Anda di halaman 1dari 15

Proceedings of the ASME 2014 Pressure Vessels & Piping Conference

PVP2014
July 20-24, 2014, Anaheim, California, USA

PVP2014-28980

ANALYSIS AND TESTING OF A RING-JOINT FLANGE

Clifford A. Hay, P.E. Warren Brown, Ph.D., P.Eng.


ExxonMobil Baton Rouge Refinery Integrity Engineering Solutions
Baton Rouge, Louisiana, USA Dunsborough, WA, Australia

ABSTRACT were also run without a spacer (i.e., using a single gasket). The
This paper presents results from analysis and testing of an mock-up was assembled twenty-six (26) times using the
NPS 20 CL600 ring-joint (RTJ) flange through assembly and a following gaskets:
thermal cycle. Using a mock-up assembly, multiple gasket 1) No gasket
materials and types including conversion gaskets were tested to 2) Soft iron (“D” material) oval RTJ
evaluate relative performance based on amount of bolt preload 3) Soft iron (“D” material) octagonal RTJ
loss as measured through the use of load-indicating studs. To 4) 5Cr (“F5” material) octagonal RTJ
simulate field geometry, most tests included a spacer in the 5) Kammpro-ORJ (i.e., octagonal RTJ gasket with serrations
flange pair, which necessitated use of two (2) gaskets in the on the four seating surfaces) with steel core and graphite
joint. ASME PCC-1-2010 [1] Appendix O calculations and covering
finite element analysis (FEA) were also performed for 6) Kammpro-ORJ with 5Cr core and graphite covering
comparison with the test results and for evaluation of options to 7) GMGC double-rail, with graphite on the grooved surface
improve flanged joint performance. Testing and analysis inside the RTJ groove and mica on the grooved surface
results are discussed, and learnings are documented. outside the RTJ groove
8) Spiral wound (SW) conversion gasket, with 316L
windings, graphite filler, and carbon steel outer ring
INTRODUCTION 9) SW conversion gasket, with 316L windings, graphite filler,
Bolted flange connections are common in the and 316L outer ring grooved with mica covering
petrochemical industry. In some higher pressure and/or higher 10) SW conversion gasket, with Inconel X750 windings,
temperature applications, ring-type joint (RTJ) flanged vermiculite filler, and 316L outer ring
connections have been used. The RTJ gasket, which can be
oval or octagonal in cross-section, fits into grooves in the Calculations based on ASME PCC-1-2010 [1] Appendix
flange face, allowing the gasket to compress and deform into O, as well as finite element analysis (FEA), were also
the grooves and establish a seal. Recent research into RTJ performed on an NPS 20 CL600 through the same thermal
gasket behavior has been documented in [2]-[5]. Nagata and cycle as the mock-up testing. This paper documents the
Sawa [2] noted that elastic-plastic material behavior contributes methodology used in the analyses and provides comparison
significantly to bolt load loss through a thermal cycle and that with the mock-up test results. Options for improving
differential thermal expansion between the flanges and bolts is performance of flanged joints with spacers were also evaluated.
a major factor.
This paper documents testing and analysis of an NPS 20 NOMENCLATURE
CL600 RTJ flange pair through a thermal cycle. The mock-up ABS Absolute value
testing was performed using a fabricated piping assembly with Avg Average
different gasket types and materials to quantify bolt load loss. GMGC Grooved metal graphite covered gasket
Bolt load measurements were taken using SPC4 load-indicating NPS Nominal pipe size
studs (from Valley Forge and Bolt Manufacturing Company), RTJ Ring-type joint
and gas-firing was utilized to achieve the thermal cycle. For StdDev Sample standard deviation
most tests, a spacer flange was present (i.e., there were two SW Spiral wound gasket
gaskets in the joint) to simulate field geometry, but a few tests

1 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


A, B, C Intermediate variables for the definition of the quality performed at 40% (bolt load was set to prevent flange groove
factor (Q) – defined in [6] damage); for flat conversion gaskets, the studs were tightened
Q Quality factor; parameter assessing the tightening in increments of 17%, 34%, and 50% using the legacy pattern,
quality in terms of average achieved bolt load and bolt with final circular passes performed at 50%. In some cases,
load uniformity – defined in [6] multiple circular passes were necessary to achieve the desired
target stud load (discussed in more detail later). For each test,
TEST ASSEMBLY SPC4 stud load readings were taken following complete
For purposes of testing multiple gaskets, a mock-up assembly; for sixteen (16) of the tests, SPC4 stud load readings
assembly was fabricated using 1.25Cr-0.5Mo pipe, flanges, and were also taken after each bolting pass to generate tightening
spacer ring. The assembly (Fig. 1) consisted of two (2) pipe charts.
spools resting on saddle supports located in a frame, which Following assembly of the center flange pair, the pipe
allowed one pipe spool to be fixed and the other to be moved portions of the mock-up assembly were insulated (i.e., flanges
into and out of position on Hilman rollers for quick gasket were left uninsulated). Internal temperature was applied to the
changes. The pipe size was NPS 20, Schedule 80 (1.031” mock-up assembly using a gas-fired burner, which is typical of
thick); the flanges were CL600 with a ring-joint (RTJ) facing; gas-fired PWHT applications, to achieve a rapid thermal
the spacer ring was fabricated using plate material (SA-387-11, transient. Using thermocouples and manual control of the gas-
Class 2) and was a “wafer” style (i.e., outside diameter was firing, the pipe external temperature (i.e., based on a
smaller than the flange bolt circle). The welded sub-assemblies thermocouple under insulation, on the upstream side near the
were post-weld heat treated (PWHT) following fabrication and center flange pair) was brought up to 400F and held for
prior to any testing. approximately 30 minutes; following the short hold time at
For each test, the center flange pair was assembled using 400F, gas-firing was increased to bring the pipe external
twenty-four (24) 1-5/8” diameter B16 load-indicating (i.e., temperature up to 950F as quickly as possible, with a limitation
SPC4) studs with Grade 4 nuts; a flat washer per ASTM F-436 of 1300F on the internal temperature (i.e., based on a
was used underneath each nut that was torqued. Tests #1-13 thermocouple attached to a small steel block placed midway
used the same stud set, and tests #14-26 used a second stud set. through the upstream pipe spool). The temperature increase
Molybdenum-disulfide lubricant was applied to stud, nut, and from 400F to 950F on the pipe outside diameter occurred in
washer bearing surfaces prior to assembly. approximately 17 minutes, for an equivalent heating rate of
The SPC4 load-indicating studs utilize a pin and cartridge approximately 1950F per hour. The pipe external temperature
assembly in which the pin recesses from the cartridge face was maintained at 950F for approximately four (4) hours; then
under bolt load. Using a mechanical or digital reader, the bolt gas-firing was turned off and the assembly was allowed to slow
load can be measured in percentage of bolt yield strength, cool in ambient conditions with the insulation still in place.
which was 105 ksi for the B16 studs used in these tests. As an See Fig. 2 for a typical heating cycle.
example, a value of 40% on the mechanical or digital reader Final SPC4 stud load readings were taken the morning
corresponds to 42 ksi stud stress (i.e., 40% times 105 ksi); after each test, approximately 14 hours after the gas-firing was
further conversion to bolt force can be made by multiplying turned off. After recording stud load readings, the mock-up
stud stress by the stud tensile area. The bolt load readings from was disassembled and then reassembled using the next test
the SPC4 studs were the measurement basis for the test results gasket set, and the test sequence was repeated.
in this paper. No attempt was made to correlate bolt load with flange
leakage rate or probability of leakage. It is acknowledged that
MOCK-UP TEST METHOD different gaskets require different seating loads, but the
A total of twenty-six (26) tests were performed using the methodology employed in this testing looked solely at the bolt
mock-up assembly. Twenty-three (23) of the tests included the load loss through a thermal cycle for different gasket types and
spacer between the center flange pair; the other three (3) tests materials.
were performed without the spacer (i.e., only one gasket).
Three (3) of the tests which included the spacer consisted of MOCK-UP TEST RESULTS
five (5) full thermal cycles; all other tests consisted of a single Desired test objectives were to understand flanged joint
thermal cycle. behavior during assembly and to compare bolt load loss due to
For each test, the center flange pair of the mock-up was thermal cycles with different gasket types and materials. For
assembled using the applicable gaskets, and the studs were each test, SPC4 stud load readings were taken following
control torqued in a criss-cross pattern using a single tool per assembly and after the thermal cycle; stud load readings were
the legacy pattern from ASME PCC-1-2010 [1]. The studs also taken after each bolting pass during assembly for tests #10,
were tightened to a specified percentage of yield strength (i.e., #11, and #13-26. For each test, flange gap measurements were
via load-control method) depending on the gasket type; for RTJ also taken at eight (8) locations around the flange
gaskets, the studs were tightened in increments of 13%, 26%, circumference using a mechanical caliper.
and 40% using the legacy pattern, with final circular passes

2 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


GMGC and SW gaskets over years of operation would likely guidance contained in WRC Bulletin 510[8]. Two thermal
be larger than predicted in these tests. cases were run, one that modeled the mock-up test and the
other that more closely matched the maximum field operational
ANALYSIS OVERVIEW temperature case. Conduction between the nut-and-flange and
The NPS 20 CL600 flanged joint was assessed using gasket-and-flange contact surfaces were included in the
calculation methods based on ASME PCC-1-2010 and using analyses. Figure 10 shows the steady-state temperature
finite element analysis (FEA). The PCC-1 calculations were to distributions for the two (2) thermal cases; there was agreement
determine an optimum assembly bolt load to provide sufficient between the FEA results and temperatures measured using an
margin against leakage. FEA was performed to understand the infrared device.
elastic-plastic behavior of the gasket and to distinguish if An additional thermal FEA case was run with a “lug-style”
material yield or creep/relaxation was the primary contributor spacer (i.e., spacer was same outside diameter as the flange) at
to loss of bolt load observed in the mock-up testing. Further the maximum field operational temperature; see Fig. 11 for the
FEA work was done to compare actual field operating resulting temperature profile. It was desired to investigate the
conditions against the results from mock-up testing. The reduction in radial shear effect with a larger diameter spacer
effects of pressure were neglected in this FEA work since the providing cooler outside material to restrain the hotter inside
analyses were primarily for comparative purposes of thermal diameter.
effects.
MECHANICAL FEA
ASME PCC-1 CALCULATIONS The FEA model geometries used in the mechanical
An approach based on ASME PCC-1-2010 Appendix O assessment were identical to the thermal assessment. The
was used to determine optimum assembly bolt load considering elements used were first-order axisymmetric elements for all
the minimum operating gasket stress, the pressure end load, components except the bolts. The bolts where modeled with
bolt load reduction due to thermal effects, and anticipated first-order 2D Plane Stress elements such that there was no
gasket relaxation. The assessment assumes that the process additional rigidity in the tangential direction. The bolt diameter
fluid operating temperature is instantaneously applied to the was adjusted to give the same effective bending inertia as the
inner surface of the flange. The transient temperature of each actual multiple cylinder case. The element thickness was
component and the overall effect of the temperature on the calculated in order to give the same total bolt tensile area as the
gasket stress levels are determined using a methodology based actual case. The gaskets were modeled as octagonal RTJ.
on WRC Bulletin 510[8]. The method is an approximation of Mechanical properties were taken from ASME Section II-
the actual behavior and provides an indication as to the severity D[9] for the flange and bolt materials. The flange and gasket
of thermal events on the ability of the joint to seal. Figure 8 materials were modeled with temperature dependent elastic-
shows a pictorial depiction of the results, with there being little plastic behavior for an initial mechanical assessment to
margin (i.e., 1% of bolt yield strength) against joint leakage. determine the effect of material yield on the operational gasket
From the assessment, it was determined that the differential stress. For creep assessments, the flange, bolt and gasket
radial expansion of the flange and spacer resulted in significant materials were assumed to behave elastically. The gasket was
radial shear of the RTJ gasket, thereby increasing the modeled by assuming either carbon steel (ASTM A-36) or 5Cr
susceptibility of the flanged joint to leakage. (ASTM A182-F5) mechanical properties. The nut-to-flange
interfaces were modeled as contact with 0.3 friction, allowing
THERMAL FEA separation after contact in order to accurately model bolt
To simulate the thermal transient, a thermal FEA was bending effects. The gasket-to-flange contact surfaces were
performed using first-order axisymmetric heat transfer elements modeled with either 0.1 or 0.3 friction, representing the
in Abaqus (Fig. 9). Material properties for the model were possible lubricated and unlubricated gasket cases. The cases
taken from ASME Section II-D [9] for the flange, spacer and analysed are detailed in Table 5.
bolt materials, with material specifications as noted in Fig. 9. The elastic-plastic assessment of the joint involved loading
For the soft iron RTJ gasket material (i.e., “D” material), the the bolts to 37% of bolt yield strength (i.e., 39 ksi assembly
properties were based on carbon steel (ASTM A-36), since stress). The bolt length was then fixed, such that elastic
there is no standard specification for soft iron. The bolt heat interaction between the components was accurately modeled.
loss coefficients and contact areas were adjusted to equal the The thermal load steps were then applied, commencing with
actual physical conditions of the joint, once adjusted for the the transient condition and ending with the steady state
model cross section and surface areas. The transient and steady condition. The component temperatures were then reduced
state thermal analyses were performed by applying heat back to ambient condition and the residual bolt load at that
transfer film coefficients to the model surfaces in order to point gives an indication of the amount of bolt relaxation
determine the operational temperatures of the joint expected due to material yield during operation. In order to
components. The heat sink temperatures and the heat transfer isolate the amount of bolt load reduction due to yield of the
films employed were determined in accordance with the gasket, the analysis was run with both elastic-plastic properties

4 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


TABLE 1 – MOCK-UP TEST RESULTS ORGANIZED BY TEST SEQUENCE
Test Gasket # circ Assembly Initial Final Delta Delta
Gasket Spacer # of Gaskets
# Type passes1 Q2 Avg3 Avg4 Avg5 %6
Oval "D"
1 RTJ Y 2 6 65.8% 35.13 20.13 -15.00 -42.7%
(excluded from analysis)
2 Oval "D" RTJ Y 2 7 67.7% 34.96 24.75 -10.21 -29.2%
SW conversion
3 Flat Y 2 2 76.2% 37.42 33.96 -3.46 -9.2%
(316L with graphite)
SW conversion
4 Flat Y 2 2 83.0% 38.92 34.54 -4.38 -11.2%
(316L with graphite)
5 None N/A Y 0 1 89.8% 39.54 35.13 -4.42 -11.2%
6 None N/A Y 0 1 80.9% 41.21 37.54 -3.67 -8.9%
7 Oct "D" RTJ Y 2 7 77.4% 35.75 23.00 -12.75 -35.7%
8 Oct "D" RTJ Y 2 8 76.9% 36.33 24.71 -11.63 -32.0%
9 Oct "F5" RTJ Y 2 6.5 75.1% 36.04 27.04 -9.00 -25.0%
10 Oct "F5" RTJ Y 2 4 67.1% 37.71 30.25 -7.46 -19.8%
11 Kamm-ORJ stl RTJ Y 2 5.25 77.4% 36.52 30.39 -6.13 -16.8%
12 Kamm-ORJ stl RTJ Y 2 5 70.8% 34.96 29.96 -5.00 -14.3%
13 Oval "D" RTJ Y 2 8 73.1% 37.29 31.21 -6.08 -16.3%
14 GMGC double-rail Flat Y 2 1 89.4% 39.46 28.58 -10.88 -27.6%
15 GMGC double-rail Flat Y 2 1 88.8% 49.25 37.63 -11.63 -23.6%
16 Oval "D" (cycle) RTJ Y 2 8 76.9% 36.54 27.25 -9.29 -25.4%
17 GMGC double-rail (cycle) Flat Y 2 1 79.4% 48.46 36.46 -12.00 -24.8%
18 Kamm-ORJ F5 (cycle) RTJ Y 2 6 77.7% 36.79 27.63 -9.17 -24.9%
19 Oval "D" (alt tight pattern) RTJ Y 2 7 84.7% 38.88 31.92 -6.96 -17.9%
20 GMGC double-rail Flat Y 2 2 79.5% 47.54 36.71 -10.83 -22.8%
21 Kamm-ORJ F5 RTJ Y 2 7 81.3% 37.50 28.63 -8.88 -23.7%
22 Oval "D" RTJ N 1 5 71.8% 36.48 35.70 -0.78 -2.1%
23 SW (mica/GMGC outer) Flat N 1 1 85.8% 48.88 43.33 -5.54 -11.3%
24 Kamm-ORJ F5 RTJ N 1 2 76.2% 37.50 35.54 -1.96 -5.2%
25 SW (mica/GMGC outer) Flat Y 2 1 68.4% 50.92 38.04 -12.88 -25.3%
SW conversion
26 Flat Y 2 2 83.1% 49.46 42.96 -6.50 -13.1%
(X750 w/ vermiculite)
Notes:
1. “# circ passes” is the number of rotational passes required, after the 3 initial criss-cross passes, to bring the bolt load to an acceptable level.
2. “Assembly Q” is the quality factor after assembly of the flanged joint.
3. “Initial Avg” is the average bolt load for the flanged joint after assembly, measured in percentage of bolt yield strength.
4. “Final Avg” is the average bolt load for the flanged joint after the thermal cycle, measured in percentage of bolt yield strength.
5. “Delta Avg” is the difference between the “Final Avg” and the “Initial Avg”.
6. “Delta %” is the “Delta Avg” divided by the “Initial Avg”.

7 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


used for the flange and gasket and then also elastic-plastic for expansion of the joint components can significantly reduce the
the gasket and elastic for the flange. The bolts had elastic bolt load from assembly levels.
material properties applied in all cases. The creep results did not reveal that a significant effect for
The creep/relaxation assessment involved similar load either the test or operating conditions would be expected. In
steps, but the temperature was held at the steady state condition Cases #6-7 using the mock-up test temperature profile, the bolt
and creep analysis was performed for a period of 7200 hours in load loss for the duration of the test was negligible (i.e., < 1%
order to examine the reduction in bolt load attributable to of bolt yield strength, or 1 ksi). In Cases #9-10 using field
component creep. The creep/relaxation method employed was operating conditions, the bolt load loss was less significant for
as per the method determined for the ASME STP-PT-036[10]. the F5 material than the D material, as was to be expected.
Since the flange and bolt materials of the actual joint case For the additional case with the lug-style spacer (Case
contribute a relatively minor amount to the relaxation and are #12), the gasket was modeled using temperature-dependent
likely to be already strain-hardened (from previous operation), elastic-plastic material properties, and the flange was modeled
only the gasket was modeled with creep properties. The gasket elastically. As seen in Table 5, the reduction in bolt load loss
relaxation properties were taken from tests on carbon steel (for was significantly less than the comparable Case #8 at the
“D” gasket material) and 1.25Cr-0.5Mo test results (for the “F” maximum field operational temperature case. Additional
gasket material) from ASTM D60. The 1.25Cr-0.5Mo analysis not documented in this paper showed that further
properties underestimate the creep/relaxation strength of the F5 cooling of the spacer outside diameter resulted in further
material, but there were insufficient test results available for the reduction of bolt load loss. Possible spacer modifications to
F5 material grade in the required temperature range to establish reduce radial thermal expansion include using a lug-style
suitable material properties. The use of the 1.25Cr-0.5Mo spacer and adding refractory to the spacer inside diameter.
results will tend to over-predict the amount of relaxation that
should occur with the F5 gasket. COMPARISON OF ANALYSIS AND TEST RESULTS
The primary goal of the mechanical FEA was to provide FEA Cases #1 and #4 most closely represented the
comparison with the mock-up test results. Multiple cases conditions in mock-up tests #7 and #8, which used octagonal
(Table 5) were performed to understand the effects of gasket “D” RTJ gaskets. As noted above, mock-up tests with the
yield and creep/relaxation on bolt load loss through a thermal octagonal “D” RTJ gaskets produced unexpected results, and
cycle and thereby better understand these effects on the the FEA results did not provide good correlation (Fig. 14).
possiblity of flange leakage. A sample von Mises stress plot Mock-up tests #2 and #16, which used oval “D” RTJ gaskets,
and bolt stress graph are shown in Fig. 12 and 13 for FEA Case provided less bolt load loss but still more than predicted by the
#1. FEA cases (Fig. 14). One possible explanation would be the
In all cases, there was a substantial reduction in bolt load assumed material properties for soft iron in the FEA cases;
levels due to gasket material yield during the operational cycle. using stronger carbon steel material properties coupled with a
The bolt load can be seen to decrease during the initial part of higher friction coefficient in the FEA cases may have reduced
the thermal transient as the joint components heat up at the gasket plastic deformation relative to the mock-up tests.
different rates and then again at the end of the thermal cycle as FEA Cases #2 and #5, with “D” RTJ gaskets and a friction
all components return to ambient temperature. Table 5 shows coefficient of 0.1, most closely resembled the conditions in
the bolt load percentage loss for each of the cases. mock-up tests #2, 13, 16, and 19, which used oval “D” RTJ
For Cases #1-3 in which only the gasket had elastic-plastic gaskets. The FEA results from Case #2 compare favorably to
material properties, the influence of the lower friction factor tests #13 and #19, in which the gasket was inserted further into
was to reduce the bolt load loss slightly; additionally there was the flange grooves (Fig. 15); tests #2 and #16 experienced more
little difference in the results between the “D” and F5 materials, bolt load loss than predicted by the FEA caes, likely due to
indicating that the bolt load reduction due to gasket yield was undergoing less plastic deformation on assembly (Fig. 15).
primarily a thermally-driven (i.e., deflection) phenomena. In FEA Case #5, which included elastic-plastic flange and gasket
Cases #4-5, the inclusion of flange material yield in the material properties, did not match any of the field tests, again
problem resulted in a slightly higher loss in bolt load, and for likely due to the carbon steel material properties used in the
these cases the lower coefficient of friction had no effect on the FEA model.
result. FEA Case #3, with octagonal F5 gaskets, most closely
For the field operating conditions evaluated in Cases #8- represented the conditions in mock-up tests #9 and #10, which
11, there were larger reductions in bolt load through the also used octagonal F5 gaskets. The FEA results compare
thermal cycle as compared to the cases using the mock-up test favorably with the test results (Fig. 16).
temperature profile. Although the assessment likely Field data from operating flanges was compared against
overestimated the actual level of the effect (due to the use of FEA Cases #8 and #11. Figure 17 shows reasonably good
minimum yield and instantaneous application of temperature), results between the FEA cases and the field flanges, although
the results indicated that gasket yield due to radial differential the field flanges showed slightly more loss on average than the
FEA results.

5 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


[3] Currie, R., “Metallic Ring Type Joints (RTJ) – Overview
CONCLUSIONS and Recent Developments,” PVP2012-78175, Proceedings
Full-scale mock-up testing and finite element analysis were of ASME PVP2012, ASME, Toronto, 2012.
performed on an NPS 20 CL600 RTJ flange pair with wafer- [4] Kondo, K. et al, “FEM Stress Analysis and the Sealing
style spacer to investigate the effects of a thermal cycle on bolt Performance Evaluation in Bolted Flange Connections
load loss. The mock-up testing included twenty-six (26) tests, with Ring Joint Gasket Subjected to Internal Pressure:
and there were twelve (12) FEA cases modeled. Following are Effect of Scatter in Bolt Preloads,” PVP2012-78420,
a few key learnings from the testing and analysis. Proceedings of ASME PVP2012, ASME, Toronto, 2012.
1) Regardless of gasket material, flanges with RTJ [5] Kondo, K. et al, “Sealing Performance Evaluation in
gaskets exhibit a sawtooth pattern in bolt load through the Bolted Flange Connections with Ring Joint Gasket
standard three (3) criss-cross passes when using a single bolt Subjected to Internal Pressure,” PVP2013-97173,
tightening tool. The first rotational pass following the criss- Proceedings of ASME PVP2013, Paris, 2013.
cross passes results in significantly less average bolt load than [6] Lejeune, H. and Riedl, A., “Investigation on Different
the target load. In this testing, multiple rotational passes were Tightening Sequences on Several Bolted Flange Types,
required to achieve values near the target bolt load. Dimensions and Their Associated Gasket Types,”
2) Higher joint assembly quality factors are more likely PVP2011-57525, Proceedings of ASME PVP2011,
with flat gaskets than RTJ gaskets. Baltimore, 2011.
3) For flanged joints with RTJ gaskets, the bolt load loss [7] Robinson, J., Lundin, M., and Spiewak, I., “Development
through a thermal cycle is dependent upon how much gasket of Ring-Joint Flanges for Use in the HRE-2,” ORNL-3165,
deformation occurs during assembly. Oak Ridge National Laboratory, Oak Ridge, Tennessee,
4) In short-term testing, identical thermal cycles in which 1961.
the flanges and bolts behave elastically will not result in [8] Brown, W., “Analysis of the Effects of Temperature on
progressive bolt load loss. Bolted Joints,” WRC Bulletin 510, Welding Research
5) ASME PCC-1-2010 Appendix O calculations can be Council, Shaker Heights, Ohio, 2006.
effective in determining key flange joint factors which may [9] “Material Properties,” ASME Boiler and Pressure Vessel
make a flange more susceptible to leakage. Code, Section II PartD, ASME, New York, 2013.
6) FEA can be used to reasonably model the behavior of [10] “Bolted Flange Connections in Elevated Temperature
bolted flange joints with RTJ gaskets, although it is highly Service,” ASME STP-PT-036, ASME, New York, 2010.
dependent upon the gasket material properties due to the
significant deformation experienced by the gasket during
assembly and operation.
7) Radial shear of RTJ gaskets can be a significant
contributor to bolt load loss through a thermal cycle.
8) For flanges with spacers, making design modifications
to reduce the radial shear effect reduces the bolt load loss
through a thermal cycle.

ACKNOWLEDGMENTS
The authors would like to thank the ExxonMobil Baton
Rouge Refinery for support and funding of this project, as well
as the coworkers who contributed their time and effort in
collecting and assessing the data. Additional thanks are offered
to the contractors (Analytic Stress, Furmanite, Lamons, Nugent
Steel, Ohmstede Ltd, and Valley Forge and Bolt) who
contributed to the mock-up test activities.

REFERENCES
[1] “Guidelines for Pressure Boundary Bolted Flange Joint
Assembly”, ASME PCC-1, ASME, New York, 2010.
[2] Nagata, S. and Sawa, T., “Effects of Temperature Change
on Bolt Load and Gasket Load of Bolted Flange
Connection with Ring Type Joint Gasket,” PVP2008-
61418, Proceedings of ASME PVP2008, ASME, Chicago,
2008.

6 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


TABLE 1 – MOCK-UP TEST RESULTS ORGANIZED BY TEST SEQUENCE
Test Gasket # circ Assembly Initial Final Delta Delta
Gasket Spacer # of Gaskets
# Type passes1 Q2 Avg3 Avg4 Avg5 %6
Oval "D"
1 RTJ Y 2 6 65.8% 35.13 20.13 -15.00 -42.7%
(excluded from analysis)
2 Oval "D" RTJ Y 2 7 67.7% 34.96 24.75 -10.21 -29.2%
SW conversion
3 Flat Y 2 2 76.2% 37.42 33.96 -3.46 -9.2%
(316L with graphite)
SW conversion
4 Flat Y 2 2 83.0% 38.92 34.54 -4.38 -11.2%
(316L with graphite)
5 None N/A Y 0 1 89.8% 39.54 35.13 -4.42 -11.2%
6 None N/A Y 0 1 80.9% 41.21 37.54 -3.67 -8.9%
7 Oct "D" RTJ Y 2 7 77.4% 35.75 23.00 -12.75 -35.7%
8 Oct "D" RTJ Y 2 8 76.9% 36.33 24.71 -11.63 -32.0%
9 Oct "F5" RTJ Y 2 6.5 75.1% 36.04 27.04 -9.00 -25.0%
10 Oct "F5" RTJ Y 2 4 67.1% 37.71 30.25 -7.46 -19.8%
11 Kamm-ORJ stl RTJ Y 2 5.25 77.4% 36.52 30.39 -6.13 -16.8%
12 Kamm-ORJ stl RTJ Y 2 5 70.8% 34.96 29.96 -5.00 -14.3%
13 Oval "D" RTJ Y 2 8 73.1% 37.29 31.21 -6.08 -16.3%
14 GMGC double-rail Flat Y 2 1 89.4% 39.46 28.58 -10.88 -27.6%
15 GMGC double-rail Flat Y 2 1 88.8% 49.25 37.63 -11.63 -23.6%
16 Oval "D" (cycle) RTJ Y 2 8 76.9% 36.54 27.25 -9.29 -25.4%
17 GMGC double-rail (cycle) Flat Y 2 1 79.4% 48.46 36.46 -12.00 -24.8%
18 Kamm-ORJ F5 (cycle) RTJ Y 2 6 77.7% 36.79 27.63 -9.17 -24.9%
19 Oval "D" (alt tight pattern) RTJ Y 2 7 84.7% 38.88 31.92 -6.96 -17.9%
20 GMGC double-rail Flat Y 2 2 79.5% 47.54 36.71 -10.83 -22.8%
21 Kamm-ORJ F5 RTJ Y 2 7 81.3% 37.50 28.63 -8.88 -23.7%
22 Oval "D" RTJ N 1 5 71.8% 36.48 35.70 -0.78 -2.1%
23 SW (mica/GMGC outer) Flat N 1 1 85.8% 48.88 43.33 -5.54 -11.3%
24 Kamm-ORJ F5 RTJ N 1 2 76.2% 37.50 35.54 -1.96 -5.2%
25 SW (mica/GMGC outer) Flat Y 2 1 68.4% 50.92 38.04 -12.88 -25.3%
SW conversion
26 Flat Y 2 2 83.1% 49.46 42.96 -6.50 -13.1%
(X750 w/ vermiculite)
Notes:
1. “# circ passes” is the number of rotational passes required, after the 3 initial criss-cross passes, to bring the bolt load to an acceptable level.
2. “Assembly Q” is the quality factor after assembly of the flanged joint.
3. “Initial Avg” is the average bolt load for the flanged joint after assembly, measured in percentage of bolt yield strength.
4. “Final Avg” is the average bolt load for the flanged joint after the thermal cycle, measured in percentage of bolt yield strength.
5. “Delta Avg” is the difference between the “Final Avg” and the “Initial Avg”.
6. “Delta %” is the “Delta Avg” divided by the “Initial Avg”.

7 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


TABLE 2 – MOCK-UP TEST RESULTS ORGANIZED BY GASKET TYPE AND MATERIAL
Gasket Type Number of Tests Avg of Delta %1 Max of Delta %2 Min of Delta %3 Avg # of Circ Passes4
RTJ Gaskets 14
Oval D 5
Spacer 4 -22.21% -16.31% -29.20% 7.5
Tests #2 & #16 2 -27.31% -25.43% -29.20% 7.5
Tests #13 & #19 2 -17.11% -16.31% -17.90% 7.5
No Spacer 1 -2.15% -2.15% -2.15% 5
Octagonal 4
D 2 -33.83% -32.00% -35.66% 7.5
F5 2 -22.38% -19.78% -24.97% 5.25
Kamm-ORJ 5
Steel w/ Graphite 2 -15.54% -14.30% -16.79% 5.125
F5 with Graphite 3
Spacer 2 -24.29% -23.67% -24.92% 6.5
No Spacer 1 -5.22% -5.22% -5.22% 2
Flat Gaskets 9
GMGC (double-rail) 4 -24.68% -22.79% -27.56% 1.25
SW Conversion 5
316L w/ Graphite 4
Normal O.R. 2 -10.24% -9.24% -11.24% 2
GMGC/mica O.R. 2
Spacer 1 -25.29% -25.29% -25.29% 1
No Spacer 1 -11.34% -11.34% -11.34% 1
X750 w/ vermiculite 1 -13.14% -13.14% -13.14% 2
No Gasket 2 -10.03% -8.90% -11.17% 1
Grand Total 25
Notes:
1. “Avg of Delta%” is the average percentage of bolt load loss through a thermal cycle for the tests represented on a given line in the table.
2. “Max of Delta%” is the maximum percentage of bolt load loss through a thermal cycle for the tests represented on a given line in the table.
3. “Min of Delta%” is the minimum percentage of bolt load loss through a thermal cycle for the tests represented on a given line in the table.
4. “Avg # of Circ Passes” is the average number of rotational passes required, after the 3 initial criss-cross passes, to bring the bolt load to an
acceptable level, for the tests represented on a given line in the table.

TABLE 3 – MOCK-UP TEST RESULTS FOR RTJ GASKET ASSEMBLY AFTER FIRST CIRCULAR PASS
Test # Gasket Spacer Actual Bolt Load after 1st Circ Pass1 Target Bolt Load1 % of Target
10 Oct "F5" Y 30 40 75%
11 Kamm-ORJ stl Y 27 40 68%
13 Oval "D" Y 22 40 55%
16 Oval "D" Y 23 40 58%
18 Kamm-ORJ F5 Y 27 40 68%
19 Oval "D" Y 29 40 73%
21 Kamm-ORJ F5 Y 27 40 68%
22 Oval "D" N 24 40 60%
24 Kamm-ORJ F5 N 33 40 83%
Note:
1. Units for bolt loads are in percentage of bolt yield strength.

8 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


TABLE 4 – FLANGE GAP DISTANCES
Gasket Types and Test Numbers Assembly Gap (inches)
RTJ Gaskets
Oval “D”
13 4.325
16 4.373
19 4.308
Octagonal “D”
7 4.415
8 4.389
Octagonal F5
9 4.423
10 4.425
Kamm-ORJ Steel w/ Graphite
11 4.352
12 4.393
Kamm-ORJ F5 w/ Graphite
18 4.385
21 4.358
No Gasket
5 4.130
6 4.134

TABLE 5 – FEA CASE LISTING WITH BOLT LOAD LOSS RESULTS


Case # Gasket Material Gasket Friction Analysis Type Temperature Case % Bolt Load Loss
1 D 0.3 Elastic-Plastic (Gasket Only) Mock-up Test 21%
2 D 0.1 Elastic-Plastic (Gasket Only) Mock-up Test 18%
3 F5 0.3 Elastic-Plastic (Gasket Only) Mock-up Test 21%
4 D 0.3 Elastic-Plastic (Gasket & Flange) Mock-up Test 22%
5 D 0.1 Elastic-Plastic (Gasket & Flange) Mock-up Test 23%
6 D 0.3 Creep (Gasket Only) Mock-up Test
7 D 0.1 Creep (Gasket Only) Mock-up Test
8 D 0.3 Elastic-Plastic (Gasket Only) Operational 64%
9 D 0.3 Creep (Gasket Only) Operational
10 F5 0.3 Creep (Gasket Only) Operational
11 D 0.3 Elastic-Plastic (Gasket & Flange) Operational 70%
12 D 0.3 Elastic-Plastic (Gasket Only) Operational 45%

9 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


FIGURE 1 – MOCK-UP TEST ASSEMBLY FIGURE 2 – TYPICAL THERMAL CYCLE CHART

FIGURE 3 – OVAL RTJ TIGHTENING PATTERN (FROM TEST #16)

10 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


FIGURE 4 – KAMMPRO ORJ (F5 WITH GRAPHITE) TIGHTENING PATTERN (TEST #21)

FIGURE 5 – GMGC DOUBLE-RAIL TIGHTENING PATTERN (TEST #15)

11 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


FIGURE 6 – SPIRAL WOUND CONVERSION GASKET (WITH GMGC OUTER RING) TIGHTENING PATTERN (TEST #25)

FIGURE 7 – NPS 12 CL1500 FLANGE ASSEMBLY IN FIELD (347SS OVAL RTJ GASKET)

12 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


FIGURE 8 – ASME PCC-1-2010 APPENDIX O CALCULATIONS FOR “D” MATERIAL RTJ GASKET

MOCK-UP TEST TEMPERATURE DISTRIBUTION

FIGURE 9 – FINITE ELEMENT MODEL MESH

OPERATING TEMPERATURE DISTRIBUTION

FIGURE 10 – FINITE ELEMENT MODEL TEMPERATURE


DISTRIBUTIONS

13 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


FIGURE 11 – FINITE ELEMENT MODEL TEMPERATURE
DISTRIBUTION FOR FULL DIAMETER SPACER CASE
(CASE #12)

FIGURE 12 – VON MISES STRESS, BEFORE AND AFTER THERMAL TRANSIENT (FEA CASE #1)

14 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms


FIGURE 13 – BOLT STRESS PLOT (FEA CASE #1)

FIGURE 14 – FEA & TEST RESULTS COMPARISON FIGURE 15 – FEA & TEST RESULTS COMPARISON

FIGURE 16 – FEA & TEST RESULTS COMPARISON FIGURE 17 – FEA & FIELD RESULTS COMPARISON

15 Copyright © 2014 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms

Anda mungkin juga menyukai