Escolar Documentos
Profissional Documentos
Cultura Documentos
Technical Manual
Version 5.5
1
UPDATE HISTORY
2
2013/07/01 STERA_3D Technical Manual Ver.3.3 is uploaded.
Missing part (2.10 Connection Panel) is supplemented. Appendix for Degrading
Tri-linear Slip Model and Modified Bi-linear Model are added.
2014/06/16 STERA_3D Technical Manual Ver.3.4 is uploaded.
Definition of effective slab width is added. Also, the Appendix for Degrading Tri-linear
Slip Model is modified. Operator splitting method is added for numerical integration
method.
2014/10/20 STERA_3D Technical Manual Ver.3.5 is uploaded.
Modified Bilinear Model for HDRB (High Damping Rubber Bearing) is added for
isolation devices.
2015/03/01 STERA_3D Technical Manual Ver.4.0 is uploaded.
Steel elements are added.
2015/03/23 STERA_3D Technical Manual Ver.4.1 is uploaded.
Column element with direct input is added.
2015/1/12 STERA_3D Technical Manual Ver.4.2 is uploaded.
K-brace (Chevron brace) is added.
2015/6/02 STERA_3D Technical Manual Ver.5.0 is uploaded.
SRC members are added for beam, column and wall.
For some isolation devices, strength reduction by dissipated energy is considered.
2015/7/10 STERA_3D Technical Manual Ver.5.1 is uploaded.
Hysteresis of Bouc-Wen Model is added for isolator and damper
2016/8/28 STERA_3D Technical Manual Ver.5.2 is uploaded.
Definition of External Spring is extended in three directions.
2016/10/23 STERA_3D Technical Manual Ver.5.3 is uploaded.
5.5 Eigen value problem, 7.4 Calculation of ground displacement are added.
2016/11/26 STERA_3D Technical Manual Ver.5.4 is uploaded.
Formulation of initial stiffness of nonlinear spring is fixed (Eqs. (3-1-34), (3-1-51))
2017/01/18 STERA_3D Technical Manual Ver.5.5 is uploaded.
“5.5 Modal analysis” is modified including participation factor, effective mass, etc.
3
INDEX
1. Basic Condition
1.1 Coordinate
2. Constitutive Equation of Elements
2.1 Beam
2.2 Column
2.3 Wall
2.4 Brace
2.5 External Spring
2.6 Base Isolation
2.7 Masonry Wall
2.8 Passive Damper
2.9 Floor Element
2.10 Connection Panel
3. Hysteresis Models of Nonlinear Springs
3.1 Beam
3.1.1 RC Beam
3.1.2 Steel Beam
3.1.3 SRC Beam
Appendix: Detail rule of degrading trilinear slip model
3.2 Column
3.2.1 RC Column
3.2.2 Steel Column
3.2.3 SRC Column
3.3 Wall
3.3.1 RC Wall
3.3.2 Steel Wall (Brace)
3.3.3 SRC Wall (Brace)
3.4 External Spring
3.5 Base Isolation
Appendix:
A-1. Hysteresis of LRB (Lead Rubber Bearing)
A-2. Hysteresis of HDRB (High Damping Rubber Bearing)
A-3. Hysteresis of Lead Damper
A-4. Hysteresis of Elastic Sliding Bearing
A-5. Hysteresis of Bouc-Wen Model
3.6 Masonry Wall
3.7 Passive Damper
4
4. Freedom Vector
4.1 Node freedom
4.2 Freedom vector
4.3 Dependent freedom
4.4 Formulation of independent displacement of the element
5. Equation of Motion
5.1 Mass matrix
5.2 Stiffness matrix
5.3 Input ground acceleration
5.4 Equation of motion
5.5 Modal analysis
5.6 Damping matrix
5.7 Numerical integration method
5.8 Energy
6. Nonlinear Static Push-Over Analysis
6. 1 Lateral distribution of earthquake force
6. 2 Capacity Curve
7. Others
7.1 Decomposition of shear and flexural deformation
7.2 P-D effect
7.3 Unbalance force correction
7.4 Calculation of ground displacement
5
1. Basic Condition
1.1 Coordinate
The global coordinate is defined as the left-hand coordinate as shown in Figure 1-1-1.
Z 3
6
X 1
4
5
Y
2
Z 3 Z Z
6
X 1 7 X X
4
5 8
Y Y Y
2
(a) lateral and rotational displacement (b) shear displacement
6
(2) Local Coordinate
The local coordinate is defined for each element. The displacement freedoms and force freedoms are named
with subscripts indicating the coordinate direction and node name. For example, the local coordinate of a
beam element in Figure 1-2 is defined to have its x-axis in the same direction of the element axis. Also the
displacement and force freedoms of a beam element are expressed as shown in Figure 1-1-2.
z
u zA u zB
yA yB x
A B
Displacement freedoms
y
Q zA Q zB
Local coordinate
M yA M yB
A B
Force freedoms
7
2. Constitutive Equation of Elements
2.1 Beam
' yA
x
M ' yA M ' yB
A B
Nx
' yB
elastic element
' yA yA
M ' yA M ' yB M ' yA M ' yB
A B A B
N 'x ' yB yB
nonlinear bending springs
l'
' yA ' yA yA yA yA
M ' yA M ' yB
' yB ' yB yB yB A yB B
' ' 0 0
x x
nonlinear shear springs
l' l'
3EI 0
6 EI y
' yA y
M ' yA
l' l'
' yB 0 M ' yB (2-1-1)
' 6 EI y 3EI y
N '
x
0 l' x
0
EA
where, E , I y , A and l ' are the modulus of elasticity, the moment of inertia of the cross-sectional area
along y-axis, the cross-sectional area and the length of the element. The rotational displacement vector of
the nonlinear bending springs is,
yA f yA 0 M ' yA
(2-1-3)
yB 0 f yB M ' yB
where, f yA and f yB are the flexural stiffness of nonlinear bending springs at both ends of the element.
The rotational displacement vector from the shear deformation of the nonlinear shear spring is,
8
1 1
yA k sz l ' k sz l ' M ' yA
1
1 M ' yB
(2-1-2)
yB
k sz l ' k sz l '
where, k sz is the shear stiffness of the nonlinear shear spring. Then, the displacement vector of the beam
element is obtained as the sum of the above three displacement vectors.
where,
l' 1 l' 1
f yA 3EI k l '
6 EI y k sz l '
0
y sz
l' 1
f B f yB 0
(2-1-5)
3EI y k sz l '
l'
sym.
EA
[ f B ] is the flexural stiffness matrix of the beam element. By taking the inverse matrix of [ f B ] , the
constitutive equation of the beam element is obtained as,
u
1 1 1 1 zA
yA l ' u zA A yA l ' u zB B yB l ' l ' 1 A B u
zB
(2-1-7)
1 1
yB u zA A yA u zB B yB
1
1
A
1 B yA
l' l' l' l' yB
9
yB
' yA yA
yB
yA ' yB
u zB
Z
u zB B l ' yB
u zA u zA A l ' yA
A B X
Al ' l' B l '
' yA ' yA
' yA 1 0 0 0
'
0 1 0 0 ' yB n ' yB
yB B (2-1-9)
' 0 0 1 1 xA xA
x xB
xB
u u zA
1 1 zA u
' yA 1 A B 0 0 u zB
l ' l ' zB
' 1 1
yB A 1 B 0 yA B yA
l' l' 0 (2-1-10)
xA yB yB
xB 0 0 0 0 1 0 xA
xA
0 0
0 0 0 1
xB xB
10
u
1 1 1 1 yA
zA l ' u yA A zA l ' u yB B zB l ' l'
1 A B u
yB
1 (2-9-7)
A 1 B zA
1 1 1
zB u yA A zA u yB B zB
l' l' l' l' zB
u yA zA
' zA u yB
zB
zB ' zB
zA
u zA
' yA 1 1 u zB
1 A B
' l ' l'
yB 1 1 yA
' zA A 1 B yB
l' l'
' zB 1
1
1 A B u yA
u
xA l'
1
l'
1 yB
xB A 1 B zA
l' l'
xA 1 zB
xB 1 xA
xB
u zA
u
zB u1
yA u
2
TixB (2-1-11)
yB
xA un
xB
The component of the transformation matrix, [TixB ] , is discussed in Chapter 4 (Freedom Vector).
11
From global node displacement to element face displacement
Transformation from the global node displacement to the element face displacement is,
u1 u1
' yA u u
2 2
'
yB n B B T
ixB T
xB (2-1-12)
'
x un un
z Z
y X
x Y
In case of Y-direction beam, the axial direction of the beam element coincides to the Y-axis in the global
coordinate, transformation of the sign of the vector components of the element coordinate is,
x 0 1 0 X
y 1 0 0 Y (2-1-13)
z
Y Beam 0 0 1 Z Global
Therefore
u zA 1 u zA u zA
u u
zB
1
0 u zB zB
yA 1 xA xA
s B (2-1-14)
yB 1 xB xB
xA 0 1 yA yA
xB Y Beam 1 yB Global yB Global
12
Transformation from the global node displacement to the element node displacement is,
u zA
u
zB u1
xA u
TiyB 2
(2-1-15)
xB
yA u n
yB
Transformation from the global node displacement to the element face displacement is,
u1 u1
' yA u u
' yB n B B s B TiyB
2
T yB
2
(2-1-16)
'
x u n u n
Constitutive equation
Finally, the constitutive equation of the X-beam is,
P1 u1 u1
P u u
2 2 2
TxB T
k B T
xB K
xB (2-1-17)
Pn un un
For Y-beam,
P1 u1 u1
P u u
2
TyB T
k B TyB 2 K yB 2
(2-1-18)
Pn un un
yA f yA 0 0 M ' yA f yA 0 0 ' yA ' yA
yB 0 f yB 0 M ' yB 0
f yB 0k B ' yB T pB
' yB (2-1-19)
1 1 1 1 ' '
y 0 N ' x 0 x x
k sz l ' k sz l ' k sz l ' k sz l '
where,
13
f yA 0 0
T
pB 0 f yB 0k B (2-1-20)
1 1
0
k sz l ' k sz l '
14
2.2 Column
Element model for column is defined as a line element with nonlinear bending springs at both ends and two
nonlinear shear springs in the middle of the element in x and y directions as shown in Figure 2-2-1.
M ' xB M ' xB
M ' yB
M ' yB
B B B
' xB
' yB
N 'z Z
l' = +
T 'z
' yA ' xA X
A A A
M ' yA Y
M ' yA
M ' xA M ' xA
15
Force-displacement relationship for nonlinear bending springs
Nonlinear interaction M x M y N z is considered in the nonlinear bending springs,
yA M ' yA
N ' zB , zB
xA f pA
M ' xA at end A (2-2-5)
N' M ' xB , xB
zA zA
M ' yB , yB
yB M ' yB
xB f pB M ' xB at end B
(2-2-6)
B
N'
zB zB
Rearrange the order of the components of the displacement vector and change the node axial displacements
into the relative axial displacement,
yA yA
yA 1 0 00 0 0
0 xA xA
yB 0 01 0 0
zA zA
xA 0 1 0 0 0 0 n p (2-2-8)
0 yB
0 0 0 1 0 yB
xB xB xB
z 0 0 1 0 0 1
zB zB
yA M ' yA M ' yA
M ' M '
yB f pA 0 yB yB
xA n p
T
n M ' f p M ' xA
f pB p xA
(2-2-9)
0 M' M '
xB
xB xB
z N ' z N ' z
16
Force-displacement relationship for nonlinear shear springs
The rotational displacement vector from the shear deformation of the nonlinear shear spring is,
1 1
yA k sx l ' k sx l ' M ' yA
1
1 M ' yB
in X-Z plane (2-2-10)
yB
k sx l ' k sx l '
1 1
xA k sy l ' k sy l ' M ' xA
1
1 M ' xB
in Y-Z plane (2-2-11)
xB
k sy l ' k sy l '
where, k sx and k sy are the shear stiffness of the nonlinear shear springs.
The displacement vector of the column element is obtained as the sum of the displacement vectors of
elastic element, nonlinear shear springs and nonlinear bending springs,
l' l'
3EI
6 EI y
y
l'
0
3EI y
l' l'
f C 3EI x 6 EI x
l'
3EI x
l'
EA
l'
sym.
GI z elestic element
17
f p11 f p12 f p13 f p14 f p15 0
f p 22 f p 23 f p 24 f p 25 0
f p 33 f p 34 f p 35 0
f p 44 f p 45 0
f p 55 0
sym. 0 bending spring
1 1
k l' k l'
sx sx
1
0
k sx l '
1 1
(2-2-13)
k sy l ' k sy l '
1
k sy l '
0
sym. 0 shear spring
By taking the inverse matrix of [ f C ] , the constitutive equation of the column element is obtained as,
' yA ' yA
'
' yA 1 ' yB yB
' 1 0 ' ' xA
yB xA
' xA 1 ' xB ' xB
nC (2-2-15)
' xB 1 zA zA
'z 0 1 1 zB zB
'z 1 1 zA zA
zB zB
18
u xA u xA
1 1
u xB u
l ' 1 A B xB
l' yA
' yA yA
' 1 1
A 1 B
0 yB
yB l ' l' yB
' xA 1
1
1 A B u yA u yA
u
' xB l' l'
C yB
yB u
1 1
zA A 1 B xA xA
l' l'
zB 1 xB
xB
zA
1 zA zA
zB 0 1 zB zB
1 zA zA
zB zB
(2-2-16)
B u xB
u yB
Bl'
yB
xB ' yB
l'
'xB
' yA Z
'xA
yA
Al ' X
xA
A u xA
u yA
Y
19
From global node displacement to element node displacement
Transformation from global node displacement to element node displacement is;
u xA
u
xB
yA
yB
u yA u1
u
u yB 2
T
iC (2-2-17)
xA
xB un
zA
zB
zA
zB
The component of the transformation matrix, [TiC ] , is discussed in Chapter 4 (Freedom Vector).
' yA
'
yB u1 u1
' xA u
u 2 2
n T
iC T
C (2-2-18)
'
xB
C C
'z u n u n
' z
Constitutive equation
Finally, the constitutive equation of the column is;
P1 u1
P u
2 2
K C (2-2-19)
Pn un
where,
20
Transformation matrix for nonlinear spring displacement
The nonlinear spring displacement vector is obtained from Equations (2-2-7), (2-2-10) and (2-2-11),
yA
f
pA 0
xA M ' yA M ' yA
zA
M ' xA M '
xA
yB
0 f N ' zA f N ' zA
pC
M ' yB
pA (2-2-21)
xB M ' yB
zB 1 1 M ' M ' xB
0 0 0 0 xB
N'
y k sx l ' k sx l '
zB N ' zB
x 0 1 1
0 0 0
k sy l ' k sy l '
Furthermore, in the same way as Equation (2-2-8),
Therefore, the nonlinear spring displacement vector is obtained from the element face displacement as,
yA
xA M ' yA ' yA ' yA
zA M ' ' '
yB yB yB
yB ' xA ' xA
' xA
f pC n' p
M
f pC n' p k C T pC (2-2-23)
xB M ' xB ' xB ' xB
zB N 'z 'z 'z
y T ' z ' z ' z
x
21
2.3 Wall
Element model for wall is defined as a line element with nonlinear bending springs at both ends and three
nonlinear shear springs; one is in the middle of the wall panel and others are in the side columns as shown
in Figure 2-3-1.
' xB ' xB Z
' yB
l' X
' xA ' yA ' xA
Y
l' l'
' yAc 3EI c 6 EI c M ' yAc
l ' M ' yBc
in wall panel (2-3-1)
' yBc l '
6 EI c 3EI c
l' l'
' xA1 3EI 1 6 EI 1 M ' xA1
l ' M ' xB1
in side column 1 (2-3-2)
' xB1 l '
6 EI 1 3EI 1
l' l'
' xA2 3EI 2 6 EI 2 M ' xA2
l ' M ' xB 2
in side column 2 (2-3-3)
' xB 2 l '
6 EI 2 3EI 2
The axial displacement is,
l'
' ' zc N ' zc (2-3-4)
EA
22
Force-displacement relationship for nonlinear bending springs
Nonlinear interaction M x M y N z is considered in the nonlinear bending springs,
A
M ' xA1 , xA1 M ' xA2 , xA2
M ' yAc , yAc
N zAc , zAc
where, [ f pA ] and [ f pB ] are the flexural stiffness matrices of the nonlinear bending springs. Therefore,
the force-displacement relationship of nonlinear bending springs is,
23
yAc M ' yAc
M '
xA1 xA1
xA2 M ' xA2
zAc f pA
0 N ' zAc
yBc 0
f pB M ' yBc
(2-3-7)
Rearrange the order of the components of the displacement vector and change the node axial displacements
into the relative axial displacement,
yAc yAc
yAc 1
xA1 xA1
yBc 1 xA2
xA1 1 xA2
zAc zAc
xB1 1 n p (2-3-8)
1 yBc yBc
xA2
xB1 xB1
xB 2 1
xB 2 xB 2
zc 1 1
zBc zBc
1 1
yAc k scl ' kscl ' M ' yAc
1 M ' yBc
in wall panel (2-3-10)
yBc 1
k scl ' kscl '
1 1
xA1 k s1l ' k s1l ' M 'xA1
1 M 'xB1
in side column 1 (2-3-11)
xB1 1
k s1l ' k s1l '
24
1 1
xA2 k s 2l ' k s 2l ' M ' xA2
1 M ' xB 2
in side column 2 (2-3-12)
xB 2 1
k s 2l ' k s 2l '
where, k sc , k s1 and k s 2 are the shear stiffness of the nonlinear shear springs.
The displacement vector of the column element is obtained as the sum of the displacement vectors of
elastic element, nonlinear shear springs and nonlinear bending springs,
l' l'
3EI
6 EI c
c
l'
3EI c
l' l'
3EI 1 6 EI 1
l'
fW
3EI 1
l' l'
3EI 2 6 EI 2
l'
sym.
3EI 2
l'
EAc elestic element
f p11 f p17
f p 71 f p 77
bending spring
25
1 1
k l' k scl '
sc
1
k scl '
1 1
k s1l ' k s1l '
1 (2-3-14)
k s1l '
1 1
k s 2l ' k s 2l '
1
sym.
k s 2l '
0 shear spring
By taking the inverse matrix of [ fW ] , the constitutive equation of the column element is obtained as,
26
u xAc u xAc
u u
1 1 xBc xBc
l ' 1 A B yAc yAc
l'
' yAc 1 1
A 1 B yBc yBc
' l ' l' u yA1 u yA1
yBc 1 1
' xA1 1 A B u yB1 u yB1
l' l'
' xB1 1 1
A 1 B xA1 W xA1
' xA 2 l' l'
1 1
xB1 xB1
' xB 2 1 A B u u yA2
l' l' yA2
' zAc 1 1 u yB 2 u yB 2
' A 1 B
zBc l' l' xA2 xA2
1
xB 2 xB 2
1
zAc zAc
zBc zBc
(2-3-17)
From global node displacement to element node displacement
Transformation from the center displacements to the node displacements is,
yc
z 2 z1
yc
z1 zc z2 w
z1 z 2
zc
2
w
1
u xAc u xA1 u xA1
u 1
xBc 1 1 zA1 zA1
yAc w w
zA2 zA 2
yBc
1 1 u
xB1 u xB1
u yA1 w w zB1 zB1
1
u yB1 zB 2 zB 2
1
u yA1 u
DW
xA1 yA1
1
xB1 u
yB1 u
yB1
1
u yA2 xA1 xA1
1
u yB 2 xB1 xB1
1
u u
xA2 1 yA2 yA2
xB 2 u u yB 2
1 yB 2
zAc 0.5 0.5 xA2 xA2
zBc 0.5 0.5
xB 2 xB 2
(2-3-18)
27
Transformation from the global node displacements to the element node displacements is;
u xA1
zA1
zA2
u xB1
zB1
zB 2 u1
u u
2
yA1
T
ixW (2-3-19)
u yB1
xA1 u n
xB1
u
yA2
u yB 2
xA2
xB 2
The component of the transformation matrix, [TixW ] , is discussed in Chapter 4 (Freedom Vector).
' yAc
'
yBc u1 u1
' xA1 u u
2 2
' xB1 nW W DW TixW TxW (2-3-20)
'
xA 2
u n u n
' xB 2
' zc
z Z
y X
x Y
28
In case of Y-direction wall, the wall panel direction coincides to the Y-axis in the global coordinate,
transformation of the sign of the vector components of the element coordinate is,
x 0 1 0 X
y 1 0 0 Y (2-3-21)
z
Y Wall 0 0 1 Z Global
Therefore
u xA1 1 u yA1 u yA1
1
zA1 zA1 zA1
zA2 1 zA2 zA2
u xB1 1 u yB1 u yB1
zB1 1 zB1 zB1
zB 2 1 zB 2 zB 2
u 1 u xA1 u
xA1
W
yA1
u yB1 1 xB1
u u xB1
xA1 1 yA1
yA1
xB1 1 yB1 yB1
u u u xA2
1 xA2
yA 2
u yB 2 1 u xB 2 u xB 2
1 yA2
xA2 yA2
xB 2 Y Wall
1 yB 2 Global yB 2 Global
(2-3-22)
Transformation from the global node displacement to the element node displacement is;
u yA1
zA1
zA2
u yB1
zB1
zB 2 u1
u u
xA1
TiyW
2
(2-3-23)
u xB1
yA1 u n
yB1
u xA2
u xB 2
yA2
yB 2
29
Transformation from the global node displacement to the element face displacement is,
' yAc
'
yBc u1 u1
' xA1 u u
'
xB1
nW W DW W T 2
ixW
T yW
2
(2-3-24)
'
xA2 u n u n
' xB 2
' zc
Constitutive equation
Finally, the constitutive equation of the wall is;
P1 u1
P u
2 2
K xW (2-3-25)
Pn u n
where,
For Y-wall,
P1 u1
P u
2
K yW
2
(2-3-27)
Pn u n
where,
K T k T
yW yW
T
W yW (2-3-28)
30
yAc
f pA 0
xA1 M ' yAc M ' yAc
xA2 M '
M ' xA1 xA1
zAc M ' xA2 M ' xA2
yBc 0 f pA
N ' zAc f
xB1 M ' yBc
pW
N ' zAc
M ' yBc
xB 2 1 1 M ' xB1 M ' xB1
zBc M '
k sc l ' k sc l ' xB 2 M ' xB 2
yc 1 1 N ' zBc N'
x1 zBc
k s1l ' k s1l '
x 2 1 1
k s 2l ' k s 2l '
(2-3-29)
Furthermore, in the same way as Equation (2-3-8),
M ' yAc
M ' M ' yAc
xA1 M '
M ' xA2 yBc
M ' xA1
N ' zAc
M '
np
T
M ' xB1 (2-3-30)
yBc M '
M ' xB1 xA2
M ' xB 2
M ' xB 2 N'
N' zc
zBc
Therefore, the nonlinear spring displacement vector is obtained from the element face displacement as,
yAc
xA1
xA2 M ' yAc ' yAc ' yAc
M ' ' '
zAc yBc yBc yBc
yBc M ' xA1 ' xA1 ' xA1
xB1 f pW n p
T
M ' xB1 f pW n p
T
kW ' xB1 T pW
' xB1
(2-3-31)
M ' ' '
xB 2 xA2 xA2 xA2
zBc M ' xB 2 ' xB 2 ' xB 2
yc N ' zc ' zc ' zc
x1
x 2
31
2.4 Brace
Element model for Brace is defined as a truss element with a nonlinear axial spring and pin-supported at
both ends as shown in Figure 2-6-1.
3 B 4
N', '
Z
h
1 2 X
A
w Y
Force-displacement relationship
~ ~ ~ ~
f x4 , u x4 f x3 , u x3
4 3
N1 , 1
x N2, 2
y x
~ ~ y ~ ~
f x1 , u x1 f x2 , u x2
1 2
(Brace 1) (Brace 2)
The relationship between axial deformation and axial force of the truss element is,
N1 k1 1 (2-4-1)
N 2 k 2 2 (2-4-2)
Replacing with the nodal force and displacement in local coordinate along the element,
~ ~
N 1 f 1x f 4 x , 1 u~4 x u~1x (2-4-3)
~ ~
N 2 f 2 x f 3 x , 1 u~3 x u~2 x (2-4-4)
32
In a matrix form,
u~1x
u~
1y
~
u1x u~2 x
~
1 1 0 0 1 u~2 x 1 0 0 0 0 0 1 0 u 2 y
u~ 0 0 1 0 1 0 0 0 u~ nb u
~
(2-4-5)
2 0 1 1 0 3x 3x
u~4 x u~3 y
~
u 4 x
u~4 y
~
f x1 1 0
~
f z1 0 0
~ ~
f 1x 1 0 f x 2 0 1
~ ~
f 2 x 0 1 N 1
~ f ~
~ f z2 0 0 N1
1 N 2
T N
nb 1 (2-4-6)
f 3x 0 1 N 2 f x3 0
N 2
~
f 1 ~
f 0
4x 0 0
~z 3
f x4 1 0
~
f 0 0
z4
From Figure 2-4-3, the relation between the nodal forces in local coordinate and those of global coordinate
is,
~
f x1 f x1 cos f z1 sin
~ for Brace 1 (2-4-7)
f y1 f x1 sin f z1 cos
and
~
f x 2 f x 2 cos f z 2 sin
~ for Brace 2 (2-4-8)
f y 2 f x 2 sin f z 2 cos
Eq. (2-4-8) can be also obtained from the Eq. (2-4-7) by replacing by and using the
formulas sin sin , cos cos .
33
f z4 ~ ~ f z3
f x4 f x3
4 3
f x4 f x3
h
l w2 h 2
f z1 ~ f z2
~ f 2x
f x1
1 2
f x2
f x1
f z1 f z2
f z1 cos f z1 sin
f x1 cos f z 2 sin
f z 2 cos
f x 2 cos
f x1
f x2
f x1 sin f x 2 sin
(Brace 1) (Brace 2)
In a matrix form,
~
f x1 c s f x1 f x1
~ s c f f
f z1 z1 z1
~
f f x2 f x2
c s
~x 2
f z2 s c f z2 f
~ Cb z 2 (2-4-9)
c s f x3
f x3 f x3
~
f s c f z3 f z3
~z 3
f x4 c s f x4 f x4
~
f s c f z 4 Global f
z 4 Local z4
where
w h
c cos , s sin
l l
Since C b C b I ,
T
Cb is an orthogonal matrix, therefore,
34
In a similar manner, from Figure 2-4-4, the relation between the nodal displacements in local coordinate
and those of global coordinate can be obtained as,
u x1 u~x1 cos u~z1 sin
for Brace 1 (2-4-11)
u u~ sin u~ cos
z1 x1 z1
and
u x 2 u~x 2 cos u~z 2 sin
for Brace 2 (2-4-12)
u u~ sin u~ cos
z2 x2 z2
Eq. (2-4-12) can be also obtained from the Eq. (2-4-11) by replacing by and using the
formulas sin sin , cos cos .
u~3
u z4 u z3
u~4
4 3
u x4 u x3
h
l w2 h 2
u~2 uz2
u z1 u~1
1 u x2
u x1 2
w
u z 2 u~x 2 sin u~z 2 cos
u~x 2
u z1 u~x1 sin u~z1 cos
In a matrix form,
u x1 c s u~x1 u~x1
u s c u~ u~
z1 z1 z1
u x 2 c s u~x 2 u~x 2
~ ~
u z 2 s c u z 2 T u z 2
Cb ~
u u~x 3
x 3
c s
~ u x 3
u z 3 s c
u z3 u~z 3
~
u x 4 c s u~x 4 u x 4
~
u u~
z 4 Local s c u z 4 Global z4
35
The stiffness matrix of brace element is,
N k~
N 1 k1 0 1
or (2-4-13)
N 2 0 k 2 2
Where
u1
u
u TixBr 2 (2-4-18)
u n
The component of the transformation matrix, [TixBr ] , is discussed in Chapter 4 (Freedom Vector).
u1 u1
u u
nb Cb u nb Cb TixBr TxBr 2
2
(2-4-19)
u n u n
Constitutive equation
Finally, the constitutive equation of the brace is;
P1 u1
P u
2 2
K xBr (2-4-20)
Pn u n
where,
36
In case of Y-direction brace
z Z
y X
x Y
In case of Y-direction brace, transformation of the sign of the vector components of the element coordinate
is,
x 0 1 0 X
y 1 0 0 Y (2-4-20)
z
Y Beam 0 0 1 Z Global
Therefore
u x1 1 u y1 u y1
u 1 u u
z1 z1 z1
u x 2 1 u y2
u y 2
u z 2 1 u z 2 u
z2 (2-4-21)
1 u y3
u x 3 u y 3
u z 3 1 u z 3 u z 3
u x 4
1 u y 4 u y 4
u
z 4 Y Brace 1 u z 4 Global u z 4 Global
Transformation from the global node displacement to the element node displacement is;
u1
u
u TiyBr 2
(2-4-22)
u n
Transformation from the global node displacement to the element face displacement is,
u1
u
T yBr
2
, T n C T
yBr b b iyBr (2-4-23)
u n
37
Constitutive equation
The constitutive equation of the Y-direction brace is;
P1 u1
P u
2
K yBr
2
(2-4-21)
Pn u n
where,
K T k T
yBr yBr
T
Br yBr (2-4-22)
38
In case of K-brace (or Cheveron brace)
3 6 4
1 5 2
3 6 6 4
l'
2
w
l' h2
2
1 5 5 2
For the left half part, as we defined before for the ordinary brace, the stiffness equation of brace element is,
where
f L f x1 f z1 f x5 f z5 f x3 f z3 f x6 f z6
T
uL u x1 u z1 u x 5 u z5 u x3 u z 3 u x6 uz6
T
k~ k1, L
0
k 2,L
, nb
1 0 0 0 0 1 0
0
0 1 0 1 0 0 0
0
L
0
c s
s c
c s
Cb
s c , c cos
w / 2 , s sin
h
c s l' l'
s c
c s
s c
39
For the right half part, in the same way, the stiffness equation of brace element is,
where
f R f x 5 f z5 f x2 f z2 f x6 f z6 f x4 f z4
T
u x1 u x1
u u
z1 z1
u x1 u x 2 u x 2
u
z1 u z 2 u z 2
u x 5 [1]
u x3 u x 3
u
u [1] z 3 D u z 3
uL z5
u x 3 u x 4 L
[1] u x 4
u z 3
[1] u z 4 u z 4
u x 6 u x 5 u x 5
u u u
z6 z5 z5
u x 6 u x 6
u u
z6 z6
u x1 u x1
u u
z1 z1
u x 5 u x 2 u x 2
u
z5 u z 2 u z 2
u x 2 [1]
u x3 u x 3
u
u [1] z 3 D u z 3
uR z2
u x 6 [1] u x 4
R
u x 4
u z 6 [1]
uz4 u z 4
u x 4 u x 5 u x 5
u u u
z4 z5 z5
u x 6 u x 6
u u
z6 z6
40
We assume the displacements of intermediate nodes, 5 and 6, are calculated from those of end nodes as
follows,
u z 2, 4
u z 5, 6
1 1
u x 5 u x1 u x 2 , u z 5 u z1 u z 2 u z1,3 u x 2, 4
2 2 2,4
1 1 u x 5, 6
u x 6 u x 3 u x 4 , u z 6 u z 3 u z 4 u x1,3
5,6
2 2
1,3
In a matrix form
u x1 u x1
u u
z1 z1
u x 5 1 / 2 0 1 / 2 0 0 0 0 0 u x 2 u x 2
u 0 1/ 2 0 1/ 2 0
z5 0 0 0 u z 2 u
hCh z 2
u x 6 0 0 0 0 1 / 2 0 1 / 2 0 u x 3 u x 3
u z 6 u z 3
0 0 0 0 0 1 / 2 0 1 / 2 u z 3
Local
u x 4 u x 4
u u
z 4 Local z 4 Local
Therefore,
u x1
u
z1
u x 2 u x1 u x1
u
u z 2 u z1 z1
u x 3 u x 2 u x 2
u z 3 I u z 2 u
TCh z 2
u x 4 hCh u x 3 u x 3
u z 4 u z 3 u z 3
u x 5 u x 4 u x 4
u u u
z5 z 4 Local z 4 Local
u x 6
u
z6
41
Therefore,
u x1 u x1
u u
z1 z1
u x1 u x 2 u x1 u x 5 u x 2 u x1
u u u u
z1 u z 2 z1 z5 u z 2 z1
u x 5 u x 3 u x 2 u x 2 u x 3 u x 2
u u u u u u
uL z 5 DL z 3 DL TCh z 2 , uR z 2 DR z 3 DR TCh z 2
u
x 3 u
x 4 u
x 3 u x 6 u x 4 u x 3
u z 3 u z 4 u z 3 u z 6 u z 4 u z 3
u x 6 u x 5 u x 4 u x 4 u x 5 u x 4
u u u u u u
z6 z5 z 4 Local z4 z5 z 4 Local
u x 6 u x 6
u u
z6 z6
f x1 u x1
f u
z1 z1
f x2 u x 2
f z2
T T T T u
TCh DL k L DL TCh TCh DR k R DR TCh z 2
f x3 u x 3
f z3 u z 3
f x4 u x 4
f u
z 4 Local z4
42
2.5 External Spring
zB
yA Z
B
A B A
xA xB X
B
Y
yB A
zA
' x u xB u xA
' y u yB u yA (2-5-2)
' z zB zA
Therefore
u xA u xA
u u
xB xB
u u
' x 1 1 0 0 0 0 yA n xE yA (2-5-3)
u yB u yB
zA zA
zB zB
u xA u xA
u u
xB xB
u yA
' y 0 0 1 1 0 0uu yA n yE (2-5-4)
yB u yB
zA zA
zB zB
43
u xA u xA
u u
xB xB
u yA u
' z 0 0 0 0 1 1 n zE yA (2-5-5)
u yB u yB
zA zA
zB zB
u xA
u
xB u1
u yA u
2
TiE (2-5-6)
u yB
zA u n
zB
The component of the transformation matrix, [TiE ] , is discussed in Chapter 4 (Freedom Vector).
u1 u1
u u
'i niE TiE TE 2 , i x, y, z
2
(2-5-7)
u n u n
Constitutive equation
The constitutive equation of the external spring is;
P1 u1
P u
2 2
K E (2-5-3)
Pn u n
where,
44
2.6 Base Isolation
zB
B Z
u xB
u yB
l X
A u xB
Y
u yA
zA
Figure 2-6-1 Element model for base isolation
Q' x
k pBI ''
x
(2-6-1)
Q' y y
Including the axial stiffness,
Q ' x k pBI
0 'x 'x
Q ' y EA ' y k BI ' y (2-6-2)
' 0 l'
' z '
z z
' x u xB u xA
' y u yB u yA (2-6-3)
' z zB zA
Therefore
u xA u xA
u u
' x 1 1
xB xB
u yA n u yA
' y 1 1 u BI (2-6-4)
' yB u yB
z 1 1
zA zA
zB zB
45
From global node displacement to element node displacement
Transformation from the global node displacement to the element node displacement is,
u xA
u
xB u1
u yA u
2
TiBI (2-6-5)
u yB
zA u n
zB
The component of the transformation matrix, [TiBI ] , is discussed in Chapter 4 (Freedom Vector).
u1 u1
' x u u
2 2
' y n BI TiBI TBI (2-6-6)
'
z u n u n
Constitutive equation
The constitutive equation of the Base isolation is;
P1 u1
P u
2 2
K BI (2-6-7)
Pn u n
where,
46
2.7 Masonry Wall
Element model for Masonry wall is defined as a line element with a nonlinear shear spring and a vertical
spring in the middle of the wall panel as shown in Figure 2-6-1.
B1 B B2
A1 A2
A
Force-displacement relationship
The relationship between the shear deformation and shear force of the nonlinear shear spring is,
47
u z 1 u xB1 u xA1 u xB 2 u xA2
(2-7-6)
z 2 l l
' xc l
l
zA2 zA1 zB 2 zB1 1 u xB1 u xA1 u xB 2 u xA2 (2-7-7)
2w 2
The axial deformation, ' z1 , ' z 2 , is,
' z1 zB1 zA1 , 'z 2 zB 2 zA2 (2-7-8)
In a matrix form,
u xA1 u xA1
zA1 zA1
l l l l u xA2 u xA2
' xc 0.5 0.5 w 0.5 0.5 0.5 0.5 0.5 0.5
w
w w
' z1 0 1 0 0 0 1 0 0 zA 2 DN zA2
u
' 0 0 0 1 0 0 0 1 xB1 u xB1
z2 zB1
zB1
u xB 2 u xB 2
zB 2 zB 2
(2-7-9)
From global node displacement to element node displacement
Transformation from the global node displacement to the element node displacement is;
u xA1
zA1
u xA2 u1
u
zA 2 2
T
ixN (2-7-10)
u xB1
zB1 u n
u xB 2
zB 2
The component of the transformation matrix, [TixN ] , is discussed in Chapter 4 (Freedom Vector).
48
u1 u1
' xc u u
2 2
'
z1 D N T
ixN T
xN (2-7-11)
'
z2 u n u n
z Z
y X
x Y
In case of Y-direction wall, the wall panel direction coincides to the Y-axis in the global coordinate,
transformation of the sign of the vector components of the element coordinate is,
x 0 1 0 X
y 1 0 0 Y (2-7-12)
z
Y Beam 0 0 1 Z Global
Therefore
Transformation from the global node displacement to the element node displacement is;
49
u yA1
zA1
u yA2 u1
u
zA2
TiyN
2
(2-7-14)
u yB1
zB1 u n
u yB 2
zB 2
Transformation from the global node displacement to the element face displacement is,
u1 u1
' xc u u
' z1 D N TiyN 2
T yN
2
(2-7-15)
'
z2 u n u n
Constitutive equation
Finally, the constitutive equation of the wall is;
P1 u1
P u
2 2
K xN (2-7-16)
Pn u n
where,
For Y-wall,
P1 u1
P u
2
K yN
2
(2-7-18)
Pn u n
where,
K T k T
yN yN
T
N yN (2-7-19)
50
2.8 Passive Damper
Element model for passive damper with a shear spring is defined as a line element with a nonlinear shear
spring as shown in Figure 2-8-1.
B1 B B2
l
Q' xc , ' xc
A1 A2
A
Force-displacement relationship
The relationship between the shear deformation and shear force of the nonlinear shear spring is,
' xc l
l
zA2 zA1 zB 2 zB1 1 u xB1 u xA1 u xB 2 u xA2 (2-8-5)
2w 2
51
The axial deformation, ' z1 , ' z 2 , is,
' z1 zB1 zA1 , 'z 2 zB 2 zA2 (2-8-6)
In a matrix form,
u xA1 u xA1
zA1 zA1
l l l l u u xA2
' xc 0.5 0.5 w 0.5 0.5 0.5 0.5 0.5 0.5 xA2
w zA2
w w zA2
' z1 0 1 0 0 0 1 0 0 DD
u xB1
u xB1
' 0 0 0 1 0 0 0 1
z2
zB1 zB1
u xB 2 u xB 2
zB 2 zB 2
(2-8-7)
From global node displacement to element node displacement
Transformation from the global node displacement to the element node displacement is;
u xA1
zA1
u xA2 u1
u
zA 2 2
TixD (2-8-8)
u xB1
zB1 un
u xB 2
zB 2
The component of the transformation matrix, [TixD ] , is discussed in Chapter 4 (Freedom Vector).
u1 u1
' xc u u
2 2
'
z1 DD T
ixD T
xD (2-8-9)
'
z2 un un
52
In case of Y-direction damper
z Z
y X
x Y
In case of Y-direction damper, the damper direction coincides to the Y-axis in the global coordinate,
transformation of the sign of the vector components of the element coordinate is,
x 0 1 0 X
y 1 0 0 Y (2-8-10)
z
Y Beam 0 0 1 Z Global
Therefore
Transformation from the global node displacement to the element node displacement is;
u yA1
zA1
u yA2 u1
u
zA 2
TiyD
2
(2-8-12)
u yB1
zB1 un
u yB 2
zB 2
53
Transformation from the global node displacement to the element face displacement is,
u1 u1
' xc u u
' z1 DD TiyD 2
TyD
2
(2-8-13)
'
z2 un un
Constitutive equation
Finally, the constitutive equation of the damper is;
P1 u1
P u
2 2
K
xD (2-8-14)
Pn un
where,
For Y-damper,
P1 u1
P u
2
K yD 2
(2-8-16)
Pn un
where,
K T k T
yD yD
T
D yD (2-8-17)
54
Appendix : Calculation of shear component
For “Masonry Wall” and “Passive Damper”, the shear deformation is defined as follows:
Shear strain is τ = Δl / l ≈θ
u x u y
If we discuss small element Eq. (2-7-4) and Eq. (2-8-2)
y x
55
This definition is necessary to remove rotational component. To explain this, suppose there is only
rotational (or bending) deformation,
-θ
For example, in the upper story of the building under horizontal deformation, the bending
component is dominant and the shear component is small. Therefore, the brace damper
doesn’t work in the upper story.
56
3) In case of damper element
θ1 θ2
θ'2
θ'1
57
2.9 Floor Element
In the default setting, STERA 3D adopts “rigid floor”. However, elastic deformation of a floor diaphragm
in-plane can be considered by the option menu selecting “flexible floor”. The stiffness matrix of the floor
element is constructed using a two dimensional isoparametric element.
P1 u1
Q v
1 1
P2 u 2
Q2 v 2
K F
P3 u 3
Q3 v3
P4 u 4
Q v
4 4
F =K u (2-9-1)
The coordinate transfer function {x, y} is expressed using the interpolation functions as follows:
4
1 1 1 1
x (r , s ) hi (r , s ) xi (1 r )(1 s ) x1 (1 r )(1 s ) x 2 (1 r )(1 s ) x3 (1 r )(1 s ) x 4
i 1 4 4 4 4
4
1 1 1 1
y (r , s ) hi (r , s ) y i (1 r )(1 s ) y1 (1 r )(1 s ) y 2 (1 r )(1 s ) y 3 (1 r )(1 s ) y 4
i 1 4 4 4 4
(2-9-2)
58
The deformation function {u, v} is also expressed using the same interpolation functions.
4
1 1 1 1
u (r , s ) hi ( r , s )u i (1 r )(1 s )u1 (1 r )(1 s )u 2 (1 r )(1 s )u 3 (1 r )(1 s )u 4
i 1 4 4 4 4
4
1 1 1 1
v ( r , s ) hi ( r , s ) v i (1 r )(1 s )v1 (1 r )(1 s )v 2 (1 r )(1 s )v 3 (1 r )(1 s )v 4
i 1 4 4 4 4
(2-9-3)
Stiffness matrix can be obtained from the “Principle of Virtual Work Method,” which is expressed in the
following form:
dv u T F
T
(2-9-4)
V
where, is a virtual strain vector, is a stress vector, u is a virtual displacement vector and F is a
load vector, respectively.
u 4
hi
i 1 x
ui
x x
v 4
hi
y vi
y i 1 y
xy u v 4 hi 4
hi
y x y u i x vi
i 1 i 1
u1
h1 h2 h3 h4 v1
0 0 0 0 u
x x x x 2
h1 h2 h3 h4 v 2
0 0 0 0 u
y y y y 3
h1 h1 h2 h2 h3 h3 h4 h4 v3
y x y x y x y x
u4
v
4
ε= B u (2-9-6)
59
In the plane stress problem, the stress-strain relationship is expressed as,
x
1 0 x
y
E 1 0 (2-9-7)
1 1 y
xy 0 0 xy
2
σ= C ε
Substituting equation (2-9-6) into equation (2-9-7),
σ= C B u (2-9-8)
Bu CBu dv u B T CBdxdy u u T F
T
T
(2-9-9)
V V ( x, y )
F Ku , K B T CBdv (2-9-10)
V
If we assume the constant thickness of the plate (= t), using the relation dv tdxdy ,
K t B
T
CBdxdy (2-9-11)
V ( x, y )
Since this integration is defined in x-y coordinate, we must transfer the coordinate into r-s coordinate to use
the numerical integration method. Introducing the Jacobian matrix,
x y
J r r ; Jacobian Matrix (2-9-12)
x y
s s
the above integration is expressed in r-s coordinate as,
1 1
( x, y )
K t B xr , s , y r , s CB xr , s , y r , s
T
drds (2-9-13)
1 1
(r , s )
where
x y
( x, y )
det J r r (2-9-14)
(r , s) x y
s s
60
Evaluation of Jacobian Matrix
y hi x hi
4 4
x
r i 1 r
i r yi
J r i 1 (2-9-15)
x y 4 hi hi
4
s
xi yi
s i 1 s i 1 s
h1 1 h1 1
(1 s ) (1 r )
r 4 s 4
h2 1 h2 1
(1 s ) (1 r )
r 4 s 4
, (2-9-18)
h3 1 h3 1
(1 s ) (1 r )
r 4 s 4
h4 1 h4 1
(1 s ) (1 s )
r 4 s 4
61
The 3 points Gauss Integration Formula is defined as:
f(0.7746)
f(t)
f(-0.7746)
f(0)
t
-1 -0.7746 0 +0.7746 +1
1 1
( x, y )
K t Bxr , s , y r , s CBxr , s , y r , s
T
drds
1 1
(r , s )
1 1
t F (r , s )drds (2-9-20)
1 1
3 3
t i j F (ri , s j )
i 1 j 1
where
( x, y )
F (r , s ) Bxr , s , y r , s CB xr , s , y r , s
T
(r , s)
62
From global node displacement to element node displacement
Transformation from global node displacements to element node displacements is,
u1
v
1
u 2 u1
u
v 2 2
T
iF (2-9-21)
u 3
v3 u n
u 4
v
4
The component of the transformation matrix, [TiF ] , is discussed in Chapter 4 (Freedom Vector).
63
2.10 Connection Panel
1) General case
In the default setting, STERA3D assumes the rigid connection zone between column and beam. You can
consider shear deformation of the connection area (we call “connection panel”) by the “Connection
member” menu.
C C
h
0.5γA 0.5γA
w w
u B 0 u C 0.5 A h
Node B: v B 0.5 A w , Node C: vC 0 (2-10-1)
0.5 0.5
B A C A
64
First we consider nodal movement without shear deformation of the connection panel. As shown in Figure
2-10-3, the displacement at node B and node C will be;
u B u A u C u A A h
Node B: v B v A A w , Node C: vC v A (2-10-2)
B A C A
Then, we consider shear deformation of the connection as shown in Figure 2-10-4. By adding Equation
(2-10-1) to (2-10-2), the displacement at node B and node C will be;
Ah
h
Aw
A
C w
vA
uA
A B
A B
65
Node B:
u
u B u A 0 uA 1 0 0 0 A
v A
v B v A A w 0.5 A w v A A w 0.5 A w 0 1 w 0.5w
0.5 0.5
B A
0 0 1
0.5 A
A A A
A
(2-10-3)
Node C:
u A
u C u A A h 0.5 A h u A A h 0.5 A h 1 0 h 0.5h
v
vC v A 0 vA 0 1 0 0 A
0.5 0.5 0 0 1 0.5 A
C A A A A
A
(2-10-4)
2) Beam element
In case of rigid connection, as described in Equation (2-1-7), the nodal displacement is expressed as,
u
1 1 1 1 zA
u u l' 1 A B u
yA
l'
zA A yA
l'
zB B yB
l' zB
1 (2-10-5)
1 1
yB u zA A yA u zB B yB
1
A
1 B yA
l' l' l' l' yB
yB
' yA yA
yB
yA ' yB
u zB
Z
u zB B l ' yB
u zA u zA A l ' yA
A B X
Al ' l' B l '
66
If we consider shear deformation of connection panel, from Figure 2-10-6,
u zA
u
1 1 zB
l' 1 A B 0.5 0.5 A 0.5 B
l ' yA
(2-10-6)
1 1
A 1 B 0.5 A 0.5 0.5B yB
l' l' yA
yB
yB
' yA yA 0.5 yA
yB 0.5 yB ' yB
yA u zB
u zB B l ' yB 0.5 yB
u zA u zA Al ' yA 0.5 yA
A B
Z
67
The transformation matrices for beam element are;
Including connection panel and node movement
u zA u zA
u u
1 1 zB zB
' yA l ' l ' 1 A B 0.5 0.5 A 0.5 B
yA yA
' 1
1
yB
l' l' A 1 B 0.5 A 0.5 0.5 B B yB
yB
xA yA
1 yA
xB yB yB
1
xA xA
xB xB
(2-10-10)
From global node displacement to element node displacement
u zA
u
zB
yA u1
u
yB 2
TixB (2-10-11)
yA
yB u n
xA
xB
u1 u1
' yA u u
2 2
'
yB n B B T
ixB T
xB (2-10-12)
'
x un un
68
In case of Y-direction beam
x 0 1 0 X
y 1 0 0 Y (2-10-13)
z
Y Beam 0 0 1 Z Global
u zA 1 u zA u zA
u 1 u u
zB zB zB
yA 1 xA xA
yB 1 xB
s B
xB
xA (2-10-14)
yA
1
xA
yB 1 xB xB
xA
1 yA yA
1 yB Global yB
xB Y Beam Global
Transformation from the global node displacement to the element node displacement is,
u zA
u
zB
xA u1
u
xB
TiyB 2
(2-10-15)
xA
xB u n
yA
yB
Transformation from the global node displacement to the element face displacement is,
u1 u1
' yA u u
' yB n B B s B TiyB
2
T yB
2
(2-10-16)
'
x u n u n
69
3) Column element
In case of rigid connection, as described in Equation (2-2-16), the nodal displacement in X-Z plane is
expressed as,
u
1 1 1 1 xA
u u l ' 1 A B u
yA
l '
xA A yA
l '
xB B yB
l' xB
1 (2-10-17)
A 1 B yA
1 1 1
yB u xA A yA u xB B yB
l' l' l' l' yB
B u xB
Bl ' yB
u xB B l ' yB ' yB
yB
l' Z
yA
u xA A l ' yA ' yA X
yA
Al '
Y
A u xA
70
If we consider shear deformation of connection panel, from Figure 2-10-8,
u xA
u
1 1 xB
l ' 1 A B 0.5 0.5 A 0.5 B yA
l'
1 1 (2-10-18)
A 1 B 0.5 A 0.5 0.5 B yB
l' l' yA
yB
B u xB
B l ' yB
u xB B l ' yB 0.5 yB
' yB
Z yB 0.5 yB
l'
X
yA 0.5 yA
' yA
u xA Al ' yA 0.5 yA
yA
Al '
u xA
A
Figure 2-9-8 Column displacement with shear deformation of connection panel (X-Z plane)
71
In the same manner, assuming rigid connection, the nodal displacement of column in Y-Z plane is
expressed as,
yA
u
1 1 1 1
xA l ' u yA A xA l ' u yB B xB l ' l ' 1 A B u
yB
(2-10-19)
A 1 B xA
1 1 1 1
xB u yA A xA u yB B xB
l' l' l' l' xB
Bl ' u yB
u yB B l ' xB
xB
'xB
l' xB
Z
xA u yA Al ' xA
X
'xA
Al '
xA A Y
u yA
72
If we consider shear deformation of connection panel, from Figure 2-10-10,
u yA
u
1 1 yB
l' 1 A B 0.5 0.5 A 0.5 B xA
l '
(2-10-20)
0.5 0.5 B xB
1 1
A 1 B 0.5 A
l' l' xA
xB
u yB B
xB Bl '
u yB B l ' xB 0.5 xB
' xB
xB 0.5 xB
Z
l'
Y
xA 0.5 xA
' xA
u yA Al ' xA 0.5 xA
xA
Al '
u yA
A
Figure 2-9-10 Column displacement with shear deformation of connection panel (Y-Z plane)
73
The transformation matrices for column element are;
Including connection panel and node movement
u xA
u
xB
yA
1 1 1 A B yB
1 A B
' yA l ' yA
l' 2 2 2
' 1 1 A 1 B
A 1 B
yB l ' l' 2 2 2 yB
' xA 1 1 1 A B u yA
1 A B u
'xB l' l' 2 2 2
yB
1 1 A 1 B
zA A 1 B xA
l' l' 2 2 2
zB
1 xB
zA 1 xA
zB 1 xB
1 zA
zB
zA
zB
u xA
u
xB
yA
yB
yA
yB
u
yA
u
C yB
xA
xB
xA
xB
zA
zB
zA (2-10-21)
zB
74
From global node displacement to element node displacement
u xA
u
xB
yA
yB
yA
yB
u u1
yA u
u yB 2
TiC
xA
xB u n
xA
xB
zA
zB
zA
(2-10-22)
zB
' yA
'
yB u1 u1
u
' xA u 2 2
n T
iC T
C (2-10-23)
'
xB
C C
'z u n u n
' z
75
4) Force-displacement relationship for the connection
x
y
The relationship between the displacement vector and force vector of the element is expressed as follows:
M Px k Px 0 x
k Py y
(2-10-24)
M Py 0
where, initial stiffness of connection area is,
k Px k Py GV (2-10-25)
u1
x u
2
T
P (2-10-26)
y
u n
The component of the transformation matrix, [TP ] , is discussed in Chapter 4 (Freedom Vector).
Constitutive equation
The constitutive equation of the external spring is;
P1 u1
P u
2 2
K P (2-10-27)
Pn u n
where,
76
3. Hysteresis model of nonlinear springs
Notation
y : Strength of rebar
Mc : Crack moment
My : Yield moment
k0 : Initial stiffness
ky : Tangential stiffness at the yield point
77
c : Crack shear deformation
78
3.1.1 Beam
3.1.1 RC Beam
a) Section properties
S
as
d1
t d1
a1
D
d2 a2 d2
B
B : Width of beam,
D : Height of beam,
S : Effective width of slab,
t : Thickness of slab
d1 : Distance to the center of top main rebars,
d2 : Distance to the center of bottom main rebars,
a1 : Area of top main rebars,
a2 : Area of bottom main rebars
as : Area of rebars in slab
AN BD S B t n E 1a1 a 2 a S (3-1-1)
where,
nE E s / Ec : Ratio of Young’s modulus between steel (Es) and concrete (Ec)
Area of section to calculate shear deformation
AS BD (3-1-2)
Moment of inertia around the center of the section
2 2
BD 3 ( S B )t 3 D
BD g S B t D g
t
Ie
12 12 2 2
2
n E 1a1 d1 g n E 1a 2 D d 2 g n E 1a S D g
2 2 t
(3-1-3)
2
where, g is the center of beam section calculated by
BD 2 / 2 ( S B )t D t / 2 n E 1a1 d 1 a 2 D d 2 a S D t / 2
g (3-1-4)
AN
79
b) Nonlinear bending spring
A
x
MA MB
A B
B
Nx
elastic element
A A
MA MB MA MB
A B A B
Nx B B
A A A A A
MA MB
B B B B A B B
0 0
x x
nonlinear shear springs
Hysteresis model of a nonlinear bending spring is defined as the moment-rotation relationship under the
anti-symmetry loading in Figure 3-1-3. The initial stiffness of the nonlinear spring is supposed to be infinite,
however, in numerical calculation, a large enough value is used for the stiffness.
M M
A B
Moment distribution
M M M
My My
6 EI = +
Mc k0 Mc
l
k 0 y k0 k0 kp
c y c y
Elastic element Nonlinear bending spring
80
Crack moment force
For reinforced concrete elements, the crack moment, M c is calculated as,
where,
B : Compression strength of concrete (N/mm2)
Z e1 , Z e 2 : Section modulus
where,
y : Strength of rebar (N/mm2)
Yield rotation
The tangential stiffness at the yield point, k y , is obtained from the following equation,:
6 Ec I e
k y y k0 , k0 (3-1-9)
l
where,
y is the stiffness degradation factor at the yield point, which is obtained from the following
empirical formulas:
y 0.043 1.63npt 0.043a / D d / D 2 , a / D 2 (3-1-10)
81
y is modified in case of tension in top main rebars as
I
y ' y e0 (3-1-12)
Ie
BD 3
where I e0 : the moment of inertia of square section without slab
12
The yield rotation of the nonlinear bending beam, y , is then obtained from,
1 My
y 1 (3-1-13)
k
y 0
In general, the relation between the rotation of bending spring and that of nonlinear bending spring is
My
(3-1-14)
k0
Crack rotation
From Figure 3-1-2, the crack rotation of the nonlinear bending beam, c , is supposed to be zero value,
however, in STERA_3D program, it is assumed as,
c 0.001 y (3-1-15)
82
Effective width of slab
Ls
B Lb
Sb
Sb B
In general, effective width of slab for the flexural behavior of a beam is assumed as,
S b 0.1 Lb D (3-1-16)
However, recent studies suggest the contribution of full length of slab to the flexural capacity, M y , of a
beam. Therefore, STERA3D adopts two types of effective widths:
S b s Ls (3-1-17)
83
Hysteresis model
To consider the difference of the flexural capacity between positive and negative side of the beam, a
degrading tri-linear slip model is developed based on the Takeda Model for the hysteresis model of the
bending springs of the beam.
M M
kp
My My
Mc Mc
ks kr
x y m
My My y Mm y
kp kr k s (3-1-18)
y y m m x m
The strength degradation under cyclic loading is considered by elongating the target displacement, m , to
be ' m as shown in the following Figure:
M
kp
Mm
My
ks kr
n m n y y m 'm
x
y
' m 1 m
n
(3-1-19)
n m
84
Relationship between curvature and rotation
c
EI
M M
A B
M
D
l
Let’s think about the relationship between curvature and rotation at the end of a beam.
In the above loading condition, the relationship between moment and rotation is
6 EI
M (3-1-20)
l
On the other hand, the relationship between moment and curvature is
M
(3-1-21)
EI
Therefore,
6
(3-1-22)
l
Assuming the neutral axis is in the middle of the section, the relationship between curvature and
compression strain at the section end is
c
(3-1-23)
D/2
Therefore, the relationship between rotation and compressive strain is
l l
c (3-1-24)
6 3D
l
Assuming D , then
9
3 c (3-1-25)
85
c) Nonlinear shear spring
Hysteresis model of nonlinear shear spring is defined as the shear force – shear rotation relationship using
an origin-oriented poly-linear model.
A
MA MB
MA MB Q
A B B l
A B l
l
nonlinear shear springs
Q Q
k y 3 0.001k 0
Qy
k y3
Qu
Qc
k 0 GA
k0
c y u
where,
86
Ultimate shear force
The ultimate shear force is, Qu , is assumed as,
Qu Qc (3-1-28)
NOTE)
In STERA_3D, the stiffness after yielding is temporary assumed to be positive to avoid instability in
numerical analysis.
Q
Qu
Qy
k y 3 0.001k 0
Qc
k 0 GA
c y u
87
d) Modification of initial stiffness of nonlinear springs
In numerical calculation, a large dummy value is used for the initial stiffness of the nonlinear spring to
represent rigid condition. This large stiffness may cause an error for estimating the force from the
displacement. One possible way to solve the problem is to reduce the initial stiffness of the nonlinear spring
to be a certain value reasonable for calculation, and on the other hand, increase the stiffness of the elastic
element so that the total initial stiffness of the beam element does not change from the original one. This
idea is proposed by K-N Li (2004) for MS model.
M M
A B
My My
6 EI = +
Mc k0 Mc
l
k 0 y k0 k0 kp
c y c y
My My 1 M
y = + y 1 y
y k0 k0 k
y 0
Increase Reduce
stiffness stiffness
M M
My
+ Mc
k k0 / 1
*
0
k *p EI / p1
* c* y* *
My 1M y 1 M
y = * + y* 1 y
y k0 k0 k
y 0
88
The idea is realized using flexibility reduction factors, 1 1, 2 1 , in the relationship between the
displacement vector and force vector of the elastic element in Equation (2-1-1) as,
l' l'
1 3EI
6 EI y
0
' yA y M ' yA
l' l'
' yB 2 0 M ' yB (3-1-32)
' 6 EI y 3EI y
N '
x l' x
0 0
EA
Also the initial flexibility matrix of the nonlinear spring can be expressed as follows, introducing the
parameters, p1 , p 2 to increase the initial flexibility.
yA p1 EI 0 M ' yA
(3-1-33)
yB 0 p2 EI M ' yB
When p1 0, p 2 0 , it represents the infinite stiffness for rigid condition. Accordingly, the crack and
yield rotation will be modified as,
Mc 1 M
c* p1 , y 1 y
*
(3-1-34)
EI k
y 0
In general, the relation between the rotation of bending spring and that of nonlinear bending spring is
My
1 (3-1-35)
k0
Making the modified flexibility matrix to be identical to the original one,
89
e) Modification of stiffness degradation factor at the yield point
(The following modification of the stiffness degradation factor, y , is suggested by Prof. Okano at Chiba
University.)
From Equations (3-1-32) and (3-1-34), the yield rotation of the member y under anti-symmetric loading
condition, M A M B M y , is calculated as,
2 1M y 1 M 1 M
y y 1 y (3-1-39)
k k
k0 y 0 y 0
where 1 2 .
The yield rotation y in Equation (3-1-39) is different from the formula in Figure 3-1-10 since the factor
is multiplied to only diagonal elements of flexural matrix in Equation (3-1-32).
1 1
1 (3-1-40)
' y y
To realize the designated value of stiffness degradation factor, y should be modified as,
1
y 1 1 (3-1-41)
'
y
For example, to realize the stiffness degradation factor ' y 0.4, assuming 0.7 , the modified y is
1
y 1 1 0.7 0.357
0.4
90
f) Modification of considering rigid zone ratio
A beam-column connection can be idealized as a rigid zone. In case of a beam element, the default length
of the rigid zone is set to be a half of the column width, and the nonlinear bending spring of the beam
element is arranged at the position of the column face.
On the other hand, if elastic deformation of the connection is considered by reducing the length of rigid
zone, the position of the nonlinear bending spring will be inside the connection area. In this case, when the
nonlinear bending spring is yielding, the moment value at the position of the column face is smaller than
the yield moment.
Column Column
My
A B
My
dA l dB
A B
d A l' d B
My' My
To make the moment at the column face to be the same as yield moment, the yield moment of the nonlinear
bending spring is increased as,
l / 2 (1 )d A
M y ' M y M y
l/2
(3-1-42)
d
1 2(1 ) A
l
For example, when l 540cm, d A 30cm, 0.75 ,
91
3.1.2 Steel Beam
a) Section properties
tf
H
tw
AN 2 Bt f H 2t f t w (3-1-44)
Area of section to calculate shear deformation ( )
AS H 2t f t w (3-1-45)
92
b) Nonlinear bending spring
A A
x
MA MB
MA MB A B
A B B
B Nx
Nx
elastic element
l A
A A A MA MB
A B
B B B B
0
x x
nonlinear bending springs
Hysteresis model of a nonlinear bending spring is defined as the moment-rotation relationship under the
anti-symmetry loading as shown in Figure 3-1-14. The initial stiffness of the nonlinear spring is supposed
to be infinite, however, in numerical calculation, a large enough value is used for the stiffness.
M M
A B
Moment distribution
M M M
My My
6 EI = +
k0
l kp
k0 k0
y y
Elastic element Nonlinear bending spring
93
Yield moment force
y
My
Figure 3-1-15
Yield rotation
From Figure 3-1-14, the yield rotation of the nonlinear bending beam, y , is supposed to be zero value,
y = 0.001 y (3-1-50)
where
6 EI
y M y / k0 , k0
l
Hysteresis model
A bi-linear model is assumed for the hysteresis model.
My
y
94
d) Modification of initial stiffness of nonlinear springs
In numerical calculation, a large dummy value is used for the initial stiffness of the nonlinear spring to
represent rigid condition. This large stiffness may cause an error for estimating the force from the
displacement. One possible way to solve the problem is to reduce the initial stiffness of the nonlinear spring
to be a certain value reasonable for calculation, and on the other hand, increase the stiffness of the elastic
element so that the total initial stiffness of the beam element does not change from the original one. This
idea is proposed by K-N Li (2004) for MS model, and can be used for nonlinear spring model also.
M M
A B
M M M
My My
6 EI = +
k0
l kp
k0 k0
y y
My My
y = + y 0
k0 k0
Increase Reduce
stiffness stiffness
M M
My
+
k0* k0 / 1 k *p EI / p1
* y* *
My 1M y My
y = * + y* 1 1
k0 k0 k0
95
The idea is realized using flexibility reduction factors, 1 1, 2 1 , in the relationship between the
displacement vector and force vector of the elastic element in Equation (2-1-1) as,
l' l'
1 3EI
6 EI y
0
' yA y M ' yA
l' l'
' yB 2 0 M ' yB (3-1-51)
' 6 EI y 3EI y
N '
x
0 l' x
0
EA
Also the initial flexibility matrix of the nonlinear spring can be expressed as follows, introducing the
parameters, p1 , p 2 to increase the initial flexibility.
yA p1 EI 0 M ' yA
(3-1-52)
yB 0 p2 EI M ' yB
When p1 0, p 2 0 , it represents the infinite stiffness for rigid condition. Accordingly, the yield
rotation will be modified as,
My
y* p1 (3-1-53)
EI
In general, the relation between the rotation of bending spring and that of nonlinear bending spring is
My
1 (3-1-54)
k0
Making the modified flexibility matrix to be identical to the original one,
96
3.1.3 SRC Beam
a) Section properties
S
as
d1
t d1
a1
D
d2 a2 d2
B
b1
B : Width of beam,
D : Height of beam,
S : Effective width of slab, tf
t : Thickness of slab
d1 : Distance to the center of top main rebars, tw h1
d2 : Distance to the center of bottom main rebars,
a1 : Area of top main rebars,
a2 : Area of bottom main rebars
as : Area of rebars in slab
b1 : Width of steel
h1 : Height of steel
tw : Thickness of web
tf : Thickness of flange
where,
n E E s / Ec : Ratio of Young’s modulus between steel (Es) and concrete (Ec)
nE 1 (3-1-60)
12
97
where, g is the center of beam section calculated by
BD 2 / 2 ( S B)t D t / 2 nE 1a1d1 a2 D d 2 aS D t / 2 aST D / 2
g
AN
(3-1-61)
b) Nonlinear bending spring
Hysteresis model of a nonlinear bending spring is the same as RC beam.
Crack moment force
For reinforced concrete elements, the crack moment, M c is calculated as,
where,
B : Compression strength of concrete (N/mm2)
Z e1 , Z e 2 : Section modulus
where
M y1, 2, RC : Yield moment of reinforced concrete (3-1-65)
M y1,RC 0.9a1 y D d1 0.9aS y D t / 2 when tension in top main rebars
M y 2,RC 0.9a2 y D d 2 when tension in bottom main rebars
where,
y : Strength of rebar (N/mm2)
1
M y ,S b1t f (h1 t f ) t w (h1 2t f ) 2 y ,S : Yield moment of steel (3-1-66)
4
where,
y,S : Strength of steel (N/mm2)
98
Appendix: Detail rule of degrading trilinear slip model
In OPTION menu in Beam Editor, you can control the shape of hysteresis loop.
Depending on the effective slab ratio R s , the yiled moment M y and the yield rotaion R y will change together
as shown in the Figure below, since the tangential stiffness at the yield point, K y , is assumed to be the
same.
Ru : Ultimate rotation angle to define the maximum moment before degradation. The default value is 1/50.
K p : The stiffness after the yield rotation angle, R y .
K u : The stiffness after the ultimate rotation angle, Ru .
It can be the negative value to consider strength degradation, however, the default value of the ratio K u / K y is
1/1000 without degradation.
99
M
kp
My
Effective slab ratio Rs
ku
( s in Eq.(3-1-17))
Stiffness over Ru
(could be negative)
ky
R
Ry Ru
Ultimate rotation angle
(recommended over 1/50)
Those parameters control the shape of hysteresis loop as descrived in Eqs. (3-1-18) and (3-1-19). That is,
M
Mm My y
My kr ( R1 )
y m
m n y ks kr Mm y
k s
n
( R2 )
x y m 'm m x m
y
' m 1 m
n
( R3 )
n m
More detail rule in the hysteresis loop is described in the following sections:
100
1. Elastic range
fy(1)
fc(1)
s0(1)
1
s0(2)
fc(2)
fy(2)
sy
4
sc
(drm(1), frm(1))
2
unloading point
3
s0
1
3
towards the maximum point of the other side
(drm(2), frm(2))
unloading point
101
3. Loading on the primary curve after yielding
(drm(1),frm(1))
fu 4
fy
fdm 5
ssd
xdm xd0
xd0 du xdm
(du,fu)
(drm,frm) (du,fu)
ssd fdm
su
5
ssd
4 xdm
(drm(2),frm(2))
If the displacement is less than the degrading point (du, fu),
the unloading point will the target point (drm, frm)
fy dy
The stiffness of unloading. ssd, will be calculated from ssd , where αis the parameter to control the stiffness
d drm
y
degradation depending on the ductility factor, (drm/dy). The default value of is α = 0.5
(du,fu) (drm,frm)
fdm
If the displacement is over the degrading point,
su
intersection of the LINE1 (degrading line) ssd
and LINE2 (unloading line) will be the target
point. LINE2 LINE1
xdm
LINE1
0.1fy
Lower boundary
LINE2
xdm
102
4. Crossing zero force line
(drm,frm)
4
fy
5
ssd
xd0
(drm(2),frm(2)) (drm(2),frm(2))
srm
(drm,frm)
(dsl,fsl)
6 5
ssl
xd0
ssl
(dsl,fsl) 6
5
srm
(drm(2),frm(2))
103
5. Calculation of slip point
srm
(dsl,fsl)
5
xd0
(dsl,fsl) 6
ssl
srm
LINE1 xd0
6
(dsl,fsl) ssl
LINE2
8
srm
ssd ssd
7 xd0 srm
ssd 6
ssl
10
8
104
dslp = s3*xd0/(s3-s2) 4
ssd
5
ssd
xd0
ssl 5
srm 6
11
9 ssd
5 8
7 xd0
6
After crossing zero force line from Lavel 7, Level 9 will be towards the
maximum point. Stiffness unloading from Level 9 (Level 11) is the same as
the stiffness of ssd (Level 5).
5
9
xd0
ssl 5
srm 6
11
5 8
7 xd0
6
105
3.2 Column
3.2.1 RC Column
a) Section properties
d1
a2
ac
x
a1
D
d2
y
B
B : Width of beam,
D : Height of beam,
d1 : Distance to the center of x-direction main rebars,
d2 : Distance to the center of y-direction main rebars,
a1 : Area of x-side main rebars,
a2 : Area of y-side main rebars,
ac : Area of corner main rebars
AN BD n E 1a1 a 2 a c (3-2-1)
Area of section to calculate shear deformation
AS BD / , 1.2 (3-2-2)
Moment of inertia around the center of the section
2
DB 3 B
Iy n E 1a c a1 d1 (3-2-3)
12 2
2
BD 3 D
Ix n E 1a c a 2 d 2 (3-2-4)
12 2
106
b) Nonlinear bending spring
To consider nonlinear interaction among M x M y N z , the nonlinear bending spring at the member
end is constructed from the nonlinear vertical springs arranged in the member section as shown in Figure
3-2-2.
N zB , zB
M xB , xB z
M yB , yB
B y
x z i xi
x
yi
i
A
y
M yA , yA
M xA , xA
N zA , zA
i z y i x xi y (3-2-5)
In a matrix form
M ' y
k i xi2 k i xi y i k x i i
y
i
y
x k p x
i i
x
M ' k y i
2
i k y i i (3-2-7)
N'
k
i i
z sym. i
z z
i
Therefore
107
y M ' y M ' y
x k p M ' x f p M ' x
1
(3-2-8)
N' N'
z z z
yA M ' yA
M '
xA xA
zA f pA 0 N ' zA
yB 0 f pB M ' yB (3-2-9)
xB M ' xB
zB N ' zB
Hysteresis model of nonlinear bending spring is defined as the moment-rotation relationship under the
anti-symmetry loading in Figure 3-2-3. The initial stiffness of the nonlinear spring is supposed to be infinite,
however, in numerical calculation, a large enough value is used for the stiffness.
A
Moment distribution
M
M M M
My My
6 EI = +
Mc k0 Mc
l
k0 k y y k0 k0 kp
c y c y
Elastic element Nonlinear bending spring
108
For reinforced concrete elements, the crack moment, M c is calculated as,
ND
M c 0.56 B Z e (3-2-10)
6
The yield moment, M y is calculated from the following formula under the balance axial force, N b ,
Nb
M y 0.8a t y D 0.5N b D 1 (3-2-11)
bD B
N b 0.4bD B (3-2-12)
Note that the balance axial force, N b , is used instead of actual axial force, N , in this formula since the
characteristics of nonlinear vertical springs in a section are determined later from the equilibrium condition
under the balance axial force.
The tangential stiffness at the yield point, k y , is obtained from the following equation,:
6 EI
k y y K0 K0 (3-2-13)
l
where,
y is the stiffness degradation factor at the yield point, which is obtained from the following
empirical formulas:
where,
k y y K0
2
M N b d 6 EI (3-2-16)
y 0.043 1.63npt 0.043 0.33 , K0
QD bD B D l
The yield rotation of the nonlinear bending beam, y , is then obtained from,
1 My
y 1
K
(3-2-17)
y 0
109
c) Nonlinear vertical springs
The nonlinear bending spring is constructed from the nonlinear vertical springs arranged in the member
section as shown in Figure 3-2-4. This model is called “Multi-spring model” proposed by S. S. Lai, G. T.
Will and S. Otani (1984) and modified by K-N. Li (1988). The section is devided in 5 areas; where 4 corner
areas have steel springs and concrete springs and the center area has one concrete spring.
The strength and the location of nonlinear springs are obtained from the equilibrium condition under
the balance axial force, N b , in Equation (3-2-8).
A1 1 2
A2
x x
5
3 4 Concrete spring
Steel spring
y y
(a) Original column section (b) Multi-spring model
(tension) (tension)
s fy
c d y
s dy
c f y
(compression) (compression)
y : Strength of rebar
110
Nb
c f y1 0.2bD B (3-2-19)
2
Therefore, the area of the corner concrete, A1 , is,
fy
A1
c
(3-2-20)
0.85 B
Nb
s f y c f y1 My
s f y c f y1
xs x
s f y1
s f y1
The area of the center concrete, A2 , is the rest of the area of the section,
A2 bD 4 A1 0 (3-2-21)
The strength of the center concrete spring is then obtained as,
c f y 2 0.85k B A2 (3-2-22)
M y x s 2 s f y c f y1 x s 2 s f y 0.5 N b (3-2-23)
Therefore,
My
xs (3-2-24)
2 s f y 0 .5 N b
Note that M y is calculated from Equation (3-2-7) for the balance axial force, N b .
111
Example)
To verify the efficiency of the Multi-Spring model for the column element, the M-N relationship is
compared between MS-model and Theory using one column element. The column section is shown in the
Figure below:
N=1000kN
300cm
50cm
Figure 3-2-6
Theoretical results of the M-N relationship are obtained from the equilibrium condition as,
if (0 N N b )
N
M y 0.8at y D 0.5 ND1 (3-2-25)
bD B
if ( N b N N max )
N N
M y 0.8at y D 0.12bD 2 B max (3-2-26)
N max N b
112
Firstly, the strengths and locations of vertical springs are calculated as
a t 15 .484 ( cm 2 ) y 1 .1 f y 32 .45 ( kN / cm 2 ) B 2 .4 ( kN / cm 2 )
M y 2 s f y 0 .5 N x s
M-N relationship
6000
Theory
5000 Multi-Spring
4000
N (kN)
3000
2000
Under-estimate
1000
0
0 10000 20000 30000 40000 50000 60000
M (kN*cm)
The results of Multi-Spring model give smaller values than theoretical results in the range 0 < N < Nb.
113
K-N. Li (1988) proposed to use the following formulation for deciding the location of vertical springs in
stead of Equation (3-2-24), as follows:
M y0
xs (3-2-29)
2 s f y 0.5 N 0
where, N 0 is the axial force from the dead loads and the live loads acting on the column ( N 0 N b ),
and M y 0 is the yield moment under the axial force N 0 , that is:
N0
M y 0 0.8a t y D 0.5 N 0 D1 (3-2-30)
bD B
x s 35.8 (cm)
M-N relationship
6000
Theory
5000 Multi-Spring
4000
N (kN)
3000
2000
1000
0
0 20000 40000 60000 80000
M (kN*cm)
It improves the results of Multi-Spring model in the range 0 < N < Nb.
114
Yield displacement of vertical spring
N0
2 fc M y0 Compression
y s dy c f y sf y
dc fc
2 s f y
xs dc c d y s d y
From the equilibrium condition under the axial force N 0 as shown in the above Figure, the yield
displacement of the tension side steel spring, s d y , is obtained as follows:
s dy d c y xs
fc
dc sdy
s f y c f y
N 0 2 s f y
fc (3-2-31)
2
y xs
sdy
N 0 2 s f y
1
2 s f y 2c f y
The yield displacement of concrete spring, c d y , is assumed to be the same as that of the steel spring,
c d y s d y (3-2-32)
115
d) Nonlinear shear spring
There are two nonlinear shear springs in x and y directions. Hysteresis model of the nonlinear shear springs
is the same as that in the beam element in Figure 3-1-4.
116
Qx
Qy
Qx y
Qy
117
e) Modification of initial stiffness of nonlinear springs
The same modification can be done for the nonlinear springs of column element as described for those of
beam element by reducing the initial stiffness of the nonlinear spring and increasing the stiffness of the
elastic element as shown in the following figure:
A
Moment distribution
M
M M M
My My
6 EI = +
Mc k0 Mc
l
k 0 y k0 k0 kp
c y c y
Elastic element Nonlinear bending spring
Increase Reduce
stiffness stiffness
M M
My
+ Mc
k *
0
k *p
c* y*
Elastic element Nonlinear bending spring
118
Introducint the concept of “plastic zones”, the initial stiffness of the i-th multi-spring can be expressed as,
Ei Ai
k 0i (3-2-39)
pz
where Ei : the material young’s modulus, Ai : the spring governed area, and pz : the length of assumed
plastic zone. When p z 0 , it represents the infinite stiffness for rigid condition.
From Equation (3-2-7), when we consider the flexural flexibility in x-z plane, the flexibility matrix for the
nonlinear MS section is,
1
y k i
0 xi2 0 pz
M ' y
E A x i i
2
i 0
M ' y
(3-2-40)
i i
z 0 1 k 0i N ' z 0 p z Ei A N ' z
i i
Also, introducing the flexibility reduction factors, 0 0, 1 0, 2 0 , the flexibility matrix of
the elastic element is,
l' l'
1 3EI
6 EI y
y
' l'
f C l
2
(3-2-41)
6 EI y 3EI y
l'
0
EA
p z1 l' l'
l' l' 1 0
E A x2
3EI 0 3 EI 6 EI y
6 EI y i i i y
y
i
l' l'
0
pz2
2 0
3EI y
i i i
E A x 2
3 EI y
l' i
sym. p z1 pz2 l'
EA original sym. 0
i Ei A i Ei A EA
mod ified
(3-2-42)
119
f) Tri-linear hysteresis for nonlinear springs
The original hysteresis models used for steel and concrete springs are bi-linear types as shown in Figure
3-2-3. To control both the initial stiffness and yield displacement, it is convenient to define tri-linear type
hysteresis.
For the steel spring, the maximum-oriented model is adopted for the hysteresis before yielding, and the
tri-linear model is adopted after yielding as shown in Figure 3-2-12.
sf y
sf y
s fy
sf y
E i Ai
k 0i s
pz
s d *y
The hysteresis of steel spring has the degradation point at the forces, sf y and s f y , where and
are the arbitrary parameters 1, 1 . The STERA_3D Program adopts the values as:
1 / 3, 0.5 (3-2-45)
*
Then, the yield deformation, s d y, may be obtained by Equations (3-2-24) and (3-2-10) considering the
reduction factor .
y* xs
sdy (3-2-46)
N 0 2 s f y
1
2 s f y 2c f y
1 1My
y* (3-2-47)
y k0
120
The hysteresis of concrete spring is also defined as tri-linear hysteresis model as shown in Figure 3-2-13.
After compression yielding, strength degradation is considered by reducing the strength of the target point
in reloading stage.
dc
fc
fy
dc
fc
fy
dm dc
fc
fy
dp
dm' dm1 r
dy dp
dm’ dm dc
fc
fy
121
References
1) S. S. Lai, G. T. Will, and S. Otani (1984), “Model for Inelastic Biaxial Bending of Concrete Members,”
Journal of Structural Division, ASCE, Vol. 110, ST1, 1984, pp.2563-2584.
2) K-N. Li (1988), “Nonlinear Earthquake Response of Reinforced Concrete Space Frames,” the
dissertation for the degree of Doctor in University of Tokyo (in Japanese), 1988.12.
3) K-N. Li (2004), CANNY, Technical Manual.
122
3.2.2 Steel Column
a) Section properties
tf t2
H t1
tw t1
B B D
AS 0.5 AN
123
3) Circle section
I
64
D 4
D 2t
4
(3-2-53)
2t1t 2 ( B t1 ) 2 ( H t 2 ) 2
J 2 2
(3-2-55)
BHt1t 2 t1 t1
5) Circle section
J
32
D 4
D t
4
(3-2-56)
124
b) Nonlinear bending spring
To consider nonlinear interaction among M x M y N z , the nonlinear bending spring at the member
end is constructed from the nonlinear vertical springs arranged in the member section as shown in Figure
3-2-16.
N zB , zB
M xB , xB z
M yB , yB
B y
x z i xi
x
yi
i
A
y
M yA , yA
M xA , xA
N zA , zA
i z y i x xi y (3-2-57)
In a matrix form
M ' y
k i xi2 k i xi y i k x i i
y
i
y
x k p x
i i
M 'x k y i
2
i k y i i (3-2-59)
N'
k
i i
z sym. i
z z
i
Therefore
125
y M ' y M ' y
x k p M ' x f p M ' x
1
(3-2-60)
N' N'
z z z
yA M ' yA
M '
xA xA
zA f pA 0 N ' zA
yB 0 f pB M ' yB (3-2-61)
xB M ' xB
zB N ' zB
Hysteresis model of nonlinear bending spring is defined as the moment-rotation relationship under the
anti-symmetry loading in Figure 3-2-17. The initial stiffness of the nonlinear spring is supposed to be
infinite, however, in numerical calculation, a large enough value is used for the stiffness.
A
Moment distribution
M
M M M
My My
6 EI = +
k0
l kp
k0 k0
y y
Elastic element Nonlinear bending spring
126
Yield moment force (full plastic moment)
1) I shape
tf
H
tw
a) When the neutral axis is inside the web, i.e., N < Aw y = t w ( H - 2t f ) y
M y M y 0 y 0 t w y
2
(3-2-62) Bt f y
1
t w ( H 2t f ) y
2
where
1 B
M y 0 Bt f ( H t f ) t w ( H 2t f ) 2 y
4
N M y0
y0
2t w y H
H H
M y B y 0 y 0 y (3-2-63) H
2 2 B y 0 y
where
2
1 N N y
y0 H
2 B y
N y 2 Bt f t w H 2t f y
My
127
2) H shape
1
M y M y 0 y 0 H y
2
(3-2-64) Bt f y
2
1
where t w ( H 2t f ) y
2
1 1 2
M y 0 B 2 t f t w ( H 2t f ) y
2 4 H
M y0
N
y0
2 H y
B
b) When the neutral axis is inside the web, i.e., N Aw y t w H y
B B 1
M y 2t f y 0 y 0 y (3-2-65) t f B y 0 y
2 2 2
where
1 N N y
y0 B
2 2t f y
H
My
N y 2 Bt f t w H 2t f y
3) Box shape
1
t w ( H 2t f ) y
t2 Bt f y 2
H M y0
t1
B
B H
128
4) Circle shape
t1
N
M y M y 0 cos (3-2-68)
2N
y
where
M y 0 D t1 t1 y
2
N y D t1 t1 y
Yield rotation
The yield rotation is
6 EI
y M y / k0 , k0 (3-2-69)
l
129
c) Nonlinear vertical springs
The nonlinear bending spring is constructed from the nonlinear vertical springs arranged in the member
section as shown in Figure 3-2-18. This model is called “fiber model”. The section is devided in several
areas which have steel springs.
30°
(H-2tf)/5 tf t2
H t1
tw t1
B B
(tension)
f y ,i
d y ,i
(compression)
f y ,i Ai y (3-2-70)
d y ,i f y ,i / k 0i , k 0i E s Ai (3-2-71)
130
The same modification can be done for the nonlinear springs of column element as described for those of
beam element by reducing the initial stiffness of the nonlinear spring and increasing the stiffness of the
elastic element as shown in the following figure:
A
Moment distribution
M
M M M
My
6 EI = +
k0
l kp
k0 k0
y y
Elastic element Nonlinear bending spring
Increase Reduce
stiffness stiffness
M M
My
+
k 0*
y*
Elastic element Nonlinear bending spring
Introducint the concept of “plastic zones”, the initial stiffness of the i-th multi-spring can be expressed as,
Ei Ai
k 0i (3-2-72)
pz
where Ei : the material young’s modulus, Ai : the spring governed area, and pz : the length of assumed
131
plastic zone. When p z 0 , it represents the infinite stiffness for rigid condition.
From Equation (3-2-7), when we consider the flexural flexibility in x-z plane, the flexibility matrix for the
nonlinear MS section is,
1
y k i
0 xi2 0 pz
M ' y
E A x i i
2
i 0
M ' y
(3-2-73)
i i
z 0 1 k 0i N ' z 0 p z Ei A N ' z
i i
Also, introducing the flexibility reduction factors, 0 0, 1 0, 2 0 , the flexibility matrix of
the elastic element is,
l' l'
1 3EI
6 EI y
y
' l'
f C l
2
(3-2-74)
6 EI y 3EI y
l'
0
EA
p z1 l' l'
l' l' 1 0
E A x2
3EI 0 3 EI 6 EI y
6 EI y i i i y
y
i
l' l'
0
pz2
2 0
3EI y
i i i
E A x 2
3 EI y
l' i
sym. p z1 pz2 l'
EA original sym. 0
i Ei A i Ei A EA
mod ified
(3-2-75)
132
3.2.2 Column with direct input
M ' xB M ' xB
M ' yB
M ' yB
B B B
' xB
' yB
N 'z Z
l' = +
T 'z
' yA ' xA X
A A A
M ' yA Y
M ' yA
M ' xA M ' xA
In case of direct input for Moment-Rotation relationship, we neglect nonlinear interaction among
M x M y N z and define the flexural stiffness of nonlinear bending springs in X and Y directions
independently. The rotational displacement vector of the nonlinear bending springs in Equation (3-2-9) will
be
yA M ' yA f yA M ' yA
M ' f xA M '
xA xA xA
zA f pA 0 N ' zA 0 N ' zA
yB 0
f pB M ' yB f yB
M ' yB
(3-2-78)
xB M ' xB f xB M ' xB
zB N ' zB 0 N ' zB
The displacement vector of the column element is obtained as the sum of the displacement vectors of
elastic element, nonlinear shear springs and nonlinear bending springs,
133
The flexural matrix [ f C ] is;
l' 1 l' 1
f yA
3EI y k sx l ' 6 EI y k sx l '
l' 1
f xA 0
3EI y k sx l '
l' 1 l' 1
f yB
f C 3EI x k sy l ' 6 EI x k sy l '
l' 1
f xB
3EI x k sy l '
l'
EA
l'
sym.
GI z
(3-2-80)
134
3.2.3 SRC Column
a) Section properties
d1
x
a2
ac
a1
y D
d2
B
B : Width of beam,
D : Height of beam,
d1 : Distance to the center of x-direction main rebars,
d2 : Distance to the center of y-direction main rebars,
a1 : Area of x-side main rebars,
a2 : Area of y-side main rebars,
ac : Area of corner main rebars
b1 : Width of steel
h1 : Height of steel
tw : Thickness of web
tf : Thickness of flange
where,
nE E s / Ec : Ratio of Young’s modulus between steel (Es) and concrete (Ec)
n f 2, nw 1 : Type1, Type2,
n f 4, nw 2 : Type3
n f 3, nw 2 : Type4, Type5
Area of section to calculate shear deformation
AS BD / , 1.2 (3-2-82)
135
Moment of inertia around the center of the section
DB 3 B
2
Iy n E 1 a c a1 d 1 I S , y (3-2-83)
12 2
BD 3 D
2
Ix n E 1 a c a 2 d 2 I S , x
(3-2-84)
12 2
where
IS : Moment of inertia of steel
I S,x I S,y
II
1
12
b1h1 b1 t w h1 2t f
3 3
IH
1
12
3
2t f b1 (h1 2t f )t w
3
Type 1
IH II
Type 2
II IH II IH
Type 3
2
II IH h
I I I H AH 1
Type 4 2
2
h II IH
I I I H AH 1
Type 5 2
136
Yield moment force
The yield moment, M y is calculated as,
M y M y , RC M y , S (3-2-86)
where
M y , RC : Yield moment of reinforced concrete
Nb
M y , RC 0.8a t y D 0.5 N b D1 (3-2-87)
bD B
M y,S : Yield moment of steel
M y,S , x M y,S , y
M yI M yH
Type 1
M yH M yI
Type 2
M yI M yH M yI M yH
Type 3
M yI M yH M yI M yT
Type 4
M yI M yT M yI M yH
Type 5
1
M yI b1t f (h1 t f ) t w (h1 2t f ) 2 y ,S
4
1 2 1 2
M yH b1 t f t w (h1 2t f ) y , S
2 4
1
M yT b1t f (h1 t f ) t w (h1 t f ) 2 y , S
2
b1 M yH h1 b1
M yI M yT
h1 b1 h1
137
3.3 Wall
3.3.1 RC Wall
a) Section properties
l w1
lw
lw2
t
x
y
l w , l w1 , l w 2 : Width of wall,
t : Depth of wall,
C1, C2 : Side columns,
aw : Area of rebars in a wall panel
where,
where,
where,
138
b) Nonlinear bending spring
To consider nonlinear interaction among M x M y N z , the nonlinear bending spring at the member
end is constructed from the nonlinear vertical springs arranged in the member section as shown in Figure
3-3-2.
z
x 2
x1 yc
zc i
yi x
xi
N ' zc , zc
M ' yc , yc
139
In the wall panel direction, all vertical springs in the nonlinear section are assumed to work against the
moment and the axial force. The equilibrium conditions are,
Nc N1 N2
M ' yc k i i xi k i i xi k i i xi
i i i
Nc N1 N2
k i ( zc xi yc ) x i k i ( zc y i x1 xi yc ) xi k i ( zc y i x 2 xi yc ) xi
i i i
yc
Nc N 1 N 2 N1 N2 Nc N 1 N 2
x1
k i xi2 k i xi y i k i xi y i i k i xi
i i i x2
zc
(3-3-5)
Nc N1 N2
N ' zc k i
i i k i i k i i
i i
Nc N1 N2
k (
i
i zc x i yc ) k i ( zc y i x1 x i yc ) k i ( zc y i x 2 x i yc )
i i
yc
Nc N 1 N 2
Nc N 1 N 2 N1 N2
x1
k i xi k i yi ki yi i k i
i i i x2
zc
(3-3-6)
where, Nc, N1 and N2 are the number of vertical springs in a wall panel, side column 1 and side column 2,
respectively.
140
In the out of wall direction, we establish the equilibrium condition for each side column independently. The
equilibrium condition for the side column 1 is,
N1
M ' x1 k i i y i
i
N1
k i ( zc y i x1 xi yc ) y i
i
yc
N1 N1 N1
x1
k i xi y i k y 2
0 k i yi
x 2
i i
i i i
zc
(3-3-7)
Also, for the side column 2,
N2
M ' x 2 k i i y i
i
N2
k i ( zc y i x1 xi yc ) y i
i
yc
N2 N2 N2
x1
k i xi y i 0 k y 2
k i yi
x 2
i i
i i i
zc
(3-3-8)
In a matrix form
Nc N 1 N 2 2 N1 N2 Nc N 1 N 2
k i xi k i xi y i k i xi y i k x i i
M ' yc N 1
i i
N1
i i
N1 yc
M ' k i xi y i k y k i yi
2 yc
0
x1 i x1
x1
i i
N2 k p
i i
N2 N2
M ' x 2 k i xi y i 0 k y 2
k i yi x 2 x2
N ' zc i zc
i i
i i zc
Nc N 1 N 2 N1 N2 Nc N 1 N 2
k i xi k i yi k i yi i k i
i i i
(3-3-9)
Therefore
yc M ' yc M ' yc
M ' M '
x1
kp 1
x1
f p
x1
(3-3-10)
x 2 M '
x2 M ' x 2
zc N ' zc N ' zc
141
yAc M ' yAc
M '
xA1 xA1
xA2 M ' xA2
zAc f pA
0 N ' zAc
yBc 0
f pB M ' yBc
(3-3-11)
For the out of wall direction, each side columns behave independently in the same way as the column
element. Therefore, we discuss here only the hysteresis model in the wall panel direction. Hysteresis model
of nonlinear bending spring is defined as the moment-rotation relationship under the symmetry loading in
Figure 3-3-5. The initial stiffness of the nonlinear spring is supposed to be infinite, however, in numerical
calculation, a large enough value is used for the stiffness.
M
A
Moment distribution
M
M M M
My My
2 EI = +
Mc k0 Mc
l
k0 k y y k0 k0 kp
c y c y
Elastic element Nonlinear bending spring
142
The yield moment, M y is obtained from the equilibrium condition in Figure 3-3-6 as,
where,
a s y
My
a w wy
lw
M c 0.3M y (3-3-13)
The tangential stiffness at the yield point, k y , is obtained from the following equation:
k y 0.2 K 0 (3-3-14)
The yield rotation of the nonlinear bending beam, y , is then obtained from,
1 My
y 1
K
(3-3-15)
y 0
y 0.02 (3-3-16)
143
c) Nonlinear vertical springs
The nonlinear bending spring is constructed from the nonlinear vertical springs arranged in the member
section as shown in Figure 3-3-6. This model is based on the concept of “Multi-spring model” and
modified for the wall element by Saito et.al. The vertical springs in the side columns are determined
independently in the same way as the Multi-spring models of columns. The wall panel section is devided in
5 areas, and a steel springs and a concrete spring are arranged at the center of each area.
l w1
lw
lw2
5 x 10
11 12 13 14 15
3 4 8 9
y Concrete spring
Steel spring
(b) Multi-spring model
(tension) (tension)
s fy
c d y
s dy
(compression) (compression)
144
Strength of steel spring in wall panel
The strength of the steel spring in the wall panel is one-fifth of total strength of rebars in the section,
a w wy
s fy (3-3-17)
5
where,
aw : Total area of vertical rebar in the wall panel
wy : Strength of rebar in the wall panel
Strength of concrete spring in wall panel
The strength of the concrete spring in the wall panel is one-fifth of total strength of concrete in the section,
0.85 A p B
c fy (3-3-18)
5
where,
Ap : Total area of wall panel section
B : Compression strength of concrete
where,
145
Crack shear force
The crack shear force is, Qc , is assumed as,
Qy
Qc (3-3-20)
3
Ultimate shear force
The crack shear force is, Qu , is assumed as,
Qu Qc (3-3-21)
Q y1 Q xc Qy2
Q y1 Q xc Qy2
146
e) Modification of initial stiffness of nonlinear springs
The same modification can be done for the nonlinear springs of wall element as described for those of
beam and column elements by reducing the initial stiffness of the nonlinear spring and increasing the
M
A
Moment distribution
M
M M M
My My
2 EI = +
Mc k0 Mc
l
k0 k y y k0 k0 kp
c y c y
Elastic element Nonlinear bending spring
Increase Reduce
stiffness stiffness
M M
My
+ Mc
k *
0
k *p
c* y*
Elastic element Nonlinear bending spring
147
Introducint the concept of “plastic zones”, the initial stiffness of the i-th multi-spring can be expressed as,
Ei Ai
k 0i (3-3-25)
pz
where Ei : the material young’s modulus, Ai : the spring governed area, and pz : the length of assumed
plastic zone. When p z 0 , it represents the infinite stiffness for rigid condition.
In the same manner of beam and column elements, introducing the flexibility reduction factors,
0 0, 1 0, 2 0 , the flexibility matrix of the elastic element is,
l' l'
1 3EI
6 EI c
c
l'
2
3EI c
l' l'
1
3EI 1 6 EI 1
l'
fW 2 (3-3-26)
3EI 1
l' l'
1
3EI 2 6 EI 2
l'
sym. 2
3EI 2
l'
0
EAc
l'
Also, adopting p z as discussed for beam and column elements, the reduction factors will be:
10
1 2 0.7 , 0 0.8 (3-3-27)
148
3.3.2 Steel Wall (Brace)
a) Axial strength
3 B 4
N', '
Z
h
1 2 X
A
w Y
N c AN 1 0.4 y ,
2
for (3-3-28)
0.6 y
N c AN , for (3-3-29)
2
where
2E
0.6 y Nc
b) Hysteresis model
A bi-linear model is assumed for the hysteresis model.
dy
149
3.3.3 SRC Wall (Brace)
a) Section properties
Qy Qy , RC Qy , S (3-3-30)
where
Qy , RC : Yield shear force of reinforced concrete
0.053 pt 0.23 ( B 18)
Qy , RC 0.85 pw wy 0.1 0 b j (3-2-31)
M /(QD) 0.12
Qy , S : Yield shear force of steel
Qy , S AS y , S cos R (3-2-32)
where,
AS : Area of steel (mm2)
y, S : Strength of steel (N/mm2)
R : Angle of steel
150
3.4 External Spring
N 'z , 'z
In STERA_3D, the external spring is attached at the base of the building to express the stiffness and
strength of the foundation of the building. In such a case, hysteresis model of the nonlinear vertical spring
is defined as the axial force – displacement relationship as shown in Figure 3-5-2; where, bilinear skeleton
is defined only in compression side, and the spring has zero stiffness in the tension side assuming that the
N 'z
(tension)
e d y
'z
e f y
(compression)
Initial stiffness
The initial stiffness of the vertical stiffness can be obtained from the following equation:
k e a F AF (3-4-1)
where,
151
3.5 Base Isolation
The element model of base isolation consists of shear springs arranged in x-y plane changing its direction
with equal angle interval as shown in Figure 3-5-1. This model is called MSS (Multi-Shear Spring) model
x x qi , u i
y i x
The hysteresis model of each nonlinear shear spring is defined as a bi-linear model as shown in Figure
3-5-2. The force and displacement vectors of i-th shear spring are expressed as,
qi , x cos i
qi (3-5-1)
q
i , y sin i
u x
u i cos isin i (3-5-2)
u y
From the relationship, qi k i u i , the constitutive equation of i-th shear spring is,
qi
fy ky qi , y
qi
qi , x
k0
ui i x
dy
152
From the sum of all nonlinear shear springs in the element, the constitutive equation of the base isolation
element is,
We assume that all nonlinear shear springs in an element have the same stiffness and strength. The initial
stiffness of the base isolation element, K 0 , is obtained from Equation (3-5-4) by substituting
u x 1, u y 0 .
N
K 0 cos 2 i k 0 (3-5-5)
i 1
Therefore, the initial stiffness of each shear spring is,
K0
k0 N
(3-5-6)
cos
i 1
2
i
The same relationship is established for the second stiffness after yielding,
Ky
ky N
(3-5-7)
cos
i 1
2
i
where, K y and k y are the second stiffness after yielding for the base isolation element and the
nonlinear shear spring, respectively.
N
Q y cos i fy (3-5-8)
i 1
Therefore, the yield shear force of each shear spring is,
Qy
fy N
(3-5-9)
cos
i 1
i
153
fy
fy
Qy i fy
fy
fy
154
Appendix :
LRB (Lead Rubber Bearing) is composed by rubber layers, steel plates and a lead plug core.
1) Bi-Linear Model
The bi-linear hysteresis of LRB is defined as a combination of an elastic model and elasto-plastic model as shown
Figure A1-2.
K2
Fy
Qd
K1
+ Kp = Keq
Qo Kr
Dy Dy Dy
Qd
RUBBER LEAD
155
where Gr is the shear modulus of the rubber, Ar is the cross section area of the rubber and Hr is the total height of
the rubber.
The elastic stiffness, Kp, from the lead plug is calculated as,
Ap
K p Gp (A1-2)
Hp
where Gp is the shear modulus of lead, Ap is the cross section area of lead plug and Hp is the total height of the plug.
The initial elastic stiffness, K1, and the secondary stiffness, K2, of the bi-linear model are then obtained as,
K1 K r K p
(A1-3)
K2 Kr
The yielding deformation, Dy, is determined from the characteristics of the lead plug. The yielding force, Fy, is
calculated as,
F y K r K p D y (A1-4)
Hysteresis of a lead rubber bearing has a characteristic of stiffness degrading according to the strain level as shown
in Figure A1-3.
Deformation, δ
K d C Kd K r K p (A1-5)
where is a strain ratio ( / H r ) and C Kd is a modification factor of the secondary stiffness, which
takes into consideration the strain dependency. Also, the intercept force is defined as,
156
Qd C Qd p A p (A1-6)
where C Qd is a modification factor of the yielding shear force and p is the yielding shear stress of lead.
F K d Qd (A1-7)
The modification factors, C Kd and C Qd , are represented by the following formulas under 15 degrees
Celsius.
The primary stiffness of the lead rubber bearing, Ku, in Figure A1-4 is determined from the secondary stiffness, Kd,
as,
Ku Kd (A1-12)
where 10 15 .
157
Following the suggestion in the manual of CANNY (K. Li, 2004), the hysteresis rules are:
a) Elastic range
Under the strain level less than γe, the hysteresis is assumed to be linear with the secant stiffness at the strain, γe,
that is:
K 0 Fe / e (A1-13)
Fe K d e e Qd e , e e H r (A1-14)
The value, e 0.1 , is adopted in STERA3D.
b) Loading on the skeleton curve after elastic range
Under the loading on the skeleton curve after elastic range, tangent stiffness is used to estimate the response at
the next step:
K dF ( ) / d (A1-15)
Reference:
Response Control and Seismic Isolation of Buildings, Edited by Masahiko Higashino and Shin Okamoto, SPON
PRESS, October 17, 2006.
Canny Technical Manual, Kangning Li, August 2004
158
A-2. Hysteresis of HDRB (High Damping Rubber Bearing)
HDRB (High Damping Rubber Bearing) is composed by rubber layers and steel plates. By adding special
ingredient in the natural rubber, rubber itself demonstrates damping characteristics.
The hysteresis of HRB is defined as a modified bilinear model as shown Figure A2-2.
K1 G1 A / H r
u heq / 2 u heq / 2
G1 Geq (A2-1)
u heq / 2
159
where : shear strain ( / H r )
Ar : cross section area of the rubber
Hr : total height of the rubber.
Example)
Diameter: 1500
Thickness of rubber: 7.5mm 20 layers 150mm
S 1 49.7
S 2 10.0
2
Nominal compression stress: 10 N / mm
160
Hysteresis of a high damping rubber bearing has a characteristic of stiffness degrading according to the strain level
as shown in Figure A2-3.
Deformation, δ
G2 1 u Geq (A2-6)
a) Elastic range
Under the strain level less than 0.1 , the hysteresis is assumed to be linear with the secant stiffness at the
strain, that is:
K 0 K h ( 0.1) (A2-10)
b) Loading on the skeleton curve after elastic range
Under the loading on the skeleton curve after elastic range, tangent stiffness is used to estimate the response at
the next step:
K dQ( ) / d (A2-11)
161
2) Consideration of strength reduction by dissipated energy
Reference
1) Takuya Nishimura et al., “Practical Response Evaluation Method for Seismic Isolation System against Long Period
Earthquake Motions - Part2- High-Damping Rubber Bearing and Lead Damper”, AIJ Annual Convention,
Architectural Institute of Japan, 2013, pp.767-768 (in Japanese)
In the above reference, the reduction factors of equivalent stiffness and equivalent damping are proposed as,
C K 0.0073 E V 1.0
E V 10.0 N mm 2 (A2-12)
C K 0.0025 E V 0.952
E V >10.0 N mm 2
C h 0.0039 E V 1.0 E V 10.0 N mm 2 (A2-13)
C h 0.0016 E V 0.977 E V >10.0 N mm 2
where E: dissipated energy, V: volume of rubber
To consider the strength reduction by energy dissipation, STERA_3D modifies the equivalent shear modulus and
the equivalent damping factor as,
162
A-3. Hysteresis of Lead Damper
Reference
1) Takuya Nishimura et al., “Experimental Study on Multi-cyclic Characteristics of Devices for Seismic Isolation
against Long Period Earthquake Motions: Part 7- Lead Damper”, AIJ Annual Convention, Architectural Institute of
Japan, 2011, pp.667-668 (in Japanese)
2) Takuya Nishimura et al., “Study on Multi-cyclic Modeling of Devices and Response Evaluation for Seismic
Isolation against Long Period Earthquake Motions: Part 5-Modeling of Lead Damper and Seismic Response
Analyses”, AIJ Annual Convention, Architectural Institute of Japan, 2012.9, pp.383-384 (in Japanese)
Cast Lead
1.2
① d=100mm d=200mm d=400mm
1
0.8
0.6 ② ③
0.4
0.2
0
④
0 2000 4000 6000 8000 10000
In the above references, from the cyclic loading test of a lead damper with the different horizontal displacement
amplitudes, three line graphs are obtained for the relationship between the dissipated hysteresis energy and the
horizontal strength reduction ratio. The breaking points of the line are proposed as follows to match the test results.
163
a) The first point of strength reduction, R1 , E1
R1 1.0
0.1 d 0.4 , 195 d <0.1, 892 d >0.4 (kNm)
E1 37 2322 d (A3-1)
The hysteresis of the lead damper is defined as a bilinear model. To consider the strength reduction by energy
dissipation, STERA_3D adopts the line of d 0.2 (m) for random amplitude. The strength of a lead damper, Qd ,
is then expressed as,
Qd R Qd 0 (A3-5)
Force, F
Skelton curve
Deformation, δ
164
A-4. Hysteresis of Elastic Sliding Bearing
Reference
1) Shigeo Minewaki et al., “Study on Multi-cyclic Modeling of Devices and Response Evaluation for Seismic Isolation
against Long Period Earthquake Motions : Part 2-Modeling of Low Friction Bearing and Viscous Damper”, AIJ
Annual Convention, Architectural Institute of Japan, 2012, pp.377-378 (in Japanese)
In the above reference, the dynamic friction coefficient changes according to the temperature of the sliding plate as,
The change of the friction coefficient is expressed as a function of the increment of temperature as
0.03 T 1
0.06
0.03 (A4-2)
On the other hand, the increment of temperature has the following relationship with the dissipated energy
E d (kNmm),
0.9
T 0.00019 E d (A4-3)
Therefore, the dynamic friction coefficient is obtained from the dissipated energy,
0 T Ed (A4-4)
The hysteresis of the elastic sliding bearing is defined as a bilinear model. In STERA_3D, the initial friction
coefficient is temporary assumed as 0 0.029 from the catalog of a manufacture. The strength reduction by
energy dissipation will be expressed as,
Qd 0 E d Qd 0 0 (A4-5)
165
A-5. Hysteresis of Bouc-Wen Model
1 1
0 0
‐3 ‐2 ‐1 0 1 2 3 ‐3 ‐2 ‐1 0 1 2 3
‐1 ‐1
Reference
1) Terje Haukaas and Armen Der Kiureghian, “Finite Element Reliability and Sensitivity Methods for
Performance-Based Earthquake Engineering”, PEER 2003/14, APRIL 2004
2) Wen, Y.-K. (1976) “Method for random vibration of hysteretic systems." Journal of Engineering
Mechanics Division, 102(EM2), 249-263.
3) Baber, T. T. and Noori, M. N. (1985). “Random vibration of degrading, pinching systems." Journal
of Engineering Mechanics, 111(8), 1010-1026.
a) Basic formulation
The basic formula of Bouc-Wen model is
f k 0 x 1 k 0 z (A5-1)
z
Ax x z
N 1
z x z
N
(A5-2)
where, , , and N are parameters that control the shape of the hysteresis loop, while A, , and
are variables that control the material degradation.
0 yN and 0 yN (A5-3)
A z
N
sgnxz z x
z x (A5-4)
x t
166
This leads to the following expression for the continuum tangent
f z A z
N
sgnxz
k k 0 1 k 0 k 0 1 k 0 (A5-5)
x x
The evolution of material degradation is governed by the following choice of equations (Baber and
Noori 1985):
A A0 A e, 1 e, 1 e (A5-6)
e 1 k 0 z x (A5-7)
z ( n 1) z ( n ) t zt n 1
(A5-9)
x( n 1) x( n ) t x t n 1
N
x( n 1) x( n )
A( n 1) z ( n 1) sgn z ( n 1) ( n 1)
t x( n 1) x( n )
z ( n 1) z ( n ) t (A5-10)
( n1) t
where
x( n 1) x( n )
e( n 1) e( n ) t 1 k0 z( n 1)
t (A5-12)
e( n ) 1 k0 z( n 1) x(n1) x(n)
Since
x( n 1) x( n )
sgn z ( n 1) sgnx( n 1) x( n ) z ( n 1) (A5-13)
t
167
f z ( n 1) z ( n1) z ( n ) x ( n 1) x( n ) 0 (A5-14)
( n1)
N
A( n 1) z ( n1) ( n 1) (A5-15)
f ' z ( n 1)
z
new old
z ( n 1) z ( n 1)
Original f z ( n 1) Function derivatives f ' z ( n 1)
x( n 1) x( n ) x( n1) x( n )
e( n 1) e( n ) t 1 k 0 z ( n 1) e' ( n 1) t 1 k 0
t t
N
A( n 1) z ( n1) ( n 1) ' A' ( n 1) N z ( n 1)
N 1
sgn z ( n 1) ( n1)
N
z ( n 1) ' ( n1)
'( n 1) '( n 1)
f z ( n 1) z ( n 1) z ( n ) x ( n 1) x( n ) f ' z( n 1) 1 x x( n )
( n1) 2( n 1) ( n 1)
(A5-18)
168
The procedure can now be summarized as follows:
N
A( n 1) z ( n1) ( n 1)
f z( n 1) z( n 1) z( n ) x( n 1) x( n ) (A5-19)
( n 1)
(b) Evaluate function derivatives
x( n1) x( n )
e' ( n 1) t 1 k 0
t
A' ( n 1) A e' ( n1)
f z( n1)
n 1) z ( n 1)
z(new (A5-21)
f ' z( n1)
n 1) z ( n 1)
z(old and z( n1) z(new
n 1) (A5-22)
169
c) Tangent stiffness
The tangent stiffness is necessary to compute the nonlinear structural analysis.
From the incremental forms:
N
x( n 1) x( n )
A( n 1) z ( n 1) sgn z ( n 1) ( n 1)
t x( n 1) x( n )
z ( n 1) z ( n ) t
( n1) t
z( n1) b4
(A5-24)
x( n1) b5
where
N
A( n 1) z ( n1) ( n 1)
b1 1 k 0 z( n1)
b2 1 k 0 x( n1) x( n )
b3
x ( n1) x( n )
( n1)
b4 b3 Ab1 b3 z( n1) b1
N
x
( n 1) x( n ) b1
2
( n 1) ( n1)
170
3.7 Masonry Wall
B1 B B2
A1 A2
A
Hysteresis model of the nonlinear shear spring is defined as the poly-linear slip model as shown in Figure
3-7-2.
Q
Q
Qy
k 0
Qc
Qu
k0
c y u
The characteristic values, Qc , Q y , Qu are obtained based on the formulation described in the reference
(Paulay and Priestley, 1992).
171
(1) Compression strength of masonry prism
The compression strength of the masonry prism ( f ' m ) is determined by the following equation (Paulay and
Priestley, 1992),
f ' cb ( f 'tb f ' j )
f 'm (3-7-2)
U u ( f 'tb f ' cb )
j
(3-7-3)
4.1hb
where,
f 'cb : Compressive strength of the brick
f 'tb : Tensile strength of the brick (= 0.1 f ' cb )
f 'j : Compressive strength of the mortar
j : Mortar joint thickness
hb : Height of masonry unit
Uu : Stress non-uniformity coefficient (=1.5)
172
(2) Shear strength by sliding shear failure
f 0 0 0 tan 0 (3-7-5)
where,
0 : Cohesive capacity of the mortar beds (=0.04 f ' m ) (Paulay and Priestly, 1992)
: Sliding friction coefficient along the bed joint
0.654 0.000515 f ' j (Chen et.al, 2003)
0 : Compression stress ( W / Aw R sin / Aw )
W
V f f AW 0 AW 0 Aw W (3-7-6)
AW
Therefore,
0 Aw
Vf (3-7-8)
1 tan
173
(3) Characteristic values of nonlinear skeleton
The shear resistance, Q y , is calculated to be the minimum value between the shear strength by sliding
shear failure, V f , and the shear strength of diagonal compression failure, Vc , that is,
Q y min(V f , Vc ) (3-7-9)
'm d m
y (3-7-10)
cos
where,
'm : Compression strain at the maximum compression stress
( ' m =0.0018, Hossein and Kabeyasawa, 2004)
Initial elastic stiffness is assumed as (Madan et al., 1997)
k 0 2Q y / y (3-7-11)
Shear resistance and displacement at the ultimate stage are assumed as (Hossein & Kabeyasawa, 2004)
Qu 0.3Q y (3-7-14)
u 3.5(0.01hm y ) (3-7-15)
where, hm is the height of masonry wall.
References:
1) T. Pauley, M.J.N. Priestley, 1992, Seismic Design of Reinforced Concrete and Masonry building, JOHN
WILEY & SONS, INC.
2) Hossein Mostafaei, Toshimi Kabeyasawa, 2004, Effect of Infill Walls on the Seismic Response of
Reinforced Concrete Buildings Subjected to the 2003 Bam Earthquake Strong Motion : A Case Study of
Bam Telephone Centre, Bulletin Earthquake Research Institute, The university of Tokyo
3) A. Madan,A.M. Reinhorn, ,J. B. Mandar, R.E. Valles, 1997, Modeling of Masonry Infill Panels for
Structural Analysis, Journal of Structural Division, ASCE, Vol.114, No.8, pp.1827-1849
174
b) Vertical springs
For the moment, the vertical springs of the element model in Figure 3-7-1 are assumed to be elastic springs.
k z E m (tl w ) / 2 (3-7-17)
where,
Em : Modulus of elasticity of masonry prism (=550 f ' m , FEMA 356, 2000)
t : Thickness of masonry wall
lw : Width of masonry wall
175
3.8 Passive Damper
a) Hysteresis damper
B1 B B2
l
Q' xc , ' xc
A1 A2
A
Different types of hysteresis model are prepared for the force-deformation relationship of the spring.
qi qi qi
fy ky ky ky
fy fy
fc k1 fc k1
k0 k0 ui k0 ui
ui
qi
fy
ui
(d) Bouc-Wen
176
b) Viscous damper
B1 B B2
A1 A2
A
Figure 3-8-4 shows the Maxwell model with an elastic spring with stiffness, K d , and a dashpot with
damping coefficient, C.
Fij, uij
Node i Node j
Fk, uk Fc, uc
Since the elastic spring and the dashpot are connected in a series,
Fk Fc Fij (3-8-1)
where, Fk : force of the elastic spring
Fc : force of the dashpot
Fij : force between i-j nodes
177
The force of the elastic spring, Fk , is obtained as,
Fk K d uk K d (uij uc ) (3-8-2)
where, u k : relative displacement of the elastic spring
u c : relative displacement of the dashpot
uij : relative displacement between i-j nodes
For an oil damper, the force-velocity relationship of the dashpot is defined as shown in Figure 3-8-5.
Fc
relief point
.
uc
178
Before the relief point of the dashpot, Equation (3-8-7) will be obtained by changing C2 C1 , Qc 0
as
K d u ij (t ) u c (t t )
u c (t ) (3-8-8)
C1
Kd
t
When the velocity of the dashpot is over the negative relief point, Equation (3-8-7) will be obtained by
changing Qc Qc ,
K d u ij (t ) u c (t t ) Qc
u c (t ) (3-8-9)
C2
Kd
t
Fij, uij
Node i Node j
Fc, uc
u ij (t ) u c (t )
Fuj Fc C 2 u c Qc
u c (t ) u ij (t )
u c (t )
t t
Therefore,
u ij (t )
Fij (t ) C 2 Qc (3-8-10)
t
179
(2) Algorithm for viscous damper devise
Figure 3-8-7 shows the Maxwell model with an elastic spring with stiffness, K d , and a dashpot with
damping coefficient, C.
Fij, uij
Node i Node j
Fk, uk Fc, uc
Since the elastic spring and the dashpot are connected in a series,
Fk Fc Fij (3-8-13)
where, Fk : force of the elastic spring
Fc : force of the dashpot
Fij : force between i-j nodes
For a viscous damper, the force-velocity relationship of the dashpot is defined as shown in Figure 3-8-8,
180
From Equations (3-8-13) and (3-8-14)
Fij (t )
u c (t ) u ij (t ) (3-8-16)
Kd
Taking time differential and substituting Equation (3-8-15) give
1/
Fij (t ) Fij (t )
Kd
sgn Fij (t )
C
u ij (t ) (3-8-17)
The numerical integration method, Runge-Kutta Method, can be used to solve the Equation (3-8-17).
1
Fij (t n 1 ) Fij (t n ) k 0 (t n ) 2k1 (t n ) 2k 2 (t n ) k 3 (t n ) (3-8-20)
6
1/
Fij (t n )
k 0 u ij (t n ) sgn Fij (t n )
C
K d t
1/
Fij (t n ) k 0 / 2
k1 u ij (t n t / 2) sgn Fij (t n ) k 0 / 2
d
K t
C
1/
Fij (t n ) k1 / 2
k 2 u ij (t n t / 2) sgn Fij (t n ) k1 / 2
d
K t
C
1/
Fij (t n ) k 2
k 3 u ij (t n t ) sgn Fij (t n ) k 2
d
K t
C
181
In this algorithm, it is assumed as,
u ij (t n ) u ij (t n t )
u ij (t n t / 2) (3-8-21)
2
182
4. Freedom Vector
Each node has six degrees of freedom and the freedom number is defined as shown in Figure 4-1-1.
Z 3 Z Z
6
X 1 7 X X
4
5 8
Y Y Y
2
(a) lateral and rotational displacement (b) shear deformation of connection
The freedom vector is defined to indicate the number of all freedoms of the structure, where the restrained
freedom is set to be zero. For the structure in Figure 4-2-1, the freedom vector has zero components for the
fixed nodes (Nodes 1-4) and eight components for other nodes (Nodes 5-8). Therefore, the total number of
freedom of the structure is 8×4 = 32.
183
Node 1 0
|
Node 4 0
1
Node 5 6
5
8
9
7 8
Node 6
16
1 2
17
Node 7
24 3 4
25
Node 8
32 shear deformation of connection
184
4.3 Dependent freedom
In the default setting, the floor diaphragm is assumed to be rigid for the in-plane deformation. Therefore,
the in-plane freedoms at the nodes in a floor are represented by the freedoms at the center of gravity of the
same floor.
6
1 4, 7
5, 8
2
(a) In-place freedoms (b) Out-of-plane freedoms
For example, the in-plane freedoms at the node, A, in Figure 4-3-2 are expressed by the in-plane freedoms
at the center of gravity, G, as follows:
u xA 1 0 l yA u xG
u yA 0 1 l xA u yG (4-3-1)
0 0 1 zG
zA
G: center of gravity
zG
G G u xG
l xA
l yA
u yG zA
A
A u xA
u yA
Figure 4-3-2 Rigid floor assumption
185
For example, in case of the structure in Figure 4-3-3, in addition to the original nodes, a new node for the
center of gravity is defined as “Node 10” in the first floor. Under the rigid floor assumption, the freedom
vector has zero components for the in-plane freedoms at the nodes except the center of gravity. Therefore,
the total number of independent freedom is 23.
0
Node 1-5 :
0
0
0
1
2
Node 6 3
0
4
5
0
0
6
Node 7 7
8
0
9 6 7
10
0 10
0
11 8 9
Node 8 12
13
0
14 1 2
15
0 5
0 3
16 4
17
Node 9 18
0
19
20 shear deformation of connection
21
22
0
0
Node 10 0
23
0
0
Freedom vector
Figure 4-3-3 Example of the freedom vector with rigid floor assumption
186
(2) Including wall element
The wall element model has rigid beams at the top and bottom of the wall, therefore, as shown in Figure
4-3-4, the rotation angles in the wall panel plane, y1 and y 2 , are dependent to the vertical
displacements, z1 and z 2 . Also, the horizontal displacement in the wall panel plane, u x 2 , is
dependent to the displacement, u x1 . The connection is assumed to be rigid.
y2
y1
z 2 z1
y1 y 2
z1 z2 w
u x1 u x2 u x1 u x 2
Figure 4-3-4 Relationship between node displacements for a wall element (X-wall)
In a matrix form;
u x1 1 0 0 u x 2
y1 0 1 / w 1 / w z1 (4-3-2)
0 1 / w 1 / w
y2 z2
x1
u y1 1 0 0 u y 2
z1
x1 0 1 / w 1 / w z1 (4-3-3)
0 1 / w 1 / w
1 x2 z 2
x2
u y1
z2
2 w
u y2
Figure 4-3-5 Relationship between node displacements for a wall element (Y-wall)
187
For example, in case of the structure in Figure 4-3-6, by eliminating dependent freedoms, the total number
of freedom becomes 17.
0
Node 1-5 :
0
0
0
1
Node 6 2
0
0
0
0
0
0
3
Node 7 4
0
0 6 7
0
0 10
0
0 8 9
5
Node 8 6
7
0 1 2
8
9 5
0
0 3 4
10
Node 9 11
12
0
13
14 shear deformation of connection
15
16
0
0
Node 10 0
17
0
0
Freedom vector
188
(3) Series of walls
In case several walls are joined together in series, it is assumed that all walls are connected by rigid beams
at the top and bottom. Therefore, as shown in Figure 4-3-7, the rotation angles in the wall panel plane, y1
and y 2 , are dependent to the vertical displacements, z1 and z 2 . Also, the horizontal displacement in
the wall panel plane, u x 2 , is dependent to the displacement, u x1 . The connection is assumed to be rigid.
yN
y2
y1
z1 z2 zN
u x1 u x2 u xN
w1
zN z1 N 1
y1 y 2 yN , L wi
L k 1
i 1
zi z1 yi Li 1 Li / L z1 Li / L zN , Li wi
k 1
u x1 u x 2 u xN
In a matrix form;
yi 1 / L 1 / L z1 (4-3-4)
zN
zi 1 Li / L Li / L z1 (4-3-5)
zN
189
In case of Y-direction wall, the relationship can be written as;
x1
z1
x2
1
u y1
z2
2 w1
u y2
zN
N z1 zN N 1
x1 x 2 xN , L wi
u yN L k 1
i 1
zi z1 xi Li 1 Li / L z1 Li / L zN , Li wi
k 1
u y1 u y 2 u yN
In a matrix form;
xi 1 / L 1 / L z1 (4-3-6)
zN
zi 1 Li / L Li / L z1 (4-3-7)
zN
190
(4) Transformation matrix for dependent freedom
In case of rigid floor assumption, Equation (4-3-1) expresses the relationship between dependent freedom
and independent freedom, that is;
u xA 1 0 l yA u xG
u yA 0 1 l xA u yG
0 0
zA 1 zG
It can be arranged into the transformation matrix between the freedom vectors of all nodes;
k l m
i u xA 1 0 l yA
u yA 0 1 l yA
zA u xG k
u l
xA yG
yA zG
zA 0 0 1
xG
yG
zG m
Since the most components of the transformation matrix, [TI ] , are zero, the components of [TI ] are
remembered using two matrices, [ N I ] and [ FI ] .
N I i k m 0
; Matrix for independent freedom numbers
FI i 1 l yA 0
; Matrix for transformation components from independent freedoms
191
It will reduce the memory size dramatically.
In the same way, for the case of including wall elements, Equation (4-3-2) expresses the relationship
between dependent freedom and independent freedom, that is;
u x1 1 0 0 u x 2
y1 0 1 / w 1 / w y1
0 1 / w 1 / w
y2 y 2
It can be arranged into the transformation matrix between the freedom vectors of all nodes;
p q r
u x1 0 1 0
j y1 1/ w 0 1/ w y1 p
y 2 1/ w 0 1/ w u x 2 q
y 2 r
N I j p r 0
; Matrix for independent freedom numbers
FI j 1 / w 1 / w 0
; Matrix for transformation components from independent freedoms
192
Initial conditions of [ N I ] and [ FI ] are:
N I i i 0 0 , FI i 1 0 0
In Figure 4-4-1 (rigid connection), the element node displacement vector of the beam element between
Node 8 and Node 9 is,
u z8 u z 9 y8 y 9 x8 x 9 T (4-4-1)
Those displacements correspond to the location numbers in the freedom vector as;
1 0
Node 1-5 : :
30 0
31 0
32 0 u z8 u z9
Node 6 33 1
34 2
35 3 x8 x9
36 0 8 9
37 0 y8 y9
38 0
Node 7 39 4
40 5
41 6
42 0
43 0 6 7
44 0
Node 8 45 7 10
46 8
47 9 8 9
48 0
49 0
50 0
Node 9 51 10 1 Z
52 11
53 12
54 0 5
55 13 3 4 X
56 14
57 0
Node 10 58 0 Y
59 0
60 15 (rigid connection)
Freedom vector
193
From rigid floor assumption, the components of independent matrices, [ N I ] and [ FI ] will be;
43 55 60 0
43 1 l y 8 0
45 45 0 0 45 1 0 0
47 47 0 0 47 1 0 0
N I
, FI (4-4-3)
49 55 60 0 49 1 l y 9 0
51 51 0 0 51 1 0 0
53 53 0 0 53 1 0 0
From the matrix, [ N I ] , the freedoms of (43) and (49) are replaced to the independent freedoms (55) and
(60). Therefore, the independent location numbers and freedom numbers of the beam element are:
u z8 u z 9 y8 y9 x8 x 9 T
45 51 47 53 43 49
T
u z 8 u z 9 y8 y9 u x10 z10
T
5 8 7 10 11 13 ;
T
freedom number
(4-4-4)
The transformation from independent displacements (= global node displacements) to element node
displacements is obtained from the matrix, [ FI ] , as follows:
u z 8 1 0 u z8 u z8
u 1 u
z9 u z9 z9
y 8 1 y 8 y 8
TixB (4-4-5)
y 9 1 y 9 y 9
x 8 1 l y 8 u x10 u x10
x 9 0 1 l y 9 z10 z10
194
The constitutive equation of the beam element and formulation of global stiffness matrix from element
stiffness matrix are shown below:
5 8 7 10 11 13
Pz 8 5 k 5,5 k 5,8 k 5, 7 k 5,10 k 5,11 k 5,13 u z 8
P 8
k 8 ,8 k 8, 7 k 8,10 k 8,11 k 8,13 u z 9
z9
M y 8 7 k 7,7 k 7 ,10 k 7 ,11 k 7 ,13 y 8
M y 9 10 k10,10 k10,11 k10,13 y 9
Px10 11 sym. k11,11 k11,13 u x10
M z10 13 k12,12 z10
Locate element stiffness
Element stiffness matrix
according to the freedom number
1 2 3 4 5 6 7 8 9 10 11 12 13
1
2
3
4
5 k 5, 5 k 5, 7 k 5,8 k 5,10 k 5,11 k 5,13
6
7 k 7,7 k 7 ,8 k 7 ,10 k 7 ,11 k 7 ,13
8 k 8,8 k 8,10 k 8,11 k 8,13
9
10 k10,10 k10,11 k10,13
11 sym. k11,11 k11,13
12
13 k13,13
In general, the transformation from independent displacements (= global node displacements) to element
node displacements for the X-beam is described as Equation (2-1-10).
u zA
u
zB u1
yA u
2
TixB (2-1-10)
yB
xA un
xB
195
And the constitutive equation of the X-beam is also described in Equation (2-1-16).
P1 u1
P u
2 2
K xB (2-1-16)
Pn u n
Using the same procedure in Figure 4-4-2, the element stiffness matrix is added into the global stiffness
matrix.
196
5. Equation of motion
In the default setting, the mass at each node is identical and equally distributed as
1
Mi M floor (5-1-1)
N floor
where, M i : mass at the node i, M floor : total mass of the floor, N floor : total number of nodes in the
floor.
However, you can change the mass at each node depending on the place of the node by setting “proportion
to influence area” in Option Menu. In this case, the mass at each node is determined from the following
equation:
Ai
Mi M floor (5-1-1)
A floor
where, Ai : influence area of node i, A floor : total area of the floor. Influence area of the node is different
depending on the place of the node as shown in Figure 5-1-1.
i j
X X
Mi
Ai, Mi Aj, Mj liy
G
IG
k lix MG
A k, M k
197
Figure 5-1-2. Influence area of the node (red)
700kN/8 50kN+12.5
= 87.5kN 87.5kN 112.5kN
= 62.5kN
87.5kN 112.5kN
87.5kN 62.5kN
87.5kN 62.5kN
87.5kN 112.5kN
198
In case of rigid floor assumption, in-plane freedoms at the nodes are dependent to the freedoms at the
center of gravity of the floor. Therefore, the mass at the center of gravity, M G , is,
M G M floor (5-1-2)
The radius of gyration at the center of gravity, IG, is obtained from the following equation:
N
I G M i lix2 liy2 (5-1-3)
i
where, N is the total number of the nodes at the floor. The radius of gyrations at other nodes are,
I i 0, i 1, , N (5-1-4)
The mass matrix is obtained as,
0 0
0
u xi Mi M i
u yi Mi M i
zi Mi M i
M (5-1-5)
xi Ii Ii
yi Ii I
i
zi Ii Ii
0
0 0
Since the mass matrix has only diagonal components, those components are saved in one-dimension vector.
For example, the mass vector of the structure in Figure 5-1-5 will be as follows:
M6
Node 6 0
M7
Node 7 0
M8
6 7
0
Node 8 0 10
M 8 9
9
0
Node 9 0
1 2
M 10
5
M 10 3 4
Node 10 I 10
199
5.2 Stiffness matrix
As shown in Figure 4-4-2, the global stiffness matrix K is formulated from element stiffness matrices.
5 8 7 10 11 13
Pz 8 5 k 5,5 k 5,8 k 5, 7 k 5,10 k 5,11 k 5,13 u z 8
P 8
k 8 ,8 k 8, 7 k 8,10 k 8,11 k 8,13 u z 9
z9
M y 8 7 k 7,7 k 7 ,10 k 7 ,11 k 7 ,13 y 8 Example of beam element
M y 9 10 k10,10 k10,11 k10,13 y 9
Px10 11 sym. k11,11 k11,13 u x10
M z10 13 k12,12 z10
Locate element stiffness
Element stiffness matrix
according to the freedom number
1 2 3 4 5 6 7 8 9 10 11 12 13
1
2
3
4
5 k 5, 5 k 5, 7 k 5,8 k 5,10 k 5,11 k 5,13
6
7 k 7,7 k 7 ,8 k 7 ,10 k 7 ,11 k 7 ,13
8 k 8,8 k 8,10 k 8,11 k 8,13
9
10 k10,10 k10,11 k10,13
11 sym. k11,11 k11,13
12
13 k13,13
200
5.3 Input ground acceleration
M i uxi X 0 uxi 1 0 0 uxi
M i uyi Y0
uyi 0 1 0
X 0 uyi X 0
M i zi Z0
zi
0
0
1
zi
Y0 M M U Y0 (5-3-1)
M M
I ixi xi 0 0 0 xi Z
0 Z 0
I iyi yi 0 0
0
yi
I izi
zi 0 0 0 zi
For example, the components of the matrix U of the structure in Figure 5-4-1 will be as follows:
201
5.4 Equation of motion
X 0
C u K u M u M U Y0 (5-4-1)
Z
0
Damping force
X0
M u C u K u M U Y0 P (5-4-2)
Z
0
202
5.5 Modal analysis
where K is the stiffness matrix and M is the mass matrix in the form;
m1 0 0
0 m2 0
M (5-5-2)
0 0 mn
Substituting into the equilibrium equation, the generalized eigen problem is obtained as,
This eigen problem yields the n eigen solutions , , , ,, , where the
2
1 1
2
2 2
2
n n
eigen vectors are M-orthonormalized as,
)
Let’s assume two different set of eigen solutions , , , .
i
2
i
2
j j
i
2
j 2 i M j 0
T
(5-5-8)
203
2) Modal decomposition of equilibrium equation
Defining a matrix whose columns are the eigenvectors and a diagonal matrix
2
12
2
1 2 n ,
2
2
(5-5-9)
2
n
u (t ) q(t ) (5-5-10)
Then,
Multiplying ,
T
where
m1
m2
M M , m T M
T
(5-5-13)
i i i
mn
k1
k2
K , k 2m
T T
K M M
i i i
kn
(5-5-14)
c1
c2
C C
T
(5-5-15)
cn
where, M , C and K are called generalized modal mass, modal damping and modal
stiffness matrix, respectively.
204
Therefore,
X0 (t )
where ri (t ) i P (t ) i M U Y0 (t )
T T
(5-5-18)
Z (t )
0
X0 (t )
qi (t ) 2hii qi (t ) i 2 qi (t ) i Y0 (t ) i , x X0 (t ) i ,yY0 (t ) i , x Z0 (t )
T
Z (t )
0
(5-5-19)
where
i T
T
T
(5-5-20)
i M i i M i
T i , x i ,y i ,z
i , x, y , z (5-5-21)
i M i
T
i , x , y , z is the coefficient when you decompose the vector U x ,y,z into mode vectors as,
U
n
)
M ,
T
Multiplying
Therefore,
M M U
1 T
x, y , z x ,y,z (5-5-24)
205
Equation (5-5-17) can be decomposed again as,
and
qi (t ) i , x xi (t ) i ,y yi (t ) i ,z zi (t ) (5-5-26)
n n n n
u (t ) q(t ) i qi (t ) i , x i xi (t ) i ,y i yi (t ) i ,z i zi (t )
i 1 i 1 i 1 i 1
(5-5-27)
M e ,i i 2 mi (5-5-32)
M e ,i i 2 mi i i T i i (5-5-33)
i M i i M i
i T
206
The sum of effective modal mass is,
n n
Therefore the ratio of effective modal mass to the total mass is used to judge the number of
significant modes that should be included in the analysis.
4) Initial condition
The initial conditions are obtained from Equation (5-5-10) as,
Therefore,
207
5.6 Damping matrix
1) Proportional damping
The mass-proportional damping and the stiffness-proportional damping are defined as,
C a0 M and C a1 K (5-6-1)
For a system with a mass-proportional damping, the generalized damping for the i-th mode in
Equation (5-5-12) is obtained as,
ci a0 mi , ci / mi 2hi i (5-6-2)
Therefore,
a0 1
a0 2hii , hi (5-6-3)
2 i
Similarly, for a system with a stiffness-proportional damping, the generalized damping for the
i-th mode is,
ci a1i mi , ci / mi 2hii
2
(5-6-4)
Therefore,
2hi a1
a1 , hi i (5-6-5)
i 2
hi
C a0 M C a1 K
a0 1 a1
hi hi i
2 i 2
In STERA_3D, you can select from the two types of stiffness-proportional damping.
One is the proportional damping using the initial stiffness matrix:
C 2h K 0 (5-6-6)
1
208
where, h: damping factor, 1 : circular frequency of the first natural mode, K 0 : the initial stiffness.
C 2h K p (5-6-7)
where, h: damping factor, 1 : circular frequency of the first natural mode, K p : the spontaneous
stiffness changing according to the nonlinearity of structural elements.
In the scene of the practical design of Japan, it is common to use the proportional damping using the
spontaneous stiffness matrix.
2) Rayleigh damping
A Rayleigh damping matrix is defined proportional to the mass and the initial stiffness
matrices as,
C a0 M a1 K 0 (5-6-8)
1 1 / 1 1 a0 h1
(5-6-10)
2 1 / 2 2 a1 h2
209
3) Damping matrix with a base isolation building
In an actual design practice for the base isolation buildings, it is common to assume zero viscous damping
for the base isolation floor. For example, in case of the stiffness-proportional damping, the damping
matrix is defined as:
C 2h K upper (5-6-12)
where, K upper : the stiffness matrix consisted with upper structures without base isolation elements.
C C pro C v (5-6-13)
where, C pro : the proportional damping matrix, C v : the global damping matrix formulated from
element damping matrices in the same manner of the global stiffness matrix.
210
5.7 Numerical integration method
Two numerical integration methods are prepared; one is the Newmark-β method with incremental
formulation using a step-by-step stiffness matrix, and another one is the Force correction method using a
step-by-step stiffness and a force vector together. In case it is difficult to define the step-by-step stiffness of
the element such as the case of using a viscous damper element, automatically the Force correction method
is selected.
a) Newmark-β method
The incremental formulation for the equation of motion of a structural system is,
where, M , C and K are the mass, damping and stiffness matrices. d i , vi , ai and
pi are the increments of the displacement, velocity, acceleration and external force vectors, that is,
d i d i 1 d i , vi vi 1 vi , ai ai 1 ai , pi pi 1 pi (5-7-2)
f is the unbalanced force vector in the previous step. Using the Newmark-β method,
1 1
d i M C K
t
2
2 t
(5-7-7)
1
p i M v i 1 a i C 1 v i 1 1a i t f
t 2 2 4
The equation can be rewritten as,
where,
211
Kˆ K 21t C 1t 2
M (5-7-9)
pˆ i pi M 1
vi 1 ai C 1 vi 1 1ai t f (5-7-10)
t 2 2 4
where, M , C and K are the mass, damping and stiffness matrices. d n1 , vn1 and a n1
are the displacement, velocity and acceleration vector at time step (n+1). f n is the restoring force
vector corresponding to d n , and f is the unbalanced force vector in the previous step. Pn1 is
the external force vector.
Using the average acceleration method,
where
Fn C vn 1 an t f n f K vn t 1 a n t 2 Pn1 (5-7-17)
2 4
212
M an1 C vn1 f n1 Pn1
from the following Figure,
f
f n1
fn K
d n d n1
Lan1 Fn
K n K n1
f n f n1
n = n+1
END
213
c) Operator Splitting method
The Operator Splitting (OS) method is a type of mixed integration method in which stiffness is divided into
linear and nonlinear (Nakashima, 1990). The explicit predictor-corrector method is employed for the
integration associated with the nonlinear stiffness, whereas the unconditionally stable Newmark-β method
is used for the integration associated with linear stiffness. The formulations are described as follows:
The equation of motion of a structural system is,
where, M , C and K are the mass, damping and initial tangential stiffness matrices. d n1 ,
~
vn1 and a n1 are the displacement, velocity and acceleration vector at time step (n+1). d n1 is
~
the predictor displacement vector, f n 1 is the restoring force vector corresponding to d n 1 , and
Pn1 is the external force vector.
Force
K d n 1
Corrector
K d~n1
f n1
Predictor
Displacement
d~
n 1 d n 1
214
Substituting Equations (5-7-19), (5-7,20) , (5-7,21) to (5-7,18),
M a n1 C vn 1 a n a n1 t K d n vn t 1 a n a n1 t 2
2 4
1 2
f n 1 K d n v n t a n t Pn 1
4
(5-7-22)
Solving for a n 1 ,
where
Step 3. Substitute f n1 to Equation (5-7-25) and solve the acceleration vector a n 1 , and obtain the
corrector displacement vector d n1 from Equations (5-7-20) and (5-7-21).
215
5.8 Energy
a) Equation of energy
X 0
M u C u K u M U Y0 P (5-8-1)
Z
0
For example, in case of a structure with a rigid floor in Figure 5-8-1, the displacement vector, u, consists
of 15 components (see RED numbers in Figure 5-8-1.)
u1
u
u 2 (5-8-2)
u15
1 4
6 7
2 5
7 15 10
10 3 6
13
8 9
8 11
9 14
12
1 2
5
3 4
Figure 5-7-1 Example of the freedom vector of a structure with a rigid floor
The equation of energy is derived by multiplying the velocity vector, u , and integrating by the time
T
range [0-t]:
t t t t
216
uT M u t uT C udt uT K u t uT Pdt
2
0
2
0
(5-8-4)
WK WD WP WI (5-8-5)
where,
WK
uT M u : Kinematic energy
2
t
WD u C udt : Damping energy
T
u K u
T
WP : Potential energy
2
t
WI u Pdt : Input energy
T
X 0
M u C u Qu, u M U Y0 P (5-8-6)
Z
0
where, Qu , u is the nonlinear restoring force vector. Then, the equation of energy can be derived as;
WK WD WP WI (5-8-7)
where,
WK
uT M u : Kinematic energy
2
t
W D u C udt
T
: Damping energy
0
t (5-8-8)
W P u Qu , u dt : Potential energy
T
0
t
W I u Pdt : Input energy
T
217
b) Decomposition of potential energy
We can decompose the restoring force vector into the restoring force of each member as,
t t n t T
n n
WP u Qu , u dt u qi u , u dt u qi u , u dt WP ,i
T T
(5-8-10)
0 0 i 1 i 1 0 i 1
where
t
WP ,i u qi u , u dt ;
T
potential energy of i-th member (5-8-11)
0
218
6. Nonlinear Static Push-Over Analysis
The static lateral load representing the earthquake force is applied at the center of gravity in each floor.
There are several formulas to define the load distribution along the height of the building. In “STERA 3D”
program, the following distributions are prepared:
1. Ai 2. Triangular 3. Uniform 4. UBC 5. Mode
(1) Ai distribution
In the “Building Standard Law” in Japan, the design shear force of i-th story, Qi, is defined as,
n
Qi C i w j , C i ZRt Ai C 0 (6-1-1)
j i
where,
Ci : design shear coefficient of i-th story,
wi : weight of i-th story,
Z: seismic zone factor,
Rt : vibration characteristic factor taking into consideration of soil condition,
Ai : lateral distribution of shear force coefficient,
C0 : design base shear coefficient (C0 =0.2 for serviceability limit, C0 =1.0 for safety limit)
If we set, Z=1.0 (Tokyo), Rt=1.0 (stiff soil, a short story building), C0=1.0 (safety design), the design shear
force distribution is simplified as,
n
Qi Ai w j (6-1-2)
j i
1 2T
Ai 1 i (6-1-3)
1 3T
i
where,
n n
i w j W , W w j : the ratio of weight upper than i-th story,
j i j 1
Fn Qn , Fi Qi Qi 1 i 1, , n 1 (6-1-4)
219
w6
F6 = Q6
Q6 = C6 w6
F5 = Q5 – Q6 w5
Q5 = C5 (w5 + w6)
w4
.
. n
. w3 Qi C i w j
. i i
. w2 C i ZRt Ai C 0
.
w1
F1 = Q1 – Q2
Q1 = C1 (w1 + w2 + ・・・+ w6)
n
Fi QB hi h
j (6-1-5)
j 1
where,
QB : base shear force
hi : the height of the i-th story from the ground
Fi
h6
h2
h1
220
(3) Uniform distribution
Uniform distribution is defined as:
Fi QB 1 n (6-1-6)
Fi
n
Fi QB Ft wi hi w h j
j (6-1-7)
j 1
Ft
Fi
h6
h2
h1
221
(4) Mode distribution
Mode distribution is defined as:
n
Fi QB wi1,i w j
1, j (6-1-9)
j 1
where,
Fi
1,i
222
6. 2 Capacity Curve
The Capacity Spectrum Method was proposed by Freeman [1978] as an approximate way to estimate the
maximum response of a structure under an earthquake ground motion. The concept was modified by
Kuramoto et.al [2000] to adopt the distribution of nonlinear story displacement as the first mode shape in
each calculation step. The method was adopted as one of the evaluation procedures in the Building
Standard Law, Japan.
The key concept of the Capacity Spectrum Method is to find out the intersection between the Demand
Spectra (= relationship between Sd (displacement spectra) and Sa (acceleration spectra)) and the Capacity
Curve (= nonlinear push-over curve of an equivalent single-degree-of-freedom system).
1400
1200
Demand Spectra
SA-SD( h=0.05)
1000
Sa (cm/sec 2)
Demand Spectra
800 SA-SD( h=0.14)
Capacity Curve
600
400
Performance Point
200
0
0 10 20
Sd (cm)
“STERA 3D” provides the menu in the static analysis to show the Capacity Curve based on the following
formula (Kuramoto et.al [2000]):
n n
mi i2 m i i
2
S a QB i 1
2 , Sd i 1
n
(6-2-1)
m
n
mi i i i
i 1 i 1
where,
mi : lumped mass in the i-th story
δi : component of the distribution of nonlinear story displacement in the i-th story
223
Sd
Fi δi
≈ M Sa
As schematically shown in Figure 8-2-2, the step-by-step results of nonlinear push-over analysis is used to
obtain the Capacity Curve of the equivalent SDOF system using Equation (8-2-1).
References
Freeman S. A. (1978), ”Prediction of Response of Concrete Buildings to Severe Earthquake Motion”,
Douglas McHenry International Symposium on Concrete and Concrete Structures, SP-55, American
Concrete Institute, Detroit, Michigan, pp. 589-605.
Kuramoto H., et.al. (200), “Predicting the Earthquake Response of Buildings using Equivalent Single
Degree of Freedom System”, 12th World Conference on Earthquake Engineering (12WCEE), Auckland
New Zealand,2000.2.
224
7. Others
The equivalent plane ( z ax by c )is obtained from the vertical displacement distribution by the least
square method:
z ax by c
zi x
L z i axi by i c
2
Minimize
z i xi x i2 x y x a
i i i
z i y i y y b
2
i i (7-1-1)
z i sym. n c
where,
n: the number of nodes in a floor
A story drift, D, can be divided into shear and flexural components as,
225
D = DS (shear) + DF (flexure) (7-1-2)
Assuming the distribution of floor deformation is expressed by an equivalent plane ( z ax by c ), the
flexural deformation, DF, can be expressed as,
DF = -a H : x-direction (7-1-3)
DF = b H : y-direction (7-1-4)
DS = D - DF (7-1-5)
D
DF DS
in x-direction
a
H
z
a z ax by c
DS DF in y-direction
b z
H
z ax by c b
y
In STERA 3D, the flexural deformation is calculated taking the average of the bottom floor angle and top
floor angle.
a i a i 1
D Fi Hi : x-direction (7-1-6)
2
bi bi 1
D Fi Hi : y-direction (7-1-7)
2
226
7.2 P-D effect
We consider equilibrium of the beam with a slight displacement with an axial load.
M + ΔM
EI v(x) M
F0 Δv
F0
x
Δx V
V
Δx
Assuming small deflection, the balance of moment on the small segment “Δx” gives
227
b) Geometric stiffness matrix of the beam with an axial load
We assume that the axial force is constant and compressive. From the general solution, Eq. (7-2-4),
at x 0
dv(0)
v(0) v1 c1 c 4 , 1 kc 2 c3 (7-2-6)
dx
Consequently, the deflected shape is
(1 C ) ( S ) c1 v1 1 L v 2
(7-2-10)
S
(1 C ) c 2 1 L 2 L
where,
C cos kL, S sin kL, kL (7-2-11)
Now we can rewrite the deflection function in terms of the nodal degrees of freedom. The moment and
shear force distributions can be obtained as
M ( x) EI
d 2v
dx 2
EI k 2 c1 cos kx k 2 c 2 sin kx (7-2-15)
d 3v
EIk 2 1 kc 2
dv
V ( x) EI 3
F0 (7-2-16)
dx dx
228
Calculating nodal loads, V (0) V1 , M (0) M 1 , V ( L) V1 , M ( L) M 1 , the stiffness matrix is
V1 2 S L(1 C ) 2S L(1 C ) v1
M
1 EI
2
L (C S ) L(1 C )
2
L2 ( S ) 1
(7-2-17)
V2 L3 2S L(1 C ) v 2
M 2 sym. L2 (C S ) 2
2 2C S
2 2 1 2 / 2 4 / 24 6 / 720 C 3 / 6 5 / 120 (7-2-18)
1 / 15 / 12
5 2
also
1 12
5 1 2 / 15
(7-2-19)
Substituting k 2 L2 F0 L / EI ,
2
k11
EI
3
12 F0 12 (7-2-21)
L L 10
In the same manner, we can expand for all the stiffness terms to get the stiffness matrix as
12 6L 12 6 L 36 3L 36 3L
4 L2 2
6 L 2 L F0 4 L2 3L L2
k EI3 (7-2-22)
L 12 6 L 30 L 36 3L
sym. 4 L2 sym. 4 L2
We can write as
k k E k G (7-2-23)
229
d) Implementation for beam element
A z
A B
B
uB
x
uA
A B y
M A 2 EI 2 1 A EI 4 L2 2 L2 A
M (7-2-24)
B L 1 2 B L3 2 L2 4 L2 B
u
1 1 A
A L 1 0
L
1
A
(7-2-25)
B 1
0 1 u B
L L B
1 1
QA L u
M L 1
1
1 A
0
0 4 L
2
2L L
2
A EI 1
L A
QB L3 1 1 2 L2
4 L2 1 1 u
0 1 B
L L L L B
M B 0 1
u 12 6 L u A
A 12
6L 6L 1 6L
4 L2 2 L2
1
EI L
1
L
0 EI
A 4 L2 6 L 2 L2 A
3
L 6 L 6 L 1 0 1 3
1 u B L
12 6 L u B
2
2L 4 L2 L L B sym. 4 L2 B
230
Therefore, the stiffness equation will be
QA 12 6L 12 6 L 36 3L 36 3L u A
M
A EI
4 L2 2
6 L 2 L F0 4 L2 3L L2 A
Q B L3 12 6 L 30 L 36 3L u B
M B sym. 4 L2 sym. 4 L2 B
231
e) Implementation for column element
B u xB
u yB
Bl' yB
xB yB
xB
l'
yA
yA
Al ' X
xA
A
xA u xA
u yA Y
M yA 2 EI 2 1 yA EI 4 L2 2 L2 yA
M in X-Z plane (7-2-27)
yB L 1 2 yB L3 2 L2 4 L2 yB
M xA 2 EI 2 1 xA EI 4 L2 2 L2 xA
M in Y-Z plane (7-2-28)
xB L 1 2 xB L3 2 L2 4 L2 xB
u
1 1 xA
yA L
1
L
0
yA
1 1 u in X-Z plane (7-2-29)
yB 0 1 xB
L L yB
u
1 1 yA
xA L 1 0
L xA
1 1 in Y-Z plane (7-2-30)
xB 0 1 u yB
L L xB
Note that the matrix for node movement in X-Z plane is different from that of beam element. The
force-deformation relationship in X-Z plane is then,
232
1 1
Q xA L u
M L 1 1 xA
yA EI 1 0 4 L
2
2 L2 L
1
L
0
yA
Q xB L3 1 1 2 L2
4 L2 1 1 u
0 1 xB
M yB L L L L yB
0 1
6 L 6 L 1 1 u
xA 12 6 L 12 6 L u xA
0 EI
2 2 L2
EI 4 L L
1
L yA 4 L2 6 L 2 L2 yA
3 u
L 6L 6 L 1 0
1 3
1 xB L
12 6 L u xB
2 yB
2L 4 L2 L L
sym. 4 L2 yB
(7-2-31)
Considering the difference of sign of stiffness matrix in X-Z plane, the geometric stiffness matrix will be
36 3L 36 3L
4 L2 3L L2
k xG F0 in X-Z plane (7-2-32)
30 L 36 3L
sym. 4 L2
36 3L 36 3L
4 L2 3L L2
F
k yG 0
30 L 36 3L
in Y-Z plane (7-2-33)
sym. 4 L2
Therefore, changing the order of vector component, the force-deformation relationship of column will be
Q xA u xA 36 36 3L 3L 0 0 0 0 0 0 0 0 u xA
Q
xB
u
xB
36 36
3L 3L 0 0 0 0 0 0 0 0 u xB
M yA yA 3L 3L 4 L 2
L 2
0 0 0 0 0 0 0 0 yA
M yB yB 3L 3L L
2
4L 2
0 0 0 0 0 0 0 0 yB
Q yA u yA 0 0 0 0 36 36 3L 3L 0 0 0 0 u yA
Q yB u yB F0 0 0 0 0 36 36 3L 3L 0 0 0 0 u yB
K
M xA
xA 30 L 0 0 0 0 3L 3L 4 L2 L2 0 0 0 0 xA
M xB xB 0 0 0 0 3L 3L L2 4 L2 0 0 0 0 xB
N zA zA
0 0 0 0 0 0 0 0 0 0 0 0 zA
N 0 0 0 0 0 0 0 0 0 0 0 0 zB
zB zB
M zA zA 0 0 0 0 0 0 0 0 0 0 0 0 zA
M
zB
zB 0 0 0 0 0 0 0 0 0 0 0 0 zB
233
u xA
u
xB
yA
yB
u yA
u
K K G yB (7-2-34)
xA
xB
zA
zB
zA
zB
where,
36 36 3L 3L 0 0 0 0 0 0 0 0
36 36 3L 3L 0 0 0 0 0 0 0 0
3L 3L 4 L 2
L 2
0 0 0 0 0 0 0 0
3L 3L L
2 2
4L 0 0 0 0 0 0 0 0
0 0 0 0 36 36 3L 3L 0 0 0 0
0 0 0 0 36 36 3L 3L 0 0 0 0
K G F0 (7-2-35)
30 L 0 0 0 0 3L 3L 4 L2 L2 0 0 0 0
0 0 0 0 3L 3L L2 4 L2 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
Then, applying translation of Equation (2-2-17), the constitutive equation of the column is;
P1 u1
P u
2 2
K
C (7-2-36)
Pn un
where,
234
7.3 Unbalance force correction
f f i 1 f i1 (7-3-1)
where, f i1 is the force on the nonlinear skeleton curve
The most preferable way to minimize the error is to adopt iterative calculations such as
Newton-Raphson method. However, this iteration may consume calculation time significantly.
Therefore, the following simple way is adopted to correct unbalance force:
d f / k (7-3-2)
where, k is the spring stiffness
2) Subtract unbalance displacement d from the increment displacement in the next step
calculation
i 1
f i 1
f
f i1
i
fi d
k
d
di d i 1
235
b) Unbalance force correction of MS model
For the Multi-spring model (MS model) of Column element, the sum of the unbalance forces of
nonlinear vertical springs in the member section is calculated as:
5 5
N f i f c ,i f s ,i (7-3-3)
i 1 i 1
5 5
D N k i N
i 1
k
i 1
c ,i k s ,i (7-3-4)
In the next step calculation, the increment displscement of each spring is ajusted as follows:
d i d i D (7-3-5)
where d i : increment displacement of i-th spring
d i : adjusted increment displacement of i-th spring
f 1
f 2 f 5
xs x
f 4
f 3
236
7.4 Calculation of ground displacement
In STERA_3D, the ground displacement is calculated from the ground acceleration data using FFT method
and filtering techniques based on the description in the following reference:
Reference: Yorihiko Osaki, “Introduction of Spectral Analysis of Earthquake Ground Motion”, Kajima
publishing corporation, 1981 (in Japanese)
N 1
1
Ck
N
x
m 0
m e i ( 2km / N ) k 0, 1, 2,, N 1 (7-4-1)
Nt N 1
y m x m dt
t mt
0
2 k 0
S k e i ( 2km / N ) m 0, 1, 2, , N 1 (7-4-3)
2v 0 N / 2 1
ImC k N 1C 0
S0 2
Nt k 0 k N
C 0
Sk 1 i cosk N i C k , S N k S N k
*
k 1, 2, , N / 2 1 (7-4-4)
N k
C 0
SN /2
N
237
The following band pass filter in frequency domain is applied to the coefficient S k .
0 0 f f L1
f f L1
0.5 0.5 cos f L1 f f L 2
L2
f f L1
h f 1 f L2 f f H1 (7-4-5)
f f H1
0.5 0.5 cos f H1 f f H 2
f H 2 f H1
0 fH2 f
h f
f ( Hz)
f L1 f L2 f H1 fH2
238
c) Calculation flow
The ground acceleration data is integrated twice to obtain displacement data. Band pass filter is applied
each time of the integration. The flow of calculation is summarized below:
x m (m 0, 1, 2, , N 1)
Ck k 0, 1, 2, , N 1
Sk k 0, 1, 2, , N 1
hk S k k 0, 1, 2, , N 1
IFFT Calculate the data of integration by Inverse Fourier transform
y m (m 0, 1, 2, , N 1)
239