Вы находитесь на странице: 1из 9

Proceedings of the Ninth (1999) International Offshore and Polar Engineering Conference.

Brest, France, May 30-J une 4, 1999


Copyright 1999 by The International Society of Offshore and Polar Engineers
ISBN 1-880653-39-7 (Set); ISBN 1-880653-43-5 (Vol. IV); ISSN 1098-6189 (Set)
Load-Deformation Relationships for Gusset-Plate to CHS Tube Joints
Under Compression Loads
M. Ari yoshi and Y. Maki no
Kumamot o University
Kumamot o, Japan
ABSTRACT
Gusset-pl ate to CHS tube joi nts are connect i ons where the end of a
pl ate is directly wel ded to the out er surface of a tube. This study
proposes equati ons for ul ti mate and yi el d strength, initial stiffness
and local deformati on at ul ti mate strength for these types of
connecti ons. These equati ons are based on a ring model wi t h an
effecti ve wi dth usi ng mul t i pl e nonl i near regressi on analyses, To
descri be the l oad- def ormat i on behavi or of gusset-pl ate to CHS tube
joi nts, each l oad- def ormat i on curve is approxi mat ed by two strai ght
lines usi ng these equati ons.
KEY WORDS: Effective Width, Strength Equation, Gusset-Plate to
CHS Tube Joint, Load-Deformation Curve, Regression Analysis, Ring
Model
INTRODUCTION
A database for gusset-plate to CHS tube joints has been compiled at
Kumamoto University to provide other researchers with a starting point
for their own work on gusset-plate to CHS tube joints. Many studies
have been carried out in Japan with results published in Japanese
journals and conferences. The language barrier has prevented these
results from being found and subsequently utilized in research problems
outside of Japan. Therefore, the authors compiled the "Database of Test
and Numerical Analysis Results for Gusset-Plate to CHS Tube Joints"
(1998).
This investigation aims to produce simple load-deformation
relationships for the joints mentioned in the database. In this paper, the
ultimate and yield strength equations and the equations for initial
stiffness and deformation values at ultimate strength for gusset-plate to
CHS tube joints under compression loads are proposed. Each load-
deformation curve is described by two straight lines as shown in Figure
1. Functions fx and f2 are derived using the geometrical and material
parameters of the connections. Hence, these relationships estimate loads
and deformations at yield and collapse for gusset-plate to CHS tube
joints of any given geometry.
54
i . e .
where
N,
Ny
N
~L
F2
5y ~i,
Fn = f o ( Nu , Ny , 5 . , S y )
Nu = fl (~,13,~/, "" ") 8u = f2 (%13, ] ," " ")
i y = f3 (ct,[~,~/,. .) 5y = f4(o%13,"/,. )
Figure 1 Load-deformation relationship
The only existing study of the strength equations of gusset-plate to CHS
tube joints to the authors' knowledge is that of Makino (1984) / Makino
(1986) in which equations for ultimate strength were developed.
However, yield strength, initial stiffness and deformation values were
not considered in Makino's paper because of insufficient test and
analysis data. Therefore in this paper, the equations are given not only
for ultimate strength but also for yield strength, initial stiffness and
deformation values. The ultimate goal of this project is to offer available
material on strength and deformation and to acquire design formulae for
every type of gusset-plate to CHS tube joint.
CLASSIFICATION OF JOINT TYPES
The joints studied in this paper consist of gusset-plates and a CHS t u b e .
The configurations are similar to tubular X or T-joints with welded
gusset-plates instead of braces. This paper deals with only those
configurations where a stiffener is added to the CHS tube to resist shear
and bending moments across the cross-section of the CHS tube. Other
configurations, where load-transmission is axial and the connection
strength is not dependent on the deformation of the cross-section, are
Table 1 Classification of Joints
L _,
T P 1
I I I
Nt Nt
L .
XP1
N~
B
Joi nt
n B "T
F]
El
N~
4b


I @ .
13
I D I
n I
Nt
XP2
N~
n
K L
TP2
Nt
XP3
T
L
TP3
Nt
XP4
L
TP4
Nt
XP5
G - - n T
L
TP5
Table 2 Numbers and parameter ranges for the data used in the re~ression anal~'ses
Numbe r s Par ame t e r Range s
Test
28
TP- ] oi nt
Anal ys i s
27
XPq o i n t
D
(ram)
30
T
(mm)
101.6-
267.1
165.2-
457.2
3.04-
16
3. 5-
6.64
N
Ak
B
(ram)
2 , I '
(ram)
85.7- 50-
342.9 148.7
100- 82.4-
450 352
457-
2150
417-
273.6
mo ( ~ y
(MP~)
308-
479
0.068- 281-
0.967 441
N
~ A
C B
NGTH
Ny
r 8
I nN
~ s " " " SCATTER BAND
'"'1" 0 . 2 5
v i n 8
Figure 2 Load-deformation curve
and definition of ultimate strength
Figure 3 Definition of yield load by Kamba Figure 4 Definition of yield load by Kurobane
excluded from this paper. The forms of gusset-plate to CHS tube joints
are XP and TP-joints, which correspond to tubular X and T-joints
respectively. Each j oi nt type is classified into five types by the shape of
the joints, as shown in Table 1. In addition, two types of gusset-plates
are distinguished: in one type the extension is oriented laterally, a rib-
plate; in the other type the extension is oriented longitudinally, parallel
to the axis of tube, a gusset-plate.
Generally, joints are subjected not only to simple loads but also to
combined loads. In this paper, only joints subjected to simple
compressive loads are considered, as presented in Table 1. This paper
uses 85 tests and analyses to develop the equations. The numbers of
each j oi nt type and the parameter ranges are shown in Table 2.
DEFINITIONS FOR ULTIMATE AND YIELD STRENGTH, INITIAL
STIFFNESS AND DEFORMATION
Ul t i mat e St rengt h
A severe stress concentration occurs in the chord near the weld toe of
the plate. The chord incurs local bending deformation that leads to
failure. This paper is limited to those joints undergoing failure in this
way. If an axial force is applied to the plate, the local deformation of the
j oi nt usually conforms to curve A of Figure 2. In this joint, ultimate
strength is defined as the maximum value of the plate axial force.
Curves B and C of Figure 2 refer to joints where in some regions
stiffness increases after severe deformation of the chord. In curve B,
after the load-deformation relationship reaches its first maximum value,
the joint becomes unstable. After a peak load, the joint shows increasing
stiffness and becomes stable again. In this case, even i f the joint is
55
subjected to larger axial forces on the plate than the fkst maximum load,
the first maximum load is defmed as the ultimate strength. For a joint
that has a load-deformation relationship described by curve C, a
practical ultimate strength is defined as the load where the stiffness
starts to increase again.
Yield Strength
Kamba (1998) defined a third of the initial stiffness as the second
stiffness, and assumed the yield strength Ny.K~nb~, as the load where the
slope of the load-deformation curve is equal to the second stiffness, as
shown in Figure 3. On the other hand, as stated below, yield strength can
also be based on results of tests on actual joints and ring models as
proposed Kurobane (1984). A load-deformation curve is plotted on
double logarithmic graph paper as illustrated in Curve A of Figure 4, but
it is also possible to approximate the load-deformation curves by two
straight lines. The yield strength Ny,Kurobane, is defined as the load at the
intersection of the two lines, representing a point of maximum curvature
variation of the load-deformation curve. However, in cases where the load-
deformation curve can not be approximated by two straight lines, as
illustrated in curve B of Figure 4, a scatter band of 0.25 compared with the
initial slope of the load-deformation curve is adopted. The load at the
intersection of this line and the load-deformation curve is defined as the
yield strength. Comparisons of Ny, Kamua and N y , Kuroban e a r e shown in Figure
5. The yield strengths determined with Kurobane' s definition are
smaller than the values given by Kamba' s formula and the former is
regarded as safer values. In this paper, Kurobane' s method is adopted, and
yield strength is based on Kttrobane' s definition.
1.2
1
0. 8
~ 0 . 6
~
0. 4
Z; 0.2
~5o o
0 XP- j o i nt
O T P - i o i n t
0 20 40 60 80 100
D / r
Figure 5 Comparisons OfNy.Kuroban e and Ny.Kamba
Initial Stiffness
Initial stiffness is defined as the slope of a line tangent at the origin of
the load-deformation curve, as shown in Figure 1,not by a line secant.
to tubular X and T-joints respectively. It is also possible to develop
strength equations of gusset-plate to e l l S tube joints using the concept
of ring models. The derivations of ring models for X and T-joints are
given next.
Ri ng Model for an X- j oi nt
An X-joint is approximated by a ring with an effective width Bo, as
shown in Figure 6 (a). Plastic strength of the ring No, is calculated using
plastic hinge theory and given by Eq. (1).
N,: B e 1
m . . . . . .
T2~y R l_d (1)
D
In a more accurate derivation of the plastic strength, it becomes clear
that the strength also changes with D/T. A mathematical model of
ultimate strength for rings is assumed as in Eq. (2) by adding a term D/T
to Eq. (1).
R N u " e O1
~ = ~ ~ ~ - f - ) ~ (2)
TEoy D 1 +131; 2 D
Here, ~ is an error term while %, % and % are regression coefficients.
That are determined by regression analysis, The results are presented in
Eq. (3).
4 4 9 ( 3 0 o ,
TE~y D 1_0. 917 d
D
(3)
The ring model method described above and multiple nonlinear
regression analyses on a large number of test results provide the basis
for the strength equations for tubular X, T, Y and K-joints suggested by
Makino et al (1984).
Ri ng Model for a T-joi nt
In a T-joint, only two concentrated forces operate on the ring model at
two points outside the brace diameter of model, as shown in Figure 6 (b).
The ring receives a shear force q as a reaction force. The T-joint is
regarded as the ring model in beam theory. The plastic failure load of a
ring with a width No, is given in Eq. (4).
T2~y R
Deformati on
In this paper, deformation is defined as the local deformation of the joint
(i.e. indentation), not as the deformation of the whole chord.
1 +cosl 3
l , l + c o s 0 ) s , _ ( , _ + c o s , , s m 0
( 4 )
TESTS OF RING MODELS AND NONLINEAR ANALYSES
Since many investigations of tubular X and T-joints have been
completed, strength equations of these joints using geometrical
parameters have been suggested. Some of these equations are based on
analytical model, e.g. simple ring models. XP and TP-joints correspond
Here, sin 13=d/D and 7 is an angle formed in a plastic hinge, y is obtained
from Eq. (5).
( 1 - 1 c o s - - [ 1 0 +cos , + c o s
56
N 3
r-yL . .
i ) _ _ ; _ _ r " . . . . .
k-.--_J
(a) X-joints
N/2
N/2
N ,';. N ~;. N ~;.
w
~ N ~ N ' 2 I - ~ - ' F F ~4-d-I~ . . . . i i 2 ~ N '
__ _ t . . . . I . . . . . N/ /2
L..J
, , / 2 i - - ; - - 1" ] . . . . . j l q ~ T
k-. --. " . . . . .
\ ]! N/2 N/Z '~ INDICATES
i i ~" i! , . o ~ , , ~ o , , o , "
N
YIELD HINGE
(b) T-joints
Fi gure 6 Ri ng model s
EQUATIONS TO DESCRIBE LOAD-DEFORMATION
RELATIONSHIPS
U l t i ma t e S t r e n g t h
( 1 ) X P - j o i n t s
It is assumed that an XP-joint under compressi on can be approxi mat ed
by a ring with an effective wi dt h Bo, as shown in Fi gure 7 (a). A
mat hemat i cal model for estimating the ultimate strength of XP-joints is
obtained from Eq. (1). A t erm D/ T is added to make the equation
applicable for a larger range.
N. x _ cq
- C (6)
T2y 1 + ~x 2 - -
D
Fr om regression analyses, t he following ultimate strength equation for
XP 1 and XP3-joints is obtained:
Nu, , : - 1.96 ( D ) ' 2 7 3 _
T2oy 1- 0" 857 c (7)
D
n = 39 mean = 1.00 CoV = 0.090
In XP3-joints, the gusset-plate protruding out of a fib-plate is regarded
as part of t he l oaded effective width, so the gusset-plate can be ignored.
The effective wi dt h of XP1 and XP3-joints is calculated by considering
Eq. (7) to be rel at ed to Eq. (3), as follows:
1 - 0 . 9 1 7 C 0.474
D 1 - 0 . 8 5 7 C
D
The XP2 and XP5-joints can be consi dered to be represent ed by a ring
wi t h an effective wi dt h (Be+B). A chord lies bet ween two gusset-plates
and extends outside of t he gusset-plates. These portions of t he chord
also collaborate to transfer the load as shown in Fi gure 7 (a). Therefore,
strength equations for XP2 and XP5-joints are based on the strength
equation deri ved for a ring model, Eq. (3).
4 4 9 ( ) 0 0 1
T 2 ~ y D 1_0. 917 c (9)
D
Beq) is calculated using Eq. (8), so that Eq. (9) becomes:
i
f - - f - ' v
. . . . I " " ' - - " " " 1
, i
: :
N[i
N ~.
N / 2 ~ N / 2
N/2 ! ! N/2
f i
N I !
( a ) X P - j o i n t s
J
'~ . . . . . !" N / ' " ' [ 2
i . . . .
11 LOCATmN o r
i , t~ Y I E L D H I N G E
(b) TP-j oints
Fi gure 7 Ring model s
N,,c, _ 1.96 (~3273 B _ _ + - - - 4.49 ( D ) -21
T 2 ~ y 1 _ 0 . 8 5 7 D 1 _ 0 . 9 1 7 (10)
D D
n = 3 me a n =0 . 9 4 5 COV=0. 030
C of XP2-joints represents the thickness of the gusset-plate and is small
enough to be safely ignored. It is possible to obtain t he ultimate strength
for XP2-joints using the following simpler equation:
- -0.273 - - - 0 . 2 0 1
Nu,c = 1 . 9 6 / D ) + 4 . 4 9 B / D 3 (11)
T 2 c r y
n = 12 mean = 1.24 CoV = 0.056
The strength equation of XP4-joints with two rib-plates is as follows:
N u , c - 1 . 9 6 a n (D3273
T2try 1_0.857C__ (12)
D
n = 6 me a n = l . 1 6 COV=0. 078
The relationship bet ween o~B and the gap of the two fib-plates B, can be
formulated as follows:
1 B
~ B = 1 + - - . ~ ( 1 3 )
4 D
I f B is large enough, it is equal to the supporting load of two
i ndependent fig-plates, that is o~B=2. On t he other hand, i f B is close to
zero, t hen the j oi nt is effectively the same as XPl - j oi nt s, whi ch means
that c~B=l. The effects of a gusset-plate can be i gnored for the same
reason as for XP3-joints.
In Fi gures 9 (a)-(c), although test and numerical results to prediction
ratios for XP2-joints are a little high, it is observed that the data and the
predicted ultimate strength values agree well. It is concluded that the
assumptions about the effective width are appropriate.
( 2 ) T P - j o i n t s
As for an XP-joint, a TP-joint is regarded as a ring with an effective
wi dt h Be, as shown in Figure 7 (b). For this type of joint, an axial force
in the brace is supported at both ends of the chord, so bending moment s
acting on the j oi nt depend on the chord length. In addition, shell-like
movement changes with the chord length. Therefore, a function (1-
0.32m0) that is a reduction factor for chord member global bending
effects is added to Eq. (6) to include these factors. The ultimate strength
is obtained as follows:
57
N, , ~ _ oq (1_ 0.32mo)
~ - ----------d (14)
T2'y 1 + a 2
D
Ultimate strengths for TP1 and TP3-joints are estimated by regression
analysis in Eq. (14).
N, , c _ 39.9 ( D) - ' 429_
( 1- 0. 32mo)
T2t~y 1 - 0 . 9 2 8 (15)
D
n = 16 mean= 1. 01 CoV =0. 157
TP4-joints are considered in a similar way as XP4-joints. This strength
equation is the following:
Nu , ~ _ 3 9 . 9 t ~ B (1 _ 0.32mo )
T2t~y 1- 0"928 C (16)
D
n = 2 mean =1. 10 CoV =0. 086
The behavior of TP2 and TP5-joints can also be approximated by a ring
with an effective width (Bo+B) similar to XP2 and XP5-joints. Their
ultimate strengths are predicted using Eq. (4).
Nu, c B~+B ( C !
(17)
In addition, the effective width of TP-joints is obtained using Eqs. (4)
and (15).
D/-0.429
1_o.92aC 0 -0 32mo)
Be D (18)
39.9
Eq. (18) is rather complicated, so for practical use it is simplified as
follows. Assign q = Bo(D/T)429/R(1-0.32m0), and plot q versus C/D as
shown in Figure 8. From t hi s figure, it has been determined that i f
C/D<0.9 then q is between 21.2 and 25.0 and almost constant. For
C/D<0.9, assume q = 22.3 and R = D/2. Consequently, Eq. (34) can be
represented as follows:
Nu , ~ _ 39.9 (1 _ 0.32m0) + B
T2cy 1_0. 928 c l l . 2 D (19)
D
n = 1 mean = 1.11
Particularly in the case of TP2-joints, C can be ignored, because the
value is small. Hence, the strength equation can be simplified further.
Nu'~ = 39.9 ( 1- 0. 32m0) + (20)
TE{y y T 11.2D
n =8 mean =0. 994 CoV=0. 110
Test results for prediction ratios for TP-joints are almost all between 0.8
and 1.2, and ultimate strengths are predicted weU, as shown in Figures
10 (a)-(c).
58
'30
e,t
~. 25
. g 20
~" lO
* 5
o g-
39.9/(1-0i928C/D)f ~
0 0. 2 0. 4 0. 6 0.8 1
C/D
Figure 8 Effective width for TP-joints
Yi el d Strength
Using the concept of effective width for a ring model makes it possible
to establish the ultimate strength equations on gusset-plate to CHS tube
joints. The same ring model methodology used above for ultimate
strength can be used to determine the yield strength for both XP and TP-
joints. The equations for yield strength are shown in Table 3, and Eqs.
(21)-(28). The test and numerical results for prediction ratios of those
equations for XP and TP-joints are plotted in Figures 11 (a)-(c) and 12
( a) - ( c) .
Ini ti al Stiffness
Kamba (1998) suggested describing the physical quantity as a non-
dimensional product of all the individual quantities related to it. For the
initial stiffness of a connection, the physical quantities are the materials'
properties and the size of the members. Young' s modulus is selected to
represent the materials' properties whereas the thickness of the chord is
taken to represent the size of a member. The initial stiffness is obtained
as follows:
=et 1 ~ (29)
E T
The following strength equation for XP1 and XP3-joints is obtained
after using Eq. (29) as a basis for regression analysis.
K"'c = 7.45 (30)
E T
n=36 mean =1.04 COV=0.342
The strength equation for XP4-joints is similar to Eq. (30) after adding a
term aB.
Kn'c - 7.45ct B (31)
E T
n=2 mean=l . 30 COV=0.298
The XP2 and XP5-joints are based on the concept of the elastic
deformation as a result of a point load acting on a cylindrical shell as
shown in Figure 7 (a). The deformation of a model loaded by a point
load is calculated using the following formula.
p R 3
8 = 0 . 1 4 9 ~ (32)
E I
From Eq. (32), the initial stiffness of this model is represented as
follows:
R K n =8. 95 ( 3 3 )
ET
As the function in Eq. (33) is considered insufficient to be applied to the
initial stiffness of XP2-joints, a second term including D/I" is added to
Eq. (33):
( D ) - 2 ( n ) ( D ) 0.409
Kn'~ - 8.95 .0.876 (34)
E T
n=10 mean=l. 01 COV=0.123
For XP5-joints, the initial stiffness is considered to be the sum of the
values calculated using Eq. (30) and the initial stiffness of a model
loaded by two point loads as shown in Figure 7 (a). The deformation
equation of the model is as follows:
5 = 0.076 - 0. 976 E I
Therefore, the initial stiffness of XP5-joints is predicted using the sum
of Eqs. (30) and (35), as shown here:
_ \ - 1 . 3 2 / - - x l .0 8 - 2 B C - 2
K,,,,:ET - 7 " 4 5 ( T) / D ) +8 " 7 8 ( D) ( - ~- ) ( - ~- - 0. 976) (36)
From Figures 13 (a)-(c), it is observed that initial stiffnesses obtained
are accurate except for one data point of an XP 1-joint. In a similar way,
equations for TP-joints are established, and shown in Table 4, as Eqs.
(37)-(40) and Figures 14 (a)-(c).
Def ormat i on at Ul t i mat e Strength
Similar to initial stiffness, the deformations values at ultimate strength
are presented in non-dimensional form by dividing the actual
deformations by the chord diameters. Referring to a mathematical model
for a ring, as in Eqs. (41) or (42), it is possible for the local deformation
at ultimate strength to be calculated.
~Suc ( D) ~2 ( C ) ~3 (41)
~ ' = O I ; 1 [ ;
D
~3u'~ - a 1 ~ (42)
D
There are few test and analysis data for some types of joints in this
paper. Therefore, equations for deformation values at ultimate strength
are applied to Eq. (41) in XP1 and TPl-joints, and Eq. (42) in XP2 and
TP2-joints respectively. The equations are shown in Table5, and Eqs.
(43)-(46). The test and numerical results for prediction ratios of those
equations for XP and TP-joints are plotted in Figures 15 (a)-(b)
n=3 mean=0.855 COV=0.154
Table 3 Ec~uations for ~ield strength
Type I I .. Yi el d St rengt h Formul ae I' No. I T y p II Yi el d Strength Formul ae No.
XP1
XP3
XP2
XP4
XP5
Ny, c _ 1 . 2 0 / D ) 0"395
T 2 ~ y 1_0. 795 C
D
n=36 mean=l. 01 Coy=0.110
TP1
(21) TP3
n=10 mean=0.893 Cov=0.146
(22)
Ny, c _ 1 . 2 0 a n I D ) 0 " 3 9 5 _
T 2 f f y 1- 0. 795 c
D
1 B
a B = 1 + - - . ~
4 D
n=2 mean=l . 13 Cov=0.164
(23)
N.v,,: _ 1.20 __ + 1.75
T 2 ~ y 1- 0. 795 C
D
n=3 mean=l . 08 Cov=0.045
(24)
TP2
TP4
TP5
N y , ~ 144 ( TD../-91
T E r r y 1_0. 828 C (1-0"32m)
D
n=15 mean=l . 02 Cov=0.188
' ( o )
N ' v ' : =144 ( 1- 0. 32m0)+
r o, f f j
n=8 mean=l . 17 Cov=0.119
B
35.2D
_i
N y , c _ 144an ( D ) -9~
T2Cr.v 1_0. 828 C (1-0"32m)
D
1 B
o~ n =14 . . . .
4 D
n= 1 mean= 1.39
N y , c _ 144 h _0. 32m0~+ 35.2 D
T 2 ~ y 1- 0. 828 c
D - -
n=l mean=0.966
(25)
(26)
(27)
( 2 8 )
59
Table 4 Ec~uations for initial stiffness
T y p e 1] I n i t i a l S t i f f n e s s F o r m u l a e
TP1
TP3
TP2
TP4
TP5
.... =2. 04
ET
n=15 mean=l . 04 Cov=0.289
Knc ( T / - 2 1 B ) I D ) 4 4 4 ( B/ - 0 " 9 8 0
.... =8. 49 -1.09
ET
n=8 mean=l . 00 Cov=0.065
0 : ' ( C / T M
.... = 2 . 0 4 ~ m
ET
1 B
o~ B = 1 + - . m
4 D
n=l mean=l . 08
Kn c (D /-0 .7 1 8 (C )1,26 ( D /- 2 ( B ) ( c )-1
.... = 2.04 +0. 667 1 5 . 1 - - - 0 . 6 3 6
ET D
n =1 mean=2.31
Table 5 Eciuations for deformation at ultimate strength
Type I I D e f o r m a t i o n a t Ul t i mat e St r engt h For mul ae No. I Type I I
XP1
XP2
~"':D ' = 0 .1 6 3 (-~ )-4 2 1 ( C ] 0"089-
n=33 mean=l . 02 Cov=0.183
8"CD' = 0 . 0 1 7 ( D) 32s (DB---) 22
n=9 mean=l. 00 Cov=0.052
(43) TP1
( 4 4 ) T P 2
N o .
(37)
( 3 8 )
( 3 9 )
( 4 0 )
D e f o r m a t i o n a t U l t i m a t e S t r e n g t h F o r m u l a e [ N o .
( ~ u c ( -- " 0"264 ( C / ' 4 4 3
1
D
n=14 mean=l. 01 Cov=0.147
~ ) u c D' 0 . 0 0 2 ( D ) ' 8 7 7 ( B ) 0"294
n=8 mean=l. 01 Cov=0.159
(45)
(46)
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
0
1.6
1.4
- " | i o x P 2 1 1
w e ; / ] 1 3 ~ - | I~XP3[ ~ " o . s
ioxP4l z" 0.6
I~<XPSl 0.4
0.2
0
20 40 60
1
1.6
1.4
1.2
0.8
Z;m 0.6
0.4
0.2
0
I
x l I t e ' e
<> XP4[
XXP51
. ,.<>~ I O X P l l
i i j ~ z ~ IZ~XP3 I
IoxP4[
IXXPSl
0 1 2 3 4 0 0.2 0.4 0.6 0.8 1
Df r B/D C/D
(a) against D/T (b) against B/D (c) against C/D
Figure 9 Test and numerical results to prediction ratios for ultimate strength (XP-joints)
1.4
1.2
1
' 0 . 8
0.6
0.4
0.2
0
1.4 1.4
~I " 1.2 41 ~) 1.2 EI
q h~ ~ _ ~ [ T P I ] 1 1 - -
l i , o l e 0 , , ' " [ ~ T P 3 ] 0.S " TP4 "
[<) TP4[ ~;* 0.6 TPS] ~ 0.6 ....
[~TpS[ 0.4 0.4
0.2 0.2
0 0
40 60 80 100 0 0.5 1 1.5 2 2.5 0 0.2 0.4 0.6
D/T C/D
(a) against D/T (c) against C/D
B/D
(b) against B/D
Figure 10 Test results to prediction ratios for ultimate strength (TP-joints)
I A] _
0 . 8
T P 1 [
ATP3[
<> TP4[
~<TP5]
60
1.4
1.2
1 ,
0.8
~ 0 . 6
0.4
0.2
0
20 40
D~
( a ) against D/T
! I . x P , I
; IOXP21
I~XP3 I
i<>xP41
I~<XeSl
1.4 1.4
1.2 1.2 ~ ) ~ "(
1 C) ~ ()
1
() ' OXP4' 0.8
0.6 I~<XeSl z o.6
0.4 0.4
0.2 0.2
0 0
60 0 1 2 3 4 0 0.2 0.4 0.6 0.8
B/D C/D
( b ) against B/D ( c ) against C/D
Figure 11 Test and numerical results to prediction ratios for yield strength (XP-joints)
OXel I
AXP3]
o x P4 I
XXP51
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
w
40
D/T
1.6 / 1.6
1.4 ~ D ~ ~ 1.4
I " TPI[ 1.2 ~ 1.2
- - ~ [OTP2] ~ 1 O ~. IoT~[ l /(,
i A TP31 0.S i < > TP41 ~ 0.S
10 TP4] 0.6 IXTP51 ~ 0.6
IXTPSI 0.4 0.4
0.2 0.2
0 0
80 100 0 0.5 1 1.5 2 2.5
B/D
0.2
( a ) a g a i n s t D / T ( b ) against B/D
Figure 12 Test results to prediction ratios for yield strength (TP-joints)
I I !
I l l
0.4 0.6
C/D
(c) against C/D
0.8
o Tel I
ATe3 I
<>TP4[
XTP51
3
2.5
2
~ 1.5
1
0.5
0
20 40
D/T
( a ) a g a i n s t D ~
3
2. 5
I
o x P , I
oxP21 ~ 2
AXP3[ 1.5
i OXP4 1 - - - - ~
I I IXXPSi
0.5
2. 5
i o x ~ I 2
O IOXP41 ~ 1.s
0.5
0 0
60 0 1 2 3 4 0 0.2
B/D
( b ) against B/D
Figure 13 Test and numerical results to prediction ratios for initial stiffness (XP-joints)
I A - - AA
| w l
o
0.4 0.6
C/I)
( c ) a g a i n s t C/D
0.8
xP1 I
xXeSI
2.5
2
~
1.5
0.5
0
(I
'
i
"0
40 60
Df r
( a ) a g a i n s t D ~
2.5 2.5
[OTPll 2 2
I OT ~ I I o m l
I A m l ~ 1.s l.s
I<>TP4I
i ~ e l I<>TP4I ~ 1 - - ~ : b - - ~ b - - It I:~TPSl 1
I~TYSI 0.5 0.5
0 0
80 100 0 0.5 1 1.5 2 2.5
B/D
(b) against B/D
0 0.2
Figure 14 Test results to prediction ratios for initial stiffness (TP-joints)
O
I
S
0.4 0.6
C/D
(c) against C/D
0.8
I ' T P l l
IATP3 I
I<>TP4I
I~TPSI
1
6 1
1.6 1.4
1.4
1.2 II 0 I 1.2 D ~
o.s - | " xvzl ~ 0.6
'~ 0.6
0.4 0.4
0.2 0.2
0 0
0 20 40 60 0 20 40 60 80 100
[
e TP1]
oTr~l
D/T D/T
(a) XP-joints (b) TP-joints
Figure 15 Test and numerical results to prediction ratios for deformation at ultimate strength
CONCLUSIONS R
T
In this paper, ring models with the effective width and regression
( Z 1, ( Z 2, ( Z 3, . . . O~ n
analyses has been used to establish equations for ultimate and yield 8u
strength, initial stiffness and deformation values at ultimate strength for
8u, c
gusset-plate to CHS tube joints. The predictions using those equations
s
are accurate for both XP and TP-joints, and are as well as values
Uy
observed in the results of tests and analyses in the database. There are
many types of gusset-plate to CHS tube joints; nevertheless, only a few
are dealt with in this paper. However, it is judged that the equations
developed in this paper may be applied to other gusset-plate to CHS
tube joints as well. Unfortunately, for some types of joints no test or
analysis data are available. Therefore, it is judged that the data are not
adequate for developing applications for all types of joints.
SYMBOLS
B
Bo
C
D
d
E
F1, F2, F3 . . . . . Fn
~, ~, ~ .....
I
K.
I~o
L
M0
m0
Nc
N~
N~
RNu
RN~,c
Ny
y, c
n
width of gusset-plate
effective width
width of rib-plate
chord diameter
brace diameter
Young's modulus
functions
functions
geometrical moment of inertia
initial stiffness under axial force
estimated initial stiffness by regression analysis
initial stiffness under axial force of ring model
chord length
bending moment in chord at center of joint
full plastic moment capacity in chord member
bending moment ratio (Mo/Mp)
plastic strength
ultimate capacity of joint measured as axial load
estimated ultimate capacity by regression analysis
ultimate capacity of ring model measured as axial
load
estimated ultimate capacity of ring model by
regression analysis
yield capacity of joint measured as axial load in
gusset-plate
estimated yield capacity by regression analysis
number of data
chord radius
chord thickness
parameters of regression analysis
local deformation of joint reached at ultimate strength
estimated local deformation by regression analysis
error team in regression analysis
yield stress of chord material
ACKNOWLEDGEMENTS
The authors would like to thank G. J. van der Vegte and Y Minehara for
their assistance and contributions to the work reported here. Without
their support, this paper would not have been possible.
REFERENCES
Kamba T., and Taclendo C., (1998). "CHS column connections without
stiffener," Proc. of 8 th Int. Symposium on Tubular Structures (ISTS),
Singapore, pp.567-576
Kurobane Y., Makino Y., and Ochi K., (1984). "Ultimate Resistance of
Unstiffened Tubular Joints," Journal of Structural Engineering, Proc. of
the American Society of Civil Engineers (ASCE), U.S.A., Vol. 110, No.2,
pp.385-400
Lu L. H., de Winkel G. D., Yu Y., and Wardenier J., (1994).
"Deformation limit for the ultimate strength of hollow section joints."
Proc. of 64 Int. Symposium on Tubular Structures (ISTS), Australia,
pp.341-347
Makino Y., (1984). "Experimental Study on Ultimate Capacity and
Deformation for Tubular Joints," Doctoral Dissertation, Osaka
University (in Japanese)
Makino Y., and Kurobane Y. (1986). "Recent Research in Kumamoto
University in Tubular Joint Design," IIW Doc. XV-615-86 / XV-E-86-
108, Japan, pp. 1-28
Makino Y., Ariyoshi M., Minehara Y., van der Vegte G. J., Wilmshurst S.
R., and Choo Y. S. (1998). "Database of Test and Numerical Analysis
Results for Gusset-Plate to CHS Tube Joints," I1W Doc. XV-E-98-237,
Singapore
Yura J. A., Zettlemoyer N., and Edward I. F., (1980). "Ultimate
Capacity Equations for Tubular Joints," Proc. of Offshore Technology
Conference (OTC), USA, Vol. 1, No. 3690, pp. 113-127
62

Оценить