Proceedings of the ASME 2019 Pressure Vessels & Piping Conference

PVP2019
July 14-19, 2019, San Antonio, Texas, USA

PVP2019-93027

Downloaded from http://asmedigitalcollection.asme.org/PVP/proceedings-pdf/PVP2019/58936/V002T02A019/6446092/v002t02a019-pvp2019-93027.pdf by ExxonMobil Upstream Research user on 05 January 2022

PERFORMANCE OF SEMI-METALLIC GASKETS WITH NUBBINS

Robert Taylor¹ David Fairbanks

3S – Superior Sealing Services LLC Sinclair Wyoming Refining Company
Houston, TX Rawlins, WY

ABSTRACT Due to this lack of test data, it was decided to

conduct a test program to understand how different
Nubbins are predominately machined into a heat semi-metallic gaskets would perform when a nubbin is
exchanger flange to create a high stress point to present.
produce a tighter seal.
Traditionally the gasket used when a nubbin is 1.1 FLANGE DESIGN WITH NUBBINS
present is a Double Metal Jacket (DMJ), which
consists of a metal jacket with a soft internal sealing
Flanges with nubbins are designed to produce a
material e.g. Graphite, PTFE. The metal jacket is
higher stress area where the nubbin is positioned.
commonly soft iron, carbon steel or a 300-series
Nubbin dimensions are usually 1/64” high by 1/8”
austenitic stainless steel.
wide.
This paper will compare the performance of a
DMJ and Kammprofile gasket, performance being –
leakage and temperature cycling and the affect the
FIGURE 1 & 2: shows a typical flange setup with a
nubbin has on the gasket after testing.
DMJ and nubbin.

INTRODUCTION

One question that is frequently asked is “What

type of gasket can I use on my heat exchanger when
the old gasket is a DMJ and one of the flanges has a
nubbin machined in?”.
DMJ gaskets have limitations on leakage rates due
to the large area of metal to metal contact.
One option available to the end user is to machine
the nubbin off the flange, which is both time
consuming and expensive. More often than not, a DMJ
gasket will be used again because of limited test data FIGURE 1: Drawing of a DMJ and a Nubbin Flange
and field applications on alternative semi-metallic (Type1)
gasket i.e. Kammprofile, CMG, SWG with nubbins.

¹Contact author email: robert.taylor@3sgaskets.com

1 Copyright © 2019 ASME

 A load versus leakage rate test (FIGURE 3) was
chosen to understand the leakage performance of the
selected semi-metallic gaskets at ambient
temperatures with Helium Gas. Test results will reveal

Downloaded from http://asmedigitalcollection.asme.org/PVP/proceedings-pdf/PVP2019/58936/V002T02A019/6446092/v002t02a019-pvp2019-93027.pdf by ExxonMobil Upstream Research user on 05 January 2022

the comparison between nubbin and flat face flanges.
Temperature cycling versus leakage testing
demonstrates the performance of the gaskets in
temperature cycling environments.

FIGURE 2: Drawing of a DMJ and a Nubbin Flange

(Type 2)

1.2 STUDY

Sinclair Refining were having issues with leakage

on their Poly Reactors using DMJ gaskets with
Nubbins. This had been occurring on a regular basis
since the units were commissioned costing the refinery
lost production and downtime.
The following test procedures were selected in
consultation with the customer.
FIGURE 3: Typical graph of leakage vs gasket
1. Understanding the leakage rate of a DMJ gasket
stress.
against the preferred gasket type, a kammprofile
in Nubbin environments.
2. Understanding the leakage of the selected
Temperature cycling versus leakage is designed to
gaskets in temperature cycling conditions. This
better gauge the gasket performance (leakage, pressure
would replicate the Poly Unit conditions.
loss) at different temperatures. FIGURE 4: illustrates
this.

1.3 GASKET SELECTION

The following gaskets were selected due to our

customers’ requirements.

- Double Metal Jacketed (DMJ)

- Kammprofile – Flexible Graphite Facing.

All gaskets tested were 316SS core or Jacket –

Inhibited Graphite sealing material.

1.4 TEST PROCEDURES FOR THIS PROJECT

Testing was carried out to the following FIGURE 4: Temperature Cycling Graph
specifications:

- Leakage at increasing gasket stresses (custom

designed per application)
- Temperature Cycling vs Leakage (custom
designed per application)

2 Copyright © 2019 ASME

1.5 LEAKAGE RESULTS
The Kammprofile gasket was not tested to
FIGURE 5 shows leakage results for the selected FIGURE 6 due to existing data showing similar or

Downloaded from http://asmedigitalcollection.asme.org/PVP/proceedings-pdf/PVP2019/58936/V002T02A019/6446092/v002t02a019-pvp2019-93027.pdf by ExxonMobil Upstream Research user on 05 January 2022

gaskets with a nubbin. At 14500 psi (100 MPa) of
gasket stress, the Kammprofile gasket is 149000 times
tighter than the DMJ gasket; 0.3 gallons per min. for
the DMJ and 0.000002 gallons per min. for the
Kammprofile gasket.
better results than shown in FIGURE 5.
The leakage results confirm a kammprofile gasket
provides a far tighter seal at all stress points as shown
in FIGURE 5.
1.6 LEAKAGE RESULTS VS TEMPERATURE
CYCLING

FIGURE 5: Leakage curve: DMJ vs Kammprofile

FIGURE 7: Temperature Cycling vs Leakage
The same test procedure as Figure. 5 was carried
out on a DJ gasket with standard flat face flanges to
FIGURE 7 shows leakage (loss of internal
compare. There was very little difference between
pressure) over two thermal cycles.
leakage rates of the DMJ gaskets for both flange types.
Even with the nubbin, which is supposed to create
Test Parameters:
a high stress point to reduce leakage the results were
Maximum Temperature: 490 Deg F (254 Deg C)
similar (within experimental error), see FIGURE 6.
Maximum Pressure: 400 psi (27.5 bar)
Flange Dimensions: 4” 300lb
Gasket Dimensions: 4.75” x 5.75”
(120mm x 146mm)
Gasket Stress: 23100 psi (160 MPa)
Bolt Stress: 79000 psi (545 MPa)
Torque 253 ft-lbs (343 Nm)
Stud Material: B7
K Factor: 0.17

Illustrated above in FIGURE 7, the Kammprofile

gasket loses a minimal amount of pressure, whereas
the DMJ loses a high percentage over the first
temperature rise.
After both tests were complete, the relaxed bolt
stress was measured to roughly calculate the
remaining gasket stress (TABLE 1).
FIGURE 6: Leakage of DMJ with and without a
nubbin.

3 Copyright © 2019 ASME

TABLE 1: COMPARISON OF THE REMAINING
GASKET STRESS UNDER TEMPERATURE
CYCLING CONDITIONS

Downloaded from http://asmedigitalcollection.asme.org/PVP/proceedings-pdf/PVP2019/58936/V002T02A019/6446092/v002t02a019-pvp2019-93027.pdf by ExxonMobil Upstream Research user on 05 January 2022

DMJ Kammprofile
Initial Bolt 79000 psi 79000 psi
Stress (545 MPa) (545 MPa)
Gasket Stress 23100 psi 23100 psi
(160 MPa) 160 MPa)
Remaining 52630 psi 53100 psi
Bolt Stress (363 MPa) (366 MPa)
Bolt Joint 33.3% 32.8%
Relaxation
Gasket Stress 15419 psi 15557 psi
After Test (106 MPa) (107 MPa)

TABLE 1 outlines the remaining gasket stress

after two temperature cycles for both gaskets were
very similar. Even though the DJ gasket lost
significantly more internal pressure in the test. This
was not due to low gasket load it was due to the style
of gasket i.e. metal to metal contact (gasket to flange)
and the well-known high leakage rates a DMJ has.
FIGURE 8b: Schematic View of the Test Machine
FIGURE 8a & 8b show the test apparatus used
for the temperature cycling test. 4” 300lb modified
Flanges, “Modified” top flange machine with a 1.7 TEST CONCLUSION
Nubbin.
Leakage Testing: The DMJ gasket showed
significantly higher leakage values than the
Kammprofile gasket. Although the leakage reduced on
the DMJ as the gasket stress increased, this was not at
the same magnitude as the Kammprofile gaskets.

There was no difference in leakage rates on a DMJ

gasket between flat face and a nubbin setup, which
indicates the nubbin is not reducing the leakage rate
due to the higher stress point.

Temperature Cycle Testing: Even though the

remaining gasket stress was very similar on both
gaskets, the pressure loss was significantly higher on
the DMJ gasket proving the Kammprofile gasket
provides a tighter seal.

No damage to gaskets or flange faces/nubbin

occurred during testing.

All results shown are from an average of three

tests per project.

FIGURE 8a: Test Machine

4 Copyright © 2019 ASME

1.8 CASE STUDY AT SINCLAIR OIL
(WYOMING)

The Poly Unit at Sinclair Wyoming Refining

Downloaded from http://asmedigitalcollection.asme.org/PVP/proceedings-pdf/PVP2019/58936/V002T02A019/6446092/v002t02a019-pvp2019-93027.pdf by ExxonMobil Upstream Research user on 05 January 2022

Company (SWRC) is a series of four vertical vessels
used to make Poly Gasoline. Four of these vessels
were designed to originally use a Double Metal Jacket
(DMJ) gasket running in a 1/2” wide groove around
the reactor flange, see FIGURE 9.

FIGURE 9: Poly Unit Flange

There have been many problems with the DMJ

gaskets sealing, this has become a major cost issue due
to the inconsistency and down time of the reactors. A
decision was made to conduct a test case putting in
Kammprofile style gasket without machining off the
nubbins and monitoring how the gaskets perform in
comparison to DMJ gaskets.
The following vessels were affected; 8EX-1710,
8EX-1711A, 8EX-1712, 8EX-1713. FIGURE 10a &
10b show a drawing of the flange and one of the
reactor vessels.
Please note; all four of the vessels listed above are
identical in terms of total size and inner/outer diameter
of the gasket.
The gaskets are replaced every 5 weeks to change
the catalyst.

FIGURE 10b: Poly Unit Drawing

FIGURE 10a: Gasket Dimensions

5 Copyright © 2019 ASME

1.9 FIELD STUDY
March to June 2018
In March of 2018, the first reactor gasket was
changed to a Kammprofile style.
Over the next four months the remaining three
reactors were converted over to Kammprofiles. During
operation the leakage reduced dramatically when
using a Kammprofile.
Inspection occurred every time the gaskets were
replaced. There was no damage to the gasket or seating
surfaces of the flanges.
Operating Conditions
Maximum Temperature: 450 Deg F (232 Deg C)
Maximum Pressure: 1000 psig (68 bar)
5-week cycles before a change in catalyst is
performed.
Bolting Information
The same bolting procedure was used on these
vessels as had been used when a DMJ gasket was in
operation. FIGURE 11 is a screenshot of the
procedure customized for these four exchanger units.
FIGURE 11: Installation Document
6 Copyright © 2019 ASME
The methodology behind the procedure is to not
change anything other than the gasket style. Ensuring
the only change in a variable is the gasket. Given we
were changing to a Kammprofile style, which tends to
Downloaded from http://asmedigitalcollection.asme.org/PVP/proceedings-pdf/PVP2019/58936/V002T02A019/6446092/v002t02a019-pvp2019-93027.pdf by ExxonMobil Upstream Research user on 05 January 2022
seal well under a wider variety of clamping force, there
was no need to change the bolting procedure.
Method of replacement
To change the style of gaskets an MOC was
written and divided into four parts. Part one was
commissioned officially on the 25th of March 2018
with the last poly reactor changed and commissioned
on the 8th of June 2018.
As of this date (January 7th, 2019) there have been
no reports of leakage, damage, or bolting problems
with the reactors. The Poly Units have run faithfully
with Kammprofile style gaskets with no machining of
the nubbins done.
1.9 FIELD STUDY CONCLUSION
Kammprofile gaskets have run in all four of the
existing poly units for approximately nine months.
There have been no issues brought up, and the existing
base material in the groove of the flange has not been
damaged or changed in any way. Our view is that this
experiment was a success and we will continue to
make case-by-case determinations of switching over
to Kammprofile from DMJ without machining off the
nubbins.
2.0 OVERALL CONCLUSION
The test study was run until both authors were
satisfied with the performance difference of the
gaskets selected and no damage to the test fixtures
occurred. It was decided no unnecessary testing
would be done outside of the actual conditions on the
Sinclair Poly Units.
The test study provided enough evidence to use a
kammprofile gasket on the units.
All parties feel this was a comprehensive test
study and the results from the study were replicated in
the actual field application.