Beruflich Dokumente
Kultur Dokumente
A A
STR/^IN DIAGRAM
L IJ N.A.
p STRESi S DIAGRAM
r
Fig. 1. Double (reduced) modulus theory
SECTION A-A
Ae
A STRAIN DIAGRAM
L IJ
^t
Acr = Et A
1
S T R E S S DIAGRAM
Fig. 2. Tangent modulus theory
Coupon Test
CO
CO
<1>
03
Strain e
Fig. 3. Stress-strain relationship for steel
umns above the tangent modulus load. Using a simpli- For any finite increase of axial load above the tangent
fied physical model, Shanley showed that bifurcation of modulus load, the column assumes equilibrium positions
equilibrium will take place when the applied load reaches with increasing deflection accompanied by a strain re-
the tangent modulus load. After bifurcation, increase in versal on the convex side of the column. Nevertheless,
lateral deflection is accompanied by a slight increase in the amount of strain reversal is less than that of the re-
load above the tangent modulus load. Thus the maxi- duced modulus theory. The readers are referred to a pa-
mum load is really slightly larger than the tangent mod- per by Johnston^ for a more thorough discussion of the
ulus load, provided the column is perfectly straight. Ex- historic highlights of the column buckling theory.
tensions of Shanley's model to describe the buckling The discussion so far pertains to columns which are
behavior of columns above the tangent modulus load were perfectly straight. Columns in reality are rarely perfectly
reported by Duberg and Wilder^ and Johnston.^ straight. Geometrical imperfection in a column tends to
In Ref. 5, it was shown that if a column were artifi- lower the maximum load of the member. As a result,
cially held in a straight position up to a load somewhere the Structural Stability Research Council (formerly the
in between the tangent modulus and reduced modulus Column Research Council) recommended the tangent
loads, then released, it would start to bend with an in- modulus load be the representative failure load of a cen-
crease in axial load. The magnitude of the increase, trally loaded column.
however, was less than that of the tangent modulus load. The reduced modulus theory and the tangent modulus
If the column was held in a straight configuration up to theory, as well as the Shanley's concept of inelastic col-
the reduced modulus load, then released, it would bend umn, are all based on physical reasoning. They provide
with no increase in axial load. Reference 6 demonstrates solutions and explanations to the behavior of perfectly
that when a column buckles at the tangent modulus load straight inelastic columns. The mathematical theory of
there is no strain reversal only for an infinitesimal in- elastic stability and the concepts of inelastic buckling are
crement of axial load. well explained in Refs. 8 and 9.
Thus, for columns with slenderness ratios less than or AISC/ASD Curve
equal to Q , the CRC curve assumes the shape of a pa- The CRC curve divided by a variable factor of safety of
rabola and for slenderness ratio exceeding C^, the CRC
curve takes the shape of a hyperbola, i.e. 5 3 [KL/r\ 1 (KL/r\^
3 8\C, / 8\C,/
(KL/rf KL _ 5 3 / \, \ 1
F. 1 sC
r ~ 3 ^ 8 VV^j " 8 VV^.
Fcr = (6)
IT^E KL in tiie inelastic range and a constant factor of safety of
\(K L/r? r >c. 23/12 in the elastic range gives the AISC Allowable Stress
1.0 Elastic
regime
0.8
J 0.6
0.4
AISC
Allowable Stress
Design Curve
0.2h
X;
1
AISC X, <V2
1 / X,
3 8 VV2/ 8 \V2
Allowable Stress Design
P 12 1
Curve X, > V2
P, 23 X;
AISC
Plastic Design
1.7I1--
Curve < 1.0 X, < V 2
P. 5^^/K
3 8 Vy2/ 8 VV2,
Plastic Design
(PD)
F.S.
//
I
JU
t
1
1
r JL
1
r
\ 1
1
1
1
is shown by dashed line 1
Load and Resistance 1
\ 1
1
1
Factor Design \ 1 1
\\
(LRFD) (^R,, > 1 7, Q. \ 1'
m m n- mw V rr
^\ 1
Tfim
\ \ 1
'Hieoretica] K value 0 5 0 7 1 0 1 0 2 0 2.0
Recommended design
tor K for the column with the particular set of end con- value when ideal ccmdi- 0 65 0 80 1 2 1 0 2 10 2.0
1 tioDS are approximated
ditions can be obtained by
"T Rotation fixed and translation fixed
Rotation free and translation fixed
K = (12) End condition code
Rotation fixed and translation free
Curve Fitting
Through D a t a Points
a.
100
Deflection ( 8 ) Slenderness Ratio ()
0.80H
0.40H
0.20H
Upon investigations of 83 end-restrained columns,^^ analysis assuming perfectly straight columns with end
the values of K for each curve do not vary significantly restraints provided by linear elastic rotational springs
over the load levels. Thus, a relationship between the K- having spring stiffness Rkj at the ends. Such comparison
factor and the magnitude of end restraint can be estab- is shown in Fig. 9. The dotted line is a plot of K versus
lished. In particular, the expression d whereas the solid lines are plots of K^i versus d. As
can be seen, K^i gives a conservative estimate of column
/ ^ - 1.0 - 0 . 0 1 7 d > 0 . 6 (14) strength provided that X is relatively low and d is rela-
where tively high.
2EL
(15)
(MXL, 2EL
MOMENT
M
in which
Ig = moment of inertia of the girder connected to
the column
Lg = length of the girder
(Mp)^ = plastic moment capacity of the column
Rki = initial connection stiffness of the connection
joining the beam to the column (Fig. 8)
was proposed^^ for non-sway columns with initial crook-
edness, residual stresses and small end restraints, taking
into account the effect of beam flexibility. Procedures
for the design of such columns have been reported in
Refs. 12, 21 and discussed in Ref. 22.
At this point, it is interesting to compare the effective ROTATION 9^
length factor K as described by Eq. 14 with the elastic
effective length factor K^i determined by an eigenvalue Fig. 8. Determination of R,,i
LU
O
LJ
U.
U.
bJ
OC
Fig. 9. Comparison of K^i and K
4. COLUMNS IN FRAMES
As mentioned earlier, columns in real structures usually
exist as part of a frame. A column in a frame is usually
subjected to the combined action of bending moments M.
and axial thrust. As a result, part of the strength of the
member is required to resist the bending moment and t I \ \ \^
only the remaining part of the strength is available to 15 ^^---J7
resist the axial force. Thus, most columns in frames must
be treated as beam-columns.
Case ^' Cm 1
Since the total deflection at midspan is the sum of the
primary and secondary deflections, i.e. 1
in which
C = 1 + it;P/P,
-i^\- -0.2 l-0.2P/Pg
dH H
Mo = HL
^PL (34)
H,=-H,
H2 ~ ~ ' ' 2
-H
Hj = - H j
TW 77/T
Br
Fig. 16. Connection moment-rotation idealization used in the power model
500
EXPONENTIAL MODEL
10 20 30 40 50 60 70 80
Fig. 17. Comparison of exponential connection model with test by Lewitt,
Chesson and Munse
TEST (RATHBUN)
EXPONENTIAL MODEL
I 3
c
O
O
o
0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008
ROTATION (radians)
Fig. 18. Comparison of exponential connection model with
test by Rathburi
60
Mc\
M
'I
40 +
-20
A '-A (3)
-40
''RL'
^ 6 5 . 5 4 K- in
0 M
-60
A- '--i\ Total
Fig. 19. Moment distributions at joint of subassemblage
; M,
r?)
MRR = Ml - M2 - M3 (38)
Moment TransferFlexible Connection
The bending moment in the column Mc can be ob-
tained by considering joint equilibrium. If the connections are not rigid, the moment transfer
mechanism between the beam and the column are more
Mr -(Mj,, + M^j,) - 2M3 (39) complicated because of the loading/unloading charac-
teristic of the connections. To demonstrate this charac-
The variation of these bending moments with the axial teristic schematically, the readers are referred to Fig. 21.
load P is plotted in Fig. 19. It can be seen from the For this subassemblage (Fig. 21a), the beams are con-
figure that the moment due to the buckling of the column nected to the column by semi-rigid connections. Beam
is not negligible. Not only does it reduce the moment of loads W, WR are first applied to simulate the dead load
the left beam, but, together with the moment arising from of the structure. Figure 21b shows the directions of mo-
joint rotation, restrains the column during the final stage ments acting on the left- and right-hand side of the joint
of loading. The moment of the right beam M^^, at first of the subassemblage. The corresponding M-6^ curves
inducing moment to the column, decreases gradually and for the left and right connections are also shown. The
at P = 26 kips reverses sign and becomes a restraining left connection will follow curve OA' and the right con-
moment to the column. On the other hand, the moment nections will follow curve OA". The moment acting on
of the left beam M^^ is always negative and thus always the column will be M^ on the left side of the joint and
restrains the column. MiR on the right side of the joint.
(a)
M2L(Zr: )
(c)
M3R
(d)
120
fhf)
^ j MB. V _L 1 ^- f
YW 1 4 x 3 0
W8x31 15' 'en
100
X^^
a /
2c 80
/
0)
f/f/i E 60
o
(a) Load Sequence 1 1
7
40
20
mu'
^ nm
^
0.01 P f 1 1 1 1
10 15 20 25
(xlO"^)
0.01 P
Rotation (radians)
(b) Load Sequence 2 In Fig. 27, the magnitude of joint translation and joint
rotation as a function of the applied force P are plotted.
Fig. 23. Flexibly connected T-shaped subassemblage Although the magnitude of joint translation for both the
rigidly connected and flexibly connected subassem-
blages are comparable, the joint rotation of the flexibly
is apparent in Fig. 25. Of particular interest is the di- connected subassemblage is significantly larger than that
rection of MBL. For the flexibly connected subassem- of the rigidly connected subassemblage. As a result, the
blage, MQI is always negative whereas for the rigidly moment induced as a result of joint rotation will out-
connected subassemblage MBL is only negative at low weigh that of joint translation, hence the final value of
values of P but becomes positive at high values of P. MBL for the flexibly connected frame is negative.
The reason for this can be explained by reference to Fig. As for the right beam, regardless of whether joint
26 in which the beam end moments at the joint are de- translation or joint rotation dominates, the induced MBR
composed. At the end of load Sequence 1, M^L is neg- is almost always negative. As a result, this beam, except
ative (i.e. counterclockwise, Fig. 26a). However, as load at the initial loading stage for the rigidly connected sub-
sequence 2 commences, the induced moment MBL may assemblage, will always provide restraint to the column
be positive (i.e. clockwise. Fig. 26b) as a result of joint regardless of whether the connection is rigid or flexible.
translation or negative (i.e. counterclockwise. Fig. 26c) It should be mentioned that unloading occurs at the con-
as a result of joint rotation. Whether the final value of nection which connects the right beam to the column as
MBL is positive or negative depends on whether joint load sequence 2 commences because the direction of
translation or joint rotation dominates. moment at this location is opposite for load sequence 1
^^
^r'^
'^BL'**^BR
Mc
2.0 2.5
(xlO"^)
Joint Rotation (radians)
(b)
Fig. 27. Joint displacement and rotation of
-400 -200 0 200 400 600
the T-shaped subassemblage
Joint Moment
. (in-kips)
(b) Flexible Connection
Fig. 25. Applied force vs. joint moment relationships and load sequence 2 (see Fig. 26). Consequently, the left
connection is offering tangent stiffness restraint to the
column whereas the right connection is offering initial
stiffness restraint.
c^; (a)
A more detailed analysis and discussion of the behav-
ior of subassemblages with flexible connections are given
elsewhere. ^^
Joint moment at the end of lood sequence 1
7. DESIGN OF COLUMNS WITH SEMI-RIGID
CONNECTIONS
0.6
C M C M
^^ my ^ ^-^ uy 1.0 (44)
where 0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0
c: 1.0 [ -^1 1 1 1 1 1 1 1 1 1 e (ft)
0
c= ^-v
P P 1 Q
CO 0.9
C=1.6- '^ u ^ y forO.5 < V ^ - 1 - 0 (45) n
L-
< X '-'
2[\n(P,Py) Q:
-J ^-^ 0.8
_5_o_ , - _ _ - ^O U.t$U
0: X X J x - ^ ^ ^o- o
- ^o .o .o T: ^ _ (ave.)
Py d (46) ^ 0.6 1
^=S
1 forV^<0.3
A r-A Q:
\
3
CD co
cr
H
CO
Q
20%
/ V/- '^^ 20%
V.^
\
W.-^ /
/
_L _L W _L _-l_ J_ w
0.6 0.7 ^1 0.8 0.9 1.0 0.6 0.7 0.8 0.9 1.0
<M 1^1 cvjl .
; CX)| CO CO! C30| OJ
dl dl dl *
WEIGHT (-^)
W =
WEIGHT (ASD)
40% Ln/Dn=3 /^\^
2
O Lr/Dn=0.2 / \ / \
for LRFD-non
3
CO / / ^ ^ forLRFD-Llnear
E 20%h
co \
Q _7 \
0.6 0.7
_L _w
0.8 0.9 1.0
col oo| cd >
dl dl dl ^
Fig. 29. Distribution of weight ratio