Beruflich Dokumente
Kultur Dokumente
HSE
The document is concerned with a test programme investigating the slip capacity of neoprene-lined
clamps. In Phase I of the project, which is reported separately, a total of sixteen full-scale tests were
conducted at Memorial University of Newfoundland, Canada. Based on the results of the Phase I
tests, interim recommendations were made for the estimation of frictional coefficients. The results
indicated some surprising effects, and further tests were recommended. The further tests have now
been conducted under Phase II of the project and are reported herein.
This report and the work it describes were funded by the Health and Safety Executive (HSE),
ExxonMobil, Shell UK Exploration and Production and MSL Engineering Limited. Its contents, including
any opinions and/or conclusions expressed, are those of the authors alone and do not necessarily
reflect HSE policy.
HSE BOOKS
or by e-mail to hmsolicensing@cabinet-office.x.gsi.gov.uk
FOREWORD
This document has been prepared by MSL for three sponsoring organisations:
ExxonMobil
The document is concerned with a test programme investigating the slip capacity of
The recommendations presented in this document are based upon the knowledge available at
the time of publication. However, no responsibility of any kind for injury, death, loss,
damage or delay, however caused, resulting from the use of the recommendations can be
accepted by MSL Engineering or others associated with its preparation.
The participants do not necessarily accept all the recommendations given in this document.
APPLICATION PHASE II
FINAL REPORT
CONTENTS
Page No
FOREWORD .............................................................................................................................3
1.
I NTRODUCTION..........................................................................................................6
1.1
General ...............................................................................................................6
1.2
2.
3.
4.
3.1
3.2
T
est Rig ............................................................................................................12
3.3
T
est Instrumentation.........................................................................................15
3.4
T
est Procedures ................................................................................................17
3.5
R
ESULTS.....................................................................................................................23
4.1
4.2
4.3
4.4
5.
6.
5.1
Discussion ........................................................................................................30
5.2
REFERENCES
1.
INTRODUCTION
1.1
General
This report is concerned with a second phase of a Joint Industry Project (JIP)
investigating the experimental slip capacity of neoprene-lined clamps.
The use of such clamps in the offshore industry has been, and continues to be,
widespread throughout the world. The following applications of such clamps may be
given:
=
Neoprene-lined clamps contain a liner that lies between the clamp steelwork and the
enclosed member. The liner provides tolerance against lack of fit of the clamp saddle
around the tubular brace. In general, the linear is made of polychloroprene (neoprene)
sheet that is bonded to the inner surface of the clamp saddle plates. The neoprene
liner is usually plain for structural connections designed to transmit axial or rotational
loads, although ribbed linings are sometimes used to accommodate potentially large
lack of fit tolerances.
Stressed neoprene-lined clamps rely on applied stud bolt pre-loads to generate
compressive forces normal to the interface between the clamp liner and the surface of
the clamped brace. The strength is considered to be dependent on the magnitude of
the normal force, the relative stiffness of the steel and liner and the effective
coefficient of friction at the liner/brace interface.
Despite the widespread use of neoprene-lined clamps through the world, there were
only limited data, and no data in the public domain, on the slip capacity of these
clamps.
As part of a Joint Industry Project conducted by MSL entitled
Demonstration Trials of Diverless Strengthening and Repair Techniques, a static
slip test on a neoprene-lined clamp exhibited a slip capacity significantly less than
that expected from the guidance available at that time.
It was against the above background that MSL launched this current Joint Industry
Project. The project was intended to generate test data so that more reliable design
guidance could be formulated, both for the rational assessment of the reliability of
neoprene-lined clamps currently in service and for the safe design of such clamps in
future applications.
With the support of HSE and two major North Sea operators, Phase I of this current
JIP was concluded in May 1999 with the issue of a final report(1) to the sponsoring
organisations. Phase I covered a programme of 16 full-scale neoprene-lined clamp
tests. The tests in Phase I encompassed both axial and torsional loading and were
designed to investigate the influence of a variety of parameters, including the bolt
6
Clamp tests with imposed interface pressures lower than those used in the
Phase I programme.
Clamp tests with neoprene liners having hardness values different to that
adopted in Phase I.
In light of the extensive use of neoprene-lined clamps, and the benefits that will result
through generation of data in the above three areas, this Phase II of the subject JIP
was instigated.
1.2
surface conditions of the clamped tubular member (Tests T1, T7 and T8)
clamp length to diameter (L/D) ratio (Tests T1, T11 and T12)
The Phase I test programme is summarised on the pullout table at the back of this
document. The primary Phase I finding is that coefficient of friction for neoprene
lined clamps is substantially below the range of values adopted in practice. Hence,
some existing structural neoprene-lined clamps potentially have capacities that may
be less than the design intent. In addition to the primary findings, the following
results were also achieved during the Phase I tests:
=
The failure load for all axial tests was defined as the position at which the
load-slip curve was seen to deviate substantially from the trendline defining its
7
initial slope. The failure torque was defined as the position with a relative
rotation of 0.45 between the clamp and tubular for pure torsional load or
when an axial displacement of 1.25 mm was reached under the combination of
tensile and torsional loads.
=
The axial tests were repeatable and similar failure loads could be derived from
tests with identical conditions according to the definition of failure.
The increase in the applied stud bolt load did not lead to a corresponding
increase in the clamp axial capacity, at least for a preload level of 40% - 60%
of the stud bold yield strength. For a given stud bolt pre-load level, significant
drop in bolt load with the increase of the slip was shown, although such
reduction was not so marked for the initial 4 to 5 mm slip.
The clamp axial slip capacity varied little with either the pipe surface
condition or the pipe radial stiffness (i.e. pipe diameter to thickness ratio)
according to the definition of failure criterion.
Application of the torsional load would reduce the clamp slip capacity.
With the above observations achieved in the Phase I test programme, interim
guidance was formulated for clamp slip capacities under axial load alone, torsional
moment alone and combined axial and torsional loadings respectively.
The developed guidance had to be considered as being of an interim nature until
further data became available due to the unexpected slip behaviour of the clamp,
particularly with regard to the relationship between applied bolt load and clamp
capacity. There were insufficient test data to permit a proper clarification of the role
of bolt load.
The interim guidance can be considered conservative, particular for small bolt pre
loads where no data exists in the Phase I test programme. For combined axial and
torsional loadings, it could also lead to a rather conservative prediction of the
torsional capacity at high values of co-existing axial load.
The combination of lower bolt pre-loads, applied loading rate effects typical of those
due to wave action, and possibly liners of greater Shore hardness may give higher
apparent coefficients of friction. It is for this purpose that Phase II of this current JIP
was launched.
The remainder of this document presents the Phase II JIP test programme in detail,
viz:
=
2.
Interface pressure - Tests T1A and T1B (along with Phase I Tests T1, T2 &
T3)
Neoprene hardness - Tests T18 and T18A (along with Test T1A)
Failure definition (full cyclic and half cyclic loading) (Test T1C)
Phase I test rig was utilised for the first two tests on T1A and T1B. In order to apply
a sinusoidal-type loading in both tensile and compressive direction (tests on T4C and
load steps 1 through 5 of tests on T1C), the Phase I test rig was slightly modified to
remove slack in the bearings.
10
Incremental
Incremental
Incremental
Incremental
Half Cyclic
T1A
T1B
T4C
T18
T18A
T1C
800
800
800
800
800
800
Clamp
Length
(mm)
324/17
324/17
324/17
324/17
324/17
324/17
Brace
D/T
(mm)
Table 2.1:
M36
M36
M36
M36
M36
M36
Bolt
Size
(nom.)
20%
20%
20%
20%
10%
20%
Bolt
Load
(% fy)
Black oxide
Black oxide
Black oxide
Black oxide
Black oxide
Black oxide
Brace
Surface
Condition
11
No.
Of
Bolts
Nature
Of Test
Loading
Test
No.
10
10
10
10
10
10
Neoprene
Thickness
(mm)
60
50
70
60
60
60
Neoprene
Hardness
(IRHD)
3.
3.1
Clamp Specimen
The clamp test specimen, as used in the Phase I trials, is illustrated in Figure 3.1. The
clamp was structurally typical of many clamps used for the retrofitting of risers to
existing installations and for the handling of pipe spools.
stud bolt centrelines (8 No. total)
600 mm
460 mm
bolt
stiffeners
neoprene
lining
saddle plates
side plates
800 mm
Figure 3.1:
3.2
Test Rig
Phase II tests were restricted to pure axial loading of the clamp along the longitudinal
axis of the pipe. The same test rig configuration used for application in axial tensile
loading in the Phase I tests, as shown in Figure 3.2 and illustrated in Figure 3.3, was
utilised in the Phase II tests. Modifications to the end connections have been made to
permit load reversal. The modified end connection and load cell tubular interface
(flange joint) are illustrated in Figure 3.4.
load
cell
hydraulic
actuator
Figure 3.2:
load
cell
hydraulic
actuator
13
tubular section
clamp test
specimen
tubular section
Figure 3.3:
14
Figure 3.4:
3.3
Test Instrumentation
Strain Measurement
=
A total of six strain gauges were mounted on the pipe to verify loading of the
specimen. The locations and identification numbers for each of strain gauges
are shown in Figure 3.5. The identification numbers of each of the stud bolts
are also shown in Figure 3.5.
Each bolt was instrumented with two strain gauges in a half bridge
configuration to monitor the total load on each bolt.
Displacement Measurement
=
A total of five LVDTs were positioned about the specimen to record the
relative displacement of the clamp with respect to the pipe. Figure 3.6
presents the locations of the LVDTs.
Sufficient travel lengths were specified for the LVDTs such that the entire
loading regime could be recorded.
15
Figure 3.5:
Figure 3.6:
Locations of LVDTs
16
Data Acquisition
3.4
Test Procedures
3.4.1 P
re-Testing Procedures
Calibration of Stud Bolts
Two strain gauges were mounted, diametrically opposite each other, at the mid length
of each stud bolt. Each bolt was calibrated by installing it into a tensile testing
machine and applying incremented tensile loading up to 80% of nominal tensile yield.
Preparation of Tubular Members
The existing uncoated specimen from the Phase I tests was utilised in the Phase II
tests. The tubular was lightly manually wire brushed immediately prior to the clamp
installation for each test.
Application of Neoprene Liner
The supply and application of the neoprene-liner (IRHD 50, IRHD 60, or IRHD 70)
for the clamp was in accordance with MSL document entitled Specification for
Clamp Lining and Bonding and consistent with procedures employed for the Phase I
tests. Every effort was made to ensure the highest quality bond of the liner to the
clamp.
Assembly of the Clamp Specimen
During steps 1 to 4, below, the neoprene liners of each of the clamp halves and the
tubular member were liberally doused with water from a hose, ensuring that all
contact surfaces remain fully soaked throughout the installation.
1.
The lower clamp half was supported at a height of approximately six inches
(152 mm) from the floor of the laboratory.
2.
The tubular member was placed into the lower clamp half.
3.
The upper clamp half was lowered onto the tubular member to align directly
above the lower clamp half.
4.
Each of the stud bolts was carefully inserted through the holes in the flange
plates of the clamp halves, ensuring that the attached strain gauges were not
damaged. Spherical washers and nuts were applied and hand tightened,
17
ensuring that the split line on each side of the clamp remained even along the
length of the clamp and approximately equal either side of the clamp. A
sufficient length of bolt protruded above the top plate of the upper clamp half
to accommodate the hydraulic stud bolt tensioning system.
5.
The leads from the strain gauges on each of the bolts and on the tubular were
connected to the appropriate channels of the data acquisition system.
6.
The operation of all instrumentation and data acquisition system was checked.
7.
8.
(b)
maximum variation of load between bolts did not exceed 10% of the target
load.
The specimen was installed into the appropriate test rig and the LVDT
instrumentation was set up.
The operation of all instrumentation and data acquisition system was checked.
The specimen was bedded down by applying load cycles not greater than 5%
of the estimated failure load.
3.4.2 T
esting Procedure
After completion of the steps described in Section 3.4.1, the application of loading
proceeded in accordance with the following procedure:
=
18
With the exception of Tests T1C and T4C, each specimen was subjected to
two loading and unloading cycles. Test T1C comprised of sinusoidal loading
cycles (at 5 different loads) and half cycle loading cycles (again at 5 different
loads). Test T4C contains 7 full cycle sinusoidal loading cycles. The actual
loading schedules used are summarised in Section 3.5.
3.5
All stud bolt nuts were completely slackened off using the Hydratight
tensioning system and readings of all instrumentation were again taken.
The electrical connection leads from the stud bolts to data-logger were
disconnected and all stud bolts were carefully removed, supporting clamp
halves appropriately.
The two clamp halves were split and the interfaces surveyed, taking
photographs and notes as appropriate.
Loading Schedule
Digital files were generated by the data acquisition system, then sampled and
converted to the control signal by the Digital/Analogue channels of the data
acquisition system. As mentioned above, for each specimen configuration except
Tests T1C and T4C, the test contained two loading cycles. Test T4C was comprised
of seven load steps and Test T1C contained ten. Table 3.1 presents an overview of
the loading schedule.
Test
Load Type
Termination Condition
T1A
Incremental Tensile
20 mm slip
T1B
Incremental Tensile
20 mm slip
T4C
Sinusoidal Tension-Compression
T18
Incremental Tensile
20 mm slip
T18A
Incremental Tensile
20 mm slip
T1C
Sinusoidal Tension-Compression
& Tensile-half of sine wave
Table 3.1:
Test Overview
19
Details of loadings are given in the Annex in the form of load-time plots. Here, the
loadings are summarised as follows:
Tests T1A, T1B, T18 & T18A
These four tests were comprised of two loading cycles each.
Loading cycle one was incremental loading of 5-minute durations commencing at 50
kN, proceeding to 75 kN, then 100 kN. After this, the load increments were reduced
to intervals of 10 kN, continuing until a relative displacement of 4 mm of the tubular
in the clamp was measured.
Loading cycle two involved a linear increase of the load, at a rate of 700 lbs
(approximately 31.2 kN) per minute, from zero load to a point where 20 mm slip was
measured between the clamp and the tubular, after which the load was gradually
reduced back to zero.
Test T4C
A total of 7 loading steps were comprised in this test. All loading steps were similar,
sinusoidal loading in tensile and compressive directions about a mean load of zero.
The only variations between load steps were frequency and amplitude of the cyclic
load and the total number of times each load was applied (cycle numbers). Figure 3.7
presents a typical loading cycle in Test T4C. Specifications of each of the 10 loading
cycles are summarised in Table 3.2. The amplitudes, periods and the steepness of
1/16 are explained below.
2
40000
1 .5
30000
1
20000
0 .5
1651
1585
1519
1453
1387
1321
1255
1189
1123
991
1057
925
859
793
727
661
595
529
463
397
331
265
199
133
-0 .5
Displacement (mm)
0
0
67
Load (lbs)
10000
LO AD LBS
AVERAG E BO LT LO AD LBS
LVDT 1 M M
-1 0 0 0 0
-1
-2 0 0 0 0
-1 .5
-3 0 0 0 0
-2
-4 0 0 0 0
-2 .5
T im e (s )
Figure 3.7:
Test T1C
Test T1C included a total of 10 load steps, the first five being sinusoidal and the latter
being the tensile load, i.e. positive portion only, of similar cycles as shown in Figure
20
55000
45000
1 .5
35000
25000
0 .5
15000
LOAD LBS
A V E RA G E BO LT LO A D LB S
LVDT 1 M M
-5 0 0 0
1376
1321
1266
1211
1156
1101
991
1046
936
881
826
771
716
661
606
551
496
441
386
331
276
221
166
111
-0 .5
56
5000
Displacement (mm)
Load (lbs)
3.8. The variations in the frequency, amplitude and number of cycles are presented in
Table 3.2.
-1
T im e (s )
Figure 3.8:
Loading
Cycle
No. of
Cycles
Amplitude
(kN)
Period
(seconds)
Comment
10
24
7.0
Steepness of 1/16
10
51
9.0
Steepness of 1/16
99
10.5
150
13.2
Steepness of 1/16
188
12.0
220
12.0
Steepness of 1/16
220
10.5
Steepness of 1/16
No. of
Cycles
Amplitude
(kN)
Period
(seconds)
Comment
1&6
10
24
7.0
Steepness of 1/16
2&7
10
51
9.0
Steepness of 1/16
3&8
99
10.5
4&9
150
13.2
Steepness of 1/16
5 & 10
188
12.0
Table 3.2:
21
The amplitudes, periods and steepness values in Table 3.2 were proposed, in
consultation with the Project Steering Committee, to be similar to what existing
clamps may experience. For this purpose, it was assumed that the test specimen had
been used to attach a 26 retrofit riser to a typical UK Southern North Sea platform.
The clamp was assumed to be located at an elevation close to the first horizontal
frame below the waterline (-8.0 m). The water depth was taken as 35.4 m. It was
assumed that the clamp had been designed with an interface pressure of 3.2 MPa (i.e.
consistent with Test T1A). It was further assumed that, in the design of the clamp, a
value of 0.2 was used for the coefficient of friction and a factor of safety of 1.7 was
applied to the extreme event load. On this basis, the clamp notional design axial slip
capacity is 188 kN.
The design wave height and period used for the 100-year event were 15.1 m and 12
seconds respectively. At 8 m water depth the lateral wave load on the riser was 25.9
kN/m. The clamp was therefore designed to resist the wave load on approximately
7.3 m of riser. The wave height and period used for the 1-year event were 10.8 m and
10.5 seconds respectively. At 8 m water depth, the load on the riser was 13.6 kN/m,
therefore, the load on the clamp during the 1-year return period wave was 99 kN. For
the determination of the wave periods associated with the intermediate load steps a
wave steepness of 1/16 has been assumed.
22
4.
RESULTS
This section presents the main results of the various tests conducted in Phase II. The
results are grouped, according to the parameter under investigation, in the following
subsections. Further details may be found in the Annex, especially of the condition of
the liner following each test.
4.1
4.2
tests that a pre-load of 60% of tensile yield caused excessive bulging of the liner
during bolt preload application and damage of the neoprene liner during slip. Hence,
50% of tensile yield was taken as the design limit of the bolt pre-load level.
Bolt Pre-load
The bolts were simultaneously tensioned using Hydratight hydraulic tensioning tools.
The loads in the bolts were continuously monitored during the tensioning operation by
means of the attached strain gauges, each bolt having been previously calibrated to
80% of yield. The average applied pre-loads at the start of Tests T1A and T1B are
presented in the table below, where the pre-loads at the start of Tests T1 and T2 in
Phase I are also included for reference.
Test ID
Total
% of tensile yield
T1A
106
848
20%
T1B
55
440
10%
T1
212
1696
40%
T2
274
2192
50%
24
Slip
The load-slip behaviour recorded in Tests T1A and T1B are shown in Figures 4.3 and
4.4 respectively. The load-slip curves for Tests T1 and T2 in Phase I are reproduced
in Figure 4.1. The figures show the mean slip of the two clamp halves for each test.
A ductile form of slip can be seen in each of the test, as already mentioned. The
linear trend line through the initial slope of the curves has been plotted on Figures 4.3
and 4.4. The failure load for each test, based on the definition discussed in Section
4.1, is given in the table below, together with the corresponding apparent friction
coefficient. Those obtained in Phase I for Tests T1 and T2 are also included for
reference. The apparent friction coefficient is defined as the failure load divided by
the total applied bolt pre-load per clamp half. As reported in Phase I final report, the
design capacities for Tests T1 and T2 were estimated as 441 kN and 552 kN
respectively ( = 0.2, factor of safety = 1.7). The corresponding estimated design
capacities for Tests T1A and T1B are 220 kN and 110 kN respectively.
Test ID
Failure Load
150
0.088
115
0.131
150
0.044
150
0.034
During the failure load determination for Tests T1A and T1B, the load cycle 2 curves
in Figures 4.3 and 4.4, which more actually represent the practical load application
rate, were utilised. The load-slip behaviour of each of the Tests T1A, T1B, T1 and
T2 are shown, for comparison, in Figure 4.7. Given the feature that all these fours
tests behaved very similar shown in Figure 4.7, the failure load of Test T1B presented
in the above table may be considered to be conservative.
It can be observed that the increase in the applied stud bolt load does not necessarily
lead to a corresponding increase in the clamp axial capacity. For a bolt load of 20%
of tensile yield and above, the nature and magnitude of clamp failure remain similar.
This, in turn, results in a reduction in the apparent friction coefficient for each of the
Tests T1A, T1 and T2.
4.3
25
Bolt Pre-Load
The bolts were simultaneously tensioned, using Hydratight hydraulic tensioning tools.
The load in the bolts was continuously monitored during the tensioning operation by
means of the attached strain gauges, each bolt having been previously calibrated to
80% of yield. The applied bolt pre-loads at the start of Tests T1A, T18 and T18A are
given in the table below.
Test
ID
Neoprene
Hardness
Average
Total
% of tensile yield
T1A
60
106
848
20%
T18
70
100
800
20%
T18A
50
105
840
20%
Neoprene
Hardness
Failure Load
(kN)
Apparent Friction
Coefficient
T1A
60
150
0.088
T18
70
160
0.100
T18A
50
130
0.077
The correlation between neoprene hardness and the apparent friction coefficient is
presented in Figure 4.9. It can be seen that the apparent friction coefficient increases
with neoprene hardness.
4.4
Test T4C with a neoprene hardness of 60 IRHD. The cyclic loading steps in Tests
T4C followed a sinusoidal tension-compression pattern with different amplitudes and
frequencies. For Test T1C ten loading steps were carried. A new neoprene liner with
hardness of 60 IRHD was fitted for Test T1C. The first five loading steps for Test
T1C were identical to those in Test T4C and represent the repeatability of the
application of cyclic loading performed in Test T4C on the neoprene-lined clamp.
The second five loading steps represent the application of half-cycle pulse loading on
the neoprene-lined clamp with the same assumptions as the first five steps. Half-cycle
wave loading shows the response of the clamp when subjected to a push-push action
instead of the push-pull action of full-cycle loading. The comparison of the results of
the two tests determined whether the clamp slip seen in the previous tests was a
product of shear deformation of the neoprene lining or actual slippage due to the
applied force. The load pattern is presented in Table 3.2 in Section 3.5.
Bolt Pre-Load
The bolts were simultaneously tensioned, using Hydratight hydraulic tensioning tools.
The load in the bolts was continuously monitored during the tensioning operation by
means of the attached strain gauges, each bolt having been previously calibrated to
80% of yield. The average pre-loads at the start of each step for Tests T4C and T1C
are presented in the following table.
Test ID
T4C Step 1
T4C Step 2
T4C Step 3
T4C Step 4
T4C Step 5
T4C Step 6
T4C Step 7
T1C Step 1
T1C Step 2
T1C Step 3
T1C Step 4
T1C Step 5
T1C Step 6
T1C Step 7
T1C Step 8
T1C Step 9
T1C Step 10
Neoprene
Hardness
60
60
60
60
60
60
60
70
70
70
70
70
70
70
70
70
70
cycle and a load loss during the compressive half-cycle were observed. The rates of
load increase and loss were similar and approximately equalled to those of bolts 1 and
8. The central bolts, numbers 2, 3, 6 and 7 see little or no loss/increase in load during
the entire tensile-compressive cycle. The above observations are similar to the
findings of the quasi-static tests as already described in Section 4.2. The variation of
the bolt pre-load for Test T1C followed a similar pattern to that for Test T4C.
Figure 4.11 presents the average bolt pre-load variations for each of the loading steps
in Tests T4C. Comparisons of the average bolt pre-load variations among the loading
steps in Test T4C, full-cycle steps in Test T1C and half-cycle steps in Test T1C are
shown in Figure 4.12. As can be seen, there is little difference in response for full- or
half-cycle loading.
Slip
As seen in the quasi-static axial loading tests, a ductile form of slip appeared in each
of the load steps of Tests T4C and T1C. The load-slip response for step 5 of Test
T4C is shown in Figure 4.13. It can be seen from Figure 4.13 that the load-slip
response forms a closed hysteresis loop. Similar load-slip behaviour was observed for
all load steps in Tests T4C.
Figure 4.14 presents a single cycle of each step in Test T4C. It can be observed that
the slope and area of the load-slip hysteresis loop depend on the frequency and
amplitude of the applied cyclic load. Generally, higher load amplitude results in a
larger loop area.
The first five loading steps of Test T1C were carried out to assess the repeatability of
the clamp slip behaviour under cyclic load. As addressed in the test procedures, the
neoprene liner was to be replaced whenever visual damage became evident or, in any
case, following three successive tests. Test T1C was conducted using a new neoprene
liner with the same hardness as that in Test T4C. The comparative load-slip
behaviour of Tests T4C and T1C is shown in Figure 4.15. Figure 4.15 reveals that the
clamp slip response in Tests T4C and T1C are similar.
As mentioned above, Test T4C had the identical specimen configurations and similar
bolt pre-load levels to Test T1A. The peak load-slip responses of each loading step in
Test T4C are compared with that of Test T1A in Figure 4.16. It can be seen from
Figure 4.16 that the clamp has similar slip behaviour under cyclic loading and quasi
static loading.
Full-cycle Tests T4C and T1C showed that the load-slip response formed a stable
hysteresis loop. The amplitudes of Steps 1 to 4 in these tests did not exceed their
static axial load capacity. It can be seen from Figure 4.15 that a residual displacement
of clamp relative to pipe is within 0.4 mm for Steps 1 to 4 when the applied load
reduced to zero. It was observed during tests that even this residual displacement was
recovered within a very short period of time.
In order to determine whether the observed residual displacements of the clamp are
purely from the shear deformation of the neoprene lining or if they represent true slip
movement (ie. sliding at interface) of the clamp, 5 half-cycle loading steps were
conducted in Test T1C. The five pulse loading steps simulate the tensile half of the
28
five full-cycle steps respectively. Figure 4.17 presents one loading cycle of Step 5 in
Test T1C.
The clamp load-slip response under the tensile half-cycle loading steps was again
observed to form a stable hysteresis loop. A typical example is shown in Figure 4.18
for loading Step 10 of Test 1C.
Comparisons of the clamp load-slip response among the full-cycles tests in T4C and
T1C and the half-cycle tests in T1C are shown in Figure 4.19. The neoprene liner
under the half-cycle load is slightly stiffer than when experiencing a full-cycle load.
In part, this may be because the loading rate in the half-cycle is slightly higher.
It can be seen from Figures 4.18 and 4.19 that the residual displacement of the clamp
under the half-cycle loading is so small that it can be neglected. The residual
displacement observed in the full-cycle tests can be taken as the pure shear of the
neoprene.
Given the above observations, the following conclusions can be drawn:
1)
A clamp under cyclic load has a similar deformation response to that when
subjected to static load.
2)
3)
With the application of axial load less than the failure load, the shear
deformation of the neoprene results in a small amount of residual
displacement, which is recoverable within a short period of time.
29
5.
This section is concerned with the development of guidance for the slip capacity of
stressed neoprene-lined clamps based on the results of the Phase I tests and of the test
programme described in Section 4.
5.1
Discussion
It is appropriate to begin with the tests that were not subject to cyclic loading. A most
illuminating, albeit short, discussion of friction behaviour, as pertains to natural
rubbers, can be found in Reference (2). It is not known how applicable it is to
synthetic rubbers such as neoprene but it would seem to explain the results of most of
the tensile tests in both Phase I and Phase II. Quoting from Reference (2) (underline
inserted):
The coefficient of friction is defined by = F / W , where F is the
tangential friction force and W the applied normal load. For rubber,
the coefficient of dry friction is not constant, but falls with increasing
normal load. At light loads the dependence is weak, but it becomes
more pronounced at high loads. The friction force F is proportional to
the real surface contact area, which for normally rough surfaces under
light loads is much less than the geometric area of contact. At very
high loads the relatively low modulus of rubber results in the real
contact area approaching the geometric area, and F tends to a limiting,
maximum value. For dry contacts, the constant of proportionality
between F and the real contact area is of the same order as the shear
modulus, but it is reduced by surface contamination.
The above passage suggests that the bolts loads in the majority of tests, although
typical of offshore practice, were sufficiently high so that the limiting value of slip
load F was reached. The evidence of liner extrusion due to preloading the bolts tends
to confirm that liner was highly stressed, and the real contact area was approaching
the geometric area. Figure 4.7 illustrates the similarity of clamp slip loads over a
wide range of bolt loads.
However, rather than taking the ultimate slip load (ie. the load occurring at a slip of
15 mm or more), a more conservative failure criterion has been used herein. The
failure load has been taken as the load when significant departure from the initial
linear elastic behaviour occurs. This corresponds to when relative displacement
occurs under sustained loads (see Test T4A curve in Figure 4.2). With this failure
criterion, most specimens again give a similar failure load, and hence approximately
similar factors of safety against true slip. In only one test (Test T1B) were bolt loads
so low that a lower failure load was inferred.
To assist in the development of design guidance, reference is made to Figure 5.1,
which is a plot of the interface shear capacity against the interface pressure. The
figure shows the results of Tests T1A, T1B, T1, T2 and T3 in which the preload was
the parameter under investigation. Superimposed are lines corresponding to the
apparent coefficients of friction inferred from the tests. Also shown is the line
corresponding to =0.8 which is typical of values suggested by liner manufacturers.
30
The results from Tests T18 and T18A for different neoprene hardness are also
presented.
0.35
=0.8
18
=0.100
0.3
T1A, =0.088
1
=0.131
0.25
T1, =0.044
T2, =0.034
18
=0.077
T3, =0.024
0.2
0.15
0.1
0.05
0
0
10
11
Figure 5.1:
The design guidance given in Section 5.2 is formulated in terms of a limiting interface
shear capacity (this is 0.29 N/mm2 for the results of Tests T1A, T1 and T2 all having
neoprene hardness IRHD = 60). A limit of 8.5 N/mm2 is put on the interface pressure
as the liner of specimen T3 in Phase I was extruded when the stud bolt preload was
applied and it also suffered damage during the slip test. For interface pressures less
than that corresponding to about that in Test T1A (q = 3 N/mm2 in fact), the limiting
interface shear capacity is ramped down in a parabola form, effectively ending with a
slope of = 0.19 at the origin. Although it is conservative compared with a slope of
0.8, uncertainties exist in the small interface pressure region, and there is no test data
available for an interface pressure less than that of Test T1B (1.7 N/mm2). The
parabola is given by:
2q q 2
0.29
9
3
5.1
The limiting interface shear capacity was observed to be a function of the neoprene
hardness, measured in International Rubber Hardness Degrees (IRHD), see Tests
T18A, T1A and T18 (IRHD values of 50,60 and 70 respectively) in Figure 5.1.
Assuming that these test results are indicative of their respective plateau regions, the
limiting interface shear capacity can be plotted against IRHD as shown in Figure 5.2.
31
0.34
0.32
T18
0.3
T1A
0.28
0.26
T18A
0.24
0.22
0.2
40
50
60
70
80
IRHD
Figure 5.2:
The curve shown in the figure is a suitably simple yet accurate approximation to the
data; its equation is given by:
=
sin (IRHD)
3
5.2
Two other sets of results, not shown on Figure 5.1, are worth mentioning. Firstly, the
results of the tests conducted at Karlsruhe fall in the region of an interface pressure of
2.5 N/mm2 at a slightly higher shear capacity than Tests T1 and T2. Secondly, some
ad hoc tests on flat plate specimens confirmed the =0.8 line for low interface
pressures. Both sets of results indicate that the design guidance is conservative.
The design/assessment provisions include a factor of safety, . For long term loads
(eg. Gravity loads) applied to the clamp, should not be taken as less than unity as
such values could lead to creep of the liner material (recall Test T4A in Figure 4.2).
Indeed, it should not be forgotten that the design/assessment provisions are essentially
based on mean values of a relatively small data sample and therefore are not even
characteristic or lower bound. However, as discussed below, environmental loads are
short term in nature, and this allows a less onerous interpretation to be assigned to .
32
Cyclic loading tests, especially those of half-cycle tests in T1C, reveal that the relative
displacement of the clamp is almost entirely due to neoprene shear deformation. The
response of the clamp took the form of closed hysteresis loops, with only negligible
deformation following load removal, see Figure 4.19(e).
Rubber-like material is highly sensitive to creep, during which the material continues
to deform under a given load. Figure 5.3, reproduced from Reference (2), shows that
for a certain types of rubber, creep varies approximately linearly with the logarithm of
time under load. It would appear that the durations of the cyclic tests were such that
no substantial creep occurred, and that this is the essential difference between the
cyclic and quasi-static tests. In the design/assessment provisions, it is therefore
recommended that the factor of safety may be taken as unity for designing clamps
subject to environmental loads. For assessment purposes of existing clamps, a factor
lower than unity may be justified for the storm event. This is because the storm event
occurs infrequently and very minor slippage (certainly less than 0.1 mm) does not
have any significant structural consequence. The data presented in Figure 4.19(e)
would suggest that a clamp load of 200 kN should be perfectly acceptable which,
when compared with the quasi-static limit of 150 kN, leads to an allowable of
150/200 = 0.75.
Figure 5.3:
The provisions in Section 5.2 for torsional and combined axial/ torsional loads follow
the Phase 1 findings and recommendations.
5.2
Design/Assessment Guidelines
Base on the above observations, design guidance can be formulated as follows. The
clamp slip capacity under axial load alone has been updated with the test results in
Phase II. The clamp slip capacities under either torsional moment or combined axial
and torsional loading are reproduced from Phase I.
33
(i)
DL
Pc =
DL 2q q 2
In the above:
is a limiting stress to be taken as =
sin (IRHD)
(degree angular measure).
3
D and L are respectively the tubular diameter and length of the clamp, both to
be expressed in units of millimetres to give Pc in unit of Newtons.
q is the radial pressure at the neoprene liner/tubular interface, to be calculated
as:
9 =
FB
, where FB is the total stud bolt load.
DL
For long term (gravity) loads, should not be taken as less than unity.
ii.
iii.
The total axial capacity of a clamp is thus 2Pc, but note that in many situations
the axial load is transferred to only one clamp half in the first instance.
The above formulation assumes that there is no interference, i.e. that the
tubular outside diameter is not greater than the inside diameter of the neoprene
liner. If there is interference, a lower capacity may result.
(ii)
Pc D
(iii)
1.0
35
6.
CONCLUDING REMARKS
A programme of slip tests have been carried out on neoprene-lined clamps as used in
offshore applications. The programme of Phase II consisted of six axial slip tests
under either quasi-static loadings or cyclic loadings that simulate wave action in the
UK Southern North Sea. The following parameters were investigated in the Phase II
test programme:
=
Bolt pre-load
Neoprene hardness
The Phase II tests, with lower bolt loads than those in Phase I, have allowed the
conservatism of the Phase I design guidelines to be removed. This is important as
many existing clamps have neoprene/steel interface pressures corresponding to lower
bolt loads.
The tests with clamps having different neoprene hardness (IRHD value) have
confirmed that hardness affects capacity.
The cyclic loading tests indicate that at the design capacity, the relative displacement
of the clamp and member is recoverable. In other words the displacement is largely
due to neoprene shear deformation as opposed to true slip. It is only when the loads
are applied statically that time dependent phenomena such as creep are manifested.
Design guidance has been formulated based on the results of both Phase I and Phase
II test programmes, see Section 5.2. The provisions encapsulate the above
observations. It is recommended that the factor of safety be adjusted depending on
whether quasi-static or dynamic loading is being considered. A further relaxation
may be used if the clamp is existing, as opposed to a new design.
36
REFERENCES
1.
2.
Figure 4.1:
Load-Slip Response for Tests T1 and T4 (Reproduced from Phase 1 Final Report)
Figure 4.2:
Load-Slip Response for Tests T1 and T4A (Reproduced from Phase 1 Final Report)
50
100
150
200
250
300
350
400
450
Figure 4.3:
10
15
20
25
50
100
150
200
250
300
350
400
Figure 4.4:
10
20
15
25
30
-5
40
50
60
70
80
90
100
110
120
130
Figure 4.5:
6
4
10
15
20
25
-5
20
40
60
80
100
120
Figure 4.6:
15
20
25
10
30
-5
-100
100
200
300
400
500
Figure 4.7:
10
15
20
25
100
200
300
400
500
600
(IRHD 70)
10
Figure 4.8:
Test T18
20
25
15
Test T1A
(IRHD60)
Test T18A
(IRHD 50)
30
35
0.02
0.04
0.06
0.08
0.1
0.12
40
Figure 4.9:
45
55
60
65
70
50
75
90
95
100
105
110
115
120
125
130
-3
-2
-1
-4
-2
-1
110
111
112
113
114
115
116
117
Step 3
Step 1
Step 2
118
Step 4
Step 7
Step 6
Step 5
Figure 4.11: Average Bolt Pre-load Variation of One Cycle for Each Step in Test T4C
-3
-0.3
-0.2
-0.1
0.1
0.2
110
111
112
113
114
115
116
117
118
0.3
-0.6
-0.4
112
0
0.2
0.4
-0.2
113
114
115
116
117
118
0.6
-1.5
-1
-0.5
0.5
112
113
114
115
116
117
118
1.5
-2
-1
-1.5
112
0
0.5
1.5
-0.5
112.5
113
113.5
114
114.5
115
115.5
116
116.5
-3
-2
111.5
0
-1
112
112.5
113
113.5
114
114.5
115
115.5
-250
-200
-150
-100
-50
50
100
150
200
250
-3
-1
Figure 4.13: Load-Slip Response of Step 5 for Test T4C (188 kN Amplitude Cycle)
-2
-4
-2
-300
-200
-100
0
0
Step 1
Step 2
Step 4
Step 5
-1
100
200
300
Step 3
Step 6
Figure 4.14: Load-Slip Response of One Cycle for Various Test T4C Load Steps
-3
Step 7
-2.5
-2
-1.5
-1
-200
-100
0
0
0.5
1.5
-0.5
100
200
T4C
T1C
2.5
-2.5
-2
-1
-0.5
-200
-100
100
200
0.5
1.5
-1.5
T4C
T1C
2.5
-2.5
-2
-1.5
-1
-200
-100
0
0
0.5
1.5
-0.5
100
200
T4C
T1C
2.5
-2.5
-2
-1.5
-1
-200
-100
0
0
0.5
1.5
-0.5
100
200
T4C
T1C
2.5
-2.5
-2
-1.5
-1
-200
-100
0
0
0.5
1.5
-0.5
100
200
T4C
T1C
2.5
-0.5
-50
50
100
150
200
250
300
0.5
2.5
1.5
3.5
4.5
-250
-200
-150
-100
-50
50
100
150
200
250
Time (Second)
10
Figure 4.17: Applied Axial Load of Step 5 in Test T1C (One Cycle)
11
12
Step 10 of Test 1C
Step 5 of Test 1C
13
14
-0.5
-100
100
200
1.5
0.5
2.5
-0.5
-50
-25
0.25
T1C-Step 6
T1C-Step 1
T4C-Step 1
Load-Slip Response of Step 1 in Tests T4C and T1C and Step 6 in Test T1C (One Cycle)
-0.25
25
50
0.5
-1
-100
-50
0
0.5
T1C-Step 7
T1C-Step 2
T4C-Step 2
Load-Slip Response of Step 2 in Tests T4C and T1C and Step 7 in Test T1C (One Cycle)
-0.5
50
100
-1
-150
-75
0.5
T1C-Step 8
T1C-Step 3
T4C-Step 3
Load-Slip Response of Step 3 in Tests T4C and T1C and Step 8 in Test T1C (One Cycle)
-0.5
75
150
-2
-1.5
-200
-100
0.5
-0.5
1.5
T1C-Step 9
T1C-Step 4
T4C-Step 4
Load-Slip Response of Step 4 in Tests T4C and T1C and Step 9 in Test T1C (One Cycle)
-1
100
200
-2
-2.5
-1
-200
-100
0
0
0.5
-0.5
1.5
2.5
T1C-Step 10
T1C-Step 5
T4C-Step 5
Load-Slip Response of Step 5 in Tests T4C and T1C and Step 10 in Test T1C (One Cycle)
-1.5
100
200
ISBN 0-7176-2577-X
RR 031
25.00
9 780717
625772