600281111
es EPA-
Environmental Protection
Agency June 1981
4>EPA Research and
Development
ANALYSIS OF SOCMI
VOC FUGITIVE
EMISSIONS DATA
Prepared for
Office of Air Quality Planning and Standards
Prepared by
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
Mt-II
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-81-111
June 1981
ANALYSIS OF SOCMI
VOC FUGITIVE EMISSIONS
DATA
Final Report
Prepared by:
G, J. Langley
S. M. Dennis
L. P. Provost
J. F. Ward
Radian Corporation
P. 0. Box 9948
Austin, Texas 78766
Contract No. 68-02-3171-28
EPA Project Officer
Dr. B. A. Tichenor
Chemical Processes Branch
EPA/IERL
Research Triangle Park, N,C. 27711
Prepared for:
Office of Air Quality Planning and Standards
-------
-------
CONTENTS
Figures . iii
Tables viii
1. Introduction 1
2. Summaries and Conclusions 4
Relationship of Leak Frequency to Process Parameters
(Section 3) 4
Emission Factor Development (Section 4) 6
Increase in Mass Emissions Due to Occurrence and
Recurrence (Section 5) ... 6
Impact of Response Adjustments on Leak Frequency
Estimation (Section 6) 9
3. Detailed Results for the Effects of Process Parameters on
Leak Frequency 12
Overview of Screening Data from 24 Chemical Units 13
Effect of Chemical Produced on Leak Frequencies 16
Effect on Leak Frequency of Primary Chemical in the
Process Line 22
Effect of Type of Valve on Leak Frequency . 31
Leak Frequency for Pump Seal Classification 43
The Effect of Line Temperature and Line Pressure 48
Effects of Line Temperature and Line Pressure on Pump
Seals, Flanges, and Open-Ended Lines 61
Effect of Ambient Temperature on Leak Frequency , 69
Effect of Elevation on Leak Frequency 71
4. Emission Factor Development for Three Processes 73
Distribution of Screening Values 73
Emission Factors and Cumulative Distributions of Total
Emissions by Screening Values 74
5. Evaluation of the Effects of'Leak Occurrence, Recurrence, and
Repair on Mass Emissions 117
Effect of Leak Occurrence on Mass Emissions 117
Effect of Leak Recurrence on Mass Emissions for
Valves 119
Further Analysis of Effect of Valve Maintenance on
Mass Emissions 121
-------
CONTENTS (CONTINUED)
6. Impact on Leak Frequency Estimates of Applying Chemical
Response Adjustments 125
Summary of Four Leak Frequency Estimates by Primary
Chemical 126
Summary of Four Leak Frequency Estimates by Process
Type 126
7. Statistical Considerations 138
Statistical Categorical Analysis Using Funcat
(Section 3) 138
Chi-Square Test for Independence (Section 3) 139
Confidence Intervals for Percent Sources Leaking
(Section 3) 140
Screening Value Distributions (Section 4) 141
Emission Factor Development (Section 4) 145
Cumulative Emission Functions 151
Increase in Mass Emissions Due to Occurrence and
Recurrence (Section 5) 155
Response Model Adjustments to Screening Values
(Section 6) 156
References 160
Appendices
A. Screening Data Summary 161
B. Detailed Information on Line Temperature and Line
Pressure 168
C. Summary Statistics and Detailed Information on the Effects
of Ambient Temperature and Elevation on Leak Frequency 190
D. Corrections to Screening Data 198
-------
FIGURES
Number Page^
3-1 Effect of Process Type on Percent of Valves Leaking 19
3-2 Categories of Sources for Further Analysis 27
3-3a Percent Leaking with 95 Percent Confidence Intervals
for Each Valve Type with Gas Service 36
3-3b Percent Leaking with 95 Percent Confidence Intervals
for Each Type of Valve with Light Liquid Service 37
3-4a Percent Leaking for Block Versus Control Valves in
Gas Service by Primary Material Group and Process
Unit Type 40
3-5 Percent Leaking with 95 Percent Confidence Intervals
for Each Type of Pump Seal with Light Liquid Service .... 47
3-6 Combined Effects of Line Temperature and Line Pressure
on Percent Leaking for Valves in Gas Service Within
Ethylene Process Units 55
3-7 Combined Effects of Line Temperature and Line Pressure
on Percent Leaking for Valves from Group 5 56
3-8 Combined Effects of Line Pressure on Percent Leaking
for Valves in Group 3 57
3-9 The Effect of Line Pressure on Percent Leaking with
95 Percent Confidence Intervals for Valves from
Group 4 and Group 8 . » 58
3-10 The Effect of Line Pressure on Percent Leaking with 95
Percent Confidence Intervals on Pump Seals in Light
Liquid Service 64
4-1 Typical Distribution of Loge (OVA Screening Value)
Ehtylene Process, Valves in Gas Service 76
4-2 Cumulative Distribution of Sources by Screening Values -
Cumene Process, Valves in Gas Service 77
111
-------
Number Page
4-3 Cumulative Distribution of Sources by Screening Values -
Cumene Process, Valves in Light Liquid Service 78
4-4 Cumulative Distribution of Sources by Screening Values -
Cumene Process, Pumps in Light Liquid Service 79
4-5 Cumulative Distribution of Sources by Screening Values -
Ethylene Process, Valves in Gas Service 80
4-6 Cumulative Distribution of Sources by Screening Values -
Ethylene Process, Valves in Light Liquid Service 81
4-7 Cumulative Distribution of Sources by Screening Values -
Ethylene Process, Pumps in Light Liquid Service 82
4-8 Cumulative Distribution of Sources by Screening Values -
Vinyl Acetate Process, Valves in Gas Service 83
4-9 Cumulative Distribution of Sources by Screening Values -
Vinyl Acetate Process,Valves in Light Liquid Service .... 84
4-10 Cumulative Distribution of Sources by Screening Values -
Vinyl Acetate Process, Pumps in Light Liquid Service .... 85
4-11 Emission Factors—Valves " . 87
4-12 Emission Factors—Pump Seals 88
4-13 Cumulative Distribution of Total Emissions by Screening
Values - Cumene Process, Valves in Gas Service 89
4-14 Cumulative Distribution of Total Emissions by Screening
Values - Cumene Process, Valves in Light Liquid Service . . 90
4-15 Cumulative Distribution of Total Emissions by Screening
Values - Cumene Process, Pumps in Light Liquid Service ... 91
4-16 Cumulative Distribution of Total Emissions by Screening
Values - Ethylene Process, Valves in Gas Service 92
4-17 Cumulative Distribution of Total Emissions by Screening
Values - Ethylene Process, Valves in Light Liquid Service . 93
4-18 Cumulative Distribution of Total Emissions by Screening
Values - Ethylene Process, Pumps in Light Liquid Service . . 94
4-19 . Cumulative Distribution of Total Emissions by Screening
Values - Vinyl Acetate Process, Valves with Light
Liquid Service 95
IV
-------
Number l?age
4-20 Cumulative Distribution of Total Emissions by Screening
Values - Vinyl Acetate Process, Valves in Light
Liquid Service 96
4-21 Cumulative Distribution of Total Emissions by Screening
Values - Vinyl Acetate Process, Pumps in Light
Liquid Service 97
4-22a Cumulative Distribution of Sources by Screening Values -
Cumene Process, Valves in Gas Service 98
4-22b Cumulative Distribution of Total Emissions by Screening
Values - Cumene Process, Valves in Gas. Service 99
4-23a Cumulative Distribution of Sources by Screening Values -
Cumene Process, Valves in Light Liquid Service 100
4-23b Cumulative Distribution of Total Emissions by Screening
Values - Cumene Process, Valves in Light Liquid Service . . 101
4-24a Cumulative Distribution of Sources by Screening Values -
Cumene Process, Pumps in Light Liquid Service 102
4-24b Cumulative Distribution of Total Emissions by Screening
Values - Cumene Process, Pumps in Light Liquid Service . . . 103
4-25-a Cumulative Distribution of Sources by Screening Values -
Ethylene Process, Valves in Gas Service 104
4-25b Cumulative Distribution of Total Emissions by Screening
Values - Ethylene Process, Valves in Gas Service 105
4-26a Cumulative Distribution of Sources by Screening Values -
Ethylene Process, Valves in Light Liquid Service 106
4-26b Cumulative Distribution of Total Emissions by Screening
Values - Ethylene Process, Valves in Light Liquid Service . 107
4-27a Cumulative Distribution of Sources by Screening Values -
Ethylene Process, Pumps in Light Liquid Service 108
4-27b Cumulative Distribution of Total Emissions by Screening
Values - Ethylene Process, Pumps in Light Liquid Service . . 109
4-28a Cumulative Distribution of Sources by Screening Values -
Vinyl Acetate Process, Valves in Gas Service 110
v
-------
Number Page
4-28b Cumulative Distribution of Total Emissions by Screening
Values - Vinyl Acetate Process, Valves in Gas Service .... Ill
4-29a Cumulative Distribution of Sources by Screening Values -
Vinyl Acetate Process, Valves in Light Liquid 112
4-29b Cumulative Distribution of Total Emissions by Screening
Values - Vinyl Acetate Process, Valves in Light Liquid
Service . 113
4-30a Cumulative Distribution of Sources by Screening Values -
Vinyl Acetate Process, Pumps in Light Liquid Service . . , . 114
4-30b Cumulative Distribution of Total Emissions by Screening
Values - Vinyl Acetate Process, Pumps in Light Liquid
Service 115
5-1 Before Minus After Maintenance Leak Rate - Valves
Screening <10,000 ppmv After Maintenance 122
5-2 Before Minus After Maintenance Leak Rate - Valves
Screening ^10,000 ppmv After Maintenance 123
6-1 OVA Reading vs. Method 1 Adjustment for Cumene Process
Valves in Gas Service 132
6-2 OVA Reading vs. Method 1 Adjustment for Cumene Process
Valves in Light Liquid Service 133
6-3 OVA Reading vs. Method 1 Adjustment for Ethylene Process
Valves in Gas Service 134
6-4 OVA Reading vs. Method 1 Adjustment for Ethylene Process
Valves in Light Liquid Service 135
6-5 OVA Reading vs. Method 1 Adjustment for Vinyl Acetate
Process Valves in .Gas Service 136
6-6 OVA Reading vs. Method 1 Adjustment for Vinyl Acetate
Process Valves in Light Liquid Service 137
B-l Distribution of Sources Screened by Line Pressure for
Ethylene and High Leaking Process Units by Chemical
Group for Valves with Gas Service 186
B-2 Distribution of Sources Screened by Line Temperature for
Ethylene and High Leaking Process Units by Chemical Group for
Valves with Gas Service 187
VI
-------
jjumber Page
B-3 Distribution of Sources Screened by Line Pressure for
Ethylene and High Leaking Process Units by Chemical
Group for Valves with Light Liquid Service 188
B-4 Distribution of Sources Screened by Line Temperature
for Ethylene and High Leaking Process Units by
Chemical Group for Valves with Light Liquid Service 189
-------
TABLES
Number
2-1 Estimated Emission Factors for Nonmethane Hydrocarbons
from Valves and Pump Seals (Ibs./hr./source and kgs./
hr. /source) 7
2-2 Summary of Percent of Sources Distribution Curves and Percent
of Mass Emissions Curves at Screening Value of 10,000.
pprav r . . . 8
2-3 Comparable Estimates for Percent Leaking (Valves) 11
3-1 Percent of Sources Leaking1 by Source 15
3-2 Percent Leaking for Each Chemical Produced as a Function of
Source Type and Stream Service 17
3-3 Definition of Chemical Process Groups. . .• 20
3-4 Leak Frequencies by Process Unit Group, Source Type and
Stream Service 21
3-5a Percent of Leaking Valves by Primary Material in Line. ... 23
3-5b Percent of Leaking Valves by Primary Material in Line. ... 24
3-6a Percent Leaking by Primary Material for Valves -
Gas Service 25
3-6b Percent Leaking by Primary Material for Valves -
Light Liquid Service 26
3-7a High Versus Low Leaking Primary Chemical Groups for High
Leaking Process Units 29
3-Tb High Versus Low Leaking Primary Chemical Groups for Ethylene
Process Units 30
3-8 Percent Leaking for All Types of Valves in Gas Service as
a Function of Process Group and Primary Material Group . . 32
3-9 Percent Leaking for All Types of Valves with Light Liquid
Stream Service by Process Group and Primary Material
Group 33
3-10 Leak Frequencies for All Types of Valves for Gas and Light
Liquid Stream Service 35
VI11
-------
Number Page
3-11 Results of Categorical Analysis on Valves1 39
3-12 Leak Frequencies for Pump Seals in Light Liquid Service ... 44
3-13 Results of Analysis of the Effects of Line Temperature and
Line Pressure on Leak Frequency for Valves 50
3-14 Line Temperature and Line Pressure and Their Combined Effects
on Valves in Gas Service within Ethylene Process Units. . . 51
3-15 Line Temperature and Line Pressure and Their Combined Effects
on Valves from Group 5* 52
3-16 Line Temperature and Line Pressure and Their Combined Effects
on Valves from Group 3* 53
3-17 Line Temperature and Line Pressure and Their Combined Effects
on Valves from Group 7A 54
3-18 Effect of Line Temperature and Line Pressure on Valves from
Group 6* 59
3-19 Effects of Line Temperature and Line Pressure on Valves from
Group 4 and Group 8 by Process Unit Group1. „..,..». 60
3-20 Effects of Line Temperature and Line Pressure on Pump Seals
with Light Liquid Service 63
3-21 Effects of Line Temperature and Line Pressure on Flanges
in Gas Service by Process Unit Group 65
3-22 Effects of Line Temperature and Line Pressure on Flanges in
Light Liquid Service by Process Unit Group 66
• 3-23 Effects of Line Temperature and Line Pressure on Open Ended
Lines in Gas Service by Process Unit Group 67
3-24 Effects of Line Temperature and Line Pressure on Open Ended
Lines in Light Liquid Service by Process Unit Group .... 68
3-25 Summary of the Effects of Ambient Temperature on Percent
Leaking 70
3-26 Summary of the Effects of Elevation on Percent Leaking. ... 72
4-1 Estimated Emission Factors for Nonmethane Hydrocarbons from
Valves and Pump Seals 86
4-2 Summary of Percent of Sources Distribution Curves and Percent
of Mass Emissions Curves at Screening Value of 10,000 pprav. 116
ix
-------
Number Page
5-1 Increase in Mass Emissions by Leak Occurrence for Valves
and Pump Seals Screening <10,000 ppmv Initially 118
5-2 Increase in Mass Emissions by Leak Recurrence for Valves
Screening <10,000 ppmv Immediately After Maintenance . . . 120
5-3 Weighted Percent Reduction in Mass Emissions for Valves
Screening ^10,000 ppmv Immediately Before Maintenance . . . 124
6-1 Percent Leaking Estimates for Valves in Light Liquid Service 128
6-2 Percent Leaking Estimates for Valves in Gas Service 129
6-3 Percent Leaking Estimates for Valves in Light Liquid
Service by Process Type 130
6-4 Percent Leaking Estimates for Valves in Gas Service by
Process Type 131
7-1 Comparison of Emission Factors with Quality Control
Estimates of Mean Leak Rates for Valves and Pump Seals . . 150
A-l Data Summary of Leak Frequencies for Various Sources in
Various Stream Services 162
B-l Summary Statistics for Line Temperature and Line Pressure
for Gas Service 170
B-2 Summary Statistics for Line Temperature and Line Pressure
for Light Liquid Service 171
B-3 Summary Statistics for Line Temperature and Line Pressure
in Heavy Liquid Service Within High and Ethylene Process
Units 172
B-4 Effects of Line Temperature and Line Pressure on Percent
Leaking for Valves in Gas Service Within Ethylene Process
Units 173
B-5 Effects of Line Temperature and Line Pressure on Percent
Leaking for Valves in Gas Service Within High Leaking
Process Units by Chemical Group 174
B-6 Effects of Line Temperature and Line Pressure on Percent
Leaking for Valves in Light Liquid Service Within
Ethylene Process Units by Chemical Group ... 175
x
-------
Number j?age
B-7 Effects of Line Temperature and Line Pressure on Percent
Leaking for Valves in Light Liquid Service Within High
Leaking Process Units by Chemical Group 176
B-8 Effects of Line Temperature and Line Pressure on Percent
Leaking for Pump Seals in Light Liquid Service 177
B-9 Effects of Line Temperature and Line Pressure for Flanges
in Gas Service From Ethylene Process Units 178
B-10 Effects of Line Temperature and Line Pressure for Flanges
With Gas Service From High Leaking Process Units by
Chemical Group 179
B-ll Effects of Line Temperature and Line Pressure for Flanges
in Light Liquid Service Within Ethylene Process Units
by Chemical Group 180
B-12 Effects of Line Temperature and Line Pressure on Percent
Leaking for Flanges in Light Liquid Service Within High
Leaking Process Units by Chemical Group 181
B-13 Effects of Line Temperature and Line Pressure on Percent
Leaking for Open Ended Lines in Gas Service Within
Ethylene Process Units 182
B-14 Effects of Line Temperature and Line Pressure on Percent
Leaking for Open-Ended Lines in Gas Service Within
High Leaking Process Units by Chemical Group 183
B-15 Effects of Line Temperature and Line Pressure on Percent
Leakage for Open-Ended Lines in Light Liquid Service
Within Ethylene Process Units by Chemical Group 184
B-16 Effects of. Line Temperature and Line Pressure on Percent
Leaking for Open-Ended Lines in Light Liquid Service
Within High Process Units by Chemical Group 185
C-l Summary of Ambient Temperature During Screening of
Various Source Types in Gas Service 192
C-2 Summary of Ambient Temperature During Screening of
Various Source Types in Light Liquid Service 193
C-3 Effect of Ambient Temperature on Percent of Sources
Leaking in Ethylene Process Units as a Function of
the Primary Chemical Groups 194
-------
Number Page
C-4 Effects of Ambient Temperature on Percent of Sources
Leaking in High Leaking Process Units as a Function
of the Primary Chemical Groups 195
C-5 Effects of Source Elevation on Percent Leaking for Ethylene
Process Units as a Function of Primary Chemical Groups . . . 196
C-6 Effects of Source Elevation on Percent Leaking in High
Leaking Process Units as a Function of Primary Chemical
Groups 197
D-l Corrections Affecting Results on Previous Reports 200
D-2 Corrections to Screening Data Sheets 202
xxi
-------
SECTION 1
INTRODUCTION
The contribution of fugitive leaks from process unit components are
being investigated as a potential source of Volatile Organic Compound (VOC)
emissions in the Synthetic Organic Chemical Manufacturing Industry (SOCMI).
The purpose of this study is to provide an in-depth analysis of data on
these emissions collected under EPA contracts 68-02-3171-1, 68-02-3173-2 and
11, 68-02-3174-5, and 68-02-3176-1 and 6 and 68-03-2776-4. These data were
collected by Radian, PEDCO, TRW and Acurex and are summarized in References
1 and 2. The results of this study will be available for use in evaluating
VOC fugitive emissions.
The study design and test procedures for the data analyzed in this
report are described in References 1 and 2. The 24 process units studied in
the data collection programs were selected to represent a cross-section of
the population of the SOCMI. Several factors were considered during process
unit selection. These factors included total annual production volume,
number of producers, process conditions, corrosivity, volatility, toxicity,
and value of the final chemical product. Factors varied widely from unit
type to unit type, so that the selected process unit types represented a
reasonable sample of the variety of chemical process units encountered in
SOCMI.
Evaluating the leak frequency in SOCMI was done by the "collection of
screening data from 24 process units, where a screening value is the maximum
repeatable concentration of total hydrocarbons detected at a source with a
portable hydrocarbon detector (Reference 1), Evaluation of maintenance was
done by measurement of fugitive emission leak rates (Ib./hour) at selected
sources before and after maintenance at six process units representing three
-------
chemical processes (Reference 2). The results of these two programs provide
the background information necessary for the current study:
• source population data
• screening value profiles for each source type
• screening-to-emission rate relationships
The screening procedures began with the definition of the process unit
boundaries. All feed streams, reaction/separation facilities, and product
and by-product delivery lines were identified on process flow diagrams and
in the process unit. Process data, including stream compositions, line tem-
peratures, and line pressures, were obtained for all flow streams.
The Century Systems Models OVA-108 and OVA-128 hydrocarbon detectors
were used for screening. The detector probe of the instrument was placed
directly on those areas of the sources where leakage would typically occur.
For example, gate valves were screened along the circumference of the annular
area around the valve stem where the stem exits the packing gland and at
the packing gland/valve bonnet interface. The actual leak rate measurements
were taken using a flow-through method described in Reference 8 and were
analyzed on Byron Total Hydrocarbon Analyzer.
All accessible sources of the following source types were screened:
• process valves,
• pump seals,
compressor seals,
agitator seals,
• relief valves,
• process drains, and
open-ended lines.
-------
Also, a randomly selected subset of flanges was screened. Originally, only
five percent of all flanges were screened. The subset was increased to 20
percent of all flanges when initial results indicated a higher frequency of
emitting flanges than had been encountered in previous programs. The impor-
tant variables available from this study are: screening value, source
category, stream service, source type, chemical produced, ambient temperature
elevation, line temperature and line pressure. For the purposes of this
report, a source is defined as "leaking" if its screening value is greater
than or equal to 10,000 ppmv.
This report is actually a presentation of four distinct data analysis
tasks. Section 2 is a short summary of the results of all four tasks. In
Section 3 a detailed analysis of the SOCMI screening data (from 24 process
units) is presented along with summaries of important correlating process
parameters (line pressure, etc.). Emission factor development for three
specific chemical processes (7 units) is presented in Section 4. The analysis
reported in Section 5, an extension of the results in Reference 2, is directed
at investigating the increase in mass emissions due to occurrence and recur-
rence of leaks. In Section 6 the impact on leak frequency from adjusting
screening values by chemical response models is investigated. The statistical
methods used in Sections 3 through 6 are presented in Section 7. Appendix A
is a statistical summary of all the screening data from the 24 untis.
Appendix B contains summary statistics and information on the effect of line
pressure and line temperature on the percent leaking. Appendix C contains
similar descriptions for ambient temperature and elevation. Appendix D is a
summary of all corrections made to the original data.
-------
-------
SECTION 2
SUMMARIES AND CONCLUSIONS
This section presents the major findings from the analyses discussed
in Sections 3 through 6.
RELATIONSHIP OF LEAK FREQUENCY TO PROCESS PARAMETERS (SECTION 3)
The process parameters that were examined for their effect on leak
frequency were: process, service, material in the line, line pressure, line
temperature, ambient temperature and source elevation. Data on four source
types (valves, pump seals, flanges and open ended lines) were used to examine
the effects of these parameters. The sources were grouped into 32 categories
(see Figure 3-2) based on source type, process type, stream service and
primary chemical in the line. These groupings were for statistical reasons
and were not based on engineering reasoning.
Stream service was defined as either gas, light liquid or heavy liquid
(Reference 1). Heavy liquids were not included in any analyses, since they
leaked so rarely regardless of the other conditions. Gas stream service
generally had a higher leak frequency than light liquid service. Proceeding
with four source types and two stream service types the data was then cate-
gorised by process unit as either ethylene processes, high leaking processes
or low leaking processes. The ethylene units were analyzed separately because
of the large number of sources in ethylene processes and the high leak fre-
quency. The high leaking group consists of all other units with greater
than 1% of all source types leaking. The low leaking group consisted of all
units with less than 1% of all source types leaking. Since there were very
few sources leaking, the low leaking process units were not considered in
further analyses. Within these process unit groups, the data was further
-------
subdivided by primary materials in the line. Caution should be used in
these evaluations, however, since other chemicals in the line may also have
an effect on leak frequency.
Examination of the data within these categories resulted in the follow-
ing conclusions for this data setr
Leak frequency was affected not only by the type of chemical process
but also by the type of primary material in the line.
• Control valves had a higher leak frequency than block valves.
For block valves, gate valves had a higher leak frequency than most
of the other types, and plug and ball valves have lower leak fre-
quencies .
On-line pump seals had an overall leak frequency of 13.1 percent
versus 4.9 percent for off-line pump seals.
These data did not show a difference in leak frequency between double
mechanical pump seals and single mechanical pump seals, although the
type of barrier fluid was unknown and therefore unaccounted for in
this analysis.
Line pressure was seen to have a statistically significant effect in
almost every case, with higher levels of pressure associated with
higher leak frequencies,
• Line temperature had no consistent effect on leak frequency. The
combined effect of line pressure and temperature was important in
some cases.
-------
Ambient temperature had a consistent effect on leak frequency,
however, the effect was not statistically significant for a majority
of the cases. Higher leak frequencies tended to be associated with
the higher ambient temperature category.
• Elevation had no consistent effect on leak frequencies. In the four
cases where a statistically significant effect was observed, sources
at ground level had a higher leak frequency than sources at higher
elevations.
EMISSION FACTOR DEVELOPMENT (SECTION 4)
The sources included in the development of the emissions factors are all
valves and pump seals screened in the seven ethylene, cumene, and vinyl
acetate process units or 51.2% (16,575) of all valves and pump seals screened
in the screening program. Since leak rate screening value models were only
developed for these three process types, emission factor estimation was
limited to these three'processes.
The emission factors developed in this study are reported in Table 2-1.
The emission factors for ethylene process are consistently higher than the
factors for the cumene and vinyl acetate processes. The vinyl acetate
process tends to have the lowest emission factors of the three process
types.
Cumulative distributions of screening values and mass emissions as a
function of screening values were also developed for each of the three
processes. Table 2-2 gives the estimates and confidence intervals from
these curves for a 10,000 ppmv screening value.
INCREASE IN MASS EMISSIONS DUE TO OCCURRENCE AND RECURRENCE (SECTION 5)
Further analysis of data collected during the EPA SOCMI maintenance
-------
TABLE 2-1. ESTIMATED EMISSION FACTORS FOR NONMETHANE HYDROCARBONS
FROM VALVES AND PUMP SEALS (Ibs./hr./source and kgs./hr;/
source)
Source Type
Emission Factor (95% Confidence Interval)
(Ibs./hr.)
(kgs./hr.)
Valves
- Gas Sen/ice
Ethylene processes
Cumene processes
Vinyl Acetate processes
- Light Liquid
Ethylene processes
Cumene processes
Vinyl Acetate processes
0.024(0.008, 0.07)
0.011(0.003, 0.05)
0.0046(0.001, 0.03)
0.020(0.007, 0.06)
0.0056(0.002, 0.02)
0.0003(0.0001, 0.002)
0.011(0.004, 0.03)
0.0052(0.001, 0.02)
0.0021(0.0004, 0.01)
0.010(0.003, 0.03)
0.0025(0.001, 0.01)
0.0001(0.00003, 0.001)
Pump Seals
- Light Liquid
Ethylene processes
Cumene processes
Vinyl Acetate processes
0.069(0.006, 0.8)
0.052(0.001, 2.7)
0.0043(0.0001, 0.1)
0.031(0.003, 0.4)
0.023(0.0004, 1.2)
0-0020(0.00006, 0.06)
-------
TABLE 2-2. SUMMARY OF PERCENT OF SOURCES DISTRIBUTION CURVES AND PERCENT
OF MASS EMISSIONS CURVES AT SCREENING VALUE OF 10,000 PPMV
Source Type
Percent of Sources
Screening ^ 10,000 ppmv
Estimate
95% Confidence
Interval
Percent of Mass Emissions
Attributable to Sources
Screening > 10,000 ppmv
95% Confidence
Estimate Interval
00
Valves
Gas
Ethylene
Cumene
Vinyl Acetate
Light Liquid
Ethylene
Cumene
Vinyl Acetate
15
16
3.7
26
12
0.2
(U, 16)
(13, 19)
(2, 5)
(24, 27)
(10, 13)
(0, 0.4)
94
94
90
89
80
25
(93, 95)
(90, 96)
(85, 94)
(87, 90)
(72, 86)
(9, 47)
Pump Seals
Light Liquid
Ethylene
t
Cumene
Vinyl Acetate
30
14
1.7
(20, 39)
(1, 27)
(0, 4)
96
89
67
(90, 98)
(50, 98)
( 5, 92)
-------
program (Reference 2) was done to estimate the effects of leak occurrence and
recurrence on mass emissions. The following conclusions are based on these
analyses:
The increase in emissions for valves for which a leak occurred over
a one to six month period was estimated to be 530% (95% confidence
interval of 200% to 900%).
• Not enough data was available to accurately quantify the effect on
emissions from leak occurrence from pump seals. However, the percent
increase estimate was 75% with a 95% confidence interval of -100%
to 6000%.
• The percent increase in emissions for valves with a leak recurrence
within the six month period was estimated to be 510% (95% confidence
interval of -100% to 1700%).
• Further analysis of the effect of valve maintenance on emissions
showed a 98% reduction in emissions for valves which were "repaired"
(screening valve <10,000 ppmv after maintenance) and a 63% reduction
for sources which were "not repaired" (screening valve remained -
^10,000 ppmv after simple, on-line maintenance).
IMPACT OF RESPONSE ADJUSTMENTS ON LEAK FREQUENCY ESTIMATION (SECTION 6)
Three different techniques were used to adjust the original screening
value for each source:
the original OVA reading adjusted for the associated OVA response
relationship of the primary chemical compound in the line,
" weighted logarithmic average of response of primary and secondary
chemicals, and
-------
weighted arithmetic average of response of primary and secondary
chemicals.
The percent of leaking valves was calculated for each of the three
estimates for both gas and light liquid services. The three estimates were
found to be similar in most cases to the leak frequency based on the original
screening valves. Table 2-3 presents the overall results.
10
-------
TABLE 2-3. COMPARABLE ESTIMATES FOR PERCENT LEAKING (VALVES)
(24 SOCMI Process Units)
Process
Stream
Gas
Light
Liquid
Number
Screened1
9374
18,133
Percent
Leaking
Based on
OVA
Readings
11.3
6.1
Percent
Leaking
Based on
Method 1
Adjustments2
10.1
5.3
Percent
Leaking
Based on
Method 2
Adiustments3
10.2
5.6
Percent
Leaking
Based on
Method 3
Adiustments1*
10.3
5.5.
sources with screening valves = 10,001 ppmv were excluded.
2Method 1 is the adjustment to the OVA Reading based on the response of the
primary chemical in the line.
3Method 2 is the mixed chemical weighted logarithmic average technique.
^Method 3 is the mixed chemical weighted average technique.
11
-------
SECTION 3
DETAILED RESULTS FOR THE EFFECTS OF
PROCESS PARAMETERS ON LEAK FREQUENCY
The effects of various process parameters on leak frequency are
evaluated in this section. The process variables analyzed are source
category, stream service, source type, chemical produced, ambient temperature,
elevation, line temperature and line pressure. Each of these variables was
examined to determine which of them is associated with high or low leak
frequencies. Leak frequency data from four source types are analyzed in
detail in this section. They are open-ended lines, valves, pumps, and
flanges. Simple summary statistics for all source types are presented in
Appendix A. The data were grouped into exclusive categories for statistical
reasons (not engineering) as outlined in the following paragraphs.
Data for this analysis come from an EPA study in which all sources
in 24 chemical process units were screened (Reference 1). In a data collec-
tion study such as this, it is possible to have several of the process
parameters confounded. This means it can be difficult to separate the
effects of one parameter from that of another. For example, if one process
source type does in fact have a high frequency of leaks, but is almost
always associated with a certain type of stream service, it may appear that
the high leak frequency is associated with the stream service. If the data
are grouped by both source type and stream service, the effect of each of
these two variables can be seen. To avoid this type of problem, the data
have been analyzed in smaller groups whenever a possibility of confounding
was suspected.
Another reason the data were grouped into subsets is that the analysis
procedure used to statistically evaluate factors affecting the leak frequency
12
-------
is very sensitive to frequencies of zero. That is, if there were no sources
leaking (or very few) in a particular category (e.g., heavy liquid) the
analysis procedure is not appropriate. To avoid this problem, the data to
be analyzed for statistical significance were first categorized to include
only groupings that displayed at least a moderate percentage of leaking
sources. Summary statistics for the groupings not statistically analyzed
(heavy liquids and process units with less than 1 percent of the sources
leaking) are presented separately in the appendices.
OVERVIEW OF SCREENING DATA FROM 24 CHEMICAL UNITS
Table 3-1 gives information on the number of sources screened, the
number that were leaking and the percentage that were leaking in the 24
chemical units screened. This information is given for each source type and
each stream service within each source category. (The stream service classi-
fications are described in Reference 1.) It can be seen from this table
that sources in the heavy liquid service category have a fairly low leak
frequency. There are also fewer heavy liquid service sources than gas or
light liquid in each source type. However, even in a group such as valves,
where there were 3,632 valves in heavy liquid service, the leak rate is
very low (0.4 percent, -or 13 leaking sources). Table 3-1 shows that valves
in gas service have both a large number and high percentage of leaking
sources. It also appears that the percent leaking varies with both source
type and stream service.
Valves as a source type had the largest number of screening values.
Flanges and open-ended lines also had a large number (although only 5 to 20
percent of the flanges were screened). For further analysis, these three
categories plus pump seals were investigated. It was felt that the sample
sizes of the other categories were too small to allow meaningful subcate-
gorization of the source type.
Since only 17 sources in heavy liquid stream service in the source types
to be further analyzed were found to be leaking, sources in this service were
13
-------
not included in further anlaysis of factors affecting the leak frequency.
However, summary statistics for this stream service are included in the later
sections, where appropriate.
14
-------
TABLE 3-1. PERCENT OF SOURCES LEAKINGJ
(24 Process Units)
BY SOURCE
Source Service
Valves Gas*
Light Liquid3
Heavy Liquid
Pump Seals Light Liquid3
Heavy Liquid
Flanges Gas1
Light Liquid3
Heavy Liquid
Open Ended Lines Gas
Light Liquid5
Heavy Liquid
Process Drains Gae
Light Liquid3
Heavy Liquid
Agitator Seals Gas
Light Liquid
Heavy Liquid
Relief Valves Gas3
Light Liquid*
Heavy Liquid
Compressors Gas5
Other2 Gas
Light Liquid1
Heavy Liquid
Number
Screened
9669
18299
3632
646
97
1450
2833
607
923
3605
477
83
496
28
7
8
1
84
68
3
22
19
34
2
Number
1103
1183
13
57
2
66
36
0
54
141
2
2
19
2
1
0
0
3
2
0
2
3
2
0
Sources with Screening
Percent
U. 4
6.5
0.4
8.8
2,1
4.6
1.3
0
5.9
3.9
1.3
2.4
3.8
7.1
14.3
0
0
3.6
2.9
0
9.1
15. B
5.9
0
Values >1 0,000
95Z Confidence Interval
for Percent >10,000
(10.8, 12.0)
(6.1, 6.9)
(0.2, 0.7)
(6.4, 11.0)
(0.3, 7.3)
(3,7, 5.7)
(0,9, 1.7)
(0, 0.6)
(4.4,7.9)
(3.4, 4.7)
(0.5, 2.8)
(0.3, 8.4)
(2.3, 5,9)
(0.9, 23.5)
(0.4, 57.9)
(0, 36.9)
(0, 100)
(0.7, 10.1)
(0.4, 10.2)
(0, 70.8)
(1.1, 26.2)
(3.4, 39.6)
(0.7, 19.7)
<0, 84.2)*
1A leaking source is defined as one with a screening value >10,000 ppmv.
2Includea filters, vacuum breakers, expansion Joints, rupture disks, sight p.iass aeals, etc.
The numbers in each column may be different from that found in Reference ; because of corrections to the original data
(See Appendix D).
-------
EFFECT OF CHEMICAL PRODUCED ON LEAK FREQUENCIES
Table 3-2 describes the screening data in terms of chemical produced
by the source types and service categories outlined earlier. Some differ-
ences between the chemical processes are apparent. The production of ethy-
lene appears to be associated with a leak frequency that is higher than that
found with the production of any of the other chemicals. Leak frequencies
from the Cumene and MEK units are also high. Other processes had very low
leak frequencies for all four of the source types. The formaldehyde unit
screened only had two leaks and the two adipic acid units had no leaks from
the four source types. Figure 3-1 graphically presents the estimated percent
leaking along with 95 percent confidence intervals for valves in gas and
light liquid service by process type.
It is clear from looking at Figure 3-1 and examining Table 3-2 that the
breakdown by process type, in addition to source type and stream service,
results in some subsets with few or no leaking sources. To avoid the problem
of analyzing such small groups, a method of grouping the chemicals produced
was devised. Three chemical process groups based on overall leak frequency
were formed. The groups are Low Leaking Process, High Leaking Process, and
Ethylene Process. Each category, and the processes and unit identification
numbers that are associated with it, is given in Table 3-3. The Low Leaking;
group contains data on chemicals whose leak frequency was less than one per-
cent for all source types and stream services. The overall leak frequencies
for the High Leaking group range from one percent to six percent.
Table 3-4 summarizes the data available for further analysis for the
subcategories formed by the source type, service category, and chemical
process groups. In the analysis of the effect of other process parameters
on leak frequency, only the High Leaking and Ethylene groups were used. The
Low Leaking group had too few leaks to adequately determine any types of
effects on leak frequency of the other variables.
16
-------
TABLE 3-2. PERCENT LEAKING FOR EACH CHEMICAL PRODUCED AS A FUNCTION OF
SOURCE TYPE AND STREAM SERVICE
Source/Chemical (units) '
Valves
Vinyl Acetate (1,3)
Kthylene (2,4,11)
Cumene (5,6)
Acttone/I'lienol (12)
lithylene bichloride (2.1,29)
Vinyl Chloride Monomer (20,28)
Formaldehyde (22)
Methyl Ethyl Ketone (31,32)
Acetaldehyde (33)
Methyl Metliacrylate (34)
Adipiu Acid (35,61)
Chlorinated Ethanes (60, 62)
Ai:rylpnitrlle (65,66)
1,1,1-Trlcliloroetlwna (&1)
Vinyl Acetate (1,3)
Ethyleme (2,1,11)'
Cumene (5,6)
Acetone/Phenol (12)
Ktliylene bichloride (21,29)
Vinyl Chloride Monomer (20,28)
Formaldehyde (22)
Methyl Ethyl Ketone (31,32)
Acetaldehyde (33)
Methyl Metliacrylate (34)
Atllpic Acid (35,64)
Chlorinated Ethanes (60,62)
Ai.'1-ylnnltrlle (65,66)
1,1 l-Triehlorocthane (filj
Number
Screened
919
6291
418
8
103
' 112
41
207
178
190
95
48
396
GAS
Number
Leaking
35
934
63
0
4
30
1
19
8
0
0
0
9
LIGHT LIQUID
Percent
Leaking
3.
14.
14.
1.
7.
2.
9.
4.
2.
•"•
7
8
1
0
0
3
4
2
5
0
0
0
3
~
_
-
-
-
-
-
-
-
-
-
~
-
-
Number
Screened
2137
4176
799
1818
2256
1209
121
671
551
1058
17
1620
1494
373
89
76
25
86
58
65
8
31
32
45
60
61
10
Number
Leaking
8
969
84
6
24
12
0
34
3
1
0
10
28
4
4
20
4
2
3
7
0
1
3
2
5
5
1
Percent
Leaking
0.
23.
10.
0.
1.
1.
5.
0.
0,
0.
0.
1.
4.
26.
16.
2.
5.
10.
3.
9.
4.
8.
a.
10.
4
2
5
3
1
0
0
1
5
1
0
6
9
1
5
3
0
3
2
8
0
2
4
4
-
3
2
0
Number
Screened.
124
1237
198
488
1478
12
95
5
15
3
36
30
B
HEAVY LIQUID
Number
Leaking
0
13
0
0
0
0
0
0
0
0
0
;
0
'I
Percent
Leaking
0
1.1
0
0
0
0
0
0
0
0
0
0
25.0
,
(Continued)
-------
TABLE 3-2. (continued)
CO
Source/Chemical (units) '
flanges
Vinyl Acetate (1,3)
Ethylene (2,4,11)
Cuinene (5,6)
Ace tone /Phenol (12)
Jithylene Bichloride (21,29)
Vinyl Chloride Monomer (20,28)
Formaldehyde (22)
Methyl Elliyl Ketone (31,32)
Acetaldeliyde (33)
Methyl Methacrylate (34)
Adiplc Acid (35,64)
Chlorinated Ethanes (60, 62)
Acrylonltrile (65,66)
J ,1,1-Ttrichloroethane (*1)
Open Ended Lines
Vinyl Acetate (1,3)
Ethylene (2,4,11)
Cumpne (5, 6)
Acetone/Phenol (12)
Ethylene Bichloride (21,29)
Vinyl Chloride Monomer (20,28)
Formaldehyde (22)
Methyl Ethyl Kutone (31,32)
Acetaldehyde (33)
Methyl Methacry late (34)
Ajlplc Acid (35,64)
Chlorinated Ethanes (60,62.)
Acrylonltrile (65,66)
l.Ul-Trlchloroethane (61.)
Number
Screened
107
634
367
25
16
2
22
32
38
49
16
142
145
305
6
2
100
55
14
37
34
63
19
27
116
GAS
Number
Leaking
3
39
19
1
2
0
0
0
0
0
0
2
8
37
0
0
0
2
0
3
3
0
0
0
1
LIGHT LIQUID
Percent
Leaking
2
6
5
--
4
12
1
•
5
12
3
8
8
0
—
.8
.2
.2
—
.0
.5
0
0
0
0
0
0
.4
—
.5
.1
0
0
0
.6
0
.1
.8
0
0
0
.9
—
Number
Screened
173
407
468
82
163
47
8
76
144
247
2
461
382
73
318
214
15
518
475
340
36 .
186
158
335
1
412
486
111
Number
Leaking
0
25
9
0
1
0
1
0
0
0
0
0
0
0
8
41
2
8
16
18
0
19
8
1
0
6
12
2
Percent
Leaking
6.
1.
0,
12.
2,
19,
13.
1,
3,
5,
10,
5,
0,
1,
2,
1
0
,1
6
0
.6
0
,5
0
0
0
0
0
0
0
.5
.2
.3
,5
.4
,3
0
,2
,1
.3
0
.5
,5
.8
Number
Screened
8
89
130
30
320
2
28
22
91
1
107
214
4
38
HEAVY LIQUID
Number
Leaking
0
0
0
0
0
0
0
2
0
0
0
0
0
4
Percent
Leaking
0
0
0
0
0
0
0
9.1
0
0
0
--_—
0
0
10.5
-------
r
Opp,r 951
Confidence
Interval
VALVES-GAS SERVICE PERCENT LEAKING BY PROCESS TYPE
35
^30
r
c 25
n __
I 15
•
? 19
' I1 ]I
I 31,32 I 34 169, 621
2,4,11 t 12 I 20,28 t 31,32 I 34
3 5,6 21,29 22 33 35,64 65,66
Process Unit Number1
ii-*-
VALVES-LIQUID SERVICE PERCENT LEAKING BY PROCESS TYPE
40
35
:3e
r
<= 25
£ 20
I 15
*
? 10
[S3 cq£3 i-4-i
n—i—i—1—i—I—i—|—i—I i | i r
1,3 I • 5,6 I 21,28 I 22 I 33 I 35,64 ( 65,66 I
2.4,11 12 20,28 31,32 34 60,62 61
Figure 3-1. Effect of Process Type on
Percent of Valves Leaking
1 See Table 3-3 for definition of process type identification numbers
19
-------
TABLE 3-3. DEFINITION OF CHEMICAL PROCESS GROUPS
Process Group
"Low Leaking"
<1% of all
source types
leaking
"High Leaking"
>1% of all
source types
leaking
"Ethylene"
Chemical Process
Adipic Acid
Acetone
Formaldehyde
Methyl Methacrylate
Trich.loroethylene/
Perchloroethylene
Vinyl/Ethylene Bichloride
Ace t aldehyde
Acrylonitrile
Vinyl Acetate
Vinyl Chloride Monomer
Ethylene Bichloride
1,1,1-Trichloroethane
Cumene
Methyl Ethyl Ketone (MEK)
Ethylene
Unit Numbers
35, 64
12
22
34
60
62
33
65, 66
1, 3
20, 28
21, 29
61
5, 6
31, 32
2, 4, 11
Percent
Leaking*
0.0
0.5
0.8
0.3
0.8
0.0
2.3
1.7
1.4
2.8
1.2
1.2
6.3
5.9
12.9
*For all source types and stream services
20
-------
TABLE 3-4. LEAK FREQUENCIES BY PROCESS UNIT GROUP,
SOURCE TYPE AND STREAM SERVICE
Ethylene Process Units
Source Type
Valves
Pump Seals
Flanges
Open Ended Lines
Stream Service
gas
light liquid
light liquid
gas
light liquid
gas
light liquid
Number Number
Screened Leaking
6294 934
4176 969
76 20
634 39
407 25
305 37
214 41
Percent
Leaking
14.8
23.2
26.3
6.2
6.1
12.1
19.2
95% Confidence
Interval for
Percent Leaking
(13.8, 15.8)
(21.8, 24.6)
(16.9, 37.7)
(4.4, 8.4)
(4.0, 8.9)
(8.6, 16.3)
(14.0, 25.3)
High Leaking Process
Source Type
Valves
Pump Seals
Flanges
Open Ended Lines
Stream Service
gas
light liquid
light liquid
gas
light liquid
gas
light liquids
Number Number
Screened Leaking
2993 168
9490 197
371 28
711 27
1626 10
493 17
2089 85
Percent
Leaking
5.6
2.1
7.5
. 3.8
0.6
3.4
4.1
Units
95% Confidence
Interval for
Percent Leaking
(4.7, 6.5)
(1-8, 2.4)
(5.1, 10.5)
(2.5, 5.5)
(0.3, 1.1)
(2.0, 5.4)
(3.3, 5.0)
Source Type
Valves
Pump Seals
Flanges
Open Ended Lines
Stream Service
gas
light liquid
light liquid
gas
light liquid
gas
light liquid
Low Leaking
Number Number
Screened Leaking
382 1
4626 16
199 9
105 0
798 1
125 0
1300 15
Process Units
Percent
Leaking
0.3
0.4
4.5
0.0
0.1
0.0
1.2
95% Confidence
Interval for
Percent Leaking
(0.01, 1.5)
(0.2, 0.6)
(2.1, 8.4)
(0.0, 3.4)
(0.0, 0.7)
(0.0, 2.9)
(0.7, 1.9)
21
-------
EFFECT ON LEAK FREQUENCY OF PRIMARY CHEMICAL IN THE PROCESS LINE
The effect on leak frequency of the primary chemicals in the process
lines is investigated in this section. The definition of primary chemical
is described in Reference 1. Only the primary chemical is investigated here;
the influence of the other chemicals in the line is not evaluated. The
results of this section should be considered with this in mind.
Tables 3-5a and 3-5b display the percent of leaking sources by their
primary chemical in the line for valves - gas service and liquid service,
respectively. Large differences in leak, frequency between primary chemicals
can be seen in these tables. Because of these differences it was decided
to further categorize the sources by the primary chemical in the line, de-
pending on the leak frequency associated with that chemical.
To do this categorization, the percent leaking data for valves associated
with primary chemicals were analyzed for the categories previously established
(source type, stream service, and process groups). Tables 3-6a and 3-6b dis-
play this data for valves. It can be seen that chemicals associated with
high percent leaking in the ethylene group were also seen to be associated
with high percent leaking in the high leaking process unit grouping. For
example, ethylene as a primary chemical in ethylene process units has a high
percent leaking and, it also was found to have a frigh percent leaking in
other process units.
Using the data from Table 3-6, the primary chemicals were grouped into
two categories. If the percent of leaking (from Tables 3-6a and 3-6b) was
above 5% the chemical was put into the high leaking chemical group. Other-
wise it was put into the low leaking group. The resulting final groupings
of the screening data for further analyses are shown in Figure 3-2.
22
-------
TABLE 3-5a. PERCENT OF LEAKING VALVES BY PRIMARY MATERIAL IN LINE
(All Process Units)
Valves - Gas Service
Chemical
Ethylene
Methane
Propylene
1,2-Ethylene Bichloride
Ethane
Benzene
Acrylonitrile
Vinyl Acetate
Ac et aldehyde
Propane
Acetic Acid
Methyl Ethyl Ketone
Vinyl Chloride
Other Chemicals
Total
Number
Screened1
3134
1849
1128
525
379
332
287
272
179
145
125
116
96
851
9418
Percent of
Total Gas
Service
Valves
33.3
19.6
12.0
5.6
4.0
3.5
3.0
2.9
1.9
1.5
1.3
1.2
1.0
9.0
100%
Number
Leaking1
498
232
207
4
35
53
0
0
4
18
1
7
0
42
1101
Percent
Leaking
15.8
12.5
18.3
0.8
9.2
16.0
0.0
0.0
2.2
12.4
0.8
6.0
0.0
4.9
11.7
Numbers displayed in this table may not add up to totals in previous
sections due to missing information on primary chemicals.
23
-------
TABLE 3-5b. PERCENT OF LEAKING VALVES BY PRIMARY
MATERIAL IN LINE
(All process units)
Chemical
1,2-Ethylene Dichloride
Propylene
Ethylene
Acetic Acid
Acrylonitrile
Vinyl Acetate
1,1,2-Trichloroethane
Cumene
Vinyl Chloride
Percfaloroethylene
Phenol
Benzene
Acetaldehyde
Methyl Ethyl Ketone
Methyl Methacrylate
Methanol
Ethane
cc-Methyl Styrene
Hydrocarbons-Cs +
Tri chlo roethylene
Acetone
Methane
Sec Butyl Alcohol
Acetone Cyanohydrin
Other Chemicals
Total
Valves
Number
Screened
2809
1604
1230
1162
1126
973
914
773
611
601
594
536
456
425
393
373
328
326
323
272
209
205
202
191
1572
18208
- Light Liquid
Percent of
Total Light
Liquid
Service
Valves
15,4
8.8
6.8
6.4
6.2
3.3
5.0
4.2
3.4
3.3
3.3
2.9
2.5
2.3
2.2
2.0
1.8
1.8
1.8
1.5
1.1
1.1
1.1
1.0
8.6
100%
Service
Number1
Leaking
32
488
321
6
6
3
4
4
4
3
0
49
2
23
1
4
92
0
8
6
5
36
10
0
69
1176
Percent
Leaking
1.1
30.4
26.1
0.5
0.5
A •>
W.J
0.4
0.-5
0.6
0,5
0.0
9.1
0.4
5.4
0.2
1.1
28.0
0.0
2.5
2.2
2.4
17.6
5.0
o.o
4.4
6.5
Numbers displayed in this table may not add up to totals in previous sections
due to missing information on primary chemicals.
24
-------
TABLE 3-6a. PERCENT LEAKING BY PRIMARY MATERIAL FOR VALVES - GAS SERVICE
ro
Ln
High Leaking Process Units
Ethylene
Methane
Propylene
1,2-Ethylene
Dichloride
Ethane
Benzene
Acrylonitrile
Vinyl Acetate
Acetaldehyde
Propane
Acetic Acid
Methyl Ethyl
Ketone
Vinyl Chloride
Other Chemicals
TOTAL
Number1
Screened
680
-
69
505
-
282
287
272
179
81
125
116
96
294
2986
Percent ,
of
Total
Screened
22.
-
2.
16.
-
9.
9.
9.
6.
2.
4.
3.
3.
9.
100.
8
3
9
4
6
1
0
7
2
8
2
9
0
Number*
Leaking
62
-
15
4
-
50
0
0
4
12
1
7
0
13
168
Ethylene Process Units .
Percent
of
Percent Number1 Total Number1
Leaking Screened Screened Leaking
9.
-
21.
0.
-
17.
0.
0.
2.
14.
0.
6.
0.
4^
5.
1 2454 40.6 436
1849 30.6 232
7 1059 17.5 192
8 -
379 6.3 35
7 50 0.8 3
0 - - _
0 - -
2 -
8 64 1.1 6
8 - -
0 - -
,0 - - -
4 195 3.2 28
6 6050 100.0 932
Percent
Leaking
17.8
12.6
18.1
-
9.2
6.0
-
-
-
9.4
_
14.4
15.4
Numbers displayed In this table may not add up to totals In previous sections due to
missing information on primary chemicals.
-------
TABLE 3-6b. PERCENT LEAKING BY PRIMARY MATERIAL FOR VALVES -
LIGHT LIQUID SERVICE
High Leaking Process Units
Primary Chemical
in the Line
Number
Screened
Percent
of
Total
Screened
Number 1
Leaking
Percent
Leaking
Ethylene Process Units
Number 1
Screened
Percent
of
Total
Screened
Number
Leaking
Percent
Leaking
1,2 Ethylene 2809 29.7 32
Dichlorlde
Propylene 253 2.7 44
Ethylene 9 0.1 0
Acetic Acid 1162 12.3 6
Acrylonitrile 1126 11.9 6
Vinyl Acetate 973 10.3 3
1,1,2 Trichlorethane - -
Vinyl Chloride 611 6.5 4
Perchloroethylene - -
Phenol - -
Benzene 432 4.6 48
Acetaldehyde 456 4.8 2
Methyl Ethyl Ketone 425 4.5 23
Methyl Methacrylate - - -
Methanol - -
Ethane -
Hydrocarbons Cj -
a-Methyl Styrene - -
Trichloroethylene - -
Acetone - -
Methane - - -
Sec Butyl Alcohol 202 2.1 10
Acetone Cyanohydrin - - -
Other Chemicals 827 8.7 _12_
TOTAL 9453 100.0 193
1.1
17.4
0.0
0.5
0.5
0.3
0.7
11.1
0.4
5.4
5.0
1.4
2.0
1351
1221
104
68
328
32.8
29.6
2.5
1.6
8.0
205 5.0
844 20.5
444
321
4
92
4121 100.0
J>8
966
32.9
26.3
1.0
5.9
28.1
36 17.6
.1
15.4
1Numbers displayed in this table may not add up to totals in previous sections due to missing
information on primary chemicals.
-------
SOURCE
TYPE
PROCESS
UNITS
STREAM
SERVICE
Valves
Pump Seals
Flanges
Open Ended Lines
Ethylene
Unit*= 2,4,11
ho
High
Leaking
GROUP 1
Low
Leaking
GROUP 2
Light Liquid
High
Leaking
GROUP 3
Low
Leaking
GROUP 4
Gas
High
Leaking
GROUP 5
High Leaking
Unit=l,3,5,6,20,21
28,29,31,32,33,
61,65,66
Light Liquid
Low
Leaking
GROUP 6
High
Leaking
GROUP 7
Low
Leaking
GROUP 8
PRIMARY
CHEMICAL
Ethylene None
Methane
Propylene
' Ethane
Benzene
Propane
Other
Ethylene
Propylene
Benzene
MEK
Ethane
Methane
Butylene
C6+
Methanol
Propane
Butanes
Butadiene
Butylenes ,
C4+
C5+
Ethanol
N Butane
Benzene
mixed
Ethylene
Propylene
Benzene
Propane
MEK
1,2-Ethy- Ethylene
lene Dich- Propylene
loride Benzene
Acryloni- MEK
trile
Vinyl Ace-
tate
Acetalde-
hyde
Acetic Acid
Vinyl Chloride
Other
1,2-Ethy lene
Bichloride
Acetic Acid
Acrylonitrile
Vinyl Acetate
Cumene
Vinyl Chloride
Acetaldehyde
Sec Butyl Alcohol
Other
1There were <4% of the sources in ethylene process units-gas streams associated with a low leaking
primary chemical.
2See Table 3-2 for definition of process unit type identification numbers
Figure 3-2. Categories of Sources for Further Analysis
-------
The major reason for grouping the sources into these eight categories
was to aggregate the sources into groups that have similar leak frequencies.
Note that this categorization was done for the data analysis and not for
engineering or physical reasons. The final 25 groups used for further
analyses are internally similar in:
source type
• stream service
• leak frequency by process type, and
• leak frequency by primary chemical in the line.
With these groupings, any differences that the analysis detects in the other
parameters" of interest (line pressure, line temperature, etc.)'will not be
confounded with these grouping parameters.
The division into categories was done separately for each combination
of stream service and process unit category. For this reason, a particular
chemical may be grouped in the high leaking group in one subset and the low
leaking group in another. Also the influence of other chemicals in the line
was not investigated here. As a result, it is difficult to quantify the
effect of a specific chemical.
Two additional comments should be made. First, the chemical groupings
were made according to valve data only, so the high-low primary chemical
breakdown for pump seals,, flanges, and open-ended lines (see Tables 3-7a and
3-7b) may not reflect a strict high versus low leaking classification in all
cases. Secondly, the numbers displayed in the tables in this section may
not add up to totals from other tables in previous sections due to missing
information on primary chemicals for some sources.
28
-------
TABLE 3-7a. HIGH VERSUS LOW LEAKING PRIMARY CHEMICAL GROUPS FOR HIGH LEAKING PROCESS UNITS
Hiah Leaking Chemicals (Group 5 and Group 71)
Source Type
Valves
Pump Seals
Flanges
Open Ended
Lines
Stream
Service
Gas
Light Liquid
Light Liquid
Gas
Light Liquid
Gas
Light Liquid
Number
Screened
1228
3299
126
391
578
146
798
Number
Leaking
146
147
14
18
10
13
47
Percent
Leaking
11.9
4.5
11.1
4.6
1.7
8.9
6.0
95Z Confidence
Intervals for
Percent Leakina
(10.2
(3. 1.
(6.;:,
<2.7,
(0.7,
(4.9.
(4.J.,
, 14.1)
5.3)
18.1)
7.1)
3.6)
14.6)
7.9)
Low
Number
Screened
1758
6154
243
316
1010
347
1291
Leaking Chemicals (Group 6 an
1 Group 81)
952 Confidence
Number Percent Intervals for
Leaking Leaking Percent Leakintt
22 1.2 (0
46 0.8 (0.
14 5.8 (3.
9 2.8 (1
0 0,0 (0
4 1.2 (0.
38 2,9 (2
7, 1.7)
6, 1.0)
3, 9.4)
2, 5.3)
0.4)
4, 3.1)
0, 4.1)
'See Figure 3-2 for explanation of groups.
Note: Chemical groupings were made according to valve data only, so other sources may not reflect
a strict high versus low leaking classification.
-------
TABLE 3-7b. HIGH VERSUS LOW LEAKING PRIMARY CHEMICAL GROUPS FOR ETHYLENE PROCESS UNITS
w
High Leaking Chemicals (Group 1 and
Source Type
Valves
-
Pump Seals
Flanges
Open Ended
Lines
Stream Number
Service Screened
Gas 6050
Light Liquid 3514
Light Liquid 61
Gas 566
Light Liquid 32 7
Gas 284
Light Liquid 151
Number Percent
Leaking Leaking
932 15.4
957 27.2
18 29.5
39 6.9
25 7.6
37 13.0
39 25.8
Group 31
)
95% Confidence
Intervals for
Percent LeaklnK
(14,
(26,
(19,
(4.9,
(5.0,
(9.4,
(19,
16)
29)
43)
9.2)
12)
17)
34)
Low Leaking Chemicals (Group 41)
95% Confidence
Number Number Percent Intervals for
Screened Leaking Leakina Percent Leakina
607 9 1.5 (0.67, 2.8)
15 2 13.3 (1.7, 40)
70 0 0.0 (0.0, 5.1)
63 2 3.2 (.39, 11)
'See Figure 3-2 for explanation of groups.
Note; Chemical groupings were made according to valve data only so other sources may not reflect
a strict high versus low leaking classification.
-------
EFFECT OF TYPE OF VALVE ON LEAK FREQUENCY
The first breakdown for studying the effect of valve type is block
valves versus control valves. Within these two categories, there are six
types of valves evaluated: gate, globe, plug, ball and butterfly, plus one
group called "other" which includes any valve type that does not fit into the
first five categories.
Tables 3-8 and 3-9 show the valve leak frequency data for gas and light
liquid stream service, respectively. Confidence intervals for percent leaking
are included to help distinguish those cases with a high percent leaking but
a small sample size.
Since some of the specific types of valves, particularly for
control valves, had very small sample sizes when the data is categorized, an
overall tabulation of valve types is given in Table 3-10. The smaller con-
fidence limits make differences by type more easily seen. Figures 3-3a and
3-3b show this information graphically. For block valves, gate valves have
the highest leak frequency while plug and ball valves have the lowest leak
frequency.
The totals for block and control valves over all individual types were
tested to evaluate the influence of both process unit and chemical in the
line. Categorical statistical analyses were performed on these data to
determine the significance of these classifications and their combined effects
(This method of analysis is described in Section 7.) Separate analyses were
performed for gas and light liquid stream service. A summary of these analy-
ses are given in Table 3-11.
The analysis of gas stream service does not include a variable to
distinguish primary material groups since no such group was defined for
ethylene process units with gas stream service. The analysis shows a sig-
nificant effect of chemical produced, block/control and their combined
31
-------
TABLE 3-8. PERCENT LEAKING FOR ALL TYPES OF VALVES IN GAS SERVICE
AS A FUNCTION OF PROCESS GROUP AND PRIMARY MATERIAL GROUP
High Leaking Process Units
Grouo 51 Prlraarv Chemicals
Valve
Function Type
Block gate
globe
plug
ball
butterfly
other
Total
Control gate
globe
plug
ball
butterfly
other
TOTAL
Number
Screened
978
9
34
102
17
1
1141
19
39
3
5
20
1
87
Number
Leaking
107
4
0
4
1
0
116
8
11
0
1
9
0
29
Percent
Leaking
10.9
44.4
0.0
3.9
5.9
0.0
10.2
42.1
28.1
0.0
20.0
45.0
0.0
33.3
952
Confidence
Interval
(9.1, 13)
(14, 79)
( 0, 10)
(1.1, 9.6)
(0.2, 29)
(0, 100)
(8.7, 13)
(20, 66)
(15, 45)
(0, 71)
(0.5, 72)
(23, 68)
(0. IQOi
(24,44)
Cr.
Number
Screened
1124
34
245
273
21
9
1706
15
21
2
1 .
13
0
52
oup 61 Pr1
Number
Leaking
20
1
0
0
0
0
21
0
1
0
0
0
-
1
Percent
Leaking
1.8
2.9
0.0
0.0
0.0
0.0
1.2
0.0
4.8
0.0
0.0
0,0
- .
1.9
95%
Confidence
Interval
(1.1, 2.8)
(0.1. 15)
(0, 1.5)
(0, 1.3)
(0, 16)
(0, 34)
(0.8, 1.9)
(0, 22)
(0.1. 24)
(0, 84)
(0, 100)
(0, 25)
-
(0.1, 10)
Ethylene Process Units
Number
Screened
4495
73
39
834
102
261
5804
24
137
5
8
56
16
246
Number
Leaking
823
10
0
14
8
16
871
7
24
0
2
26
2
61
Percent
Leaking
18.3
13.7
0.0
1.7
7.8
6.1
15.0
29.2
17.5
0
25.0
46.4
12.5
24.8
952
Confidence
Interval
(17, 19)
(6-7, 24)
(0, 9.0)
(0.9, 2.9)
(3.3, 15)
(3.7, 10)
(14, 16)
(13, 51)
(11, 24)
(0, 52)
(3.2, 65)
(34, 62)
(1.6, 38)
(19, 30)
See Figure 3-2 for explanation of groups.
-------
TABLE 3-9. PERCENT LEAKING FOR ALL TYPES OF VALVES WITH LIGHT LIQUID STREAM
SERVICE BY PROCESS GROUP AND PRIMARY MATERIAL GROUP
High Leaking Process Units
Group 7 Primary Chemicals
Valve
Type
Clock
Gate
Globe
Plug
Ball
Butterfly
Other
TOTAL
Control
Gate
Globe
Plug
Ball
Butterfly
Other
TOTAL
Number
Screened
2330
37
470
255
63
5
3160
58
40
26
2
11
2
138
Number
Leaking
122
4
2
0
0
0
128
11
6
1
0
1
0
19
%
Leaking
5.2
10.8
0.4
0.0
0.0
0.0
4.1
19.0
15.0
3.8
0.0
9.1
0.0
13.8
95X
Confidence
Interval
(4.3, 6.2)
(3.0, J5)
(0.1, 1.5)
(0, 1.5)
(0, 5.7)
(0. 52)
(3.4. 4.9)
(9.9. 3D
{5.7, 30)
(0.1. 20)
(0, 85)
(0.2, 41)
(0. 85)
(8.4. 20)
Group 8 Primary Chemicals
Numb jr
Screened
332J
187
1030
1263
13
33
587 J
65
121
5)
25
12
5
281
Number
Leaking
38
1
0
0
0
0
39
2
5
0
0
0
0
7
X
Leaking
1.4
0.5
0.0
o.o
o.o
0.0
0.66
3.1
4.1
0.0
0.0
0.0
0.0
2.5
95Z
Confidence
Interval
(0.8, 1.6)
(0.01,2.9)
(0. .36)
(0, .29)
(0, 11)
(0, 9.25)
(0.5, ,91)
(0.4, 11)
(1.3, 9.5) ,
(0, 6.7)
(0. 14)
(0. 26)
(0. 52)
(1.0, 5.0)
-------
TABLE 3-9. (continued)
OJ
Ethylene Procese Units
Group 3 Primary Chemicals
Valve
Type
Block
Gate
Globe
Plug
Ball
Butterfly
Other
TOTAL
Control
Gate
Globe
Plug
Ball
Butterfly
Other
TOTAL
Number
Screened
3125
45
3
19
A
94
3320
20
162
2
3
7
0
194
Number
Leaking
877
2
0
4
2
15
900
9
46
0
0
2
55
%
Leaking
28.
. 4.
0.
18.
50.
16.
27.
45.
28.
0.
0.
28.
__
28.
1
4
0
2
0
0
1
0
4
0
0
6
-
4
. 95X
Confidence
Interval
(26,
(0.5
(0,
(2.3
(6.8
(9.2
(25,
(23,
(22,
(0,
(0,
(3.7
—
(23,
30)
, 15)
71)
, 20)
, 93)
, 25)
29)
68)
3f>
84)
71)
, 71)
-
35)
Group 4 Primary Chemicals
HI mibc r
Screened
529
15
2
8
1
• JO
575
7
20
5
0
0
0
32
Number
Leaking
7
0
0
0
0
0
7
0
2
0
0
2
2
Leaking
1.3
0.0
0.0
0.0
0.0
0.0
1.2
0.0
10.0
0.0
6.2
95%
Confidence
Interval
(0.
(0,
(0,
(0,
(0,
-------
TABLE 3-10. LEAK FREQUENCIES FOR ALL TYPES OF VALVES FOR GAS AND
LIGHT LIQUID STREAM SERVICE
UJ
Ln
Gaa Service
Type
Block
Gate
Globe
Plug
Ball
Butterfly
Other
TOTAL
Control
Gate
Globe
Plug
Ball
Butterfly
Other
TOTAL
Number
Screened
6976
145
440
1272
160
275
9268
61
207
10
15
91
17
401
Number
Leaking
952
15
0
18
9
16
1010
15
36
0
4
35
3
93
Percent
Leaking
13.7
10.3
0.0
1.4
5.6
5.8
10.9
24.6
17.4
0.0
26.7
38.5
17.6
23.2
95X
Confidence
Limits
(13. 15)
(5.9, 17)
(0, 0.8)
(0.7, 2.5)
(2.6, 10)
(3.5, 9.2)
(10, 11)
(14, 37)
(13, 24)
(0, 31)
(7.8, 55)
(28, 49)
(3.8, 43)
(19, 28)
Nuiaber
Screened
11017
755
:'.-'. 7 9
2732
157
378
17518
182
417
91
33
34
25
782
Liaht Liquid Service
Number
Leaking
1059
8
2
4
2
17
1092
22
61
3
1
3
1
91
Percent .
Leaking
9.6
1.1
0.1
0.2
1.3
4.5
6.2
12.1
14.6
3.3
3.0
8.8
4.0
11.6
95X
Confidence
Limits
(8.6, 9.8)
(0.5, 2.1)
(0.01, 0.3)
(0.04, 0.4)
(0.15. 4.4)
(2,7, 7.8)
(5.9, 6.6)
(7.8, 18)
(12, 19)
(0.7, 9.3)
(0.1, 16)
(0.7, 20)
(0, 20)
(10, 15)
-------
VALVES IN GAS SERVICE
I_1KJ
p
e
r
c 40
e
n
t
30
1
e
a 20
k
i
n
g 10
-
-
—
_
-
_
-
_
-
r —
i
— i
" LJ3-J
-
-
-
i
L_
i
i
_i
*~
.
— i ^.
. J
i — i
i
i _
t
1 — • i_
i
L '
, rf_, rm ' '
GA
TE
•*!
PLUG
B-F
GLOBE BALL
i __
RI nru'
"LY
GA
TE
r—
i —
— i
i
_i
* r
PL
— i
t
— ,
i
i — i
i
i — i
i
t
i— — i
i —
i
_j
i »
UG
B-F
"LY
OTHER GLOBE BALL OTHER
rnMTPni
Figure 3-3a. Percent Leaking with 95 Percent Confidence Intervals for Each Valve Type with
Gas Service
-------
VALVES IN LIQUID SERVICE
-
-
:S
—
-
-
: 3
i
GATE
c
• 'Hi I"T
PLUG
i
l
i
7-1 i —
B-F
GLOBE BALL
RI nr.K
r—
— , t
t_
i
_i
'LY
i—
i
-i "-
i —
~i
i
i t_
— i
i —
i
_j
r~
—i
r
L_
— i
1 i
_i
i -
GA
TE
PL
UG
r~
— i
i
i
L_
— > i —
i
t
— i
, — » '
B-F
"LY
— i
i
i
OTHER GLOBE BALL OTHER
:_ - ^- — rnMToni
p
e
r
c
e
n
t
a
k
i
n
15
Figure 3-3b. Percent Leaking with 95 Percent Confidence Intervals for Each Type of Valve with
Light Liquid Service
-------
effect on the leak frequency. For valves in gas stream service, Table 3-8
shows control valves with a higher percent leaking than block valves for each
group. The significant combined effect indicates that the difference between
block and control valves is significantly greater in the high leaking process
units than in the ethylene process units.
The second analysis summarized in Table 3-11 is for light liquid
service. The variables used here are the same as for gas service with the
addition of a category by primary material in the line and all of the two-way
combined effects. All of the main effects and two of the combined effects
are highly significant. The combined effect of primary material in the line
and block/control is also statistically significant. Table 3-9 shows the
direction of these differences. It can be seen from this table that the per-
centage of sources leaking for control valves from high leaking process units
with high leaking primary materials is about three times that of block valves
in the same group. For the low leaking primary material group, it was about
four times. Ethylene process units with high leaking primary materials in
the line had similar leak frequencies between block and control groups.
Figure 3-4a and Figure 3-4b provide a graphical display of the differ-
ences .in leak frequency between the block and control valves for each process
and primary material category.
Since the analysis found a significant different in leak frequency
between block and control valves, the comparison by specific type of valve
was done for each of these valve classifications. Figures 3-3a and 3-3b
show leak frequencies for each valve type with 95 percent confidence intervals
for the leak frequencies. The larger number of block valves tested makes
this group the easier one to examine for differences by valve type. For
both gas and light liquid service, gate valves have the highest leak fre-
quency, and plug and ball valves have the lowest leak frequency. Globe and
butterfly valves in light liquid service also have low leak frequencies,
The^e last two types of valves have comparatively wide confidence intervals
for gas service because of the small number of valves of those types found.
38
-------
TABLE 3-11. RESULTS OF CATEGORICAL ANALYSIS ON VALVES1
Source
Gas Stream Service"
Chemical Process
Block/Control
Combined Effects
Light Liquid Stream Service
Chemical Process
Primary Material
Block/Control
Combined Effects
Process by Material
Process by Block/Control
Material by Block/Control
Chi-Square
Statistic
655.40
19.44
13.81
207.1
60.1
25.4
64.9
15.7
4.4
Probability of
No Effect
<0.01
<0.01
<0.01
<0.01
-------
VALVES IN GAS SERVICE
O
45
40
P
e 35
r
• 30
n
t 25
1 20
i
n 10
9
5
i
[i
BLOCK CONTROL
Group 5
P rimary Ch emic als
BLOCK
CONTROL
Group 6
Primary Chemicals
BLOCK CONTROL
Group 1
Primary Chemicals
Figure 3-4a.
'High Leaking Processes ' — ' ' ' Ethylene Processes
Percent Leaking for Block Versus Control Valves in Gas Service by Primary
Material Group and Process Unit Type
Note: See Figure 3-2 for explanation of primary material groups.
-------
r
VALVES IN LIQUID SERVICE
-TO
40
P
e 35
r
c 30
e
n ^
t
I 20
e
a 15
k
i
n '
9
5
-
-
-
-
-
-
_.
"
_,
i —
-
__
- i
-
t_
-
f nn
i —
T1
T
ul_i '
-. «-
I
_j
- I
r~
— i
1
(
_t
j ,
t *""
i_i_i
— i
I
— j
BLOCK CONTROL BLOCK CONTROL BLOCK CONTROL BLOCK CONTROL
Group 7 Group 8
Primary Chemicals Primary Chemicals
Group 3
Primary Chemicals
Group 4
Primary Chemicals
High Leaking Processes
Ethylene Processes
Figure 3-4b. Percent Leaking for Block Versus Control Valves in Light Liquid Service
by Primary Material Group and Process Unit Type
Note: See Figure 3-2 for explanation of primary material groups.
-------
For gas service, they appear to fall in the middle range between the high
leaking gate valves and low leaking plug and ball valves. A comparison of
types of control valves is more difficult because of the small sample sizes.
The actual percent leaking for gate valves is higher than that of plug and
ball valves, but there are overlapping confidence intervals. In the group
of control valves in light liquid service, globe valves show a significantly
higher leak rate than plug valves from the same group.
42
-------
LEAK FREQUENCY FOR PUMP SEAL CLASSIFICATIONS
Pump seals comprise a much smaller group of sources than valves. For
this reason, the groupings by Process uti.it type and primary materials in the
line are not reported in this section. When these subcategorizations were
examined, the small sample size in these categories resulted in such large
confidence limits that no statistical differences could be seen. Combining
the categories did not effect any of the trends observed.
The primary classifications for pump seals are on-line versus off-line,
single versus double seals, mechanical versus packed seals, and location of
the emission point. Table 3-12 gives the number of pump seals screened, the
number leaking, the percentage leaking and appropriate 95% confidence inter-
vals for these classifications of pump seals. On-line and off-line single
mechanical seals with emission point at seal are the two largest groups.
A chi-square test (see Section 7 for details) was performed to determine
if there was a statistically significant difference in the leak frequencies
between on-line and off-line pump seals when the emission point was at the
seal. The outcome is below:
NO LEAK
Number %
LEAK
Number %
On-line
Off-line
Total
271 86.9
232 95.1
41 13.1
12 4.9
503 90.5
Chi-Square _ -,Q .
Statistic
53
9.5
TOTAL
312
244
556
This test indicates that there is a significant difference between
on-line and off-line pump seals, with the leak frequency for off-line pumps
43
-------
TABLE 3-12. LEAK FREQUENCIES FOR PUMP SEALS IN LIGHT LIQUID SERVICE
On Line/
Off Line
On-Line
Off-Line
Mechanical/ Single/
Packed Double
Mechanical Single
Mechanical Double
Packed Single
Mechanical Single
Mechanical Double
Packed Single
Emission
Point
Seal
Vent
Other
Seal
Vent
Other
Seal
Vent
Other
TOTAL AT
THE SEAL
Seal
Vent
Other
Seal
Vent
Other
Seal
Vent
Other
TOTAL AT
THE SEAL
Numb e f
Screened
215
24
30
92
3
3
5
0
1 '
312
139
9
17
86
0
1
19
2
0
244
Number
Leaking
28
0
2
13
1
1
0
-
0
41
9
0
0
3
-
0
0
0
_
12
Percent
Leaking
13.0
0.0
6.7
14.1
33.3
33.3
o.o
_
0.0
13.1
6.5
0.0
0.0
3.5
-
0.0
0.0
0.0
-
4.9
95Z Confidence
Interval
(9, 20)
(0, 14)
(0.8, 22)
(7.7, 23)
(0.8, 91)
, (0.8, 91)
(0, 52)
-
(0. 100)
(9.0, 17)
(3.0, 12)
(0, 34)
(0, 20)
(0.7, 9.9)
-
(0, 100)
(0, 18)
(0, 84)
_
(2.6, 8.6)
-------
about one-third that of on-line pumps.
A Chi-square test was also used to compare the leak frequency for the
single mechanical pump seals to double mechanical pump seals. Separate tests
were performed for on-line and off-line seals. Only data with the emission
point at the seal was considered. Single packed pump seals had no leaks in
either case and so could not be included in the test. The following table
describes this test:
On-Line Pump
Seals
NOT LEAKING
Numb er %
266
86.6
LEAK
Number %
Single
Mechanical
Double
Mechanical
187 87.0
79 85.9
28 13.0
13 14.1
Chi-Square Statistic = 0.07
p >0.10
215
98
41 13.4 307
Off-Line Pump
Seals
NOT LEAKING
Number %
LEAKING
Number %
Single
Mechanical
Double
Mechanical
130 93.5
83 96.5
9 6.5
3 3.5
139
87
240. 94.7 12 5.3 252
Chi-Square Statistic = 0.94 p >0.10
-------
The leak frequency for single mechanical versus double mechanical was
not significantly different for either the on-line or the off-line data.
Figure 3-5 shows this same information graphically. Note that the presence
or type of barrier fluids was generally not recorded for this data. This
may have been a factor in the lack of a significant difference between single
mechanical and double mechanical pump seals.
46
-------
PUMP SEALS
45
p
e
35
r
c
e 30
n
t 25
1 20
e
a
k '5
n 10
9
5
A-J-
A-L
SING-MECH
SING-PACK
DOUB-MECH
ON-LINE -
DOUB-MECH
SING-MECH SING-PACK
OFF-LINE
Figure 3-5. Percent Leaking with 95 Percent Confidence Intervals for Each Type
of Pump Seal with Light Liquid Service
-------
THE EFFECT OF LINE TEMPERATURE AND LINE PRESSURE
The effects of line temperature and line pressure are examined in paral-
lel in this section. The effects- of these two variables are evaluated for
the four major source types: flanges, open-ended lines, valves, and pump
seals. It was found, that different levels of temperature and pressure are
present in the 24 units studied depending both on the type of chemical produced
and also on the primary material in the line. The data are grouped by these
variables as they have been defined earlier in this report. Appendix B
contains summary statistics in tabular form for line pressure and line temper-
ature for each of the groups.
Since valves constituted the largest group by source type, the effects
of line temperature and line temperature on leak frequency for valves could
be studied in the greatest detail. Categorical statistical analysis,
described in Section 7, was used to determine the significance on leak fre-
quency of line temperature and line pressure and their combined effects
(interaction). This method of analysis is biased by empty cells (any tempera-
ture and pressure categories with no leaks). As a result, the only groups
to be studied for possible combined effects were valves with high leaking
primary chemical groups (see Figure 3-2 for an explanation of groups). The
results of this analysis are given in Table 3-13.
The results of the categorical analysis show that for ethylene process
units with valves in gas service, both line temperature and line pressure,
and also their combined effect, were significant. For the valves in light
liquid service, pressure and the combined effect of temperature and pressure
were significant. Both of the groups from the high leaking process units
showed only pressure to have a significant effect on leak frequency. Tables
3-14 to 3-17 give the data used in this analysis. Figures 3-6 to 3-9 graph-
ically show the results of these analyses. Figure 3-6 provides a good
example of significant combined effects (interaction) of line temperature
and line pressure. It shows that the effects of increased pressure on the
percent leaking is not the same for all temperature groups. If there was no
48
-------
significant combined effect, the lines would be parallel.
Tables 3-18 and 3-19 show the effects of line temperature and line
pressure on the valve for primary material groups 4, 6 and 8. The categories
of temperature and pressure were chosen to agree with those in Appendix B
It appears that pressure may .have an effect on each of these three groups
(group 4, group 6 and group 8). Temperature appears to have an effect on
valves from group 4.
In summary, higher levels of pressure appear to result in higher leak
frequency in almost every instance. For example, valves from Primary
Material Group 1 have a 4.1 percent leaking in the "less than 25 psig"
pressure group and 25.8 percent leaking in the "greater than 200 psig"
pressure group. In those cases where this is not seen, it may be due to
the smaller sample sizes. Temperature appears to be significant in only a
few cases. In those cases, it was the middle range of line temperature
rather than the extremes that was associated with higher leak frequency.
Valves in gas service from ethylene process units had the greatest percent
leaking (16.2%) at temperatures between 0°F and 49°F. The combined effects
of line temperature and line pressure could only be studied for valves. It
had a significant effect for the ethylene process units only. Higher leak
frequencies for high pressure and middle level temperature were found.
Figures 3-6 to 3-8 graphically show the effect of the interaction. The
significant combined effect is apparent in the fact that the lines for
levels of line temperature are not parallel.
49
-------
TABLE 3-13. RESULTS OF ANALYSIS OF THE EFFECTS OF LINE
TEMPERATURE AND LINE PRESSURE ON LEAK FREQUENCY
FOR VALVES
Process
- Unit
Group
Ethylene
Processes
High
Leaking
Processes
Primary
Material
Group 1
Group 1
Group 3
Group 5
Group 7
Stream
Service
Gas
Light
Liquid
Gas
Light
Liquid.
Source
Temperature
Pressure
Combined effects
Temperature
Pressure
Combined effects
Temperature
Pressure
Combined effects
Temperature
Pressure
Combined effects
Degrees
of
Freedom2
3
3
9
2
2
4
2
2
4
2
2
4
Chi-
Square
9.1
268.6
42.4
3.4
67.2
25.7
3.2
7.6
3.1
3.4
13.9
3.1
Signifi-
cance
A
AA
AA
n.s.
AA
AA
n.s .
A
n.s.
n.s.
A
n.s.
1See Figure 3-2 for explanation of groups.
2The total degrees of freedom for gas service in the ethylene process units
is higher than the total for the other groups because, in this group, four
levels of both temperature and pressure could be used without producing any
empty cells in the analysis.
•s
*probability of no significant effect is less than 0.05
**probability of no significant effect is less than 0.01
n.s.-no significant effect
50
-------
TABLE 3-14. LINE TEMPERATURE AND LINE PRESSURE AND THEIR COMBINED EFFECTS
ON VALVES IN GAS SERVICE WITHIN ETHYLENE PROCESS UNITS
Temper atu
Pressure
-15-25
25-99
100-199
200-1050
TOTAL
re <°F>
Number
Screened
348
180
55
415
998
-267-0
Number
Leaking
1-8
30
12
80
140
Percent
Leaking
5.2
16.7
21.8
19.3
14.0
Hunter
Screened
506
449
104
392
1451
0-49
Number
Leaking
29
54
8
144
235
Percent
Leaking
5.7
12.0
7.7
36.7
16.2
Number
Screened
542
881
332
491
2246
50-99
Number
Leaking
14
96
71
146
327
Percent
Leaking
2.6
10,9
21.4
29.7
14.6
Number
Screened
244
276
350
714
1584
100-1570
Number
Leaking
7
IB
55
149
229
Percent
Leaking
2.9
6.5
15.7
20.9
14.5
Number
Screened
1640
1786
841
2012
6279
TOTAL
Number
Leakina
68
198
146
519
931
Percent
Leaking
4.1
11.1
17.4
25.8
14.8
-------
TABLE 3-15. LINE TEMPERATURE AND LINE PRESSURE AND THEIR COMBINED EFFECTS
ON VALVES FROM GROUP 5*
Temper at ur
Pressure
(pslg)
-15-99
100-199
200-1050
TOTAL
e <°F)
Number
Screened
120
16
73
209
-267-99
Number
Leaking
8
4
10
22
Percent
Leaking
6.7
25.0
13.7
10.5
Number
Screened
141 •
83
91
315
100-149
Number
Leaking
8
7
14
29
Percent
Leaking
5.7
8.4
15.4
9.2
Number
Screened
187
118
236
541
150-1570
Number
Leaking
22
15
45
82
Percent
Leaking
11.8
12.7
19.1
15.2
Number
Screened
448
217
400
1065
TOTAL
Number
Leaking
38
26
69
133
Percent
Leaking
8.5
12.0
17.2
12.5
*See Figure 3-2 for explanation of groups. Group 5 Is the high leaking primary chemical group from high leaking processes in gas service.
-------
TABLE 3-16. LINE TEMPERATURE AND LINE PRESSURE AND THEIR COMBINED
EFFECTS ON VALVES FROM GROUP 3*
Temperatur
Pressure
(psig)
-15-99
100-199
200-1050
TOTAL
e {op)
Number
Screened
683
65
1325
2073
-267-49
Number
Leaking
106
6
360
647
Percent
Leaking
15.5
9.2
2?. 2
31.2
Number
Screened
56
282
490
828
50-99
Number
Leaking
1
90
196
287
Percent
Leaking
1.8
31.9
40.0
34.7
Number
Screened
104
108
396
608
100-1570
Number
Leaking
16
14
163
198
Percent
Leaking
15.4
13.0
42.4
32.6
Number
Screened
843
455
2211
3509
TOTAL
Number
Leaking
123
110
724
957
Percent
Leaking
14.6
24.2
32.7
27.3
w
*See Figure 3-2 for explanation of groups. Group 3 is the high leaking chemical group from light liquid service from ethylene processes.
-------
TABLE 3-17. LINE TEMPERATURE AND LINE PRESSURE AND THEIR COMBINED
EFFECTS ON VALVES FROM GROUP 7*
Temper a tur
Pressure
(palg)
-15-99
100-199
200-1050
TOTAL
» CF)
Number
Screened
795
216
143
1154
-267-99
Hunter
Leaking
14
12
19
45
Percent
Leaking
1.8
5.6
13.3
3.9
Number
Screened
322
234
263
8.9
100-149
Number
Leaking
4
12
20
36
Percent
Leaking
1.2
5.1
7.6
4.4
Number
Screened
663
245
390
1298
150-1510
Number
Leaking
24
12
27
63
Percent
Leaking
3.6
4.9
6.9
4.8
Number
Screened
1780
695
796
3271
TOTAL
Number
Leaking
42
36
66
144
•
Percent
Leaking
2.4
5.2
8.3
4.4
*See Figure 3-2 for explanation of groups. Group 7 Is the high leaking primary chemical group for high leaking processes in light liquid service.
-------
p
E
R
C
E
N
T
L
E
A
K
I
N
G
40
32-
24-
16-
8
0
COMBINED EFFECTS OF TEMPERATURE AND PRESSURE
ETHYLENE PROCESS UNITS — GAS SERVICE
-15 to 25
26 to 99
100 to 199
200+
~ TEMPC<0:>
••' TEMPC0-49D
~' TEMPC50-995
— TEMPO 100)
PRESSURE
Figure 3-6. Combined Effects of Line Temperature and Line Pressure on Percent
Leaking for Valves in Gas Service Within Ethylene Process Units.
-------
Ui
Os
R
C
E
N
T
L
E
A
K
I
N
G
30
24-
8"
12-
6-
0
COMBINED EFFECTS OF TEMPERATURE AND
PRESSURE ON VALVES IN GROUP 5*
/ x
S
V •
. *N
-15 to 90
TEMPO 00-149)
TEMP01495
1
100 to 199
PRESSURE:
1
200+
Figure 3-7. Combined Effects of Line Temperature and Line Pressure on
Percent Leaking for Valves from Group 5*.
*See Figure 3-2 for explanation of groups.
-------
COMBINED EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE
ON VALVES IN GROUP 3*
Ui
ss
11
100-199
>199
TEMP
TEMP CG0-99D
TEMP 099)
Figure 3-8. Combined Effects of Line Pressure on Percent Leaking for Valves in Group 3.*
*See Figure 3-2 for explanation of groups.
-------
00
30
r
c
e
n
t
I
e
o
k
I
n
9
20
10
1
CED
nri
•15 to 49 100 to 249 -15 to 49 100 to 249
50 to 99 250 «• 50 to 99 250 •*-
i ETHYLENE PROCESS UNITS " HIGH LEAKING PROCESS UNITS [
GROUP 4 Primary Chemicals GROUP 8 Primary Chemicals
Figure 3-9. The Effect of Line Pressure on Percent Leaking With 95 Percent
Confidence Intervals for Valves from Group 4 and Group 8.*
*See figure 3-2 for explanation of groups
-------
TABLE 3-18. EFFECT OF LINE TEMPERATURE AND LINE PRESSURE ON
VALVES FROM GROUP 6*
Pressure
-15-49
50-99
100-249
250-1050
Total
Temperature
-267-49
50-99
100-199
200-1570
Total
Number
Screened
1267
141
188
162
1758
45
335
823
555
1758
Number
Leaking
7
2
13
0
22
0
4
2
16
22
Percent
Leaking
0.5
1.4
6.9
0.0
1.2
0.0
1.2
0.2
2.9
1.2
95%
Confidence
Limits
(0.2, 1.1)
(0.2, 5.1)
(3.7, 12)
(0, 2.3)
(0.78, 1.9)
(0, 7.9)
(0.3. .3. 2)
(0.03, 0.9)
(1.7, 4;?>
(0.8, 1.9>
*See Figure 3-2 for explanation of groups
59
-------
TABLE 3-19. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON VALVES
FROM GROUP 4 AND GROUP 8 BY PROCESS UNIT GROUP1
Group 41 PRIMARY CHEMICALS
Number Number
Pressure (psig) Screened Leaking
-15-49
50-99
100-249
250-1050
O Total
Temperature (°F)
-267-49
50-99
100-199
200-1570
Total
173 0
215 4
181 4
38 1
607 9
•
29 0
127 2
341 '6
110 1
607 9
95%
Number Confidence
Leaking Intervalu
0.0 (0, 2.1) ;.
1.9 (0.5, 4.8)
2.2 (0.8, 5,6)
' 2.6 (0.1, 14)
1.5 (0.7, 2.8)
0.0 (0, 12)
1.6 (0.2, 5.5)
1.8 (0.6, 3.8.1
0.9 (0.0, 5.0)
1<5 (0.7, 2.8)
Group 81 PRIMARY CHEMICALS
Number
Screened
2432
1567
1916
205
6120
96
1912
2583
1563
6154
Number
Leaking
16
5
18
7
46
2
11
16
17
46
Number
Leaking
0.7
0.3
0.9
3.4
0,8
2.1
0.6
0.6
1.1
0.8
95%
Confidence
Intervals
(0.4,
(0.1,
(0.6,
(1.4,
(0.5,
(0.3,
(0.3,
(0.4,
(0.6,
(0.5,
1.1)
0.7)
1.5)
7.1)
1.0)
7.3)
1.0)
1.0)
1.7)
1.0)
*See Figure 3-2 for explanation of groups
-------
EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON PUMP SEALS, FLANGES, AND
OPEN-ENDED LINES
There was not enough data available to study the possible combined
effect of line temperature and line pressure for the remaining source types,
and so the effects of these two variables were examined separately. Catego-
ries were used that would conform to earlier tables and at the same time
provide an approximately even distribution of sources screened.
Table 3-20 shows this information for pump seals for both groups of
process units and all primary materials. The 95 percent confidence intervals
indicate that no significant effects of temperature or pressure can be seen.
If the overall number screened was increased, the size of the confidence
intervals would be decreased and it is possible that some significant dif-
ferences might then be seen. Figure 3-10 shows the percent leaking with
95 percent confidence intervals as a function of pressure for this source
type.
Tables 3-21 and 3-22 give the leak frequencies by line temperature and
line pressure for flanges in gas and light liquid services, respectively.
Increasing levels of line pressure result in increased leak frequency. The
effect is most clear for gas service streams. These data are presented by
process unit group since for the light liquid service, there appears to be
some differences between these two groups.
Tables 3-23 and 3-24 show the leak frequency of open-ended lines by
line temperature and line pressure. Ethylene process units services in gas
service show a higher leak frequency at the highest level of pressure.
Otherwise the gas service show overlapping confidence intervals. Open-ended
lines in light liquid service within ethylene process units show an increased
leak frequency at higher pressure levels and also a higher frequency at the
upper two pressure levels when compared to the high leaking process units.
The leak frequency from the high leaking process units does not appear to
61
-------
be affected by line temperature or line pressure. Line temperature does
not appear to have an effect on open-ended lines within ethylene process
units either.
62
-------
TABLE 3-20. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON PUMP
SEALS WITH LIGHT LIQUID SERVICE
Pressure
(psig)
-15 - 49
50 - 99
100 - 249
250 - 1050
Number
Screened
146
115
116
65
Number
Leaking
10
19
11
12
Percent
-Leaking
6.8
16.5
9.5
18.5
95% Confidence
Intervals
(3.3, 12)
CIO, 25)
(4.9, 16)
(9.9, 30)
Total 442
Temperature (°F)
-267 - 49 53
50 - 99 148
100 - 199 146
200 - 1570 100
Total 447
52
12
14
13
_i
48
11.8
22.6
9.5
8.9
0.9
10.7
(9.0, 15)
(12, 36)
(5.2, 14)
(4.8, 14)
(4.2, 16)
(8.0, 14)
63
-------
PUMP SEALS — LIGHT LIQUID
35
P 30
e
r
c 25
n
t 20
e ' ">
o
k
i '
n
9 5
ii
u
15 to 49 50 to 99 100 to 249
PRESSURE
250 •*•
Figure 3-10. The Effect of Line Pressure on Percent Leaking with 95 Percent Confidence Intervals
on Pump Seals in Light Liquid Service
-------
TABLE 3-21. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON FLANGES
IN GAS SERVICE BY PROCESS UNIT GROUP
1 Ethylene Process Units
Number
Pressure (pain) Screened
-15-49
50-99
100-249
250-1050
o\
Ln
TOTAL
Temperature (°F)
-267-49
50-99
100-199
200-1570
210
102
136
182
630
129
335
155
15
Number
Leaking
3
4
8
24
39
17
15
5
2
Percent
Leaking
1.4
3.9
5,9
13.2
6.2
13.2
4.5
3.2
13.3
95%
Confidence
Intervals
(0.3,
(1.1,
(2.5,
(8.7,
(4.4,
(7.8,
(2.6,
(1-0,
(1.7,
4.3)
9.6)
11)
19)
8.0)
•
20)
7.8)
7.4)
41)
Number
Screened
301
76
117
217
711
16
99
268
321
High Leaking Process Units
Number
Leaking
6
2
2
17
27
0
1
4
22
Percent
Leaking
2.0
2.6
1.7
7.8
3,8
o.o
1,0
1.5
6.8
95*
Confidence
Intervals
(0.
to.
(0.
(4,
(2.
(0,
(0.
(0.
(4.
7, 4.3)
3, 9.2)
2, 6.1)
8, 13)
5, 5.5)
21)
5.5)
4, 3.7)
4, 10)
TOTAL
634
39
6.2
(4.4, 8.0)
704
27
3.8
(2.5, 5.5)
-------
TABLE 3-22. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON FLANGES
IN LIGHT LIQUID SERVICE BY PROCESS UNIT GROUP
Number
Pressure (psig) Screened
-15-49
50-99
100-249
250-1050
TOTAL
Temperature (°'F)
-267-49
50-99
100-199
200-1570
TOTAL
70
52
74
200
396
133
141
121
9
404
Ethylene
Number
Leaking
1
0
5
19
25.
13
10
2
0
25
Process Units
Percent
Leaking
1.4
0,0
6.8
9.5
6.31
9.8
7.1
1.7
o.o-
6.2
95%
Confidence Huriber
Intervals Screened
(0,
, (0,
(2.
(5.
(4.
(5.
(3.
CO.
(0,
(4.
7.7) !i83
6.8) 364
2, 15) 372
8, 14) 27<<
1, 9.1) 1596
3, 16) 2£
5, 13) 493
2, 5.8) 638
34) 452
1, 9.1) 1607
High Leaking Process
Number
Leaking
0
0
2
8
10
' 0
2
6
2
10
Percent
Leaking
0.0
0.0
0.5
2.9
0.6
0.0
0.4
0.9
0.4
0.6
Units
95%
Confidence
Intervals
(0,
CO,
CO.
(1-
(0.
CO,
Co.
(0.
(o.
(0.
0.7)
1.1)
1, 2.0)
3, 5.7)
3, 1.2)
13)
1, 1.4)
3, 2.1)
1, 1.6)
3, l.D
-------
TABLE 3-23. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON OPEN ENDED
LINES IN GAS SERVICE BY PROCESS UNIT GROUP
Number
Pressure (psig) Screened
-15-49
50-99
100-249
250-1050
TOTAL
Temperature < °F)
-267-49
50-99
100-199
200-1570
TOTAL
160
53
35
55
303
173
96
31
5
305
Ethylene
Number
Leaking
6
7
3
20
36
16
17
4
0
37
Process Units
Percent
Leaking
3.8
.13.2
8.6
36.4
11.9
9.2
17.7
12.9
0.0
12.1
95%
Confidence
Intervals
(1.
(5.
(1.
(24
(8.
(5.
(11
(3-
CO,
(8.
4, 8.0)
5, 25)
8, 23)
, 50)
6, 16)
5, 15)
,0, 27)
6, 30)
52)
6, 16)
Number
Screened
272
59
74
51
456
13
83
225
134
455
High
Number
Leaking
5
4
1
4
14
0
8
5
1
14
Leaking Process Units
Percent
Leaking
1.8
6.8
1.3
7.8
3.1
0.0
9.6
2.2
0.8
3.5
95%
Confidence
Intervals
CO
(i
(0
(2
(1
(0
(4
(0
(0
(1
.6, 4.3)
.9, 16)
, 7.3)
.2, 19)
.7, 5.8)
, 25)
.2, 18)
.7, 5.1)
, 4.1)
.7, 5.8)
-------
TABLE 3-24. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON OPEN ENDED
LINES IN LIGHT LIQUID SERVICE BY PROCESS UNIT GROUP
CO
Ethylene Process Units
Pressure (psig)
-15-49
50-99
100-249
250-1050
TOTAL
Temperature ( F)
-267-49
50-99
100-199
200-1570
TOTAL
Number
Screened
30
48
38
98
214
75
56
62
21
214
Number
Leaking
0
2
7
32
41
15
13
11
2
41
Percent
Leaking
0.0
4.2
18.4
32.6
19.2
20.0
23.2
17.7
9.5
19.2
95%
Confidence
Intervals
(0,
(0.5
(7.7
(24,
(15,
(12,
(13,
(9.2
(1.2
(15,
12)
, I*)
, 34)
«)
26)
31)
36)
, 30)
, 30)
26)
Number
Screened
813
479
474
125
1891
56
668
685
^88
1897
High Leaking Process Units
Number
Leaking
27
22
14
3
66
0
28
24
14
66
Percent
Leaking
3
4
2
2
3
0
4
3
2
3
.3
.6
.9
.4
.5
.0
.2
.5
.9
.5
9^2
J Jn
Confidence
Intervals
(2.2,
(2.9,
(1.7,
(0.5,
(2.7.
(0, 6.
(2.8,
(2.4,
(1.6,
(2.7,
4.7)
6.8)
4.9)
6.9)
4.4)
4)
5.9)
5.2)
4.8)
4.4)
-------
EFFECT OF AMBIENT TEMPERATURE ON LEAK FREQUENCY
This section evaluates the effects on leak frequency of the ambient
temperature. The ambient temperature was measured at the same time that the
source was screened. Ambient temperature was measured as a continuous
variable, but to evaluate its effect on leak frequency, it was grouped as
less than 70 F or greater than or equal to 70°F. Appendix C contains summary
statistics for this variable.
Statistical tests were performed for each primary material group to
determine if there was a significant difference in leak frequencies between
the two classifications of ambient temperature.
Table 3-25 gives a summary of the effects of ambient temperature on
leak frequencies of sources. In those cases where the percent leaking was
not affected by the primary material in the line or the type of process unit
or both, groups were combined. Those groups that did show a significant
effect of ambient temperature are noted with asterisks. In one case (open-
ended lines in gas service from high leaking process units) a significant
difference in leak frequencies was seen when the primary material groups were
combined (but the differences were not significant when they were separated) .
Overall, ten of the 25 groupings of sources showed a statistically
significant effect of ambient temperature on the leak frequency. Four of
the seven comparisons for valves were significant. Generally higher leak
frequencies were associated with the high ambient temperature classification.
Differences in leak frequencies between the two ambient temperature categories
range from three percent leaking to 14 percent leaking.
69
-------
TABLE 3-25. SUMMARY OF THE EFFECTS OF AMBIENT TEMPERATURE ON PERCENT LEAKING
Source
Type
Valves
Pump
Seals
Flanges
Open Ended
Lines
Stream
Service
Gas
Light
Liquid
Light
Liquid
Gas
Light
Liquid
Gas
Light
Liquid
Process
Group
E thy I en e
High
Leaking
Ethylene
High
Leaking
Both
Both
Both
Ethylene
High
Leaking
Both
Primary
Material Ambient
Group Temperature
Group
1
Group 5
and Group 6
Group 3
and Group 4
Group
-Group
Group
Group
Group
Group
Group
Group
Group
7
a
3,4,
7,8
1,5,6
3,4,
7,8
1
5
6
3,4,
7,8
<70°
70"+
<70"
70°+
<70°
70e+
<70*
70°+
<70"
70°+
<70"
70°+
<70"
70°+
<70°
70"+
<70°
70"+
<70"
70°+
<70°
70e+
<70°
70°+
Number
Screened
3760
2534
1591
1402
1906
2215
2435
803
2861
3293
245
202
288
1057
457
1576
223
82
71
75
204
143
1288
1015
Number
Leaking
474
460
67
J01
448 .
518
52
95
17
29
21
27
14
52
7
28
19
18
4
9
1
3
84
42
Significant
Percent Effect of
Leaking Temperature
12.
18.
4.
7.
23.
23.
2.
11.
0.
0.
9.
13.
4.
4.
1.
1.
8.
22.
5.
12.
0.
2.
6.
4.
6 **
2
2 ,
2 **
5
4
1 **
8
6
9
0
0
7
9
5
6
5
0 **
6
0 2
5
1
5
1
952 Confidence
Interval
(11.
(17,
(3.3
(5.8
(21,
(21,
(1.6
(9.6
(0.4
(0.6
(5.3
(9.1
(2.6
(3,7
(0,6
(1.2
(5.2
(14,
(1.6
(5.6
(0,
(0,5
(5.2
(3.0
14)
20)
, 5.4)
, 8.7)
26)
26)
, 2.8)
, 14)
, 1.0)
, 1.2)
, 13)
, 19)
, 7.8)
, 4.3)
, 3.2)
, 2.6)
. 13)
32)
, 14)
, 22)
2.8)
, 6.1)
, 8.1)
, 5.6)
1-See Figure 3-2 for explanation of groups.
2-Showed significance when groups 5 and 6 were combined.
**Probabillty of no difference in leak frequency between ambient temperature categories is less than one percent.
-------
EFFECT OF ELEVATION ON LEAK FREQUENCY
This section evaluates the effect of source elevation on leak frequency.
The elevation of each screened source was recorded at the time of screening.
This elevation was expressed as the process unit landing level closest to
the screened source. For analysis in this report, the source elevation was
categorized as either ground level or above ground. The data is presented
in the same format as that for ambient temperature. Appendix C contains
summary statistics for elevation and the results of all statistical tests
on the elevation categories.
Table 3-26 gives the same type of summary for the effects of elevation
that was given for ambient temperature. Groups of process units and primary
materials were combined wherever the effect of elevation was consistent.
Only five of the 25 source type/primary chemical groups evaluated
indicated a significant effect of elevation on leak frequency. In all of
these cases, the sources at ground level had a higher leak frequency than
the sources above ground. The differences between the elevation categories
for those groups ranged from 1.7 to 6.0 percent leaking.
71
-------
TABLE 3-26. SUMMARY OF THE EFFECTS OF ELEVATION ON PERCENT LEAKING
ro
Source Stream .
Type Service
V«lvea Gaa
Light
Liquid
pump Light
Seals Liquid
Flanges Gas
Light
Liquid
Open Ended
Lines Gas
Light
Liquid
Prln*ry
Proceaa Material
Group Group
Ethylene Group 1
High Group S
Group 6
Ethylene Group 3,4
High Group 7
Group 8
Both Group 3,4
7.8
Both Group 1,5.6
Both Croup 3,4,
7,8
Ethylene Croup 1
High Group 5
Group 6
Ethylene Group 3,4
Htgh Group 7
Group 8
Elevation
Ground .
Abova
Ground
Above
Ground
Above
Ground
Above
Ground
Above
Ground
Above
Ground
Above
Ground
Above
Ground
Above
Ground
Above
Ground
Above
Ground
Above
Ground
Above
Ground
Above
Ground
Above
Number
Screened
3298
3977
479
• 749
423
1333
3123
1041
2494
795
4394
1743
437
10
481
863
1414
610
235
69
59
87
72
274
163
51
623
172
949
340
Number
Leaking
475
453
54
92
12
10
727
238
121
25
35
10
48
0
22
44
25
9
25
12
6
7
2
2
31
10
45
2
29
9
Significant
Percent Effect of
Leaking Elevation
14.4
15.2
11.3
12.3
2.8 **
0.7
23.3
22.9
4.8 *
3.1
0.8
0.6
11.0
0.0
4.6
5.1
1.8
1.5
10.6
17.4
10.2
8.0 2
2.8
0.7
19.0
19.6
7.2 »*
1.2
3.1
2.6
95 J Confidence
Interval
(13, 16)
(14, ">
(8.5, 14)
(10, 14)
(1.4, 4.9)
(0,4, 1.4)
(22. 25)
(20, 26)
(4.0, 5.8)
(2.0, 4.6)
(0.6, 1.1)
(0.3, 1.0)
(B.3, 14)
(0, 3D
(3.0, 7.8)
(3.7. 7.0)
(1.1. 2.6)
(0.7, 2.8)
(7.1. 16)
(9.3, 28)
(3.8, 21)
(3.3. 16)
(0.3, 9.7)
(0.1, 2.6)
(13, 2«)
(9.8, 33)
(5.3, 9.4)
(0.1, 4.1)
{2.0. 4.4)
(1.2, 4.9)
1-See Figure 3-2 for explanation of groups.
2-Therc wao a Qigniflcant difference between elevation categories when chemical groups 5 and 6 were combined.
*Probability of no difference in leak frequency between elevation categories Is less.than five percent.
**Probablllty of no difference in leak frequency between elevation categories is less than one percent.
-------
-------
SECTION 4
EMISSION FACTOR DEVELOPMENT FOR THREE PROCESSES
This section presents detailed results of the investigations in the
following three areas:
* Distribution of screening values,
Estimation of emission factors, and
Mass emission distribution over the range of screening
values.
DISTRIBUTION OF SCREENING VALUES
Distributions of OVA screening values were examined for each process,
source type (valve or pump seal) and service. From past experience with
the refining industry, it was expected that the distributions of the nonzero
screening values could be modeled with a lognormal distribution. It was
anticipated that censoring above 100,000 ppmv would occur due to the inconsis-
tent use of a secondary OVA dilution probe. Figure 4-1 shows a typical
histogram of the logarithms of the nonzero screening values with a pattern
that frequently occurred: a large number of observations nominally at 100,000
ppmv with positive skewness (more large values occurring than expected from
a normal distribution) and negative kurtosis (flatter peak and shorter tails
than a normal distribution). Further examination of the distributions by
primary material classification showed similar departures from the lognormal
distribution. Two approaches were subsequently taken in modeling the distri-
bution of screening values: fitting an empirical cumulative distribution,
which reflects detailed features of the data, and fitting a cumulative log-
normal distribution to the nonzero screening values, with adjustment for
censoring of the data. Section 7 contains a more detailed discussion of these
73
-------
distribution models. Figures 4-2 through 4-10 compare the lognormal models
with the empirical distributions. The departure from lognormality of the
screening data does not appear large in magnitude. The lognormal model was
therefore used in the development of both screening value distributions and
distributions of mass emissions.
EMISSION FACTORS AND CUMULATIVE DISTRIBUTIONS OF TOTAL EMISSIONS BY
SCREENING VALUES
This section briefly describes the estimated emission factors and mass
emission functions. A more detailed discussion of the statistical methods
and assumptions employed is found in Section 7.
Table 4-1 presents estimated emission factors for nonmethane hydrocarbon
fugitive emissions from valves and pump seals. Figures 4-11 and 4-12 show
graphically how these emission factors compare between the three processes
considered in this report.
Comparison of emission factors among the three processes by means of
their related confidence intervals shows only one difference that can be
considered to be statistically significant: ethylene has a significantly
larger emission factor than vinyl acetate for valves with light liquid service
Note, however, that ethylene consistently shows the largest emission factor,
followed by cumene and vinyl acetate. For pump seals, ethylene and cumene
have about the same emission factor. With the exception of vinyl acetate,
pump seals have larger emission factors than do valves. Finally, for com-
parable sources, gas service has higher emission factors than light liquid
service.
Fugitive emissions may also be compared by means of the cumulative
distribution of total mass emission by screening value. These curves relate
the OVA screening value to the percentage of the total mass emission which
can be expected from all sources with screening values greater than any given
74
-------
value. These cumulative functions have been estimated for each process,
source type, and service. Figures 4-13 through 4-21 display the cumulative
mass emission estimated by an empirical function, with the lognormal model
superimposed for comparison. Both the lognormal model and the empirical
function are described more fully in Section 7.
Confidence bounds are given to indicate how well the cumulative mass
function has been estimated from the data collected in both the screening
and maintenance programs. The development of these intervals is discussed
in Section 7. In using these estimated functions and confidence intervals,
it should be kept in mind that the relationship between screening values and
mass emissions is imperfect. Also, the true distribution of screening values
is not known precisely: it is estimated from the observed screening value
distribution. These two sources of variation contribute to the width of the
confidence bands shown in the figures.
Figures 4-22 through 4-30 show the cumulative distribution functions for
screening values (Part a) and mass emissions (Part b) based on the lognormal
model for the screening values.
Application of Figures 4-22 through 4-30 may be illustrated through the
use of Table 4-2, which exhibits point estimated and 95% confidence intervals
for both the percentages of sources screening ^ 10,000 ppmv and the percentage
of total mass emissions attributable to sources screening ^ 10,000 ppmv. For
example, approximately 15% of ethylene process valves in gas service can be
expected to have screening values above 10,000 ppmv (Figure 4-25a). However,
these 15% of the valves are responsible for an estimated 94% of the mass
emissions (Figure 4-25b). In the same manner, other specific screening
values (or percentage of sources) could be chosen and the corresponding
percentage of mass emissions found.
75
-------
bno
2 Oil
inr
NOTE:
Log (10) = 2,30 '
Log (100) = 4.61
Log (1000) = 6.91
Log (10,000) = 9.21
Log (100,000) = 11.51
Log (1,000,000) = 13.82
I
1
1
t-
1
1
1
1
1
I
1
t
1
1
I
1
+
1
1
1
1
•t
1
1
1
1
*****
*****
•1 t * * *•
*****
-t **+•*
M 1 t * -t t +- 1 *
t * t * * t****
*+M* *****
***** •«*•***
**.!** J t * * *
** * **
*****
*****
*****
*****
***» * •
***.* +
*****
*****
*****
******
** *** '
*****
** * **
*****
*** **
*****
*** t *
*****
*****
*****
**•** *
*****
*****
**•»-* -I
*****
t** **
* *Jt **
*.*» * t
*****
*****
»,*-, *• *
*****
*** * *•
*****
*****
*****
**»* *
*****
** * 4 *
*****
* 4 ^ t *
*****
**** *
*****
*****
*****
*****
*****
*****
*****
* ****
* * * ^* *
*****
*****
*****
*****
*****
*****
*****
** ***
*****
* * * * #
*****
*****
*****
*****
*****
*****
*****
** ***
*****
*****
*****
*****
*****
*****
*****
**-***
*****
** ***
**•«*:*:
*****
*****
*****
*****
*****
*****
*****
*****
*****
* ****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*•**.**
*****
' *****
*****
*****
*****
*****
*****
***** *****
10
11
12
LOGP (OVA SCREENING VALUE)
Figure 4-1. Typical Distribution of Log (OVA Screening Value)
Ethylene Process, Valves in Gas Service
-------
R i
C 35 +
El
N I
T I
30 +
Si
C
I
R t
E 25 +
El
Ml
II
W 20 +
C I
I
G 1
R 15 +
El
A I
T I
E 10 +
I
R
* *
x x ***
xxx
xxx
**x x
x '
* x
* *
X x , ,
XX.. .
* XXX
***** X ....
+* X
* * XXX ., .
.. XX ...
. . . , ****XXX . ,
.. * XXX.
. . , ** XX
, ***
... *
x = lognormal model
* = empirical function
* = 95% confidence limits
for empirical function
XX . *
* XX . . .
** X
*** XX ....
* * XX . .
** X
... ** X X
* *XX
XX *
X
1.0
1.5
2,0
2.b
i.O
3.5 4.0
SCREENING VALUE)
H,5
5.0
5.0
t>.U
Figure 4-2. Cumulative Distribution of Sources by Screening Values -
Cumene Process, Valves in Gas Service
-------
35
P
E 30
R
C
E
M
T 25
S
C
ft
E 20
E
N
I
N
& 15
G
H
E
A 10
T
E
R
5
0 +•
* . •
X X*
xxx*** ...
xxxx
. . XXX ..
.... *xx ..
.. XX ..
... ** XX .
.i ** XX.
. ** XV,
. . XXX
.. ** XX. . .
* xx..
.< ** X *.
*+** xx,.
* xx..
... * XX. ,
*** xx..
... ** X ...
. ** XX .,
... ** xx
+ *xx
*xxx...
*xx . ..
* XX
* *** X
.. * *x
. *
X = lognormal model
* = empirical function
• = 95% confidence limits
for empirical function
x** **
XX ****
xxxx
XXX
l.o
i.a
a.o
i.o
i.b
t.b
6.U
Figure 4-3. Cumulative Distribution of Sources by Screening Values -
Cumene Process, Valves in Light Liquid Service
-------
64
60
p
E
R
C 56
E
M 52
36
44
R
E 40
E
N
I
N 32
G
28
G
R 24
E
A 20
T
E 16
K
12 +
I
8 -f
I
lognormal model
empirical function
95% confidence limits
for empirical function
x *
X
X
*
5.0
1.0
2,U
3.0
3.b 4.0
SCKELNING VALUE)
4.5
5.3 b.U
Figure 4-4. Cumulative Distribution of Sources by Screening Values
Cumene Process, Pumps in Light Liquid Service
-------
45
40
P
t
R
C 35
E
N
T
JO
S
c
R
E 25
E
N
I
Co N 20
0 G
U
R 15
(,
A
T
E 10
K
5
0
XX. •*••..
, XXXXK+**..
jXxxx ^ = l°gnormal model
..xxx * = empirical function
•**x*x • = 95% confidence limits
± .*. « *t y y v
xxx for empirical function
«... xxx
.**** . xxxx
..*** AX
xxx
****.. XX
«.«***.. XXX
* w Y • T 4 M «t rt A
.***.. XX
•..**... XX
.»***,, XX
...***.. XX
..***.. xxx
*.***•*. *xx
.,****, ..XX
*'"!!!!I**xx.
. ,***tx. .
...**XKK.*.*
. ..*xxx*. ....
....•OK***.
* t*XXt******
XXA...4
xxx.
xxx
XX
XXX AX
**** * *
** *** «»
l.C 1 . iJ «!,U 2,t> 4.0 3.5 t+.O 4.5 5.0 5.3 fa.U
Figure 4-5. Cumulative Distribution of Sources by Screening Values -
Ethylene Process4 Valves in Gas Service
-------
00
H1
70
P
L 60
R
C
E
N
T 50
S
C
R
E 40
E
N
N
G 30
G
K
A 20
T
E
f<
10
* ** *** ,4,,
X XX XXX X***...
1.0
l,b
,.****,XXX*
...,**..xxxx
...**#. xxx
X = lognormal model
* - empirical function
• = 95% confidence limits
for empirical function
,.****xx
...****xx
...*«xxxx
.****.xx.
...»***!**.,.
»**«X
X X
X X
xxx
. ..•• •*.*
t * * * * **#*
._•!_.
2.0
2.i> 3.0
L.OG1UIOVA
3,5
1.0
VALUE)
H,5
5.0
5.3 6.U
Figure 4-6. Cumulative Distribution of Sources by Screening Values -
Ethylene Process, Valves in Light Liquid Service
-------
CO.
70
E 60
R
C
E
N
T 50
S
C
il
E HO
E
N
I
i-J
G 30
6
R
E
A 20
r
E
10
*
X X*
X
X
* *
* XX
* X
* * X
*x
* * »
*xx
*
X
* XX
X
* x
X = lognormal model
* = empirical function
• = 95% confidence limits
for empirical function
i.o
1.3
2.0
3.0
3.b
4.0
5.0
fa.U
Figure 4-7. Cumulative Distribution of Sources by Screening Values -
Ethylene Process, Pumps in Light Liquid Service
-------
17.5
E ib.o
R
C
E
N
T 12.5 -f
S
C
R
E 10,0
E
N
I
N
G 7.5
G
H
E
A 5,0
T
E
R
2,5 *
XX**
X X *
. . x
0.0
XX,.
*XX
* XXX.
. * X
**
. .
XXX
If
* ».
** . xxx
* . .x
* .XX
**
. * *.
*, **
** X . *
* * XXX
* * * x
** *x
.. * X
= lognormal model
= empirical function
= 95% confidence limits
for empirical function
XKX
*** *
xxx
X
• - + -.
b.U
3.0
LOGlO(UV/\ y
Figure 4-8. Cumulative Distribution of Sources by Screening Values -
Vinyl Acetate Process, Valves in Gas Service
-------
00
p
E 6
R
C
E
N
T 5
S
C
R
E 1
E
N
I
W
ti 3
G
R
E
A 2
T
E
R
1
**
XX
1.0
X M
X*
X.
X!
X.
X
[* X i
* X .
* .
X
. ** X
* *
I *
.. *
XX
*+x
X ,,
* X .
** XX .,
•. * X ..
** XX
X
***x
. **
X = lognormal model
* = empirical function
• = 95% confidence limits
empirical function
..
* xx *
X * *
l.b
2.0
4.0 3.5 4.0
IU&1UIOVA sCKEENlNf, VALUE)
. 5
.- + _.
5.0
•- + -•
b.5
.- + -.
d.U
Figure 4-9. Cumulative Distribution of Sources by Screening Values -
Vinyl Acetate Process, Valves in Light Liquid Service
-------
00
27.
25.
24.
23.
22.
21.
20.
19.
Id.
16.
15.
1H.
13.
12.
11.
10.
9.
7.
6.
5.
4.
3.
2.
1.
0.
oon
875
750
625
500
375
250
125
000
875
750
625
500
375
25n
125
000
875
750
625
500
375
250
125
000
X = lognormal model
* = empirical function
• = 95% confidence limits
for empirical function
K X
* X
i.o
* X
* X
* X
* X
*
*
X X
* X
XX
*
* X
* X
*
1.5
3.U 3.5
UUdlOtOVA
4.0
VALUt)
H.5
5.0
6.U
Figure 4-10.
Cumulative Distribution o£ Sources by Screening Values
Vinyl Acetate Process, Pumps in Light Liquid Service
-------
TABLE 4-1. ESTIMATED EMISSION FACTORS FOR NONMETHANE HYDROCARBONS
FROM VALVES AND PUMP SEALS
Source Type
Emission Factor (Confidence^
Number Screened
(Ibs./hr./source)
(kg,/hr./source)
00
Valves
Gas
Ethylene
Cumene
Vinyl Acetate
Light Liquid
Ethylene
Cumene
Vinyl Acetate
6,294
448
949
i
4,176
799
2,137
0.024(0.008, 0.07)
0.011(0.003, 0.05)
0.0046(0.001, 0.03)
0.020(0.007, 0.06)
0.0056(0.002, 0.02)
0.0003(0.0001, 0.002)
0.011(0.004, 0.03)
0.0052(0.001, 0.02)
0.0021(0.0004, 0.01)
0.010(0.003, 0.03)
0.0025(0.001, 0.01)
0.0001(0.00003, 0.001)
Pump Seals
Light Liquid
Ethylene
Cumene
Vinyl Acetate
76
25
89
0.069(0.006, 0.8)
0.052(0.001, 2.7)
0.0043(0.0001, 0.1)
0.031(0.003, 0.4)
0.023(0.0004, 1.2)
0.0020(0.00006, 0.06)
-------
00
GJ
H
id
o
WJ
o
•a
a
D
P.
0. t
0.01
0.001
0.000
CUMENE
V. ACETATE
ETHYLENE
J-, Qas Source
ETHYLENE
CUMENE V. ACETATE
Light Liquid Service 1
Figure 4-11. Emission Factors—Valves
-------
g-
p-
en
en
H-
O
3
88
pounds/hour/source (log scale)
©
G3
i ~r
OS
C
i-l
ffi
o
C
2
rn
Z
rn
r
m
z
rn
CO
ro
H
o
R
-I
>
-I
rn
-------
00
100
90
60
70
T 60
0
T
A 50
L
30
20
10
X**
.....
lognormal model
empirical function
95% confidence limits
for empirical function
***** XX
* **XXX
*** * X
* * *»X
. ***X
* X
*
*•
1.0
1.5 2.0 2.5 3.0 3.5
LOGIOIOVA
H.O
4.5
b.O
6.0
Figure 4-13. Cumulative Distribution of Total Emissions by Screening Values
Cumene Process, Valves in Gas Service
-------
\
I
I
I
100 * * X X « XXXXXXXXXXm* •
i .*****XKXXXXXXXX
I ****XXKXXXX ...
90 * ... »*XXXX
P| **KXXX
E I .. ***XX
R tiO + ... ' *XxX
C I .*• *XXX
El .. *** X
ij 70 + . . * »X
T I .... #* **
I ... X* *»
T £.0 t •. X *
01 X **
T I , XX *
ft 50 + ' X
LI ..XX
I •
E HO + .
Ml
II ' -.
s 30 + x * lognormal model
[i * « empirical function *\
a 20 + . - 95% confidence limits
(J [ for empirical function *
10 +
i
i
o + .
i
I
I
i
1.0 1.5 2.0 2.5 3.0 3.5 4.0 1.5 5.0 5.5 6.0
LOGlbfOVA SCREENING VA|_UE)
Figure 4-14. Cumulative Distribution of Total Emissions by Screening Values
Cumene Process, Valves in Light Liquid Service
-------
100
90
E
K (10
•C
E
N 70
T 60
0
T
A 50
L
50
0 20
* X X
*
lognormal model
empirical function
95% confidence limits
for empirical function
1.0 1.5 2.0 2.5 3.C 3.5 4.0 4.5 5." 5,5 6.0
LOGlOtOVA SCREENING
Figure 4-15. Cumulative Distribution of Total Emissions by Screening Values
Cumene Process, Pumps in Light Liquid Service
-------
I
I
I
t
100 * x***ju*»***m»****x*******x******if *******
i XXKXXM**
90 * ... XXX«**»*
P I .... AXXX ****
E I ... XXX »***
H ao + ' ... xx *
C.I ..XX
El ..XX
N 70 + ..XX
T I ..XX
I .. X
T 60 + ..X
0 I X
T I ... X
A 50 + ' , . X
LI . . X
I
C 40 + X ..
HI X ...
\ ' X = lognormal model
si * = empirical function ' »**» * x
I' . = 95% confidence limits * * ** x
" 20 \ for empirical function *** *
!1 I X «
I
10 +
I X
I X
0 + .
I
t
I
3.0 3.5 1.0 4.5 &,0 5.b 6.1)
tOGtU
Figure 4-16* Cumulative Distribution of Total Emissions by Screening Values
Ethylene Process, Valves in Gas Service
-------
loo +
p
E
K 60 +
C
e
N 70
T 60 +
0
T
A 50
L
E 40
30
I
S
S
I
0 20 +
ft
10 +
*K
lognormal model
empirical function
95% confidence limits
for empirical function
,**xx*xxx
»XXXXM..»,
.... *XXX
X**
xxx***
,. XX ***
.. XX
.. XX
.. x
XX
...X
* * * * » ***
*
1.0 1.3 2.0 2.5 3.0 3.5
LOG10(OtfA
4.0
4.5
b.O
Figure 4-17. Cumulative Distribution of Total Emissions by Screening Values
Ethylene Process, Valves in Light Liquid^Service
-------
100
90
P
E
R 60
C
E
M 70
T
T 60
0
T
A 50
L
E tO
N
I
S 30
S
I
0 20
N
10
0
Kilt *
• • *
X * lognormal model
* = empirical function
. = 95% confidence limits
for empirical function
** X
*
1.5 ?,0 2.5 3.0 3.5 4,0
UOG1CHDVA SCREENING
5.0
S.5
6.0
Figure 4-18. Cumulative Distribution of^Total Emissions by Screening Values
Ethylene Process, Pumps in Light Liquid Service
-------
100
90
flO
70
50
30
20
1.0
*H X K X «** *
XX XH * **
x * *
XX * +
.. x * *
. . X
XX
X
X = lognormal model
* = empirical functipn
. = 95% confidence limits
for empirical function
xx
i.o
2.0 2.5 3.0 3.5 1,0
UOGlOfOUA SCREENING VA| Ui
5.0
5.5
Figure 4-19. Cumulative Distribution of Total Emissions by Screening Values
Vinyl Acetate Process, Valves with Light Liquid Service
-------
I
I
I
100 + XX » X t
| XXM VXX
I .. . . ***X XX
go + * x *
PI . x x*«
El . XX ** ' .
R 80 t . XX*
C I .. X**
El . XX *
ftl 70 + . **
T I . XX * .
I • X **
T 60 + . X **
01 . XX *
T I XX *
A 50 + . ' * *
U I .. *
I
E 40 + . X «
HI
II X *
S 30 + . X
S I X.
I I X = lognormal model
^ ° J * = empirical function '
I . " 95% confidence limits .. x
10 * for empirical function
i x
I x
I
I
I
I
1.0 1.5 £.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
LOt](J(OVA SCREENING VA(.UE)
Figure 4-20. Cumulative Distribution of Total Emissions by Screening Values
Vinyl Acetate Process, Valves in Light Liquid Service
-------
p
E
R
C
F.
iJ
T
T
0
T
A
L
E
\o ,
-J I
S
S
I
0
•'
1
1
1
1
1 X * XX
1 * X
90 +. X
1
1
00 + .
1 *
1 X
70 + .
1 . X
I
60 +
1 •
t
50 + "X
1
1 *
HO +
1
1
30 + ,
|
1 .
20 +
1
*
10 +
1
(I •*• *.*• .(•
1
1
1
1
1.0 1.5 2.0 2.5 3,0 3.5 4.0 4.5 5.0 5.5 6.0
LOGJOIOVA SCREENING VALUE)
Figure 4-21. Cumulative Distribution of Total Emissions by Screening Values
Vinyl Acetate Process, Pumps in Light Liquid Service
-------
CO
p
E
R
C
E
N
T
S
C
R
E
E
N
I
N
G
G
R
E
A
T
E
R
40
0
•7— Estimated Percent of Sources
Screening Greater Than the Selected Source
— — 95X Confidence Limits
345
LOG10COVA SCREENING VALUED
Figure 4-22a. Cumulative Distribution of Sources by Screening Values
Cumene Process, Valves in Gas Service
-------
P
E
R
C
E
N
T
T
0
T
A
L
E
M
I
S
S
I
0
N
100
40
0
~^Estimated Percent of Total
Mass Emissions Attributable
to Sources With Screening
Values Greater Than the
Selected Value
— 95% Confidence Limits
0
2345
LOG10COVA SCREENING VALUE)
Figure 4-22b. Cumulative Distribution of Total Emissions by Screening Values
Cumene Process, Valves in Gas Service
-------
o
o
p
E
R
C
E
N
T
S
C
R
E
E
N
I
N
G
G
R
E
A
T
E
R
80
40
20
—— Ectiiutad Percent of Sour CM
Screening Greater Than the Selected Source
— — 95X Confidence Unite
345
LOGI0COVA SCREENING VALUE)
6
Figure 4-23a,
Cumulative Distribution of Sources by Screening Values
Cumene Process, Valves in Light Liquid Service
-------
E
R
C
E
N
T
T
0
T
A
L
E
M
I
S
S
I
0
N
40
0
0
— Estimated Percent of Total
Mass Emissions Attributable
to Sources With Screening
Values Greater Than the
Selected Value
— 95% Confidence Limits
2345
LOG10COVA SCREENING VALUE?
6
Figure 4-23b. Cumulative Distribution of Total Emissions by Screening Values
Cumene Process, Valves in Light Liquid Service
-------
P
E
R 100
c
E
N
T 80
S
C
R
E
E
N
I
N
G
G
R
E
A
T
E
R
60
40
0
• ' Eetlnated Percent of Sourcee
Screening Greater Than Che Selected Source
95X Confidence Limits
2 34 5
LOG10COVA SCREENING VALUE)
6
Figure 4-24a.
Cumulative Distribution of Sources by Screening Values
Cumene Process, Pumps in Light Liquid Service
-------
P
E
R
C
E
N
T
T
0
T
A
L
E
M
I
S
s
I
0
N
00
60
40
— Estimated Percent of Total
Mass Emissions Attributable
to Sources With Screening
Values Greater Than the
Selected Value
— 95% Confidence Limits
12345
LOG10COVA SCREENING VALUE)
Figure 4-24b. Cumulative Distribution of Total Emissions by Screening Values
Curaene Process, Pumps in Light-Liquid Service
-------
O
•P-
P
E
R
C
E
N
T
S
C
R
E
E
N
J
N
G
G
R
E
A
T
E
R
100
60
40
20
0
— Estimated Patcant of Sources
Sct««nlng Gteatar Than the Selectad Source
— 95X Confidence Limits
345
LOG10COVA SCREENING VALUE}
Figure 4-25a. Cumulative Distribution of Sources by Screening Values
Ethylene Process, Valves in Gas Service
-------
o
Ul
p
E
R
C
E
N
T
T
0
T
A
L
E
M
I
S
S
I
0
N
103
40
— Estimated Percent of Total
Mass Emissions Attributable
to Sources With Screening
Values Greater Than the
Selected Value
-- 95% Confidence Limits
0
0
2345
LOG10COVA SCREENING VALUE)
6
Figure 4-25b,
Cumulative Distribution of Total Emissions by Screening Values
Ethylene Process, Valves in Gas Service
-------
P
E
R
C
E
N
T
S
C
R
E
E
N
I
N
G
G
R
E
A
T
E
R
100
— Estimated Percent of Soureaa
Screening Greater Than th* Selected Source
— 951 Confidence Limits
60
40
20
0
345
LOG10COVA SCREENING VALUE)
Figure 4-26a.
Cumulative Distribution of Sources by Screening Values
Ethylene Process, Valves in Light Liquid Service
-------
p
E
R
C
E:
N
T
T
0
T
A
L
E
M
I
S
S
I
0
N
108
60
40
0
— Estimated Percent of Total
Mass Emissions Attributable
to Sources With Screening
Values Greater Than the
Selected Value
— 95% Confidence Limits
0
2345
LOG10COVA SCREENING VALUE)
6
Figure 4-26b,
Cumulative Distribution of Total Emissions by Screening Values
Ethylene Process, Valves in Light Liquid Service
-------
o
O3
p
E
R
C
E
N
T
S
C
R
E
E
N
I
N
G
6
R
E
A
T
E
R
80
60
20
0
— Estimated Percent of Sources
Screening Greater Than ehe Selected Source
95Z Confidence Limit*
234 5
LOGI0COVA SCREENING VALUE)
Figure 4-27a. Cumulative Distribution of Sources by Screening Values
Ethylene Process, Pumps in Light Liquid Service
-------
100
R
C
E
N
T
T
0
T
A
U
E
M
I
S
S
I
0
N
Estimated Percent of Total
Mass Emissions Attributabl*
to Sources With Screening
Values Greater Than the
Selected Value
95% Confidence limits
2345
LOG10COVA SCREENING VALUE)
6
Figure 4-27b. Cumulative Distribution of Total Emissions by Screening Values
Ethylene Process, Pumps in Light Liquid Service
-------
p
E
R 100
C
E
N
T 80
S
C
R
E
E
N
I
N
G
6
R
E
A
T
E
R
60
40
20
0
— Estimated Percent of Sources
Screening Greater Than the Selected Source
— 95% Confidence Limits
345
LOG10COVA SCREENING VALUED
Figure 4-28a.
Cumulative Distribution of Sources by Screening Values
Vinyl Acetate Process, Valves in Gas Service
-------
E
R
C
E
N
T
T
0
T
A
L
E
H
I
S
S
I
0
N
813
40
0
—- Estimated Percent of Total
Mass Emissions Attributable
to Sources With Screening
Values Greater Than the
Selected Value
— 95% Confidence Limits
12345
LOG10COVA SCREENING VALUE)
6
Figure 4-28b, Cumulative Distribution of Total Emissions by Screening Values
Vinyl Acetate Process, Valves in Gas Service
-------
p
E
R
C
E
N
T
S
C
E
E
N
I
N
G
G
R
E
A
T
R
108
80
60
Estimated Percent of Sources
Screening Greater Than the Selected Source
— — 951 Confidence Limits
20
0
345
LOG10COVA SCREENING VALUE)
6
Figure 4-29 a.
Cumulative Distribution of Sources by Screening Values
Vinyl Acetate Process, Valves in Light Liquid
-------
Estimated Percent of Total
Mass Emissions Attributable
to Sources With Screening
Values Greater Than the
Selected Value
95X Confidence Limits
12345
LOG10COVA SCREENING VALUE)
Figure 4~29b. Cumulative Distribution of Total Emissions by Screening Values
Vinyl Acetate Process, Valves in Light liquid Service
-------
p
E
R
C
E
N
T
S
C
R
E
E
N
I
N
G
G
R
E
A
T
E
R
100
—— Estimated Percent of Sources
Screening Greater Than Che Selected Source
95X Confidence Limits
40
20
0
3 4.5
LOG10COVA SCREENING VALUE)
Figure 4-30a.
Cumulative Distribution of Sources by Screening Values
Vinyl Acetate Process, Pumps in Light Liquid Service
-------
H
K
Ui
E
R
C
E
N
T
T
0
T
A
L
E
M
I
S
S
I
0
N
100
80
60
40
0
—- Estimated Percent of Total
Mass Emissions Attributable
to Sources With Screening
Values Greater Than the
Selected Value
— 95X Confidence Limits
2345
LOG10COVA SCREENING VALUE}
Figure 4-30b. Cumulative Distribution of Total Emissions by Screening Values
Vinyl Acetate Process, Pumps in Light Liquid Service
-------
TABLE 4-2. SUMMARY OF PERCENT OF SOURCES DISTRIBUTION CURVES AND PERCENT
OF MASS EMISSIONS CURVES AT SCREENING VALUE OF 10,000 PPMV
Percent of Sources
Screening > 10,000 ppmv
Source Type
Valves
Gas
Ethylene
Cumene
Vinyl Acetate
Light Liquid
Ethylene
Cumene
Vinyl Acetate
Pump Seals
Light Liquid
Ethylene
Cumene
Vinyl Acetate
Estimate
15
16
3.7
26
12
0.2
30
14
1.7
95% Confidence
Interval
(14
(13
(2,
(24
(10
(0,
(20
(1,
(0,
» 16)
> 19)
5)
, 27)
, 13)
0.4)
, 39)
27)
4)
Percent of Mass Emissions
Attributable to Sources
Screening ^ 10,000 ppmv
Estimate
94
94
90
89
80
25
96
89
67
95% Confidence
Interval
(93,
(90,
(85,
(87,
(72,
( 9,
(90,
(50,
( 5,
95)
96)
94)
90)
86)
47)
98)
98)
92)
-------
SECTION 5
EVALUATION OF THE EFFECTS OF LEAK OCCURRENCE,
RECURRENCE, AND REPAIR ON MASS EMISSIONS
This section presents the results of investigations of leak occurrence,
recurrence, and maintenance effects on VOC mass emissions. The analysis is an
extension of previous work on these aspects of fugitive emissions control
presented in Reference 2 (maintenance study).
EFFECT OF LEAK OCCURRENCE ON MASS EMISSIONS
Leak occurrence was defined in Reference 2 for sources initially
screening < 10,000 ppmv as the first occurrence of a leak (screening ^ 10,000
ppmv) at any time after the initial screening. In the maintenance study,
described in Reference 2, there were 651 valves and 89 pumps which screened
below 10,000 ppmv initially, and were subsequently rescreened two to six times
over a six month period. Estimated leak occurrence rates were developed in
Reference 2 for both valves and pump seals. This section presents estimates
of the effect on mass emissions from those leak occurrences. The statistical
procedures used to develop these estimates are discussed in Section 7.
Table 5-1 snow estimates of the weighted percent increase (WPI) and the
increase in mass emissions for the sources for which leaks did and did not
occur. The WPI is applicable as an estimate of the effect of leak occurrence
on mass emissions. The mean emission estimates (Ib/hr/source) are applicable
only to the data from the maintenance study since they represent the combined
data from three specific chemical processes.
117
-------
TABLE 5-1. INCREASE IN MASS EMISSIONS BY LEAK OCCURRENCE1' FOR VALVES
AND PUMP SEALS SCREENING < 10,000 ppmv INITIALLY
H
H*
Oo
Source Category
(number of sources
in category)
Sources with leak
occurrences
Valves (30)
Pump Seals (15)
Sources without
leak occurrence
Valves (621)
Pump Seals (74)
Weighted Percent
Increase (%)
530
(200, 900)
75
(-100, 6000)
-37
(-56, -18)
-47
(-100, 11)
Mean Emissions
At Initial Screening
0.0052
(0.001, 0.03)
0.013
(0.001, 0.1)
0.00065
(0.0002, .002)
0.0014
(0.0002, 0.01)
(Ibs/hr/source) Mean Emissions
At First Leak Occurrence
or Last Screening
0.033
(0.006, 0.1)
0.99
(0.005, 10)
0.00041
(0.0001, 0.002) (-0
0.00075
(0.00004, 0.005)
Increase
(Ibs/hr/source)
0.028
(0.005, 0.2)
0.98
(0.02, 20)
-0.00024
.001, 0.00002)
-0.00066
(-0.02, 0)
1) Screening ^ 10,000 ppmv the first time following initial screening
Note: leak rates estimated at initial screening and measured (or estimated) at either (1) time of
first occurrence of (2) time of last screening
Note: estimates are reported with an approximate 95% confidence interval
-------
The valves with a leak occurrence had a WPI in emissions of 530% while
the valves without leak occurrence showed a slight decrease in emissions
(WPI = -37%). These estimates can be combined with the occurrence rates
estimates in Reference 2 to estimate the total..impact of leak occurrence on
mass emissions- The confidence intervals for these estimates should be con-
sidered in analyses of this type. The confidence intervals for the WPI
estimates for pumps are quite large and include zero (no increase).
EFFECT OF LEAK RECURRENCE ON MASS EMISSIONS FOR VALVES
Leak recurrence was defined in Reference 2 for maintained valves
(screening value < 10,000 ppmv immediately after maintenance) as a leak
(screening value ^ 10,000 ppmv) at any time after maintenance. In the
maintenance study (Reference 2) there were 28 valves with the potential for
leak recurrence (i.e., with screening value ^ 10,000 ppmv before maintenance
and < 10,000 ppmv immediately after maintenance). Eight valves exhibited a
leak recurrence during the six month period after maintenance. Leak recur-
rence rates for valves were estimated in Reference 2 using these data. This
section presents estimates of the effect on mass emissions from these leak
recurrences.
Table 5-2 shows estimates of the weighted percent increase (WPI) and
estimates of the mean emissions before maintenance, after maintenance, and
after recurrence or at time of last screening. As with the occurrence esti-
mates, the mean emission estimates are applicable only to the data from the
maintenance study. The confidence intervals for the WPI estimate include
zero in both cases due to the small number of sources studied for recurrence,
The estimates can be combined with recurrence rate estimates in Reference 2
to evaluate the impact of recurrence on emissions from valves, but the con-
fidence intervals should be considered in these evaluations.
119
-------
TABLE 5-2. INCREASE IN MASS EMISSIONS BY LEAK RECURRENCE FOR VALVES
SCREENING < 10,000 ppmv IMMEDIATELY AFTER MAINTENANCE
Source Category
Valves with leak
recurrence
(8 valves)
Valves without
leak recurrence
(20 valves)
Weighted Percent
Increase (%)
510
(-100, 1700)
-50
(-96, -5)
Mean Emissions (Ib/hr/source)
Before After At first Recurrence
Maintenance Maintenance or Last Screening
0.26 0.0033 0.02
(0, 0.6) (0, 0.02) (0, 0.08)
0.024 0.0016 0.0008
(0.01, 0.04) (0.0001, 0.01) (0.0001, 0.002)
Mean Emissions
Increase at
Recurrence or
Last Screening
(Ib/hr/source)
0.017
(-0.04, 0.2)
-0.0008
(0.008, 0.002)
1) Screening > 10,000 ppmv the first time following after maintenance screening
Note: leak rates measured (or estimated) after maintenance and at either (1) time of first
recurrence or (2) time of last measurement
Note: estimates are reported with an approximate 95% confidence interval
-------
FURTHER ANALYSIS OF EFFECT OF VALVE MAINTENANCE ON MASS EMISSIONS
A statistical analysis was done to expand on the analysis of the
immediate effect of valve maintenance in Reference 2. Reference 2 reported
a weighted percent reduction (WPR) of 71% (95% confidence interval of 54%
to 88%) for 155 valves for which maintenance was performed. The WPR for the
97 valves with a before maintenance screening valve of > 10,000 ppmv was 70%
(95% confidence interval of 46% to 95%). Reference 2 also reported that
only 29% of the 97 valves were "repaired" by simple on-line maintenance,
where a "repair" is defined as screening below 10,000 ppmv immediately after
maintenance. This analysis compares the reduction for the 29% of the sources
repaired with the 71% not repaired.
Table 5-3 summarizes this comparison and Figures 5-1 and 5-2 show the
before minus after maintenance leak rates plotted against the before mainten-
ance leak rates for the "repaired" and "non-repaired" valves. The weighted
percent reduction for repaired valves was 97.7% (95%, 100%) compared with
62.6% (41%, 85%) for non-repaired valves. This significant difference in
emissions reduction between the two groups of valves can be seen by comparing
the data plots in Figure 5-1 and 5-2.
Table 5-3 also contains estimates of the mean emissions from the valves
before and after maintenance. These estimates are only applicable to the
sources in the data base since they represent the combined data from valves
from three specific chemical processes.
121
-------
ro
to
lO.OOOo
B
E
F 1*0000
0
R
E 0*1000
H
I 0*0100
N
U
s o.noio
A
F O.OOOi
T
E
R Q.QOOo
E -O.OOOl
A
K
-0*0010
H
A
T -0.010Q
-O.lOOfl
"Repaired" Valves
LEGENDI A s 1 DBS, B = 2 QbSt ETC.
perfect repair
(100% reduction)
--t + +—„ ——-+ + +- --+.._—„—+—^—--+ --^—+-
0,0001 0.0nl)3 0.0010 0.0030 0.0100 0.0300 0*1000 0.30UO 1.0000 3.0000
BEFORt MAINTENANCE LEAK HATE
Figure 5-1. Before Minus After Maintenance Leak Rate -
Valves Screening < 10,000 ppmv After Maintenance
-------
NJ
Lo
10.0000
1.0000 +
I
0,0001 +
1
I
R O.OOOQ +•
I
L I
E -O.OQOi +
A I
K I
-0.001Q +
K !
A I
T -O.Olflo +
E I
I
-O.lOOQ +
I
0.0001
"Non-Repaired" Valves
LtGENOI A = 1 DBS* B = 2 0HS, ETC.
perfect repair
(100% reduction)
A
AA
0.0003 0.0010 0.0030 0.0100 0.0300 0*1000
BEFORE MAINTENANCE LEAK RATE
0.5000 1.0000 3.0000
Figure 5-2. Before Minus After Maintenance Leak Rate -
Valves Screening ^ 10,000 ppmv After Maintenance
-------
TABLE 5-3. WEIGHTED PERCENT REDUCTION IN MASS EMISSIONS FOR VALVES
SCREENING > 10,000 ppmv IMMEDIATELY BEFORE MAINTENANCE
Source Category
Mean Emissions (Ib/hr/source)
Mean Emissions
Reduction
Weighted Percent
Reduction (%) Before Maintenance After Maintenance (Ib/hr/source)
Sources repaired'
(28 valves)
97.7
(95, 100)
0.09
(0, 0.2)
0.0002
(0, 0.02)
0.088
(-0.007, 0.2)
Sources not repaired
(69 valves)
62.6
(41, 85)
0.10
(0.04, 0.2)
0.038
(0.02, 0.05)
0.062
(0.006, 0.12)
Total
(97 valves)
70.1
(46, 95)
1) Screening < 10,000 ppmv Immediately after maintenance
Note: leak rates measured before and after maintenance
Note: estimates are reported with approximate 95% confidence interval
-------
SECTION 6
IMPACT ON LEAK FREQUENCY ESTIMATES OF
APPLYING CHEMICAL RESPONSE ADJUSTMENTS
The goal of the analysis in this section was to investigate the effect of
applying chemical(s) specific response adjustments to the OVA readings to
estimate the frequency of leaks from SOCMI process units. This was accom-
plished by calculating adjusted screening values based on the original screen-
ing value and chemical response factor corrections. For the purposes of this
study a source is said to be leaking if its screening value is ^ 10,000 pprov.
Three different techniques were used to adjust the original OVA screening
value:
1) the original OVA reading adjusted for the associated OVA
response relationship of the primary chemical compound in
the line (see Section 7 for more detail),
2) weighted logarithmic average of response of primary and
secondary chemicals (see Section 7 for more detail),
3) weighted arithmetic average of response of primary and
secondary chemicals (see Section 7 for more detail).
The percent of valves leaking was calculated for each of the three
estimates for both gas and light liquid services. The three estimates were
found to be similar to the leak frequency estimate based on the original
screening value.
*
It should be noted that the total number of valves used in this analysis
may not match totals from previous sections of this report. The reason is
that in certain process units a dilution probe was not used. This resulted
in 119 sources having a recorded OVA reading of 10,001 ppnrv (indicating a
concentration above 10,000 ppmv). Many of the adjustments of these data
125
-------
resulted in estimates just below 10,000 (i.e., -9,997) ppmv). Therefore,
all sources with OVA readings equal to 10,001 ppmv were excluded from the
analysis. These 119 observations came from the following process types:
Acrylonitrile - 37 observations,
Chlorinated Ethanes - 11 observations,
Ethylene Dichloride - 28 observations,
• Formaldehyde - 1 observation, and
Vinyl Chloride Monomer - 42 observations.
Because of these deletions, the percent leaking estimates will have a small
negative bias. However, the comparison of the four estimates is still valid
since the relative sizes of the estimates is the important aspect to be
evaluated.
SUMMARY OF FOUR LEAK FREQUENCY ESTIMATES BY PRIMARY CHEMICAL
The percent leaking estimates resulting from the three adjustment
methods are presented in Tables 6-1 and 6-2. Also included is the percent
leaking estimate from unadjusted OVA readings for comparison purposes. As
seen in the tables, the three leak frequency estimates based on adjusted
screening values are similar to the unadjusted estimates.
SUMMARY OF FOUR LEAK FREQUENCY ESTIMATES BY PROCESS TYPE
The main question to be answered by this investigation is, "If the OVA
readings for a given process unit are adjusted for chemical response, will
significantly different estimates of the percent of leaking sources result?"
From the summarizations shown in Tables 6-3 and 6-4, it is evident that there
are no drastic changes in the estimates of percent leaking. However, there
is a general trend for a small reduction in the estimated frequencies.
To show the relationship between OVA readings and Method 1 estimates,
126
-------
plots of these two variables are shown for specific process types in Figures
6-1 through 6-6. The effects of specific chemicals with a process type can
be seen as straight lines. This is especially apparent in Figures 6-3 and
6-4. The northwest quadrant of these plots indicate valves where the original
screening value was below 10,000 ppmv and the Method 1 estimates are ^
10,000 ppmv. The southeast quadrant represents the opposite situation. The
other two quadrants indicate no change in the leak designation for those
valves.
For the high leaking processes the adjustments to the gas service valves
result in consistently lower percent leaking estimates. These estimates
are approximately 3 percentage points lower. The estimates in all other cases
are almost indistinguishable from the unadjusted estimate.
127
-------
TABLE 6-1. PERCENT LEAKING ESTIMATES FOR VALVES IN LIGHT LIQUID SERVICE
N>
oo
Chemical
Propylene
Ethane
Ethylene
Methane
Benzene
Methyl pthyl Ketone
Sec Butyl Alcohol
Hydrocarbons-Cs
Acetone
Methanol
Acetic Acid
Cumene
Acetaldehyde
Tr ichloroethylene
Vinyl Acetate
Methyl Methacrylate
Perchloroethylene
1,1,2 Tr Ichloroethane
1,2 Ethylene Dlchloride
Acrylonltrlle
Vinyl Chloride
Phenol
ct-Methyl Styrene
Acetone Cyanohydrln
Other Chemicals
TOTAL
OVA
Response
Factor @
10,000 ppmv
Responae
0^80
0.65
0.70
1.00
0.29
0.60
0.76
0.52
0.80
4.39
1.60
1.87
1.14
0.95
1.30
0.99
2.97
1.25
0,95
0.97
0.80
AAA4
113.9
3.51
Number
Screened
1583
328
1230
205
536
425
202
323
209
373
1162
773
456
267
973
393
599
911
2777
1120
607
594
326
191
15 JO
18,133s
Percent Leaking
Based on
OVA. Readings
Number
Leaking
467
92
321
36
49
23
10
a
5
4
6
4
2
1
3
1
1
1
0
0
0
0
0
0
67
1101
Percent
Leaking
29.50
28.05
26.10
17.56
9.14
5.41
4.95
2.48
2.39
1.07
0.52
0.52
0.44
0.37
0.31
0.25
0.17
0.11
0
0
0
0
0
0
4.27
6.07
Percent Leaking
Baaed on Method 1
Adjustments1
Number
Leaking
417
70
271
36
26
14
8
7
4
11
7
9
2
0
3
0
6
1
0
0
0
1
3
0
69
965
Percent
Leaking
26.34
21.34
22,03
17.56
4.85
3.29
3.96
2.17
1.91
2.95
0.60
1.16
0.44
0
0.31
0
1.00
0.11
0
0
0
0.17
0.92
0
4.39
5.32
Percent Leaking
Baaed on Method 2
Adjustments2
Number
Leaking
446
75
273
47
30
10
7
7
5
11
8
11
4
1
4
1
6
1
5
1
0
2
0
0
68
1023
Percent
Leaking
28.17
22.87
22.20
22.93
5.60
2.35
3.47
2.17
2.39
2.95
0.69
1.42
0.88
0.37
0.41
0.25
1.00
0.11
0.18
0.09
0
0.34
0
0
4.33
5.64
Percent Leaking
Baaed on Method 3
Adjustments3
Number
Leaking
431
76
2^3
38
28
13
7
6
4
11
6
9
2
0
3
1
5
1
0
0
0
0
1
0
65
990
Percent
Leaking
27.23
23,17
23.01
18.54
5,22
3,06
3,47
1,86
1.91
2,95
0,52
1,16
0,44
0
0.31
0.25
0.83
0,11
0
0
0
0
0.31
0
4.14
5.46
?S°3 i I" ^ a?Ju't[n*nt to1the °VA read1"8 baBed °n the «sponee of the primary chemical in the line.
Method 2 la the mixed chemical weighted logarithmic average technique.
Method 3 ia the mixed chemical weighted average technique.
A response of 10,000 ppmv for Phenol was experimentally unattainable.
74 aourcea with OVA Reading - 10,001 ppmv were excluded.
-------
TABLE 6-2. PERCENT LEAKING ESTIMATES FOR VALVES IN GAS SERVICE
ho
Chemical
Propylene
Benzene
Ethyl ene
Methane
Propane
Ethane
Methyl Ethyl Ketone
Acetaldehyde
Acetic Acid
1,2-Ethylene Dichlorlde
Acrylonitrile
Vinyl Acetate
Vinyl Chloride
Other Chemicals
TOTAL (
OVA
Response
Factor @
10,000 ppmv
Response
0.80
0.29
0.70
1.00
0.60
0.65
0.60
1.14 .
1.60
0.95
0.97
1.30
0.80
—
Percent
Baaed
Leaking
on
OVA. Read inns
Number
Screened
1119
33Z
3104
1849
145
379
116
179
125
521
287
272
96
850
93741*
Number
Leaking
198
53
468
232
18
• 35
7
' 4
1
0
0
0
0
41
1057
Percent
Leaking
17.69
15.96
15.08
12.55
12.41
9.23
6.03
2.23
0.80
0
0
0
0
4.82
11.28
Percent
Based on
Leaking
Method 1
Adjustments'
Number
Leaking
168
31
422
232
19
25
4
4
1
0
0
0
0
40
946
Percent
Leaking
15.01
9.34
13.60
12.55
13.10
6.60
3.45
2.23
0.80
0
0
0
0
5.29
10.09
Percent
Baaed on
Leaking
Method 2
Adjustments2
Number
Leaking
189
36
425
213
18
31
3
4
1
0
0
0
1
40
961
Percent
Leaking
16.89
10.84
13.69
11.52
12.41
8.18
2.59
2.23
0.80
0
0
0
1.04
4.71
10.25
Percent
Baaed an
Leaking
Method 3
Adjustments3
Number
Leaking
173
32
437
234
18
29
3
4
1
0
0
0
0
37
968
Percent
Leaking
15.46
9.64
14.08
12.66
12.41
7.65
2.59
2.23
0.80
0
0
0
0
4.35
10.33
1 Method 1 is the adjustment to the OVA reading based on the response of [he primary chemical in the line.
2 Method 2 is the mixed chemical weighted logarithmic average technique.
1 Method 3 la the mixed chemical weighted average technique.
* 45 sources with OVA Readings - 10,001 ppmv were excluded.
-------
TABLE 6-3. PERCENT LEAKING ESTIMATES FOR VALVES IN LIGHT LIQUID SERVICE BY PROCESS TYPE
CO
o
Percent Leaking
Based on
OVJL Read Inns
Process (unit #'B)
Ethylene (2,4,11)
Cumene (5,6)
Methyl Ethyl Ketone (31,32)
Acetaldehyde (33)
Vinyl Acetate (1,3)
Acetone/Phenol (12)
Chlorinated Ethanes (60,61,62)
Methyl Methacrylate (34)
1,2-Ethylene Dichlorlde (21,29)
Acrylonitrile (65,66)
Vinyl Chloride Monomer (20,28)
Formaldehyde (22)
Adiplc Acid (35,64)
TOTAL
Number
Screened
4121
762
671
551
2137
1818
1982
1058
2232
1466
1197
121
17
18,133
Number
Leaking
966
80
34
3
8
6
3
1
0
0
0
0
0
1101
Percent
Leaking
23.44
10.50
5.07
0.54
0.37
0.33
0.15
0.09
0
0
0
0
0
6.07
Percent Leaking
Baaed on Method 1
Adjustments
Number
Leaking
852
62
23
3
9
9
7
0
0
0
0
0
0
965
Percent
Leaking
20.67
8.14
3.43
0.54
0.42
0.50
0.35
0
0
0
0
0
0
5.32
Percent Leaking
Based on Method 2
Adjustments
Nunber
Leaking
895
69
18
4
12
9
8
1
4
2
1
0
0
1023
Percent
Leaking
21.72
9.06
2.6B
0.73
0.56
0.50
0.40
0.09
0.18
0.14
0.08
0
0
5.64
Percent Leaking
Based on Method 3
Adjustments
Number
Leaking
882
63
,21
2
9
6
6
1
0
0
0
0
0
990
Percent
Leaking
21.40
8.27
3.13
0.36
0.42
0.33
0.30
0.09
0
0
0
0
0
5.46
Method
1 la the adjustment to the OVA reading baaed on the response of the primary chemical in the line.
2 Method 2 is the mixed chemical weighted logarithmic average technique.
1 Method 3 is the mixed chemical weighted average technique,
-------
TABLE 6-4. PERCENT LEAKING ESTIMATES FOR VALVES IN GAS SERVICE BY PROCESS TYPE
Percent
Baaed
Leaking
on
OVA. Readings
Process (unit #'s)
Ethylene (2,4,11)
Cumene (5,6)
Methyl Ethyl Ketone (31,32)
Acetaldehyde (33)
Vinyl Acetate (1,3)
1,2-Ethylene Dlchloride (21,29)
Acrylonitrile (65,66)
Vinyl Chloride Monomer (20,28)
Methyl Mathacrylate (34)
Adiplc Acid (35,64)
Chlorinated Ethanes (60,61,62)
Formaldehyde (22)
Ace tone/ Phenol (12)
TOTAL
Mumber
Screened
6050
1443
207
178
949
397
387
382
190
95
48
40
8
9374
Number
Leaking
932
63
19
8
35
0
0
0
0
0
0
0
0
1057
Percent
Leaking
15.40
14.22
9.18
4.49
3.69
0
0
0
0
0
0
0
0
11.28
Percent
Based on
Leaking
Method 1
Adjustments'
Number
Leaking
849
45
13
8
31
0
0
0
0
0
0
0
0
946
Percent
Leaking
14,03
10.16
6.28
4.49
3.27
0
0
0
0
0
0
0
0
10.09
Percent
Based on
Leaking
Method 2
Adjustments2
Number
Leaking
856
49
12
8
33
1
1
0
0
0
0
1
0
961
Percent
Leaking
14,15
11.06
5.80
4.49
3.48
0.25
0.26
0
0
0
0
2.50
0
10.25
Percent
Based on
Leaking
Method 3
Adjustments1
Number
Leaking
873
44
11
8
32
0
0
0
0
0
0
0
0
968
Percent
Leaking
14.43
9.93
5.31
4.49
3.37
0
0
0
0
0
0
0
0
10.33
Method 1 id the adjustment to the OVA reading based on the response of the primary chemical in the line.
Method 2 is the mixed chemical weighted logarithmic average technique.
Method 3 is the mixed chemical weighted average technique.
-------
u>
N>
loouooon
1U 0 0 0 0 It
10000(1
AOOOO
loon
100
1
1 10
Legend: A * 1 OBS, B - 2 OBS, etc.
"•*•-•
ion
1000
OVtt
lUOOo
100000
1UUOOUO 10000000
Figure 6-1. OVA Reading vs. Method 1 Adjustment for
Cumene Process Valves in Gas Service
-------
I-1
10001)00!,
1UUOOOC
10001)
100P
100
10
*-••— 4_-*-'I<»te-*l»*' — .^**— •— • — ^l-*-.f.«—_»»_» — <-4->f«*
\ 10 100 1000
Legend: A * 1 DBS, B = 2 DBS, etc. f'VA
1UOOO
• --_ -f ••>.
100000
10000UO 100000UO
Figure 6-2. OVA Reading vs. Method 1 Adjustment for
Cumene Process Valves in Light Liquid Service
-------
1000 f) 000
jUUOOOu
10000(1
11) 0 0 0
1000
11)
1
1 10 100 1000
Legund: A " 1 OBS, B - 2 OBS, etc. OVA
100000
10000UU 1000QOUO
Figure 6-3. OVA Reading vs. Method 1 Adjustment for
Ethylene Process Valves in Gas Service
-------
LO
lOOUOOOf) •*•
1 0 U 0 0 0 0 +•
1UOOOO
ioono +
1001)
100
10 t
1
A Slhli
AARLKF VflZM
fJ?trj tiftM
PZG1 Dj^tVI
HIM rvQifi
IAL nt^iist:
AF A KfJFCVtDl
UntlAflAHA
PL
AH K> B
1 10
Legend: A = 1 DBS, B = 2 OBS, etc.
100
1000
lUOOf)
10 () 0 0 0
10UOOUO 10000DUO
Figure 6-4. OVA Reading vs. Method 1 Adjustment for
Ethylene Process Valves in Light Liquid Service
-------
r
1 U I) 0 0 0 0
1U 0 0 0II
iuoon
1000
100
A
AD
3 C A
U&C /A
fCF, A A
HEAHAA
A
Hl*C A A
A H AllA R
AA A
A
c ' „
1 10
Legend: A = 1 DBS, -B - 2 DBS, etc.
100
1POO
UVA
1UOOQ
100000
1QUUOUU 1QOOOUUO
Figure 6-5. OVA Reading vs. Method 1 Adjustment for
Vinyl Acetate Process Valves in Gas Service
-------
luouoooo •*
OJ
--J
1
1
h
ft
r
t
u
K
p
I*
M
V
"'I' H "Jo ioou iuooo""""ioDOOo" iouooou loooouuo
Legend: A = 1 DBS, 6 = 2 DBS, etc.
OVA Kt.«Ult|l»
Figure 6-6. OVA Reading vs. Method 1 Adjustment for
Vinyl Acetate Process Valves in Light Liquid Service
-------
-------
SECTION 7
STATISTICAL CONSIDERATIONS
This section discusses the assumptions and technical details of the
statistical methods employed in the analysis of data within this study.
STATISTICAL CATEGORICAL ANALYSIS USING FUNCAT (SECTION 3)
The Funcat procedure in SAS (a computer software system) was used to
test for significance in leak frequency between categories. This procedure
is. used in Section 3 to consider leak frequency as a function of line temper-
ature and pressure for valves in gas stream service.
The analysis is based on fitting a log-linear model to the cell frequen-
cies. The model is:
ln(F; :, ) = 9 + a. + Y; + ay, - where:
•Cj hi -C j A^j
F. -, = expected cell frequency of leaking or not leaking at each
level of temperature and pressure
0 = intercept term
a- = main effect of factor a at level - (in this case, temperature)
••C A^.
Y- = main effect of factor y at level . (in this case, pressure)
o^Y;; = interaction (combined effects) of temperature and pressure
*-j
The program tests the significance of the main effects and interaction
via x2 tests. The resulting analysis tables and interpretations are similar
to analysis of variance tables and their interpretations.
138
-------
The following table is typical of the form of the output from Funcat
Analysis.
Source elf CHI-SQUARE P_
Temperature 3 54.35 0.0001
Pressure 3 252.31 0.0001
Interaction 9 39.69 0.0001
In this example the main effects, temperature and pressure, are
significant as is the interaction of these two variables. The P_ statistics
is the probability of making an incorrect significance statement. The
interaction, or combined effect, can be seen graphically when either pressure
or temperature is plotted against percent leaking, with a separate line
drawn for each level of the other variable. Where there is significant
interaction, the lines will be non-parallel.
CHI-SQUARE TEST FOR INDEPENDENCE (SECTION 3)
The two-way chi-square test is a technique for testing that two
characteristics are independent. Here the term "independent" means the
distribution of one characteristic should be the same regardless of the
level of the other characteristic. This test is used in Section 3.
When there are two levels of both variables in the two-way classifi-
cation, the computational formula for testing the hypothesis of independence.
is:
X2 =
N AD-BC -j)
(A+B) (C+D) (A+C) (B+D)
where the letters A through D refer to the cell frequencies, N is the total
number of observations and the data is tabulated in a 2 x 2 table as shown:
139
-------
VARIABLE I
VARIABLE II
A
C
B
D
The degrees of freedom for the x2 calculated from this formula is one. A x
value which exceeds the tabulated value (the specified probability (P) point
of a chi-squre distribution) indicates a dependence of one variable on the
other.
CONFIDENCE INTERVALS FOR PERCENT SOURCES LEAKING (SECTION 3)
Confidence intervals for the percent of leaking sources were computed
using the Binomal Distribution. The Binomial is used to model data when a
random sample is selected and each item is classified into one of two cate-
gories (leaking or non-leaking here). Exact confidence limits (level 1-a)
for the estimate of percent leaking can be obtained by iteration, solving for
PT in
n
X
i-k
(1 - P,) = —r for the lower limit and for P in
L • 2 u
k
I
i=o
(1 - P ) = —r for the upper limit,
where n = number of sources screened and k = number of leaking sources.
Tables of these solutions (Reference 6), available for most cases, were used
to develop 95% confidence intervals reported in Section 3.
140
-------
SCREENING VALUE DISTRIBUTIONS (SECTION 4)
In order to utilize the results of previous work on the estimation of
mass emissions over a range of screening values (Reference 8) » it was neces-
sary to confirm that the screening values followed a distribution close to
lognormal in form. Summary statistics for Log (OVA screening value) were
e
generated, including coefficients of skewness and kurtosis, for cumene units,
ethylene units, vinyl acetate units, and for each source type and service.
From an earlier study (Reference 2) , it was decided that since the detection
limit of the OVA is approximately 10 ppmv, that this number would be used to
define an emitting (not leaking) source. Separate statistics for screening
values below 10 ppmv and for screening values between 10 and 100,000 ppmv were
generated to evaluate the effect of a larger-than-expected number of observa-
tions at 100,000 ppmv. The patterns of skewness and kurtosis were similar in
both cases. An empirical approach was taken in the development of the screen-
ing value distributions and their confidence intervals for comparison with the
lognormal models (described later) .
Chi-square tests were performed to compare the percentage of each
screening value category (processes and units within process by source type
and service) > 10 ppmv. The statistic computed was
X2(r-l,d.f.) =
All
Categories
where
3, comparing 3 processes
2 or 3 » comparing 2 or 3 units ,
within a process
0 = Observed number of sources < 10 ppmv.
or ^ 10 ppmv., for each process or unit,
141
-------
E = Expected number of sources < 10 ppmv.
or ^ 10 ppmv., for each process or unit, and
d.f. = the degrees of freedom.
The empirical cumulative distribution function (CDF) of screening
values, defined by
T?f \ Number of sources screening ^ xn
Total number of sources
for each category (process by source type by service) was computed. The
curves displayed in Figures 4-2 through 4-10 are of the reverse cumulative
distribution functions (RCDF):
s\
(1-F (x0)) x 100 vs. Log10(x0)
showing the percentage of sources screening greater than given value XQ.
Confidence limits for the RCDF were constructed using a Kolmogorov
2-sided critical value, w (using tables from Reference 4). Upper and lower
'approximate 95 percent confidence limits for F(x) (UCL and LCL, respectively)
were obtained using the two following equations:
F(x) + w, if F(x) + w < 1
^
^.0 » if F(X) + w > 1
and
( , _ J F(x) - w, if F(x) - w > 0
* W
, if F(x) - w < 0
142
-------
where w is the tabulated critical value and n is the number of sources
screened in the given category. The resulting limits for the RCDF are
UCL(x) =
x 100
and
LCL-(x) = (l-Fu(x)) x 100,
A lognormal distribution was used to model the distribution of screening
values greater than 10 ppmv. This distribution has the property that when
the original data are transformed by taking natural logarithms, the trans-
formed data will follow a normal distribution. The lognormal distribution is
often appropriate when the standard error of an individual value is propor-
tional to the magnitude of the value. The form of the lognormal distribution
is as follows:
f(x) =
[-
x -
X
2a2
G/2?
for 0 > x > °°
for x < 0 .
143
-------
In order to develop cumulative screening value distribution curves,
the "non-emitting" sources (with screening values less than 10 ppmv) also
had to be modeled. A mixed distribution, specifically a lognormal distribu-
tion with a discrete probability mass at 0, was used for this purpose. Let-
ting p equal the fraction of non-emitting sources in the population, this
mixed-lognormal distribution has the following form:
f(x) =
fi ^
(1-p) exp -
Mean = (1-p) exp
+
for 10 < x < «>
for 0 < x < 10
for x < 0
Another set of curves (4-22a through 4-30a) contains the estimated
cumulative distribution of log screening values. The curves show 100 percent
minus the cumulative percent, or the estimated percent of sources which would
have screening values greater than any particular screening value. These
cumulative distribution functions were estimated by fitting a lognormal
distribution, as described above, to the screening data and then generating
the cumulative distribution.
There was some difficulty in fitting the lognormal distribution to the
screening values. Figure 4-1 shows a typical histogram of log screening
values for valves in gas service. The histogram appears to approximate a
normal distribution adequately up to 100,000 ppmv (5.0 on Iog10 scale). The
spike at 100,000 ppmv was due to the inability of the screening device to
measure beyond 100,000 without a modification to the dilution probe. The
modified dilution probe was used in only a few cases in the screening process
during this program.
144
-------
To overcome the bias caused by this spike, only log screening values
less than 5.0 were used to estimate the parameters of this distribution.
Formulas from "censored" normal distribution theory (discussed in Reference
3) were then used to arrive at unbiased estimates of the entire distribution.
These estimates were used to generate the cumulative distribution function
for each source type/process stream grouping.
Confidence intervals for these cumulative-functions were obtained using
the Binomial Distribution. The 95 percent confidence interval for individual
probabilities were approximated using
± 1.96 [p(l - p)/n]
I/ 2
where p is the estimated cumulative percent and n is the number of screening
values for each particular source type and stream group.
The estimated lognormal cumulative distribution functions were compared
with the empirical distribution function and appeared to fit the data reason-
ably well. Figures 4-2 through 4-10 show the lognormal and empirical dis-
tributions for the source type and service classifications. Discrepancies
were found at the 100,000 ppmv screening value (5.0 log screening value)
in almost all cases, but this was to be expected since the sample function
had a big jump at this point.
EMISSION FACTOR DEVELOPMENT (SECTION 4)
Predicted log leak rates were generated for all sources with screening
values greater than 10 ppmv, using the prediction equations (Reference 2)
developed from modeling the available measured leak rates with asso-
ciated OVA screening values. Emission factors were estimated from the
predicted leak rates using
(leak rate) = a + 6 [Log10 (OVA Value)] + Z (standard error)
where a and 3 are model parameters developed in the maintenance study
145
-------
(Reference 2) Z is a standard normal random number and the standard error
is associated with the prediction equation. Because the true leak rate/
screening relationship is unknown, there is a potential bias introduced when
these predicted leak rates are used in developing emission factors. This
potential bias was taken into account in developing confidence intervals
discussed below.
As described in the previous subsection, a lognormal distribution was
used to model the distribution of leak rates for emitting sources (i.e.,
sources with a screening value > ppmv). Coefficients of skewness and kurtosis
for Log (leak rates) were computed and histograms examined for normality.
This assumption is adequate for the generated emissions data. To account for
the non-leaking sources, a mixed distribution with a discrete probability mass
at zero was fit to the data. The precise form of this distribution was given
earlier in this section. The best, unbiased estimator of the population
mean emission rate from this distribution is:
- *
m = C(l-p) exp(y)] g
2 / *
where
g —— - bias correction factor (discussed in detail in Reference 8).
Confidence intervals for the percent of sources screening > 10 ppmv
were computed using the Binomial Distribution. Binomial Confidence Interval
tables, available for most cases, were used for computing 97.5 percent con-
fidence intervals which were then used in developing confidence intervals for
emission factors. The 97.5 percent was selected so that approximate 95
percent confidence intervals for emission factors would result when the
estimated percent leaking was combined with the estimated mean leak rate
(0.975 x 0.975 = 0.95).
146
-------
The 97.5 percent confidence intervals were computed for the average, y,
of the Log leak rate estimates using:
2 i/2
= lower limit = y - 2.24 [s2/(n - r)]
and
2 1/2
G = upper limit = y + 2.24 [s2/(n - r)]
where
f}
s = the variance of the log leak rate estimates and
e
n - r)= the number of leaking sources.
Then confidence intervals for the mean leak rate (emission factor estimate)
was computed using
C' = lower limit = exp[C»] g(s2/2)
X* )fj
and
" = upper limit = exp[C ] g(s2/2)
where
(-^—-) is the bias correction factor.
To obtain 95 percent confidence limits for the emission factors, the
confidence limits for the percent leaking and for the mean leak rate were
combined as follows:
lower 95% limit for emission factor = p (Cj)
upper 95% limit for emission factor = P (C')
rr • u u
147
-------
These confidence intervals are conservative in the sense that 95 percent
is a lower bound for the confidence coefficient for the intervals. The
intervals consider random sampling variation and random test error, with
no adjustments for potential bias in the estimation of the log rates.
An adjustment was applied to the emission factor confidence intervals
to account for the potential bias due to estimating leak rates. The stan-
dard error of the predicted average Iog10 leak rate (SEP) was calculated from
where n = number of leaking sources,
k = number of data pairs used to estimate
the prediction equation, and
x. = ith screening value (Log10 scale) used to estimate the prediction
equation.
The reported confidence intervals for the emission factors were widened by a
factor of
10
2(SEP)
A similar procedure was used to adjust the confidence intervals for the mean
emission estimate in Section 5.
As a Quality Control measure on the emission factor estimation, an
alternative approach to estimate emission factors was also explored. The
alternative model was
n
E.F. alternative = -i_ Y"* C-10logl° (leak rate)
n •—•. '
148
-------
where n = number of sources and C = bias correction factor. This approach
employs an estimator based on the arithmetic mean computed in arithmetic
scale of the leak rates. (The bias correction factor is actually part of
the predicted leak rate in the arithmetic scale as discussed in Reference 2) .
This alternative estimator is unbiased regardless of the distribution of leak
rates and avoids using the generated error term Z (standard error of estimate)
for the predicted leak rates. Comparison of the results of these approaches
is shown in Table 7-1.
Confidence limits for the alternative estimates of mean leak rates
were based on computing the mean leak rates for each of the two limiting
distributions of screening values given by the confidence bounds for the
empirical CDF as described earlier in this section. These bounds were
further adjusted to account for the potential bias in using predicted leak
rates. For the alternative estimates, the standard error of the mean (SEM)
was calculated from
SEM - CIO CLOS10 (leak rate)]' 1/2
where cr = residual error for the fitted prediction equation. The confidence
limits were adjusted by adding or subtracting 2.24 SEM to the upper or lower
confidence limits, respectively. The results (Table 7-1) are an attempt to
approximate 95 percent confidence limits for these alternative estimates.
Note that the lower confidence is zero in most cases for the alternative
estimate indicating that the ± SEM limits do not adequately reflect the
skewness of the distribution of the alternate estimates.
The emission factor estimates are not consistently higher or lower than
the quality control estimate.
149
-------
TABLE 7-1. COMPARISON OF EMISSION FACTORS WITH QUALITY CONTROL
ESTIMATES OF MEAN LEAK RATES FOR VALVES AND PUMP SEALS
Process
Source Type
Service
Emission Factor
(Ibs./hr./source)
Quality Control Estimate'
(Ibs./hr./source)
Cumene
Ethylene
Vinyl Acetate
Ui
o
Valves
Pump Seals
Valves
Pump Seals
Valves
Pump Seals
Gas -
Light Liquid
Light Liquid
Gas
Light Liquid
Light Liquid
Gas
Light Liquid
Light Liquid
0.011(0.003, 0.05)
0.0056(0.002, 0.02)
0.052(0.001', 2.7)
0.024(0.008, 0.07)
0.020(0.007, 0.06)
0.069(0.006, 0.8)
0.0046(0.001, 0.03)
0.0003(0.0001, 0.002)
0.0043(0.0001, 0.1)
0.0079(0, 0.02)
0.0061(0, 0.02)
0.030(0, 0.3)
0.010(0.004, 0.02)
0.013(0,008, 0.02)
0.085(0,0.7)
0.0027(0, 0.03)
0.0003 (0,0.005)
0.0051(0,0.06)
Emission factor reported with 95% confidence interval
r\
Quality Control estimate reported with approximate 95% confidence interval
based on estimate ±2 (standard error of the estimate)
-------
CUMULATIVE EMISSION FUNCTIONS
A cumulative function for the percentage of total mass emissions for
all sources screening greater than a given value was estimated by integrating
the leak/screening regression relationship over a lognormal distribution of
screening values. This function has the following form:
CF m J° C(10)Nx)Bl
where
So = selected upper screening value for integration,
C = log/arithmetic scale bias correction factor,
BO = log10 regression intercept term,
BI = Iog10 regression slope term,
u = mean of the log (screening values),
a2 - variance of the log (screening values),
x = screening values over which the integration is being done, and
CF = cumulative function described above in Ibs/hr
D = numerator of CF evaluated at S0 = 1,000,000
The form of the cumulative function can be simplified by algebraic
reduction and change of variables to obtain:
, ,.^.,-J - u - BiCT | /^ | Jin(1,000,000) - u - Bier'
CF —
-------
in each case. The log/log least-squares regression estimates were used for
the scale bias correction factor and for BO and BI. Division by the numer-
ator of the function evaluated at one million ppmv forced the function to
1.0 at one million ppmv. These function values were then subtracted from
1.0 and multiplied by 100.0 to obtain the functions shown in Figures 4-22£>
through 4-30b.
The estimated lognormal cumulative emissions functions were compared
with the empirical functions (discussed below) and found to adequately
approximate the data. Figures 4-13 through 4-21 show the lognormal and
empirical functions for the source type and service classifications.
The biggest discrepancies were near the 100,000 ppmv screening value where
the sample function has a big jump. This area is more critical for this
function than the cumulative distribution function since most of the emis-
sions are attributable to sources with screening values greater than 100,000
ppmv. It is important to note that very little screening data are available
with screening values greater than 100,000 ppmv. Thus, this portion of the
curve is based on extrapolations using models developed from screening values
less than 100,000 ppmv.
This cumulative function is a very complex nonlinear function of three
sample statistics. Due to the complexity of this function, it was not
possible to derive a closed-form analytical expression for the confidence
intervals. Thus, a Monte-Carlo computer method was used to generate the
confidence intervals.
This method involved regenerating the cumulative function 200 times.
Each time, the data collected in the project (the number of sources with
screening values greater than 10 ppmv) were regenerated, except with an
independent set of random variations. The distributional properties of the
leak rate and screening data were used in computing the required random
numbers.
152
-------
For each of the 200 trials, sample estimates of the three parameters
required to compute the cumulative function were computed. Then these
estimates were used to generate a new cumulative function. The one percent
lower result and the 99 percent upper result from the 200 trials for any
given screening value were then selected as approximate 95 percent confidence
limits for the population cumulative function.
Since these confidence limits address the uncertainty in the cumulative
function for the entire sampled population of a particular source type, they
are not necessarily applicable to a finite sample of sources in a particular
situation. The variation of this function depends on the number of sources
in a complex manner, so it is not possible to draw a general conclusion for
the effect of sample size.
Empirical functions computing the percentage of total mass emissions
for all sources screening greater than a given screening value were devel-
oped using the estimator
£
- x < x0
G (-xuJ
Z-^
all x
c • 10 Lo§1° (leak rate)
c . 10 Log10 (leak rate)
in addition to the approach based on the lognormal distribution discussed
^N
earlier. Note that the denominator of G(XO) is an expression for the total
mass emission used in the quality control check for estimating emission
factors.
A ^
Confidence bounds for G(XO) were obtained by evaluating G(x0) for the
screening value distributions corresponding to the confidence bounds shown
in Figures 4-13 through 4-21. Applying a standard approximation to calculate
153
-------
the variance of a ratio (Reference 5), the following expression for the stan-
/\
dard error of G(XQ) was derived:
SE = CT(Loge(10))
x
\A11
x
where A = 10
[Log10 (leak rate)]
LJ
x
\
A-
x> x x
A _^ A
x
-------
INCREASE IN MASS EMISSIONS DUE TO OCCURRENCE AND RECURRENCE (SECTION 5)
Increase (or reduction) of mass emissions reported in this study has
been expressed in two ways:
• Weighted percent increase (reduction)
• Mean increase (reduction)
The first measure of change in mass emissions has been discussed in detail in
Reference 2 as Weighted Percent Reduction (WPR), defined as the percent of
total emissions reduced due to maintenance:
n n
-__. _ (£ mass emissions before - Z mass emissions after)
WrK. — X 100%
n
I mass emissions before
(See Reference 2 also for a discussion of the development of confidence
intervals for this estimate.) Weighted percent increase (WPI) is defined as
WPI = -WPR, where "before" and "after" refer to before and after leak occur-
rence, recurrence, or maintenance, depending upon the application.
The mean increase is defined as the difference in mass of the
average before and after maintenance emissions from individual sources.
Confidence intervals for this measure of increase in mass emissions are
given by:
*
Mean increase ± £0.975 x SE,
SE = Standard deviation of increase// n , where n = number of sources.
Reference 2 provides details of the estimation of nonmeasured leak rates
from screening values via prediction equations appropriate to source type and
service type. The equations, as expressed in arithmetic scale, were applied
as discussed in Reference 2, except for sources in unit 1. In unit 1 there
was no information on service type for valves in the occurrence analysis.
155
-------
The following equation was used for these valves:
Predicted Leak Rate = 5.08 • 10 C-5'22 + °"67 Log"
-------
The other two adjustment methods used to estimate actual concentration take
into account the mixed nature of the chemical composition. The OVA response
to a mixture of compounds is intermediate to the individual responses to
each compound at the same concentration. Using this concept, one of the
mixed chemical adjustment methods (Method 3) used a weighted average of the
responses to estimate actual concentration. An estimate of the weighted
response is
where, R = the estimated weighted average response,
£~L
p. = the fraction of the mixture total concentration accounted for
i
by compound i (P. = C./C ),
a. = exp (A) with "A" from Brown, et al (1980) for component i,
i - '
b. = coefficient "B" from Brown, et al (1980) for component i,
i --
s, = parameters "SE" from Brown, et al (1980) for component i,
i " —
C_ = EC., the total concentration, and
C: = the concentration in the mixture of compound i.
The coefficients A, B, and SE can be found in Tables 5-169 and 5-170 of
Brown, ot_ al (1980) (Reference 7) for selected compounds.
The above discussion involves the prediction of an instrument response
when the actual concentration of mixture components are known. For this
study, the reverse is the case: the response is observed and it is desired
to estimate the total concentration of the constituents. Basically, this
cannot be done without some additional information. The compound identifi-
cation of the constituents must be known. If the constituent proportions
are also know, the total concentration can be computed assuming the above
model is correct. The total concentration (C ) is estimated by solving
equation 1.
157
-------
Equation 1 cannot be solved explicitly for total concentration. An
iterative solution is required. This can be done using the Newton-Raphson
method. Let
f«v>= E p±ai cibi e%s x- R
where R is the observed instrument response, and
f'(CT) -
.-l
Then the iteration formula is
C -C -f(C)/f(C),
A reasonable starting value C is R, the observed instrument response.
o
The other mixed chemical adjustment method (Method 2) used for estimat-
ing actual concentration was a weighted logarithmic average. In this case
(2)
log(RL) = Yl P± [log a± + %a| + b± log
where R^ is the estimated instrument response using a weighted logarithmic
average
In contrast to the previously given weighted arithmetic average model
(equation 1) , this weighted logarithmic average model (equation 2) has an
explicit solution for actual total concentration:
= exp
log R - ?.. (log a
158
-------
Both of the chemical mixture methods used the information on primary
and secondary chemicals and their percentage of the total concentration.
If their percentages did not total 100 percent, (i.e., there were other
chemical compounds in the line) the rest of the percentage was assigned a
response factor of 1.
159
-------
REFERENCES
1. Blacksmith, J. R., G. E, Harris, and G. J. Langley. Frequency of Leak
Occurrence for Fittings in Synthetic Organic Chemical Plant Process Units:
Final Report. EPA-600/2-81-003 (NTIS Ho. PB81-14566), U. S. Environmental
Protection Agency, Research Triangle Park, N. C., 1980.
2. Langley, G. J. and R. G. Wetherold. Evaluation of Maintenance for
Fugitive VOC Emissions Control: Final Report. EPA-600/S2-81-080, U. S.
Environmental Protection Agency, Cincinnati, Ohio, 1981.
3. Wetherold, R, G, and L. P. Provost. Emission Factors and Frequency of
Leak Occurrence for Fittings in Refinery Process Units. EPA-600/2-79-044,
U. S. Environmental Protection Agency, Washington, D. C., 1979.
4. Conover, "W. J. Practical Nonparametric Statistics. John "Wiley and Sons
Inc., New York, 1971.
5. Mood, A. M., F. A. Graybill, and D. C. Boes. Introduction to the Theory
of Statistics, Third Edition. McGraw-Hill Book Company, New York, 1974.
6. Beyer, W. H. (ed.). CRC Handbook of Tables for Probability and Statistics,
Second Edition. The Chemical Rubber Co., Cleveland, Ohio, 1968.
7. Brown, G. E., D. A. DuBose, W. R. Phillips, and G. E. Harris. Response
Factors of VOC Analyzers Calibrated with Methane for Selected Organic
Chemicals. EPA-600/2-81-002 (NTIS No. PB81-136194), U. S. Environmental
Protection Agency, Research Triangle Park, N.C., 1980.
8. Wetherold, R, G, and D. D, Rosebrook, Environmental Assessment of Atmo-
spheric Emissions from Petroleum Refinery. EPA-600/2-80-075a (NTI'S No.
PB80-225-253), U. S. Environmental Protection Agency, Research Triangle
Park, N. C., 1980.
160
-------
-------
APPENDIX A
SCREENING DATA SUMMARY
Section 3 of this report contains an analysis of screening data collected
on an earlier EPA project (Reference 1). This appendix gives detailed source
type groupings of this data. Table A-l gives the number of sources screened,
the number leaking and the percent leaking for each possible source type and
for each type of stream service, including heavy liquids.
161
-------
TABLE A-l. DATA SUMMARY OF LEAK FREQUENCIES FOR VARIOUS
SOURCES IN VARIOUS STREAM SERVICES
Source
Flange
Process Drain
Open-End Line
Agitator Seal
Relief Valve
Block Valve-
Gate Type
Block Valve-
Globe Type
•
Block Valve-
Plug Type
Block Valve-
Ball Type
Service
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
C-as
T.ighf Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Number
Screened
1450
2833
607
83
496
28
923
3605
477
7
8
1
84
68
3
6976
11017
2034
1 /. C
-ICC
129
440
2479
1031
1272
2732
251
Number
Leaking1
66
36
0
2
19
2
54
141
6
1
0
0
3
2
0
952
1059
9
L3
8
0
0
2'
0
18
4
4
Percent
Leaking1
4.5
1.3
0.0
2.4
3.8
7.1
5.8
3.9
1.3
14.3
0.0
0.0
3.6
2.9
o.o
13.6
9.6
0.4
10.3
1. 1
0.0
0.0 .
0.1
0.0
1.4
0.1
1.6
leaking defined as OVA reading >10,000 ppmv.
(continued)
162
-------
TABLE A-l. DATA SUMMARY OF LEAK FREQUENCIES FOR VARIOUS
SOURCES IN VARIOUS STREAM SERVICES (CONTINUED)
Source
Block Valve -
Butterfly Type
Block Valve -
Other Types
Control Valve -
Gate Type
Control Valve -
Globe Type
Control Valve -
Plug Type
Control Valve -
Ball Type
Control Valve - •
Butterfly Type
Control Valve -
Other Tir**es
On-Line Pump Seals
Single Mechanical-
Emission
Point at Seal
Service
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Number
Screened
160
157
8
275
378
35
61
182
27
207
417
107
10
91
0
15
33
3
91
34
6
•17
25
1
215
60
Number
Leaking
9
2
0
16
17
0
15
22
0
36
61
0
0
3
4
1
0
35
3
0
3
1
0
28
2
Percent
Leaking
5.6
1.3
0.0
5.8
4.5
0.0
24.6
12.1
0.0
17.4
14.6
0.0
0.0
3.0
26.7
3.0
0.0
38.5
8.9
0.0
17. 6
i.n
0.0
13.0
3.3
leaking defined as OVA reading ^10,000 ppmv.
(continued)
163
-------
TABLE A-l. DATA SUMMARY OF LEAK FREQUENCIES FOR VARIOUS
SOURCES IK VARIOUS STREAM SERVICES (CONTINUED)
Source
Service
Number
Screened
Number
Leaking,1
Percent
Leaking1
On-Line Pump Seals
Single Mechanical -
Emission
Point at Vent
On-Line Pump Seals
Single Mechanical -
Other
Emission Point
On-Line Pump Seals
Double Mechanical -
Emission
Point at Seal
On—Line Pump Seals
Double Mechanical -
Emission
Point at Vent
On-Line Pump Seals
Double Mechanical -
Other
Emission Point
On-Line Pump Seals
Single, Packed,
Emission
Point at Seal
On—Line Pump Seals
Single, Packed,
Emission
Point at Vent
On-Line Pump Seals
Single, Packed,
Other
Emission Point
On-Line Pump Seals
Sealess Pumps
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light-Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
24
0
30
0
92
2
13
0
0.0
6.7
14.1
0.0
33.3
0.0
0.0
0.0
1 Leaking defined as OVA reading 3*10,000 ppmv.
(continued)
164
-------
TABLE A-l. DATA SUMMARY OF LEAK FREQUENCIES FOR VARIOUS
SOURCES IN VARIOUS STREAM SERVICES (CONTINUED)
Source
Service
Number
Screened
Number
Leakingl
Percent
Leaking1
Off-Line Pump Seals
Single Mechanical -
Emission
Point at Seal
Off-Line Pump Seals
Single Mechanical -
Emission
Point at Vent
Of f-Lin& Pump Seals
Single Mechanical -
Other
Emission Point
Off-Line Pump Seals
Double Mechanical -
Emission
Point at Seal
Off-Line Pump Seals
Double Mechanical -
Other
Off-Line Pump Seal
Single, Packed
Emiss ion
Point at Seal
Off-Line Pump Seals,
Single, Packed
Emission
Point at Vent
Off-Line Pump Seals,
Single, Packed
Other
Emission Point
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
139
24
17
1
86
1
19
6.5
0.0
0.0
0.0
0.0
3.5
0.0
0.0
0.0
0.0
0.0
1Leaking defined as OVA reading 10,000 ppmv.
(continued)
165
-------
TABLE A-l. DATA SUMMARY OF LEAK FREQUENCIES FOR VARIOUS
SOURCES IN VARIOUS STREAM SERVICES (CONTINUED)
Source
Service
Number
Screened
Number
Leaking1
Percent
Leaking!
Off-Line Pump Selas,
Sealess Pumps
On-Line Compressor
Seals, Single,
Mechanical, Emission
Point at Seal
Light Liquid
Heavy Liquid
Gas
On-Line Compressor
Seals, Single,
Mechanical, Emission
Point at Vent Gas
On-Line Compressor
Seals, Single,
Mechanical, Other
Emission Point
Gas
33.3
On-Line Compressor
Seals, Double,
Mechanical, Emission
Point at Seal Gas
0.0
On—Line Compressor
Seals, Double,
Point it Vent
0.0
On—Line Compressor
Seals, Double,
Mechanical, Other
Emission Point
Gas
0.0
On—Line Compressor
Seals, Single,
Packed, Emission
Point at Seal
Gas
0.0
On-Line Compressor
Seals, Single,
Packed, Emission
Point at Seal
Gas
100.0
'Leaking defined as OVA reading ^10,000 ppmv.
(continued)
166
-------
TABLE A-l. DATA SUMMARY OF LEAK FREQUENCIES FOR VARIOUS
SOURCES IN VARIOUS STREAM SERVICES (CONTINUED)
Hunfoer Humber Percent
jaource Service Screened Leaking * Leaking
On-Line Compressor
Seals, Single,
Packed, Other
Emission Point Gas 1 0 ' 0.0
Other Source Types Gas 19 3 15.8
Light Liquid 33 2 6.1
Heavy Liquid 2 0 0.0
1Leaking defined as OVA reading ^10,000 ppmv.
167
-------
-------
APPENDIX B
DETAILED INFORMATION ON
LINE TEMPERATURE AND LINE PRESSURE
Section 3 contains an analysis of the effect of temperature and
pressure on leak frequency. This appendix contains statistical information
on temperature and pressure.
Tables B—1 to B-3 contain summary statistics for line pressure and line
temperature for gas, light liquid, and heavy liquid stream services. Separate
values are given for each of the major source types. Differences between the
types of processes, ethylene versus high leaking, and also between the groups
of primary chemicals in the line can be seen at this stage. For example, the
average line temperature for the high leaking process units appears to be much
higher than that for the ethylene units. The minimum temperature for the
ethylenes is also much lower. Line pressure seems to differ more by type of
chemical in the line. The heavy liquids are not broken down by primary
material groups in the line since they had a low leak frequency.
Although line temperature and pressure were recorded as continuous
variables, they are grouped for evaluating leak frequency. Tables B-4 to B-16
#
show the number screened, percent screened, number leaking, and percent
leaking at different levels of temperature and pressure. This information is
given for ethylene process units and for high leaking process units and also
for primary "-material groups for all source types but pump seals. Possible
reasons for some of the differences in leak frequencies for the different
categories can be seen from these tables. None of the high leaking group
sources are at very low temperatures. This group also has some screening
values for each source type at the higher temperatures. The ethylene group
exhibits a different distribution of temperatures. There are some values in
168
-------
the very low temperature group; and on the average, the temperatures found in
the ethylene unit sources are lower. If the data were not separated into
these groups, differences that were actually attributable to the type of
process unit might appear to be due to line temperature.
Figures B-l to B-4 show the distributions of the sources screened as a
function of line temperature and line pressure for valves in gas and light
liquid service.
169
-------
TABLE B-l. SUMMARY STATISTICS FOR LINE TEMPERATURE
AND LINE PRESSURE FOR GAS SERVICE
H
^J
O
High Leaking
Variable
Line Temperature
(°F)
,.
Line Pressure
(psig)
PRIMARY CHEMICAL Group * :
Type Statistics
Valves Average
Standard Deviation
Minimum
Maximum
Flanges Average
Standard Deviation
Minimum
Maximum
Open Ended Average
Lines Standard Deviation
Minimum
Maximum
Valves Average
Standard Deviation
Minimum
Max imum
Flanges Average
t Standard Deviation
Minimum
Maximum
Open Ended Average
Lines Standard Deviation
Minimum
Maximum
Group 5
210.8
153.2
30
825
271.3
146.8
30
800
128.2
82.1
30
392
184.7
167.4
-10
600
273.6
184.9
-9
600
132.0
142.5
0
600
Process Units
Group 6
217.0
195.4
20
1000
235.3
228.5
20
1000
218.1
190.6
20
1000
56.4
99.5
-15
650
37.7
85.0
-15
590
52.1
92.9
-15
450
Ethvlene Process Units
Group 1 Group 2
60.7
101.8
-267
1570
73.3
94.0 v •-
-267
750
46.1
74.7
-267 - .
720
166.7
178.7
0
1050
184.6
160.0
0
805
120.6
160.3
1
805
lSee Figure 3-2 for explanation of groups.
-------
TABLE B-2. SUMMARY STATISTICS FOR LINE TEMPERATURE
AND LINE PRESSURE FOR LIGHT LIQUID SERVICE
Variable
Line Temperature
<8F)
Line Pressure
(psig)
'
PRIMARY CHEMICAL Group1:
Type Statistics
Valves Average
Standard Deviation
Minimum
Maximum
Pump Seals Average
Standard Deviation
Minimum
Maximum
Flanges Average
Standard Deviation
Minimum
Maximum
Open Ended Average
Lines Standard Deviation
Minimum
Maximum
Valves Average
Standard Deviation
Minimum
Maximum
Pump Seals Average
Standard Deviation
Minimum
Maximum
Flanges Average
Standard Deviation
Minimum
Maximum
Open Ended Average
Lines Standard Deviation
Minimum
Maximum
High Leaking
Group 7
147.2
94.1
20
500
129.5
82.4
32
540
165
93
20
500
142.8
100.6
30
500
161.0
190.5
-10
740
116.2
152.8
0
720
247,3
210.4
-9
740
123.7
171.1
0
740
Processing Units
Group 8
145.5
90.8
15
1000
133.8
68.2
32
345
148.3
102.8
30
1000
137.2
83
20
1000
80.2
78.5
-20
700
79.3
77.5
0
700
70,6
77.0
-20
700
66,9
67.4
-20
700
Ethylene Process Units
Group 3
20.3
86.4
-267
190
8.0
73.6
-145
118
44.7
71.9
-212
190
44.4
84.5
-267
190
372.0
368.1
0
2270
512.2
427.6
80
1960
380.7
396.3
0
2270
379.4
383,9
0
2270
Group 4
129-4
57.2
40
235
112.6
70.8
40
235
128.4
49.6
40
235
154,6
74.7
40
235
101.8
111.6
2
500
65.1
57.4
2
165
79.2
79.8
2
500
75.5
96.6
0
500
*See Figure 3-2 for explanation of groups.
-------
TABLE B-3. SUMMARY STATISTICS FOR LINE TEMPERATURE AND LINE PRESSURE IN HEAVY
LIQUID SERVICE WITHIN HIGH AND ETHYLENE PROCESS UNITS
Variable Type Statistics
Line Temperature Valves Average
(°F) Standard Deviation
Minimum
Maximum
Pump Seals Average
Standard Deviation
Minimum
Maximum
Flanges Average
Standard Deviation
Minimum
Maximum
Open Ended Average
Lines Standard Deviation
Minimum
Maximum
Line Pressure Valves Average
(psig) Standard Deviation
Minimum
Maximum
Pump Seals Average
Standard Deviation
Minimum
Maximum
Flanges Average
Standard Deviation
Minimum
Maximum
Open Ended Average
Lines Standard Deviation
Minimum
Maximum
High Leaking
Process Units
228.8
177.7
60
600
153.6
146.8
72
460
219.9
63.4
60
500
93.9
41.1
60
260
58.5
55.4
1
230
78.6
10.2
62
92
63.9
67.3
1
230
50.0
40.4
1
120
Ethylene Leaking
Process Units
128.1
70.0
25
370
168.1
60.5
90
300
124.8
63.8
25
300
156.6
90.1
60
30
97.0
^ 110.4
0
540
48.4
57.7
0
170
89.5
112.9
0
320
68.8
95.9
0
480
-------
TABLE B-4. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE
ON PERCENT LEAKING FOR VALVES IN GAS SERVICE
WITHIN ETHYLENE PROCESS UNITS
Pressure (psie)
-15 - -1
0 - 49
50 - 99
100 - 149
150 - 199
200 - 249
250 - 299
300 - 349
350 - 399
400 - 449
450 - 499
500 - 549
550 - 999
1000 - 1050
Number
Screened
0
2123
1072
681
157
321
502
196
144
94
267
316
167
4
Group
Percent
of
Total
Screened
0
35.1
17.8
11.3
2.6
5.3
8.3
3.2
2.4
1.6
4.4
5.2
2.8
0.1
1s PRIMARY CHEMICALS
Number
Leaking
_
94
171
110
36
73
136
51
54
20
54
91
37
2
Percent
Leaking
_
4.4
16.0
16.2
22.9
22.7
27.1
26.0
37.5
21.3
20.2
28.9
22.2
50.0
TOTAL
6043
929
15.4
Temperature p
-267 -
0 -
50 °F -
100 °F -
150 °F -
200°F -
250°F -
300 °F -
350°F -
400 °F -
TOTAL
-1
49
99
149
199
249
299
349
399
1570
998
1452
2035
1011
373
78
32
4
19
48
6050
16.5
24.0
33.6
16.7
6.2
1.3
0.5
0.1
0.3
0.8
140
236
327
111
72
27
8
0
7
4
932
14.0
16.2
16.1
11.0
19.3
34.6
25.0
0.0
36.8
8.3
15.4
JSee Figure 3-2 for explanation of groups.
173
-------
TABLE B-5. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE
ON PERCENT LEAKING FOR VALVES IN GAS SERVICE
WITHIN HIGH LEAKING PROCESS UNITS BY CHEMICAL
GROUP
Group1:
Pressure (pslg)
-15 -
0 -
50 -
100 -
150 -
200 -
250 -
300 -
350 -
400 -
450 -
500 -
550 -
1000 -
-1
49
99
149
199
249
299
349
399
449
499
549
999
1050
Hunter
Screened
5
203
282
267
53
87
136
10
15
0
80
37
53
0
'.Group 31 PRIMARY CHEMICALS
Percent
of
Total
Screened
0.4
16.5
23.0
21.7
4.3
7.1
11.1
0.8
1.2
0.0
6.5
3.0
4.3
0.0
Number
Leaking
0
9
31
30
6
9
13
0
4
-
21
9
14
-
Percent
Leaking
O.D
4.4
11.0
11.2
11.3
10.3
9.5
0.3
26.7
-
26.2
24.3
26.4
~
Number
Screened
66
1201
141
136
34
18
20
84
13
13
6
18
8
0
Group -41 PRIMARY CHEMICALS
Percent
of
Total
Screened
3.8
68.3
8.0
7.7
1.9
1.0
1.1
4.8
0.7
0.7
0.3
1.0
0.5
0.0
Number
Leaking
0
7
2
12
1
0
0
0
0
0
0
0
0
—
Percent
Leaking
0.0
0.6
1.4
8.8
2.9
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
~
TOTAL
1228
146
11.9
1758
22
1.2
Temperature (Pp>
-267 -
0 -
50 °F -
100 "F -
150 °F -
200 8F -
250°F -
300 °F -
350 "F -
400°F -
TOTAL
-1
49
99
149
199
249
299
349
399
1570
0
12
243
355
127
43
77
109
113
146
1225
0.0
1.0
19.8
29,0
10.4
3.5
6.3
8.9
9.2
11.9
-
1
27
31
9
2
8
14
22
31
145
-
8.3
11.1
8.7
7.1
4.6
10.4
12.8
19.5
21.2
11.8
0
45
335
454
369
156
113
63
7
216
1758
0.0
2.6
19.1
25.8
21.0
8.9
6.4
3.6
0.4
12.3
-
0
4
2
0
2
1
1
0
12
22
-
0.0
1.2
0.4
0.0
1.3
0.9
1.6
0.0
5.6
1.2
See Figure 3-2 for explanation of groups.
-------
TABLE B-6. EFFECTS OF LINE TEMPERATURE AND LINE
PRESSURE ON PERCENT LEAKING FOR VALVES
IN LIGHT LIQUID SERVICE WITHIN ETHYLENE
PROCESS UNITS BY CHEMICAL GROUP
GROUP 31 PRIMARY CHEMICALS
GROUP 4 PRIMARY CHEMICALS
Ui
Pressure (pale)
-15 -
50 -
50 -
100 -
150 -
200 -
250 .-
300 -
350 -
400 -
450
500 -
550 -
1000 -
-1
49
99
149
199
249
299
349
399
449
. 499
549
999
1050
TOTAL
Number
Screened
0
480
363
226
229
138
477
141
310
109
273
242
312
211
3511
Percent
of
Total
Screened
0.0
13,7
10.3
6.4
6.5
3.9
13.6
4.0
8.8
3.1
7.8
6.9
8.9
6.0
Number
Leaking
_
57
66
54
56
55
156
43
94
28
59
93
97
99
957
Per :ent
Le iking
111.9
18. 2
11,9
>4.4
VI.9
C1.7
Jtl.5
311.3
. 25.7
.21.6
38.4
31.1
46.9
2/..3
Number
Screened
0
173
215
107
54
20
0
0
0
0
0
38
0
_JL
607
Percent
of
Total
Screened
0.0
28.5
. 35.4
17.6
8.9 ,
3.3
0.0
0.0
0.0
0.0
0.0
6.3
0.0
0.0
Number
Leaking
0
4
1
0
3
_
_
-
_
_
1
-
-
9
•
Percent
Leaking
0.0
1.9
0.9
0.0
15.0
_
_
_
_
_
2.6
_
-
1.5
Temperature (°F)
-267 -
0 -
50 -
100 -
150 -
200 -
250 -
300 -
350 -
400 -
-1
49
99
149
199
249
299
349
399
1570
TOTAL
1349
724
829
500
108
0
0
0
0
0
3510
38.4
20.6
23.6
14.2
3.1
0.0
0.0
0.0
0.0
0.0
265
207
287
177
21
-
-
-
-
-
957
l'J.6
28.6
34.6
33.4
1').4
_
~
-
27.3
0
29
127
252
89
110
0
0
0
0
607
0.0
4.8
20.9
41.5
14.7
18.1
0.0
0.0
0.0
0.0
_
0
2
3
3
1
_
_
_
-
9
_
0.0
1.6
1.2
3.4
0.9
_
_
_
_
1.5
'See Figure 3-2 fur explanation of groups.
-------
TABLE B-7. EFFECTS OF LINE TEMPERATURE AND LINE
PRESSURE ON PERCENT LEAKING FOR VALVES
IN LIGHT LIQUID SERVICE WITHIN HIGH
LEAKING PROCESS UNITS BY CHEMICAL GROUP
Pressure (pela)
-15
0
50
100
150
200
250
300
350
400
450
500
550
1000
- -1
- 49
- 99
- 149
- 199
- 249
- 299
- 349
- 399
- 449
- 499
- 549
- 999
- 1050
TOTAL
Number
Screened
6
1075
702
383
317
100
45
168
29
33
36
174
224
3292
Group
Percent
of
Total
Screened
0.2
32.6
21.3
11.6
9.6
3.0
1.4
5.1
0.9
1.0
.1.1
5.3
6.8
7 '* PRIMARY CHEMICALS
Number
Leaking
1
20
21
20
16
9
6
10
4
5
3
15
17
147
Percent
Leaking
16.7
1.9
3.0
5.2
5.0
9.0
13.3
6.0
13.8
15.2 '
8.3
8.6
7.6
A. 5
Number
Screened
21
2411
1567
1034
621
261
21
140
24
A
0
0
16
6120
Group 81 PRIMARY CHEMICALS
Percent
of
Total
Screened
0.3
39.4
25.6
16.9
10.2
4.3
0.3
2.3
0.4
0.1
0.0
0.0
0.3
Number
Leak Inn
0
16
5
13
3
2
0
5
0
0
-
-
2
46
Percent
Leaking
0.0
0.7
0.3
1.3
0.5
0.8 i
0.0
3.6
0.0
. 0.0
-
-
12.5
0.8
Temperature (°F)
-267 -
0 -
50"F -
100 °F -
150 °F -
200"F -
250°F -
300 °F -
350°F -
400°F -
-1
49
99
149
199
2A9
299
349
399
1570
TOTAL
0
122
1039
819
287
634
89
161
15
112
3278
0.0
3.7
31.7
25.0
8.8
19.3
2.7
4.9
0.5
3. A
-
7
38
36
22
2A
3
11
1
2
144
-
5.7
3.7
A. A
7.7
3.8
3. A •
6.8
6.7
1.8
A. A
0
96
1912
1922
661
743
512
158
68
82
6154
0,0
1.6
31.1
31.2
10.7
12.1
8.3
2.6
1.1
1.3
-
2
11
10
6
3
9
0
0
5
A6
-
2.1
0.6
0.5
0.9
0.4
1.8
0.0
0.0
6.1
0.8
See Figure 3-2 for explanation of groups.
-------
TABLE B-8. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON
PERCENT LEAKING FOR PUMP SEALS IN LIGHT LIQUID
SERVICE
Pres:
-15
0
50
100
150
200
250
300
350
400
450
500
550
1000
sure (psig)
-1
- 49
99
- 149 •
- 199
- 249
- 299
- 349
- 399
- 449
- 499
- 549
- 999
- 1050
TOTAL
Number
Screened
0
146
115
66
44
6
8
15
9
6
8
6
13
0
442
Percent of
Total
Screened
0.0
33.0
26.0
14.9
9.9
1.4
1.8
3.4
2.0
1.4
1.8
1.4
2.9
0.0
Number
Leaking
10
19
3
8
0
3
0
1
3
1
2
2
-
52
Percent
Leaking
6.8
16.5
4.5
18.2
0.0
37.5
0.0
11.1
50.0
12.5
33.3
15.4
-
11.8
Temperature (°F)
-267
0
50
100
150
200
250
300
350
400
-1
- 49
99
- 149
- 199
- 249
- 299
- 349
- 399
- 1570
TOTAL
26
27
148
112
34
73
21
4
0
2
447
5.8
6.0
33.1
25.1
7.6
16.3
4.7
0.9
0.0
0.4
8
4
14
11
2
8
1
0
—
0
48
30.8
14.8
9.5
9.8
5.9
11.0
4.8
0.0
—
0.0
10.7
177
-------
TABLE B-9. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE
FOR FLANGES IN GAS SERVICE FROM ETHYLENE PRO-
CESS UNITS
GROUP I1 PRIMARY MATERIALS
Pressure (psig)
-15
0
50
100
150
200
250
300
350
400
450
500
550
1000
-1
- 49
- 99
- 149
- 199
- 249
- 299
- 349
- 399
- 449
- 499
- 549
- 999
- 1050
Number
Screened
0
145
102
36
34
65
66
30
22
6
18
29
11
0
Percent
of
Total
Screened
0.0
25.7
18.1
6.4
6.0 .
11.5
11.7
5.3
3.9
1.1
3;2
5.1
2.0
0.0
Number
Leaking
3
4
2
4
2
13
4
3
1
1
2 '
0 '
-
Percent
Leaking
_
2.1
3.9
5,6
11.8
3.1
19.7
13.3
13.6
16.7
5.6
6.9
0.0
-
TOTAL
564
39
6.9
Temperature (°F)
•267' -
0' -
50 -
100 -
150
200
250
300
350
400 -
TOTAL
-1
49
99
149
199
249
299
349
399
1570
71
58
270
117
35
8
0
1
1
5
566
12.5
10.2
47.7
20.7
6.2
1.4
0.0
0.2
0.2
0.9
7
10
15
4
1
1
-
0
0
1
39
9.9
17.2
5.6
3.4
2.9
12.5
-
0.0
0.0
20.0
6.9
See Figure 3-2 for explanation of groups
178
-------
TABLE B-10. EFFECTS OF LINE TEMPERATURE AND LINE
PRESSURE FOR FLANGES WITH GAS SERVICE
FROM HIGH LEAKING PROCESS UNITS BY
CHEMICAL GROUP
GROUP 51 PRIMARY CHEMICALS
Pressure (psig)
-15
0
50
100
150
200
250
300
350
400
450
500
530
1000
-1
- 49
- 99
- 149
- 199
- 249
- 299
- 349
- 399
- 449
- 499
- 549
- 999
- 1050
Number
Screened
1
34
53
46
5
52
72
10
2
0
52
33
31
0
Percent
of
Total
Screened
0.3
8.7
13.6
11.8
1.3
13.3
18.4
2.6
0.5
0.0
13.3
8.4
7.9
0.0
Number
Leaking
0
1
1
1
0
1
2
0
0
-
6
5
1
-
Percent
Leaking
0.0
2.9
1.9
2.2
0.0
1.9
2.8
0.0
0.0
-
11.5
15.2
3.2
-
GROUP 61 PRIMARY CHEMICALS
Suiter
Screened
17
245
23
7
6
1
2
11
0
0
0
0
4
0
Percent
of
Total
Screened
5.4
77.5
7.3
2.2
1.9
0.3
0.6
3.5
0.0
0.0
0.0
0.0
1.3
0.0
Number
Leaking
~0
5
1
0
0
0
0
1
—
-
-
-
2
-
Percent
Leaking
0.0
2.0
4.4
0.0
0.0
0.0
0.0
9.1
-
-
-
-
0.0
-
TOTAL
391
18
4.6
316
2.8
Temperature (CF)
•267
0
5"
100 -
150 -
200 -
250 -
30C -
350 ' -
400 -
TOTAL
-1
49
99
149
199
249
299
349
399
1570
0
1
44
75
36
11
22
68
51
83
391
0.0
0.3
11.2
19.2
9.2
2.8
5.6
17.4
13.0
21.2
0
0
1
2
0
0
0
3
5
7
18
0.0
0.0
2.3
2.7
0.0
0.0
0.0
4.4
9.8
8.4
4.6
0
15
51
95
62
22
5
15
3
48
316
0.0
4.8
16.1
30.1
19.6
7.0
1.6
4.8
1.0
15.2
-
0
0
2
0
3
0
2
0
2
9
-
0.0
0.0
2.1
0.0
13.6
0.0
13.3
0.0
4.2
2.8
'See Figure 3-2 for explanation of groups.
179
-------
TABLE B-ll. EFFECTS OF LINE TEMPERATURE AND LINE
PRESSURE FOR FLANGES IN LIGHT LIQUID
SERVICE WITHIN ETHYLENE PROCESS UNITS
BY CHEMICAL GROUP
GROUP 31 PRIMARY CHEMICALS
GROUP 4l PRIMARY CHEMICALS
Pressure (psig)
-15 -
0 -
50 -
100 -
150 -
200 -
250 -
300 -
350 -
400 -
450 -
500 -
550 -
1000 -
-1
49
99
149
199
249
299
349
399
449
499
549
999
1050
Number
Screened
0
39
32
24
16
15
42
19
32
4
32
21
23
25
Percent
of
Total
Screened
0.0
12.0
9.9
7.4
4.9
4.6
13.0
5.9
9.9
1.2
9.9
6.5
7.1
7.7
Number
Leaking
_
1
0
2
3
0
1
2
2
0
6
4
0
4
Percent
Leaking
_
2.6
0.0
8.3
18.7
0.0
2.4
10.5
6.2
0.0
18.7
19.0
0.0
16.0
Number
Screened
0
31
19
10
5
4
0
0
0
0
0
1
0
0
Percent
of
Total
Screened
0.0
44.3
27.1
14.3
7.1
5.7
0.0
0.0
0.0
0.0
0.0
1.4
0.0
0.0
Number
Leaking
_
0
0
0
0
0
-
-
-
-
-
0
-
-
Percent
Leaking
_
0.0
0.0
0.0
0.0
0.0
-
-
-
_
-
0.0
_
-
TOTAL
Temperature C°F)
324
TOTAL
324
25
7.7
70
25
7.7
70
0.0
•267 -
0 -
50 -
100 -
150 -
200 -
250 -
300 -
350 -
400 -
-1
49
99
149
199
249
299
349
399
1570
66
63
125
60
10
0
0
31
36
44
20.4
19.4
38.6
18.5
3.1
0,0
0.0
20.5
23.8
29.1
7
6
10
2
0
-
-
7
8
13
2.2
9.5
8.0
3.3
0.0
-
-
22.6
22.2
29.5
0
2
8
40
11
9
0
0
0
0
0.0
2.9
11.4
57.1
15.7
12.9
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
1See Figure 3-2 for explanation of groups.
180
-------
TABLE B-12. EFFECTS OF LINE TEMPERATURE AND LINE
PRESSURE ON PERCENT LEAKING FOR FLANGES
IN LIGHT LIQUID SERVICE WITHIN HIGH
LEAKING PROCESS UNITS BY CHEMICAL GROUP
GROUP 71 PRIMARY CHEMICALS
GROUP 81 PRIMARY CHEMICALS
Pressure (psig)
-15
0
50
100
150
200
250
300
350
WO
450
500
550
1000
- -I
- 49
- 99
- 149
- 199
- 249
- 299
- 349
- 399
- 449
- 499
- 549
- 999
- 1050
Hunter
Screened
1
116
82
63
30
35
36
30
16
13
22
70
54
0
Percent
of
Total
Screened
0.2
20.4
14.4
11.1
5.3
6.2
6.3
5.3
2.8
2.3
3.9
12.3
9.5
0.0
Hunter
Leak! OR
0
0
0
2
0
0
1
3
0
0
0
2
2
-
Percent
Leakirg
0.0
0.0
0.0
3.2
0.0
0.0
2.8
10.0
0.0.
0.0
0.0
2.9
3.7
—
Humber
Screened
6
446
282
117
61
40
10
23
4
0
0
0
1
0
Percent
of
Total
Screened
0.6
45.0
28.5
11.8
6.2
4.0
1.0
2.3
0.4
0.0
0.0
0.0
0.1
0.0
Number
Leaking
0
0
0
0
0
0
0
0
0
-
-
-
0
-
Percent
Leaking
0.0
0.0
0.0
0.0
0.0
o.o
0.0
0.0
0.0
_
-
-
0.0
_'
TOTAL
568
10
1.8
990
0.0
Temperature (°F)
-267
0
50
100
150
200
250
300
350
400
-1
- 49
- 99
- 149
- 199
- 249
- 299
- 349
- 399
- 1570
TOTAL
0
14
143
125
80
125
21
32
14
18
572
0.0
2.4
25.0
21.8
14.0
21.8
3.7
5.6
2.4
3.2
-
0
2
0
6
1
1
0
0
0
10
-
0^0
1.4
0.0
7.5
0.8
4.8
0.0
0.0
0.0
1.8
0
12
310
309
124
134
69
12
7
33
1010
0.0
1.2
30.7
30.6
12.3
13.3
6.8
1.2
0.7
3.3
-
0
0
0
0
0
0
0
0
0
0
_
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
o.o
0.0
See Figure 3-2 for explanation of groups.
181
-------
TABLE B-13. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON
PERCENT LEAKING FOR OPEN-ENDED LINES IN GAS
SERVICE WITHIN ETHYLENE PROCESS UNITS
GROUP I1 PRIMARY CHEMICALS
Pressure (psig)
-15 - -1
0 - 49
50 - 99
100 - 149
150 - 199
200 - 249
250 - 299
300 - 349
350 - 399
400 - 449
450 - 499
500 - 549
550 - 999
1000 - 1050
TOTAL
Temperature (°F)
•267 - -1
0-49
50 ' - 99
100 - 149
150 - 199
200 - 249
250 - 299
300 - 349
350 - 399
400 - 1570
TOTAL
Number
Screened
0
139
53
12
3
20
20
5
2
2
12
2
12
0
282
42
131
77
17
12
3
0
1
0
1
284
Percent of
'Total
Screened
0..0
49.3
18.8
4.3
1.1
7.1
7.1
1.8
0.7
0.7
4.3
0.7
4.3
0.0
14.8
- 46.1
27.1
6.0
4.2
1.1
0.0
0.4
0.0
0.4
Number
Leaking
—
6
7
3
0
0
5
3
0
1
5
2
4
-
36
8
8
17
3
1
0
-
0
-
0
37
Percent
Leaking
—
4.3
13.2
25.0
0.0
0.0
25.0
60.0
0.0
50.0
41.7
100.0
33.3
-
12.8
19.0
6.1
22.1
17.6
8.3
0.0
-
0.0
-
0.0
13.0
1See Figure 3-2 for explanation of groups.
182
-------
TABLE B-14. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON PERCENT LEAKING
FOR OPEN-ENDED LINES IN GAS SERVICE WITHIN HIGH LEAKING PROCESS
UNITS BY CHEMICAL GROUP
GROUP 51 PRIMARY CHEMICALS GROUP 61 PRIMARY CHEMICALS
Pressure (psig)
-15 - -1
0 - 49
50 - 99
100 - 149
150 - 199
200 - 249
250 - 299
300 - 349
350 - 399
400 - 449
450 - 499
500 - 549
550 - 999
1000 - 1050
TOTAL
Temperature (°F)
-267 - _1
0-49
50 - 99
100 - 149
150 - 199
200 - 249
250 - 299
300 - 349
350 - 399
400 - 1570
TOTAL
Number
Screened
0
29
28
42
5
3
7
0
4
0
0
. 0
4
0
122
0
2
37
49
13
6
4
1
9
0
121
Percent of
- Total
Screened
0.0
23.8
23.0
34.4
4.0
2.5
5.0
0.0
3.3
0.0
0.0
0.0
3.3
0.0
0.0
1.6
30.6
40.5
10.7
5.0
3.3
0.8
7.4
0.0
Number
Leaklnit .
-
3
3
1
0
0
2
-
0
-
-
-
2
-
11
-
0
8
2
0
1
0
0
0
_
11
Percent
LeakinR
-
D.3
L0.7
1. 4
0.0
0.0
ifl.6
-
0
. -
-
-
5D. 0
9.0
-
0,0
n.,6
/..i
0.0
lb.7
II .0
n.o
0.0
-
!'.09
Number
Screened
16
227
31
5
12
7
2
29
1
3
1
0
0
0
334
0
11
46
84
79
36
21
17
0
40
334
Percent of
Total
Screened
4.8
68.0
9.3
1.5
3.6
2.1
0.6
8.7
0.3
0.9
0.3
0.0
0.0
0.0
0.0
3.3
13.8
25.2
23.6
10.8
6.3
5.1
0.0
12.0
Number
LeakinR
0
2
1
0
0
0
0
0
0
0
0
-
-
-
3
-
0
0
2
1
0
0
0
-
0
3
Percent
Leaklna
0.0
0.9
3.2
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
-
-
-
0.9
-
0.0
0.0
2.4
1.3
0.0
0.0
0.0
-
0.0
0.9
'See Figure 3-2 for explanation of groups.
-------
TABLE B-15. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON PERCENT LEAKAGE
FOR OPEN-ENDED LINES IN LIGHT LIQUID SERVICE WITHIN ETHYLENE
PROCESS UNITS BY CHEMICAL GROUP
GROUP 31 PRIMARY CHEMICALS
GROUP 41 PRIMARY CHEMICALS
00
Number
Screened
Percent of
Total
Screened
Number
Leaking
Percei
Leak!
Pressure (pale)
-15
0
50
100
150
200
250
300
350
400
450
500
550
1000
-1
- 49
- 99
- 149
- 199
- 249
- 299
- 319
- 399
- 449
- 499
- 549
- 999
- 1050
TOTAL
0
18
9
16
7
6
14
12
23
1
12
12
7
14
151
0.0
11.9
6.0
10.6
4.6
4.0
9.3
8.0
15.2
0.7
8.0
8.0
4.6
9.3
_
0
0
3
2
2
6
6
5
0
1
2
5
7
39
_
0.0
0.0
18.7
28.6
33.3
42.9
50.0
21.7
0.0
8.3
16.7
71.4
50.0
25.8
Temperature (°F)
-267
0
50
100
150
200
250
300
350
400
- -1
- 49
- 99
- 149
- 199
- 249
- 299
- 349
- 399
- 1570
31
36
44
36
4
0
0
0
0
0
20.5
23.8
29.1
23.8
2.6
0.0
0.0
fi.O
0.0
0.0
7
8
13
11
0
_
-
,
-
-
22.6
22.2
29.5
30.6
0.0
-
_
-
-
-
Number
Screened
0
12
39
6
3
0
0
0
0
0
0
3
0
0
63
0
8
12
12
10
21
0
0
0
0
Percent of
Total
Screened
0.0
19.0
61.9
9.5
4.8
0.0
0.0
0.0
0.0
0.0
0.0
4.8
0.0
0.0
0.0
12.7
19.0
19.0
15.9
33.3
0.0
0.0
0.0
0.0
Number
Leaking
0
2
0
0
-
-
_
-
-
-
0
-
-
2
_
0
0
0
0
2
-
-
-
-
Percent
Leakitifi
0.0
5.1
0.0
0.0
-
-
-
-
-
-
0.0
-
-
3.2
_
0.0
0.0
o.o
0.0
9.5
-
-
-
-
TOTAL
151
39
25.8
63
3.2
See Figure 3-2 for explanation of groups.
-------
TABLE B-16. EFFECTS OF LINE TEMPERATURE AND LINE PRESSURE ON PERCENT LEAKING
FOR OPEN-ENDED LINES IN LIGHT LIQUID SERVICE WITHIN HIGH PROCESS
UNITS BY CHEMICAL GROUP
GROUP 71 PRIMARY CHEMICALS
GROUP 81 PRIMARY CHEMICALS
Number
Screened
Percent of •
Total
Screened
Pressure (palg)
-15
0
50
100
150
200
250
300
350
400
450
500
550
1000
- -1
- 49
- 99
- 149
- 199
- 249
- 299
- 349
- 399
- 449
- 499
- 549
- 999
- 1050
TOTAL
0
289
140
54
86
12
2
28
1
5
5
19
36
0
677
0.0
42.7
20,7
8,0
12,7
1.8
0.3
4.1
2.0
0.7
0.7
2,8
5.3
0.0
Number
Leaking
12
10
2
3
1
0
1
0
0
0
1
0
_
31
Percent
Leaking
.
4.2
7.1
3.7
3.5
8.3
0,0
3,6
• 0.0
20.0
0.0
5.3
0,0
_
4.6
Number
Screened
5
519
339
174
138
10
0
23
4
1
0
0
1
0
1214
Percent of
Total
Screened
0.4
42.8
27.9
14.3
11.4
0.8
0.0
1.9
0.3
0.1
0.0
0.0
0.1
0.0
Number
Leaking
0
15
12
7
1
0
-
0
0
0
-
-
0
-
35
Percent
Leaking
0.0
2.9
3.5
4.0
0.7
0.0
-
0.0
0.0
0.0
-
-
0.0
-
2.9
Temperature ("F)
-267
0
50
100
150
200
250
300
350
400
-1
- 49
- 99
- 149
- 199
- 249
- 299
- 349
- 399
- 1570
TOTAL
0
41
228
•154
40
133
15
40
0
21
672
0.0
6.1
33.9
22.9
6.0
19.8
2.2
6.0
0.0
3.1
-
0
13
6
2
7
1
2
. -
0
31
-
0.0
5.7
3.9
5.0
5.3
6.7
5.0
-
0,0
4.6
0
15
440
330
161
178
74
15
3
9
1225
0.0
1.2
35.9
26.9
13.1
14.5
6.0
1.2
0.2
0.7
-
0
15
12
4
2
1
0
0
1
35
-
0.0
3.4
3.6
2.5
1.1
1.4
0.0
0.0
11.1
2.9
'see Figure 3-2 for explanation of groups.
-------
VALVES—GAS
00
p
E
R
C
E
N
T
S
C
R
E
E
N
E
D
0-49
Group 1
Group 6
Group 7
100-149 200-249 300-349 400-449 500-549
PRESSURE
10001-
Figure B-l. Distribution of Sources Screened by Line Pressure for Ethylene and High
Leaking Process Units by Chemical Group for Valves with Gas Service
*See Figure 3-2 for explanation of groups.
-------
VALVES—GAS
P
E
R
C
E
N
T
S
C
R
E
E
N
E
P
0-49
Group 1
Group 5
Group 6
100-149 208-249
TEMPERATURE CO
300-349
400+
Figure B-2. Distribution of Sources Screened by Line Temperature for Ethylene and
High Leaking Process Units by Chemical Group for Valves with Gas Service
*See Figure 3-2 for explanation of groups.
-------
VALVES—LIGHT LIQUID
M
OO
oo
50-
P
E 40-
R
C
E
N
T 30-
L
E
A
K '20-
I
N
G
10~
0«— .
. — . _ . _ . _ .
-
i
' ,
/ \ •'
/ Y* '•
/ "^. * \
//.\ *
>:\\\
'/; \ \ .
/'" ' •
//• v. V.
//•' \ ^'
y /- 'x- x
/;' / ^ *** • A. / \
•* / ^ "-\\ / \ -^
h ^ — ^ <• i x ,' Nv '""*""-."]-• x"-
*i -V N^ / •'•" !. N
f v"x^^^-'- -^ ''-. x\
IiiiiiiiiTii.il
< -1 50-99 150-199 250-299 350-399 450-499 550-999
r-Dnift-49 100-149 200-249 300-349 400-449 500-549 1000+
bKuUP o
GROUP 4
GROUP 7 PRESSURE
y^»r*i^*il 11^* x-.
Figure B-3. Distribution of Sources Screened by Line Pressure for Ethylene and High
Leaking Process Units by Chemical Group for Valves with Light Liquid Service
*See Figure 3-2 for explanation of groups.
-------
VALVES—LIGHT LIQUID
O3
50 ~
P
E 40 ~
R
C
E
N
T 30 ~
L
E
A
K - 20 "
I
N
G
10"
ra —
t
•"•
\ .' *.
\
\
\ t_ — • — l •
x ; x \ '•
/ :\ \ -.
1 J-; N-\-
\ ^ -j A \ \ .
// •' x ^ \ '. /'
' S \ '• - • " '^
f •" "\ ' \ " ' • '\
/(•' \ .v-C—-
-------
APPENDIX C
SUMMARY STATISTICS AND DETAILED INFORMATION ON THE EFFECTS
OF AMBIENT TEMPERATURE AND ELEVATION ON LEAK FREQUENCY
This appendix contains detailed information on the effects of ambient
temperature and source elevation on leak frequency. Tables C-l and C-2 con-
tain summary statistics for ambient temperature for the groupings of sources
described in Section 3.
Ambient temperature was measured as a continuous variable, but to
evaluate its effect on leak frequency, it was grouped as less than 70 °F or
greater than or equal to 70°F. Tables C-3 and C-4 give the number of sources
screened, number leaking and percent leaking for both of the groups of ambient
temperatures. The statistics are categorized by source type, stream service,
the type of process unit, and the primary material group. Table C-3 contains
the data for ethylene process units and Table C-4 contains the data for high
leaking process units,
Chi-square tests were performed on each group to determine if there was
a significant difference in leak frequencies between the two categories of
ambient temperatures. The results are given in Tables C-3 and C-4. For the
ehtylene process units (Table C-3), the leak frequencies of valves are signif-
icantly different in all categories. For both gas and light liquid service
in the high leaking primary material group, higher leak frequencies were
found at the higher levels of ambient temperature. For the low leaking pri-
mary material group, the higher leak frequencies occurred at the lower ambient
temperature level. The only other group in the ethylene process units that
showed a significant effect of the ambient temperature is open-ended lines in
gas stream service in the high leaking primary material group.
190
-------
Table C-4 contains the same information for the high leaking process
units. Valves in group 7 showed the only significant effect on leak frequen-
cies of ambient temperature. The higher level of ambient temperature was
associated with the higher leak frequency.
Tables C-5 and C-6 contain the data on the effects of elevation on
leak frequency for ethylene and high leaking process units, respectively.
Chi-square tests were performed to determine differences in percent leaking
for the two levels. There were no significant differences in leak frequencies
for any source types in the ethylene process units. The high leaking process
units showed a few significant effects of elevation on leak frequencies.
These effects were seen for valves and open-ended lines irx light liquid stream
service and with high leaking primary materials in the line. For the even
numbered groups only valves in gas stream service were significantly affected.
In each of these cases the higher leak frequency occurred at the ground level.
191
-------
TABLE C-l. SUMMARY OF AMBIENT TEMPERATURE DURING SCREENING
OF VARIOUS SOURCE TYPES IN GAS SERVICE
HIGH LEAKING PROCESS
SOURCE
VARIABLE TYPE
VALVES
FLANGES
OPEN-ENDED
LINES
SUMMARY STATISTICS FOR
AMBIENT AIR TEMPERATURE
Average ("F)
Standard Deviation
Minimum
Maximum
Average (OF)
Standard Deviacion
Minimum
Maximum
Average (°F)
Standard Deviation
Minimum
Maximum.
HIGH LEAKING
PRIMARY MATERIALS
73.1
18.5
33
104
86.2
13.0
33
102
69.5
17.5
33
100
LOW LEAKING
PRIMARY MATERIALS
63.6
19-5
30
100
79-8
16.4
33
100
66.0
20.4
30
100
ETHYL ENE
HIGH LEAKING
PRIMARY MATERIALS
58.
20.
11
187
73.
14.
20
120
48.
20.
20
90
.3
,6
.2
.5
.2
,1
192
-------
TABLE C-2. SUMMARY OF AMBIENT TEMPERATURE DURING SCREENING OF VARIOUS
SOURCE TYPES IN LIGHT LIQUID SERVICE
HIGH LEAKING PROCESS
SOURCE SUMMARY STATISTICS FOR
TYPE AMBIENT AIR TEMPERATURE
VALVES
PUMP SEALS
FLANGES
OPEN-ENDED
LINES
Average (°F)
Standard Deviation
Minimum
Maximum
Average (°F)
Standard Deviation
Minimum
Maximum
Average (°F)
Standard Deviation
Minimum
Maximum
Average (°F)
Standard Deviation
Minimum
Maximum
HIGH LEAKING
PRIMARY MATERIALS
57.8
19.0
29
100
54.5
15.1
32
98
77.6
19.0
29
100
52.2
13.5
29
98
LOW LEAKING
PRIMARY MATERIALS
72.4
17.7
29
101
74.2
17.6
32
100
82.0
15.4
29
102
77.6
17.4
29
10ft
ETHYLENE
HIGH LEAKING
PRIMARY MATERIALS
62.8
19.9
21
91
56.6
21.2
22
85
74.1
13.6
24
91
51.3
21.1
24
91
LOW LEAKING
PRIMARY MATERIALS
65.
20,
22
91
62,
21,
36
88
78,
14,
30
91
48.
19.
22
90
,5
,7
.9
.2
.1
.7
,2
,4
-------
TABLE C-3. EFFECT OF AMBIENT TEMPERATURE ON PERCENT OF SOURCES LEAKING IN
ETHYLENE PROCESS UNITS AS A FUNCTION OF THE PRIMARY CHEMICAL
GROUPS
PRIMARY CHEMICAL
SOURCE
TYPE
VALVES
PUMP SEALS
FLANGES
OPEN-ENDED
LINES
STREAM
SERVICE
Gas
Light Liquid
Light Liquid
Gas
Light Liquid
Gas
Light Liquid
AMBIENT
TEMPERATURE,
"F
<70°
70°+
<70°
70°+
<70°
70"+
<70"
70°+
<70"
70°+
<70°
70°+
<70°
70"+
NUMBER
SCREENED
3760
2534
1666
1848
29
32
165
469
68
259
223
82
110
41
NUMBER
LEAKING
474
460
446
511
7
11
10
29
6
19
19
18
30
9
Group 1
PERCENT
LEAKING
12.6
18.2
26.8
27.6
24.1
34.4
6.1
6.2
8.8
7.3
8.5
22.0
27.3
21.9
and Group 3
CHI-SQUARE p2
36.8 <0.001
0.34 >0.05
0.4 >0.05
0.01 >0.05
0.16 >0.05
10.1 <0.01
0.04 >0.05
NUMBER
SCREENED
240
367
7
8
9
61
50
13
PRIMARY
NUMBER
LEAKING
2
7
0
2
0
0
2
0
CHEMICAL Group 41
PERCENT
LEAKING CHI-SQUARE
0.8 1.1
1.9
0.0 2.1
25.0
0.0 *
0.0
4.0 0.54
0.0
P2
>0.05
>0.05
>0.05
'See Figure 3-2 for explanation of groups.
Probability of no significant difference in leak frequency due to ambient temperature.
*Expected values were too low for Chi-square test.
-------
TABLE C-4. EFFECTS OF AMBIENT TEMPERATURE ON PERCENT OF SOURCES LEAKING IN HIGH
LEAKING PROCESS UNITS AS A FUNCTION OF THE PRIMARY CHEMICAL GROUPS
PRIMARY CHEMICAL Gronp 5 and' Group 71 PRIMARY CHEMICAL Grouo 6
SOURCE STREAM
TYPE SERVICE
VALVES Gaa
Light Liquid
PUMP SEALS Light Liquid
FLANGES Gas
Light Liquid
OPEN-ENDED
LINES Gas
Light Liquid
AMBIENT
TBMPERATURE,
°F
<70°
70°+'
<70°
70°+
<70°
70°+
<70°
70°+
<70e
70"+
<70°
70°+
<70"
70"+
NUMBER
SCREENED
499
729
2435
803
108
18
46
345
161
417
71
75
713
85
HUMBER
LEAKING
50
96
52
95
11
3
1
17
1
9
4
9
45
2
PEFCENT
LEAKING
10.0
13.2
2.1
11.8
10.2
16.7
2.2
4.9
0.6
2.2
5.6
12.0
6,3
2.4
NUMBER
CHI-SQUARE P1 SCREENED
2.8 >0.05 1090
668
135.6 <0.001 2861
3293
0.7 >0.05 101
142
0.7 >0.05 77
239
0.6 >0.05 219
791
1.8 >0,05 204
143
2.1 >0.05 415
876
NUMBER
LEAKING
17
5
17
29
3
11
3
6
0
0
1
3
7
31
PERCENT
LEAKING
1.6
0.8
0.6
0.9
3.0
7.8
3.9
2.5
0.0
0.0
.5
2.1
1.7
3.5
and Group S1
CHI-SQUARE P1
2.2 >0.05
1.7 >0.05
2.5 >0.05
0.4 >0.05
*
1.9 >0.05
3.3 >0.05
'See Figure 3-2 for explanation of groups.
^Probability oE no significant difference in leak frequency due to ambient temperature.
*
Expected values were too low for Chi-square test.
-------
TABLE C-5. EFFECTS OF SOURCE ELEVATION ON PERCENT LEAKING FOR ETHYLENE PROCESS
UNITS AS A FUNCTION OF PRIMARY CHEMICAL GROUPS
PRIMARY CHEMICAL Group 1 and Group 31 PRIMARY CHEMICAL Group 21
SOURCE
TYPE
VALVES
PUMP SEALS
FLANGES
OPEN-ENDED
LINES
SERVICE ELEVATION
Gas • Ground
Above
Light Liquid Ground
Above
Light Liquid Ground
Above
Gas Ground
Above
Light Liquid Ground
Above
Ga^ Ground
Above
Light Liquid Ground
Above
NUMBER NUMBER
SCREENED LEAKING
3298
2844
2578
926
61
0
246
387
234
91
235
69
109
42
475
452
716
237
' 18
-
13
26
16
a
25
12
29
10
PERCENT
LEAKING CHI-SQUARE pa
14.9 1.2 >0.05
15.9
27.8 1.6 >0.05
25.6
29.5 *
-
5.3 0.53 >0.05
6.7
6.8 0.37 >0.05
8.8
10.6 2.3 >0.05
17.4 2.0
26.6 0.12 >0.05
23.8
NUMBER NUMBER PERCENT
SCREENED LEAKING LEAKING CHI-SQUARE p2
494 8 1.6 0.34 >0.05
113 1 0.9
15 2 13.3 *
0 — —
55 0 0.0 *
15 0 0.0
54 2 3.7 0.3 >0.05
9 0 0.0
See Figure 3-2 for explanation of groups.
Probability of no significant difference in leak frequency due to source elevation.
Insufficient data for Chi-square test
-------
TABLE C-6. EFFECTS OF SOURCE ELEVATION ON PERCENT LEAKING IN HIGH LEAKING PROCESS
UNITS AS A FUNCTION OF PRIMARY CHEMICAL GROUPS
PRIMARY CHEMICAL Group 5 and Group 7 PRIMARY CHEMICAL Group 6
B
SOURCE
TYPE SERVICE ELEVATION
VALVES Gas Ground
Above
Light Liquid Ground
Above
PUMP SEALS Light Liquid Ground
Above
FLANGE Gaa Ground
t-1
^D Above
-J
Light Liquid Ground
Above
OPEN-ENDED
LINE Gas Ground
Above
Light Liquid Ground
Above
NUMBER
SCREENED
479
749
2494
795
122
4
155
236
417
160
59
87
623
172
NUMBER
LEAKING
54
92
121
25
14
0
5
13
9
1
6
7
45
2
PERCENT
LEAKING CHI-SQUARE p 2
11.3 0.3 >0.05
12.3
4.fl 4.1 <0.05
3.1
11.5 *
3.2' 1.0 >0.05
5.!.
2.1! 1.6 >0.05
O.li
10.!! 0.2 >0.05
8.0
7.;; 8.9 0.05
*
2.1 >0.05
*
2.1 >0.05
0.1 >0.05
'See Figure 3-2 for explanation of groups.
^Probability of no significant difference in leak frequency due to source elevation.
*Insufficlent data for Chi-squares test
-------
APPENDIX D
CORRECTIONS TO SCREENING DATA
During the SOCMI fugitive emission screening project (Reference 1),
occasional corrections were required on the original data sheets. These cor-
rections were subsequently documented along with an explanation of why they
were necessary.
One clarification that affected almost all of the units screened was
dua to the decision to exclude water when calculating the primary chemical
concentration. To make this adjustment all primary material concentrations
that were below thirty percent (30%) were changed to reflect ninety to one-
hundred percent (90-100%), if no secondary material other than water existed.
That is, if the primary chemical was twenty percent (20%) of the stream and
water made up the other eighty percent (80%), then the primary chemical con-
centration was adjusted to one hundred percent (100%). So, the concentration
number was adjusted to reflect the percent of total VOC's.
Table D-l and D-2 summarize all adjustments and corrections made to the
original data sheets. Table D-l is a summary of the detailed corrections
listed in Table D-2 which affected the overall number screened, the number
not screened and the number screened greater than or equal to 10,000, as
reported earlier (Reference 1). After the corrections described in this
appendix were made, the data reported in this reference were used for the
analyses in Section 3 of this report.
Other clarifications, mostly due to miscoding by the recorder, are
listed in Table D-2, by unit, then source identification sequence. Coding
corrections covered a wide range of source identification codes and few were
changed from the same code number. Therefore, it was not feasible to list
198
-------
all the old codes along with their corrections
199
-------
TABLE D-l. CORRECTIONS AFFECTING RESULTS ON PREVIOUS REPORTS
SOURCE/SERVICE
NUMBER
SCREENED
OLD NEH
1443 1450
Light Liquid 2897 2833
NUMBER
HOT SCREENED
OLD NEW
No Change
76 142
HU10JER
SCREENED >°10000
OLD NEW
NCI Change
Nc Chinge
O
O
Open-Ended Lines
Light Liquid 3603 3605
417 415
No Change'
Relief Valves
Gas
Light Liquid
85
84
69 68
226 227
47 48
No Change
No Ch.-mge
EXPLANATION
Unit 2 had 4 sources
and Unit 4 had 3 sources
which were reclassified
from compressors.
Unit 60 had 64 sources
in which the values were
changed from 0 to miss-
ing. Unit 4 had 2 sources
which were reclassl£led
from pumps and their
values changed from 0 to
missing. Unit 60 had 2
sources which should have
been recorded with Unit 61.
Unit 29 had 2 sources in
which the values were
changed from missing to 0.
Unit 60 had 2 sources which
should have been recorded
with Unit 61.
Unit 20 had 1 source in which
the value was changed from 0
to missing.
Unit 20 had 1 source in which
the value was changed from 0
to missing.
-------
TABLE D-l (continued)
NUMBER
SCREENED
NUMBER
HOT SCREENED
SOURCE/SERVICE OLD HEW OLD NEW
H'JMT.FR
SCREENSD :-10000
OLD NEW
EXPLANATION
9668 9669
Light Liquid 18294 18300
2047 2046
2553 2548
Ho Ch.inge
1174 U83
to
O
PUIDDB
Light Liquid
Compressors
Gaa
647 646
29 22
29
28
No Change
Ho Change
Other
Light Liquid
33 34
No Change
No Change
Unit 29 had 1 source in
which the valve was changed
from missing to (J.
Unit 29 had 6 sources in
which the values were changed
from missing to (J. Unit 20
had 1 source in which the value
was changed from 0 Co missing.
Unit 3 had 1 source in which the
value was changed from missing to
0, Unit 60 had 9 sources which
had values of 10,000 that were
not included in the number screen-
ed over 10,000. Unit 60 had 8
sources which should have been
recorded with Unit 61, 1 of which
had a value of 10,000.
Unit 29 had 1 source in which the
value was changed from missing to
0. Unit 4 had 2 sources which
were reclasslfied as flanges.
Unit 2 had 4 sources and Unit 4
had 3 sources which were reclassi-
fied to flanges. Unit 64 had 1
source in which the value was
changed from missing to (t. Unit
4 had 1 source which was reclassl-
fled to other.
Unit 4 had 1 source which was
reclassified from compressors.
-------
TABLE D-2. CORRECTIONS TO SCREENING DATA SHEETS
Unit
Source ID
Change
33
133-140
534, 1052, 1355-1356, 1393
365
1552
1580-1581
Elevation to 1
Comment to missing
Secondary material concentration to 4
Comment to 1
Deleted
Source type to 54
Line temperature to 140
Service to 2
918
1574-1576, 1594-1596,
1606-1608, 1624, 1633,
1646, 1723
1756
1757
1758, 1822
1873, 1895
2215-2219
2440, 2487
3004
3005-3020
3021-3024
3025
Screening value to missing
Comment to 1 (inaccessible)
Source type to 2
Source type to 1
Line temperature to 143
Line pressure to 525
Source type to 1
Source type to 52
Line temperature to 143
Line pressure to 525
Source type to 1
Line temperature to 188
Line pressure to 155
Screening value to missing
Comment-to 1 (inaccessible)
Source type to 2
Line temperature to 300
Line pressure to 6
Line pressure to 6
Primary material to 3
Primary material concentration to 3
Secondary material to 1
Secondary material concentration to 2
Screening value to missing
Comment to 1 (inaccessible)
(continued)
202
-------
TABLE D-2. CORRECTIONS TO SCREENING DATA SHEETS (CONTINUED)
Unit
Source ID
Change
3028-3030
3031-3033
3034-3038, 3121
3236-3298
3239
3240
3241-3243
3289
3300, 3309-3317,
3332-3347, 3350-3373
3378-3388
3447
3463-3469, 3473-3477
3478-3498
3552-3578
3604
3605-3607
3608-3630
3631-3634
3635-3660
Primary material to 3
Primary material concentration to 3
Secondary material to 1
Secondary material concentration to 2
Line temperature to 260
Line temperature to 260
Primary material to 3
Primary material concentration to 3
Secondary material to 1
Secondary material concentration to 2
Line temperature to 5
Source type to 40
Line temperature to 260
Line temperature to 5
Line temperature to 5
Screening value to missing
Comment to 1 (inaccessible)
Service to 1
Service to 1
Line pressure to 2
Screening value to missing
Comment to 1 (inaccessible)
Primary material to 1
Primary material concentration to 7
Line temperature to 25
Service to 1
Line pressure to 500
Service to 2
Service to 1
Line pressure to 2
Service to 2
Service to 1
Line pressure to 2
(continued)
203
-------
TABLE D-2. CORRECTIONS TO SCREENING DATA SHEETS (CONTINUED)
Unit
Source ID
Change
3683-3684
3700-3714
3716, 3718
3719-3720
3721-3722
3734-3740
Service to 1
Primary material concentration to 5
Primary material concentration to 5
Secondary material to 2
Secondary material concentration to 2
Primary material concentration to 4
Secondary material to- 2
Secondary material concentration to 2
Primary material to 3
Primary material concentration to 5
Secondary material to 1
Secondary material concentration to 3
Primary material concentration to 5
Secondary material concentration to 3
3751-3754
3755-3758
3762-3783
3801-3809
3882
3901
3913-3917
3968-3975, 3987
4668, 5048
5339
5690
3 1215-1216
1349
1638
1639
Service to 2
Service to 1 t
Secondary material concentration to 1
Service to 1
Source type to 42
Source type to 30
Line temperature to 62
Line pressure to 480
Service to 1
Source Type to 35
Screening value to 10,000
Source type to 52
Service to 1
Primary material to 5
Line temperature to 195
Line pressure to 225
Comment to 1
Secondary material concentration to 0
Source type to 32
Source type to 42
(continued)
204
-------
TABLE D-2. CORRECTIONS TO SCREENING DATA SHEETS (CONTINUED)
Unit
Source ID
Change
11
12
97-98, 849, 956.
956-959
1120-1139
2161, 2173
3375
5192, 5837
6760
If Service=missing
21-40
607 (2nd)
608 (2nd)
613 (2nd)
735 (2nd)
885
1745
1953
2358-2359
2426-2429
2566-2585, 2626-2641
2731-2734
2889
3156
3409, 3411, 3418, 3422
If screening team=13 or 15
If source type=*
31-100
Comment to 1 (inaccessible)
Deleted
Source type to 1
Source type to 32
Source type to 1
Service to 1
Source type to 1
Service to 2
Day to 21
Source id to 4105
Source id to 4106
Source id to 4107
Source id to 4108
Source type to 10
Unit to 11
Source type to 10
Service to 10
Comment to 3
Secondary material concentration to
missing
Comment to 3
Deleted
Service to 2
Comment" to 3
Source id to: id plus 1959
Source type to 1
Month to 3
(continued)
205
-------
TABLE D-2. CORRECTIONS TO SCREENING DATA SHEETS (CONTINUED)
Unit
Source ID
Change
12
87
221-240
310, 338, 339, 362
561-562
565-580
2730-2739
2788-2791
3416
Source type to 40
Service to 2
Source type to 40
Service to 2
Primary material to 40
Primary material concentration to 6
Secondary material to 45
Secondary material concentration to 3
Line temperature to 170
Line pressure to 100
Ambient air temperature to 60
Service to 2
Primary material to 42
Primary material concentration to 9
Secondary material to missing
Secondary material concentration to
missing
Primary material to 41
Primary material concentration to 9
Secondary material to missing
Secondary material concentration to
missing
Source type to 3
20* If screening value=
missing
1049, 1081
1239, 1251, 1314
3156
3201
Comment to 1
Source typ'e to 33
Screening value to missing
Service to 2
Screening value to missing
21*
947
Service to 2
22
154
221-240
Comment to 1 (inaccessible)
Ambient air temperature
*Pedco used duplicate unit #*s, so the VMC unit screened between 2-14 and 2-20 was
changed to unit 28 and the EDC unit screened between 2-12 and 2-15 was changed to
unit 29.
(continued)
206
-------
TABLE D-2. CORRECTIONS TO SCREENING DATA SHEETS (CONTINUED)
Unit
Source ID
Change
28 (20)* If primary material=19
2710-2720
2777
2793
2855
3257-3276
29 (21)* If screening value is blank
19
359-360
432-440
626-640
308-809
'865-880
1017, 1025
2103-2120
2135-2140
3116
3181-3200
3239-3240
Primary material to 14
Primary material concentration to 9
Secondary material to missing
Secondary material concentration to
missing
Secondary material to missing
Secondary material concentration to
missing
Screening value to missing
Source type to 45
Screening value to 30
Line temperature Co 355
Screening value to 0
Source type to 35
Comment to 1
Secondary material to missing
Secondary material concentration to
missing
Secondary material to missing
Secondary material concentration to
missing
Comment to 1 (inaccessible)
Secondary material to missing
Secondary material concentration to
missing
Source type to 35
Comment to missing
Secondary material to missing
Secondary material concentration to
missing
Comment to 1
Instrument to 2
Secondary material to missing
Secondary material concentration to
missing
*Pedco used duplicate unit :.-'s, so the VMC unit screened between
2-14 and 2-20 was changed to unit 28 and the EEC unit screened
between 2-12 and 2-15 was changed to unit 29.
207
-------
TABLE D-2. CORRECTIONS TO SCREENING DATA SHEETS (CONTINUED)
Unit
Source ID
Change
32
435, 463
646-647
Comment to 2
Secondary material concentration to 2
33
331, 337-338
Comment to 1
34
741-760
1521
Elevation to 2
Service to 2
60
1-505
(excluding 361-364j
398-401, 425-428)
39
506-696
897-913
1044-1057
1058-1068
1269-1278
1301-1720
1810 (2nd)
1829-1831
1941-1960
2005-2080
2081-2123
2354-2406, 2421-2449.
2436-2500
2591-2600
Primary material to 6
Source type to 30
Primary material to 6
Primary material concentration to 6
Secondary material to 14
Secondary material concentration to 2
Service to 2
Secondary material to 9
Elevation" to 3
Secondary material to 9
Secondary material to 8
Primary material to 6
Source id to 1811
Secondary material to missing
Secondary material concentration to
missing
Primary material to 6
Primary material to 4
Secondary material to 15
Primary material to 4
Secondary material to 15
Secondary material concentration to 0
Primary material to 5
Secondary material to missing
Secondary material concentration to
missing
(continued)
208
-------
TABLE D-2. CORRECTIONS TO SCREENING DATA SHEETS (CONTINUED)
Unit
Source ID
Change
60
61
2628
2743-2747, 2787-2800
2890-2900
4113-4130
6970-6974
3021-3027, 3044-3060
3053
3121-3127
3188
3233, 3235-3237
3246-3260
3261-3280
3281-3297
3298-3305
3306-3328
3329-3335
3521-3716
3742-3760
3781-3830
3805-3806
Source type to 1
Secondary material to missing
Secondary material concentration to
missing
Secondary material to missing
Secondary material concentration to
missing
Primary material to 5
Secondary material to missing
Secondary material concentration to
missing
Secondary material to 3
Source type to 1
Screening value to 3600
Primary material to 6
Screening value to 200
Screening value to 0
Secondary material to 6
Secondary material to 3
Primary material to 3
Primary material to 3
Secondary material to 6
Primary material to 6
Primary material to 3
Secondary material to 6
Primary material to 3
Secondary material missing
Secondary material concentration to
missing
Primary material to 3
Secondary material to 6
Comment to 1
(continued)
209
-------
TABLE D-2. CORRECTIONS TO SCREENING DATA SHEETS (CONTINUED)
Unit
Source ID
Change
62
64
•
65
'
3841-3860
If service=2 and
primary material is
2, 4, or 5
3933
3941-395Q
4357, 4362
4604-4609, 4611-4612
4631-4640
4801-4820
4900 (2nd)
5241-2
5243
5244-5280
5286-5290, 5299-5300
5340
5354-5356, 536*4-5367
5393
5404, 5413-5414,
5464-5466
5468
5511
5521-5540
5979-5986, 5988-6000
6011-6013
6021-6039
Unit to 62
Service to 3
Service to 1
Primary material concentration to 1
Secondary material concentration to 0
Secondary material to 3
Secondary material concentration to 2
Service to 3
Comment to 1 (inaccessible)
Source id=id + 46
Ambient air temperature to 8
Source id to 6751
Service to 2
Service to "2
Elevation to 2
Service to 2
Comment to 1 (inaccessible)
Source type to 1
Screening value to missing
Source type to 10
Screening value to missing
Service to 1
Comment to 2
Line pressure to 45
Line pressure to -5
Comment to 1 (inaccessible)
Ambient air temperature to 84
(continued)
210
-------
TABLE D-2. CORRECTIONS TO SCREENING DATA SHEETS (CONTINUED)
Unit
Source ID
Change
65
6040
6113-6114
6221-6222, 6224-6225,
6243-6249
6601-6620
.Ambient air temperature to 84
Comment Co 2
Screening value to missing
Commenc Co I {Inaccessible)
Line pressure to -5
66
7083
7141-7160
7173, 7249, 7264, 7336,
7354
8574, 8808-8810
3850, 3362
Source type to 30
Unit to 66
Source type co 30
Secondary material to missing
Secondary material concentration to
missing
Source type to 30
211
-------
TECHNICAL REPORT DATA
(Please read httrucrions on the reverse before completing}
. REPORT NO.
EPA-600/2-81-111
3. RECIPIENT'S ACCESSION- NO.
TITLE ANDSUSTITLE
Analysis of SOCMI VOC Fugitive Emissions Data
5. REPORT DATE
June 1981
S. PERFORMING ORGANIZATION CODE
. AUTHORlS)
B. PERFORMING ORGANIZATION REPORT NO.
G. J.Langley, S.M.Dennis, J.F.Ward, and
L. P. Provost
. PERFORMING ORGANIZATION NAME AND ADDRESS
Radian Corporation
P.O. Box 9948
Austin, Texas 78766
10. PROGRAM ELEMENT NO.
C9HA1A
11. CONTRACT/GRANT NO.
68-02-3171, Task 28
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOOCOVER6D
Task Final; 12/80 - 5/81
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES TERL_RTP prOject officer is Bruce
919/541-2547.
A. Tichenor, Mail Drop 63,
16. ABSTRACT ,pne rep0rt gives results of an examination of fugitive emission data from
Synthetic Organic Chemical Manufacturing Industry (SOCMI) processing units (col-
lected under earlier EPA studies) for correlations between process variables and
leak frequency. Although line temperature did not have a consistent relationship with
leak frequency, the data showed that leak frequency increased with increasing line
pressure. Also, emission factors for three process types (vinyl acetate, cumene,
and ethylene) were developed and presented. Increases in mass emissions due to
occurrence and recurrence of leaks for these three process types are also estima-
ted. Finally, the effect of adjusting portable hydrocarbon readings by chemical
response factor curves on leakage frequency estimates is investigated. Despite the
wide range of response factors encountered, the adjusted leak frequencies were
essentially the same as the unadjusted frequencies.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution Cumene
Volatility Ethylene
Organic Compounds
Processing Hydrocarbons
Leakage
Vinyl Acetate
Pollution Control
Stationary Sources
Volatile Organic Com-
pounds
Fugitive Emissions
13B
20M
07C
13H
14G
13. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
225
20. SECURITY CLASS {Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (»-73)
212
-------
-------
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Center for Environmental Research Information
Cincinnati, Ohio 45268
OFFICIAL BUSINESS
PENALTY FOR PRIVATE USE S3OO
AN EQUAL OPPORTUNITY EMPLOYER
POSTAGE AND FEES PAID
US ENVIRONMENTAL PROTECTION AGENCY
EPA-335
If your address is incorrect, please change on the above label
tear off; and return to the above address.
If you do not desire to continue receiving these technical
reports. CHECK HffffD; tear off label, and return it to the
above address.
Publication No. EPA-600/2-8i-m
------- |