW
NEW
NEW
NEU
NEW
NEW
NEW
OLD
NEW
NEW
OLD
OLD
OLD
NEW
NEW
NEW
NEW
NEW
NEW
NEW
OLD
OLD
NEW
NEW
NEW
OLD
NEW
NEW
NEW
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
NEW
OLD
OLD
OLD
OLD
NEW
OLD
Measured
Emission
Rate (kg/hr)
0.0000045549
0.0000056834
0.0000061124
0.0000070548
0.0000074252
0.0000080241
0.0000083624
0.0000118648
0.0000128110
0.0000137662
0.0000149663
0.0000166075
0.0000175591
0.0000214657
0.0000220929
0.0000243523
0.0000246644
0.0000263657
0.0000285391
0.0000298709
0.0000357822
0.0000359337
0.0000365393
0.0000395358
0.0000421641
0.0000440123
0.0000445925
0.0000523996
0.0000557747
0.0000617007
0.0000647076
0.0000724907
0.0000779572
0.0000833618
0.0000996210
0.0001071514
0.0001137777
0.0001197735
0.0001341897
0.0001376705
0.0001518078
0.0001625511
0.0001720041
0.0001766026
Screening
Value (ppmv)
5.40
9.50
4.00
2.10
17.50
3.40
40.15
20.46
8.50
83.90
4952.69
4954.50
1007.37
698.50
20.50
850.00
144.50
139.25
15.50
109.00
2987.55
2497.04
598.00
3.50
98.50
2282.07
17.50
78.00
119.00
2670.91
1740.60
680.87
1315.53
290.43
700.59
4740.81
4385.68
474.40
987.15
496.21
1224.74
24157.28
498.75
7061.58
service=u -
Natural
Log of
Emission
Rate
(kg/hr)
-12.2993
-12.0780
-12.0052
-11.8618
-11.8106
-11.7331
-11.6918
-11.3419
-11.2652
-11.1933
-11.1097
-11.0057
-10.9499
-10.7491
-10.7203
-10.6229
-10.6101
-10.5434
-10.4642
-10.4186
-10.2381
-10.2338
-10.2171
-10.1383
-10.0739
-10.0310
-10.0179
-9.8566
-9.7942
-9.6932
-9.6456
-9.5321
-9.4594
-9.3923
-9.2141
-9.1413
-9.0813
-9.0299
-8.9163
-8.8906
-8.7929
-8.7245
-8.6680
-8.6416
Natural
Log of
Screening
Value Outlier
(ppmv) Flag
1.6864
2.2513
1.3863
0.7419
2.8622
1.2238
3.6926
3.0184
2.1401
4.4296
8.5077
8.5081
6.9151
6.5489
3.0204
6.7452
4.9733
4.9363
2.7408
4.6913
8.0022
7.8229
6.3936
1.2528
4.5901
7.7328
2.8622
4.3567
4.7791
7.8902
7.4620
6.5234
7.1820
5.6714
6.5519
8.4640
8.3861
6.1621
6.8948
6.2070
7.1105
10.0923
6.2121
8.8624
B-27
-------
Table B-1-1. Bagging data used to develop the correlation equations
equipment iype=v«tvt
(continued)
Plant
Type
BO
RE
RE
RE
RE
RE
RE
RE
RE
RE
BD
RE
RE
BD
EO
RE
EO
BD
RE
BD
RE
RE
BD
RE
RE
RE
RE
RE
RE
RE
RE
RE
BD
BD
RE
RE
RE
RE
RE
BD
RE
BD
EO
RE
Data
Origin
NEW
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
NEW
OLD
OLD
NEW
NEW
OLD
NEW
NEW
OLD
NEW
OLD
OLD
NEW
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
NEW
NEW
OLD
OLD
OLD
OLD
OLD
NEW
OLD
NEW
NEW
OLD
Measured
Emission
Rate (kg/hr)
0.0001866845
0.0001904680
0.0001964120
0.0001977607
0.0002152405
0.0002180108
0.0002232184
0.0002275124
0.0002307162
0.0002322459
0.0002437423
0.0002528838
0.0002757637
0.0002760188
0.0002904846
0.0003425098
0.0003724437
0.0003991030
0.0004050504
0.0004404057
0.0004427801
0.0004461460
0.0004471948
0.0004520589
0.0004529831
0.0004536846
0.0004640417
0.0004685177
0.0004728028
0.0005228957
0.0005323154
0.0005465275
0.0005634682
0.0005651718
0.0005730494
0.0005839129
0.0005991093
0.0006007199
0.0006146615
0.0006404920
0.0006448431
0.0007363507
0.0009188385
0.0009212745
Screening
Value (ppmv)
824.40
1643.51
1423.98
24689.43
1556.44
2095.88
3292.43
6482.10
4804.03
4368.95
499.40
928.66
877.50
6695.10
8998.00
2139.46
9998.00
394.00
9863.86
1999.00
4287.44
18661.82
799.00
55794.96
4949.37
3965.77
560.84
4279.25
14956.09
4399.96
2867.11
16699.10
2999.70
247.00
2037.49
35105.41
246.51
27836.27
1592.14
2743.50
2313.46
1247.00
3448.00
2316.36
aervice=u -
Natural
Log of
Emission
Rate
(kg/hr)
-8.5861
-8.5660
-8.5353
-8.5285
-8.4438
-8.4310
-8.4074
-8.3883
-8.3743
-8.3677
-8.3194
-8.2826
-8.1960
-8.1950
-8.1440
-7.9792
-7.8954
-7.8263
-7.8115
-7.7278
-7.7224
-7.7149
-7.7125
-7.7017
-7.6997
-7.6981
-7.6755
-7.6659
-7.6568
-7.5561
-7.5383
-7.5119
-7.4814
-7.4784
-7.4645
-7.4458
-7.4201
-7.4174
-7.3944
-7.3533
-7.3465
-7.2138
-6.9924
-6.9898
Natural
Log of
Screening
Value Outlier
(ppmv) Flag
6.7147
7.4046
7.2612
10.1141
7.3502
7.6477
8.0994
8.7768
8.4772
8.3823
6.2134
6.8337
6.7771
8.8091
9.1048
7.6683
9.2101
5.9764
9.1966
7.6004
8.3634
9.8342
6.6834
10.9294
8.5070
8.2855
6.3294
8.3615
9.6129
8.3894
7.9611
9.7231
8.0063
5.5094
7.6195
10.4661
5.5074
10.2341
7.3728
7.9170
7.7465
7.1285
8.1455
7.7478
B-28
-------
Table B-1-1. Bagging data used to develop the correlation equations
equipment iype-VALve
(continued)
Plant
Type
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
BD
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
BD
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
Data
Origin
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
NEW
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
NEW
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
Measured
Emission
Rate (kg/hr)
0.0009386789
0.0009859662
0.0011533445
0.0011636438
0.0011668930
0.0011712242
0.0017829290
0.0019401846
0.0020010182
0.0022581253
0.0022870889
0.0025260448
0.0025348896
0.0026295658
0.0027833322
0.0029409798
0.0031312882
0.0031778789
0.0033409352
0.0033838729
0.0036846059
0.0036971583
0.0039426484
0.0039504089
0.0040050325
0.0041065399
0.0041660267
0.0046273787
0.0051511364
0.0060064387
0.0064640997
0.0067947745
0.0086599432
0.0102338821
0.0112479155
0.0150883255
0.0192079955
0.0212769340
0.0262475666
0.0265051976
0.0277367164
0.0342721260
0.0449106195
0.0645502674
Screening
Value (ppmv)
7331 .62
32119.44
2785.34
2797.20
203224.00
21751.69
67504.85
56199.96
8684.64
4284.86
3791 .44
3163.33
534.08
50201.19
20393.42
4530.72
1860.09
4297.80
219611.97
23015.69
17536.22
16495.48
12647.22
34241.04
1333.88
4005.05
2803.86
20516.30
3629.80
760.42
61150.08
102781.04
287461.04
12994.00
9730.32
749143.47
191834.63
29340.67
189629.11
2373.75
820321 .32
90882.86
17031.74
16874.50
aervice=u •
Natural
Log of
Emission
Rate
(kg/hr)
-6.9710
-6.9219
-6.7651
-6.7562
-6.7534
-6.7497
-6.3295
-6.2450
-6.2141
-6.0932
-6.0805
-5.9811
-5.9776
-5.9409
-5.8841
-5.8290
-5.7663
-5.7515
-5.7015
-5.6887
-5.6036
-5.6002
-5.5359
-5.5339
-5.5202
-5.4952
-5.4808
-5.3758
-5.2685
-5.1149
-5.0415
-4.9916
-4.7490
-4.5821
-4.4876
-4.1938
-3.9524
-3.8501
-3.6402
-3.6304
-3.5850
-3.3734
-3.1031
-2.7403
Natural
Log of
Screening
Value Outlier
(ppmv) Flag
8.9000
10.3772
7.9321
7.9364
12.2221
9.9874
11.1200
10.9367
9.0693
8.3628
8.2405
8.0594
6.2805
10.8238
9.9230
8.4186
7.5284
8.3659
12.2996
10.0439
9.7720
9.7108
9.4452
10.4412
7.1958
8.2953
7.9388
9.9290
8.1969
6.6339
11.0211
11.5404
12.5688
9.4722
9.1830
13.5267
12.1644
10.2867
12.1528
7.7722
13.6175
11.4173
9.7428
9.7336
B-29
-------
Table B-1-1. Bagging data used to develop the correlation equations
(continued)
Plant
Type
RE
RE
Data
Origin
OLD
OLD
Measured
Emission
Rate (kg/hr)
0.1109042134
0.1K0677949
Screening
Value (ppmv)
326432.21
20836.56
Natural
Log of
Emission
Rate
(kg/hr)
-2.1991
-2.1710
Natural
Log of
Screening
Value
(ppmv)
12.6960
9.9445
Outlier
Flag
179 (0 outliers)
B-30
-------
Table B-1-1. Bagging data used to develop the correlation equations
Plant
Type
EO
EO
EO
EO
BD
BD
EO
BD
EO
EO
EO
EO
EO
BD
BD
BD
BD
EO
BD
EO
EO
EO
EO
BD
EO
EO
BD
BD
BD
BD
EO
BD
Data
Origin
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
K •
NEW
NEW
NEW
NEW
cquipim
Measured
Emission
Rate (kg/hr)
0.0000001148
0.0000001182
0.0000001490
0.0000001545
0.0000001546
0.0000001705
0.0000001748
0.0000001777
0.0000002092
0.0000002655
0.0000002662
0.0000002674
0.0000002973
0.0000003209
0.0000003246
0.0000003272
0.0000003761
0.0000004160
0.0000004269
0.0000005550
0.0000006711
0.0000006800
0.0000007182
0.0000007281
0.0000007741
0.0000007760
0.0000009403
0.0000009766
0.0000010750
0.0000013768
0.0000014189
0.0000017270
;ni iype-vALvc ;
Screening
Value (ppmv)
2.00
0.40
0.70
7.00
2.00
2.25
13.50
1.50
0.90
24.25
34.00
119.00
1.00
0.25
14.00
145.00
1.00
1.10
2.50
0.60
2.00
1547.50
2.80
1.30
1.85
0.45
2.25
3.25
3.50
6.45
398.00
4.00
>ervice=LL -
Natural
Log of
Emission
Rate
(kg/hr)
-15.9798
-15.9509
-15.7195
-15.6828
-15.6825
-15.5843
-15.5593
-15.5431
-15.3801
-15.1418
-15.1392
-15.1344
-15.0285
-14.9523
-14.9406
-14.9326
-14.7934
-14.6925
-14.6668
-14.4043
-14.2144
-14.2011
-14.1465
-14.1328
-14.0715
-14.0691
-13.8770
-13.8391
-13.7432
-13.4957
-13.4656
-13.2691
Natural
Log of
Screening
Value Outlier
(ppmv) Flag
0.6931
-0.9163
-0.3567
1 .9459
0.6931
0.8109
2.6027
0.4055
-0.1054
3.1884
3.5264
4.7791
0.0000
-1.3863
2.6391
4.9767
0.0000
0.0953
0.9163
-0.5108
0.6931
7.3444
1.0296
0.2624
0.6152
-0.7985
0.8109
1.1787
1.2528
1.8641
5.9865
1.3863
B-31
-------
Table B-1-1. Bagging data used to develop the correlation equations
(continued)
Plant
Type
BO
EO
BD
EO
EO
RE
BD
BD
BD
BD
BD
EO
EO
BD
EO
EO
BD
RE
EO
EO
BD
BD
BD
BD
BD
EO
EO
BD
EO
RE
BD
BD
BD
EO
RE
EO
BD
BD
RE
RE
RE
BD
BD
EO
Data
Origin
NEW
NEW
NEW
NEW
NEW
OLD
NEU
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
OLD
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
NEW
OLD
NEW
NEW
NEW
NEW
OLD
NEW
NEW
NEW
OLD
OLD
OLD
NEW
NEW
NEW
Measured
Emission
Rate (kg/he)
0.0000021600
0.0000026370
0.0000026381
0.0000028522
0.0000031653
0.0000032615
0.0000034734
0.0000034854
0.0000036357
0.0000036487
0.0000038172
0.0000038185
0.0000045401
0.0000048429
0.0000053288
0.0000054257
0.0000054590
0.0000061236
0.0000063620
0.0000076923
0.0000079625
0.0000080291
0.0000081895
0.0000087183
0.0000090393
0.0000096017
0.0000106063
0.0000114056
0.0000116662
0.0000118300
0.0000123249
0.0000130315
0.0000136318
0.0000138914
0.0000150006
0.0000150217
0.0000150810
0.0000155478
0.0000185551
0.0000191256
0.0000196624
0.0000200735
0.0000212478
0.0000226439
Screening
Value (ppmv)
209.00
6.70
18.50
51.20
21.80
2740.82
13.50
486.75
1.40
3.05
0.20
45.00
21.50
21.50
11.40
1.00
44.90
5194.17
4.80
30.00
195.50
20.85
17.75
0.25
7.00
0.90
29.00
2.40
90.00
97.72
21.90
20.00
49.80
39.40
500.63
108.00
32.50
54.50
78.10
191501.42
4878.72
250.00
67.00
44.10
>ervice=LL -
Natural
Log of
Emission
Rate
(kg/hr)
-13.0454
-12.8459
-12.8455
-12.7674
-12.6633
-12.6333
-12.5704
-12.5669
-12.5247
-12.5211
-12.4760
-12.4756
-12.3026
-12.2380
-12.1424
-12.1244
-12.1182
-12.0034
-11.9652
-11.7753
-11.7408
-11.7324
-11.7127
-11.6501
-11.6139
-11.5536
-11.4541
-11.3814
-11.3588
-11.3449
-11.3039
-11.2481
-11.2031
-11.1842
-11.1074
-11.1060
-11.1021
-11.0716
-10.8948
-10.8645
-10.8368
-10.8161
-10.7593
-10.6956
Natural
Log of
Screening
Value
(ppmv)
5.3423
1.9021
2.9178
3.9357
3.0819
7.9160
2.6027
6.1878
0.3365
1.1151
-1.6094
3.8067
3.0681
3.0681
2.4336
0.0000
3.8044
8.5553
1.5686
3.4012
5.2756
3.0374
2.8764
-1.3863
1 .9459
-0.1054
3.3673
0.8755
4.4998
4.5821
3.0865
2.9957
3.9080
3.6738
6.2159
4.6821
3.4812
3.9982
4.3580
12.1627
8.4926
5.5215
4.2047
3.7865
Outlier
Flag
OUTLIER
B-32
-------
Table B-1-1. Bagging data used to develop the correlation equations
equipment iype=v«Lvc :
(continued)
Plant
Type
EO
EO
BD
RE
BD
EO
EO
BD
BD
RE
EO
EO
RE
EO
BD
BD
RE
BD
BD
BD
RE
RE
EO
RE
RE
EO
EO
EO
RE
RE
RE
RE
EO
RE
EO
BD
BD
BD
EO
BD
RE
BD
BD
RE
Data
Origin
NEW
NEW
NEW
OLD
NEW
NEW
NEW
NEW
NEW
OLD
NEW
NEW
OLD
NEW
NEW
NEW
OLD
NEW
NEW
NEW
OLD
OLD
NEW
OLD
OLD
NEW
NEW
NEW
OLD
OLD
OLD
OLD
NEW
OLD
NEW
NEW
NEW
NEW
NEW
NEW
OLD
NEW
NEW
OLD
Measured
Emission
Rate (kg/hr)
0.0000228716
0.0000242425
0.0000244394
0.0000269514
0.0000298536
0.0000301615
0.0000330901
0.0000336994
0.0000354699
0.0000378083
0.0000382742
0.0000383797
0.0000387557
0.0000387574
0.0000407202
0.0000415953
0.0000417925
0.0000429883
0.0000443510
0.0000462778
0.0000470621
0.0000482670
0.0000508340
0.0000529921
0.0000546755
0.0000561055
0.0000569507
0.0000626293
0.0000626636
0.0000654535
0.0000660567
0.0000664281
0.0000713497
0.0000749810
0.0000778658
0.0000893438
0.0000936958
0.0001029548
0.0001063538
0.0001147397
0.0001266782
0.0001377292
0.0001972580
0.0002313295
Screening
Value (ppmv)
74.80
2.40
35.50
5443.31
298.90
148.00
59.25
92.50
28.50
604.46
657.80
243.60
242.12
48.90
29.00
1349.80
42609.46
248.00
99.00
1.75
906.10
10833.21
79.00
890.55
1193.53
348.00
60.00
163.70
1985.67
318.60
5226.31
4914.24
343.00
1458.90
148.50
350.00
199.75
872.75
148.75
499.50
1183.21
73.00
174.75
50044.57
>ervice=LL -
Natural
Log of
Emission
Rate
(kg/hr)
-10.6856
-10.6274
-10.6193
-10.5215
-10.4192
-10.4089
-10.3163
-10.2980
-10.2468
-10.1830
-10.1707
-10.1680
-10.1582
-10.1582
-10.1088
-10.0875
-10.0828
-10.0546
-10.0234
-9.9808
-9.9640
-9.9388
-9.8869
-9.8454
-9.8141
-9.7883
J.7733
-9.6783
-9.6777
-9.6342
-9.6250
-9.6194
-9.5479
-9.4983
-9.4605
-9.3230
-9.2755
-9.1812
-9. 1487
-9.0728
-8.9739
-8.8902
-8.5310
-8.3717
Natural
Log of
Screening
Value Outlier
(ppmv) Flag
4.3148
0.8755
3.5695
8.6021
5.7001
4.9972
4.0818
4.5272
3.3499
6.4043
6.4889
5.4955
5.4894
3.8898
3.3673
7.2077
10.6598
5.5134
4.5951
0.5596
6.8091
9.2904
4.3694
6.7918
7.0847
5.8522
4.0943
5.0980
7.5937
5.7639
8.5615
8.4999
5.8377
7.2854
5.0006
5.8579
5.2971
6.7716
5.0023
6.2136
7.0760
4.2905
5.1634
10.8207
B-33
-------
Table B-1-1. Bagging data used to develop the correlation equations
(continued)
Plant
Type
BO
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
BO
RE
RE
BD
RE
RE
RE
BD
RE
RE
RE
BD
RE
RE
RE
RE
RE
BD
RE
RE
RE
RE
RE
RE
RE
RE
RE
Data
Origin
NEW
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
NEW
OLD
OLD
NEW
OLD
OLD
OLD
NEW
OLD
OLD
OLD
NEW
OLD
OLD
OLD
OLD
OLD
NEW
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
Measured
Emission
Rate (kg/hr)
0.0002317965
0.0002524777
0.0002580228
0.0002594664
0.0002714139
0.0002825941
0.0002947841
0.0003011106
0.0003056054
0.0003367527
0.0003494725
0.0003655199
0.0003726697
0.0003738730
0.0003743390
0.0003964414
0.0004653107
0.0004698821
0.0004809845
0.0004922594
0.0005246367
0.0005251847
0.0005308943
0.0005614771
0.0005705547
0.0006267770
0.0006426106
0.0006597100
0.0006830173
0.0007019466
0.0007129023
0.0007649183
0.0007702967
0.0008350761
0.0008369235
0.0008536995
0.0008577230
0.0009616788
0.0010351161
0.0010736310
0.0012337497
0.0012793343
0.0013448227
0.0013933013
Screening
Value (ppmv)
180.00
12405.49
44328.29
510.60
185.88
6976.92
1516.43
44592.42
181.92
88.38
1041.01
17367.57
856.19
8088.28
1959.19
4048.28
35414.65
1543.75
4284.78
104088.32
2645.50
1151.37
14765.02
97.30
358.30
1565.55
5861.53
1793.09
94.75
8827.10
9940.79
25559.24
14.18
1281.36
6097.00
2810.09
6709.07
46673.57
71798.27
3136.03
8519.07
16658.85
962.89
1602.40
>ervice=LL -
Natural
Log of
Emission
Rate
(kg/hr)
-8.3697
-8.2842
-8.2625
-8.2569
-8.2119
-8.1715
-8.1293
-8.1080
-8.0932
-7.9962
-7.9591
-7.9142
-7.8948
-7.8916
-7.8903
-7.8330
-7.6728
-7.6630
-7.6397
-7.6165
-7.5528
-7.5518
-7.5409
-7.4849
-7.4689
-7.3749
-7.3500
-7.3237
-7.2890
-7.2617
-7.2462
-7.1757
-7.1687
-7.0880
-7.0858
-7.0659
-7.0612
-6.9468
-6.8732
-6.8367
-6.6977
-6.6614
-6.6115
-6.5761
Natural
Log of
Screening
Value Outlier
(ppmv) Flag
5.1930
9.4259
10.6994
6.2356
5.2251
8.8504
7.3241
10.7053
5.2036
4.4816
6.9479
9.7624
6.7525
8.9982
7.5803
8.3060
10.4749
7.3420
8.3628
11.5530
7.8806
7.0487
9.6000
4.5778
5.8814
7.3560
8.6762
7.4917
4.5512
9.0856
9.2044
10.1488
2.6518
7.1557
8.7156
7.9410
8.8112
10.7509
11.1816
8.0507
9.0501
9.7207
6.8699
7.3793
B-34
-------
Table B-1-1. Bagging data used to develop the correlation equations
Equipment iype=VALVt :
(continued)
Plant
Type
RE
RE
RE
RE
RE
RE
BO
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
BO
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
BD
RE
RE
RE
RE
Data
Origin
OLD
OLD
OLD
OLD
OLD
OLD
NEW
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
NEW
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
NEW
OLD
OLD
OLD
OLD
Measured
Emission
Rate (kg/hr)
0.0014732045
0.0016009142
0.0018373887
0.0018697565
0.0021076721
0.0022196068
0.0023716142
0.0026041383
0.0026564280
0.0030068935
0.0030297587
0.0032025436
0.0032489277
0.0032868739
0.0034814651
0.0034830527
0.0035502018
0.0036059944
0.0037109239
0.0037115648
0.0038957946
0.0038969686
0.0039248950
0.0040089261
0.0042596218
0.0043498677
0.0043951332
0.0046094493
0.0046247477
0.0046555934
0.0047542941
0.0049436538
0.0049687260
0.0055770694
0.0059962681
0.0066867186
0.0073478291
0.0076182294
0.0078722531
0.0079621021
0.0095095298
0.0102176741
0.0105761365
0.0126755860
Screening
Value (ppmv)
22177.98
22172.87
1769.15
25877.90
93629.13
4376.80
1495.00
1313.08
52084.68
45068.90
6771.42
9836.80
140865.29
134149.17
284948.25
59618.63
4839.96
5555.74
72002.57
24755.46
9810.65
1544.40
7476.44
13953.59
30597.64
2026.05
4587.13
73036.68
2875.27
3279.62
5891.43
2135.71
9436.54
80485.19
19368.05
28552.82
129657.01
194.63
3118.82
9500.00
2553.37
44254.56
20652.95
960160.86
>ervice=LL -
Natural
Log of
Emission
Rate
(kg/hr)
-6.5203
-6.4372
-6.2994
-6.2819
-6.1622
-6.1104
-6.0442
-5.9507
-5.9308
-5.8068
-5.7993
-5.7438
-5.7294
-5.7178
-5.6603
-5.6598
-5.6408
-5.6252
-5.5965
-5.5963
-5.5479
-5.5476
-5.5404
-5.5192
-5.4586
-5.4376
-5.4273
-5.3796
-5.3763
-5.3697
-5.3487
-5.3097
-5.3046
-5.1891
-5.1166
-5.0076
-4.9134
-4.8772
-4.8444
-4.8331
-4.6555
-4.5836
-4.5492
-4.3681
Natural
Log of
Screening
Value Outlier
(ppmv) Flag
10.0069
10.0066
7.4783
10.1611
11.4471
8.3841
7.3099
7.1801
10.8606
10.7159
8.8205
9.1939
11.8556
11.8067
12.5601
10.9957
8.4847
8.6226
11.1845
10.1168
9.1912
7.3424
8.9195
9.5435
10.3287
7.6138
8.4310
11.1987
7.9639
8.0955
8.6813
7.6666
9.1523
11.2958
9.8714
10.2595
11.7726
5.2711
8.0452
9.1590
7.8452
10.6977
9.9356
13.7749
B-35
-------
Table B-1-1. Bagging data used to develop the correlation equations
(continued)
Plant
Type
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
RE
Data
Origin
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
OLD
Measured
Emission
Rate (kg/hr)
0.0128994159
0.0134752877
0.0137156706
0.0190054451
0.0194889771
0.0220373843
0.0220386022
0.0221003955
0.0248459751
0.0254155227
0.0263386824
0.0272779071
0.0283621432
0.0283930499
0.0293848208
0.0303470196
0.0305360632
0.0372725448
0.0410821388
0.0468639667
0.0687821973
0.0713743302
0.0838252864
0.1027415340
0.2448798474
Screening
Value (ppmv)
301945.80
28558.21
114.30
1649.34
518201.90
213772.09
7980.81
362645.26
9843.83
41862.00
659517.01
1399.25
288.41
352.85
480.98
562236.45
21853.55
122666.22
62573.58
393961.70
49473.43
36751.32
360547.09
53569.80
371111.15
>ervice=LL -
Natural
Log of
Emission
Rate
(kg/hr)
-4.3506
-4.3069
-4.2892
-3.9630
-3.9379
-3.8150
-3.8150
-3.8122
-3.6951
-3.6724
-3.6367
-3.6017
-3.5627
-3.5616
-3.5273
-3.4951
-3.4888
-3.2895
-3.1922
-3.0605
-2.6768
-2.6398
-2.4790
-2.2755
-1.4070
Natural
Log of
Screening
Value Outlier
(ppmv) Flag
12.6180
10.2597
4.7388
7.4081
13.1581
12.2727
8.9848
12.8012
9.1946
10.6421
13.3993
7.2437
5.6644
5.8660
6.1758
13.2397
9.9921
11.7172
11.0441
12.8840
10.8092
10.5119
12.7954
10.8887
12.8243
N = 233 (1 outliers)
B-36
-------
Table B-l-2. Comparison of regression results for the old, new, and
combined bagging data sets.
Equipment Type/Service: Connectors/All
Statistical Parameter
Data Used in Regression
New
Number of data pairs
Regression intercept
Regression slope
Regression R
Regression correlation coefficient
Regression mean square error
Regression root mean square error
Average In screening value
Sum of squares of In screening values
Scale bias correction factor
Correlation equation constant
107
-14.815
0.885
0.525
0.725
4.355
2.087
3.472
646.821
8.298
3.05E-6
Equipment Type/Service: Pumps/Light Liquid
Data Used in Regression
Statistical Parameter
Olda
New
Combined
Number of data pairs
Regression intercept
Regression slope
Regression R
Regression correlation coefficient
Regression mean square error
Regression root mean square error
Average In screening value
Sum of squares of In screening values
Scale bias correction factor
Correlation equation constant
51
-12.827
0.865
0.613
0.783
2.246
1.499
8.582
233.223
2.941
7.91E-6
68
-12.515
0.907
0.644
0.803
3.783
1.945
5.393
548.793
6.149
2.26E-5
117
-12.142
0.824
0.710
0.842
2.591
1.610
6.783
1071.500
3.563
1.90E-5
a Indicates that the parameter were derived from the
digitized data pairs for the OLD regression.
B-37
-------
Table B-l-2. (continued)
Equipment Type/Service: Valves/Gas
Data Used in Regression
Statistical Parameter
Number of data pairs
Regression intercept
Regression slope
Regression R
Regression correlation coefficient
Regression mean square error
Regression root mean square error
Average In screening value
Sum of squares of In screening values
Scale bias correction factor
Correlation equation constant
Olda
95
-12.848
0.661
0.359
0.599
2.767
1.663
8.823
329.550
3.858
1.02E-5
New
84
-14.936
0.750
0.516
0.711
4.392
2.096
3.691
682.442
8.311
2.71E-6
Combined
179
-15.033
0.873
0.715
0.846
3.745
1.935
6.415
2186.020
6.315
1.87E-6
a Indicates that the parameter were derived from the
digitized data pairs for the OLD regression.
Equipment Type/Service: Valves/Light Liquid
Data Used in Regression
Statistical Parameter
Number of data pairs
Regression intercept
Regression slope
Regression R
Regression correlation coefficient
Regression mean square error
Regression root mean square error
Average In screening value
Sum of squares of In screening values
Scale bias correction factor
Correlation equation constant
Olda
126
-10.585
0.452
0.194
0.441
4.413
2.101
8.978
644.683
8.608
2.18E-4
New
107
-14.137
0.721
0.502
0.709
3.115
1.765
3.300
633.647
4.580
3.32E-6
Combined
232
-13.975
0.797
0.677
0.823
4.088
2.022
6.345
3110.310
7.520
6.41E-6
a indicates that the parameter were derived from the
digitized data pairs for the OLD regression
B-38
-------
APPENDIX B: ATTACHMENT 2
This attachment lists the data used to develop the default-zero
emission leak rates in Table B-2-1. Table B-2-2 lists summary
information on the default-zero development.
B-39
-------
Table B-2-1. Data used for default zero calculations.
PLTJTYPE
EO
EO
EO
EO
EO
EO
BD
BD
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
BD
EO
EO
BD
EO
BD
BD
BD
BD
BD
BD
BD
BD
BD
-Equipment Type=CONNECTORS Service=ALL
Natural Log
Mass of Mass
Emission Emission
Rate (kg/hr) Rate (kg/hr
Screening
Value (ppmv)
0.00 0.0000000475
0.00 0.0000000608
0.00 0.0000000613
0.00 0.0000000790
0.00 0.0000000988
0.00 0.0000001027
0.00 0.0000001033
0.00 0.0000001037
0.00 0.0000001065
0.00 0.0000001079
0.00 0.0000001085
0.00 0.0000001089
0.00 0.0000001112
0.00 0.0000001113
0.00 0.0000001115
0.00 0.0000001120
0.00 0.0000001125
0.00 0.0000001133
0.00 0.0000001146
0.00 0.0000001146
0.00 0.0000001150
0.00 0.0000001166
0.00 0.0000001176
0.00 0.0000001177
0.00 0.0000001178
0.00 0.0000001181
0.00 0.0000001189
0.00 0.0000001213
0.00 0.0000001234
0.00 0.0000001240
0.00 0.0000001296
0.00 0.0000001320
0.00 0.0000001349
0.00 0.0000001376
0.00 0.0000001390
0.00 0.0000001412
0.00 0.0000001413
0.00 0.0000001440
0.00 0.0000001446
0.00 0.0000001448
0.00 0.0000001454
0.00 0.0000001455
0.00 0.0000001485
0.00 0.0000001490
0.00 0.0000001497
0.00 0.0000001505
-16.86331619
-16.61499543
-16.60715372
-16.35377339
-16.13056673
-16.09179287
-16.08517422
-16.08139097
-16.05508510
-16.04208307
-16.03689892
-16.03320436
-16.01231281
-16.01113856
-16.00911113
-16.00437388
-16.00075170
-15.99300732
-15.98221965
-15.98146212
-15.97834935
-15.96444127
-15.95559511
-15.95545662
-15.95391595
-15.95192362
-15.94478891
-15.92488652
-15.90745448
-15.90308275
-15.85882804
-15.84081663
-15.81855266
-15.79862472
-15.78899513
-15.77318199
-15.77244897
-15.75326730
-15.74929429
-15.74817023
-15.74382504
-15.74329360
-15.72271562
-15.71949421
-15.71483698
-15.70909501
B-40
-------
Table B-2-1. Data used for default zero calculations.
Equipment Type=CONNECTORS Service=ALL
(continued)
Natural Log
Mass of Mass
Screening Emission Emission
PLT_TYPE Value (ppmv) Rate (kg/hr) Rate (kg/hr
EO
EO
BD
EO
BD
BD
BD
BD
BD
BD
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
BD
BD
EO
EO
EO
BD
BD
BD
EO
EO
EO
EO
EO
EO
EO
BD
EO
BD
EO
EO
0.00 0.0000001511
0.00 0.0000001544
0.00 0.0000001547
0.00 0.0000001563
0.00 0.0000001573
0.00 0.0000001574
0.00 0.0000001596
0.00 0.0000001614
0.00 0.0000001621
0.00 0.0000001625
0.00 0.0000001631
0.00 0.0000001636
0.00 0.0000001641
0.00 0.0000001642
0.00 0.0000001648
0.00 0.0000001648
0.00 0.0000001650
0.00 0.0000001650
0.00 0.0000001651
0.00 0.0000001657
0.00 0.0000001657
0.00 0.0000001660
0.00 0.0000001688
0.00 0.0000001692
0.00 0.0000001717
0.00 0.0000001741
0.00 0.0000001747
0.00 0.0000001750
0.00 0.0000001807
0.00 0.0000001812
0.00 0.0000001904
0.00 0.0000001920
0.00 0.0000001932
0.00 0.0000001990
0.00 0.0000002086
0.00 0.0000002194
0.00 0.0000002431
0.00 0.0000002476
0.00 0.0000002508
0.00 0.0000002570
0.00 0.0000002585
0.00 0.0000002593
0.00 0.0000002594
0.00 0.0000002602
0.00 0.0000002607
-15.70514515
-15.68403336
-15.68204363
-15.67144879
-15.66508859
-15.66465227
-15.65073157
-15.63962500
-15.63500235
-15.63229582
-15.62914831
-15.62557049
-15.62273582
-15.62198449
-15.61837621
-15.61837621
-15.61705986
-15.61705962
-15.61656953
-15.61295101
-15.61295101
-15.61112981
-15.59463081
-15.59241662
-15.57752890
-15.56347827
-15.56001908
-15.55828552
-15.52620814
-15.52341721
-15.47417798
-15.46559058
-15.45958528
-15.43018880
-15.38283699
-15.33220908
-15.22964242
-15.21159451
-15.19874994
-15.17423032
-15.16823490
-15.16532554
-15.16500428
-15.16174131
-15.15994436
B-41
-------
Table B-2-1. Data used for default zero calculations.
Equipment Type=CONNECTORS Service=ALL
(continued)
Natural Log
Mass of Mass
Screening Emission Emission
PLTJTYPE Value (ppmv) Rate (kg/hr) Rate (kg/hr
EO
EO
EO
EO
BD
EO
BD
EO
BD
BD
BD
BD
BD
BD
BD
BD
BD
BD
BD
EO
BD
EO
BD
EO
EO
EO
EO
EO
BD
EO
EO
EO
EO
EO
BD
EO
EO
EO
EO
BD
EO
BD
BD
BD
BD
0.00 0.0000002626
0.00 0.0000002626
0.00 0.0000002659
0.00 0.0000002664
0.00 0.0000002959
0.00 0.0000003055
0.00 0.0000003140
0.00 0.0000003276
0.00 0.0000003303
0.00 0.0000003315
0.00 0.0000003346
0.00 0.0000003436
0.00 0.0000003436
0.00 0.0000003442
0.00 0.0000003461
0.00 0.0000003504
0.00 0.0000003672
0.00 0.0000003946
0.00 0.0000004121
0.00 0.0000004133
0.00 0.0000004212
0.00 0.0000004468
0.00 0.0000004720
0.00 0.0000005089
0.00 0.0000005180
0.00 0.0000005187
0.00 0.0000005908
0.00 0.0000006166
0.00 0.0000006960
0.00 0.0000007110
0.00 0.0000007192
0.00 0.0000008267
0.00 0.0000009572
0.00 0.0000010002
0.00 0.0000010065
0.00 0.0000010071
0.00 0.0000011795
0.00 0.0000011927
0.00 0.0000021315
0.00 0.0000023492
0.00 0.0000024557
0.00 0.0000024895
0.00 0.0000025620
0.00 0.0000030901
0.00 0.0000033269
-15.15272411
-15.15272411
-15.13996186
-15.13812531
-15.03330632
-15.00115460
-14.97386313
-14.93133352
-14.92340849
-14.91955531
-14.91035517
-14.88372774
-14.88368692
-14.88192105
-14.87648133
-14.86410580
-14.81747447
-14.74527193
-14.70207785
-14.69904106
-14.68010001
-14.62113094
-14.56621062
-14.49108397
-14.47320006
-14.47197698
-14.34186784
-14.29899587
-14.17794549
-14.15652787
-14.14510177
-14.00581175
-13.85929011
-13.81535039
-13.80901606
-13.80841513
-13.65045667
-13.63931593
-13.05868377
-12.96141917
-12.91711588
-12.90342759
-12.87473675
-12.68731235
-12.61346713
B-42
-------
Table B-2-1. Data used for default zero calculations.
Equipment Type=CONNECTORS Service=ALL
(continued)
Natural Log
Mass of Mass
Screening Emission Emission
PLTJTYPE Value (ppmv) Rate (kg/hr) Rate (kg/hr
BD
BD
BD
BD
BD
BD
BD
BD
BD
BD
N = 146
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Equi
0.0000037589
0.0000040185
0.0000042414
0.0000044626
0.0000066833
0.0000075709
0.0000105577
0.0000144776
0.0000154005
0.0000165494
-12.49138454
-12.42460572
-12.37062573
-12.31978282
-11.91589131
-11.79119727
-11.45865639
-11.14290744
-11.08111125
-11.00916328
Equipment Type=PUMP Service=LL
PLTJTYPE
EO
EO
BD
BD
BD
BD
EO
BD
Screening
Value (ppmv)
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Mass
Emission
Rate (kg/hr)
0.0000002532
0.0000002674
0.0000003397
0.0000006493
0.0000013801
0.0000031715
0.0000061497
0.0000978267
Natural Log
of Mass
Emission
Rate (kg/hr
-15.18920187
-15.13444207
-14.89520337
-14.24738145
-13.49334976
-12.66130995
-11.99910617
-9.232313175
N = 8
B-43
-------
Table B-2-1. Data used for default zero calculations.
Equipment Type=VALVE Service=G
PLT_TYPE
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
BD
BD
EO
BD
EO
BD
BD
EO
EO
BD
EO
BD
EO
BD
BD
EO
EO
BD
EO
BD
EO
BD
BD
BD
BD
BD
BD
BD
N = 40
Screening
Value (ppmv)
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Mass
Emission
Rate (kg/hr)
0.0000000591
0.0000000722
0.0000000737
0.0000000786
0.0000000790
0.0000000796
0.0000001079
0.0000001081
0.0000001083
0.0000001312
0.0000001321
0.0000001325
0.0000001382
0.0000001436
0.0000001446
0.0000001516
0.0000001581
0.0000001595
0.0000001602
0.0000001750
0.0000002350
0.0000002539
0.0000002612
0.0000002633
0.0000002674
0.0000003272
0.0000003339
0.0000003878
0.0000004091
0.0000004607
0.0000006457
0.0000007014
0.0000009932
0.0000009955
0.0000022122
0.0000022562
0.0000025712
0.0000033699
0.0000044219
0.0000106176
Natural Log
of Mass
Emission
Rate (kg/hr
•16.64400086
-16.44327301
-16.42283692
-16.35920326
-16.35376554
-16.34647953
-16.04237697
-16.04053084
-16.03863245
-15.84631356
-15.83996505
-15.83639998
-15.79429751
-15.75651804
-15.74956966
-15.70207714
-15.65972752
-15.65122577
-15.64710329
-15.55828552
-15.26347692
-15.18638489
-15.15814418
-15.14979281
-15.13444207
-14.93266093
-14.91228255
-14.76283680
-14.70928502
-14.59056027
-14.25286952
-14.17014032
-13.82235860
-13.81999480
-13.02153380
-13.00184573
-12.87114036
-12.60062417
-12.32894306
-11.45299698
B-44
-------
Table B-2-1. Data used for default zero calculations.
PLTJTYPE
EO
EO
EO
EO
EO
BD
BD
BD
BD
BD
EO
EO
EO
EO
EO
EO
EO
EO
EO
EO
BD
EO
EO
BD
BD
EO
BD
EO
EO
EO
EO
EO
EO
EO
EO
BD
BD
BD
BD
BD
BD
EO
EO
EO
EO
BD
Equipment Type=VALVE Service=LL
Screening
Value (ppmv)
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Mass
Emission
Rate (kg/hr)
0.0000001121
0.0000001173
0.0000001211
0.0000001229
0.0000001337
0.0000001440
0.0000001461
0.0000001498
0.0000001503
0.0000001513
0.0000001642
0.0000001644
0.0000001644
0.0000001645
0.0000001648
0.0000001654
0.0000001656
0.0000001657
0.0000001660
0.0000001663
0.0000001669
0.0000001758
0.0000001758
0.0000001780
0.0000001804
0.0000001827
0.0000001853
0.0000002507
0.0000002568
0.0000002623
0.0000002645
0.0000002654
0.0000002657
0.0000002664
0.0000002750
0.0000002786
0.0000002807
0.0000002831
0.0000003292
0.0000003296
0.0000003327
0.0000003803
0.0000003997
0.0000004350
0.0000004933
0.0000005121
Natural Log
of Mass
Emission
Rate (kg/hr
-16.00352165
-15.95857877
-15.92634574
-15.91229458
-15.82756192
-15.75311308
-15.73913742
-15.71376221
-15.71042334
-15.70424314
-15.62246991
-15.62066973
-15.62066973
-15.62017964
-15.61837621
-15.61475957
-15.61343643
-15.61294634
-15.61112981
-15.60930997
-15.60596798
-15.55382679
-15.55382679
-15.54144504
-15.52802656
-15.51543605
-15.50155175
-15.19885548
-15.17511567
-15.15362868
-15.14545135
-15.14208066
-15.14094135
-15.13812531
-15.10635430
-15.09348218
-15.08603323
-15.07737541
-14.92670035
-14.92525863
-14.91592554
•14.78222371
-14.73266021
-14.64784669
-14.52205744
•14.48467228
B-45
-------
Table B-2-1. Data used for default zero calculations.
(continued)
Equipment Type=VALVE Service=LL
PLT_TYPE
EO
BD
BD
EO
BD
BD
EO
N = 53
Screening
Value (ppmv)
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Mass
Emission
Rate (kg/hr)
0.0000007099
0.0000011219
0.0000022380
0.0000028444
0.0000041389
0.0000053490
0.0000121637
Natural Log
of Mass
Emission
Rate (kg/hr
-14.15820731
-13.70046348
-13.00992148
-12.77016392
-12.39507152
-12.13860411
-11.31705756
B-46
-------
f
+»
in
CD
2
^
Q_
UJ
1
|
-o
c
*
"o
8
o
^^
^
O-
UJ
1o
c
.?
J?
4-*
It-
co
0)
4->
g
n
i5
w
M
£
Q
0)
's
**-
o
1.
o
o
.
+j
«U
i1
I
CD
1
UJ
UJ
a£
"
14-
10
*5
CO
0)
cc
Ol
III
I*-5
1. 01
01 X O
KI in c
en o> *>
? fccl
K) O. C — 1
H- D. O
0) =3 U
0)
X 0
m c
en 01 4->
iil
o o
— I U
0 5
^ ^T
10 ^~ *^ oi
Jn? ^"^
"Slg
S § u
£55
u u
cu o) 10
r- 1- U.
10 I-
0 0
CO U
CO
>t- O
0 -^
t- 10
01 >
1 Si
O
01
O 4->
0) OC
C c-
TJ 4-> o JT
O *3 ^> O>
IO 10 -*
O UJ
0)
o
Ol
in
^
c
Is
F
CO CM
»H CO
o d
r^ in
i i
UJ UJ
in r-<
^f •— *
^ »
r— co
i I
UJ UJ
I
S K
CM <*
CO 00
^H
ir> m
i i
UJ UJ
^- »-t
PO O5
o> ro
O
UJ
z a.
§ §
U a.
—1 O>
CO O
0 0
r«<. ^
i I
UJ UJ
r-. CM
CD «r
ai co
rv r«k
i I
UJ UJ
in F-.
CO CO
^- CO
i i
UJ UJ
Sin
CO
CO T
o> m
-H CO
CM ^4
O CO
^- in
in ^^
i i
UJ UJ
^H CM
co in
CO ^>
CO —1
UJ UJ
5 5
n
I
1
o
u
Sc
oS
*J CO
"« 0)
3 CO
M t-
|JI
10 t—
tn
^
= 4^
«
^e
n t-
£ o
u u
•a ai
l^
* i
10 a.
JZ Q.
Ol 0) CO
= 5S
10 o 15
at c
c •»- cn
C ^3 *^
ft) t- (U
•^ IM
01
01 ** 4->
»— t- 3
IO 10
O 01
U>— "D
= ss
c •»• a>
^.
U V) 0)
CO CO 4->
"o -^
U 0 4->
01 >- CO
N 01 01
N
4-> O
3 "3 -a
>t- a ai
01 >t- CO
•D 0) 3
s -o
01 01 c.
iE£5!
B-47
-------
APPENDIX B: ATTACHMENT 3
This attachment summarizes information on each of the
screening data sets. Table B-3-1 summarizes data used to revise
the SOCMI emission factors. Figures B-3-1 through B-3-4 plot the
screening value distributions for each data set.
B-48
-------
Connectors
% of Sources
c
(0
r
L
0
s
in
(D
u
L
1-
0
100
1000
10000
100000
1000DOO
Screening Values
D 24 Unit -f EO/BD 0 Corrbtned
Figure B-3-1. Connectors
B-49
-------
c
id
r
U)
u
L
3
0
w
1-
0
as
Light Liquid Pumps
% of Sources
10 -
100
1000
10000
100000
1000000
Screen i ng VaIues (ppmv)
D 24 Unit + EO/BD 0 Combined
Figure B-3-2. Light Liquid Pumps
B-50
-------
Gas Valves
% of Sources
c
(0
£
L
0
w
.c
in
Q)
u
L
a
o
w
1-
0
10 100 1000 1QOQD 10QQQD
Screening Values Cppmv)
D 24 Unit + EO/BD <> Combined
1QQQOOO
Figure B-3-3. Gas Valves
B-51
-------
c
(0
.c
L
0
5
ID
U
3
0
I/)
Light Liquid Va ves
% of Sources
100
1000
10000
100000 1000000
Screening Values (pprnv)
D 2A Unit + EO/BD 0 Combined
Figure B-3-4. Light Liquid Valves
B-52
-------
Table B-3-1. Emission Factors Calculated From Revised SOCMI
Correlation Equations.
Screening
data set
24 UNIT
24 UNIT
24 UNIT
24 UNIT
EO/BD
EO/BD
EO/BD
EO/BD
COMBINED
COMBINED
COMBINED
COMBINED
Equipment
type
CONNEC
PUMP
VALVE
VALVE "
CONNEC
PUMP
VALVE
VALVE
CONNEC
PUMP
VALVE
VALVE
Phase
ALL
LL
G
LL
ALL
LL
G
LL
ALL
LL
G
LL
old
emission
factor
(kg/hr)
8.30E-04
4.94E-02
5.60E-03
7.10E-03
8.30E-04
4.94E-02
5.60E-03
7.10E-03
8.30E-04
4.94E-02
5.60E-03
7.10E-03
Total
number of
screening
values
4,283
646
9,669
18,300
3,562
252
6,507
15,810
7,845
898
16,176
34,110
Average
nonzero
emission
rate
(kg/hr)
2.50E-02
5.36E-02
2.47E-02
2.99E-02
3.76E-04
7.12E-03
2.83E-03
3.26E-03
5.28E-03
3.73E-02
1.75E-02
1.46E-02
Number of
zero
screening
values
3,740
335
5,962
14,292
1,381
85
4,685
10,429
5,121
420
10,647
24,721
Default
zero
emission
rate
(kg/hr)
6.12E-07
7.45E-06
6.56E-07
4.85E-07
6.12E-07
7.45E-06
6.56E-07
4.85E-07
6.12E-07
7.45E-06
6.56E-07
4.85E-07
Average
emission
factor
(kg/hr)
3.16E-03
2.58E-02
9.45E-03
6.55E-03
2.30E-04
4.72E-03
7.92E-04
1.11E-03
1.83E-038
1.99E-028
5.97E-038
4.03E-038
8 These average emission factors are the revised SOCMI average emission factors.
B-53
-------
APPENDIX C: RESPONSE FACTORS
-------
APPENDIX C
RESPONSE FACTORS
The response factors presented in Table C-l were taken from
two separate sources. The response factors at an actual
concentration of 10,000 ppmv are from the EPA document entitled,
"Response Factors of VOC Analyzers Calibrated with Methane for
Selected Organic Chemicals," EPA-600/2-81-002 (September 1980).
The document presents results of analytical tests performed to
determine the response factors at 10,000 ppmv of two portable
monitoring instruments—the Foxboro OVA-108 and the Bacharach
TLV-108. Both instruments were calibrated with methane.
The response factors at a concentration of 500 ppmv are from
the document entitled "Method 21 Evaluation for the HON,
"90-ME-07)" (March 1991) prepared for the Emission Measurement
Branch of the U.S. Environmental Protection Agency. This
document presents the results of analytical tests performed to
determine the response factors at an actual concentration of
500 ppmv of several emission monitors including the Foxboro
OVA-108, two of Foxboro OVA-128 units, the Heath Detecto-PAK III,
and the HNU Systems HW-101. The two Foxboro OVA-128 instrument
response factors are presented in the table to indicate the
variability of individual instruments. To determine the response
factor for the OVA-128, the average of the two instrument
response factors should be used. All of the instruments except
the HNU HW-101 were calibrated with methane. The HNU HW-101 was
calibrated with benzene.
A dashed line in Table C-l indicates that the study did not
test that particular chemical. If the emission monitor did not
respond to a chemical, N/R was recorded to indicate no response.
Operators of portable leak detection devices should be
thoroughly familiar with their instrumentation. Even under the
best of circumstances, no two analyzers will perform exactly the
same and the effect of changes in instrument parameters upon
C-l
-------
accuracy can be significant. Other external quality controls,
such as a checklist for periodically noting battery condition,
fuel pressure, post-survey calibration checks, etc., will support
the validity of the data. An audit program testing both the
operator and the analyzer should be a requirement whenever a
situation warranting an exacting determination of a fugitive
emission is encountered.
In general, the response factors follow the pattern which
would be predicted for increasing flame ionization detector
response with increasing hydrocarbon character for the molecule.
The sequence of compounds methyl chloride, methylene chloride,
chloroform, and carbon tetrachloride exhibits progressively
decreasing response on the OVA detectors (response factors
ranging from 2 to 12) as the substitution on the methyl carbon
atom increases (i.e., decreasing hydrocarbon character for the
molecule). In general, increasing electronegativity of the
substituent decreases the system response: methyl chloride,
response factor approximately 2; methyl bromide, response factor
approximately 5; iodomethane, response factor approximately 8.
Carbon tetrachloride exhibits a response factor of 12 or more,
but tetrachloroethylene has a response factor of 2 or less. The
lack of carbon-hydrogen bonds in tetrachloroethylene is
apparently compensated by the presence of a site of unsaturation
in the molecule (chlorobenzene, response factor 0.60 vs.
trichlorobenzene, response factor of 12 or greater). The
difficulty of obtaining a reproducible and useful response factor
for compounds of insufficient volatility such as nitrobenzene,
m-cresol, and oxygenated compounds such as acrylic acid
demonstrates that there is a point dictated by vapor pressure or
possibly boiling point where an accurate measurement cannot be
made using the portable field analyzers. With compounds which
are not very volatile, the portable field analyzers can be used
C-2
-------
only qualitatively, at best; if a large amount of the compound is
present in the air, the compound will be observed but not with a
proportionate quantitative response.
C-3
-------
o
o
in
O
o
o
fc
o
CO
o
H
H
§
o
I
§
in
§
|
I
u
|
§
|l
u -
o
a
3
*
JD_
5
IT
1
•
I:
o
0«
IT
£l
i*
*i
si
l~
"i
\i
CO
e|
£I
>.
i— *
"o
1
u
«
£
&:::.'!
•0
S: : i : *•
1111 CM
• •
in •-
s : ' ' •' fc
«::::«
• i i i
«~ i i i i O
•* CM
CO ^
; 6 S JM S t
in CM «- r>- «^
: s 8 e Of g
•- •- O M 0
1 t
| ?! 1 r
3 < c f
•n v Of •»-
2 o o £ E c
«J .^ — 6 O O
tl V « « V 41
<<<<<<
or- i «- in o
1 1 ^ 1 1 1
t^> O^ CM «^ >Q in
0 «- i ^ 0 0
i i CO i i i
in •* o N. in in
P^ ^ ^ *O S- r^
fe : : R 5
l*O CM tO
K ' ' ^ =
O* ' ' K- ^O
CVJ K1 CO
3 : : 3 S
CM in o
CM ' ' «X -
' in
o «- o ^
-i — i ,_ — i
X J » ^ _1
S 'i 2
c j= 4! a
S u c c
f V — < •
3 • S •
"
•* • • «n • '
In ' ' tn ' •
•^ CM
S' ' tn ' '
' ' ". ' '
r CM
11 it
in • i ^- i i
in K
^ rg
eg ; e s
CM in •- «-
g is c? : s K
000 00
_J 0 _l _1 _1 _l
= 1 « J-
1- O O f
4-> 0 — • O
§"" "S "5 i 4P
£ * £
III!}}
*r "° «- °
K» CO T ?
»- «- in «-
o o o «-
CO
«
O
EH
U
£
H
CO
CO
H
I
U
H
J
i
C-4
-------
^
o
u
PH
CU
O
o
in
(X
(X
o
o
o
^
o
H
fe
O
(0
2
O
U
55
8
EH
U
CO
§
85
w
CO
§
PH
to
w
u
1
§
in
|
•
i
S
u
1
u
1
CD >
4-* 5.
§1
^o-
1
u
o
2
a
8 =
X Q>
•
jj™
s ^
8
i*
o .^
"1
1:
g
""1
•g
ft
i
8*
i *
«£s
8
X
.5"
§
i
u
WE
d
(O
V»» ' 'o' ' ^ ' CM in • ' '
N ' ' o ' ' Jo • 5> " ' • i
in «- «^ o CM
u
X-< 'ca1 'in'CMOO' ' '
K ' ' S ' ' IP ' «- 5! ' ' '
•* o o in •"
"op ' 'O1 ' »• • «B K ' ' '
CM o ^ in CM
tfl ' *Nt' ' (V ' «- O> ' ' '
sr1 'in1 ' Ur ' N. >o ' ' '
O O »- -O (M
(>•! 1* 1 III
«^i i^Oi ifOiO^1 i *
*» in ^ o> ^
•* o -r- in CM
i oo^roo^^O i oaoo^o
i -o KI o «- ^ eo »- I o^eorM
CM>O**C>^O^«— ^OOCM^~
I CM>Q*->*OO rvcOKtp
1 OKiCMCM^tCMIO i K1M»*P-
OCMOCMCM
l^fMl^i*^l iCXI^l i
lOvinKi^co^^ino^o^^ocM
mi i^-ioaieocMf ifOO^
lOOiOiOi i«O>Oi i
CMOO«~OCOOO|nOOv~00
u
1 ' tn ' '
JQ • •
01
: : g : ;
to
t i (^ i i
w
CM
, «^
i I ^ I i
CM
1 1 O 1 1
i i r«» i i
Kl
i W C>
CM CM «- «-
^ K ' S ^
do o o
«rt 0 J ^
U I
CM X
i « ttl
4J «« ft| 4^
(. 0 <-• «
«i C f\J ^ fM
O * * i *
* CO O ^> (NJ
m O^ N» co po
•O * i t i
t S> «- K1 «-
p O ^ N ^
n
C-5
-------
4J
C
O
u
o
o
10
ft
o
o
o
CO
§
H
o
2
O
O
i
CO
CO
S5
2
CO
U
jQ_
O
1 ,
i
1 1=
§
in
* Ot
a Q *~
I £|
I e*
* X '
£l
2|
§1
"1
•g
§ l3>
!| 5*"
r i;
X
-8
i
u
*
o
CO
1 1 • 1 1 1 1 1
Illlllll
Illlllll
Illlllll
Illlllll
Illlllll
1 1 1 1 1 1 1 •
Illlllll
>OM*-oor\toco
3«-o«inooN*coin
§••••••*
p^»
^IM«O>OiniMtoh-
IMOOOOr^^o
i
i i t. i
U I f ft Z
I ft U t. h-
Z Cfl i ft ft
J J " ". "8* =
VVftU.ftw.co
5 £-.£.£.£ g 1 «
-uiiil«.iJ
33333333
CD 00 CO CD 00 00 CD 00
O K) * i I ^1 (M
^K ^fc (M ^^ ^O ^Nb ON
i i in «o o i i
M ON ON 1 1 M h»
>» o ro in co PJ o
^ »- r- R. ON «- «-
! ^ 5ff •* ! *~ "S ! in in
1 K o •- • CM o ' m O «- Sf ' CO CO ' CO CO
1 <3 «-; x ' S3 to ' K RJ
K «^ »^ o «^ »»
in »•
1 >£ rj SI ' N/J -* ' •- IM
' z K z ' 55 in ' o> in
to *Z o ^ i»!
. at . .
. 49>^! £l<2 ! CO«O
1 c5 S » ' &• >o ' P5 P-
to in «- o CM ON
' « ° S. ! S. ^ ! S. fc.
RIM IO v- O (M 1^
^~ ^5
R $ : : o : § 2 fc •
>O ' ' ' O ' -C t*- CO '
•N* • * * -NT ' tO ^ "* '
o to o o •*
-S ."2
•:: -S o 1-
ju o ft X -• v
•e _ -e ^ f u
jiEgiEs"^.,
•>-31-3IOX«iC X
l_«4J
CDUUUUUUUUUO
0 «- O h-
i o in i i ON i to to
>*• in t eo o « o i i
N> i tOinfM<»tONO
-------
o
u
O
O
m
H
U
1
i
§
in
£
V
*•»
I
i
**
u
i
p
o o
|s
1"
*
o
ae >
1
•^
i:
o
85
£«
g
E*
"8 *
•2 5
§
|"§
""1
j|
CD
a
Eg
i •
£l
X
— *
TJ
1
1*
u
*
o
z
g
• KJ ' • ' ' • ' » i ' co '
W" ^"
: fe : s : : : : I S : R :
o o
«- in v
i 1 i i » i i U 9? * *
i * ^ ' i $3 00 i
> O
in ogjog»oo*-oo«-
(N*
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
g g a s s
«* M «^ «^ O
SO 0 to O
in •* •* o
o «- o o o
Z _1 _1 _1 X
1
| | 1 i
X X X X
01 01 « V
1 f 1 ? 1
u u u u o
o «- eo eo in
i i i i i
e» •* Q r *
O^ v* 00 O* ^~
i i i i i
§ 0 ° 0 ^
C-7
-------
o
u
o
o
in
&
o
o
o
V
o
H
fe
O
CO
§
H
EH
H
O
§
O
CO
g
w
CO
§
ft
CO
w
H
O
.o_
1°
i
l 1=
§
in
c
5 I?
| £I
I g8
"s
21
!;
"s
m. U
1 fe?;
«> o1-
u B S
4-» D. CD
C O.
° d" g S
r -T
< £l
*
£
1
4k
O
2
CO
2
III Ifllll jt 1
111 iiiiii )C *
CO
iii 111111 Si* i
iii 111111 eo '
*•
ill 111111 ^m *
iii 111111 5r '
CM
III llllll ^» •
iii 111111 ji i
•
so
111 111111 U i
111 111111 CM I
8*
O* CO OT CO CM CO O O* CO i *O
O f\J J* ^^r-^f>3^^ K1
K 5 S SSfc«S5 !S
«^«-^«^ ddd<3«^«-^ IM
• t>
—• C w- *- * •• ^
O V * « V CJ «^
£ fi. 4! 4! > - f P «> 4!
*5 !T £ t 41 0) S C C
u ^ 5 S c c •- — _, e
* ^- S £ •£ •£ 5 •£ (M w «
§' -fi-fil!l!4;4!"4ll
•^O OOOOOO*~OO
4-*4^O OOOOOO»OO
O O O (M O O O O O OH- O O
IM CO «- IM 0 0 •*
II i ^ K) i i i i
CM K» M i i M3 O> O> >»
^o S-o^omin ^
tt i in Ki * i i i
S«- «- I i ^ O O *-
Kl •$ ir\ in o •& *t «-
•-^ in c> K <- in in ^
: s : : R
«- «n
: R : : g
: s : : S
""
: § : : S
r\i X)
i ii
1 S ' ' ~
IM -O
§ ; R 3 S
«-^ *• v- CM
° : ~ ~ *
"- O «- K»
§
II CM -
CM M • Z
^ «" »
. of |8
C- U — " O
S. 5. x x *-
1 fe 1 1 1
•g -g £ i i
o o o o o
CO
I ° IM
1C rL fj.
CM «- *T
•«*• in CO
in (M 9
~
CM
g
CM
8
K>
3
CM
S
CM
g
CM
g
«J—
-J
1
*^
«r-
jf
N
C9
i.
"x
JC.
I
0
N.
i
i
N.
in
08
-------
§
o
o
o
U*)
o
o
o
CO
§
H
EH
I
W
K
O
W
CO
%
o
ex,
W
H
u
CO
I
g
in
g
i
I
1
a >
4* 5.
O O
!°
o
3*~
3.
^
o
17
s
*•* M
i:
o
j*
£i
1*
O --
-I
si
8<
"1
I.
"5 H-
.
o g
IT
£i
X
°T a)
.7; «
*• —
"S
1
ac
d
x
H
:
.
*
;
:
8
~*
§
0
X
•S
'i
»•-
"3
M
1
i
O
in
3
i
N-
3 8 3 :
•- «- CM
~ ~ « ''
ro ^ ^
s & s :
ro CM CM
I*. — .» '
CM «» in '
•* CM CM
S S fc :
ro CM CM
R g ; R
^ CM 0
S K ! in.
«- «^ o
j d w
1
CM
c ^
1 *U *~
**. ^ «r
- f 1
* - I 2
§"5 x a
•- fi ^
O UJ UJ U
«- CO 1^
i i i O
S i i ft
- S • '• 8 R : : S : 8 Q | 5
"~ «- CM »- O «- 0
:g: :ss: :s:»2:5
CM" - ^ d - - R
!°! I*01"! IS}!1*1*-"
'in' ' r-
ro
iini iQ>COi i N- i K> S- •-
in •* ^ h- o ro to
10 CM «-^ 0 CM *^ •#
i«rrorOi ir^^"«*f>i i i
, •- , '•**"; ""I**1 ' '
- ro 10 ^ ^ ro ^
Seo^rjiM'N-^opj' i *
^GOOSv'OOr^in' ' '
oj^csroo «-^ooo
~ 1 H ^ T o
* V V E O — •
_ tltlJIU O — . O
Q u u 4-> *o a i-fu
C <-> ux
f^4->XOOVW T3 T3 D>
WtlVU— <— «£N
ox--8«"£lllli
UJ ty UJ UJ (u uj UJ UJ U) UJ UJ UJ UJ
in>o^in in *t ** CNJ *-
in i i i i K> i r^- i «- i i i
• o so h- eb i Ch < ^ i ro ^) *—
h-COr^O^OOOKi^*^lA^Of>J
«- i O < i O t fS* i 00 * i i
i o ^ ^- o i in i o i "O h- N.
^•w-N*%t>d-inooO'*4-ooo
C-9
-------
o
o
in
o
o
o
to
§
H
EH
O
§
8
CO
to
§
&
to
W
K
H
I
O
I
JQ_
0
sT
i
1 *=
1 i:
o
i
1 1*
S £ <
1 g
1 E%
< | .
£i
sl
£ <
""§
i h
it 2
"°" g|
1 !"•"
5 £l
>
*£ •
~o
1
i
0
z
<
»"•'•* • ' ^J» O\ ' ' CM
>o • ' 3 • • rt 5 • • Rj
"O "* CX *~ O
z ' • 8 : : 8 8 : : R
• • • • •
«- « 2 ° **
3 : : * : : S R : : «
CM Is! J5 *~ *"*
fc : : ft : : ^ 3 ' : s
CM «^ «\I «^ «0
IO IM
ii i i O * i
O > > M ' • «O CM i i •£
^ CO CM ^ O
N CO NO 1^ CO
35S- RtC-KSIS?'
CMCMKt mo O(MO>
K P i fe : 3 s : pj s g :
csi «~ ** to o o o co
«dzl d d d d d
u
6
< 1 -I1 ! §
"Si-oKi-S £ *> v
O •-•—*'>• O i i V C
•pox: i- z «- o ^XE EE U4-* 4D 01 tt> C-S
£fE EC >. a x x xp-o
*'*'OOO-<««I«I«XO
K> in in M >o
eo » *o i i ro * i «^
I in i CM CM < •* «- >
•-•-co incoeain** ob
CM i •— t i «O i i 00
t N- > fN-oo *^
p-»-*o in «— oo •— in S-
1 1 «Q 1 .A * «
• • § • fS
0 1^
1 * ^ ' At ' '
^» O^
• e- • ^ •
• S • z
0
: '• s '- ?; : '•
- s
ii i ii
• s • g • •
s
5 R : : : w «
d <> «^ «-
O CM > OO < O CO
*? * ' . ' . .
O CM O O Q
« <= J J J J
01
+J
s . - 1
s £ g . 1 1 %
S if S I fe 8- 1
3 9 y «- -c i- i-
1^88-1-8-8-
« WJ W) Ul W) tft tft
in i i in «- o i
cv i i r^- m •O i
i in o i * i eo
in r» -^ ep po r^- o
N* v in iv N» ^> <—
C-10
-------
C
O
U
o
O
in
04
o
o
o
CO
§
H
u
§
u
I
•«
CO
CO
§
ft
CO
H
O
H
.J
9
I
§
in
g
4-»
§
u
1
<
g
h
°0
1"
*
o
?
tf
£ M
+* »-i
M
8*
|7
°f
2%
O ^™
"1
||
g '
""1
h
8"
m
ji
£I
Is
*J **
"o
1
u
•
o
vt
5
1 1 1 1 1
( 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
I 1 1 1 1
I 1 1 1 1
1 1 I 1 1
1 1 1 1 1
I 1 1 1 1
0. 2* CM 0 «-
6^ o «- •- *o
O <\l K) Kl "O
N in N K «-
^d in «-
o o «- «- o
—i — i -J _i ac
! i
1 1 a * :
_^ jg o *w
"x 2 ^ 1 x
o" — * X u o
c. • f a a
--««x
fO ^^
^) i i K> >*
o» co p* m ^*
N i * 00 -4-
i O r- i i
in o^ ^^ co p1^
N> tn v~ ^ ^
o> ^ I ! ^ t> ra •
in to • • ^ K » '
>* «vi «^ ^ f«j
R 5 : ! 3 !5 Z '
»^ N i\j ^ ^
IM
1 1 : : 3 fc 8 :
M - -
* ^ : : *? s. a. :
K 0- Kt CM »-
Oil i
•» «- i i •- h. CO i
«\J -8 J! «>
«j — -5 .^- v>
a a X ••- i- — > o
£««§££-!
tl U U 1- £ 4-> O
_. , O O U O t> *•
g^--__.--
££££££££
££££££££
«-?O.K(>KIK|rr
i O * > t i i ^*
'O^O^MKWfO
in i O ^~ ^* ^^ ^» ^* ^* ^~
8 S sS S :
ro «^ «^ «-^
^ S Si K :
in o o o
^555
in »- «- ^
0 0 -0 in '
in v «> '
£ •- 6 V «M
i
f.
££££££
^ '
-------
C
o
CJ
o
o
in
o
o
o
CO
§
U
2
O
I
u
s
o
E-i
O
CO
2
2
CO
H
I
CJ
1
§
in
|
4-*
i
1
|
S
<
g
It
o o
o o
u *
o
1"
*
o
gT
i
1=
£0.
o
2 S
l<
"-8
28
o •"
li
a
£I
h
o '-
CD
_
P o
|T
£i
>•
|
u
*
0
Vt
s
1 1 1 1 1 1
1 1 1 1 1 1
1 1 1 1 1 1
• till*
1 1 1 1 1 1
1 1 1 1 1 1
1 1 1 1 1 1
«* i v- m in CM
in i ** CM eo CM
•- -r- in O CM
Q* *~ >Q W CO W
>*• in ** *- to K*
O 0 «- sf 0 O
1
i CM
» *r-
C * > 0> V
55. « 8 S
i I |||
oj ro -- c x >•
i i a a u o
.c .c .e .c £ !"
£ z 2 * 2 S
in op PJ
ti. ^- s.
i i i
S& o to
Si i i VL i ' o ' ' * * * *^i '^* • Sr
S • • ' o ' • S ' ' z
° s -1
JJ I I ' Z ' ' S ' ' O
u
K ' ' • in ' ' S> ' '00
s. : : : 'i- ', : « : : »
pJ«-»-o«Min«-inp>i
•or^-*(MNoro-ofV-* *
rMT-oJO-K-j-KieS^oeS '
PJ^OOO>«-KI«-»-^
u
*U f\J L.
£ o .' . S
•Sli Sc^S!
«!^x£Kg£8. S
cv4-*"-c.c*'Oz v^
•JO-ia— uwtii.!
f£fa.i-i-t.i.a u^K}Oin^cMrMeoo
o K oo i c> CM in ro i •
• • i O i i i Kl *~ «—
LnN»op«-coc>.Ln>*«-«-
^» n» V* ^™ v* r*» r*" CM ^~ *""
i
i
i
i
9
1
S
C3
PJ
O
"
!
o
«
1
-
C-12
-------
£
o
o
On
A
O
o
in
04
o
o
o
**
o
H
fa
O
CO
§
t1
£
u
*
o
CJ
o
EH
CO
!Z
O
04
CO
u
PC:
u
n
I
§
in
|
*J
1
s
2
1
" d"
V
u
ja
if
1
| =
«
£i
o%
-1
I1
£I
Bacharach
TLV"
2§
""1
st
53
o
u
o
CO
2
: : K - ' ' ' & 8 K :
«» ^ Kl CM
i • m * • ' ' ** *A m •
1 ' K • ' ' ' OS IM KJ '
CM O >r *^
' ' in ' ' ' ' 1 CO «» CM '
• ' CM ' ' ' ' «3r »™ in '
•* *- CM *-
ii i i i i i
i iv~i i i to^cMini
O *4" O P*
«* r CM ^
I^K^in^inN-o i in • ^
^ co J «- ! »^
^-O«-K»«-inCM «- «-
^^ CO O* '& ^~ ^f O^ ' O ' *""
ro « ^ KI a >* S. ' oo ' •*
o d i-^ ^ o ca d d d
^^jj^jjjj j
T> CM
i o "~
"0 TJ — • • f TJ «T
S- — o «> u •" e
1 £ U f C ••" X *^
CM 13 C O S T3 O E
S'sll'ISilSls
O Q.O-O>&&&£&&""
u ooooooo oot.
i »o i ** co • i in o oo i
Sico* tin^i i i*v
COKIO>KI-«-o«-coinin»-
S 8 : K ^
•j^ fO ^~ Q
in >O ' •- CM
2 I • 5 ^
SO J O c>
•^ ^ o
»^ CNI «^ CM
0 «T I ^f t-
•^ t.
ff fi tl • «l «. OP
in u -*
£ £ S s
i O Kl i
§i i N
CM
KJ
in
d
fS
d
N.
d
«
d
CM
KJ
CM
Kl
Kl
d
0
Kt
§
1
I
C-13
-------
4J
i
cm
cu
o
o
to
o
o
o
*.
o
H
fn
O
(0
§
O
§
o
s
CO
K
g
U
g
H
CO
I
CO
H
P4
H
U
H
.4
I
I
8
in
g
*•»
4-»
<
|
* ^
W
|"
*
O
ae i
2
li?
o
-,
ij
«i
i -
£|
i*
"i
ji
i1
£i
-8
-j Q
O
j
d
2
S 8 R 8 ! R fe fc 5 ! g S
«O «~ ^ •- O IM «- CM v-O
•4 p ps in ! in co co -'O ^ »~ * f1^ OQ
r-«-cM3S' -j- K? ^t »- • o- 3
>O »- »- CM OKICMCM CMO
in 2 g;
mo^cM -* -o «- o h-eo
CM^«-CM OKICMCM CMO
O> •- co in ro <-o>,opcoin
Kl -» ,»-N-KIKI
OCMKIKIIM «-KI >-OCMin
inp^>o^in «o«*|in^mN-
KI^-CMOO* •*KI'OO"-CM
•- O •- O O O «- OOr-CM
•J- 1 1
£ • • c «r « *u
gSJ5| J!S°
S^l^l 2S-82§«
V 4-* +J +> ^ *•• «B *^ W *•—
i 4> 4| v 0- £ *^ 6 o a j* •
ooooo a «i o — - o c o.
t- u t. c. L. — • o t- j: t- $
o o o o o x « .0 o a. TJ »
O U U O O » 41 > X > >> X S
«- CO ** «M Kl
i *O IA ^ >^ i i t «^ <^i
Siii* ** m o i i CM
inov-co >* o 'O ^- iA^t
i in O o «— i i i o Kl *
O » « i i *• 00 K* i i *Q
rsi «- p- ex >o rxioottn in o
«^ ^^ ^^ ^^ ^ ^~ ^~ in i^ ^* ^~
$ g
O w^
•*
ae n N N
znj J3 u
C-14
-------
APPENDIX D
SELECTION OF SAMPLE SIZE FOR SCREENING CONNECTORS
-------
APPENDIX D
SELECTION OF SAMPLE SIZE FOR SCREENING CONNECTORS
In estimating emissions for a given process unit, all
equipment components must be surveyed for each class of
components. The one exception to this "total component
screening" criterion is the category of connectors. Note
however, that if the process unit is subject to a standard which
requires the screening of connectors, then all connectors must be
screened. In typical process units, connectors represent the
largest count of individual equipment components, making it
costly to screen all components. The purpose of this appendix is
to present a methodology for determining how many connectors must
be screened to constitute a large enough sample size to identify
the actual screening value distribution of connectors in the
entire process unit. Please note that the sampling is to be a
random sampling throughout the process unit.
The basis for selecting the sample population to be screened
is the probability that at least one "leaking" connector wixl be
in the screened population. The "leaker" is used as a
representation of the complete distribution of screening values
for the entire class of sources. The following binomial
distribution was developed to approximate the number of
connectors that must be screened to ensure that the entire
distribution of screening values for these components is
represented in the sample:
n > N * (1 - (1 -p)1/D]
where:
N = Number of connectors;
D = (fraction of leaking connectors) * N; and
p > 0.95.
D-l
-------
Refer to Figure D-l, which shows the fraction of leaking
connectors at several leak definitions based on currently
available data. Since the fraction of leaking connectors will
most likely not be known prior to screening, the leaking fraction
at the intersection of the SOCMI average emission factor line and
applicable leak definition line on Figure F-l can be used to
estimate what the fraction of leaking connectors will be.
Entering this value into equation D-l for at least a 95 percent
confidence interval (p = 0.95) will give the minimum number of
connectors that need to be screened. A larger sample size will
be required for units exhibiting a lower fraction of leaking
connectors.
After 'n1 connectors have been screened, an actual leak
frequency should be calculated as follows:
Leaking frequency = Number of leaking connectors
n
Then, the confidence level of the sample size can be calculated
using the following equation, based upon a hypergeometric
distribution:
P = 1 - (N-D'll (N-nl!
N! (N-D'-n)!
where:
N = Total population of connectors;
n = Sample size; and
D1 = Number of leaking connectors * jj
n
If 'p1 calculated in this manner is less than 0.95, then a less
than 95 percent confidence exists that the screening value
distribution has been properly identified. Therefore, additional
connectors must be screened to achieve a 95 percent confidence
level. The number of additional connectors required to satisfy
the requirement for a 95 percent confidence level can be
calculated by solving Equation (D-l) again, using the leak
D-2
-------
SOCMI Connector Equations
n
-------
frequency calculated in Equation (D-2), and subtracting the
original sample size. After this additional number of connectors
have been screened, the revised fraction of leaking components
and the confidence level of the new sample size (i.e., the
original sample size plus the additional connectors screened)
should be recalculated using Equation (D-3). The Agency requires
sufficient screening to achieve a 95 percent confidence level,
until a maximum of 50 percent of the total number of connectors
in the process unit have been screened. The EPA believes that 50
percent of the total connector population is a reasonable upper
limit for a sample size. If half of the total number of
connectors are screened, no further connector screening is
necessary, even if a 95 percent confidence level has not been
achieved.
D-4
-------
APPENDIX E
REFERENCE METHOD 21
(Code of Federal Regulations, Title 40, Part 60,
Appendix A. Reference Method 21, Determination of
Volatile Organic Compound Leaks. Washington, D.C.,
U.S. Government Printing Office. Revised
June 22, 1990.)
-------
STATIONARY SOURCES
S-923
120:1281
%COj =Measured CO: concentration
measured, dry basis, percent.
7.4 Average Adjusted NO, Concentra-
tion. Calculate the average adjusted NO,
concentration by summing the adjusted
values for each sample point and dividing
by the number of points for each run.
7.5 NO, and SO2 Emission Rate Calcu-
lations. The emission rates for NO, and
SC>2 in units of pollutant mass per quanti-
ty of heat input can be calculated using
the pollutant and diluent concentrations
and fuel-specific F-factors based on the
fuel combustion characteristics. The mea-
sured concentrations of pollutant in units
of parts per million by volume (ppm)
must be converted to mass per unit vol-
ume concentration units for these calcula-
tions. Use the following table for such
conversions:
CONVERSION FACTORS FOR CONCEN-
TRATION
E-C»F.
100
%COt.
Eq. 20-8
From
Ib/scl
ppm (SO«)
ppm(NO>)
PpmfSC*)
Ppm(NOJ
To
ng/«m*. ._ ....
nQ/sfn*
ng/sm*
lb/»cf
to/set
Multiply by
10*
10*
1«02x 10*
2660 x 10*
1 912 x 10*
1.660x10*
1.194x10*
7.5.1 Calculation of Emission Rate Us-
ing Oxygen Correction. Both the Oj con-
centration and the pollutant concentration
must be on a dry basis. Calculate the pol-
lutant emission rate, as follows:
20.9
20.9- %O|
Eq. 20-6
where:
E=Mass emission rate of pollutant, ng/J
(lb/106 Btu).
7.5.2 Calculation of Emission Rate Us-
ing Carbon Dioxide Correction. The CO:
concentration and the pollutant concen-
tration may be on either a dry basis or a
wet basis, but both concentrations must
be on the same basis for the calculations.
Calculate the pollutant emission rate us-
ing Equation 20-7 or 20-8:
Eq. 20-7
where:
Cw=Pollutant concentration measured on
a moist sample basis, ng/sm3 (Ib/scf).
%CO2w=Measured COa concentration
measured on a moist sample basis, per-
cent.
8. Bibliography
1. Curtis, F. A Method for Analyzing
NO, Cylinder Gases-Specific Ion Elec-
trode Procedure, Monograph available
from Emission Measurement Laboratory,
ESED, Research Triangle Park, NC
2771 I.October 1978.
2. Sigsby, John E., F. M. Black, T. A.
Bellar, and D. L Klosterman. Cbemilumi-
nescent Method for Analysis of Nitrogen
Compounds in Mobile Source Emissions
(NO, NOz, and Ns). "Environmental Sci-
ence and Technology," 7:51-54. January
1973.
3. Shigehara, R.T., R.M. Neulicht, and
W.S. Smith. Validating Orsat Analysis
Data from Fossil Fuel-Fired Units. Emis-
sion Measurement Branch, Emission
Standards and Engineering Division, Of-
fice of Air Quality Planning and Stan-
dards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
27711. June 1975.
METHOD 21—DETERMINATION OF VOLA-
TILE ORGANIC COMPOUNDS LEAKS
1. Applicability and Principle
1.1 Applicability. This method applies
to the determination of volatile organic
compound (VOC) leaks from process
equipment. These sources include, but are
not limited to, valves, flanges and other
connections, pumps and compressors,
pressure relief devices, process drains,
open-ended valves, pump and compressor
seal system degassing vents, accumulator
vessel vents, agitator seals, and access
door seals.
1.2 Principle. A portable instrument is
used to detect VOC leaks from individual
sources. The instrument detector type is
not specified, but it must meet the specifi-
cations and performance criteria con-
tained in Section 3. A leak definition con-
centration based on a reference compound
is specified in each applicable regulation.
This procedure is intended to locate and
classify leaks only, and is not to be used as
a direct measure of mass emission rates
from individual sources.
2. Definitions
2.1 Leak Definition Concentration. The
local VOC concentration at the surface of
a leak source that indicates that a VOC
emission (leak) is present. The leak defini-
tion is an instrument meter reading based
on a reference compound.
2.2 Reference Compound. The VOC
species selected as an instrument calibra-
tion basis for specification of the leak defi-
nition concentration. (For example: If a
leak definition concentration is 10,000
ppmv as methane, then any source emis-
sion that results in a local concentration
that yields a meter reading of 10,000 on
an instrument calibrated with methane
would be classified as a leak. In this exam-
ple, the leak definition is 10,000 ppmv,
and the reference compound is methane.)
2.3 Calibration Gas. The VOC com-
pound used to adjust the instrument me-
ter reading to a known value. The calibra-
tion gas is usually the reference com-
pound at a concentration approximately
equal to the leak definition concentration.
2.4 No Detectable Emission. Any VOC
concentration at a potential leak source
(adjusted for local VOC ambient concen-
tration) that is less than a value corre-
sponding to the instrument readability
specification of section 3.1.1(c) indicates
that a leak is not present.
2.5 Response Factor. The ratio of the
known concentration of a VOC compound
to the observed meter reading when mea-
sured using an instrument calibrated with
the reference compound specified in the
application regulation.
2.6 Calibration Precision. The degree
of agreement between measurements of
the same known value, expressed as the
relative percentage of the average differ-
ence between the meter readings and the
known concentration to the known con-
centration.
2.7 Response Time. The time interval
from a step change in VOC concentration
at the input of the sampling system to the
time at which 90 percent of the corre-
sponding final value is reached as dis-
played on the instrument readout meter.
3. Apparatus
3.1 Monitoring Instrument.
3.1.1 Specifications.
a. The VOC instrument detector shall
respond to the compounds being pro-
cessed. Detector types which may meet
[Part 60, Appendix A, Method 21]
10-16-92
Copyright © 1992 by The Bureau of National Affairs, Inc.
0013-9211/92/SO-K 50
669
-------
120:1282
FEDERAL REGULATIONS
this requirement include, but are not lim-
ited to, catalytic oxidation, flame ioniza-
tion, infrared absorption, and photoioniza-
tion.
b. Both the linear response range and
the measurable range of the instrument
for each of the VOC to be measured, and
for the VOC calibration gas that is used
for calibration, shall encompass the leak
definition concentration specified in the
regulation. A dilution probe assembly
may be used to bring the VOC concentra-
tion within both ranges; however, the
specifications for instrument response
time and sample probe diameter shall still
be met.
c. The scale of the instrument meter
shall be readable to ± 2.5 percent of the
specified, leak definition concentration
when performing a no detectable emission
survey.
d. The instrument shall be equipped
with an electrically driven pump to insure
that a sample is provided to the detector
at a constant flow rate. The nominal sam-
ple flow rate, as measured at the sample
probe tip, shall be 0.10 to 3.0 liters per
minute when the probe is fitted with a
glass wool plug or filter that may be used
to prevent plugging of the instrument.
e. The instrument shall be intrinsically
safe as defined by the applicable U.S.A.
standards (e.g., National Electric Code
by the National Fire Prevention Associa-
tion) for operation in any explosive atmo-
spheres that may be encountered in its
use. The instrument shall, at a minimum,
be intrinsically safe for Class 1, Division L
conditions, and Class 2, Division 1 condi-
tions, as defined by the example Code.
The instrument shall not be operated with
any safety device, such as an exhaust
flame arrestor, removed.
f. The instrument shall be equipped
with a probe or probe extension for sam-
pling not to exceed !/4 in. in outside diam-
eter, with a single end opening for admis-
sion of sample.
3.1.2 Performance Criteria.
(a) The instrument response factors for
each of the VOC to be measured shall be
less than 10. When no instrument is avail-
able that meets this specification when
calibrated with the reference VOC speci-
fied in the applicable regulation, the avail-
able instrument may be calibrated with
one of the VOC to be measured, or any
other VOC, so long as the instrument
then has a response factor of less than 10
for each of the VOC to be measured.
(b) The instrument response time shall
be equal to or less than 30 seconds. The
instrument pump, dilution probe (if any),
sample probe, and probe filter, that will
be used during testing, shall all be in
place during the response time determina-
tion.
c. The calibration precision must be
equal to or less than 10 percent of the
calibration gas value.
d. The evaluation procedure for each
parameter is given in Section 4.4.
3.1.3 Performance Evaluation Require-
ments.
a. A response factor must be deter-
mined for each compound that is to be
measured, either by testing or from refer-
ence sources. The response factor tests are
required before placing the analyzer into
service, but do not have to be repeated at
subsequent intervals.
b. The calibration precision test must
be completed prior to placing the analyzer
into service, and at subsequent 3-month
intervals or at the next use whichever is
later.
c. The response time test is required
prior to placing the instrument into ser-
vice. If a modification to the sample
pumping system or flow configuration is
made that would change the response
time, a new test is required prior to fur-
ther use.
3.2 Calibration Gases. The monitoring
instrument is calibrated in terms of parts
per million by volume (ppmv) of the refer-
ence compound specified in the applicable
regulation. The calibration gases required
for monitoring and instrument perfor-
mance evaluation are a zero gas (air, less
than 10 ppmv VOC) and a calibration gas
in air mixture approximately equal to the
leak definition specified in the regulation.
If cylinder calibration gas mixtures are
used, they must be analyzed and certified
by the manufacturer to be within ± 2 per-
cent accuracy, and a shelf life roust be
specified. Cylinder standards must be ei-
ther reanalyzed or replaced at the end of
the specified shelf life. Alternately, cali-
bration gases may be prepared by the user
according to any accepted gaseous stan-
dards preparation procedure that will
yield a mixture accurate to within ±2
percent. Prepared standards .must be re-
placed each day of use unless it can be
demonstrated that degradation does not
occur during storage.
Calibrations may be performed using a
compound other than the reference com-
pound if a conversion factor is determined
for that alternative compound so that the
resulting meter readings during source
surveys can be converted to reference
compound results.
4. Procedures
4.1 Pretest Preparations. Perform the
instrument evaluation procedures given in
Section 4.4 if the evaluation requirements
of Section 3.1.3 have not been met.
4.2 Calibration Procedures. Assemble
and start up the VOC analyzer according
to the manufacturer's instructions. After
the appropriate warmup period and zero
internal calibration procedure, introduce
the calibration gas into the instrument
sample probe. Adjust the instrument me-
ter readout to correspond to the calibra-
tion gas value.
NOTE: If the meter readout cannot be adjusted to
the proper value, a malfunction of the analyzer is
indicated and corrective actions are necessary
before use.
4.3 Individual Source Surveys.
4.3.1 Type I—Leak Definition Based
on Concentration. Place the probe inlet at
the surface of the component interface
where leakage could occur. Move the
probe along the interface periphery while
observing the instrument readout. If an
increased meter reading is observed, slow-
ly sample the interface where leakage is
indicated until the maximum meter read-
ing is obtained. Leave the probe inlet at
this maximum reading location for ap-
proximately two times the instrument re-
sponse time. If the maximum observed
meter reading is greater than the leak def-
inition in the applicable regulation, record
and report the results as specified in the
regulation reporting requirements. Exam-
ples of the application of this general
technique to specific equipment types are:
a. Valves—The most common source of
leaks from valves is at the seal between
the stem and housing. Place the probe at
the interface where the stem exits the
packing gland and sample the stem cir-
cumference., Also, place the probe at the
interface of the packing gland take-up
flange seat and sample the periphery. In
addition, survey valve housings of mul-
tipart assembly at the surface of all inter-
faces where a leak could occur.
[Part 60, Appendix A, Method 21]
10-16-92
Environment Reporter
0013-9211/92/W+.50
670
-------
STATIONARY SOURCES
S-923
120:1283
b. Flanges and Other Connec-
tions—For welded flanges, place the
probe at the outer edge of the flange-gas-
ket interface and sample the circumfer-
ence of the flange. Sample other types of
nonpermanent joints (such as threaded
connections) with a similar traverse.
c. Pumps and Compressors—Conduct a
circumferential traverse at the outer sur-
face of the pump or compressor shaft and
seal interface. If the source is a rotating
shaft, position the probe inlet within 1 cm
of the shaft-seal interface for the survey.
If the housing configuration prevents a
complete traverse of the shaft periphery,
sample all accessible portions. Sample all
other joints on the pump or compressor
housing where leakage could occur.
d. Pressure Relief Devices—The config-
uration of most pressure relief devices
prevents sampling at the sealing seat in-
terface. For those devices equipped with
an enclosed extension, or horn, place the
probe inlet at approximately the center of
the exhaust area to the atmosphere.
e. Process Drains—For open drains,
place the probe inlet at approximately the
center of the area open to the atmosphere.
For covered drains, place the probe at the
surface of the cover interface and conduct
a peripheral traverse.
f. Open-Ended Lines or Valves—Place
the probe inlet at approximately the cen-
ter of the opening to the atmosphere.
g. Seal System Degassing Vents and
Accumulator Vents—Place th: probe in-
let at approximately the center of the
opening to the atmosphere.
h. Access Door Seals—Place the probe
inlet at the surface of the door seal inter-
face and conduct a peripheral traverse.
4.3.2 Type II—"No Detectable Emis-
sion".
Determine the local ambient concentra-
tion around the source by moving the
probe inlet randomly upwind and down-
wind at a distance of one to two meters
from the source. If an interference exists
with this determination due to a nearby
emission or leak, the local ambient con-
centration may be determined at dis-
tances closer to the source, but in no case
shall the distance be less than 25 centime-
ters. Then move the probe inlet to the
surface of the source and determine the
concentration described in 4.3.1. The dif-
ference between these concentrations de-
termines whether there are no detectable
emissions. Record and report the results
as specified by the regulation.
For those cases where the regulation
requires a specific device installation, or
that specified vents be ducted or piped to
a control device, the existence of these
conditions shall be visually confirmed.
When the regulation also requires that no
detectable emissions exist, visual observa-
tions and sampling surveys are required.
Examples of this technique are:
(a) Pump or Compressor Seals—If ap-
plicable, determine the type of shaft seal.
Preform a survey of the local area ambi-
ent VOC concentration and determine if
detectable emissions exist as described
above.
(b) Seal System Degassing Vents, Ac-
cumulator Vessel Vents, Pressure Relief
Devices—If applicable, observe whether
or not the applicable ducting or piping
exists. Also, determine if any sources exist
in the ducting or piping where emissions
could occur prior to the control device. If
the required ducting or piping exists and
there are no sources where the emissions
could be vented to the atmosphere prior to
the control device, then it is presumed
that no detectable emissions are present.
If there are sources in the ducting or pip-
ing where emissions could be vented or
sources where leaks could occur, the sam-
pling surveys described in this paragraph
shall be used to determine if detectable
emissions exist.
4.3.3 Alternative Screening Procedure.
A screening procedure based on .„.: for-
mation of bubbles in a soap solution that
is sprayed on a potential leak source may
be used for those sources that do not have
continuously moving parts, that do not
have surface temperatures greater than
the boiling point or less than the freezing
point of the soap solution, that do not
have open areas to the atmosphere that
the soap solution cannot bridge, or that do
not exhibit evidence of liquid leakage.
Sources that have these conditions present
must be surveyed using the instrument
techniques of 4.3.1 or 4.3.2.
Spray a soap solution over all potential
leak sources. The soap solution may be a
commercially available leak detection so-
lution or may be prepared using concen-
trated detergent and water. A pressure
sprayer or a squeeze bottle may be used to
dispense the solution. Observe the poten-
tial leak sites to determine if any bubbles
are formed. If no bubbles arc observed;
the source is presumed to have no detect-
able emissions or leaks as applicable. If
any bubbles are observed, the instrument
techniques of 4.3.1 or 4.3.2 shall be used
to determine if a leak exists, or if the
source has detectable emissions, as appli-
cable.
4.4 Instrument Evaluation Procedures.
At the beginning of the instrument perfor-
mance evaluation test, assemble and start
up the instrument according to the manu-
facturer's instructions for recommended
warmup period and preliminary adjust-
ments.
4.4.1 Response Factor. Calibrate the
instrument with the reference compound
as specified in the applicable regulation.
For each organic species that is to be mea-
sured during individual source surveys,
obtain or prepare a known standard in air
at a concentration of approximately 80
percent of the applicable leak definition
unless limited by volatility or explosivity.
In these cases, prepare a standard at 90
percent of the saturation concentration, or
70 percent of the lower explosive limit,
respectively. Introduce this mixture to the
analyzer and record the observed meter
reading. Introduce zero air until a stable
reading is obtained. Make a total of three
measurements by alternating between the
known mixture and zero air. Calculate
the response factor for each repetition and
the average response factor.
Alternatively, if response factors have
been published for the compounds of in-
terest for the instrument or detector type,
the response factor determination is not
required, and existing results may be ref-
erenced. Examples of published response
factors for flame ionization and catalytic
oxidation detectors are included in Bibli-
ography.
4.4.2 Calibration Precision. Make a to-
tal of three measurements by alternately
using zero gas and the specified calibra-
tion gas. Record the meter readings. Cal-
culate the average algebraic difference
between the meter readings and the
known value. Divide this average differ-
ence by the known calibration value and
mutiply by 100 to express the resulting
calibration precision as a percentage.
4.4.3 Response Time. Introduce zero
gas into the instrument sample probe.
When the meter reading has stabilized,
switch quickly to the specified calibration
gas. Measure the time from switching to
when 90 percent of the final stable read-
[Part 60, Appendix A, Method 21]
10-16-92
Copyright © 1992 by The Bureau of National Affairs, Inc.
0013-9211/92/SO-K50
671
-------
120:1284
FEDERAL REGULATIONS
ing is attained. Perform this test sequence
three times and record the results. Calcu-
late the average response time.
5. Bibliography
1. DuBose, D.A., and G.E. Harris. Re-
sponse Factors of VOC Analyzers at a
Meter Reading of 10,000 ppmv for Se-
lected Organic Compounds. U.S. Environ-
mental Protection Agency, Research Tri-
angle Park, NC. Publication No. EPA
600/2-81-051. September 1981.
2. Brown, G.E., et al. Response Factors
of VOC Analyzers Calibrated with Meth-
ane for Selected Organic Compounds.
U.S. Environmental Protection Agency,
Research Triangle Park, NC. Publication
No. EPA 600/2-81-022. May 1981.
3. DuBose, D.A., et al. Response of
Portable VOC Analyzers to Chemical
Mixtures. U.S. Environmental Protection
Agency, Research Triangle Park, NC.
Publication No. EPA 600/2-81-110. Sep-
tember 1981.
METHOD 22—VISUAL DETERMINATION
OF FUGITIVE EMISSIONS FROM MA-
TERIAL SOURCES AND SMOKE EMIS-
SIONS FROM FLARES
1. Introduction
This method involves the visual deter-
mination of fugitive emissions, i.e., emis-
sions not emitted directly from a process
stack or duct. Fugitive emissions include
emissions that (1) escape capture by pro-
cess equipment exhaust hoods; (2) are
emitted during material transfer; (3) are
emitted from buildings housing material
processing or handling equipment; and
(4) are emitted directly from process
equipment. This method is used also to
determine visible smoke emissions from
flares used for combustion of waste pro-
cess materials.
This method determines the amount of
time that any visible emissions occur dur-
ing the observation period, i.e., the accu-
mulated emission time. This method does
not require that the opacity of emissions
be determined. Since this procedure re-
quires only the determination of whether
a visible emission occurs and does not re-
quire the determination of opacity levels,
observer certification according to the
procedures of Method 9 are not required.
However, it is necessary that the observer
is educated on the general procedures for
determining the presence of visible emis-
sions. As a minimum, the observer must
be trained and knowledgeable regarding
the effects on the visibility of emissions
caused by background contrast, ambient
lighting, observer position relative to
lighting, wind, and the presence of un-
combined water (condensing water va-
por). This training is to be obtained from
written materials found in Citations 1 and
2 of Bibliography or from the lecture por-
tion of the Method 9 certification course.
2. Applicability and Principle
2.1 Applicability. This method applies
to the determination of the frequency of
fugitive emissions from stationary sources
(located indoors or outdoors) when speci-
fied as the test method for determining
compliance with new source performance
standards.
This method also is applicable for the
determination of the frequency of visible
smoke emissions from flares.
2.2 Principle. Fugitive emissions pro-
duced during material processing, han-
dling, and transfer operations or smoke
emissions from flares are visually deter-
mined by an observer without the aid of
instruments.
3. Definitions
3.1 Emission Frequency. Percentage of
time that emissions are visible during the
observation period.
3.2 Emission Time. Accumulated
amount of time that emissions are visible
during the observation period.
3.3 Fugitive Emissions. Pollutant gen-
erated by an affected facility which is not
collected by a capture system and is re-
leased to the atmosphere.
3.4 Smoke Emissions. Pollutant gener-
ated by combustion in a flare and occur-
ring immediately downstream of the
flame. Smoke occurring within the flame,
but not downstream of the flame, is not
considered a smoke emission.
3.5 Observation Period. Accumulated
time period during which observations are
conducted, not to be less than the period
specified in the applicable regulation.
4. Equipment
4.1 Stopwatches. Accumulative type
with unit divisions of at least 0.5 seconds;
two required.
4.2 Light Meter. Light meter capable
of measuring illuminance in the 50- to
200-lux range; required for indoor obser-
vations only.
5. Procedure
5.1 Position. Survey the affected facili-
ty or building or structure housing the
process to be observed and determine the
locations of potential emissions. If the af-
fected facility is located inside a building,
determine an observation location that is
consistent with the requirements of the
applicable regulation (i.e., outside obser-
vation of emissions escaping the build-
ing/structure or inside observation of
emissions directly emitted from the affect-
ed facility process unit). Then select a po-
sition that enables a clear view of the po-
tential emission point(s) of the affected
facility or of the building or structure
housing the affected facility, as appropri-
ate for the applicable subpart. A position
at least 15 feet, but not more than 0.25
miles, from the emission source is recom-
mended. For outdoor locations, select a
position where the sun is not directly in
the observer's eyes.
5.2 Field Records.
5.2.1 Outdoor Location. Record the fol-
lowing information on the field data sheet
(Figure 22-1): company name, industry,
process unit, observer's name, observer's
affiliation, and date. Record also the esti-
mated wind speed, wind direction, and
sky condition. Sketch the process unit be-
ing observed and note the observer loca-
tion relative to the source and the sun.
Indicate the potential and actual emission
points on the sketch.
5.2.2 Indoor Location. Record the fol-
lowing information on the field data sheet
(Figure 22-2): company name, industry,
process unit, observer's name, observer's
affiliation, and date. Record as appropri-
ate the type, location, and intensity of
lighting on the data sheet. Sketch the pro-
cess unit being observed and note observ-
er location relative to the source. Indicate
the potential and actual fugitive emission
points on the sketch.
5.3 Indoor Lighting Requirements. For
indoor locations, use a light meter to mea-
sure the level of illumination at a location
as close to the emission source(s) as is
feasible. An illumination of greater than
100 lux (10 foot candles) is considered
necessary for proper application of this
method.
5.4 Observations. Record the clock
time when observations begin. Use one
stopwatch to monitor the duration of the
observation period; start this stopwatch
when the observation period begins. If the
observation period is divided into two or
more segments by process shutdowns or
observer rest breaks, stop the stopwatch
when a break begins and restart it without
[Part 60, Appendix A, Method 22]
10-16-92
Environment Reporter
0013-9211/92/SO+.50
672
-------
APPENDIX F
DEVELOPMENT OF LEAK RATE
VERSUS FRACTION LEAKING EQUATIONS
AND DETERMINATION OF LDAR CONTROL EFFECTIVENESS
-------
APPENDIX F
The purpose of this appendix is to provide additional
information on the approach used to develop the average leak rate
versus fraction leaking equations presented in Chapter 5.0.
Also, background information is presented on the determination of
control effectiveness of LDAR programs at SOCMI process units and
refinery process units.
F.I DEVELOPMENT OF AVERAGE LEAK RATE VERSUS FRACTION LEAKING
EQUATIONS
In Chapter 5.0, Tables 5-4 and 5-5 present equations that
predict average leak rate based on the fraction leaking at SOCMI
process units and refinery process units, respectively.
Equations are presented for gas valves, light liquid valves,
light liquid pumps, and connectors, and each of the equations are
plotted in Figures 5-1 through 5-8.
The equations are expressed in the following format:
Average Leak Rate = (Slope * Fraction Leaking) + Intercept
The average leak rate has units of kilograms per hour per source.
The fraction leaking is the fraction of sources that screen
greater than or equal to the applicable leak definition. The
leak definition is the screening value at which a leak is
indicated. (For example an equipment leak regulation may have a
leak definition of 10,000 ppmv.) Equations were developed for
several possible leak definitions.
Using the applicable equation, if it is known what
percentage of sources screen greater than or equal to the leak
definition, then an overall average leak rate for all sources can
be estimated. If the fraction leaking before and after an LDAR
program is implemented are known, then the average leak rates
before and after the program can be determined. These average
leak rates before and after the program are used to calculate the
control efficiency of the program.
F-l
-------
The leak rate versus fraction leaking equations were
developed using the following procedure:
STEP 1: Determine average emission factors for
(1) screening values greater than or equal to the
applicable leak definition, and (2) screening
values less than the applicable definition.
STEP 2: The average emission factor for screening values
less than the leak definition is the intercept in
the equation.
STEP 3: The average emission factor for screening values
greater than or equal to the leak definition minus
the average emission factor for screening values
less than the leak definition is the slope in the
equation.
An example of the above steps is presented for gas valves in a
SOCMI process units for a leak definition of 10,000 ppmv. From
Table 2-4 the gas valve >10,000 ppmv emission factor is
0.0782 kg/hr and the <10,000 ppmv factor is 0.000131 kg/hr.
Thus, the equation relating average leak rate to fraction leaking
for SOCMI gas valves with a leak definition of 10,000 ppmv is as
follows:
Avg Leak Rate (kg/hr) = [(0.0782-0.000131) * FL] + 0.000131
= (0.0781 * FL) + 0.000131
where:
FL = Fraction leaking.
Notice that when applying the above equation if 100 percent of
the gas valves screened less than 10,000 ppmv, the equation
predicts an average leak rate equal to the <10,000 ppmv factor.
Similarly, if 100 percent of sources screened greater than or
equal to 10,000 ppmv, the equation predicts an average leak rate
equal to the >10,000 ppmv factor.
For SOCMI process units, equations were developed for each
of the equipment types for leak definitions of 500 ppmv,
1,000 ppmv, 2,000 ppmv, 5,000 ppmv, and 10,000 ppmv. For each of
the leak definitions, the greater than or equal to factors and
the less than factors were developed by entering the applicable
F-2
-------
screening data from the combined screening data set into the
applicable revised SOCMI correlation equation (see Appendix B).
For example, the <500 ppmv factor for connectors was estimated by
entering all connector screening data with values less than
500 ppmv from the combined screening dataset into the revised
SOCMI connector correlation equation. The sum of total emissions
divided by the number of screening values gives the <500 ppmv
connector average emission factor.
For refinery process units, equations were developed for
each of the equipment types for leak definitions of 500 ppmv,
1,000 ppmv and 10,000 ppmv. The refinery >10,000 ppmv and
<10,000 ppmv emission factors had previously been developed and
are presented in Table 2-5. The same approach used to develop
the >:10,000/<10,000 ppmv refinery factors was used to develop the
factors for leak definitions of 500 ppmv and 1,000 ppmv. This
approach involves using information from the Refinery Assessment
Study (EPA-600/2-80-075c) on the cumulative distribution of
emissions and screening values.
F.2 CONTROL EFFECTIVENESS CALCULATIONS
In addition to the equations described in Section F.I,
Chapter 5.0 presents estimated control effectiveness values at
SOCMI and refinery process units for control equivalent to:
(1) Monthly LDAR program with a leak definition of
10,000 ppmv;
(2) Quarterly LDAR program with a leak definition of
10,000 ppmv; and
(3) Control equivalent to the LDAR program required by the
proposed hazardous organic NESHAP equipment leaks
negotiated regulation.
Tables F-l and F-2 summarize how the control effectiveness values
of the above LDAR programs were determined for SOCMI and refinery
process units, respectively.
The approach for calculating the control effectiveness of a
LDAR program is discussed in detail in Chapter 5.0. The approach
involves determining the average leak rate before and after the
LDAR program is implemented. The average leak rates before and
F-3
-------
CO
H
CO
CO
W
u
O
H
§
O
CO
EH
CO
CO
H
1-1
EFFECT:
o
«
H
§
O
S
Q
i-5
,
O
O
H
E-i
H
Q
H
1
fa
W
£?
^4
E-i
_, M
2 | £
gj|
ili
s *•
11*1
ul •"* S 5
^
i i?
u •
£ *N
• 4-1
So X J
Si £Sg
5; 1-2 §
*1 l^s
i 1 *" ~*
41 — •
•M 0,
5.E i«
1 53J
I ffel
5 l^1
0
"3 « ^
4-> ^
"^
JC
S2
_ c
— ** o
ID U ^
.?*: i
'c **
C
j< *J ?
51 i
ol
u u
s?
u Q.
„
i *
& &
'!"
UJ
a»- CO N. N. /
•o So eo >o
•* O> o in in
>o in in F. o>
g § i I §
000 00
•* O O O> Kl
in v o s. K>
O «- «- O CM
s s g as
O CM CM ^ Kl
s s; g a §
0 O O O O
o o o S S
o o o o o
o o o o o
Kt KI in in in
•* •* co r^ s.'
o o o o o
§00 o o
o in o o
o o o
"i* - $
2 tl 01 f «l
4-» 4-» 41 4-» 4J
C "- <- C L.
« 1 g l 1
Z
(I)
« >
— ~a
a >
a>
-i a
—I U
/
8!
in
0
g
g
CM
g
O
g
O
o
•o
Kl
O
o
in
1
S1
s
i
•
o
fc
in
K!
g
o
i
o
o
in
F*-
g
O
|
J
3/
i
0
o
rC
0
in
K
g
0
i
o
o
in
r>^
o
o
o
o
X
Quarte
rO
*•
o
o
8j
•*
CO
o
o
o
1
o
0
*-t
!*•
g
o
t
?
8
I
•
O
!Q
o
o
in
o
o
o
o
CO
i
0
0>
K»
g
in
I
I
to
0
y
1
u
F-4
-------
CO
H
CO
CO
W
u
CU
>4
i
H
S5
H
fa
W
|^4
CO
CO
H
B
H
fa
H
i
8
B$
1
fo
O
o
H
EH
s
H
i
Q
CM
fa
W
OQ
E-i
2*C
g>£
8-52
||1
l!i
"" 8 vi
w
- <0
&2
•
^
** ** c
.2 «- i
I1!
li •
• ""* a
i *•
•S •
-- »J *»
•- o ro o o
o «- o o «- o
o> OK OK co oo co
o o o o o o
o o o o o o
o o in o e o
«- »- CO O O ~»
«- »- CM ^ «- CM
too o o o
o o o o o
o in o o in
* * t » s t
f. t> a> £ « a
Ml M|
! 1
— • a
S *
d I
j
eg uf
N. 0
- g
s §
O 0
st 53
iC
0
o
in
o
g
0
1
0
K
O
o
in
1
n
|
1
$
u
F-5
-------
after implementing the LDAR program are estimated by entering the
fraction leaking before and after implementing the program into
the equations described in Section F.I.
For SOCMI process units, the fraction leaking before
implementing the LDAR program was based on the percentage of
equipment screening above the applicable leak definition in the
combined SOCMI screening dataset. (See Appendix B.) Similarly,
the initial fraction leaking for refinery process units was based
on data from the Refinery Assessment Study on the percentage of
equipment screening above the applicable leak definition. Note
that each of the initial leak fractions predict leak rates equal
to the applicable SOCMI or refinery average emission factors
(Tables 2-1 and 2-2) when entered into the applicable equation
described in Section F.I. In other words, when estimating the
control effectiveness for the SOCMI and refinery LDAR programs,
it has been assumed that prior to implementing the program
equipment leak emissions are equivalent to emissions that would
be predicted by the average emission factors.
The fraction leaking after implementing the LDAR program is
assumed to be the average of the "steady-state" fraction leaking
immediately before and after a monitoring cycle (see discussion
in Chapter 5.0). The following parameters are used to estimate
the steady-state leak fractions:
• recurrence rate,
• unsuccessful repair rate, and
• occurrence rate.
The values used for these parameters are summarized in
Table F-3 for both SOCMI and refinery process units.
The paragraphs below summarize the approach used to
determine the above parameters. First, the approach used to
determine the parameters in a program with a leak definition of
10,000 ppmv is described. Then, the approach used to determine
the parameters in a program equivalent to the proposed hazardous
organic NESHAP equipment leaks negotiated regulation is
described.
F-6
-------
TABLE F-3.
PARAMETERS USED TO CALCULATE STEADY-STATE LEAK
FRACTION AFTER LDAR PROGRAM IS IMPLEMENTED
Equipment type
PARAMETER VALUES
LI Valves
Gas Valves
LL Pumps
Connectors
PARAMETER VALUES
LL Valves
Gas Valves
LL Pumps
Connectors
Leak
Control definition
program (ppmv)
Recurrence
rate8
(percent)
Unsuccessful
repair rate8
(percent)
Initial leak
fraction"
(percent)
Occurrence
rate0
(percent)
FOR SOCHI PROCESS UNITS
Monthly
Quarterly
HON reg neg
Monthly
Quarterly
HON reg neg
Monthly
Quarterly
HON reg neg
HON reg neg
FOR REFINERY PROCESS
Monthly
Quarterly
HON reg neg
Monthly
Quarterly
HON reg neg
Monthly
Quarterly
HON reg neg
HON reg neg
10000
10000
500
10000
10000
500
10000
10000
1000
500
UNITS
10000
10000
500
10000
10000
500
10000
10000
1000
500
14
H
0
H
14
0
0
0
0
0
U
H
0
14
14
0
0
0
0
0
10
10
0
10
10
0
0
0
0
0
10
10
0
10
10
0
0
0
0
0
4.3
4.3
8.5
7.5
7.5
13.6
7.5
7.5
17.1
3.9
11.0
11.0
28.5
10.0
10.0
24.0
24.0
24.0
48.0
1.7
0.68
2.03
2.00
1.00
2.97
2.00
3.53
7.50
8.04
0.50
1.34
3.97
2.00
1.24
3.67
2.00
11.28
24.00
10.00
0.50
a The recurrence rate and unsuccessful repair rate for valves and pumps in LDAR programs with a leak
definition of 10,000 ppmv was obtained from the SOCMI Fugitives AID (EPA-450/3-82-010). For the HON
reg neg, a simplifying assumption was made that the recurrence rate and unsuccessful repair rate equal
zero percent for all equipment types.
b The initial leak fraction for SOCMI process units is based on the combined screening dataset. The initial
leak fraction for refinery process units is based on data collected in the Refinery Assessment Study
(EPA-600/2-8--075C).
c The occurrence rate for LDAR programs with a leak definition of 10,000 ppmv is calculated as a function of
the initial leak fraction. The relationship is based on data collected in the Six Unit Maintenance Study
(EPA-600/S2-081-080). The equations for valves and pumps are as follows:
Valve 30 Day Occurrence rate * 0.0976 * leak fraction + 0.264.
Pump 30 Day Occurrence rate = 0.47 * leak fraction.
The quarterly occurrence rate is approximately 3 times the 30-day occurrence rate. In cases where the
quarterly occurrence rate exceeded the initial leak fraction, it was set equal to the initial leak
fraction. The occurrence rate for the HON reg neg LDAR programs is set equal to the performance level,
except for pumps in SOCMI process units. For pumps in SOCMI process units the occurrence rate is
calculated using the equation above.
F-7
-------
F.2.1 LDAR Program with Leak Definition of 10,000 ppmv.
Estimates for the recurrence rate and unsuccessful repair
rate were obtained from the Fugitive Emissions Additional
Information document (EPA-450/3-82-010). In this document, data
collected for LDAR programs with a leak definition of 10,000 ppmv
were summarized. It was concluded that the recurrence rate for
valves was 14 percent and the unsuccessful repair rate for valves
10 percent. It was assumed that all pumps are replaced with a
new seal and for that reason the recurrence rate and unsuccessful
repair rate for pumps were both assumed equal to zero percent
(i.e., all pumps are successfully repaired and leaks do not
recur). Data were unavailable for connectors for an LDAR program
with a leak definition of 10,000 ppmv, and, for this reason,
control efficiency for connectors in an LDAR program with a leak
definition of 10,000 ppmv have not been estimated.
Estimates for the occurrence rate were based on data
collected in the Six Unit Maintenance Study (EPA-600/S2-081-080).
Data from this study indicated that the occurrence rate is a
function of the initial leak fraction. For valves this
relationship was expressed by the following equation:
OCCvalve = 0.0976 (LF) + 0.264
where:
OCCvalve = Monthly occurrence rate for valves;
and
LF = Initial leak fraction.
For pumps, the relationship was as follows:
OCCpump = °'47 * LF
where:
occpunp = MontnlY occurrence rate for pumps; and
LF = Initial leak fraction.
F-8
-------
For both pumps and valves, the monthly occurrence rate was used
to estimate the quarterly occurrence rate using the following
equation:
Q = M + M (1 - M) + M {1 - [M + M (1 - M)]}
where:
M = Monthly occurrence rate; and
Q = Quarterly occurrence rate.
Note that in cases where the estimated quarterly occurrence rate
exceeded the initial leak fraction, it was set equal to the
initial leak fraction.
F.2.2 Control Equivalent to the LDAR Program Required by the
Proposed Hazardous Organic NESHAP Equipment Leaks
Negotiated Regulation
For each of the equipment types, the proposed hazardous
organic NESHAP LDAR program requirements include a performance
level requirement. This performance level specifies the
allowable leak fraction once the program is in place. For
example, the performance level for valves is 2 percent. Because
the proposed hazardous organic NESHAP rule contains the
performance level requirement and because limited data are
available on LDAR programs with the leak definitions of the
proposed hazardous organic NESHAP rule, simplifying assumptions
were made when estimating the recurrence rate, unsuccessful
repair rate, and occurrence rate.
For each of the equipment types, it was assumed that the
recurrence rate and unsuccessful repair rate were equal to zero
percent. These two parameters have the least impact on the
predicted control efficiency.
For valves and connectors, the proposed hazardous organic
NESHAP rule allows for reduced monitoring frequency if the leak
fraction remains below the performance level. For this reason,
it was assumed that process units would monitor valves and
connectors at whatever monitoring frequency (i.e., monthly,
quarterly, annually, etc.) that allows them to meet the
performance level. Thus, for valves and connectors the
F-9
-------
occurrence rate was set equal to the performance level. Note
that in cases where process units remain below the performance
level this may overestimate the occurrence rate. However, this
is offset by the assumption that the recurrence rate and
unsuccessful repair rate are equal to zero percent.
For pumps the proposed hazardous organic NESHAP rule
requires monthly monitoring. For this reason the occurrence rate
was calculated using the same equation for pumps as presented in
Section F.2.1 for LDAR programs with a leak definition of
10,000 ppmv. Note, however, that the initial leak fraction used
in the equation was the leak fraction associated with the leak
definition of the proposed hazardous organic NESHAP rule
(1,000 ppmv). For refineries, the predicted occurrence rate for
pumps exceeded the performance level, and for this reason the
occurrence rate was set equal to the performance level.
F-10
-------
TECHNICAL REPORT DATA
iPleatt retd Instructions on ikt rtvtrse btfort compliant/
1. REPORT NO.
3. RECIPIENT'S ACCESSION NO
4. TITLE AND SU8TITL6
Protocol for Equipment Leak Emission Estimates
5. REPORT DATE
June 1993
6. PERFORMING ORGANIZATION CODE
7. AOTHOR(S)
Kenneth J. Hausle
a. PERFORMING ORGANIZATION REPORT \
DCN: 93-239-026-85-02 '
. PERFORMING ORGANIZATION NAME AND ADDRESS
Radian Corporation
3200 E. Chapel Hill Road/Nelson Highway
P.O. Box 13000
Research Triangle Park, NC 27709
10. PROGRAM ELEMENT NO
11 CONTRACT/GRANT NO.
68-D1-0117, WA 85
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
14 SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
Project Officer is David W. Markwordt, Mail Drop 13, (919)541-0837
16. ABSTRACT
The report presents standard protocols for estimating mass emissions from
equipment leaks. Different approaches for estimating equipment leak emissions are
described and several topics relevant to estimating equipment leak emissions (such
as speciating emissions) are addressed. Information on how to perform a screening
survey at a process unit is presented. Information on how a process unit can collect
equipment leak rate data by enclosing individual equipment pieces and measuring
mass emissions is provided. Also, information is provided which can be used to
estimate the control efficiency of equipment leak control techniques. The document
will help facilities generate accurate plant-specific equipment leak emissions
estimates.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFfERS/OPEN ENDED TERMS
c. COSATi Field.Croup
Synthetic Organic Chemical
Industry
Petroleum Refining
Organic Compounds
Gas Plants
Manufacturing
Fugitive Emissions
Equipment Leaks
Response Factors
Screening Data
Bagging Data
Leak Detection and
Repair
18. DISTRIBUTION STATEMENT
Release to Public
19 SEO'JRITY CLASS tTliis Report!
Unclassified
21. NO. Of 'AC
20 SECURITY CLASS iT/ns pagti
Unclassified
22. PRICE
EPA
2220-) (*•». 4-77) P*«VIOU» eoi TION is O«SOU«TS
------- |