Beruflich Dokumente
Kultur Dokumente
Introduction
The sizing of tubular joints for ultimate strength is an element model would have to include nonlinear material
important step in the design of offshore platforms. This behavior, a fracture criterion, and, possibly, nonlinear
process is performed once platform members have been geometric terms to account for local buckling. Thus,
designed to withstand operational and severe environ- for the short term, the' empirical approach seems most
mental loading conditions. Even though nominal mem- practical.
ber stresses may be at reasonable levels, the complex This paper presents a new set of joint-strength equa-
behavior of tubular intersections can result in high- tions that are helpful for platform design. The expres-
stress amplification that can lead to failure. In many sions are based on more data and cover more joint config-
cases, the walls of joint cans must be thickened or urations than present procedures. The formulas are
member diameters must be increased to provide adequate presented along with some comparisons of the new pre-
strength. dictions with those of current methods. An example
The procedure for designing tubular joints must be problem is also given to illustrate how the equations are
simple and should lend itself to automation to handle the used in joint design.
large number of joints in most platforms. Considerable
computer output must be screened to determine max- Tubular-Joint Failure Behavior
imum applied loads at each joint. Predicted joint strength Part of the problem in analytically predicting tubular-
must be calculated for each member intersection and joint strength is that many competing failure modes are
compared with the applied loads. Joints with deficient possible in even the simplest of joints. Thus, it is help-
capacity are then sized again. ful for designers to have some insight into failure
At present, empirical expressions represent the state of mechanisms to produce both safe and efficient designs.
the art for predicting the ultimate strength of tubular In the following sections we review a number of potential
joints. They are based on axially loaded laboratory joint failure modes for simple T-, X-, and K-joints to point out
tests. Very little ultimate-strength data exist for bending trends in joint behavior and important design considera-
or combined axial and bending loads. Consequently, tions. For purposes of this discussion, the chord and
bending and combined loading effects are usually ac- branch members are as defined in Fig. 1 (Type 1). The
counted for in an heuristic manner. Analytical methods chord generally has a large diameter and larger wall
have not been successful because of the geometric, com- thickness than the branches. The branch members are
putational, and analytical complexities involved. The coped and welded to the chord wall to form the joint.
finite-element method probably could be applied but
would be too costly, time-consuming, and complicated T- and X-Joint Failure Modes
for practical joint design. To be acceptable, a finite- Fig. I summarizes a number of possible failure types
Tubular-joint ultimate-strength equations that are helpful for offshore platform design are
presented. The formulas are based on datafrom 346 joint tests. They are easy to use and cover
more joint configurations than do present design formulas.
I
(BRANCHl{]
CJ.
. / ' FAILURE AT BRANCH ADJACENT
,.--- TO WELD TOE
o !
/FAILURE IN WELD
b, --------'"'
TYPE 1
•· d
TYPE 2
TYPE 3
1
r--- A
EXAGGERATED DEFORMED
I SHAPE
(
\
A-A PLANE
L.::..- A
LAMELLAR TEARING
OF CHORD WALL
U
.-Wi
I
LOCAL COLLAPSE
DF e"ORD WALL
\ e"DRD WALlOUe'LE
TYPE 4 TYPE 5
Fig. 1-Some possible failure modes for simple, axially loaded, tubular T- and X-joints.
same T- or X-joint (see Figs. 5 and 6). Consequently, branch loading. Joint strength strongly depends on this
the compressive loading generally governs in design. factor, so that increasing the diameter of a branch
member or flaring a branch diameter near a joint is often
K-Joint Failure Modes an effective strengthening approach.
A fairly large number of tubular K-joints has been tested The intersection angle, 8, can be viewed as affecting
to failure with one branch loaded in tension and the other joint strength in two ways. First, a sin8 term is used to
in compression. The typical failure* will resemble that resolve the axial load into a component that is perpen-
shown in Fig. 2A (previously reported by Reber2 ). The dicular to the axis of the chord member. The second
compression branch produced large plastic deformation effect is a measure of how the axial joint load tends to
of the chord wall. A fracture was initiated in the chord increase in-plane stresses in the chord. This in-plane
wall at the point of maximum tensile stress at the weld toe effect is oflesser importance and was only included in the
of the tension branch. The chord wall was fabricated from K-joint equations of this paper.
ASTM A 537 Grade B steel. Fig. 2B shows a companion The parameter glD applies only to the K-joint case. It
specimen in which the chord wall was fabricated from tends to be a measure of some additional strengthening
ASTM A 36 steel. Even though the maximum branch that occurs as the branches of a K-joint are moved closer
loads for this member were only slightly less, the member together. Above glD values of about 0.2 the failure load
failed in a brittle manner with only slight plastic deforma- appears to be independent of g/D. However, as the
tion around the compressive branch. This again under- branches come closer and eventually overlap, a marked
lines the need for fabricating joint cans from material strengthening is noticed (see Fig. 3).
with good fracture toughness. Ultimate Strength Correlations
Geometric Parameters and Tubular-Joint Experimental data obtained from 214 K-joint (Table 1)
Failure Behavior and 132 T-, Y·, and X-joint load tests (Table 2) are used
as the basis for this correlation. Ultimate strength is
Empirical equations predicting tubular-joint strength
assumed to depend on the yield strength,** (jy, of the
generally correlate the ultimate load with a number of
chord wall and the geometric parameters discussed in
dimensionless ratios. It is important to know how ca-
the previous section. The form of the expression is
pacity depends on these parameters so that joints can be
resized efficiently during design. The geometric parame- P u = (j yT 2 fl(dID) f2(TID) f3 (8) figlD) (1)
ters considered in this paper are the following:
Each dimensionless function, Iv 12' 13, and 14, is taken
TID = ratio of the chord-wall thickness to its outer to have an independent influence on joint strength and,
diameter. thus, can be evaluated separately while holding the others
diD = ratio of the outer diameter of the branch to the constant. This approach extends the work of Washio
outer diameter of the chord. et aI. 3 and Gibstein 4 by considering new configurations
8 = like intersecting angle between the axis of a and incorporating additional data.
given branch and the axis of the chord.
glD = ratio of the gap (clear space distance between K-Joint Formula
the intersections of two branches with the Washio et al. 3 found that for K-joints the ultimate
chord wall) and the outer diameter of the strength, P u' can be written as
chord. This ratio applies only to K-joints.
The parameter TID gives a measure of the radial stiff-
P u sin8 = (j yT2 YRIT (1 +6.52dID)
ness of the chord wall. This parameter has a reasonably (1-0.26 cos 2 8)ligID) , (2)
strong effect on joint strength over s~me ranges and, where R=D/2 is the outside chord radius. To determine
thus, thickening of the chord wall can be effective in the influence of the gap ratio parameter, g ID, it is as-
improving joint strength.
The ratio diD is a measure of the concentration of the "The yield strength is the average of (Tu determined in directions parallel and perpen-
dicular to the direction of rolling. If this information is not available, the nominal yield
'Failure modes similar to Types 1, 2, and 5 can also occur in K-joints. strength should be used.
~1 ( t .. · ,
~
N
PUK-:= OyT'- y'lUT (1+6.52 d/DII1-.26 COS' (3){1.0544-1.59 g/Dl/SIN
O~----!..;;-----A;:-_---J=-~~_ _~_--.J:--_-L_ _...L.._--.J_ _-L_ _..L_--l_ _.-J
e
....> -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 .6 .7 .8
o 9/D
Fig. 3-lnfluence of the gap ratio parameter, 9 10, on the ultimate strength of tubular K-joints.
5 COMPRESSION
o KANATANI
cJISC
, 0 SEAL AND TOPRAC
\ "SMU
TENSIOI\f
• KANATANI
·JISC
• SEAL AND TOPRAC
.. TOPRAC AND NATARAJAN
m = 3.42 API safety factor would produce safe but, in many cases,
)( I;)
:::l
D.
very conservative designs. The same comment generally
80 = 22.75 ..
applies to the Det norske Veritas equations when the
10 safety factors of 1.5 for environmental loads and 2.0 for
operational loads are applied.
The T- and Y-joint comparisons are shown in Figs. 10
5 and 11. For some practical ranges of parameters it can be
seen that both API and Det norske Veritas predictions
.1 .5
may fall above the lower bound to available strength data .
dID Finally, Fig. 12 compares the new X-joint formula
Fig. 6-Ultimate strength of tubular X-joint. with the equivalent Det norske Veritas formula. The Det
VALIDITY RANGE
JOINT AND LOAD TYPE ULTIMATE STRENGTH FORMULA
19 ~ DIT ~ 93
.19 ~ diD ~ 1
P
U
= 30 ay T 2 (d/D)·64 ISIN e .8 ~ diD ~ 1
x
.8 ~ diD ~ 1
300 ~ e ~ 90 0
~ .59 glD) J
that the joint also can withstand these same design loads.
First, the computer output is screened to determine the X (1.05 _ 2/3 (5)
maximum applied loads. These are shown in the follow-
ing table. For an applied load of 904 kips on Member A, the
2r--------------------,
diD> .6 OR e < 90°
diD"';; .6
ii:
2r----------------------. « >-
:::l 1
e = 45° >- D.
:::l
D.
9/0> .16
o0=---:'=---:!=----=L:----!':40:---.l:--~-~-----I80
RIT
Fi~. 1O-Comparisonof API rule with T- and Y-joint
API S.F. 1.8 NOT INCLUDeD ultimate-strength formula (for compressive loading).
2,--------------------,
O=--_-:-:-_~-----II:_-...l.---L-----l.--..l.---.J
o ~ ~ ~ m 00
RIT COMPRessION
FI'g. a-Comparison of API rule with nonoverlapping K-joint
ultimate-strength formula. 1.5
'.5 >c:
'? >-
>- :::l
:::l D.
D.
.5
DnV - S.F. NOT INCLUDED
.5
O'----'-_--'-_--'--_-'----l_-.L_-l.-_..l..----JL.---..l
o0=-----.-:--,----..J,.2~-9-/D----'.3L----
.. L4- - - - - - J .5 o ~
diD
Fig. 9-Comparison of Det norske Veritas rule with Fig. 11-Comparison of Det norske Veritas rule with T- and
nonover1apping K-joint ultimate-strength formula. V-joint ultimate-strength formula.
1.5~--------------------'
50 KIPS
>t:
1
9 )(
)( ::>
::> a..
a..
.5
OL-_L----'_--i._-..L_--'-_---l.-_-'-_~ _ __'___'
o .5
diD
54" ¢ x 0.75" W.T.
Fig. 12-Comparison of Oet norske Veritas rule with X-joint
ultimate-strength formula (for compressive loading). Fig. 13-Example problem.