United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-453/R-95-017
November 1995
Air
Protocol for Equipment Leak
Emission Estimates

-------
                                  EPA-453/R-95-017
1995  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
                  November 1995

-------
     This report has been reviewed by the Emission Standards
Division of the Office of Air Quality Planning and Standards, the
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-95-017

-------
                             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,
the 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 procedures
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 recently collected in the petroleum
industry has been used to revise the existing refinery
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 	  2-3

          2.2.1  Equipment Leak Emission Estimation
                 Approaches	2-3
          2.2.2  Overview of Equipment Leak Data Collection .  2-6

     2.3  Approaches for Estimating Equipment Leak
          Emissions	2-10

          2.3.1  Average Emission Factor Approach 	  2-10
          2.3.2  Screening Ranges Approach  	  2-18
          2.3.3  EPA Correlation Approach	2-24
          2.3.4  Unit-Specific Correlation Approach ....  2-38

     2.4  Special Topics	2-44

          2.4.1  Speciating Emissions 	  2-45
          2.4.2  Using Response Factors 	  2-46
          2.4.3  Monitoring Instrument Type and Calibration
                 Gas	2-51
          2.4.4  Estimating Emissions for Equipment Not
                 Screened	2-51
          2.4.5  Using Screening Data Collected at Several
                 Different Times  	  2-52
          2.4.6  Estimating VOC Emission Rates from Equipment
                 Containing Non-VOC's 	  2-52
          2.4.7  Estimating Equipment Leak Emissions of
                 Inorganic Compounds  	  2-53

     2.5  References	2-54

3.0  SOURCE SCREENING 	  3-1

     3.1  Introduction	3-1
     3.2  Monitoring Instruments  	  3-2

          3.2.1  Operating Principles and Limitations of
                 Portable VOC Detection Devices 	  3-2
                               11

-------
                  TABLE OF CONTENTS (Continued)

Section                                                      Page
          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
                 EPA Reference 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	3-15
          3.3.3  Data Handling	3-25

     3.4  References	3-27

4.0  MASS EMISSION SAMPLING	4-1

     4.1  Introduction	4-1
     4.2  Sampling Methods  	  4-2

          4.2.1  Vacuum Method	4-4
          4.2.2  Blow-Through Method  	  4-8

     4.3  Source Enclosure  	   4-15

          4.3.1  Valves	4-15
          4.3.2  Pumps and Agitators	4-16
          4.3.3  Compressors	4-17
          4.3.4  Connectors	4-17
          4.3.5  Relief Valves	4-18

     4.4  Analytical Techniques 	   4-18

          4.4.1  Analytical Instrumentation 	   4-19
          4.4.2  Calibration of Analytical Instruments  .  .   4-19
          4.4.3  Analytical Techniques for Condensate .  .  .   4-21
          4.4.4  Calibration Procedures for the Portable
                 Monitoring Instrument  	   4-21

     4.5  Quality Control and Quality Assurance Guidelines   4-21

          4.5.1  Quality Control Procedures 	   4-22
          4.5.2  Quality Assurance Procedures 	   4-25

     4.6  References	4-29

5.0  ESTIMATION OF CONTROL EFFICIENCIES FOR EQUIPMENT LEAK
     CONTROL TECHNIQUES 	  5-1

     5.1  Introduction	5-1


                               iii

-------
                  TABLE OF CONTENTS (Continued)

Section                                                      Page


     5.2  Equipment Modification Control Efficiency 	 5-1

          5.2.1  Closed-Vent Systems  	 5-2
          5.2.2  Pumps	5-2
          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 	 5-6
          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  	 5-8
          5.3.2  Example Application of Approach  	   5-57

     5.4  References	5-61


APPENDICES

  A  Example Calculations 	 A-l

  B  Leak Rate Screening Value Correlation Development and
     Revision of SOCMI Correlations and Emission Factors  .   . B-l

  C  Revision of Petroleum Industry Correlations and Emission
     Factor	C-l

  D  Response Factors 	 D-l

  E  Selection of Sample Size for Screening Connectors  .  .   . E-l

  F  Reference Method 21  	 F-l

  G  Development of Leak Rate Versus Fraction Leaking
     Equations and Determination of LDAR Control
     Effectiveness  	 G-l
                                IV

-------
                         LIST OF TABLES

Table                                                        Page

 2-1      SOCMI Average Emission Factors  	   2-12

 2-2      Refinery Average Emission Factors 	   2-13

 2-3      Marketing Terminal Average Emission Factors .  .  .   2-14

 2-4      Oil and Gas Production Operations Average Emission
          Factors	2-15

 2-5      SOCMI Screening Ranges Emission Factors 	   2-19

 2-6      Refinery Screening Ranges Emission Factors  .  .  .   2-20

 2-7      Marketing Terminal Screening Ranges Emission
          Factors	2-21

 2-8      Oil and Gas Production Operations Screening Ranges
          Emission Factors  	   2-22

 2-9      SOCMI Leak Rate/Screening Value Correlations  .  .   2-26

 2-10     Petroleum Industry Leak Rate/Screening Value
          Correlations  	   2-27

 2-11     Default-Zero Values:   SOCMI Process Units ....   2-33

 2-12     Default-Zero Values:   Petroleum Industry  ....   2-34

 2-13     10,000 and 100,000 ppmv Screening Value Pegged Emission
          Rates For SOCMI Process Units	2-36

 2-14     10,000 and 100,000 ppmv Screening Value Pegged Emission
          Rates For the Petroleum Industry	2-37

 3-1      Performance Criteria for Portable VOC Detectors  .  .3-6

 3-2      Portable VOC 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	4-20

                                v

-------
                   LIST OF TABLES  (Continued)

Table                                                        Page
 4-4      Example Data Collection Form For Fugitive
          Emissions Bagging Test (Blow-Through Method)   .  .   4-23

 4-5      Example Data Collection Form For Fugitive
          Emissions Bagging Test (Vacuum Method)   	   4-24

 4-6      Example Drift Test Report Form	4-28

 5-1      Summary of Equipment Modifications  	 5-3

 5-2      Control Effectiveness For an LDAR Program At a
          SOCMI Process Unit	5-9

 5-3      Control Effectiveness For an LDAR Program At a
          Refinery Process Unit	5-10

 5-4      Equations Relating Average Leak Rate to Fraction
          Leaking at SOCMI Units  	   5-46

 5-5      Equations Relating Average Leak Rate to Fraction
          Leaking at Refinery Units 	   5-47

 5-6      Equations Relating Average Leak Rate to Fraction
          Leaking At Marketing Terminal Units 	   5-48

 5-7      Equations Relating Average Leak Rate to Fraction
          Leaking At Oil and Gas Production Operating
          Units	5-50

 5-8      Values Used in Example Calculation	5-58

 5-9      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	5-60
                               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
          (TOO Leak Rate to Screening Value:
          0 - 1,000 ppmv	2-28

 2-3      SOCMI Correlations Relating Total Organic Compound
          (TOO Leak Rate to Screening Value:
          1,000 - 1,000,000 ppmv	2-29

 2-4      Petroleum Industry Correlations Relating Total Organic
          Compound (TOO  Leak Rate to Screening Value:
          1,000 - 1,000,000 ppmv	2-30

 2-5      Petroleum Industry Correlations Relating Total Organic
          Compound (TOO  Leak Rate to Screening Value:
          1,000 - 1,000,000 ppmv	2-31

 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 	   3-23

 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      Marketing Terminal Gas Fittings Average Leak Rate
          Versus Fraction Leaking at Several Leak
          Definitions	5-19

 5-10     Marketing Terminal Light Liquid Fittings Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-20

 5-11     Marketing Terminal Gas Others Average Leak Rate Versus
          Fraction Leaking at Several Leak Definitions  .   .  5-21

 5-12     Marketing Terminal Light Liquid Others Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-22

 5-13     Marketing Terminal Light Liquid Pumps Average Leak Rate
          Versus Fraction Leaking at Several Leak
          Definitions	5-23

 5-14     Marketing Terminal Gas Valves Average Leak Rate Versus
          Fraction Leaking at Several Leak Definitions  .   .  5-24

 5-15     Marketing Terminal Light Liquid Valves Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-25

 5-16     Oil and Gas Production Gas Connectors Average Leak Rate
          Versus Fraction Leaking at Several Leak
          Definitions	5-27
                              VI11

-------
                   LIST OF FIGURES (Continued)
Figure
 5-17     Oil and Gas Production Light Oil Connectors Average
          Leak Rate Versus Fraction Leaking at Several Leak
          Definitions	5-28

 5-18     Oil and Gas Production Water/Oil Connectors Average
          Leak Rate Versus Fraction Leaking at Several Leak
          Definitions	5-29

 5-19     Oil and Gas Production Gas Flanges Average Leak Rate
          Versus Fraction Leaking at Several Leak
          Definitions	5-30

 5-20     Oil and Gas Production Light Oil Flanges Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-31

 5-21     Oil and Gas Production Gas Open-Ended Lines Average
          Leak Rate Versus Fraction Leaking at Several Leak
          Definitions	5-32

 5-22     Oil and Gas Production Heavy Oil Open-Ended Lines
          Average Leak Rate Versus Fraction Leaking at Several
          Leak Definitions	5-33

 5-23     Oil and Gas Production Light Oil Open-Ended Lines
          Average Leak Rate Versus Fraction Leaking at Several
          Leak Definitions	5-34

 5-24     Oil and Gas Production Water/Oil Open-Ended Lines
          Average Leak Rate Versus Fraction Leaking at Several
          Leak Definitions	5-35

 5-25     Oil and Gas Production Gas Other Average Leak Rate
          Versus Fraction Leaking at Several Leak
          Definitions	5-36

 5-26     Oil and Gas Production Heavy Oil Other Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-37

 5-27     Oil and Gas Production Light Oil Other Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-38

 5-28     Oil and Gas Production Water/Oil Other Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-39
                                IX

-------
                   LIST OF FIGURES (Continued)

Figure                                                       Page


 5-29     Oil and Gas Production Gas Pumps Average Leak Rate
          Versus Fraction Leaking at Several Leak
          Definitions	5-40

 5-30     Oil and Gas Production Light Oil Pumps Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-41

 5-31     Oil and Gas Production Gas Valves Average Leak Rate
          Versus Fraction Leaking at Several Leak
          Definitions	5-42

 5-32     Oil and Gas Production Heavy Oil Valves Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-43

 5-33     Oil and Gas Production Light Oil Valves Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-44

 5-34     Oil and Gas Production Water/Oil Valves Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-45

 5-35     Marketing Terminals Light Liquid Pumps Average Leak
          Rate Versus Fraction Leaking at Several Leak
          Definitions	5-55
                                x

-------
                         1.0  INTRODUCTION


    This document is an update to the EPA equipment leaks

protocol document ("Protocol for Equipment Leak Emission

Estimates," EPA-453/R-93-026, June 1993).   The purpose of this

document is the same as the original protocol document and

subsequent revisions- to present standard procedures 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 and

subsequent revisions were published, and also expands on some of

the topics that were covered in the original protocol.

    Some of the new features of the updated protocol are:

    (1) New correlation equations, default zero emission  rates,
       and pegged emission rates for the petroleum  industry that
       replace the refinery  correlations previously published
       are presented.  The correlations relate screening values
       obtained using a portable monitoring  instrument to mass
       emissions.

    (2) The document has been expanded to include emission
       factors for marketing terminals and for oil  and gas
       production operations.  The  refinery  emission factors
       were not revised due  to an unavailability of new  data.

    (3) Pegged emission rates for pegged readings at 10,000 ppmv
       have been added for SOCMI process units.

    (4) Several of the equations in  this version of  the protocol
       have been revised by  simplifying the  symbols to more
       clearly communicate the concept being conveyed.

    (5) An adjustment has been added to the blow-through  method
       of calculating mass emissions.  This  adjustment more
       accurately accounts for the  total flow through the bag.
                               1-1

-------
   As with the original protocol document, this document
presents standard procedures 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
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.
   Chapter 2.0 presents the four approaches for estimating total
organic emissions from equipment leaks.  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 volatile organic compounds, estimating
emissions of inorganic compounds,  and other topics not
specifically related to any one of the four approaches.
   Chapter 3.0 explains 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 the
EPA Reference Method 21.  Additionally, in chapter 3.0,  guidance
is provided on how to set up a screening program and how to
screen different types of equipment.
                               1-2

-------
   Chapter 4.0, explains how to collect equipment leak rate data
(bagging data) by enclosing individual equipment in a "bag" and
measuring mass emissions.  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 bagging data to generate unit-specific
correlations.  These steps are intended to ensure that the data
are of high quality.
   Chapter 5.0, explains how to estimate the control efficiency
of equipment leak emission control 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 G  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 presents the rationale
for the development of the petroleum industry correlations, as
well as the background for the development of marketing terminal
and oil and gas production operations emission factors.
Appendix D summarizes available data on response factors.
Appendix E provides guidance on how to collect representative
screening data for connectors.   Appendix F contains a copy of the
EPA Reference Method 21.  Finally, appendix G 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 the methods for
estimating mass 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 the
petroleum industry and the synthetic organic chemical
manufacturing industry  (SOCMI).  The SOCMI emission factors and
correlations were revised and introduced in the 1993 update of
this document.  The refinery correlations that have been revised
and expanded to include the entire petroleum industry are
introduced in this document.  Additionally, emission factors for
marketing terminals are introduced in this document.  Emission
factors for gas plants that have been updated and expanded to
included oil and gas production operations are also introduced in
this document.  The procedures in this document estimate
emissions of total organic compounds (TOC's).  However, special
procedures are also described for the purpose of estimating
volatile organic compounds  (VOC's).   As defined by the EPA, VOC's
                               2-1

-------
include all organic compounds except those specifically 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;
    •  Using screening data collected at different  times;
    •  Estimating VOC emissions from equipment  containing
       organic compounds excluded from  the EPA's  classification
       of TOC's; and
    •  Estimating emissions from equipment containing  inorganic
       compounds.
    Appendices A through E 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 presents
background information on the development of average emission
factors and correlation equations for the petroleum industry.
Appendix D offers a detailed listing of available response
factors.   Appendix E 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

-------
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 TOC or 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
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).
                               2-3

-------
        Count Equipment Components
           (by type and service)
            Conduct Complete
            Screening Survey
                6
                6
                            Approach 1
                                                  Apply Average Emission Factors
                                                  and Composite Total Emissions
                                                    Inventory
  Approach 2
     Apply>10,000/
<10,000 ppmv Emission Factors
and Composite Total Emissions
                                                    Inventory
  Approach 3
                          Apply EPA Correlations and
                          Composite Total Emissions
                           Inventory
          Bag Components for Each
         Equipment Type and Service
           Develop Unit-Specific
              Correlations
                           Apply New Correlations
                         and Composite Total Emissions
                           Inventory    Section 2.3.4
Figure  2-1.
Overview of  Data  Collection  and Analysis
Approaches  for  Developing  Equipment  Leak
Emissions  Inventory
                                                 2-4

-------
    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 procedures 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.  Procedures for collecting bagging data
are described in detail in chapter 4.0.
    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 SOCMI, refineries, marketing terminals
and oils and gas production operations have yielded emission
factors and correlations for these source categories.  Emission
factors and correlations for 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
                               2-5

-------
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
    Data on equipment leak emissions of organic compounds have
been collected from refineries, marketing terminals, oil and gas
production operations, and SOCMI process units.  Emission factors
and correlations have been developed for the following equipment
types:  valves, pumps, compressors,  pressure relief valves,
connectors, flanges, and open-ended lines.  An "others" category
has also been developed for the petroleum industry.  For sampling
connections,  an average emission factor has been developed that
estimates the typical amount of material purged when a sample is
collected.  A brief history of the development of these factors
and correlations is presented below.
    2.2.2.1  Refinery Assessment Study.1'2  In the late 1970s,
the 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
                               2-6

-------
(1)  the phase of the process stream (service),  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  Revised Petroleum Industry Correlations and Emission
Factors.   During the early-1990's,  new petroleum industry
equipment leak bagging data were collected and analyzed.  The
Western States Petroleum Association  (WSPA)  and the American
Petroleum Institute  (API)  jointly commissioned the 1994 refinery
equipment leak report4 to evaluate fugitive emissions collected
from five petroleum refineries.  The API also commissioned the
1993 marketing terminal equipment leak report,5 which included
bagging data from three marketing terminals, and, along with the
Gas Research Institute (GRI),  jointly commissioned the 1993 and
1995 oil and gas production operations reports,  which included
bagging data from 24 facilities . 6/7  In addition to the bagging
data, screening data were also collected from 17 marketing
terminals8 and 24 oil and gas production facilities.6/7  Data
from gas/vapor,  light liquid,  and/or heavy liquid streams were
collected for these studies from non-flanged connectors, flanges,
open-ended lines,  pumps,  values,  instruments, loading arms,

                               2-7

-------
pressure relief valves, stuffing boxes, vents,  compressors,  dump
lever arms, diaphrams, drains, hatches, meters,  and polished
rods.
    A specific goal of the above studies was to collect high
quality data to enhance or replace the previously published
refinery correlations.  As a result of the analyses discussed in
appendix C, the bagging data collected from refineries, marketing
terminals,  and oil and gas production facilities during the
early-1990's were combined to replace the previously published
refinery correlations with correlations applicable to the entire
petroleum industry.  In addition, the new correlations apply
across all services for a given equipment type.   The previously
published refinery correlations were specific to service and
equipment.
    The screening data were used to develop average emission
factors for marketing terminals and for oil and gas production
operations.  The average emission factors for oil and gas
production operations replace the gas plant factors published in
previous versions of this document and apply to light crude,
heavy crude, gas plant, gas production and off shore facilities.
No new screening data were available for refineries, therefore
the previously published refinery average emission factors remain
unchanged in this version of the protocol.  Appendix C contains
more detailed information on how the new petroleum industry
correlations, marketing terminal emission factors, and oil and
gas production operations emission factors were developed.
    2.2.2.4  Original SOCMI Average Emission Factors and
Correlations.  In 1980, two studies were coordinated by the EPA
to collect data from SOCMI process units.  These studies were the
24-Unit Study,9 and the Six-Unit Maintenance Study.10  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
                               2-8

-------
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 = (F x RLF)  + (1 - F) x RNLF
where:
    F     =  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
                               2-9

-------
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.
   2.2.2.5  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 for estimating TOC
emissions.  Each of the approaches are demonstrated in example
calculations contained in appendix A.  Special topics at the end
of the chapter have been included to address how to estimate VOC
emissions when some of the organic compounds in the stream are
not classified as VOC's and also how to speciate emissions for
individual chemicals from equipment containing a mixture.
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 TOC concentration
                               2-10

-------
of the stream (and VOC or HAP concentrations if speciation is to
be performed) ,  and (4) the time period each component was in that
service.  The average emission factors for SOCMI process units,
refineries, marketing terminals,  and oil and gas production
operations are presented in tables 2-1, 2-2, 2-3,  and 2-4
respectively.  The SOCMI, marketing terminal, and oil and gas
production operations 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 TOC in weight fraction within the
equipment is needed because equipment with higher TOC
concentrations tend to have higher TOC leak rates.   When using
the Average Emission Factor Approach, equipment should be grouped
into "streams" where all the equipment within the stream have
approximately the same TOC weight fraction.
    To apply the average emission factors,  use the following
equation to estimate TOC mass emissions from all of the equipment
in a stream of a given equipment type:
                           = FA x WF>POC x
where :
                    Emission rate of TOC from all equipment in
                    the stream of a given equipment type (kg/hr) ;
                               2-11

-------
            TABLE  2-1.   SOCMI AVERAGE  EMISSION  FACTORS
Equipment type
Valves

Pump sealsb
Compressor seals
Pressure relief valves
Connectors
Open-ended lines
Sampling connections
Emission factor3
Service (kg/hr/source)
Gas
Light liquid
Heavy liquid
Light liquid
Heavy liquid
Gas
Gas
All
All
All
0
0
0
0
0
0
0
0
0
0
.00597
.00403
.00023
.0199
.00862
.228
.104
.00183
.0017
.0150
aThese factors are for total organic compound emission
 rates.

bThe light liquid pump seal factor can be used to estimate the
 leak rate from agitator seals.
                               2-12

-------
          TABLE  2-2.   REFINERY AVERAGE  EMISSION  FACTORS3
Equipment type
Valves
Pump seals0
Compressor seals
Pressure relief valves
Connectors
Open-ended lines
Sampling connections
Emission factor
Service (kg/hr/source) b
Gas
Light liquid
Heavy liquid
Light liquid
Heavy liquid
Gas
Gas
All
All
All
0
0
0
0
0
0
0
0
0
0
.0268
.0109
.00023
.114
.021
.636
.16
.00025
.0023
.0150
aSource : Reference 2.
^These factors are for non-methane organic compound
 emission rates.

cThe light liquid pump seal factor can be used to estimate the
 leak rate from agitator seals.
                               2-13

-------
     TABLE 2-3.  MARKETING TERMINAL AVERAGE EMISSION FACTORS


                                         Emission factor
 Equipment type             Service       (kg/hr/source)a
Valves
Pump seals
Others (compressors
and others )k
Fittings (connectors
and flanges)0
Gas
Light Liquid
Gas
Light Liquid
Gas
Light Liquid
Gas
Light Liquid
1.3E-05
4.3E-05
6.5E-05
5.4E-04
1.2E-04
1.3E-04
4.2E-05
8.0E-06
aThese factors are for total organic compound emission rates
 (including non-VOC's such as methane and ethane).

bThe "other" equipment type should be applied for any equipment
 type other than fittings, pumps, or valves.

c"Fittings" were not identified as flanges or non-flanged
 connectors; therefore,  the fitting emissions were estimated by
 averaging the estimates from the connector and the flange
 correlation equations.
                               2-14

-------
  TABLE  2-4.   OIL AND  GAS  PRODUCTION OPERATIONS  AVERAGE  EMISSION
 	FACTORS  (kg/hr/source)	

                                              Emission Factor
Equipment Type
Valves

Pump seals
Others0

Connectors

Flanges
Open-ended lines
Service3
Gas
Heavy Oil
Light Oil
Water/Oil
Gas
Heavy Oil
Light Oil
Water/Oil
Gas
Heavy Oil
Light Oil
Water/Oil
Gas
Heavy Oil
Light Oil
Water/Oil
Gas
Heavy Oil
Light Oil
Water/Oil
Gas
Heavy Oil
Light Oil
Water/Oil
(kg/hr/source) b
4.5E-03
8.4E-06
2.5E-03
9.8E-05
2 .4E-03
NA
1.3E-02
2 .4E-05
8.8E-03
3.2E-05
7.5E-03
1.4E-02
2 .OE-04
7.5E-06
2 .1E-04
1.1E-04
3.9E-04
3.9E-07
1.1E-04
2.9E-06
2.0E-03
1.4E-04
1.4E-03
2 .5E-04
aWater/Oil emission factors apply to water streams in oil service
 with a water content greater than 50%, from the point of origin
 to the point where the water content reaches 99%.  For water
 streams with a water content greater than 99%,  the emission rate
 is considered negligible.
bThese factors are for total organic compound emission rates
 (including non-VOC's such as methane and ethane)  and apply to
 light crude, heavy crude, gas plant, gas production, and
 off shore facilities.  "NA" indicates that not enough data were
 available to develop the indicated emission factor.
cThe "other" equipment type was derived from compressors,
 diaphrams, drains, dump arms, hatches, instruments, meters,
 pressure relief valves, polished rods, relief valves, and vents.
 This "other" equipment type should be applied for any equipment
 type other than connectors, flanges, open-ended lines, pumps, or
 valves.
                               2-15

-------
    FA          =   Applicable average emission factor for the
                    equipment type (kg/hr/source);

                    FOR REFINERIES ONLY:  The emission factor
                    "FA" must be adjusted to account for all
                    organic compounds in the stream because the
                    refinery factors are only valid for
                    non-methane organic compounds (percents up to
                    a maximum of 10 percent by weight methane are
                    permitted):

                                         WFTOC
                         FA = FA x 	 ;
                                    WFTOc - WFmethane

    WFTOC       =   Average weight fraction of TOC in the stream;

    WFmethane   =   Average weight fraction of methane in the
                    stream; and

    N           =   Number of pieces of equipment of the
                    applicable equipment type in the stream.


Note that the emission factor "FA" is defined differently for

refineries than for SOCMI, marketing terminals,  or oil and gas

production operations when calculating TOC mass emissions.  It is

necessary to adjust the "FA" term when applied to refineries,

because when the refinery factors were developed,  the methane was

subtracted out of the organic total.   Adjusting the "FA" term for

refineries is a way to correct for this.  Two guidelines when

correcting the "FA" term when applied to refineries are as

follows:

    •    The correction should only be applied to equipment
         containing a mixture of organics 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).


    Thus, at a SOCMI process unit, if there were 100 gas valves

in a stream containing, on average, 90 weight percent TOC and

10 weight percent water vapor,  emissions would be calculated as

follows:

                              2-16

-------
            = -^A x WF>POC x N
            = 0.00597 kg/hr/gas valve x 0.9 x 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 non-methane TOC,
10 weight percent water vapor, and 10 weight percent methane
(thus, the TOC weight percent would be 90), emissions would be
calculated using the above equation as follows:
                          WFTOC
                    WFTOc - WFmethane
            = 0.0268 kg/hr/gas valve x (0.9/0.9-0.1) x 0.9 x
              100 gas valves
            = 2.71 kg/hr of VOC from gas valves in the stream

    If there are several streams at a process unit, the total TOC
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 TOC
emission rate from leaking equipment.
    Assuming all of the organic compounds in the stream are
classified as VOC's, the total VOC emission for each stream is
calculated as the sum of TOC emissions associated with each
specific equipment type in the stream.  Section 2.4.6 discusses
an adjustment that can be made to predict the VOC emission rate
if some of the organic compounds in the stream are not classified
as VOC's (such as methane and ethane) .
    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

                               2-17

-------
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
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 is
included in this section primarily to aid in the analysis of old
datasets which were collected for older regulations that used
10,000 ppmv as the leak definition.   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, refineries,
marketing terminals, and oil and gas production operations for
these two ranges of screening values are presented in tables 2-5,
2-6, and 2-7, and 2-8, respectively.  As with the average
factors, the SOCMI,  marketing terminal, and oil and gas
production operations screening range factors predict total
                               2-18

-------
       TABLE 2-5.  SOCMI SCREENING RANGES EMISSION FACTORS
 Equipment type
Service
 >10,000 ppmv
Emission factor
(kg/hr/source)a
  <10,0 0 0 ppmv
Emission factor
(kg/hr/source)a
Valves

Pump sealsb
Compressor
Gas
Light liquid
Heavy liquid
Light liquid
Heavy liquid
Gas
0
0
0
0
0
1
.0782
.0892
.00023
.243
.216
.608
0
0
0
0
0
0
.000131
.000165
.00023
.00187
.00210
.0894
 seals

 Pressure
 relief valves

 Connectors

 Open-ended
 lines
  Gas


  All

  All
      1.691


      0.113

      0.01195
    0.0447


    0.0000810

    0.00150
aThese factors are for total organic compound emission rates.

bThe light liquid pump seal factors can be applied to estimate
 the leak rate  from agitator seals.
                               2-19

-------
     TABLE 2-6.  REFINERY SCREENING RANGES EMISSION FACTORS3
Equipment type
Valves

Pump seals0
Compressor seals
Pressure relief
valves
Connectors
Open-ended lines
aSource : Reference
>10,000 ppmv
Emission factor
Service (kg/hr/source) b
Gas
Light liquid
Heavy liquid
Light liquid
Heavy liquid
Gas
Gas
All
All
6.
0
0
0
0
0
1
1
0
0

.2626
.0852
.00023
.437
.3885
.608
.691
.0375
.01195

< 1 0 , 0 0 0 ppmv
Emission factor
(kg/hr/source) b
0
0
0
0
0
0
0
0
0

.0006
.0017
.00023
.0120
.0135
.0894
.0447
.00006
.00150

^These factors are for non-methane organic compound emission
 rates.

cThe light liquid pump seal factors can be applied to estimate
 the leak rate from agitator seals.
                               2-20

-------
 TABLE  2-7.  MARKETING  TERMINAL  SCREENING  RANGES  EMISSION  FACTORS
Equipment
type
Valves
Pump seals
Other
(compressors
and others )b
Fittings
(connectors
and flanges)0
>10,000 ppmv
Emission factor
Service (kg/hr/source) a
Gas
Light Liquid
Light
liquid
Gas
Light liquid
Gas
Light liquid
2
7
3
3
6
NA
.3E-
. VE-
NA
.4E-
.4E-
.5E-
02
02
02
02
03
< 1 0 , 0 0 0 ppmv
Emission factor
(kg/hr/source) a
1
1
2
1
2
5
7
.3E-
.5E-
.4E-
.2E-
.4E-
.9E-
.2E-
05
05
04
04
05
06
06
aThese factors are for total organic compound emission rates
 (including non-VOC's such as methane and ethane).   "NA"
 indicates that not enough data were available to develop the
 indicated emission factor.

bThe "other" equipment type should be applied for any equipment
 type other than fittings, pumps, or valves.

c"Fittings" were not identified as flanges or connectors;
 therefore, the fitting emissions were estimated by averaging the
 estimates from the connector and the flange correlation
 equations.
                               2-21

-------
TABLE 2-8.
              OIL AND  GAS  PRODUCTION OPERATIONS  SCREENING RANGES
                        EMISSION FACTORS
Equipment type
Valves

Pump seals
Others0

Connectors

Flanges
Open-ended lines
>10,000 ppmv
Emission factor
Service^ (kg/hr/source) a
Gas
Heavy Oil
Light Oil
Water/Oil
Gas
Heavy Oil
Light Oil
Water/Oil
Gas
Heavy Oil
Light Oil
Water/Oil
Gas
Heavy Oil
Light Oil
Water/Oil
Gas
Heavy Oil
Light Oil
Water/Oil
Gas
Heavy Oil
Light Oil
Water/Oil
9.8E-02
NA
8.7E-02
6.4E-02
7.4E-02
NA
l.OE-01
NA
8.9E-02
NA
8.3E-02
6.9E-02
2.6E-02
NA
2.6E-02
2.8E-02
8.2E-02
NA
7.3E-02
NA
5.5E-02
3.0E-02
4.4E-02
3.0E-02
< 1 0 , 0 0 0 ppmv
Emission factor
(kg/hr/source) a
2 .5E-05
8.4E-06
1.9E-05
9.7E-06
3 .5E-04
NA
5.1E-04
2 .4E-05
1.2E-04
3.2E-05
1.1E-04
5.9E-05
l.OE-05
7.5E-06
9.7E-06
l.OE-05
5.7E-06
3 .9E-07
2 .4E-06
2.9E-06
1.5E-05
7.2E-06
1.4E-05
3.5E-06
aThese factors are for total organic compound emission rates
 (including non-VOC's such as methane and ethane) and apply to
 light crude, heavy crude, gas plant, gas production, and
 offshore facilities.  "NA" indicates that not enough data were
 available to develop the indicated emission factor.
^Water/Oil emission factors apply to water streams in oil service
 with a water content greater than 50%, from the point of origin
 to the point where the water content reaches 99%.  For water
 streams with a water content greater than 99%,  the emission rate
 is considered negligible.
cThe "other" equipment type  was derived from compressors,
 diaphrams, drains, dump arms, hatches, instruments, meters,
 pressure relief valves,  polished rods, relief valves, and vents.
 This "other" equipment type should be applied for any equipment
 type other than connectors, flanges, open-ended lines, pumps, or
 valves.
                               2-22

-------
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.
    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.  Also, for
refineries, the screening range emission factors must be adjusted
for methane in the equipment because when the refinery factors
were developed, the methane was subtracted out of the organic
total.
    To calculate TOC emissions using the Screening Ranges
Approach, the following equation is used:
ETOC =
                                          x  NL)
where:
    ETOC

    FG
 TOC emission  rate  for an equipment  type
  (kg/hr);
 Applicable emission  factor  for  sources with
 screening values greater than or equal to
 10,000  ppmv  (kg/hr/source);
 FOR REFINERIES ONLY: The emission factor "FG"
 must be adjusted to  account  for all organic
 compounds in  the stream because the refinery
 factors are only valid for  non-methane
 organic compounds  (percents  up  to a maximum
 of 10 percent by weight methane are
 permitted):

                      WPTOc
    WPTOC
       FG  =  FG  x 	
                 WPTOC  - wpmethane
 Average  weight percent of TOC  in the  stream;
                               2-23

-------
    wpmethane   =   Average weight percent of methane in the
                    stream;
    NG          =   Equipment count (specific equipment type)  for
                    sources with screening values greater than or
                    equal to 10,000 ppmv;
    FL          =   Applicable emission factor for sources with
                    screening values less than 10,000 ppmv
                    (kg/hr/source)
                    FOR REFINERIES ONLY:  The emission factor
                    "FL" must be adjusted to account for all
                    organic compounds in the stream because the
                    refinery factors are only valid for
                    non-methane organic compounds (percents up to
                    a maximum of 10 percent by weight methane are
                    permitted) :
                                         WPTOC
                         FL = FL x - ;  and
                                          - wpmethane
    N           =   Equipment count (specific equipment type)  for
                    sources with screening values less than
                    10,000 ppmv .
    Assuming all of the organic compounds in the stream are
classified as VOC's, the total VOC emission for each stream is
calculated as the sum of TOC emissions associated with each
specific equipment type in the stream.  Section 2.4.6 discusses
an adjustment that can be made to predict the VOC emission rate
if some of the organic compounds in the stream are not classified
as VOC's (such as methane and ethane)  .
    The screening range emission factors are a better indication
of the actual leak rate from individual equipment than the
average emission factors.  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 the EPA
relating screening values to mass emission rates for SOCMI

                               2-24

-------
process units and for petroleum industry process units are
presented in tables 2-9 and 2-10,  respectively.  Correlations for
the petroleum industry apply to refineries, marketing terminals
and oil and gas production operations.  Figures 2-2 through 2-5
plot the correlations.  Both the SOCMI and petroleum industry
correlations predict total organic compound emission rates.
Appendix B.I contains additional information on the general
development of correlation equations.  Additionally, appendix B.2
contains information about the development of the SOCMI
correlations and appendix C contains information about the
development of the petroleum industry correlations.
         The EPA Correlation Approach is  preferred when actual
screening values are available.  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.  This approach involves entering the
non-zero, non-pegged screening value into the correlation
equation, which predicts the TOC mass emission rate based on the
screening value.  Default zero emission rates are used for
screening values of zero ppmv and pegged emission rates are used
for "pegged" screening values  (the screening value is beyond the
upper limit measured by the portable screening device).
    Correlations for SOCMI are available for (1) gas valves;
(2) light liquid valves; (3) connectors;  and (4) light liquid
pump seals.  Correlations for the petroleum industry are
available for (1) valves; (2) connectors;  (3) pumps;  (4) flanges;
(5) open-ended lines; and (6) "others" (derived from instruments,
loading arms, pressure relief valves, stuffing boxes, and vents).
    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
                               2-25

-------
     TABLE  2-9.   SOCMI  LEAK RATE/SCREENING VALUE CORRELATIONS


 Equipment  type                      Correlation3/b

 Gas  valves          Leak rate (kg/hr)  = 1.87E-06 x (SV)°-873

 Light  liquid  valves  Leak rate (kg/hr)  = 6.41E-06 x (SV)°-797


 Light  liquid  pumpsc  Leak rate (kg/hr)  = 1.90E-05 x (SV)°-824


 Connectors	Leak rate (kg/hr)  = 3.05E-06 x (SV)°-885

aSV = Screening value in ppmv.

bThese correlations predict total organic compound emission
 rates.

cThe correlation for light liquid pumps can be applied to
 compressor seals, pressure relief valves, agitator seals, and
 heavy liquid pumps.
                               2-26

-------
    TABLE 2-10.  PETROLEUM INDUSTRY LEAK RATE/SCREENING VALUE
                          CORRELATIONS3


 Equipment
 type/service                       Correlation13/c

 Valves/all            Leak rate (kg/hr) = 2.29E-06 x (SV)°-746

 Pump  seals/all       Leak rate (kg/hr) = 5.03E-05 x (SV)°-610

 Othersd               Leak rate (kg/hr) = 1.36E-05 x (SV)°-589

 Connectors/all       Leak rate (kg/hr) = 1.53E-06 x (SV)°-735

 Flanges/all           Leak rate (kg/hr) = 4.61E-06 x (SV)°-703

 Open-ended lines/all  Leak rate (kg/hr) = 2.20E-06 x (SV)°-704

aThe correlations presented in this table are revised petroleum
 industry correlations.

ksv = Screening value in ppmv.

GThese correlations predict total organic compound emission
 rates (including non-VOC's such as methane and ethane).

dThe "other"  equipment type was derived from instruments,
 loading arms, pressure relief valves, stuffing boxes, and
 vents.  This  "other" equipment type should be applied to any
 equipment type other than connectors, flanges, open-ended
 lines, pumps, or valves.
                               2-27

-------
 A/
 id
 OJ
 u
 0
                      S-Q
                                                BDD
1DDD
    D  Gas Valves   +
Figure 2-2. SOCMI  Correlations relating total organic compound
             (TOO  leak rate to screening value:
            0  -  1,000  ppmv
                                2-28

-------
u
o
                                                       —H
                                               BDD
1DDD
Figure 2-3. SOCMI Correlations relating total organic  compound
            (TOO leak rate to screening value:
            1,000 - 1,000,000 ppmv
                               2-29

-------
     CD

     ccs
     cr  D , DD1
     O)
     O
                                       BDD
Figure 2-4. Petroleum Industry  Correlations relating total
            organic compound  (TOO  leak rate to screening value:
            1,000 - 1,000,000 ppmv
                               2-30

-------
                                               F  anges

                                               Other
Figure 2-5.  Petroleum Industry Correlations relating total
            organic compound (TOO leak rate to screening value:
            1,000 - 1,000,000 ppmv
                               2-31

-------
estimate emissions for compressor seals and pressure relief
valves in SOCMI process units.  Because bagging data were limited
and the frequency of occurrence of some equipment types was
small, a correlation for an "other" equipment type was developed
for the petroleum industry correlations to apply to any equipment
type other than connectors, flanges,  open-ended lines, pumps,  or
valves.
    Bagging data for agitator seals at petroleum industry 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.
    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 and petroleum industry
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.
    Table 2-11 lists the SOCMI default-zero leak rates and
table 2-12 presents the petroleum industry default-zero leak
rates for each of the equipment types with correlation equations.
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 portable monitoring device used to collect the
default-zero data was sufficiently sensitive to indicate a
                               2-32

-------
      TABLE  2-11.   DEFAULT-ZERO  VALUES:   SOCMI  PROCESS  UNITS


                                    Default-zero  emission  rate
 Equipment type                          (kg/hr/source)a

 Gas valve                                   6.6E-07

 Light liquid valve                          4.9E-07

 Light liquid pumpb                          7.5E-06

 Connectors	6 .1E-07	

aThe default zero emission rates are for total organic compounds
 (including non-VOC's such as methane and ethane).

bThe light liquid pump default zero value can be applied to
 compressors, pressure relief valves, agitators,  and heavy
 liquid pumps.
                               2-33

-------
      TABLE 2-12.  DEFAULT-ZERO VALUES:  PETROLEUM INDUSTRY
 Equipment type/service
Default-zero emission ratesa'b
        (kg/hr/source)
 Valves/all

 Pump seals/all

 Othersc/all

 Connectors/all

 Flanges/all

 Open-ended lines/all
            7.8E-06

            2.4E-05

            4.0E-06

            7.5E-06

            3.1E-07

            2.0E-06
aDefault zero emission rates were based on the combined
 1993 refinery and marketing terminal data only  (default zero
 data were not collected from oil and gas production
 facilities).

bThese default zero emission rates are for total organic
 compounds (including non-VOC's such as methane and ethane).

cThe "other"  equipment type was developed from instruments,
 loading arms, pressure relief valves, stuffing boxes, vents,
 compressors,  and dump lever arms.  This "other" equipment type
 should be applied to any equipment type other than connectors,
 flanges, open-ended lines, pumps, or valves.
                               2-34

-------
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 tables 2-11 and 2-12 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.
    The "pegged" emission rate is the mass emission rate
associated with a screening value that has "pegged" the meter on
the portable screening device (i.e. the screening value is beyond
the upper limit measured by the portable screening device).   In
the case of a screening value pegged at 10,000 ppmv, a dilution
probe should be used to extend the upper limit of the portable
screening device to 100,000 ppmv.  Thus, screening values can be
reported up to 100,000 ppmv before pegging the instrument and the
correlation equation can be used to estimate the mass emissions.
However, in the case of previously-collected data or in the
absence of a dilution probe, pegged readings of 10,000 ppmv are
sometimes reported.  In such cases, the 10,000 ppmv pegged
emission rates can be used to estimate the mass emissions.
    Table 2-13 presents the 10,000 ppmv and 100,000 ppmv pegged
emission rates for SOCMI process units and table 2-14 presents
the 10,000 ppmv and 100,000 ppmv pegged emission rates for
petroleum industry process units.  These pegged emission rates
are to be used to estimate emissions when instrument readings are
pegged and a dilution probe is not used.
    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).
    To summarize the correlation approach, each equipment piece
with a screening value of zero is assigned the default-zero leak
                              2-35

-------
 TABLE  2-13.   10,000  PPMV AND  100,000  PPMV SCREENING  VALUE  PEGGED
              EMISSION  RATES FOR  SOCMI PROCESS  UNITS


                      10,000 ppmv pegged    100,000 ppmv pegged
                        emission rate         emission rate
 Equipment type        (kg/hr/source)a/b        (kg/hr/source)a
Gas valves
Light liquid
valves
Light liquid pump
sealsb
Connectors
0.024
0.036
0.14
0.044
0.11
0.15
0.62
0.22
aThe SOCMI pegged emission rates are for total organic compounds.

bThe 10,000 ppmv pegged emission rate applies only when a
 dilution probe cannot be used or in the case of
 previously-collected data that contained screening values
 reported pegged at 10,000 ppmv.

cThe light liquid pump seal pegged emission rates can be applied
 to compressors, pressure relief valves, and agitators.
                               2-36

-------
 TABLE  2-14.   10,000 ppmv  and  100,000  PPMV  SCREENING  VALUE  PEGGED
            EMISSION RATES FOR THE PETROLEUM INDUSTRY


                        10,000 ppmv pegged   100,000 ppmv pegged
 Equipment                emission rate        emission rate
 type/service           (kg/hr/source)a/b     (kg/hr/source)a
Valves/all
Pump seals/all
Othersd/all
Connectors/all
Flanges/all
Open-ended lines/all
0.064
0.074
0.073
0.028
0.085
0.030
0.140
0.160C
0.110
0.030
0.084
0.079
aThe petroleum industry pegged emission rates are for total
 organic compounds (including non-VOC's such as methane and
 ethane).

bThe 10,000 ppmv pegged emission rate applies only when a
 dilution probe cannot be used or in the case of
 previously-collected data that contained screening values
 reported pegged at 10,000 ppmv.  The 10,000 ppmv pegged emission
 rate was based on components screened at greater than or equal
 to 10,000 ppmv; however, in some cases, most of the data could
 have come from components screened at greater than 100,000 ppmv,
 thereby resulting in similar pegged emission rates for both the
 10,000 and 100,000 pegged levels (e.g., connector and flanges).


C0nly 2 data points were available for the pump seal
 100,000 pegged emission rate; therefore the ratio of the pump
 seal 10,000 pegged emission rate to the overall 10,000 ppmv
 pegged emission rate was multiplied by the overall 10,000 ppmv
 pegged emission rate to approximate the pump 100,000 ppmv pegged
 emission rate.

dThe "other" equipment type was developed from instruments,
 loading arms, pressure relief valves, stuffing boxes, vents,
 compressors,  dump lever arms, diaphrams, drains, hatches,
 meters, and polished rods.  This "other" equipment type should
 be applied to any equipment type other than connectors, flanges,
 open-ended lines, pumps, and valves.
                               2-37

-------
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.
It 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.
Finally,  each equipment piece with a screening value reported as
pegged is assigned the appropriate pegged emission rate.
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 procedures 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
procedures 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
                               2-38

-------
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
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
                               2-39

-------
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 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 the 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
                               2-40

-------
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
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
"pegged" emission rates shown in table 2-13 for SOCMI process
units, and in table 2-14 for petroleum industry process units.
                               2-41

-------
    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 pegged emission rate presented in table 2-13 for
SOCMI process units,  or table 2-14 for petroleum industry process
units.  An advantage of using the greater than or equal to
10,000 ppmv pegged emission rates 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 for
this alternative.
    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:

       Logio (leak rate  [in kg/hr])  = Po + Pi x Log^g (SV)
where:
    PC/ Pi  = Regression constants;  and
    SV      = Screening value.
                               2-42

-------
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
back to the original units.  The transformed equation is the
unit-specific correlation, and is expressed as:
                   Leak rate = SBCF x 10^° x
where:
   Leak rate    =   Emission rate of TOC's from the individual
                    equipment piece (kg/hr);
    SBCF        =   Scale bias correction factor;
    Po/Pi       =   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,

                               2-43

-------
1 kg/hr, and 10 kg/hr.  In log space, these emission rates
correspond to log^g (0.1)  = -1, log^g (1) = 0, and
1°910 (1Q)  = !' respectively.  Thus, the geometric mean of these
three points is (-1 + 0 + 1)/3 = 0.   Directly transforming this
geometric mean to normal space predicts an emission rate for a
screening value of 500,000 ppmv of 101-1 = 1 kg/hr, whereas the
arithmetic mean of the emission rates is
(0.1 + 1 + 10)/3 = 3.7 kg/hr.  From this 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.
                               2-44

-------
Each of these topics above are addressed in the following
sections .
2.4.1  Speciating Emissions
    For each of the four approaches for estimating equipment leak
emissions, the equations presented are used to estimate TOC
emissions for estimating equipment leak 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 equation is used to speciate emissions from a single
equipment piece:

                     Ex = ETOC x  (WPx/WPTOc)
where :
    Ex      =   The mass emissions of organic chemical  "x" from
                the equipment  (kg/hr) ;
     ETOC   =   Tne TOC mass emissions from the equipment
                 (kg/hr) calculated from either the Average
                Emission Factor,  Screening  Ranges, Correlation,
                or Unit-Specific  Correlation approaches;
    WPX     =   The concentration of organic chemical  "x" in the
                equipment  in weight percent; and
    WPTOC   =   Tne TOC 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.

                               2-45

-------
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
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 TOC concentration do not respond to
different TOC's equally.  (This is discussed in more detail in
chapter 3.0).   To demonstrate this point, consider a monitoring
                               2-46

-------
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
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.
                               2-47

-------
    A detailed listing of published RF's is contained in
appendix D.  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 D-2 in appendix D, 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
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 D are for pure compounds.  Those RF's
can be used to estimate the RF for a mixture using the equation:
                              1 = 1

where:
    RFm  =   Response factor of the mixture;
    n    =   Number of components in the mixture;
    xj_   =   Mole fraction of constituent i in the mixture; and
    RFj_  =   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.
                               2-48

-------
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
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 D;  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 the 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

                              2-49

-------
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
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
                               2-50

-------
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
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 E 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 E 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

                               2-51

-------
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
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 TOC emissions have been estimated by using either the
Average Emission Factor, the Screening Ranges, the Correlation,
or the Unit-Specific Correlation approaches, the VOC emissions
from a group of equipment containing similar composition can be
calculated using the equation:

                   EVOC = ETOC x (WPVOC/WPTOC)
where:
    EVOC    =   Tne VOC mass emissions from the equipment
                 (kg/hr);
    ETOC    =   Tne TOC mass emissions from the equipment
                 (kg/hr) calculated form  either the Average
                               2-52

-------
                Emission Factor, Screening Ranges, Correlation,
                or Unit-Specific Correlation approaches;
    WPVOC   =   Tne concentration of VOC in the equipment in
                weight percent;
    WPTOC   =   Tne TOC concentration in the equipment in weight
                percent .

2.4.7  Estimating Equipment Leak Emissions of Inorganic
     Compounds
    The majority of data collected for estimating equipment leak
emissions has been for TOC's or 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 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.
                               2-53

-------
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.  1993 Study of Refinery Fugitive Emissions from Equipment
    Leaks, Volumes I,  II, and III, Radian DCN No. 93-209-081-02,
    Radian Corporation, Prepared for Western States Petroleum
    Association,  Glendale, CA, and American Petroleum Institute,
    Washington,  B.C.,  1994.

5.  Development of Fugitive Emission Factors and Emission
    Profiles for Petroleum Marketing Terminals, Volumes I and II,
    API 4588, Radian Corporation, Prepared for American Petroleum
    Institute,  1993.

6.  Fugitive Hydrocarbon Emissions from Oil and Gas Production
    Operations,  API 4589, Star Environmental, Prepared for
    American Petroleum Institute, 1993.

7.  Emission Factors for Oil and Gas Production Operations,
    API 4615, Star Environmental, Prepared for American Petroleum
    Institute,  1995.

8.  Letter from Robert Strieter, API, to David Markwordt,
    U.S. EPA, regarding the submittal of gasoline distribution
    facility fugitive equipment screening data.  April 26,  1994.

9.  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.

10. Langley, G.J.  and R.G. Wetherhold.   (Radian Corporation.)
    Evaluation of Maintenance for Fugitive VOC Emissions Control.


                               2-54

-------
    Prepared for U.S.  Environmental Protection Agency.  Research
    Triangle Park,  NC.   Publication No. EPA-600/S2-81-080.
    May 1981.

11.  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.
                               2-55

-------
                      3.0  SOURCE SCREENING

3.1  INTRODUCTION
    This chapter presents procedures 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 the 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 the EPA Reference 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.
                               3-1

-------
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 the EPA
Reference Method 21, section 3.0.1.  Reference Method 21 is
included in this document as appendix F.
    In general, portable VOC monitoring instruments are equipped
with a probe that is placed at the leak interface of a piece of
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 (CC>2) do not produce

                               3-2

-------
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 non-methane organics or individual
organic components.  Certain organic compounds containing
nitrogen, oxygen, or halogen atoms give a reduced response when
sampled 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 CC>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
                               3-3

-------
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.
3.2.2  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 the EPA Reference Method 21.
(See appendix F.)  Reference Method 21 specifies the requirements
that must be met when a facility is collecting screening data to
comply with a regulation.  The requirements of the EPA Reference
Method 21 are also applicable when screening data are collected
for the sole purpose of estimating emissions.  When the
requirements of Reference 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.
    Reference Method 21 requires that the analyzer meet the
following specifications:1
    •    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;
                               3-4

-------
    •    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.
Note that the suction flow rate span allowed by Reference
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 D.)   The RF tests are

                               3-5

-------
   TABLE  3-1.   PERFORMANCE  CRITERIA  FOR  PORTABLE  VOC  DETECTORS3
      Criteria           Requirement           Time interval

 Instrument          Must be  <10  unless   One time, before
 response factor     correction curve  is  detector is put in
                     used                service.

 Instrument          Must be  <30  seconds  One time, before
 response time                           detector is put in
                                         service. If
                                         modification to sample
                                         pumping or flow
                                         configuration is made,
                                         a new test is required.

 Calibration         Must be  <10  percent  Before detector is put
 precision           of  calibration gas   in service and at
                     value               3-month intervals or
                                         next use, whichever is
	later.	

aSource:   Reference 1.
                               3-6

-------
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 Reference Method 21,  the RF
can either 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 laboratory
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 D 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 Reference 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
                               3-7

-------
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.
    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.
                               3-8

-------
    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
calibration gas is injected.  The response time 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
                               3-9

-------
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
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 1 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 EPA Reference
       Method 21 Requirements
    In some cases, a monitoring device may not be available that
meets all of the performance specifications of the EPA Reference

                               3-10

-------
                       TABLE 3-2.   PORTABLE VOC DETECTION  INSTRUMENTS
Manufacturer Model no .
Bacharach Instrument Co. , L
Santa Clara, California
TLV Sniffer
Foxboro, OVA- 12 8
S . Norwalk, Connecticut
OVA- 10 8
Miran IBX

Health Consultants Detecto -
PAK III
HNU Systems, Inc. HW-101
Newton Upper Fal 1 s ,
Massachusetts

U>
1 Mine Safety Appliances Co., 40
|— i Pittsburgh, Pennsylvania
I—1
Survey and Analysis, Inc., On Mark Model 5
Northboro, Massachusetts
Pollutant (s)
detected
Combustible gases

Combustible gases
Most organic compounds

Most organic compounds
Compounds that absorb
infrared radiation
Most organic compounds

Chlorinated hydrocarbons,
aromatics, aldehydes,
ketones, any substance
that UV light ionizes

Combustible gases

Combustible gases

Principle of
operation
Catalytic combustion

Catalytic combustion
FID/GC

FID/GC
NDIR

FID/GC

Photoionization




Catalytic combustion

Thermal conductivity

Range
(ppm)
0-100% LELa




0-1, 000 and 0-10, 000
0-1, 000

0-10, 000
Compound specific

0-10, 000








0-20, 0-200, 0-2, 000




0-10% and
0-100% LELa
0-5 and 0-100% LEL







a

aLEL = Lower explosive limit.

-------
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 Reference
Method 21 instrument because it cannot meet the RF requirement.
If an instrument fails to meet Reference Method 21 requirements,
it can still be used for the purpose of estimating emissions if
its reliability can be documented.
    Two primary steps must be taken to document the reliability
of an analyzer that fails to meet the Reference 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 Reference 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
                               3-12

-------
TABLE 3-3.  EQUIPMENT LEAK EMISSION SOURCES


              Equipment types

              Agitator  seals
              Compressor seals
                 Connectors
                 Diaphrams
                  Drains
              Dump lever arms
                  Flanges
                  Hatches
                Instruments
                Loading  arms
                  Meters
              Open-ended lines
               Polished rods
          Pressure relief devices
                 Pump  seals
              Stuffing  boxes
                  Valves
                   Vents

                  Service
                 Gas/vapor
                Light  liquid
                Heavy  liquid
                    3-13

-------
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
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.
                               3-14

-------
3.3.2  Procedure for Screening
    Once the equipment to be screened has been identified, the
procedures outlined in the EPA Reference 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
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
                               3-15

-------
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
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 and flanges, pumps and compressors,
open-ended lines, and other potential sources of VOC leakage,
such as pressure relief devices, loading arms, stuffing boxes,
instruments,  vents,  dump lever arms, drains, diaphrams, hatches,
notes,  or polished rods. 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 and Flanges.  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
                               3-16

-------
               Rising Stem Type
                                                  Stem
                                         Packing Nut
                                        Packing Gland





                                          Packing
                                              Disk or Wedge
                                                                                Packing Out
                                                             Nonrising Stem Type
                                                                                    Disk or Wedge
Figure  3-1.  Gate  Valves
                                            3-17

-------
 through
       Stem
      Packing
      Gland
      Packing
 Screen
'  Here

Packing Nut

Screen
Here
                                       Body
               Manual Globe Valve
Figure 3-2.  Globe  Valves
                                                          Globe Type Control Valve
                                         3-18

-------
                                              Screen Here

                       i
                                                        1
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
                  Weir-Type Diaphragm Valve
                               Screen Here
                                 Screen
                                 Here
                                               Screen Here
                                                  Check Valves


Figure 3-5. Weir-Type Diaphragm Valve  and Check  Valves

                                     3-21

-------
threaded connections, are sampled with a similar traverse.
    3.3.2.3  Pumps, 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.
    3.3.2.4  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  Open-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 I 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.
                               3-22

-------
                                                      Moving Shaft





                                                         Screen Here
                                                     Stationary Casing
                    Vertical Centrifugal Pump
                                  Horizontal Centrifugal Pump
Figure 3-6.  Centrifugal  Pumps




                                         3-23

-------
                           Screen Here
            Alternate Screening
             Area if Horn
             Inaccessible
                                                             Tension Adjustment
                                                                Thimble
                                                                Spring
                                                                   Disk
                                                To Process
Figure  3-7.   Spring-Loaded  Relief Valve

                                          3-24

-------
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.)
    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 (ppmv) .
    10.  Background concentration (ppmv).
    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

-------
                          TABLE  3-4.    EXAMPLE  FIELD  SHEETS  FOR EQUIPMENT SCREENING DATA
    Detector model no.

    Operator name 	
                                                                                            Screening
       Component          Component         Location/                          Operating           value          Background
          ID               Type             Stream           Service            hrs/yr            (ppmv)            (ppmv)
U>
 I
DO

-------
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 EMISSION 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.
                               4-1

-------
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
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:
                               4-2

-------
     •    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
     •    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
                               4-3

-------
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
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 the 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.
                               4-4

-------
                         This line should
                           be as short
                           as possible
                           Cold Trap in Ice Bath
                          (Optional)
Trap
                                             Dry Gas
                                              Meter
                                                                   Hg Manometer
                                                           u
                                                                  Control
                                                                  Valve
                                                        Small
                                                      Diaphragm
                                                        Pump
                                                                         Filter
                                                                         Sample Bag
                                                                                 Vacuum
                                                                                  Pump
                   Two-V\fey Valve
Figure  4-1.  Sampling  train  for bagging a source  using  the vacuum
                method.
                                          4-5

-------
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
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
(f/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
                               4-6

-------
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
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).

                               4-7

-------
     (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).
     (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 f/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 pound 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
                               4-8

-------
 TABLE 4 -1 .
CALCULATION PROCEDURES FOR LEAK RATE WHEN USING THE
             VACUUM METHOD
     Leak Rate
      (kg/hr)

where:

9.63 x ICr10
     Q

     MWa
     T

     P

     VL

     16.67
  =  9.63  x 10-10  (Q)(MW)(GC)(P)   +  (p)(VL)
            T +  273.15               16.67(t
       A conversion factor using the gas constant:

             °K x 10^ x kg-mol x min
                                        /
                 0  x hour x mmHg

       Flow rate out of bag (f/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/kg-mol);

       Sample bag organic compound concentration
       (ppmv) minus background bag organic compound
       concentration0 (ppmv);

       Absolute pressure at the dry gas meter
       (mmHg);

       Temperature at the dry gas meter (°C);

       Density of organic liquid collected  (g/mf);

       Volume of liquid collected (mf ) ;

       A conversion factor to adjust term to units
       of kilograms per hour  (g x hr)/(kg x min)

       Time in which liquid is collected (min); and
aFor mixtures calculate MW as :
                         n
                                        n
                                     /
     where :
          MWj_ = Molecular weight of organic compound i;
           Xj_ = 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 .
cCollection of a background bag is optional.  If a background bag
 is not collected, assume the background concentration is zero.
                               4-9

-------
   Carrier
   Gas
   Source
           Activated
           Ch arcoal
          (if needed)
Tape or Adjustable Metal Band
                                                                           Sample Port for Collecting Data on
                                                                         -Temperature
                                                                         - Hydrocarbon Concentration
                                                                         - Oxygen Concentration
  Tape or Com pressed Foil
Figure  4-2.  Equipment  Required  for  the  Blow-Through
                 Sampling Technique
                                           4-10

-------
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
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.
                               4-11

-------
     (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.

     (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.  An adjustment is provided

for the leak rate equation in table 4-2 to account for the total

flow through the bag.  This adjustment is recommended and

represents an improvement over previous versions


                               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)    (  106ppmv
           -
                                       +
           =
   (kg/hr)  -          T + 273.15           16.67 (t)
                                                      i0ppmv-GC
where :
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 x 106  x  kg-mol
     Q         =  flow rate  out  of  tent  (m3/hr);

             N2 Flow Rate (C/min)                [0.06  (m3/min)]
      1 -  [Tent Oxygen Cone,  (volume %)/21]          (0/hr)

     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);

     GCb       =  Sample  bag organic compound concentration
                  (ppmv),  corrected for background bag organic
                  compound concentration (ppmv);c

     T         =  Temperature in tent  (°C);

     p         =  Density of organic liquid collected (g/mf);

     VL        =  Volume  of liquid  collected (mf);

     16.67     =  A  conversion factor  to adjust  term to units of
                  Kilograms per hour  (g x hr)/(kg x min);  and

     t         =  Time in which liquid is collected (min).
aFor mixtures calculate MW as:
                              n              n
                              Z   MWi Xi  /  Z   Xi
                              1=1            1=1
     where:
          MWj_ = Molecular weight of organic compound i;
                               4-13

-------
         TABLE  4-2.   CALCULATION PROCEDURES  FOR LEAK RATE
                     WHEN USING THE  BLOW-THROUGH METHOD
                           (Continued)
           Xj_ = 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.

cCollection 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   _ OR _(TENT     '
                      (ppmv) ~ SB  I— x BG
     where:
          SB   = Sample bag concentration  (ppmv);
          TENT = Tent oxygen concentration  (volume %); and
          BG   = Background bag concentration  (ppmv)
                               4-14

-------
of this document for quantifying mass emissions from 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 the EPA method 18 is one way to
check 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

                               4-15

-------
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
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
                               4-16

-------
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
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
                               4-17

-------
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.
4.3.5  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
                              4-18

-------
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
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
                               4-19

-------
          TABLE 4-3.  EXAMPLE GC CALIBRATION DATA  SHEET
Plant ID
Instrument ID
Analyst Name
                      Certified    Instrument
                     Gas Cone.       Reading
   Date      Time      (ppmv)        (ppmv)          Comments
                               4-20

-------
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
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

                               4-21

-------
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

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 in the
          event that this is not possible, 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 the 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.

                               4-22

-------
 TABLE 4-4.
     EXAMPLE DATA COLLECTION FORM FOR FUGITIVE EMISSIONS
        BAGGING TEST (BLOW-THROUGH METHOD)
Equipment Type 	
Equipment Category
Line Size
Stream Phase  (G/V, LL, HL)
Barometric Pressure 	
Ambient Temperature 	
Stream Temperature 	
Stream Composition  (Wt %]
                                Component ID
                                Plant ID 	
                                Date
                                Analysis Team
                                Instrument ID 	
                                Stream Pressure
Time
                   Bagging Test Measurement Data

Initial Screening  (ppmv) Equipment Piece 	 Bkgd.
Background Bag Organic Compound Cone.  (ppmv)a 	
Dilution Gas Flow Rate  (f/min) 	
Sample Bag 1 Organic Compound Cone.  (ppmv) 	
02 Concentration (volume %) 	
Bag Temperature  (°C) 	
        Dilution Gas Flow Rate  (f/min)
        Sample Bag 2 Organic Compound Cone.  (ppmv)
        02 Concentration  (volume %) 	
        Bag Temperature  (°C) 	
Condensate Accumulation:  Starting Time
Organic Condensate Collected  (mf) 	
Density of Organic Condensate  (g/mf) 	
                                        Final Time
        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.
                               4-23

-------
 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)k 	
	 Dry Gas Meter Reading  (f/min) 	
	 Sample Bag 1 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) 	

	 Dry Gas Meter Reading  (f/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  (mf) 	
Density of Organic Condensate  (g/mf) 	
       Final Screening  (ppmv) Equip. Piecea 	 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.

                               4-24

-------
     •    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.
     •    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.
                               4-25

-------
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.
     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

                               4-26

-------
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
at 1,000 ppmv 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.
                               4-27

-------
            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.
                               4-28

-------
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.
                               4-29

-------
          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

                               5-1

-------
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.
5.2.1  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
                               5-2

-------
          TABLE  5-1.   SUMMARY OF  EQUIPMENT MODIFICATIONS
 Equipment type
          Modification
                                                    Approximate
                                                      control
                                                     efficiency
 Pumps
 Compressors
 Pressure relief
 devices


 Valves

 Connectors

 Open-ended
 lines

 Sampling
 connections
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

Closed-vent system

Rupture disk assembly

Sealless design

Weld together

Blind, cap, plug,  or second
valve

Closed-loop sampling
100a

 90^

100




 90^

100
 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.

GControl 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.
                               5-3

-------
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.
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
                               5-4

-------
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
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.
                               5-5

-------
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
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
                               5-6

-------
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
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:  (1)  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
                               5-7

-------
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,
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.  Figures 5-9 through 5-15 are graphs
presenting mass emission rate versus leak frequency for gas
fittings, light liquid fittings, gas others, light liquid others,
light liquid pumps,  gas valves, and light liquid valves, for

                               5-8

-------
         TABLE 5-2.  CONTROL EFFECTIVENESS FOR AN LDAR PROGRAM AT A SOCMI PROCESS UNIT
                                                Control effectiveness (%)
      Equipment  type  and
            service
Monthly monitoring
 10,000 ppmv leak
    definition
Quarterly monitoring
  10,000 ppmv leak
     definition
HON reg nega
Valves - gas
Valves - light liquid
Pumps - light liquid
Connectors - all
87
84
69
b
67
61
45
b
92
88
75
93
Ul
   a Control effectiveness attributable to the requirements of the proposed hazardous
     organic NESHAP equipment leak negotiated regulation are estimated based on equipment-
     specific leak definitions and performance levels.
     Data are not available to estimate control effectiveness.

-------
        TABLE  5-3.   CONTROL  EFFECTIVENESS  FOR  AN LDAR  PROGRAM AT A REFINERY PROCESS  UNIT
(Jl
                                                Control effectiveness  (%)
       Equipment  type  and
            service
                           Monthly monitoring
                            10,000 ppmv leak
                               definition
Quarterly monitoring
  10,000 ppmv leak
     definition
HON reg nega
Valves - gas
Valves - light liquid
Pumps - light liquid
Connectors - all
88
76
68
b
70
61
45
b
96
95
88
81
a Control effectiveness attributable to the requirements of the proposed hazardous
  organic NESHAP equipment leak negotiated regulation are estimated based on equipment-
  specific leak definitions and performance levels.
0  b Data are not available to estimate control effectiveness.

-------
  L
  r
                      D.D5
                                   0,25
Figure 5-1.
SOCMI Gas Valve Average Leak Rate Versus  Fraction
Leaking at Several Leak Definitions.
                               5-11

-------
    (0
    [I
                                  kraction keak 1ng

           D  5DD ppmv kk  Def,    +  1.DDD ppmv kk  Def,   0  2, ODD ppmv kk  Def

           A  5,ODD ppmv kk  Def,   X  ID.DDD  ppmv kk  Def   V  SOCMI Avg, kactor
Figure  5-2.
SOCMI  Light  Liquid Valve Average Mass  Emission
Rate Versus  Fraction  Leaking  at Several Leak
Definitions
                                  5-12

-------
Figure 5-3.
SOCMI Light Liquid Pump Average Leak Rate Versus
Fraction Leaking at Several Leak Definitions
                               5-13

-------
  0)
  -l-l
  [fl
  cc
  M
  id
         0,0'
Figure 5-4.    SOCMI Connector Average Leak Rate Versus Fraction
               Leaking  at  Several  Leak Definitions.
                               5-14

-------
                           Refinery Gas Valve Equations
Figure 5-5.
             V  Refinery Avg Factor
Refinery  Gas Valve Average Leak Rate  Versus
Fraction  Leaking at Several Leak Definitions
                                 5-15

-------
          D.D1
                                                           0,5
Figure 5-6.
             V  Refinery Avg Factor
Refinery Light  Liquid Valve Average  Leak Rate
Versus Fraction Leaking at Several Leak
Definitions
                                5-16

-------
          0,35
Figure 5-7.
             V  Refinery Avg Factor
Refinery Light  Liquid Pump Average Leak  Rate
Versus Fraction Leaking at Several Leak
Definitions
                                5-17

-------
                           Refinery Connector Equations

0,002
D.OD1B
n
0)
u 0,0016
L
3
0
o 0,0014
\^
L
v 0,0012
0)
A/
u 0,001
0)
, i
-H
a; D.ODOB
gj 0,0006
j
N 0,0004
id
i_
0) 0,0002
\
<
n

-
— /^
/
/
— /
/
/
/
- /
/
— /
/
/
- 0
/
/ ^-
	 / ^-^
/ ~~^
/ ^^^^^ ^~~~^
/ ^~~^^ -~-~~~^^~^
/ /-^^^^^
/ -~^~^~^ .^~~~^~~^
vi / -~~~~^ -~~~^~^ r7 vi
/ ,-~-^~ ,-^~~' "'
f^^^
                                                           D.D5
Figure 5-8.     Refinery Connector Average Leak Rate Versus
                Fraction Leaking  at Several Leak  Definitions
                                5-18

-------
                  DO   D,D5   D,10   0,15
Figure 5-9.
Marketing Terminal Gas Fittings Average Leak Rate
Versus Fraction Leaking at Several Leak
Definitions
                               5-19

-------
          0
          CO
          0)
   D , DD13


   D , DD12


   D , DD 1 1


   D , DD1D


m  D , DDD9


E  D , DDDB


?  D , DDD7
cr  D , DDD5


   D,DDD4


   D , DDD3


   D , DDD2


>  D,DDD1


   D , DDDD

        D , DD
                             D , 1D
Figure 5-10.
      Marketing Terminal  Light Liquid Fittings Average
      Leak Rate Versus  Fraction Leaking at Several  Leak
      Definitions
                                5-20

-------
          cd D , DDD4
          CL
                                               25  D,3D
Figure 5-11.
Marketing Terminal Gas Others Average Leak Rate
Versus Fraction Leaking at Several Leak
Definitions
                               5-21

-------
Figure 5-12.
Marketing Terminal Light Liquid Others Average
Leak Rate Versus Fraction Leaking at Several Leak
Definitions
                              5-22

-------
                  DO   D,D5   D,10   0,15
Figure 5-13.
Marketing Terminal Light Liquid Pumps Average Leak
Rate Versus Fraction Leaking at Several Leak
Definitions
                               5-23

-------
          o
          en
          cd
          CO
          0)
          0>
          CQ
                                                25  D,3D
Figure 5-14.
Marketing Terminal Gas Valves  Average Leak Rate
Versus Fraction Leaking at  Several  Leak
Definitions
                               5-24

-------
Figure 5-15.
Marketing Terminal Liquid Light Valves Average
Leak Rate Versus Fraction Leaking at Several Leak
Definitions
                              5-25

-------
marketing terminal process units.  Figures 5-16 through 5-34
present mass emission rate as a function of leak frequency for
connectors,  flanges, open-ended lines, others, pumps, and valves
at oil and gas production operations.  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, 5-5, 5-6, and
5-7 present equations for the lines in each of the SOCMI,
refinery, marketing terminal, and oil and gas production
operations,  and graphs,  respectively.  Appendix G describes the
approach that was used to develop the equations.
     Figure 5-35 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 5-35, 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 5-35.  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 cycle);
     •    Point Z - leak frequency immediately preceding a
          monitoring cycle; and
     •    Point F - average leak frequency between monitoring
          cycles  (final leak frequency).
                               5-26

-------
             D , D D B
             D , D D 6
          [0
          a:
          CD
          0>
          to
             D , D D4
             D , D D 3
             D , D D 2
             D , D D 1
             D , D D D
                 D , D D   0 , D 5  D.1D  0,15   0,20  0,25  0,30
                           Fraction  Leaking


                            Leak  Definition
Figure 5-16.   Oil and Gas Production Gas Connectors  Average Leak
               Rate Versus Fraction Leaking  at  Several  Leak
               Definitions
                               5-27

-------
             D , D D B
             D , D D 6
          [0
          a:
          CD
          0>
          to
             D , D D4
             D , D D 3
             D , D D 2
             D , D D 1
             D , D D D
                 D , D D   0 , D 5  D.1D  0,15   0,20  0,25  0,30
                           Fraction  Leaking


                            Leak  Definition
Figure 5-17.   Oil and Gas Production Light Oil  Connectors
               Average Leak Rate Versus Fraction Leaking  at
               Several Leak Definitions
                               5-28

-------
          CD
          U
          [0
          a:
          CD
          0>
          [0
             D , D D 9
             D , D D B
             D , D D 7
             D , D D B
             B , D B 5
               D B4
               D B 3
               D B ;
             B , D B 1
             B , D B B
                 0,00   0,05  B , 1 B  D.15   0,20  0,25  0,30


                           Fraction  Leaking


                            Leak  Definition
Figure 5-18.   Oil and Gas Production Water/Oil  Connectors
               Average Leak Rate Versus  Fraction Leaking at
               Several Leak Definitions
                               5-29

-------
0 2 B -

024 -
022 -
D 2 B -
D 1 B -
D 1 B -
014 -
012 -
D 1 B -
D B B -
D B B -
D B4 -
D B 2 -
OBB |

A
/
/ /

/ / A
/ / /
y& / /
/ / / p-
/ / / //
/ / / / &
/ / 5<7 / ./
/ / / \— */ s'
/ / / /*— ^ s/
/ / / / /^
/y!) / / //
/ / yf ;/ /^

// S j/ L-'M
.^^
IF
                 D.DD  B.D5  D . 1 D   D . 1 5  0,20   0.2'.


                          Fraction Leaking


                            Leak Definition
                                                  B , 3 B
Figure 5-19.    Oil and Gas Production Gas Flanges Average Leak
               Rate Versus Fraction Leaking at Several Leak
               Definitions
                              5-30

-------
             0,022
             0 , 0 1 8
          CD  D
          U
D 1 B
             D , D 1 D
             0,008
          [0
          cr  0
               DDE
          CD
          0>
          [0
             D , D D4
             D , D D ;
             D , D D D
                 D , D D   0 , D 5  D.1D  0,15   0,20  0,25  D.3D


                           Fraction  Leaking


                            Leak  Definition
Figure 5-20.   Oil and Gas Production  Light  Oil  Flanges Average
               Leak Rate Versus Fraction  Leaking at Several Leak
               Definitions
                               5-31

-------





r~\
CD
u
D
o
m
^_

u>
^j
CD
+J
[0
cr

to
CD
_i
CD
o>
[0
I
CD
>
2

3
3
3
3

3
3

3

3
3
3
3

3

3
3

3
3

3


017 -

D 1 B -
015 -
014 -
013 -

012 -
011 -

010 -

DOS -
D 3 B -
007 -
D D B -

005 -

D 04 -
003 -

D 0 2 ft
001 -

D 0 3 i

A
/
//
/ / A
/ / X
/ / /
/ / /
/ / / £
// / /%
/ / / / /
// / //
/ / /X //
// / //
///)&
/////
/////
i*y / //
/VL)///
//Yf//
////M>
// / /y
///^
//^ n

•of
1
I I I I I I
0,00 0,05 0,10 0,15 0,20 0,25 2,3
                          Fraction Leaking


                            Leak Definition
Figure 5-21.
Oil and Gas Production Gas Open-Ended Lines
Average Leak Rate Versus Fraction Leaking at
Several Leak Definitions
                              5-32

-------
            D , D D 9
            D , D D 8
            D , D D 7
          0)
          u
          O
          LO
              D D 3
          [0
          cc
          [0
          0)
          D)
          [0
            D , D D 1
            D , D D D
                 D , D D   D , D 5  D.1D  D.15   0,20  0,25  0,30


                           Fraction  Leaking


                            Leak  Definition
Figure 5-22.   Oil and Gas Production Heavy Oil Open-Ended Lines
               Average Leak Rate  Versus  Fraction Leaking at
               Several Leak Definitions
                               5-33

-------




0
u
D
o
LO
--
JZ

D)
^
0
-I-J
[0
cr
A/
[0
o

0
D)
[0
L
>
D
D
0
0

0
0


D

D

0
0

0
0


0
D


D

014 -
013 -
012 -
011 -

010 -
009 -


DOB -

007 -

006 -
005 -

004 -
003 -


002 -
A
001 -


ODD i

/
/ i —
/ / £
/ / /
/ / / F
y/7/ xS

/ ^/ / / /
// / //
//$>//
///!//

/////
/////
%////

////!§>
/////
/////
///^
///^ A A A

=*»#-
T?
i I I I I I
0,00 0,05 0,10 0,15 0,20 0,25 0,3
                          Fraction Leaking


                           Leak Definition
Figure 5-23.
Oil and Gas Production Light Oil Open-Ended Lines
Average Leak Rate Versus Fraction Leaking at
Several Leak Definitions
                               5-34

-------
            D , D D 9
                 D , D D  D , D 5  D . 1 D  D . 1 5  0,20   0,25  0,30
                          Fraction Leaking


                           Leak Definition
Figure 5-24.    Oil and Gas Production Water/Oil Open-Ended Lines
               Average Leak Rate Versus Fraction Leaking at
               Several Leak Definitions
                              5-35

-------





CD
u
D
o
m
^
-C
^
CD
[0
cr

to
0

CD
0>
[0
(
CD
>


B
B
B
B

B
B


B
B
B
B
B

B


B
B


B


Average Leak Rate Curves
Oil & Gas Production — Gas Othen
0 2 B -
D 2 B -
024 -
022 -

D 2 B -
0 1 B -


D 1 B -
014 -
012 -
D 1 B -
D B B -

DBS -


0 B4 -
D B 2 -


0 B B i
A
/ £
/ /
/ /

/ / XE
V7 / / s~
/ / / /
/ / / / &r
/ iZ5 ////
// $////
///%S
/ / / / )
-------
             D , D D D 3 6
             D , D D D 3 2
          0)
          [0
          cr
          Q)
          re
             D , D D D 1 6
             D , D D D 1 2
             D , D D D D B
              D D D D4
             D , D D D D D
                   D , D D D , Q 5  D . 1 D  0,15  0 , 2 D  Q , 2 5  D , 3 0


                            Fraction  Leaking


                            Leak  Definition


                    ~ ~ ~             n n n I.DDD ppmv
Figure 5-26.   Oil and Gas Production Heavy Oil Other Average
               Leak Rate Versus Fraction Leaking  at  Several  Leak
               Definitions
                               5-37

-------
                                                   ^
                   BB  B , 0 5  D.1D   D.15  0,20   B , 2 5  B , 3 B


                          Fraction Leaking


                            Leak Definition
Figure 5-27.    Oil and Gas Production Light Oil Other Average
               Leak Rate Versus Fraction Leaking at Several Leak
               Definitions
                              5-38

-------
                                                  D , 3 D
                                 o n
                                          ng
                            Leak Definition
Figure 5-28.
Oil and Gas Production Water/Oil Other Average
Leak Rate Versus Fraction Leaking at Several Leak
Definitions
                              5-39

-------
            D , D 24
            D , D 2 D


            0 . D 1 B


            D


            D


            D , D 1 2
              D D (
              DOB
              D D4
              D D ;
              ODD
                 0,00  0,05  D.1D   0,15  0,20   0,25  0,30


                          Fraction Leaking


                            Leak Definition
Figure 5-29.    Oil and Gas Production Gas Pump Average Leak Rate
               Versus Fraction Leaking at Several Leak
               Definitions
                               5-40

-------
                                       D , 2 D   D,2;
                                                  D , 3 D
                          Fraction Leaking


                            Leak Definition
Figure 5-30.
Oil and Gas Production Light Oil Pumps Average
Leak Rate Versus Fraction Leaking at Several Leak
Definitions
                               5-41

-------




CD
u
D
o
m
L
D>
^
CD

[0
cr
to
CD
CD
0>
[0
I
0
0
0
0
D

0

0
D
0
0

D

D
D
D
D

D
D 3 D -
D 2 B -
D 2 B -
D 24 -
D 2 2 -

D 2 B -

D 1 B -
D 1 B -
D 14 -
D 1 2 -

D 1 B -

D B B -
D B B -
D B4 -
D B 2 -

D B B |
A
/ fi~
/ / A
/ / /
// /
// / /^
////%.
ft/ / / /
/r/ / / /
//.///
///^/
/////®
/ / / / /
// ///
f///
//
/
/////
/////
M7 A
^7
t/'
1
                 D.DD  D.D5  D.1D   0,15  D.2D   D.25  0,30
                          Fraction Leaking


                            Leak Definition
Figure 5-31.
Oil and Gas Production Gas Valves Average Leak
Rate Versus Fraction Leaking at Several Leak
Definitions
                              5-42

-------
                                           ng
                           Leak Definition
Figure 5-32.
Oil and Gas Production Heavy Oil Valves Average
Leak Rate Versus Fraction Leaking at Several Leak
Definitions
                              5-43

-------





CD
u
D
o
m
^
-C
^
CD
[0
cr

to
o

CD
0>
[0
(
CD
>
B
B

B
B

B
B


B
B
B
B
B
B


B
B


B


0 2 B -
D 2 B -

024 -
022 -

D 2 B -
0 1 B -


D 1 B -
014 -
012 -
D 1 B -
D B B -
DBS -


0 B4 -
D B 2 :


0 B B i

£
//z
//
///
// / ]7
// / /~
/ / / /
VJ"/ / / /^
/Oj / / /
/ / / / /
////^/
/////
////
/$ / / /
//fy\/ /
///^Q
//////
/////
/////
/s///
" '// A A A

w'
i
I I I I I I
0 , B B B , 0 5 D . 1 D D . 1 5 0,20 B , 2 5 D . 3
                          Fraction Leaking


                            Leak Definition
Figure 5-33.
Oil and Gas Production Light Oil Valves Average
Leak Rate Versus Fraction Leaking at Several Leak
Definitions
                               5-44

-------
                                             D , 2 ;
                                                  D , 3 D
                          Fraction Leaking
                            Leak Definition
Figure 5-34.
Oil and Gas Production Water/Oil Valves Average
Leak Rate Versus Fraction Leaking at Several Leak
Definitions
                              5-45

-------
      TABLE 5-4.  EQUATIONS  RELATING AVERAGE LEAK RATE TO FRACTION LEAKING  AT SOCMI UNITS
(Jl
CTl
Leak definition
Equipment type (ppmv)
Gas valve 500
1000
2000
5000
10000
Light liquid valve 500
1000
2000
5000
10000
Light liquid pump 500
1000
2000
5000
10000
Connector 500
1000
2000
5000
10000
Equations3
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(
(
(
<
(0
(0
(0
(0
(
.044
.050
.057
.068
.078
.047
.053
.061
.077
.089
.095
0.11
0.13
0.20
0.24
.047
.060
.073
.092
0.11
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
LKFRAC)
+ 1
+ 2
+ 4
+ 8
+ 1
+ 2
+ 3
+ 5
+ 1
+ 1
+ 3
+ 4
+ 6
+ 1
+ 1
+ 1
+ 2
+ 3
+ 5
+ 8
.7E-05
.8E-05
.3E-05
.IE-OS
.3E-04
.7E-05
.9E-05
.9E-05
.1E-04
.7E-04
.1E-04
.6E-04
.7E-04
.4E-03
.9E-03
.7E-05
.5E-05
.5E-05
.4E-05
.IE-OS
   aALR = Average leak rate  (kg/hr per source)  and LKFRAC = leak fraction.

-------
     TABLE  5-5.   EQUATIONS RELATING AVERAGE LEAK RATE TO FRACTION LEAKING AT REFINERY UNITS
        Equipment type
Leak definition
     (ppmv)
(Jl
Equation3
Gas valve 500
1000
10000
Light liquid valve 500
1000
10000
Light liquid pump 500
1000
10000
Connector 500
1000
10000
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
(0.11
(0.13
(0.26
(0.038
(0.042
(0.084
(0.20
(0.23
(0.43
(0.014
(0.017
(0.037
X
X
X
X
X
X
X
X
X
X
X
X
LKFRAC) -
LKFRAC) -
LKFRAC) -
LKFRAC) -
LKFRAC) -
LKFRAC) -
LKFRAC) -
LKFRAC) -
LKFRAC) -
LKFRAC) -
LKFRAC) -
LKFRAC) -
f 8
t- 1
f 6
t- 2
t- 2
t- 1
t- 1
t- 2
t- 1
t- 1
t- 1
f 6
.8E-
.4E-
.OE-
.OE-
.8E-
.7E-
.3E-
.OE-
.2E-
.3E-
.8E-
.OE-
05
04
04
04
04
03
03
03
02
05
05
05
   aALR = Average leak rate  (kg/hr per source) and  LKFRAC  =  leak fraction.

-------
TABLE 5 - 6.
EQUATIONS RELATING AVERAGE LEAK  RATE  TO FRACTION
 LEAKING  AT  MARKETING TERMINAL UNITS
Equipment
Type
Gas
Connector



Light
Liquid
Connector



Gas Other




Light
Liquid
Other



Light
Liquid Pump



Leak
Definition
(ppmv)
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
1.
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
.017
.017
.034
.034
.034
.0021
.0028
.0042
.0058
.0065
.0018
.0021
.0023
.0029
2E-04
.019
.022
.025
.034
.034
.014
.018
.029
.051
.077
Equation3
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)

x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)

+ 5.3E-06
+ 5.3E-06
+ 5.9E-06
+ 5.9E-06
+ 5.9E-06
+ 7.0E-06
+ 7.1E-06
+ 7.1E-06
+ 7.2E-06
+ 7.2E-06
+ 3.1E-05
+ 4.0E-05
+ 4.8E-05
+ 8.4E-05

+ 2.1E-05
+ 2.2E-05
+ 2.2E-05
+ 2.4E-05
+ 2.4E-05
+ 9.6E-05
+ 1.2E-04
+ 1.6E-04
+ 2.1E-04
+ 2.4E-04
                             5-48

-------
  TABLE  5-6.   EQUATIONS RELATING AVERAGE LEAK RATE TO FRACTION
         LEAKING AT MARKETING TERMINAL UNITS  (CONTINUED)
Equipment
Type
Gas Valve




Light
Liquid
Valve



Leak
Definition
(ppmv)
500
1000
2000
5000
10000
500
1000
2000
5000
10000
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
(0
(0
(0
(0
1.
(0
(0
(0
(0
(0
.0012
.0017
.0017
.0017
3E-05
.0045
.0052
.0077
.013
.023
Equation3
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)

x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
+ 8
+ 9
+ 9
+ 9

+ 9
+ 9
+ 1
+ 1.
+ 1.
.9E
.2E
.2E
.2E

.5E
.8E
.IE
2E-
5E-
-06
-06
-06
-06

-06
-06
-05
05
05
aALR = Average leak rate  (kg/hr per source)
 LKFRAC = Leak fraction.
                               5-49

-------
TABLE 5-7.  EQUATIONS RELATING AVERAGE  LEAK RATE TO FRACTION
      LEAKING  AT OIL AND GAS PRODUCTION OPERATION UNITS
Equipment
Type
Gas
Connector



Light Oil
Connector



Water/Oil
Connector



Gas Flange




Light Oil
Flange



Leak
Definition
(ppmv)
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
.016
.018
.020
.023
.026
.016
.021
.022
.025
.026
.013
.014
.016
.023
.028
.043
.051
.059
.075
.082
.037
.046
.055
.068
.073
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Equation3
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
+- 7
f 8
f 8
f 9
+- 1
+- 7
f 8
f 8
f 9
f 9
+- 7
+- 7
f 8
f 9
+- 1
+- 1
+- 1
+- 2
+- 4
f 5
f 9
+- 1
+- 1
+- 2
+- 2
.7E-
.OE-
.5E-
.4E-
.OE-
.7E-
.3E-
.6E-
.2E-
.7E-
.8E-
.9E-
.3E-
.4E-
.OE-
.1E-
.8E-
.6E-
.7E-
.7E-
.4E-
.2E-
.6E-
.1E-
.4E-
06
06
06
06
05
06
06
06
06
06
06
06
06
06
05
06
06
06
06
06
07
06
06
06
06
                             5-50

-------
TABLE 5-7.  EQUATIONS RELATING AVERAGE LEAK RATE TO FRACTION
     LEAKING AT OIL AND  GAS  PRODUCTION OPERATION UNITS
                         (CONTINUED)
Equipment
Type
Gas
Open- Ended
Line



Heavy Oil
Open- Ended
Line



Light Oil
Open- Ended
Line



Water/Oil
Open- Ended
Line



Gas Other




Leak
Definition
(ppmv)
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
.037
.039
.045
.051
.055
.012
.015
.020
.030
.030
.030
.032
.036
.040
.044
.030
.030
.030
.030
.030
.055
.061
.066
.078
.089
Equation3
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
LKFRAC) •
+- 4
f 5
+- 7
+- 1
+- 1
+- 4
+- 4
f 6
+- 7
+- 7
f 3
+- 4
f 6
f 9
+- 1
f 3
f 3
f 3
f 3
f 3
+- 1
f 3
+- 4
f 8
+- 1
.1E-
.OE-
.5E-
.2E-
.5E-
.3E-
.9E-
.OE-
.2E-
.2E-
.8E-
.7E-
.7E-
.7E-
.4E-
.5E-
.5E-
.5E-
.5E-
.5E-
.8E-
.1E-
.5E-
.2E-
.2E-
06
06
06
05
05
06
06
06
06
06
06
06
06
06
05
06
06
06
06
06
05
05
05
05
04
                             5-51

-------
TABLE 5-7.  EQUATIONS RELATING AVERAGE LEAK RATE TO FRACTION
      LEAKING AT OIL AND GAS PRODUCTION OPERATION UNITS
                         (CONTINUED)
Equipment
Type
Heavy Oil
Other



Light Oil
Other



Water/Oil
Other



Gas Pump




Light Oil
Pump



Leak
Definition
(ppmv)
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
Equation3
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
(0
(0
3.
3.
3.
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
.0011
.0011
2E-05
2E-05
2E-05
.053
.058
.067
.075
.083
.066
.066
.066
.066
.069
.027
.052
.052
.074
.074
.071
.079
.082
.10 x
.10 x
x LKFRAC)
x LKFRAC)



x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
x LKFRAC)
LKFRAC) +
LKFRAC) +
+ 2.1E-05
+ 2.1E-05



+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+





3
4
6
8
1
2
2
2
2
5
1
2
2
3
3
7
1
1
5.
5.



.4E
.4E
.4E
.6E
.4E
.5E
.5E
.5E
.5E
.9E
.IE
.3E
.3E
.5E
.5E
.9E
.5E
.9E
1E-
1E-



-05
-05
-05
-05
-04
-05
-05
-05
-05
-05
-04
-04
-04
-04
-04
-05
-04
-04
04
04
                             5-52

-------
  TABLE  5-7.   EQUATIONS RELATING AVERAGE LEAK RATE TO FRACTION
        LEAKING AT OIL AND GAS PRODUCTION OPERATION UNITS
                            (CONTINUED)
Equipment
Type
Gas Valve




Heavy Oil
Valve



Light Oil
Valve



Water/Light
Oil Valve



Leak
Definition
(ppmv)
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
500
1000
2000
5000
10000
Equation3
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
ALR =
(0
(0
(0
(0
(0
(0
(0
(0
8.
8.
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
.070
.076
.083
.092
.098
.0013
.0013
.0013
4E-06
4E-06
.059
.069
.075
.083
.087
.022
.022
.064
.064
.064
x
x
X
X
X




X
X
X
X
X
X
X
X
X
X
LKFRAC) +
LKFRAC) +
LKFRAC) +
LKFRAC) +
LKFRAC) +
x LKFRAC)
x LKFRAC)
x LKFRAC)


LKFRAC) +
LKFRAC) +
LKFRAC) +
LKFRAC) +
LKFRAC) +
LKFRAC) +
LKFRAC) +
LKFRAC) +
LKFRAC) +
LKFRAC) +
9
1
1
1
2
+
+


9
1
1
1
1
8
8
9
9
9
.1E-
.1E-
.4E-
.9E-
.5E-
7.8E
7.8E
7.8E


.4E-
.2E-
.4E-
.7E-
.9E-
.1E-
.1E-
.7E-
.7E-
.7E-
06
05
05
05
05
-06
-06
-06


06
05
05
05
05
06
06
06
06
06
aALR = Average
 LKFRAC = Leak
leak rate (kg/hr per
fraction.
source]
                               5-53

-------
     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 5-35.  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.
     On figures 5-1 through 5-34, 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, 5-5, 5-6, and 5-7.
     The leak frequency immediately after a monitoring cycle is
Point Y on figure 5-35.  After an LDAR program is implemented for
a given time period, point Y will reach a "steady-state" value.
As presented in figure 5-35, 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.
                               5-54

-------
  0)
  c
  !2
  a
  a
  LL
        0.08
        0.07
        0.06
        0.05
0.04
        0.03
        0.02
        0.01
                Point X (initial leak frequency)
                                             Point F (final leak frequency)
                                     Monitoring Cycle
Figure 5-35.
          Simplified Graphical Presentation of Changes  in
          Leak Frequency After Implementation  of an LDAR
          Program
                                    5-55

-------
Based on these assumptions the value for point Y is calculated
using the following equation:
              Yi  =  zi - (FR x zi) + (FR x Zi x R)
where:
     Yj_   =  Leak fraction immediately after monitoring cycle i;
     Zj_   =  Leak fraction immediately preceding monitoring cycle
            i (note that Z-\_ equals point X.);
     R    =  Fraction of repaired sources for which a leak
            immediately recurs; and
     FR   =  Fraction of leaking sources successfully repaired.
     Point Z on figure 5-35 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 5-35, 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+1  =  Oc x (1 - Yi)  + Yi
where:
     zi+l   =   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 5-35.   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
                               5-56

-------
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-FLR)/ILR x 100
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 G.  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 5-8 presents the SOCMI gas valve occurrence rate,
recurrence rate, unsuccessful repair rate, and initial leak
frequency.   (See appendix G 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
                               5-57

-------
         TABLE 5-8.  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:^  7.5%
aSee appendix F for information on how the occurrence rate,
recurrence rate,  and unsuccessful repair rate were determined.
bBased on the SOCMI average emission factor for gas valves.
                               5-58

-------
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,
the steady-state leak frequency immediately after monitoring (see
point Yg on figure 5-35)  equals 0.29 percent.  The steady-state
leak frequency prior to monitoring  (see point Zg on figure 5-35)
equals 1.29 percent.  This gives an average of 0.79 percent as
the final leak frequency (see point F on figure 5-35).  The
calculations performed to determine the final leak frequency are
shown in table 5-9.  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 x 100
                = 87  percent.
                               5-59

-------
   TABLE 5-9.  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  PPMVa
              Starting parameters
                                                 Resulting parameters
(Jl
i
CTl
O
         Leak definition:
           = 10,000 ppmv
         Leak occurrence (Oc):
           = 1.00 percent
         Leak recurrence (R):
           = 14 percent
         Successful repair rate
           = 90 percent
         Initial leak frequency (Point X)
           = 7.5 percent
                       (FR)
            Steady-state leak frequency after
            monitoring  (Point Yg):
              = 0.29 percent
            Steady-state leak frequency immediately
            prior to monitoring   (Point Zg):
              = 1.29 percent
            Final leak frequency   (Point F)b:
              = 0.79 percent
                                          Calculations
Monitoring cycle
 Leak  frequency after
monitoring:  Point Yj_
      (percent)c
Leak frequency prior to
 monitoring:  Point Zj_
       (percent)d
1
2
3
4
5
6e
1
0
0
0
0
0
.70
.61
.36
.31
.29
.29
7
2
1
1
1
1
.50
.67
.60
.36
.30
.29
   aRefer to Figure 5-4 for graphical presentation of all points  identified  in  this  table.
          Leak Frequency equals the average of the prior to monitoring  and  after  monitoring
    steady- state leak frequencies.
   °Yi = zi - (FR * zi) +  (FR * zi * R)
   dZi + 1 = Oc * (1 - Yj_) + Yi

   eAfter the sixth monitoring cycle, the values for Yj_ and  Zj_ reach  steady- state .

-------
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.
                              5-61

-------