MIDWEST RESEARCH INSTITUTE
Suite 350
401 Harrison Oaks Boulevard
Cary, North Carolite 27513-2412
Telephone JJB19) 677-0249
FAX (6191677-0065
Date: February 23, 1996
Subject: Final Deck Fitting Loss Factors for AP-42 Section 7.1
EPA Contract 68-D2-0159; Work Assignment No. Ill-01
MRI Project No. 4603-01-02
From: Amy Parker
To: Dennis Beauregard
EFIG (MD-14)
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
I. Background and Introduction
The purpose of this memo is to document recommendations for
the final deck fitting loss factors for incorporation into the
February 1996 version of AP-42 Section 7.1, Organic Liquid
Storage Tanks. As documented in a May 25, 1994 memo from
R. Jones and D. Wallace, MRI, to A. Pope, EPA, and memos from
A. Parker, MRI, to the project file, analyses were conducted by
the American Petroleum Institute (API) and MRI to determine loss
factors for deck fittings under various control configurations.
Several discussions were held between EPA, API, and MRI
regarding the best way to analyze the data, particularly the
slotted guidepole data, and present the factors in the AP-42
document and in Chapter 19.2 of API's Manual of Petroleum
Measurement Standards. In an April 27, 1995, conference call, it
was recommended that data for several of the slotted guidepole
configurations be combined when developing the fitting loss
factors, because the fittings either exhibited similar levels of
control or were similar in configuration. Data for slotted
guidepoles with gasketed and ungasketed sliding covers were
combined, as were data for configurations that varied only with
the height of the float (at or 1 inch above the sliding cover)
during testing. These factors were published in the draft
Section 7.1.
During a conference call held on November 29, 1995,
combining fitting Nos. 24 {gasketed, with float at cover
elevation, pole sleeve, and pole wiper at sliding cover
elevation) and 29 (gasketed, with float 1 inch above cover, pole
sleeve, and pole wiper 6 inches above cover elevation) into a
single set of loss factors was discussed. It had been decided in
the April conference call that float height was not significant
-------
(between 0 and 1 inch above the sliding cover) and it was noted
in the November call that the pole wiper had been elevated only
for ease of construction of the test assembly. Also, the
evaporative loss factors developed separately for the two fitting
configurations showed comparable emissions. Therefore, it was
recommended that the data for these fittings be combined to
develop a single set of loss factors.
During the call, participants also recommended a footnote
to the loss factor table that cautions the user against applying
the slotted guidepole loss factors to configurations where the
float wiper is below the sliding cover, unless a pole sleeve is
employed, since the fitting tests for configurations without a
pole sleeve were only conducted for float wipers at or 1 inch
above the sliding cover. It was postulated that when a pole
sleeve is used, the height of the float wiper (above or below the
sliding cover) is not critical, since the pole sleeve restricts
the flow of vapor from the well vapor space into the slotted
guidepole. Data from a slotted guidepole configuration tested by
Chicago Bridge and Iron (CBI) (fitting No. COD that included a
gasket, "short" float (5 inches below the sliding cover
elevation), pole sleeve, and pole wiper confirmed this
assumption. These data were analyzed by MRI in January 1996 and
combined with the other slotted guidepole, gasketed with float,
pole sleeve, pole wiper data (fitting Nos. 24 and 29) to develop
a single set of loss factors for a slotted guidepole with a
gasketed sliding cover, float, pole sleeve, and pole wiper. This
single set of loss factors is recommended for use in AP-42.
In a January 23, 1996 meeting of API's Committee on
Evaporative Loss Measurement (CELM) attended by EPA and MRI, an
additional test (denoted as fitting No. 31) for a slotted
guidepole with a pole sleeve was introduced by CBI. This test
was a duplicate of fitting No. 2, but was conducted after the
initial test program was completed. The data for fitting No. 31
were analyzed by MRI and combined with the data for fitting No. 2
to obtain a revised set of loss factors for a slotted guidepole
with a gasketed sliding cover and pole sleeve. This revised set
of loss factors is recommended for use in AP-42.
Also discussed at the January meeting was the designation
of all slotted guidepole configurations as "gasketed or
ungasketed." Only two of the configurations (with and without a
float) were tested in both the gasketed and ungasketed condition.
These two cases provided the basis for combining the gasketed and
ungasketed data. Therefore, some API committee members reasoned
that only those two configurations should be grouped and
designated as "gasketed or ungasketed," and the other four
configurations should be designated only as "gasketed." Upon
reviewing the underlying reasons for grouping the configurations
and the previous data analyses, there is no strong reason to
object to API's proposal and MRI recommends that only the two
-------
configurations that were actually tested in both conditions be
designated as "gasketed or ungasketed."
Royce Laverman of CBI also presented a method for
developing loss factors for center-area deck legs with gaskets
and with socks (test data are available only for ungasketed
center-area deck legs). He applied "control efficiencies" for
gaskets and for socks (based on a comparison of the loss factors
for pontoon-area deck legs with gaskets and socks to the
ungasketed pontoon-area deck leg loss factors) to the ungasketed
center-area deck leg factors and obtained new factors for
center-area deck legs with gaskets and with socks. For example,
applying a gasket to a pontoon-area deck leg reduces KFa by
35 percent, Kpb by 78 percent, and m by 29 percent from the
factors for the ungasketed pontoon-area deck leg. These
reductions were applied to the loss factors for the ungasketed
center-area deck leg to obtain loss factors for a gasketed
center-area deck leg. The same method was applied to obtain
factors for an ungasketed center-area deck leg with a sock.
The justification for estimating loss factors for these
configurations is that currently, facilities receive no credit
for controlling their center-area deck legs on floating roof
tanks. The CBI method was reviewed by MRI, but this approach is
not recommended due to reservations about its mathematical
validity and inconsistency with the methodology used for
estimating rim seal loss factors.
The methodology was discussed with Mr. Eob Ferry of TGB; we
agreed that a slight modification to the extrapolation method
used with rim seal loss factors is more appropriate; the
methodology and results are discussed in attachment 1.
The results obtained using this method are similar to the
CBI results, as shown in Table 1. The method presented here is
recommended to maintain consistency. However, because test data
for these configurations are not currently available, no
conclusion can be drawn with respect to the accuracy of either
method.
II. Deck Fitting Loss Factors
Table 2 presents the recommended final deck fitting loss
factors for floating roof tanks for AP-42 Section 7.1, Organic
Liquid Storage Tanks, and the software program TANKS 3.0.
Table 3 indicates the slotted guidepole configurations that were
combined to develop a single set of factors.
-------
TABLE 1. DECK LEG LOSS FACTORS, KFa, Kph, AND m
Configuration
Method
Center- area leg, ungasketed
Center- area leg,
gasketed
Center- area leg, with
sock
CBI
TGB/MRI
CBI
TGB/MRI
KFS
0.82
0.53
0.53
0.49
0.49
^Fb
0.53
0.11
0.11
0.20
0.16
m
0.14
0.10
0.13
0.10
0.14
-------
TABLE 2. DECK FITTING LOSS FACTORS, KFa, K^,
AND m, AND TYPICAL NUMBER OF DECK FITTINGS, NFa
Fitting Type And Construction Details
Access hatch (24-inch diameter well)
Bolted cover, gasketed
Unbolted cover, ungasketed
Unbolted cover, gasketed
Fixed roof support column well
Round pipe, ungasketed sliding cover
Round pipe, gasketed sliding cover
Round pipe, flexible fabric sleeve seal
Built-up column, ungasketed sliding cover*
Built-up column, gasketed sliding cover
Unslotted guide-pole and well (8-inch
diameter unslotted pole, 21-inch
diameter well)
Ungasketed sliding cover1*
Ungasketed sliding cover w/pole sleeve
Gasketed sliding cover
Gasketed sliding cover w/pole wiper
Gasketed sliding cover w/pole sleeve
Slotted guide-pole/sample well (8-inch
diameter slotted pole, 21 -inch
diameter well)*
Ungasketed or gasketed sliding cover
Ungasketed or gasketed sliding cover,
with floats
Gasketed sliding cover, with
pole wiper
Gasketed sliding cover, with
pole sleeve
Gasketed sliding cover, with
float and pole wiper8
Gasketed sliding cover, with
float, pole sleeve, and pole wiper**
Gauge-float well (automatic gauge)
Unbolted cover, ungasketecr
Unbolted cover, gasketed
Bolted cover, gasketed
Gauge-hatch/sample port
Weighted mechanical actuation,
gasketedb
Weighted mechanical actuation,
ungasketed
Slit fabric seal, 10% open area'
Vacuum breaker
Weighted mechanical actuation,
ungasketed
Weighted mechanical actuation, gasketed b
Deck drain (3-inch diameter)
Openb
90% closed
Sbb Drain (1-inch diameter^
%
(Ib-mole/yr)
1.6
36°
31
31
25
10
47
33
31
25
25
14
8.6
43
31
41
11
21
11
14e
4.3
2.8
0.47
2.3
12
7.8
6.2s
1.5
1.8
1.2
Loss Factors
Kpb
(lb-mole/(mph)m-yr)
0
5.9
5.2
150
2.2
13
3.7
12
270
36
48
46
7.9
9.9
5.4
17
0
0.02
0
0.01
1.2
0.21
0.14
m
(djmcnsionless)
0
1.2
1.3
1.4
2.1
2.2
0.78
0.81
1.4
2.0
1.4
1.4
1.8
0.89
1.1
0.38
0
0.97
0
4.0
0.94
1.7
1.1
Typical Number Of
Fittings, NF
1
NC
(Table 7. 1-11)
I
f
1
1
Nvb (Table 7.1-13)*
Nd (Table 7.1-13)
Nd (Table 7. 1-15)
-------
TABLE 2 (CONT.).
Fitting Type And Construction Details
Deck leg (3-inch diameter)
Adjustable, internal floating deck0
Adjustable, pontoon area - ungasketed"
Adjustable, pontoon area - gasketed
Adjustable, pontoon area - sock, ungask.
Adjustable, center area - ungasketed"
Adjustable, center area - gasketed01
Adjustable, center area - sock, ungask.
Adjustable, double-deck roofs
Fixed
Rim Vent"
Weighted mechanical actuation, ungasketed
Weighted mechanical actuation, gasketed
Ladder well
Sliding cover, ungasketed0
Sliding cover, gasketed
(Ib-mole/yr)
7.9
2.0
1.3
1.2
0.82
0.53
0.49
0.82
0
0.68
0.71
76
56
Loss Factors
(lb-moie/(mph)m-yr)
0.37
0.08
0.14
0.53
0.11
0.16
0.53
0
1.8
0.10
m
(dimensionless)
0.91
0.65
0.65
0.14
0.13
0.14
0.14
0
1.0
1.0
Typical Number Of
Fittings, NF
Nj (Table 7.1-15),
(Table 7.1-14)
1
ld
Note: The deck fitting loss factors, KFa,
15 miles per hour.
, and m, may only be used for wind speeds below
Reference 5, unless otherwise indicated.
blf no specific information is available, this value can be assumed to represent the most common or
typical roof fitting currently in use for external and domed external floating roof tanks.
clf no specific information is available, this value can be assumed to represent the most common or
typical roof fitting currently in use for internal floating roof tanks.
dColumn wells and ladder wells are not typically used with self supported fixed roofs,
"References 16,20.
fA slotted guide-pole/sample well is an optional fitting and is not typically used.
gTests were conducted with floats positioned with die float wiper at and 1 inch above the sliding
cover. The user is cautioned against applying these factors to floats that are positioned with the
wiper or top of the float below the sliding cover ("short floats"). The emission factor for such a
float is expected to be between the factors for a guidepole without a float and with a float, depending
upon the position of the float top and/or wiper within the guidepole.
hTests were conducted with float wipers positioned at varying heights with respect to the sliding
cover. This fitting configuration also includes a pole sleeve which restricts the airflow from the well
vapor space into the slotted guidepole. Consequently, the height of the float within the guidepole (at,
above, or below the sliding cover) is not expected to significantly affect emission levels for this
fitting configuration.
JNvb 1 for internal floating roof tanks.
kStub drains are not used on welded contact internal floating decks.
These loss factors were projected using the results from pontoon-area deck legs.
"Rim vents are used only wkh mechanical-shoe primary seals.
-------
TABLE 3. SLOTTED GUIDEPOLE CONFIGURATION COMBINATIONS
Fitting No.
1
25
3
26
20
2
31
23
4
24
29
C01
Configuration
Ungasketed sliding cover
Gasketed sliding cover
Ungasketed sliding cover, with float 1 inch above sliding cover
Gasketed sliding cover, with float 1 inch above sliding cover
Gasketed sliding cover, with pole wiper
Gasketed sliding cover, with pole sleeve
Gasketed sliding cover, with pole sleeve
Gasketed sliding cover, with float at cover elevation, pole wiper
Gasketed sliding cover, with float 1 inch above sliding cover, pole wiper
Gasketed sliding cover, with float at cover elevation, pole sleeve, pole wiper
Gasketed sliding cover, with float 1 inch above cover, pole sleeve, pole
wiper six inches above sliding cover
Gasketed sliding cover, with float 5 inches below cover, pole sleeve, pole
wiper
-------
Attachment 1. Deck leg loss factors extrapolation methodology
The methodology recommended for extrapolating loss factors
for center-area deck legs with gaskets and with socks is
illustrated below. First, loss factors for the controlled
center-area deck leg are projected by applying the percent
reduction achieved by gaskets and socks on pontoon-area deck legs
to the ungasketed center-area deck leg, using the following
equation.
Exo -
(1)
where:
E
xc
Eye/ Ey
the emissions level at a given wind
speed for the control configuration
under consideration, determined by
factoring the emissions level of the
uncontrolled case, Ex, by the ratio of
the emission levels from a similar
device, with and without the control
the emissions level at the indicated
wind speed for the device under
consideration, without the subject
control; and
the emissions levels at the indicated
wind speed for the similar device, with
(Eyc) and without (Ey) the subject
control.
Using the rim seal methodology, values of Exc are projected
at three wind speeds (0, vi( and Vj), and the form of the loss
equation is assumed to be:
E = K
Kbv*
(2)
The zero miles per hour value for E is assigned to Ka/ and
using the values of Exc and v at wind speeds
values for Kb and m are determined as follows:
and
the
Exc -
(3)
A-l
-------
E = Bxe - K (4)
log(EneC) = log Kb + m log v (5)
Having projected values for E and v at wind speeds v,^ and
v j , as well as for Ka (v = 0), the slope, m, of the log(Enet)
versus log{v) curve is determined as follows:
log IS ,/E )
Finally, Kb is determined as follows:
= * W
When using the above method to estimate rim seal loss
factors, the wind speeds selected were vi Ŧ 4 mph and
Vj = 10 mph. When this method is applied to deck leg loss
factors, however, the estimated emissions decrease at high wind
speeds. This result was deemed unrealistic, and the method was
modified by setting v^ ŧ 0 mph and Vj = 4 mph. Since using
VA = 0 results in dividing by zero to determine m, zero was
approximated as 1 x io~10" (\rą = i x 10~100 mph) . The results are
summarized in Tables 1 and 2.
A-2
-------
Table 1. Extrapolation of loss factors for center-area deck legs, gasketed
Center-area, ungasketed
Pontoon-area, gasketed
Pontoon-area, ungasketed
Ex
Eye
Ey
Exc
Ka
0.82
1,3
2
0.5330
Ka
0.53
Kb
0.53
0.08
0.37
Kb
0.11
m
0.14
0.65
0.91
m
0.13
Ex =
Eyc =
Ey =
Exc =
Enet =
VI
1E-100
0.8200
1.3000
2.0000
0.5330
3.44E-15
vj
4
1.4635
1.4970
3.3064
0.6626
0.1296
Table 2. Extrapolation of loss factors for center-area deck legs, sock
vi vj
Ka Kb m 1E-100 4
Center-area, ungasketed Ex 0.82 0.53 0.14 Ex= 0.8200 1.4635
Pontoon-area, sock Eye 1.2 0.14 0.65 Eyc= 1.2000 1.5447
Pontoon-area, ungasketed Ey 2 0.37 0,91 Ey = 2.0000 3.3Q64
Exc 0.4920 Exc= 0.4920 0.6837
Enet= 3.22E-15 0.1917
Ka Kb m
0.49 0.16 0.14
A-3
------- |