United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S2-81-109 Sept. 1981
Project Summary
VOC Fugitive Emissions
Data—High Density
Polyethylene Process Unit
Cynthia M. Harvey and A. Carl Nelson, Jr.
The report gives data from a 10-
month study of volatile organic com-
pound (VOC) fugitive emissions from
a high density polyethylene process
unit. It gives statistics on leak fre-
quency, leak occurrence, and leak
recurrence, with a leak defined as
having a screening value equal to or
greater than 10,000 ppmv. The
statistics are broken down into gas
and light-liquid service. The report
also includes a valve maintenance
program; the data provide statistics on
repair effectiveness.
This Project Summary was develop-
ed by EPA's Industrial Environmental
Research Laboratory. Research Tri-
angle Park, NC, to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
Allied Chemical Corporation collected
VOC fugitive emissions data at its HOPE
(high density polyethylene) plant in
Baton Rouge, Louisiana, from October
1979 to August 1980. The data were
collected on Reactor F, which uses the
Phillips Petroleum HOPE process. Table
1 provides a summary of these data,
which were analyzed to provide infor-
mation on leak frequency, rate of leak
occurrence, rate of leak recurrence,
repair effectiveness, and emission rate
of leaking source.
This summary presents the results of
these analyses.
Results
Leak Frequency
The leak frequency is defined as the
percentage of screened sources with
meter readings ^ 10,000 ppm. Table 2
shows the leak frequency for each of the
six tests by source and service.
Rate of Leak Occurrence
The rate of leak occurrence (RLO) is
defined as the percentage per unit time
of sources screened at §10,000 ppm
that had earlier been screened at
<10,000 ppm. Note that only those
sources screened in all six surveys were
included in the analysis. Thus, this
statistic is calculated from the number
of nonleakers obtained during Test 1.
For each successive test, the cumulative
number of new leakers (CNML) is
divided by the number of nonleakers
obtained during Test 1 and converted to
a percentage. This percentage is further
divided by the average number of days
after the end of Test 1 and multiplied by
30 days to express a monthly RLO. Table
3 presents the 30-day RLO for each test
by source and service. Average rates of
leak occurrence based on percent of
original (i.e., at Test 1) nonleakers are
calculated for each test period and are
normalized to a 30-day period. These
values represent the average rates of
leak occurrence over the total period
from time zero (i.e., at Test 1) to the end
of the test for which they were computed.
This does not mean that the leak rate
was assumed constant over that time
-------
Table 1.
Test
No.
1
2
3
4
5
6
Summary of VOC Fugitive Emissions Test Results for Reactor F of Allied Chemical
Sources in Repair Type of
Test Sources all previous Leakers Leak Attempted Successful effective- mainten-
dates tested tests (^10,000 ppm) freq., % repairs repairs ness, % ance'
10/31
thru 902 902
12/05
1/07
thru 908 883
2/01
3/03
thru 923 881
3/24
4/12
thru 927 854
5/01
5/12
thru 940 845
6/15
7/07
thru 927 818
8/11
148 16.4 83 46 43.4 U
148 16.3 148 123 83.1 D
89 9.6 89 45 49.4 U
124 13.4 124 55 44.3 U
66 7.0 66 40 60.6 U
77 8.3 77 33 42.8 U
*D - Directed maintenance.
U = Undirected maintenance.
period. These values can be used to
approximate the rates of leak occurrence
over any time period by applying the
appropriate average 30-day rate. For
example, the percent of leaks occurring
over the first 180 days for valves in
liquid service can be approximated by
multiplying the 30-day rate for the fifth
survey (5.0 percent) by six (i.e., 180/30)
which gives 30 percent. Based on the
original number of nonleakers of 126,
one would expect approximately 38
(0.30x126 = 37.8) leak occurrences
after 180 days.
Rate of Leak Recurrence
The rate of leak recurrence (RLR) is
defined as the percentage per unit time
of repaired sources (i.e., those with
leaks reduced by maintenance below
10,000 ppm) that were subsequently
screened at §10,000 ppm. Allied
performed directed maintenance only
after Test 2. Therefore, these statistics
are computed on the number of repaired
sources obtained during Test 2 that
were rescreened in each successive
test. For each successive test, the
cumulative number of recurring leakers
(CNRL) is divided by the number of
repaired sources and converted to a
percent. The percentage is further
divided by the number of days to
recurrence and multiplied by 30 days to
express a monthly RLR. Table 4 presents
the 30-day RLR by source and service at
various intervals throughout the study.
Repair Effectiveness
The repair effectiveness is defined as
the percentage of attempted repairs that
were successful. Allied performed
directed maintenance only after Test 2.
Table 5 indicates the repair effective-
ness of directed maintenance after Test
2 by source and service. Table 6 shows
repair effectiveness after other tests by
source and service.
Emission Rates of Leaking
Sources
The emission rate is defined as the
mass of VOC fugitive emissions per unit
time per leaker. Table 7 presents an
arithmetic average of the results by
source and service for those sources
which were measured before mainten-
ance. This average is an estimate of the
emission rate for leaking sources, and it
should not be compared to an emission
factor which applies to all sources.
-------
Leak Frequency by Source and Service
Liquid service
Source
Flanges
Valves
Statistic
No. of leakers
No. of sources
tested
Leak frequency.
%
/.a," %
UCL,C %
No. of leakers
No. of sources
tested
Leak frequency.
Test 1
(0)'
69
471
14.6
11.7
18.1
54
197
Test 2
(58)
72
474
15.2
12.2
18.7
47
206
Test3
(110)
47
482
9.8
7.4
12.7
22
208
Test 4
(148)
71
463
15.3
12.3
18.9
34
206
TestS
(190)
31
469
6.6
4.7
9.2
26
209
Test 6
(250)
33
464
7.1
5.1
9.8
28
200
LCL*'
UCL,C
27.4 22.8
21.7 17.6
34.0 29.2
10.6
7.1
15.5
16.5
12.1
22.2
12.4
8.6
17.6
Gas service
14.0
9.9
19.5
Flanges
Valves
No. of leakers
No. of sources
tested
Leak frequency.
%
LCL* %
UCL,C %
No. of leakers
No. of sources
tested
Leak frequency.
%
LCL* %
UCL,C %
11
149
7.4
4.2
12.7
13
74
17.6
10.6
27.8
14
146
9.6
5.8
15.5
13
71
18.3
11.0
28.8
13
148
8.8
5.2
14.4
6
74
8.1
3.8
16.6
13
169
7.7
4.6
12.7
6
77
7.8
3.6
16.0
7
173
4.0
2.0
8.1
2
76
2.6
0.7
9.1
9
175
5.1
2.7
9.5
7
76
9.2
4.5
17.8
*The day number in parentheses indicates the number of days from the end of Test 1 to
the end of the specified test.
"/.Diver 95 percent confidence limit.
cUpper 95 percent confidence limit.
-------
Table 3.
Source
Thirty-Day Rate of Leak Occurrence by Source and Service
Liquid service
Statistic
Test 2 Test 3 Test 4 Test 5 Test 6
(58f (110) (148) (190) (250)
Flanges
N = 362"
Valves
N = 126"
Flanges
N= 127*
Valves
N = 55*
No. of new leakers
Cum. No. leakers
%/30 days
LCLC %
UCL," %
No. of new leakers
Cum. No. leakers
%/30 days
LCL,C %
UCL," %
No. of new leakers
Cum. No. leakers
%/30 days
LCL,C %
UCL," %
No. of new leakers
Cum. No. leakers
%/30 days
LCL,C %
UCL," %
43
43
6.1
4.6
8.1
20
20
8.2
5.4
12.0
8
8
3.3
1.7
6.2
6
6
5.6
2.6
11.3
19
62
4.7
3.7
5.8
8
28
6.1
4.3
8.2
5
13
2.8
1.7
4.6
3
9
4.5
2.4
7.7
21
83
4.6
3.8
5.6
8
36
5.8
4.3
7.5
Gas service
2
15
2.4
1.5
3.8
0
9
3.3
1.8
5.7
5
88
3.8
3.2
4.6
4
40
5.0
3.8
6.4
0
15
1.9
1.2
2.9
0
9
2.6
1.4
4.5
8
96
3.2
2.7
3.8
3
43
4.1
3.2
5.1
1
16
1.5
0.9
2.3
3
12
2.6
1.6
4.1
'The day number in parentheses indicates the number of days from the end of Test 1
to the end of the specified test.
b/V = number of nonleaking sources in the specified category at the initial
screening after elimination of all sources which were not screened in all six tests.
"Lower 95 percent confidence limit.
"Upper 95 percent confidence limit.
-------
Table 4. Thirty-Day Rate of Leak Recurrence by Source and Service"
Liquid service
Source
Flanges
N = 54C
Valves
N = 31C
Statistic
No. of leakers
CNRL
%/30 days
LCL,' %
UCL,1 %
No. of leakers
CNRL"
%/30 days
LCL,9 %
UCL,' %
52(3)
14
14
15.0
9.3
22.5
5
5
9.3
4.1
18.8
Number
90(4)
16
30
18.5
14.1
22.7
7
12
12.9
7.9
18.7
of days*
132(5)
3
33
13.9
10.9
16.6
1
13
9.5
6.0
13.5
192(6)
2
35
10.1
8.0
11.9
5
18
9.1
6.4
11.5
Gas service
Flanges
N= Wc
Valves
N = 11C
No. of leakers
CNRL"
%/30 days
LCL* %
UCL,1 %
No. of leakers
CURL"
%/30 days
LCL* %
UCL,' %
2
2
11.5
3.3
29.4
1
1
5.2
0.9
21.8
2
4
13.3
5.6
22.9
1
2
6.1
1.7
15.9
0
4
9.1
3.8
15.6
1
3
6.2
2.2
12.9
0
4
6.2
2.6
10.7
0
3
4.3
1.5
8.8
'Results are based on the leakers for Test 2 which were successfully repaired
by directed maintenance (106 sources}.
^Number of days is determined from the end of Test 2 to the end of the specified
test in parentheses.
CN = number of nonleakers after directed maintenance.
Cumulative number of recurring leakers.
"Lower 95 percent confidence limit
'Upper 95 percent confidence limit.
Table 5. Repair Effectiveness of Directed Maintenance After Test 2 by
Source and Service
Gas and
Source
Flanges
Valves
Statistic
No. of attempted
repairs
No. of successful
repairs
Repair eff, %
LCL,' %
UCL* %
No. of attempted
repairs
No. of successful
repairs
Repair eff. %
LCL* %
UCL* %
Liquid service
72
59
81.9
71.1
90.0
47
37
78.7
64.2
87.9
Gas service
14
12
85.7
57.2
98.2
13
13
100.0
75.3
100
liquid service
86
71
82.6
72.5
89.7
60
50
83.3
70.0
90.6
'Lower 95 percent confidence limit.
''Upper 95 percent confidence limit.
-------
Table 6. Repair Effectiveness of Undirected Maintenance by Source and
Service'
Liquid service
Source
Flanges
Valves
Statistic
No. of
attempted
repairs
No. of
successful
repairs
Repair
eff. %
No. of
attempted
repairs
No. of
successful
repairs
Repair
eff,%
Test 1
38
18
47.4
21
8
38.1
Test 3
45
20
44.4
24
13
54.2
Test 4
71
32
45.1
34
15
44.1
TestS
31
19
61.3
26
15
57.7
Test 6
34
14
41.2
27
13
48.1
Gas service
Flanges
Valves
No. of
attempted
repairs
No. of.
successful
repairs
Repair
eff, %
No. of
attempted
repairs
No. of
successful
repairs
Repair
eff, %
11 13
13
10
36654
27.3 46.2 46.2 71.4 40.0
12 6 6 2 6
65212
50.0 83.3 33.3 50.0 33.3
"Repair effectiveness results are based on the number of successful repairs that
Allied personnel found during the rescreening after each test. A leak could
have recurred after the maintenance and before the rescreening, because the
rescreening was conducted up to 1 month after the maintenance in some tests.
-------
Table 7. Leak Rates for Leakers by Source and Service
Liquid service
Gas service
Source
Ib/h/source kg/h/sourceIb/h/source kg/h/source
Flanges, mean
LCL," %
UCL.h%
No. tested*
Valves, mean
LCL* %
UCL* %
No. tested*
0.040
0.025
0.063
70
0.047
0.030
0.076
56
0.018
0.011
0.029
0.021
0.014
0.034
0.013
0.008
0.022
0.0049
0.0027
0.0086
0.0059
0.0036
0.0100
20
0.0022
0.0012
0.0039
7
"Results were determined by bagging tests on a sample of 153 leakers
(^10,000 ppmv).
*The lower (LCL) and upper (UCL) 95 percent confidence limits are determined
by using a pooled estimate of the variance for the four data sets.
cThe number of bagging tests by source and service.
Cynthia M. Harvey and A. Carl Nelson, Jr. are with PEDCo Environment, Inc.,
1006 N. Bowen Road, Arlington, TX 76012.
Bruce A. Tic/tenor is the EPA Project Officer (see below).
The complete report, entitled "VOC Fugitive Emissions Data—High Density
Polyethylene Process Unit," (Order No. PB 81 -234 254; Cost: $11.00, subject
to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
it US GOVERNMENT PRINTING OFFICE, 1981 —757-012/7335
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
0000329
------- |