Professional Documents
Culture Documents
STERA3D Technical Manual
STERA3D Technical Manual
Technical Manual
Version 7.0
1
UPDATE HISTORY
2
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 Modal analysis”, “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.
2017/03/20 STERA_3D Technical Manual Ver.5.6 is uploaded.
“7.4 Calculation of ground displacement” is modified changing band-pass filter.
2017/10/08 STERA_3D Technical Manual Ver.5.7 is uploaded.
Ground springs are added.
2017/10/24 STERA_3D Technical Manual Ver.5.8 is uploaded.
“4.6 Mass matrix corresponding to independent degrees of freedom” is added.
2019/02/12 STERA_3D Technical Manual Ver.6.0 is uploaded.
2019/05/20 Radiation damping for ground springs is added.
2019/07/25 External force by Wind is added.
2019/10/08 Buckling hysteresis of a brace is added.
2020/03/16 Pile foundation is included for ground springs.
Air spring is added for an external spring.
2021/10/10 STERA_3D Technical Manual Ver.7.0 is uploaded.
For RC column and RC wall, the nonlinear bending springs independent in x and y
directions are introduced.
For Steel beam, the nonlinear shear spring for hysteresis damper is introduced.
Damage indices of members are introduced.
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
2.11 Ground Spring
3. Nonlinear Element Models
3.1 Beam
3.1.1 RC Beam
3.1.2 Steel Beam
3.1.3 SRC Beam
Appendix 3.1:
A-1. Hysteresis of Degrading Trilinear Slip Model
3.2 Column
3.2.1 RC Column
3.2.2 Steel Column
3.2.3 SRC Column
Appendix 3.2:
Plastic theory model for nonlinear interaction between shear and axial strength
3.3 Wall
3.3.1 RC Wall
3.3.2 Steel Wall (Brace)
3.3.3 SRC Wall (Brace)
3.4 External Spring
3.4.1 Lift up spring
3.4.2 Air spring
3.5 Base Isolation
Appendix 3.5:
A-1. Hysteresis of LRB (Lead Rubber Bearing)
4
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
3.8 Ground Spring
3.8.1 Soil structure interaction
3.8.2 Cone model to calculate the static stiffness
3.8.3 Embedded foundation
3.8.4 Radiation damping
3.8.5 Complex stiffness with material damping
3.8.6 Impedance matrix
3.8.7 Pile foundation
3.8.8 Equivalent period and damping factor considering soil structure interaction
4. Freedom Vector
4.1 Node freedom
4.2 Freedom vector
4.3 Dependent freedom
4.4 Transformation matrix of dependent freedom
4.5 Stiffness matrix corresponding to independent degrees of freedom
4.6 Mass matrix corresponding to independent degrees of freedom
5. Equation of Motion
5.1 Mass matrix
5.2 Stiffness matrix
5.3 Modal analysis
5.4 Damping matrix
5.5 Input ground acceleration
5.6 External force by vibrator
5.7 External force by wind
5.8 Numerical integration method
5.9 Energy
6. Nonlinear Static Push-Over Analysis
6. 1 Lateral distribution of earthquake force
6. 2 Capacity Curve
7. Lumped Mass Model
7.1 Decomposition of shear and flexural deformation
7.2 Lumped mass model with shear and flexural stiffness
8. P-D effect
5
9. Unbalance force correction
10. Calculation of ground displacement
11. Damage Index
6
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
7
(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
θ yA θ yB x
A B
Displacement freedoms
y
Q zA Q zB
Local coordinate
M yA M yB
A B
Force freedoms
8
2. Constitutive Equation of Elements
2.1 Beam
τ ' yA
δx
M ' yA M ' yB
A B
Nx
τ ' yB
elastic element
θ ' yA φ yA
M ' yA M ' yB M ' yA M ' yB
A B A B
N 'x θ ' yB φ yB
nonlinear bending springs
l'
θ ' yA τ ' yA φ yA η yA η yA
M ' yA M ' yB
θ ' yB = τ ' yB + φ yB + η yB A η yB B
δ ' δ ' 0 0
x x
nonlinear shear springs
l' l'
3EI − 0
6 EI y
τ ' yA y
M ' yA
l' l'
τ ' yB = − 0 M ' yB (2-1-1)
δ ' 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,
φ yA f yA 0 M ' yA
= (2-1-2)
φ 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.
9
The force-deformation relationship of shear spring is
Qz = k z sz or sz = (1 k z ) Qz
M' yA
Qx = [1 l ' 1 l ']
M' yB
The end rotational displacement due to shear deformation is obtained as,
1 1
η yA 1 l' 1 l' 1 l' M' yA k sz l'2 ksz l'2 M' yA
=
= k z ) Qz (1 k z ) [1 l' 1 l'] =
sz 1 l' (1 =
η
yB 1 l ' 1
l ' M' yB 1 1 M' yB
k l'2 ksz l'2
sz
(2-1-3)
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'2 − + 2
0
6 EI y k sz l'
y sz
l' 1
=[ fB ] f yB + + 2
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,
10
Including rigid parts and node movement
Including rigid parts and node movement as shown in Figure 2-1-2, the rotational displacement vector is,
u
1 1 1 1 zA
θ +
yA l ' zAu + λ θ − u + λ θ
B yB − 1 + λ λ B u
A yA zB A
zB
= l'
= l' l'
(2-1-7)
1 1 1 1 θ
θ yB + u zA + λ Aθ yA − u zB + λ Bθ yB − λA 1 + λ B yA
l' l' l' l' θ yB
θ yB
θ ' yA θ yA τ
θ yB
θ yA τ θ ' yB
u zB
Z
u zB − λ B l 'θ yB
u zA u zA + λ A l 'θ yA
A B X
λ Al ' l' λB l '
θ ' yA θ ' yA
θ ' yA 1 0 0 0
θ '
= 0 1 0 0 θ ' yB = [n ]θ ' yB (2-1-9)
yB B
δ ' 0 0 − 1 1 δ xA δ xA
x δ δ xB
xB
u zA u zA
1 1 u
θ ' yA l ' − l ' 1 + λ A λB 0 0 u zB
zB
θ ' 1 1 θ θ
yB
= l' − l' λ A 1 + λB 0 0 yA = [Λ B ] yA (2-1-10)
δ θ yB θ yB
xA 0 0
δ xB 0 0 1 0 δ xA
δ xA
0 0
0 0 0 1
δ xB δ xB
11
Out of plane deformation of beam
If we consider out-of-plane deformation of beam in case of flexible floor, as shown in Figure 2-1-4, the
rotational displacement vector is,
u
1 1 1 1 yA
θ zA − l ' u yA + λ Aθ zA + l ' u yB + λBθ zB − l ' 1 + λA λB u
l' yB
= = 1 (2-1-11)
1 1 1
θ zB − u yA + λ Aθ zA + u yB + λBθ zB − λ A 1 + λB θ zA
l' l' l' l' θ zB
u yA θ zA u yB
θ ' zA τ
θ zB
τ θ zB θ ' zB
θ zA
u
1 1 zA
− l ' 1 + λA λB u zB
l'
θ ' yA 1 1
θ
θ ' − λ A 1 + λB yA
yB l ' l' θ yB
θ 'zA 1 1
= − 1 + λA λB u yA (2-1-12)
θ '
zB l' l' u yB
δ xA 1 1
− λ A 1 + λB θ zA
l' l'
δ xB θ zB
1
δ xA
1
δ xB
12
From global node displacement to element node displacement
Transformation from global node displacements to element node displacements is,
u zA
u
zB u1
θ yA u
2
= [TixB ] (2-1-13)
θ yB
δ xA un
δ xB
The component of the transformation matrix, [TixB ] , is discussed in Chapter 4 (Freedom Vector).
u1 u1
θ ' yA u u
2 2
θ ' yB = [nB ][Λ B ][TixB ] = [TxB ] (2-1-14)
δ '
x un un
φ yA f yA
0 0 M' yA
φ yB = 0 f yB 0 M' yB (2-1-15)
s 1 1
N'
z 0 x
ksz l' ksz l'
13
In case of Y-direction beam
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 (2-1-16)
z
Y − Beam 0 0 1 Z Global
Therefore
u zA 1 u zA u zA
u u
zB
1 0 u zB zB
θ yA −1 θ xA θ xA
= = [s B ] (2-1-17)
θ yB −1 θ xB θ xB
δ xA 0 1 δ yA δ yA
δ xB Y − Beam 1 δ yB Global δ yB Global
Transformation from the global node displacement to the element node displacement is,
u zA
u
zB u1
θ xA u
= TiyB [ ] 2
(2-1-18)
θ xB
δ yA u n
δ yB
Transformation from the global node displacement to the element face displacement is,
u1 u1
θ ' yA u u
[ ]
θ ' yB = [n B ][Λ B ][s B ] TiyB
2
[ ]
= T yB
2
(2-1-19)
δ '
x u n u n
14
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 (2-1-20)
Pn un un
For Y-beam,
P1 u1 u1
P u u
2
= TyB [ ] T
[ ] [ ]
[k B ] TyB 2 = K yB 2
(2-1-21)
Pn un un
15
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
θ ' xB
θ ' yB
N 'z Z
l' = +
T 'z
θ ' yA θ ' xA X
A A A
M ' yA Y
M ' yA
M ' xA M ' xA
l' l'
−
τ ' xA 3EI x 6 EI x M ' xA
=
l ' M ' xB
in Y-Z plane (2-2-2)
τ ' xB − l '
6 EI x 3EI x
The axial displacement is,
l'
δ ' 'z = N 'z (2-2-3)
EA
The torsion angle by torque force is,
l'
θ 'z = T 'z (2-2-4)
GI z
where, G and I z are the shear modulus and the pole moment of inertia of the cross-sectional area.
16
Force-displacement relationship for nonlinear bending springs
Case 1: In the case that bending springs in x and y directions are independently defined
M ' yB M ' xB
B B
φ yB φxB
φ yA φxA
A A
M ' yA
M ' xA
The rotational displacement vector of the nonlinear bending spring is defined independently,
where, f yA , f xA , f yB , and f xB are the flexural stiffness of nonlinear bending springs at both ends of
the element.
It can be expressed as
φ yA M ' yA f yA
[ ]
φ xA = f pA
M ' xA , f pA =
f xA
at end A (2-2-7)
ε N' 0
zA zA
φ yB M ' yB f yB
φxB = f pB M 'xB , f pB = f xB at end B (2-2-8)
ε N' 0
zB zB
17
Case 2: In the case that nonlinear interaction between moment and axial components is considered
φ yA M ' yA
N ' zB , ε zB
[ ]
φ xA = f pA
M ' xA at end A (2-2-9)
ε N' M ' xB , φ xB
zA zA
M ' yB , φ yB
φ yB M ' yB
[ ]
φ xB = f pB
M ' xB at end B (2-2-10)
B
ε N'
zB zB
Rearrange the order of the components of the displacement vector and change the node axial displacements
into the relative axial displacement,
φ yA φ yA
φ yA 1 0 0 0 0 0 φ
φ 0 φ xA
yB 0 0 1 0 0 xA
ε zA ε zA
φ xA = 0 1 0 0 0 0 = n p [ ] (2-2-12)
φ 0 φ yB φ yB
0 0 0 1 0
xB φ xB φ xB
ε z 0 0 − 1 0 0 1
ε zB ε zB
φ yA M ' yA M ' yA
φ M ' M '
[ ][ ] [ ][ ]
yB f pA 0 yB yB
φ xA = n p
T
[ ]
n M ' = f p M ' xA
f pB p xA
(2-2-13)
φ 0 M' M '
xB
xB xB
ε z N ' z N ' z
18
Force-displacement relationship for nonlinear shear springs
Case 1: In the case that shear springs in x and y directions are independently defined
X-Z plane Y-Z plane
M ' yB M ' xB
B B
η yB
η xB
ksx
Qx , sx ksy
Qy , s y
η yA η xA
A A
M ' yA
M ' xA
Figure 2-2-4 Element model for column
M' yA
Q'x 1 l ' 1 l ' 0 0 M' yB
= (2-2-15)
Q' y 0 0 1 l ' 1 l ' M'xA
M'xB
The end rotational displacement due to shear deformation is obtained as,
19
1 1
k l '2 0 0
ksx l '2
sx
1 1
k l '2 0 0
ksx l '2
[ f s1 ] =
sx
where (2-2-17)
1 1
0 0
ksy l '2 ksy l '2
1 1
0 0
ksy l '2 ksy l '2
Case 2: In the case that nonlinear interaction between shear and axial components is considered
N 'zB
M 'xB
M ' yB
B
N z , ε sz
Qx , sx
Qy , s y
A
M ' yA
M 'xA
N 'zA
Q'x sx l ' η y η y
Q' y =
ksp s y ksp η x k 'sp η x
= l '= (2-2-18)
1 ε sz
N' z ε sz ε sz
η y Q'x
−1
=η x =
f sp Q' y , f sp k 'sp (2-2-19)
ε N'
sz z
20
From the relationship between shear force and moment,
M' yA M' yA
Q'x 1 l ' 1 l ' 0 0 0 M' yB M'
yB
=
Q' y 0 0 1 l ' 1 l ' 0 M'xA [ Ls ] M'xA (2-2-20)
N' 0 0 0 0 1 M'xB M'
z xB
N'z N'z
The end rotational displacement due to shear deformation is obtained as,
η yA 1 1 M' yA M' yA
1 η 1
η yB y M' yB M' yB
η xA 1 η x sp [ Ls ] M' xA [ f s 2 ] M'xA
= = 1 f= (2-2-21)
η M' M'
xB 1 ε sz 1 xB xB
ε sz 1 1 N'z N'z
where
1 f11 f12 f13
1 f f12 f13 f f12 f13 1 l ' 1 l ' 0 0 0
11 11
1 f 21 f 22
f 23 [ Ls ] = f 21 f 22
f 23 0 0 1 l ' 1 l ' 0
1 f31 f32 f33 f 21 f 22 f 23 0 0 0 0 1
1 f31 f32 f33
(2-2-22)
f11 l ' f11 l ' f12 l' f12 l' f13
f l' f l' f l' f12 l' f13
11 11 12
η yA M' yA
η M'
yB yB [ f ] 0
η xA = [ f s ] M'xA , [ f s ] = s1 (case 1), [ fs ] = [ fs2 ] (case 2) (2-2-23)
η M' 0 0
xB xB
esz N'z
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,
21
θ ' yA τ ' yA ϕ yA η yA M' yA
θ ' τ ' ϕ η M'
yB yB yB yB yB
θ 'xA τ 'xA ϕ xA η M'
θ '
=
τ '
+
ϕ
+ xA
η
= [ fC ] M' xA (2-2-24)
xB xB xB xB xB
δ 'z δ ''z εz ε sz N 'z
θ 'z θ 'z elastic element 0 bending spring 0 shear spring T 'z
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
[ f s ] 0
(2-2-25)
0 0 shear spring
By taking the inverse matrix of [ f C ] , the constitutive equation of the column element is obtained as,
22
Including rigid parts and node movement
Change relative axial displacement and torsion displacement into node displacement,
θ ' yA θ ' yA
θ '
θ ' yA 1 θ ' yB yB
θ ' 1 0 θ ' θ ' xA
yB xA
θ ' xA 1 θ ' xB θ ' xB
= = [nC ] (2-2-27)
θ ' xB 1 δ zA δ zA
δ 'z 0 −1 1 δ zB δ zB
θ 'z − 1 1 θ zA θ zA
θ zB θ zB
Including rigid parts and node movement,
u xA u xA
1 1
u xB u
− 1 + λA λB xB
θ ' yA l ' l'
θ yA θ yA
θ ' − 1 1
λA 1 + λB
0 θ yB
yB l ' l' θ yB
θ ' xA 1 1 u yA u yA
− 1 + λA λB
l' l' u
θ ' xB u
= [Λ C ] yB
yB
= 1 1
δ zA − λA 1 + λB θ xA θ xA
l' l' θ
δ zB 1 θ xB
xB
θ zA 1 δ zA δ zA
θ δ
zB 0 1 δ zB zB
1 θ zA θ zA
θ θ
zB zB
(2-2-28)
B u xB
u yB
λB l '
θ yB
θ xB θ ' yB
l'
θ 'xB
θ ' yA Z
θ 'xA
θ yA
λ Al ' X
θ xA
A u xA
u yA
Y
23
From global node displacement to element node displacement
Transformation from global node displacement to element node displacement is;
u xA
u
xB
θ yA
θ yB
u yA u1
u
u yB 2
= [T ]
iC (2-2-29)
θ xA
θ xB un
δ zA
δ zB
θ zA
θ
zB
The component of the transformation matrix, [TiC ] , is discussed in Chapter 4 (Freedom Vector).
θ ' yA
θ '
yB u1 u1
θ ' xA u
u 2 2
= [n C ][Λ C ][T ]
iC = [T ]
C (2-2-30)
θ ' xB
δ 'z u n u n
θ ' z
Constitutive equation
Finally, the constitutive equation of the column is;
P1 u1
P u
2 2
= [K C ] (2-2-31)
Pn un
where,
24
Transformation matrix for nonlinear spring displacement
The nonlinear spring displacement vector is obtained from
(2-2-33)
and
M' yA M' yA
M' yA M' M'
yB yB
η y Q'x M' yB
M' M'
f sp Q' y f sp [ =
η x =
= Ls ] M'xA f sp [ Ls ] [ I ] =
0 xA [ f sC ] M' xA
ε M' M'xB xB
sz N' z xB
N 'z N 'z
N'z
T 'z T 'z
(2-2-34)
25
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.
θ ' xB θ ' xB Z
θ ' yB
l' X
θ ' xA θ ' yA θ ' xA
Y
l' l'
−
τ ' yAc 3EI c 6 EI c M ' yAc
=
l ' M ' yBc
in wall panel (2-3-1)
τ ' yBc − l '
6 EI c 3EI c
l' l'
−
τ ' xA1 3EI 1 6 EI 1 M ' xA1
=
l ' M ' xB1
in side column 1 (2-3-2)
τ ' xB1 − l '
6 EI 1 3EI 1
l' l'
−
τ ' xA2 3EI 2 6 EI 2 M ' xA2
=
l ' M ' xB 2
in side column 2 (2-3-3)
τ ' xB 2 − l '
6 EI 2 3EI 2
The axial displacement is,
l'
δ ' ' zc = N ' zc (2-3-4)
EA
26
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 , φ xA1 M ' xA2 , φ xA2
M ' yAc , φ yAc
N zAc , ε 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,
27
φ yAc M ' yAc
φ M '
xA1 xA1
φ xA2 M ' xA2
[ ]
ε zAc f pA
0 N ' zAc
=
φ 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,
φ yAc φ yAc
φ yAc 1 φ
φ φxA1 xA1
yBc 1
φxA2 φxA2
φxA1 1
ε zAc ε zAc
φxB1 = 1 φ = n p φ [ ] (2-3-8)
φ 1 yBc yBc
xA2
φxB1 φxB1
φxB 2 1
ε φxB 2 φxB 2
zc −1 1 ε
ε zBc zBc
M' yAc
Q'xc = [1 l ' 1 l ']
M' yBc
The end rotational displacement due to shear deformation is obtained as,
28
1 1
η yAc 1 l ' 1 l ' k l '2 ksc l'2 M'xAc
=
η
= s
xc (1 k
= sc ) Q' xc
sc in wall panel (2-3-10)
yBc 1 l ' 1 l ' 1 1 M'xBc
k l'2 ksc l'2
sc
1 1
η xA1 ks1l' ks1l'2 M'xA1
2
1 1
η xA 2 ks 2l' ks 2l'2 M'xA 2
2
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,
29
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
1 1
k l '2 ksc l'2
sc
1
ksc l'2
1 1
ks1l'2 ks1l'2
1 (2-3-14)
ks1l'2
1 1
ks 2l'2 ks 2l'2
1
sym.
ks 2l'2
0 shear spring
By taking the inverse matrix of [ fW ] , the constitutive equation of the column element is obtained as,
30
Including rigid parts and node movement
Change relative axial displacement and torsion displacement into node displacement,
(2-3-17)
From global node displacement to element node displacement
Transformation from the center displacements to the node displacements is,
θ yc
δ z 2 − δ z1
θ yc =
δ z1 δ zc δ z2 w
δ +δ
δ zc = z1 z 2
2
w
31
1
u xAc u xA1 u xA1
u 1 δ δ
xBc 1 1 zA1 zA1
θ yAc − w w δ
zA 2 δ zA 2
1 1 u
θ yBc − xB1 u xB1
u yA1 w w δ zB1 δ zB1
1 δ
u yB1 1 zB 2 δ zB 2
θ u yA1 u
= [DW ]
yA1
xA1
= 1
θ
xB1 u
yB1 u
yB1
1
u yA 2 θ xA1 θ xA1
1
u yB 2 1 θ xB1 θ xB1
θ u u
xA 2 1 yA 2 yA2
θ xB 2 u yB 2 u yB 2
δ 1 θ
θ
zAc 0.5 0.5 xA2 xA2
δ zBc θ xB 2 θ xB 2
0.5 0.5 (2-3-18)
Transformation from the global node displacements to the element node displacements is;
u xA1
δ
zA1
δ zA 2
u xB1
δ zB1
δ zB 2 u1
u u
2
= [TixW ]
yA1
(2-3-19)
u yB1
θ xA1 u n
θ xB1
u
yA2
u yB 2
θ
xA2
θ xB 2
The component of the transformation matrix, [TixW ] , is discussed in Chapter 4 (Freedom Vector).
32
θ ' yAc
θ '
yBc u1 u1
θ ' xA1 u u
2 2
θ ' xB1 = [nW ][Λ W ][DW ][TixW ] = [TxW ] (2-3-20)
θ '
xA 2
u n u n
θ ' xB 2
δ'
zc
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 (2-3-21)
z
Y −Wall 0 0 1 Z Global
Therefore
33
u xA1 1 u yA1 u yA1
δ 1 δ δ
zA1 zA1 zA1
δ zA 2 1 δ zA 2 δ zA 2
u
u xB1 1 yB1 u yB1
δ zB1 1 δ zB1 δ zB1
δ zB 2 1 δ zB 2 δ zB 2
u −1 u xA1 u
xA1
yA1
= u = [ε ]
W
u yB1 −1 xB1 u xB1
θ xA1 1 θ θ yA1
yA1
θ xB1 1 θ yB1 θ yB1
u u u xA2
−1 xA2
yA2
u yB 2 −1 u xB 2 u xB 2
θ
θ
1 θ yA2
xA2 yA2
θ θ
1 yB 2 Global θ yB 2 Global
xB 2 Y −Wall
(2-3-22)
Transformation from the global node displacement to the element node displacement is;
u yA1
δ
zA1
δ zA 2
u yB1
δ zB1
δ zB 2 u1
u u
xA1
u
[ ]
= TiyW
2
(2-3-23)
xB 1
θ yA1 u n
θ yB1
u xA2
u xB 2
θ
yA2
θ yB 2
34
Transformation from the global node displacement to the element face displacement is,
θ ' yAc
θ '
yBc u1 u1
θ ' xA1 u u
[n ][ ][D ][ε ][T ] 2
θ ' xB1 = W Λ W W W ixW = T yW [ ]
2
(2-3-24)
θ '
xA 2
u n u n
θ ' xB 2
δ'
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)
35
ϕ yAc M ' yAc M ' yAc
ϕ M ' M '
xA1 xA1 xA1
ϕ xA 2 M 'xA 2 M 'xA 2
f pA 0 N 'zAc
ε zAc N 'zAc
= [ f mW ] M '
ϕ yBc 0
f pB M ' yBc
yBc
ϕ xB1 M 'xB1 M 'xB1
ϕ xB 2 M 'xB 2 M 'xB 2
ε N' N'
zBc zBc zBc
M ' yAc
M '
xA1
M '
sxc 1 ksc Q'xc 1 ksc 1 l ' 0 1 l' 0 xA 2
= N'
s y1 1 ks1 Q' y1 1 ks1 1 l' 0 1 l' 0 zAc
M'
sy 2 1 ks 2 Q' y 2 1 k s 2 1 l' 0 1 l ' 0 yBc
M 'xB1
M 'xB 2
N'
zBc
f pA 0
ϕ yAc
ϕ
xA1 M ' yAc M ' yAc
ϕ xA 2 M ' M '
xA1 xA1
ε zAc M 'xA 2 M 'xA 2
ϕ yBc 0 f pA
N 'zAc = f N 'zAc
ϕ xB1 = M ' yBc pW M ' yBc
ϕ
xB 2
M 'xB1 M 'xB1
ε zBc 1 0
1
0 M '
s ksc l' ksc l' xB 2 M 'xB 2
xc
1 1 N 'zBc N'
zBc
s y1 0 0
k s1l' ks1l'
sy2
1 1
0 0
k s 2 l' k s 2 l'
(2-3-29)
36
Furthermore, in the same way as Equation (2-3-8),
M ' yAc
M ' M ' yAc
xA1 M '
M ' xA2 yBc
M ' xA1
N ' zAc
= np [ ] T
M ' xB1 (2-3-30)
M ' 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,
ϕ yAc
ϕ
xA1
ϕ xA 2 M ' yAc
M '
ε zAc yBc
ϕ yBc M 'xA1
T
ϕ xB1 = f pW n p M 'xB1 (2-3-31)
ϕ M '
xB 2 xA 2
ε zBc M 'xB 2
η N'
yc zc
η x1
ηx2
37
In case of direct input wall
Direct input wall model is defined as a line element with a nonlinear shear spring and a nonlinear bending
spring in the middle of the element as shown in Figure 2-3-1.
M ' yBc
kn Z
ks h X
kb
Y
M ' yAc
This model can be used as an alternative model so called the lumped mass model representing the restoring
force characteristics of each layer in the analysis of high-rise building as shown below. The detail of the
model is described in Chapter 7.1
38
STERA_3D adopts the formulation to have nonlinear shear and bending springs at the middle of the
element.
δ sx
B u xBc
Z
θ yBc ks
h φ yb θ yAc X
u xAc kb Y
A
Figure 2-3-7 Nonlinear bending and shear springs
Force-displacement relationship
The relationship between the displacement and force of the springs is,
Qxs k s 0 0 δ xs δ xs
= M yb
=
0 kb 0 φ yb [ kW ] φ yb (2-3-32)
N
zn 0 0 kn ε zn ε zn
u xAc u xAc
u u
δ xs −1 1 0 h 0 0 xBc xBc
θ yAc
θ yAc
φ yb =
0 0 −1 1 0 0
θ [
= Λ W ]
θ
(2-3-34)
ε 0 0 0 0 −1 1 yBc yBc
zn δ δ zAc
zAc
δ zBc
δ zBc
39
From global node displacement to element node displacement
Transformation from the center displacements to the node displacements is,
θ yc
δ z 2 − δ z1
θ yc =
δ z1 δ zc δ z2 w
δ +δ
δ zc = z1 z 2
2
w
1
u xAc 1 u xA1 u xA1
u δ
xBc 1 1 δ zA1 zA1
θ yAc − δ δ zA 2
w w
= zA 2 [ DW ] u (2-3-35)
θ yBc
−
1 1 u xB1 xB1
δ zAc w w δ zB1 δ zB1
0.5 0.5
δ zBc δ zB 2 δ zB 2
0.5 0.5
Transformation from the global node displacements to the element node displacements is;
u xA1
δ
zA1 u1
δ zA 2
u 2
[ ixW ]
= T (2-3-36)
u xB1
δ zB1 un
δ zB 2
The component of the transformation matrix, [TixW ] , is discussed in Chapter 4 (Freedom Vector).
u1 u1
δ xs u
2 u 2
yb [ W ][ W ][ ixW ] [ xW ]
φ =
Λ D T T
= (2-3-37)
ε
zn
un un
40
In case of Y-direction wall
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 (2-3-38)
z
Y −Wall 0 0 1 Z Global
Therefore
Transformation from the global node displacement to the element node displacement is;
u yA1
δ
zA1 u1
δ zA 2 u
2
= TiyW (2-3-40)
u yB1
δ zB1 un
δ zB 2
Transformation from the global node displacement to the element face displacement is,
u1 u1 u1
δ ys u u u
2 2 2
xb [ W ][ W ][ W ][ ixW ] [ W ][ W ][ ixW ] yW
φ =
Λ D ε T =
Λ D T
=T (2-3-41)
ε
zn un un un
41
Constitutive equation
Finally, the constitutive equation of the wall is;
P1 u1
P u
2 2
= [K xW ] (2-3-42)
Pn u n
where,
For Y-wall,
P1 u1
P u
2
[
= K yW ]
2
(2-3-44)
Pn u n
where,
[K ] = [T ] [k ][T ]
yW yW
T
W yW (2-3-45)
42
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', δ '
Z
h
1 2 X
A
w Y
Force-displacement relationship
~ ~ ~ ~
f x4 , u x4 f x3 , u x3
4 3
N1 , δ 1
x N2, δ2
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,
N 1 = k1δ 1 (2-4-1)
N 2 = k 2δ 2 (2-4-2)
Replacing with the nodal force and displacement in local coordinate along the element,
~ ~
N 1 = − f1x = f 4 x , δ 1 = u~4 x − u~1x (2-4-3)
~ ~
N 2 = − f 2 x = f 3 x , δ 1 = u~3 x − u~2 x (2-4-4)
43
In a matrix form,
u~1x
u~
1y
~
u1x u~2 x
~
δ 1 − 1 0 0 1 u~2 x − 1 0 0 0 0 0 1 0 u 2 y ~
=
δ 0 − 1 1 0 u~ = 0 0 − 1 0 1 0 0 0 u~ = [nb ]{u } (2-4-5)
2 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
T N
= [nb ] 1 (2-4-6)
f 3x 0 1 N 2 f x3 0 1 N 2 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 θ + f z1 sin θ
~ for Brace 1 (2-4-7)
f y1 = − f x1 sin θ + f z1 cos θ
and
~
f x 2 = − f x 2 cos θ + f z 2 sin θ
~ for Brace 2 (2-4-8)
f y 2 = − f x 2 sin θ − f z 2 cos θ
Eq. (2-4-8) can be also obtained from the Eq. (2-4-7) by replacing θ by (π − θ ) and using the
formulas sin (π − θ ) = sin θ , cos(π − θ ) = − cos θ .
44
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 θ f z1 sin θ
θ f x1 cos θ f z 2 sin θ θ
f z 2 cos θ
− f x 2 cos θ
θ f x1 θ
− f x2
f x1 sin θ − f x 2 sin θ
(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 x2 f x2
−c s f x2
~
f z2 −s −c f z2 f
~ =
= [Cb ] z 2 (2-4-9)
f x3
f x3
−c s
f x3
~ f z3
f −s −c 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 θ = , s = sin θ =
l l
Since [C b ][C b ] = I ,
T
[Cb ] is an orthogonal matrix, therefore,
45
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 θ − u~z1 sin θ
for Brace 1 (2-4-11)
u = u~ sin θ + u~ cos θ
z1 x1 z1
and
u x 2 = −u~x 2 cos θ − u~z 2 sin θ
for Brace 2 (2-4-12)
u = u~ sin θ − u~ cos θ
z2 x2 z2
Eq. (2-4-12) can be also obtained from the Eq. (2-4-11) by replacing θ by (π − θ ) and using the
formulas sin (π − θ ) = sin θ , cos(π − θ ) = − cos θ .
uz4 u z3
u~4 u~3
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 θ − u~z 2 cos θ
u~x 2
u z1 = u~x1 sin θ + u~z1 cos θ
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~ u~x 2
x2
~ ~
u z 2 s −c u z 2 T u z 2
= = [C b ] ~
u u~x 3
x 3
− c − s
~ u x 3
u z 3 s −c u z 3 u~z 3
~
u x 4
c − s u~x 4 u x 4
~
u
s c u z 4 Global u~
z 4 Local z4
46
The stiffness matrix of brace element is,
{N } = [k~ ]{δ }
N 1 k1 0 δ 1
= or (2-4-13)
N 2 0 k 2 δ 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
{δ } = [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,
47
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
{δ } = [T yBr ]
2
, [T ] = [n ][C ][T ]
yBr b b iyBr (2-4-23)
u n
48
Constitutive equation
The constitutive equation of the Y-direction brace is;
P1 u1
P u
2
[
= K yBr ]
2
(2-4-24)
Pn u n
where,
[K ] = [T ] [k ][T ]
yBr yBr
T
Br yBr (2-4-25)
49
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
[k~]L
k1, L
=
0
k 2, L
, [nb ] =
− 1 0 0 0 0 0 1 0
0 0 0 − 1 0 1 0 0 0
c s
− s c
−c s
[Cb ] =
−s −c , c = cosθ =
(w / 2) , s = sin θ =
h
−c s l' l'
−s −c
c s
− s c
50
For the right half part, in the same way, the stiffness equation of brace element is,
where
{ f }R = { f x5 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 z 5 [1] u
{u}L = = z 3 = [DL ] z 3
u x 3 [1] u x 4 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 z 2 [1] u
{u}R = = z 3 = [DR ] z 3
u x 6 [1] u x 4 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
51
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 (2-4-28)
u x 6 0 0 0 0 1 / 2 0 1 / 2 0 u x 3 u x 3
u z 3
u z 6 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 (2-4-29)
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
52
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 z 5 u z 3 u z 2 u z 2 u z 3 u z 2
{u}L = = [DL ] = [DL ][TCh ] , {u}R = = [DR ] = [DR ][TCh ]
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
(2-4-30)
Finally the force-displacement relationship of Cheveron brace is,
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 (2-4-32)
f x3 u x 3
f z3 u z 3
f x4 u x 4
f u
z 4 Local z4
53
2.5 External Spring
δ zB
δ yA Z
B
A B A
δ xA δ xB X
B
Y
δ yB A
δ zA
δ ' x = u xB − u xA
δ ' y = u yB − u yA (2-5-2)
δ ' z = δ zB − δ zA
Therefore
u xA u xA
u u
xB xB
u u
{δ ' x } = [− 1 1 0 0 0 0] yA = [n xE ] yA (2-5-3)
u yB u yB
δ zA δ zA
δ zB δ zB
u xA u xA
u u
xB xB
u yA
{δ ' y } = [0 0 − 1 1 0 0]uu yA = n yE [ ] (2-5-4)
yB u yB
δ zA δ zA
δ zB δ zB
54
u xA u xA
u u
xB xB
u yA u
{δ ' z } = [0 0 0 0 − 1 1] = [n zE ] yA (2-5-5)
u yB u yB
δ zA δ zA
δ zB δ zB
u xA
u
xB u1
u yA u
2
= [TiE ] (2-5-6)
u yB
δ zA u n
δ zB
The component of the transformation matrix, [TiE ] , is discussed in Chapter 4 (Freedom Vector).
u1 u1
u u
{δ '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,
55
2.6 Base Isolation
δ zB
B Z
u xB
u yB
l X
A u xB
Y
u yA
δ zA
Figure 2-6-1 Element model for base isolation
Q' x
= k pBI [ ]δδ ''
x
(2-6-1)
Q' y y
Including the axial stiffness,
Q ' x
[
k pBI ] 0 δ ' x
δ ' x
Q ' y = EA δ ' y = [k BI ]δ ' y (2-6-2)
δ ' 0
z l'
δ ' z
δ '
z
δ ' x = u xB − u xA
δ ' y = u yB − u yA (2-6-3)
δ ' z = δ zB − δ zA
Therefore
u xA u xA
u u
δ ' x − 1 1
xB xB
u yA = [n ]u yA
δ ' y = −1 1 u BI (2-6-4)
δ ' yB u yB
z − 1 1
δ zA δ zA
δ zB δ zB
56
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
δ zA u n
δ zB
The component of the transformation matrix, [TiBI ] , is discussed in Chapter 4 (Freedom Vector).
u1 u1
δ ' x u u
2 2
δ ' y = [n BI ][TiBI ] = [TBI ] (2-6-6)
δ '
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,
57
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,
∂δ z 1 δ zA 2 − δ zA1 δ zB 2 − δ zB1
≈ + (2-7-5)
∂x 2 w w
58
∂u z 1 u xB1 − u xA1 u xB 2 − u xA2
≈ + (2-7-6)
∂z 2 l l
γ ' xc = τ l =
l
(δ zA2 − δ zA1 + δ zB 2 − δ zB1 ) + 1 (u xB1 − u xA1 + u xB 2 − u xA2 ) (2-7-7)
2w 2
The axial deformation, ε ' z1 , ε ' z 2 , is,
ε 'z1 = δ zB1 − δ zA1 , ε 'z 2 = δ zB 2 − δ zA 2 (2-7-8)
In a matrix form,
u xA1 u xA1
δ δ
zA1 zA1
l l l l u xA2 u xA2
γ ' xc − 0.5 − 0.5 w − 0.5 0.5
w
0.5 − 0.5
w
0.5 0.5
w δ
δ
ε ' z1 = 0 −1 0 0 0 1 0 0 zA 2 = [DN ] zA 2
u
ε ' 0 0 0 −1 0 0 0 1 xB1 u xB1
z2 δ zB1
δ zB1
u xB 2 u xB 2
δ δ
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
δ
zA1
u xA2 u1
u
δ zA 2 2
= [T ]
ixN (2-7-10)
u xB1
δ zB1 u n
u xB 2
δ
zB 2
The component of the transformation matrix, [TixN ] , is discussed in Chapter 4 (Freedom Vector).
59
u1 u1
γ ' xc u u
2 2
ε '
z1 = [D N ][T ]
ixN = [T ]
xN (2-7-11)
ε '
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;
60
u yA1
δ
zA1
u yA2 u1
u
δ zA 2
u
[ ]
= TiyN
2
(2-7-14)
yB1
δ zB1 u n
u yB 2
δ
zB 2
Transformation from the global node displacement to the element face displacement is,
u1 u1
γ ' xc u u
ε ' z1 = [D N ] TiyN[ ] 2
[ ]
= T yN
2
(2-7-15)
ε '
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)
61
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 , γ ' xc
A1 A2
A
Force-displacement relationship
The relationship between the shear deformation and shear force of the nonlinear shear spring is,
∂δ z 1 δ zA 2 − δ zA1 δ zB 2 − δ zB1
≈ + (2-8-3)
∂x 2 w w
γ ' xc = τ l =
l
(δ zA2 − δ zA1 + δ zB 2 − δ zB1 ) + 1 (u xB1 − u xA1 + u xB 2 − u xA2 ) (2-8-5)
2w 2
62
The axial deformation, ε 'z1 , ε 'z 2 , is,
ε 'z1 = δ zB1 − δ zA1 , ε 'z 2 = δ zB 2 − δ zA 2 (2-8-6)
In a matrix form,
u xA1 u xA1
δ δ
zA1 zA1
l l l l u u xA2
γ ' xc − 0.5 − 0.5 w − 0.5 0.5
w
0.5 − 0.5
w
0.5 0.5 xA2
w δ zA 2
δ zA 2
ε ' z1 = 0 −1 0 0 0 1 0 0 = [DD ]
u xB1 u xB1
ε ' 0 0 0 −1 0 0 0 1
z2 δ zB1
δ zB1
u xB 2 u xB 2
δ δ
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
δ
zA1
u xA2 u1
u
δ zA 2 2
= [TixD ] (2-8-8)
u xB1
δ zB1 un
u xB 2
δ
zB 2
The component of the transformation matrix, [TixD ] , is discussed in Chapter 4 (Freedom Vector).
u1 u1
γ ' xc u u
2 2
ε '
z1 = [DD ][T ]
ixD = [T ]
xD (2-8-9)
ε '
z2 un un
63
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
δ
zA1
u yA2 u1
u
δ zA 2
u
[ ]
= TiyD
2
(2-8-12)
yB1
δ zB1 un
u yB 2
δ
zB 2
64
Transformation from the global node displacement to the element face displacement is,
u1 u1
γ ' xc u u
ε ' z1 = [DD ] TiyD[ ] 2
[ ]
= TyD
2
(2-8-13)
ε '
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)
65
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 τ= + Eq. (2-7-4) and Eq. (2-8-2)
∂y ∂x
66
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.
67
3) In case of damper element
θ1 θ2
θ'2
θ'1
68
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)
69
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:
∫ε σ dv = u T F
T
(2-9-4)
V
where, ε is a virtual strain vector, σ is a stress vector, u is a virtual displacement vector and F is a
load vector, respectively.
∂u 4
∂hi
∑
i =1 ∂x
ui
ε x ∂x
∂v 4
∂hi
εy = = ∑ vi
γ ∂y i =1 ∂y
xy ∂u ∂v 4 ∂hi 4
∂hi
∂y + ∂x ∑ ∂y u i + ∑ ∂x vi
i =1 i =1
u1
v
∂h1 ∂h2 ∂h3 ∂h4 1
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)
70
In the plane stress problem, the stress-strain relationship is expressed as,
σ x
1 ν 0 ε x
E ν 1 ε
σ y =
0
y
(2-9-7)
τ 1 − ν 0 0
1 −ν
γ xy
xy 2
σ= C ε
Substituting equation (2-9-6) into equation (2-9-7),
σ= C B u (2-9-8)
T
∫ (Bu ) (CBu )dv = u B T CBdxdy u = u T F
∫
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 ,
∫B
T
K =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( x(r , s ), y (r , s )) CB( x(r , 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
71
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
72
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 ∫ ∫ B( x(r , s ), y (r , s )) CB( x(r , 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 ∑∑ α iα j F (ri , s j )
i =1 j =1
where
∂ ( x, y )
F (r , s ) = B(x(r , s ), y (r , s )) CB( x(r , s ), y (r , s ))
T
∂ (r , s)
73
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).
74
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 − 0.5γ A h
Node B: ∆v B ≈ − 0.5γ A w , Node C: ∆vC ≈ 0 (2-10-1)
∆θ 0.5γ ∆θ − 0.5γ
B A C A
75
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 − θ A h
Node B: v B ≈ v A + θ A w , Node C: vC ≈ v A (2-10-2)
θ θ θ θ
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;
− θ Ah
h
θ Aw
θA
C w
vA
uA
A B
A B
76
Node B:
u
u B u A 0 uA 1 0 0 0 A
v A
v B ≈ v A + θ A w + − 0.5γ A w = v A + θ A w − 0.5γ A w = 0 1 w − 0.5w
θ
θ θ
B
0.5γ θ + 0.5γ
A
0 0 1
0.5 A
A A A
γ A
(2-10-3)
Node C:
u A
u C u A − θ A h − 0.5γ A h u A − θ A h − 0.5γ A h 1 0 h − 0.5h
v
vC ≈ v A + 0 = vA = 0 1 0 0 A
θ θ − 0.5γ θ − 0.5γ 0 0 1 − 0.5 θ A
C A A A A γ
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
θ +
yA l ' zAu + λ θ − u + λ θ
B yB − 1 + λ λ B u
A yA zB A
zB
= l'
= l' l'
(2-10-5)
1 1 1 1 θ
θ yB + u zA + λ Aθ yA − u zB + λ Bθ yB − λA 1 + λ B yA
l' l' l' l' θ yB
θ yB
θ ' yA θ yA τ
θ yB
θ yA τ θ ' yB
u zB
Z
u zB − λ B l 'θ yB
u zA u zA + λ A l 'θ yA
A B X
λ Al ' l' λB l '
77
If we consider shear deformation of connection panel, from Figure 2-10-6,
u zA
u
1 1 zB
l' − 1 + λA λB 0.5 − 0.5λ A − 0.5λ B θ
l ' yA
= θ (2-10-6)
1 1
− λ A 1 + λB − 0.5λ A 0.5 − 0.5λ B yB
l' l' γ yA
γ yB
θ yB
τ
θ ' yA θ yA + 0.5γ yA
θ yB + 0.5γ yB θ ' yB
τ
θ yA u zB
u zB − λ B l ' (θ yB − 0.5γ yB )
u zA u zA + λ Al ' (θ yA − 0.5γ yA )
A B
Z
78
The transformation matrices for beam element are;
Including connection panel and node movement
u zA u zA
u u
1 1 zB zB
θ ' yA l ' − l ' 1 + λ A λB 0.5 − 0.5λ A − 0.5λ B θ yA
θ yA
θ ' 1
θ
1 θ
yB
= l' − l' λA 1 + λB − 0.5λ A 0.5 − 0.5λ B = [Λ B ] yB
yB
δ γ yA γ yA
1
xA
δ xB γ yB γ yB
1
δ xA δ xA
δ δ
xB xB
(2-10-10)
From global node displacement to element node displacement
u zA
u
zB
θ yA u1
u
θ yB 2
= [TixB ] (2-10-11)
γ yA
γ yB u n
δ xA
δ
xB
u1 u1
θ ' yA u u
2 2
θ '
yB = [n B ][Λ B ][T ]
ixB = [T ]
xB (2-10-12)
δ '
x un un
79
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
θ yA −1 θ xA θ xA
θ yB − 1 θ θ xB
= = [s B ]
xB
(2-10-14)
γ xA
γ yA
−1
γ xA
γ yB −1 γ xB γ xB
δ xA
1 δ yA δ yA
δ 1 δ yB Global δ yB
xB Y − Beam Global
Transformation from the global node displacement to the element node displacement is,
u zA
u
zB
θ xA u1
u
θ xB
= TiyB [ ] 2
(2-10-15)
γ xA
γ xB u n
δ yA
δ yB
Transformation from the global node displacement to the element face displacement is,
u1 u1
θ ' yA u u
[ ]
θ ' yB = [n B ][Λ B ][s B ] TiyB
2
[ ]
= T yB
2
(2-10-16)
δ '
x u n u n
80
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
θ − u + λ θ + u + λ θ − l ' 1 + λA λ B u
yA
l '
xA A yA
l '
xB B yB
l' xB
= = 1 (2-10-17)
1 1 1
θ yB − u xA + λ Aθ yA + u xB + λ Bθ yB − λ A 1 + λ B θ yA
l' l' l' l' θ yB
B u xB
λB l ' θ yB
u xB + λ B l 'θ yB θ ' yB
θ yB
τ
l' Z
τ
θ yA
u xA − λ A l 'θ yA θ ' yA X
θ yA
λ Al '
Y
A u xA
81
If we consider shear deformation of connection panel, from Figure 2-10-8,
u xA
u
1 1 xB
− l ' 1 + λA λB − 0.5 + 0.5λ A 0.5λ B θ yA
= l' (2-10-18)
1 1 θ
− λ A 1 + λB 0.5λ A − 0.5 + 0.5λ B yB
l' l' γ yA
γ yB
B u xB
λB l ' θ yB
u xB + λ B l ' (θ yB + 0.5γ yB )
θ ' yB
Z θ yB − 0.5γ yB
τ
l'
X
τ
θ yA − 0.5γ yA
θ ' yA
u xA − λ Al ' (θ yA + 0.5γ yA )
θ yA
λ Al '
u xA
A
Figure 2-9-8 Column displacement with shear deformation of connection panel (X-Z plane)
82
In the same manner, assuming rigid connection, the nodal displacement of column in Y-Z plane is
expressed as,
u
1 1 1 1 yA
θ xA + l ' u yA + λ Aθ xA − l ' u yB + λ Bθ xB l ' − l ' 1 + λ A λB u
yB
= =
1 (2-10-19)
1 1 1
θ xB + u yA + λ Aθ xA − u yB + λ Bθ xB − λ A 1 + λ B θ xA
l' l' l' l' θ xB
λB l ' u yB
u yB − λ B l 'θ xB
θ xB
θ 'xB
l' θ xB
τ Z
τ
θ xA u yA + λ Al 'θ xA
X
θ 'xA
λ Al '
θ xA A Y
u yA
83
If we consider shear deformation of connection panel, from Figure 2-10-10,
u yA
u
1 1 yB
l' − 1 + λA λB − 0.5 + 0.5λ A 0.5λ B θ xA
= l ' (2-10-20)
1 1
− λA 1 + λB 0.5λ A − 0.5 + 0.5λ B θ xB
l' l' γ xA
γ xB
u yB B
θ xB λB l '
u yB − λ B l ' (θ xB + 0.5γ xB )
θ ' xB
τ θ xB − 0.5γ xB
Z
l'
τ
Y
θ xA − 0.5γ xA
θ ' xA
u yA + λ Al ' (θ xA + 0.5γ xA )
θ xA
λ Al '
u yA
A
Figure 2-9-10 Column displacement with shear deformation of connection panel (Y-Z plane)
84
The transformation matrices for column element are;
Including connection panel and node movement
u xA
u
xB
θ yA
1 1 1 λA λB θ yB
− 1 + λA λB − +
θ ' yA l ' l' 2 2 2
γ yA
θ ' − 1 1 λA 1 λB
λ A 1 + λB − + γ
yB l ' l' 2 2 2 yB
θ ' xA 1 1 1 λA λB u yA
− 1 + λA λB − + u
θ ' xB l' l' 2 2 2
= 1 1 λA 1 λB yB
δ zA − λ A 1 + λB − + θ xA
l' l' 2 2 2
δ zB
1 θ xB
θ zA 1 γ xA
θ
zB 1 γ xB
1 δ zA
δ
zB
θ zA
θ zB
u xA
u
xB
θ yA
θ yB
γ yA
γ yB
u
yA
u yB
= [Λ C ]
θ xA
θ xB
γ xA
γ
xB
δ zA
δ
zB
θ zA (2-10-21)
θ zB
85
From global node displacement to element node displacement
u xA
u
xB
θ yA
θ yB
γ yA
γ yB
u u1
yA u
u yB 2
= [TiC ]
θ xA
θ xB u n
γ xA
γ
xB
δ zA
δ
zB
θ zA
(2-10-22)
θ zB
θ ' yA
θ '
yB u1 u1
θ ' xA u
u 2 2
= [n C ][Λ C ][T ]
iC = [T ]
C (2-10-23)
θ ' xB
δ 'z u n u n
θ ' z
86
4) Force-displacement relationship for the connection
γx
γy
The relationship between the displacement vector and force vector of the element is expressed as follows:
M Px k Px 0 γ x
= (2-10-24)
M Py 0 k Py γ y
k Px = k Py = GV (2-10-25)
u1
γ x u
2
= [T ]
P (2-10-26)
γ
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,
87
2.11 Ground Spring
Foundation
K Hx M yG , θ yG
PxG , u xG
G
CSx K Ry CRy
Foundation
K Hy
CSy
M xG , θ xG
G
CRx Z
PyG , u yG
K Rx
X
88
Therefore
u xG u1
u u
yG 2
= [TG ] (2-11-4)
θ yG
θ xG un
The component of the transformation matrix, [TG ] , is discussed in Chapter 4 (Freedom Vector).
Constitutive equation
The constitutive equation of the ground spring is;
P1 u1 u1
P u u
2
=
[ G ] [ G ] 2
K
2
+ C (2-11-5)
Pn un un
where,
[ KG ] [T=
G ] [ kG ][TG ] , [ CG ] [TG ] [cG ][TG ]
T T
(2-11-6)
89
3. Nonlinear Element Models
Notation
σy : Strength of rebar
Mc : Crack moment
My : Yield moment
k0 : Initial stiffness
90
γc : Crack shear deformation
91
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 − 1)(a1 + 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 t
Ie = + + BD g − + (S − B )t D − − g +
12 12 2 2
2
t
(n E − 1)a1 (d 1 − g ) + (n E − 1)a 2 (D − d 2 − g ) + (n E − 1)a S D − − g
2 2
(3-1-3)
2
where, g is the center of beam section calculated by
BD 2 / 2 + ( S − B ) t ( D − t / 2 ) + ( nE − 1) {a2 d 2 + a1 ( D − d1 ) + aS ( D − t / 2 )}
g= (3-1-4)
AN
92
b) Nonlinear bending spring
τA
δx
MA MB
A B
τB
Nx
elastic element
θA φA
MA MB MA MB
A B A B
Nx θB φB
θ A τ A φ A η A ηA
MA MB
θ B = τ B + φB + η B A ηB B
δ δ 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.
φ θ
M τ τ M
A B
θ φ
Moment distribution
M M M
My My
6 EI = +
Mc k0 = Mc
l
k 0 α y k0 k0 kn ≈ ∞
θc θy θ τ φc φ y φ
Elastic element Nonlinear bending spring
93
Crack moment force
For reinforced concrete elements, the crack moment, M c is calculated as,
where,
σB : Compression strength of concrete (N/mm2)
Z e1 , Z e 2 : Section modulus
where,
σy : 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 = α y k0 , k0 = (3-1-9)
l
where,
α y is the stiffness degradation factor at the yield point, which is obtained from the following
empirical formulas:
α y = (0.043 + 1.63npt + 0.043a / D )(d / D )2 , (a / D ≤ 2) (3-1-10)
94
α y is modified in case of tension in top main rebars as
I
α y ' = α 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, φ y , is then obtained from,
1 My
φ y = − 1
k
(3-1-13)
αy 0
In general, the relation between the rotation of bending spring and that of nonlinear bending spring is
My
φ =θ − (3-1-14)
k0
Crack rotation
From Figure 3-1-2, the crack rotation of the nonlinear bending beam, φ c , is supposed to be zero value,
however, in STERA_3D program, it is assumed as,
φ c = 0.001 φ y (3-1-15)
95
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 = η s Ls (3-1-17)
96
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
φ φ
φx φy φm
α β
My My φy Mm φy
kp =η kr = k s = (3-1-18)
ϕ φ φ
y y m φm − φ x φm
The strength degradation under cyclic loading is considered by elongating the target displacement, φ m , to
be φ ' m as shown in the following Figure:
M
kp
Mm
My
ks kr
φ
n φm n φy φ φy φm φ 'm
x
φy
φ ' m = 1 + γ
n
φ m (3-1-19)
n φm
97
Relationship between curvature and rotation
εc
θ EI
M M
A B
θ M
D
φ
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= θ (3-1-20)
l
On the other hand, the relationship between moment and curvature is
M
φ= (3-1-21)
EI
Therefore,
6
φ= θ (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
εc
φ= (3-1-23)
D/2
Therefore, the relationship between rotation and compressive strain is
l l
θ= φ= εc (3-1-24)
6 3D
l
Assuming D ≈ , then
9
θ = 3ε c (3-1-25)
98
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.
ηA s
MA MA + MB
MB Q=
A ηB B l
s
η=
A η=
B
l l
nonlinear shear springs
Q Q
k y 3 = 0.001k 0
Qy
k y3
Qu
Qc s
k0 = GA l
k0
sc sy su s
where,
99
Ultimate shear force
The ultimate shear force is, Qu , is assumed as,
Qy + k y 3 ( su − s y )
Qu = (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
k0 = GA l
sc sy su s
100
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.
φ θ
M τ τ M
A B
θ φ
My My
6 EI = +
Mc k0 = Mc
l
k 0 α y k0 k0 kn ≈ ∞
θc θy θ τ φc φ y φ
My My 1 M
θy = = τ= + φy = − 1 y
α y k0 k0 α k
y 0
Increase Reduce
stiffness stiffness
M M
My
+ Mc
k = k0 / γ 1
*
0
kn* = EI / p1
τ* φ c* φ y* φ*
My γ 1M y 1 M
θy = = τ* = + φ y* = − γ1 y
α y k0 k0 α k
y 0
101
The idea is realized using flexibility reduction factors, γ 1 (< 1), γ 2 (< 1) , in the relationship between the
displacement vector and force vector of the elastic element in Equation (2-1-1) as,
l' l'
γ 1 3EI −
6 EI y
0
τ ' yA y M ' yA
l' l'
τ ' yB = − γ2 0 M ' yB (3-1-32)
δ ' 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.
φ yA p1 EI 0 M ' yA
= (3-1-33)
φ 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
φ c* = p1 , φy = − γ1 y
*
(3-1-34)
EI α k
y 0
In general, the relation between the rotation of bending spring and that of nonlinear bending spring is
My
φ = θ − γ1 (3-1-35)
k0
Making the modified flexibility matrix to be identical to the original one,
102
e) Modification of stiffness degradation factor at the yield point
(The following modification of the stiffness degradation factor, α y , is suggested by Prof. Okano at Chiba
University.)
From Equations (3-1-32) and (3-1-34), the yield rotation of the member θ y under anti-symmetric loading
condition, M A = M B = M y , is calculated as,
(2γ − 1)M y 1 M 1 M
θy = + −γ y = + γ − 1 y (3-1-39)
α k k
k0 y 0 αy 0
where γ1 = γ 2 = γ .
The yield rotation θ y in Equation (3-1-39) is different from the formula in Figure 3-1-10 since the factor
γ is multiplied to only diagonal elements of flexural matrix in Equation (3-1-32).
1 1
= + γ − 1 (3-1-40)
α ' y α y
To realize the designated value of stiffness degradation factor, α y should be modified as,
1
αy = 1 +1− γ (3-1-41)
α'
y
For example, to realize the stiffness degradation factor α ' y = 0.4, assuming γ = 0.7 , the modified α y is
1
αy =1 + 1 − 0.7 = 0.357
0.4
103
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
ηd A l' ηd 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 − η )d A
M y '= M y = ξM y
l/2
(3-1-42)
d
ξ = 1 + 2(1 − η ) A
l
For example, when l = 540cm, d A = 30cm, η = 0.75 ,
104
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)
105
b) Nonlinear bending spring
θA τA
δx
MA MB
MA MB A B
A B τB
θB Nx
Nx
elastic element
l φA
θ A τ A φ A MA MB
A B
θ B = τ B + φ B φB
δ δ 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.
φ θ
M τ τ M
A B
θ φ
Moment distribution
M M M
My My
6 EI = +
k0 =
l kn ≈ ∞
k0 k0
θy θ τ φy φ
Elastic element Nonlinear bending spring
106
Yield moment force
σy
My
Figure 3-1-15
Yield rotation
From Figure 3-1-14, the yield rotation of the nonlinear bending beam, φ y , is supposed to be zero value,
φ y = 0.001 θ y (3-1-50)
where
6 EI
θ y = M y / k0 , k0 =
l
Hysteresis model
A bi-linear model is assumed for the hysteresis model.
My
φy
107
c) 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.
φ θ
M τ τ M
A B
θ φ
M M M
My My
6 EI = +
k0 =
l kn ≈ ∞
k0 k0
θy θ τ φy φ
My My
θy = = τ= + φy ≈ 0
k0 k0
Increase Reduce
stiffness stiffness
M M
My
+
k0* = k0 / γ 1 kn* = EI / p1
τ* φ y* φ*
My γ 1M y My
θy = = τ* = + φ y* = (1 − γ 1 )
k0 k0 k0
108
The idea is realized using flexibility reduction factors, γ 1 (< 1), γ 2 (< 1) , in the relationship between the
displacement vector and force vector of the elastic element in Equation (2-1-1) as,
l' l'
γ 1 3EI −
6 EI y
0
τ ' yA y M ' yA
l' l'
τ ' yB = − γ2 0 M ' yB (3-1-51)
δ ' 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.
φ yA p1 EI 0 M ' yA
= (3-1-52)
φ 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
φ y* = p1 (3-1-53)
EI
In general, the relation between the rotation of bending spring and that of nonlinear bending spring is
My
φ = θ − γ1 (3-1-54)
k0
Making the modified flexibility matrix to be identical to the original one,
109
d) Shear spring for damper using low yield strength steel
ηA
s
MA MB
A ηB B
MA + MB s
Q= , η=
A η=
B
l l
Nonlinear relationship between shear force Q and deformation s is defined.
110
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,
nE = E s / Ec : Ratio of Young’s modulus between steel (Es) and concrete (Ec)
111
where, g is the center of beam section calculated by
BD 2 / 2 + ( S − B)t (D − t / 2 ) + (nE − 1)(a1d1 + 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,
σB : 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 = 0.9a1σ y (D − d1 ) + 0.9aS σ y (D − t / 2 ) when tension in top main rebars
M y 2, RC = 0.9a2σ y (D − d 2 ) when tension in bottom main rebars
where,
σy : Strength of rebar (N/mm2)
1
M y ,S = b1t f (h1 − t f ) + t w (h1 − 2t f ) 2 σ y ,S : Yield moment of steel (3-1-66)
4
where,
σ y, S : Strength of steel (N/mm2)
112
Appendix 3.1:
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.
113
M
kp
My
Effective slab ratio Rs
ku
( η 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,
Mm My φy
α
My kr = (α = R1 )
φ φ
y m
n φm n φy ks kr Mm φy
β
φ k s = ( β = R2 )
φx φy φm φ 'm φm − φ x φm
φy
φ ' m = 1 + γ
n
φ m (γ = R3 )
n φm
More detail rule in the hysteresis loop is described in the following sections:
114
1. Elastic range
fy(1)
fc(1)
s0(1)
1
s0(2)
fc(2)
fy(2)
sy
4
sc
2 (drm(1), frm(1))
unloading point
3
s0
1
3
towards the maximum point of the other side
(drm(2), frm(2))
unloading point
115
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 α
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
LINE1
point. LINE2
xdm
LINE1
0.1fy
Lower boundary
LINE2
xdm
116
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))
117
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
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
119
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
120
b) Nonlinear bending spring
Hysteresis model of nonlinear bending spring is defined as the moment-rotation relationship under the
anti-symmetry loading in Figure 3-2-2. 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 kn
c y c y
Elastic element Nonlinear bending spring
Ze : Section modulus
N : Axial load
121
The yield moment, M y is calculated from the following formula under the axial load, N
if (0 N N b )
N
M y 0.8at y D 0.5 ND1 (3-2-6)
bD B
if ( N b N N max )
N N
M y 0.8a t y D 0.12bD 2 B max (3-2-7)
N max N b
The tangential stiffness at the yield point, k y , is obtained from the following equation,
6 EI
k y y K0 K0 (3-2-10)
l
where,
y is the stiffness degradation factor at the yield point, which is obtained from the following
empirical formulas:
(3-2-12)
where,
122
The yield rotation of the nonlinear bending beam, y , is then obtained from,
1 My
y 1
K
(3-2-13)
y 0
Reference:
AIJ Standard for Structural Calculation of Reinforced Concrete Structures, Architectural Institute of Japan,
2018 (in Japanese)
123
Case 1: In the case that bending springs in x and y directions are independently defined
M ' yB M ' xB
B B
yB xB
yA xA
A A
M ' yA
M ' xA
The rotational displacement vector of the nonlinear bending spring is defined independently,
where, f yA , f xA , f yB , and f xB are the flexural stiffness of nonlinear bending springs at both ends of
the element, and
f yA 1 k yA , f xA 1 k xA , f yB 1 k yB , f xB 1 k xB (3-2-16)
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
(3-2-17)
xB M ' xB f xB M ' xB
zB N ' zB 0 N ' zB
124
The hysteresis model for M relationship is the degrading tri-linear slip model as used for the
hysteresis model of the bending springs of the RC beam.
M M
kp
My
My
Mc Mc
ks kr
x y m
M My y Mm y
kp y kr k s (3-2-18)
y
y m m x m
125
Case 2: In the case that nonlinear interaction between moment and axial components is considered
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
3-2-4.
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 (3-2-18)
In a matrix form
M ' y
k i xi2 k i xi y i k x i i
y y
i
x k p x
i i
M 'x k y i
2
i k y i i (3-2-20)
N'
k
i i
z sym. z z
i
i
126
Therefore
y M ' y M ' y
x k p M ' x f p M ' x
1
(3-2-21)
N' N'
z z z
yA M ' yA
M '
xA xA
zA f pA 0 N ' zA
yB 0 f pB M ' yB (3-2-22)
xB M ' xB
zB N ' zB
127
c) Nonlinear vertical springs
The nonlinear bending spring is constructed from the nonlinear vertical springs arranged in the member
section as shown in Figure 3-2-5. 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 .
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
128
Strength of concrete spring
As shown in Figure 3-2-6, the strength of the corner concrete spring is obtained from the equilibrium
condition in the vertical direction under the balance axial force, N b 0.4bD B , that is,
Nb
c f y1 0.2bD B (3-2-24)
2
Therefore, the area of the corner concrete, A1 , is,
c fy
A1 (3-2-25)
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 (3-2-26)
The strength of the center concrete spring is then obtained as,
c f y 2 0.85k B A2 (3-2-27)
M y x s 2 s f y c f y1 x s 2 s f y 0.5 N b (3-2-28)
Therefore,
My
xs (3-2-29)
2 s f y 0 .5 N b
Note that M y is calculated from Equation (3-2-6) for the balance axial force, N Nb .
129
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 3-2-8
a t 15 .484 ( cm 2 ) y 1 .1 f y 32 .45 ( kN / cm 2 ) B 2 .4 ( kN / cm 2 )
s f y 251.2 (kN ) f
c y1 1200 (kN ) f
c y2 390 (kN ) x s 30 (cm)
M y 2 s f y 0 .5 N x s
which is plotted as the solid line in Figure 3-2-8. The results of Multi-Spring model give smaller values
than theoretical results in the range 0 < N < Nb.
130
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)
K-N. Li (1988) proposed to use the following formulation for deciding the location of vertical springs
instead of Equation (3-2-29), as follows:
M y0
xs (3-2-30)
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
bD B
x s 35.8 (cm)
The yield moment is plotted as the solid line in Figure 3-2-9. It improves the results of Multi-Spring model.
131
M-N relationship
6000
Theory
5000 Multi-Spring
4000
N (kN)
3000
2000
1000
0
0 20000 40000 60000 80000
M (kN*cm)
132
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)
133
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.
Qc 0.3Qy (3-2-34)
Qu Qy k y 3 su s y (3-2-35)
Qy
k y3
Qu
Qc 0.001k0
k0 GA l
sc sy su s
134
Crack shear deformation
The crack shear deformation is obtained as,
Qc
sc cl , c (3-2-36)
GA
Yield shear displacement
The yield shear deformation is assumed as,
1
s y yl , y (3-2-37)
250
Ultimate shear displacement
The ultimate shear deformation is assumed as,
1
su u l , u (3-2-38)
100
The poly-linear slip model (see Appendix) is adopted for the hysteresis of the shear spring.
The parameters on the backbone curve can be changed in the Option Menu of Column element. The default
values are given as follows:
135
d-2) Shear spring model 1
Case 1: In the case that shear springs in x and y directions are independently defined
Qx
Qy
sx z
sy
x
Qx y
Qy
Qx k sx 0 sx
(3-2-39)
Qy 0 ksy s y
136
d-3) Shear spring model 2
Case 2: In the case that nonlinear interaction between shear and axial components is considered
N 'zB
M 'xB
M ' yB
B
N z , sz
Qx , sx
Qy , s y
A
M ' yA
M 'xA
N 'zA
Qx sx
Qy k sp s y (3-2-40)
N
z sz
The stiffness matrix k sp is obtained by the Plastic Theory as explained in the Appendix (not
implemented).
137
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 kn
c y c y
Elastic element Nonlinear bending spring
Increase Reduce
stiffness stiffness
M M
My
+ Mc
k *
0
kn*
c* y*
Elastic element Nonlinear bending spring
138
Introducint the concept of “plastic zones”, the initial stiffness of the i-th multi-spring can be expressed as,
Ei Ai
k 0i (3-2-41)
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 (3-2-20), 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
i i
(3-2-42)
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 ' l' (3-2-43)
f C 2
6 EI y 3EI y
l'
0
EA
p z1 l' l'
l' l' 1 0
E A x2
3EI 0 3EI y 6 EI y
6 EI y i i i
y
i
l' pz2 l'
0 2 0
3EI y
Ei Ai xi
2
3EI y
l' i
sym. p z1 pz2 l'
EA original sym. 0
i Ei A i Ei A EA
mod ified
(3-2-44)
Since Ax
i
2
i i I y , this gives the flexivility reduction factors as:
3 3 1
1 1 p z 1 , 2 1 p z 2 , 0 1 p z1 p z 2 (3-2-45)
l' l' l'
l'
Adopting p z1 p z 2 as discussed for beam element, the reduction factors will be:
10
1 2 0.7 , 0 0.8 (3-2-46)
139
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.
140
3.2.2 Steel Column
a) Section properties
tf t2
H t1
tw t1
B B D
AS 0.5 AN
141
3) Circle section
I
64
D 4
D 2t
4
(3-2-55)
2t1t 2 ( B t1 ) 2 ( H t 2 ) 2
J 2 2
(3-2-57)
BHt1t 2 t1 t1
5) Circle section
J
32
D 4
D t
4
(3-2-58)
142
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 the
following Figure.
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 (3-2-59)
In a matrix form
M ' y
k i xi2 k i xi y i k x i i
y y
i
x k p x
i i
M 'x k y i
2
i k y i i (3-2-61)
N'
k
i i
z sym. i
z z
i
Therefore
143
y M ' y M ' y
x k p M ' x f p M ' x
1
(3-2-62)
N' N'
z z z
yA M ' yA
M '
xA xA
zA f pA 0 N ' zA
yB 0 f pB M ' yB (3-2-63)
xB M ' xB
zB N ' zB
Hysteresis model of nonlinear bending spring is defined as the moment-rotation relationship under the
anti-symmetry loading as shown in Figure 3-2-20. 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
144
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
(3-2-64) 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 (3-2-65) H
2 2 B y 0 y
2
where
1 N N y
y0 H
2 B y
N y 2 Bt f t w H 2t f y
My
145
2) H shape
1
M y M y 0 y 0 H y
2
(3-2-66) Bt f y
2
where 1
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 (3-2-67) 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
146
4) Circle shape
t1
N
M y M y 0 cos (3-2-70)
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 (3-2-71)
l
147
c) Nonlinear vertical springs
The nonlinear bending spring is constructed from the nonlinear vertical springs arranged in the member
section as shown in Figure 3-2-21. 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 (3-2-72)
d y ,i f y ,i / k 0i , k 0i E s Ai (3-2-73)
148
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 (3-2-74)
pz
where Ei : the material young’s modulus, Ai : the spring governed area, and pz : the length of assumed
149
plastic zone. When p z 0 , it represents the infinite stiffness for rigid condition.
From Equation (3-2-20), 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
i i
(3-2-75)
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 ' l'
f C 2
(3-2-76)
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' pz2 l'
0 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
(3-2-77)
150
3.2.2 Column with direct input
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 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
(3-2-80)
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,
151
The flexural matrix [ f C ] is;
l' 1 l' 1
f yA 3EI k l'2
6 EI y k sx l'2
y sx
l' 1
f xA 0
3EI y k sx l'2
l' 1 l' 1
f yB
3EI x k sy l'2 6 EI x ksy l'2
C
f
l' 1
f xB
3EI x ksy l'2
l'
EA
l'
sym.
GI z
(3-2-82)
152
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
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 (3-2-84)
153
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 (3-2-85)
12 2
BD 3 D
2
Ix n E 1 a c a 2 d 2 I S , x
(3-2-86)
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
154
Yield moment force
The yield moment, M y is calculated as,
M y M y , RC M y , S (3-2-88)
where
M y , RC : Yield moment of reinforced concrete
Nb
M y , RC 0.8a t y D 0.5 N b D1 (3-2-89)
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
155
Appendix 3.2:
A-1. Hysteresis of Steel and Concrete Springs of Multi-Spring Models for RC elements
a) Steel spring
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 3-2-15.
φ sf y
−φ sf y
s fy
ν sf y
E i Ai
k 0i = s
pz
s d *y
The hysteresis of steel spring has the degradation point at the forces, ν sf y and φ s f y , where ν and φ
are the arbitrary parameters (ν < 1, φ < 1) . The STERA_3D Program adopts the values as:
ν = 1 / 3, φ = 0.5 (3-2-47)
*
Then, the yield deformation, s d y, may be obtained by Equations (3-2-31) and (3-2-13) considering the
reduction factor γ.
φ y* x s
sdy = (3-2-48)
N 0 +2 s f y
1+
2 s f y +2c f y
1 1My
φ y* = − (3-2-49)
α y γ k0
156
b) Concrete spring
The hysteresis of concrete spring is also defined as tri-linear hysteresis model as shown in Figure 3-2-16.
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
157
A-2. Hysteresis of Poly-linear Slip Model for Shear Springs for RC and Masonry members
Force
FY
(DRX, FRX)
SY
FC SU
SF
S2 S1
SC
-DU -DY 0.15 FRX
DISP
0.15 FRN DY DU
-FC
(DRN, FRN)
-FY SE
(DRX, FRX)
FRX + FRN
SE =
DRX + DRN
S1= 2 × SE S1
S1
2 1.2 × SE
S=
(DL2, FL2) (DL1, FL1)
FL
= 1 FL
= 2 0.15 × FRX
FL
= 3 FL
= 4 0.15 × FRN
(DL4, FL4) (DL3, FL3)
S1
S1
(DRN, FRN)
158
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,
159
b) Nonlinear bending spring
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
φ τ
θ
θ
φ τ
A
Moment distribution
M
M M M
My My
2 EI
Mc k0 = = + Mc
l
k0 k y = α y k0 k0 kp ≈ ∞
θc θy θ τ φc φ y φ
Elastic element Nonlinear bending spring
The yield moment, M y is obtained from the equilibrium condition in Figure 3-3-6 as,
where,
as : Total area of rebar in the side column
σy : Strength of rebar in the side column
aw : Total area of vertical rebar in the wall panel
σ wy : Strength of rebar in the wall panel
N : Axial load from the dead load
160
N
a sσ y
My
a wσ wy
lw
M c = 0.3M y (3-3-5)
The tangential stiffness at the yield point, k y , is obtained from the following equation:
k y = 0.2 K 0 (3-3-6)
The yield rotation of the nonlinear bending beam, φ y , is then obtained from,
1 My
φ y = − 1 (3-3-7)
K
α y 0
α y = 0.02 (3-3-8)
161
Case 1: In the case that bending springs are independently defined
y
φxA1 φ yAc φxA 2
A1 Ac A2
M ' yAc
M 'xA1 M 'xA 2
The rotational displacement vector of the nonlinear bending spring is defined independently,
(3-3-11)
162
The hysteresis model for M − φ relationship is the degrading tri-linear slip model as used for the
hysteresis model of the bending springs of the RC beam.
M M
kp
My
My
Mc Mc
ks kr
φ φ
φx φy φm
α β
My My φy Mm φy
kp =η kr = k s = (3-3-12)
ϕ φ φ
y y m φm − φ x φm
163
Case 2: In the case that nonlinear interaction between moment and axial components is considered
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
φx 2
φx1 φ yc
ε zc εi
yi x
xi
N ' zc , ε zc
M ' yc , φ yc
164
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 ε i xi + ∑ k i ε i xi + ∑ k i ε i xi
i i i
Nc N1 N2
= ∑ k i (ε zc + xiφ yc ) x i + ∑ k i (ε zc − y iφ x1 + xiφ yc ) xi + ∑ k i (ε zc − y iφ x 2 + xiφ yc ) xi
i i i
φ yc
Nc + N 1+ N 2 N1 N2 Nc + N 1+ N 2
φ x1
= ∑ k i xi2 − ∑ k i xi y i − ∑ k i xi y i ∑i k i xi φ
i i i x2
ε zc
(3-3-14)
Nc N1 N2
N ' zc = ∑ k i ε i + ∑ k i ε i + ∑ k i ε i
i i i
Nc N1 N2
= ∑ k i (ε zc + xiφ yc ) + ∑ k i (ε zc − y iφ x1 + xiφ yc ) + ∑ k i (ε zc − y iφ x 2 + xiφ yc )
i i i
φ yc
Nc + N 1+ N 2 φ
Nc + N 1+ N 2 N1 N2
= ∑ k i xi − ∑ k i yi − ∑ k i yi ∑i k i φ x1
i i i x2
ε zc
(3-3-15)
where, Nc, N1 and N2 are the number of vertical springs in a wall panel, side column 1 and side column 2,
respectively.
165
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 ε i y i
i
N1
= −∑ k i (ε zc − y iφ x1 + xiφ yc ) y i
i
φ yc
N1 N1 N1
φ x1
= − ∑ k i xi y i ∑k y i
2
i 0 − ∑ k i yi
i i i φ x 2
ε zc
(3-3-16)
Also, for the side column 2,
N2
M ' x 2 = −∑ k i ε i y i
i
N2
= −∑ k i (ε zc − y iφ x1 + xiφ yc ) y i
i
φ yc
N2 N2 N2
φ x1
= − ∑ k i xi y i 0 ∑k y i
2
i − ∑ k i yi
i i i φ x 2
ε zc
(3-3-17)
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 φ φ yc
M ' − ∑ k i xi y i − ∑ k i yi
yc
x1 i ∑k y i
2
i 0
φ x1
φ
= N2
i
N2 N2
i
[ ] x1
φ = k p φ
M ' x 2 − ∑ k i xi y i 0 ∑k y i
2
i − ∑ k i yi x 2 x2
N ' zc i i i
ε zc ε 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
(3-3-18)
Therefore
φ yc M ' yc M ' yc
φ M ' M '
x1
= kp [ ] −1
x1
= f [ ]
x1
p (3-3-19)
φ x 2 M ' x 2 M ' x 2
ε zc N ' zc N ' zc
166
φ yAc M ' yAc
φ M '
xA1 xA1
φ xA2 M ' xA2
[ ]
ε zAc f pA
0 N ' zAc
=
φ yBc 0 [ ]
f pB M ' yBc
(3-3-20)
167
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)
168
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 wσ wy
s fy = (3-3-21)
5
where,
aw : Total area of vertical rebar in the wall panel
σ 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 pσ B
c fy = (3-3-22)
5
where,
Ap : Total area of wall panel section
σB : Compression strength of concrete
where,
169
Crack shear force
The crack shear force is, Qc , is assumed as,
Qy
Qc = (3-3-24)
3
Ultimate shear force
The crack shear force is, Qu , is assumed as,
Qu = Qc (3-3-25)
Q y1 Q xc Qy2
− Q y1 − Q xc − Qy2
170
Q Q
k y 3 = 0.001k 0
Qy
k y3
Qu
Qc s
k0 = GA l
k0
sc sy su s
171
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
φ τ
θ
θ
φ τ
A
Moment distribution
M
M M M
My My
2 EI
Mc k0 = = + Mc
l
k0 k y = α y k0 k0 kp ≈ ∞
θc θy θ τ φc φ y φ
Elastic element Nonlinear bending spring
Increase Reduce
stiffness stiffness
M M
My
+ Mc
k *
0
k *p
τ φ c* φ y* φ
Elastic element Nonlinear bending spring
172
Introducint the concept of “plastic zones”, the initial stiffness of the i-th multi-spring can be expressed as,
Ei Ai
k 0i = (3-3-29)
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,
γ 0 (< 0), γ 1 (< 0), γ 2 (< 0) , the flexibility matrix of the elastic element is,
l' l'
γ 1 3EI −
6 EI c
c
l'
γ2
3EI c
l' l'
γ1 −
3EI 1 6 EI 1
l'
[ fW ] = γ2 (3-3-26)
3EI 1
l' l'
γ1 −
3EI 2 6 EI 2
l'
sym. γ2
3EI 2
l'
γ0
EAc
l'
Also, adopting p z = as discussed for beam and column elements, the reduction factors will be:
10
γ 1 = γ 2 = 0.7 , γ 0 = 0.8 (3-3-27)
173
f) Reduction factor of shear stiffness
If shear cracking occurs in the reinforced concrete wall, the shear stiffness decreases. The following graph
shows the test results of the relationship between the stiffness reduction factor β and the lateral drift
−3
angle R × 10 (referred from “Standard for Structural Calculation of Reinforced Concrete Structure”,
Architectural Institute of Japan).
For example, if the lateral drift angle is over than 1/1000, the reduction factor becomes less than 0.2.
Therefore, STERA_3D assumes the “Reduction Factor for Stiffness” is 0.2 in the default setting for the
option of the RC wall element.
174
3.3.1 Direct Wall
Direct Wall identifies the force-displacement points in the back-bone curves of the nonlinear shear spring
and the nonlinear bending spring.
M ' yBc
h X
M ' yAc
Different types of hysteresis model are prepared for the force-deformation relationship of the spring.
qi qi
ky ky
fy fy
fc k1 fc k1
k0 ui k0 ui
175
3.3.2 Steel Wall (Brace)
a) Buckling of brace
3 B 4
N', δ '
Z
h
1 2 X
A
w Y
Under the compression load, the stress of buckling failure is calculated theoretically as
π 2E
σE = 2 ,
λ
L
where λ = : slenderness ratio
i
If σ E > σ y (strength of steel), the compression failure will occur before buckling.
176
The AIJ (Architectural Institute of Japan) guideline adopts the following equation for the stress of buckling.
0.6
σ cr = σy, for λ > λ p (3-3-31)
(λ λ ) p
π 2E
where λp = : Critical slenderness ratio
0.6σ y
b) Hysteresis model
The hysteresis model proposed by Wakabayashi et. al. is adopted in STERA_3D (hereinafter referred to as
Wakabayashi model). The model consists of four Stages A, B, C and D.
σ [N/mm2] σ [N/mm2]
300.00 300.00
Stage A
250.00 250.00
200.00 200.00
150.00 λ=60 λ=60
150.00
100.00 100.00
50.00 50.00 Stage D
0.00 0.00
-50.00 -50.00
-100.00 -100.00
-150.00 -150.00
-200.00 -200.00
-250.00
ε -250.00 ε
-0.004 -0.002 0.000 0.002 0.004 -0.004 -0.002 0.000 0.002 0.004
σ [N/mm2] σ [N/mm2]
300.00 300.00
250.00 250.00
200.00 200.00
λ=60 150.00 λ=60
150.00
Stage C
100.00 100.00
Stage D
50.00 50.00
0.00 0.00
-50.00 -50.00
-100.00 -100.00
-150.00 Stage B -150.00
-200.00 -200.00
-250.00
ε -250.00 ε
-0.004 -0.002 0.000 0.002 0.004 -0.004 -0.002 0.000 0.002 0.004
Stage B: buckling failure and strength reduction Stage C: tension stage after buckling
177
The compression curve (Stage B) and the tension curve (Stage C) are defined using the nondimensional
strength and deformation as,
n = N / N 0 : nondimensional strength
δ =∆ / ∆ 0 : nondimensional deformation
where N : axial load, N 0 = Aσ y : axial strength (A: area, σ y : yielding stress of steel)
∆ : displacement, ∆=
0 Lε=
y Lσ y E : yield deformation
n 1 ( aδ + b )
r
=
n 1 ( p1δ + p2 )
12
=
10 nE − 1
where p1
= , = p2 4 nE + 0.6
3
Compression strength nc is also on this curve, therefore,
nc 1 ( p1δ c + p2 )
12
=
or p1nc 2δ c + p2 nc − 1 =0
Nc E ( ∆c / L ) A ∆c
Since nc
= = = = δ
N0 E (∆ / L) A ∆
p1nc 3 + p2 nc − 1 =0
n 1 ( p3δ + 1)
32
=
1
where p3 =
3.1 nE + 1.4
178
b-3) Movement of Tension Curve
Movement of tension curve x is defined as follows:
=x ln ( q1δ a + 1) − q2 s δa s x
3 − 1 nE
where
= q1 = , q2 0.115 nE + 0.36
10
y
y0
179
b-5) Movement of Compression Curve
The point shifting from the unloading Stage D to Stage C is obtained by assuming that the
plastic tension deformation δ t is proportional to the plastic compression deformation δ c as
δ t = q3δ c
δt
where
= q3 0.3 nE + 0.24
δc
Example
λ = 60
References
M. Shibata, T. Nakayama and M. Wakabayashi, "Mathematical Expression of Hysteretic Behavior of
Braces", Research Report, Architectural Institute of Japan, No. 316, pp.18-24, 1982.6 (in Japanese)
180
3.3.3 SRC Wall (Brace)
a) Section properties
Figure 3-3-17 Element model for SRC wall (RC wall with steel brace)
Qy = Qy , RC + Qy , S (3-3-32)
where
Qy , RC : Yield shear force of reinforced concrete
Qy , S = AS σ y , S cos R (3-3-34)
where,
AS : Area of steel (mm2)
σ y, S : Strength of steel (N/mm2)
R : Angle of steel
181
3.4 External Spring
δ zB
δ yA Z
B
A B A
δ xA δ xB X
B
Y
δ yB A
δ zA
N
tension
δ
K0
compression
ground
In STERA_3D, if there is no building element at one end of the external spring, this end is considered fixed. Such
spring is used to express the stiffness the ground attached to the building. In such a case, as the relationship
between axial force and deformation of the spring, the linear stiffness is defined only in compression side and zero
stiffness in the tension side as shown in Figure 3-4-2, assuming that the building detaches from the ground.
182
3.4.2 Air spring
Reference:
1) Marin Presthus, “Derivation of Air Spring Model Parameters for Train Simulation”, Master of Science
Programme, Department of Applied Physics and Mechanical Engineering, Luleå University of Technology,
Sweden, 2002
An effective area Ae is introduced to express the volume change of air bag ∆Vb as
∆Vb =
Ae z (3-4-1)
When the initial pressure of air spring is p0 , after the deflection, the pressure will change as
p=
b p0 + ∆pb for air bag (3-4-2a)
p=
r p0 + ∆pr for reservoir (3-4-2b)
The volume will also change as
Vb = Vb 0 − zAe + zs As for air bag (3-4-3a)
V=
r Vr 0 − zs As for reservoir (3-4-3b)
where
zs : the movement of air mass through orifice
As : area of surge pipe
The pressure and the volume of the isentropic process can be described by
where
p1 , V1 : initial pressure and volume
p2 , V2 : final pressure and volume
183
n : ratio of specific heat = 1.4 for Air
Applying the above equation to the air bag
n
− zA + z A
( p0 + ∆pb ) ⋅ 1 + e s s = p0 (3-4-5b)
Vb 0
∆pb n ( − zAe + zs As )
1 + ⋅ 1 + =1 (3-4-5c)
p0 Vb 0
∆pb − zAe + zs As
Assuming ≈0
p0 Vb 0
∆pb n ( zAe − zs As )
≈ (3-4-5d)
p0 Vb 0
Using the same procedure for the reservoir
( p0 + ∆pr ) ⋅ (Vr 0 − zs As )=
n
p0 ⋅ Vrn0 (3-4-6a)
∆pr nzs As
≈ (3-4-6b)
p0 Vr 0
From the Bernoulli equation, the difference of the pressure between the left and right of the pipe speeds up a
portion of gas through the orifice. The force balance in the pipe is given by
where
β : viscous damping parameter determined by experiment
Substituting Eq. (3-4-5d) and (3-4-6b),
zA − z A z A
p0 As n e s s − s s =Cs zs β (3-4-7b)
Vb 0 Vr 0
np0 As Ae Vb 0 As 1 1
z− + Cs zs β
zs = (3-4-7c)
Vb 0 Ae Vb 0 Vr 0
184
The force balance for the piston can be expressed as
( p0 + ∆pb − patm ) Ae
F
=z
∆pb Ae + ( p0 − patm ) Ae
=
n ( zAe − zs As )
= p0 Ae + ( p0 − patm ) Ae
Vb 0
np0 Ae 2 As
= z − zs + ( p0 − patm ) Ae (3-4-9)
Vb 0 Ae
From Eq. (3-4-7c)
np0 As Ae As Vb 0 + Vr 0
z − Cs zs β
zs = (3-4-10)
Vb 0 Ae Vr 0
Vr 0
λ=
Vb 0 + Vr 0
np0 Ae 2 As Ae
z− zs = Cs zs β (3-4-11)
Vb 0 Ae λ As
np0 Ae 2 λ 1 As
=Fz z− zs + ( p0 − patm ) Ae (3-4-12)
Vb 0 λ Ae λ
1 1
Fz = K v z − y + ( p0 − patm ) Ae = K v ( z − y ) + K v − 1 z + ( p0 − patm ) Ae (3-4-14)
λ λ
Therefore Fz
K v ( z − y ) = Cβ ⋅ y β (3-4-15)
Kv z
Ke
F=
z K v ( z − y ) + K e z + ( p0 − patm ) Ae (3-4-16) Cβ y
185
Incremental form of equation is
z( n +=
1) z( n ) + ∆t ⋅ z ( tn +1 )
y( n +=
1) y( n ) + ∆t ⋅ y ( tn +1 )
Then
β
y( n +1) − y( n )
Cβ = K v ( z( n +1) − y( n +1) ) (3-4-18)
∆ t
new
f ( y( n +1) )
y= y( n +1) old −
f ' ( y( n +1) )
( n +1)
where the prime f ' ( y( n +1) ) denotes derivative with respect to y( n +1) ,
f ( y)
y
y( n +1) new y( n +1) old
186
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 developed by
Wada et al.
x x qi , u i
y θi
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 θ i
= qi (3-5-1)
qi , y sin θ i
u x
u i = [cos θ isin θ 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 θi
ui x
dy
187
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 θ 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
θi
The same relationship is established for the second stiffness after yielding,
Ky
ky = N
(3-5-7)
∑ cos
i =1
2
θi
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 θ i f y (3-5-8)
i =1
Therefore, the yield shear force of each shear spring is,
Qy
fy = N
(3-5-9)
∑ cosθ
i =1
i
188
fy
fy
Qy θi fy
fy
fy
189
Appendix 3.5:
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
190
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 (γ ) = C Kd (γ )(K r + K p ) (A1-5)
where γ is a strain ratio ( γ = δ / H r ) and C Kd (γ ) is a modification factor of the secondary stiffness, which
takes into consideration the strain dependency. Also, the intercept force is defined as,
191
Qd (γ ) = C Qd (γ )σ p A p (A1-6)
where C Qd (γ ) is a modification factor of the yielding shear force and σ p is the yielding shear stress of lead.
F (γ ) = K d (γ ) δ + Qd (γ ) (A1-7)
The modification factors, C Kd (γ ) and C Qd (γ ) , are represented by the following formulas under 15 degrees
Celsius.
C Kd (γ ) C Qd (γ ) =
6 1.2
5 1
4 0.8
3 0.6
2 0.4
1 0.2
γ γ
0 0
0 1 2 3 0 1 2 3
192
Under a different temperature, t, the secondary stiffness and the yielding shear force are to be corrected by the
following formulas:
The primary stiffness of the lead rubber bearing, Ku, in Figure A1-4 is determined from the secondary stiffness, Kd,
as,
Ku = β Kd (A1-12)
where 10 ≤ β ≤ 15 .
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 / γ e (A1-13)
Fe = K d (γ e ) δ e + Qd (γ e ), δ e = γ e H r (A1-14)
The value, γ e = 0.01 , 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 (γ ) / dγ (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
193
2) Consideration of strength reduction by dissipated energy
Reference
1) Masanori Iiba, et.al., “Research on Characteristics of Isolators and Dampers under Multi-cyclic Earthquake Motions
and Effects on Response of Seismically Isolated Buildings”, Building Research Institute, National Research and
Development Agency, Building Research Data, No. 170, April 2016 (in Japanese).
2) Haruyuki Kitamura and Miyuki Omiya, “Design method for long period ground motion - Points to note when
dealing with long-period ground motion”, The Kenchiku Gijyutsu, No. 815, pp.116-125, 2017.12 (in Japanese)
From Reference 1), the yield shear stress of lead plug, τ , is expressed as
τ= {
τ 0 1 − (T TL )
αT
}, 0.4 + 0.25 (T TL )
αT = (A1-16)
Where,
τ 0 : Design value of the yield shear stress of lead plug = 15.0 (N/mm2)
T : Average temperature of lead plug
TL : Melting point of lead plug = 327.5 (oC)
For example, when T =20 (oC), τ is calculated to be 10.3 (N/mm2).
Qd ' ( γ ) = µ Qd ( γ ) (A1-17)
1 W pb
µ=
−0.06 + 1.25exp − (A1-18)
360 V pb
where
Qd ( γ ) : Intercept force without reduction
µ : Reduction factor
W pb : Dissipated energy
π
V pb = D pb 2 hpb : Volume of lead plug
4
hpb = ntr + ( n − 1) ts : Height of lead plug
n : number of rubber layer, tr : thickness of rubber layer, ts : thickness of steel plate
8.33 1 W pb
µ= −0.06 + 1.25exp − f ( D pb ) (A1-19)
7.97 360 V pb
where
f ( D pb ) = 0.16 D pb 0.31 : Correction value by the diameter of the lead plug, D pb (mm)
194
STERA_3D adopts Equation (A1-19). The reduction factor µ is plotted as a function of energy dissipation as
follows.
195
250 250
200 200
150 150
100 100
50 50
0 0
-30 -20 -10 -50 0 10 20 30 -30 -20 -10 -50 0 10 20 30
-100 -100
-150 -150
-200 -200
-250 -250
196
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 (γ ) = G1 (γ )× A / H r
u (γ ) − π × heq (γ ) / 2 + u (γ )× π × heq (γ ) / 2
G1 (γ ) = × Geq (γ ) (A2-1)
u (γ ) − π × heq (γ ) / 2
197
where γ : shear strain ( γ = δ / H r )
Ar : cross section area of the rubber
Hr : total height of the rubber.
Example)
Diameter: φ1500
Thickness of rubber: 7.5mm × 20 layers = 150mm
S1 = 49.7
S 2 = 10.0
2
Nominal compression stress: 10 N / mm
198
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 (γ ) = (1 − u )× Geq (γ ) (A2-6)
a) Elastic range
Under the strain level less than γ = 0.01 , the hysteresis is assumed to be linear with the secant stiffness at the
strain, that is:
h (γ
K 0 K=
= 0.01) (A2-10)
199
2) Consideration of strength reduction by dissipated energy
Reference
3) 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,
Geq ( γ ) C=
= K Geq ( γ 0.01) (A2-14)
heq ( γ ) C=
= h heq ( γ 0.01) (A2-15)
200
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.
201
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
202
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
∆µ = 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,
µ = µ 0 (T ) + ∆µ (E d ) (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 µ 0 = 0.029 from the catalog of a manufacture. The strength reduction by
energy dissipation will be expressed as,
Qd = (µ 0 + ∆µ (E d )) (Qd 0 µ 0 ) (A4-5)
203
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 = α k 0 x + (1 − α ) k 0 z (A5-1)
z =
{
Ax − β x z
N −1
z + γ x z
N
}ν (A5-2)
η
where, β , γ , and N are parameters that control the shape of the hysteresis loop, while A, ν , and η
are variables that control the material degradation.
β = β 0 δ yN and γ = γ 0 δ yN (A5-3)
A− z
N
{β sgn (xz ) + γ }ν ∂z ∂x
z = x = (A5-4)
η ∂x ∂t
204
This leads to the following expression for the continuum tangent
∂f ∂z A− z
N
{β sgn (xz ) + γ }ν
k= = α k 0 + (1 − α ) k 0 = α k 0 + (1 − α ) k 0 (A5-5)
∂x ∂x η
The evolution of material degradation is governed by the following choice of equations (Baber and
Noori 1985):
A = A0 − δ A e, ν = 1 + δν e, η = 1 + δ η e (A5-6)
e = (1 − α ) k 0 z x (A5-7)
z ( n +1) = z ( n ) + ∆t z (t 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) β sgn z ( n +1) + γ ν ( n +1)
∆t x( n +1) − x( n )
z ( n +1) = z ( n ) + ∆t (A5-10)
η ( n +1) ∆t
where
x( n +1) − x( n )
e( n +1) = e( n ) + ∆t (1 − α ) k0 z( n +1)
∆t (A5-12)
= e( n ) + (1 − α ) k0 z( n +1) (x(n +1) − x(n) )
Since
x( n +1) − x( n )
sgn z ( n +1) = sgn{(x( n +1) − x( n ) )z ( n +1) } (A5-13)
∆t
205
Φ
f (z ( n +1) ) = z ( n +1) − z ( n ) − (x ( n +1) − x( n ) ) = 0 (A5-14)
η ( n +1)
N
Φ = A( n +1) − z ( n +1) Ψ ν ( n +1) (A5-15)
z
new
z ( n +1) z old
( n +1)
(
Original f z ( n +1) ) Function derivatives f ' z ( n +1) ( )
x( n +1) − x( n ) x( n +1) − x( n )
e( n +1) = e( n ) + ∆t (1 − α ) k 0 z ( n +1) e' ( n +1) = ∆t (1 − α ) k 0
∆t ∆t
N
Φ = A( n +1) − z ( n +1) Ψ ν ( n +1) Φ ' = A' ( n +1) − N z ( n +1)
N −1
( )
sgn z ( n +1) Ψ ν ( n +1)
N
− z ( n +1) Ψν ' ( n +1)
Φ Φ 'η( n +1) − Φη '( n +1)
f (z ( n +1) ) = z ( n +1) − z ( n ) − (x ( n +1) − x( n ) ) f ' (z( n +1) ) = 1 − (x − x( n ) )
η ( n +1) η 2 ( n +1) ( n +1)
(A5-18)
206
The procedure can now be summarized as follows:
N
Φ = A( n +1) − z ( n +1) Ψ ν ( n +1)
Φ
f (z( n +1) ) = z( n +1) − z( n ) − (x( n +1) − x( n ) ) (A5-19)
η( n +1)
(b) Evaluate function derivatives
x( n +1) − x( n )
e' ( n +1) = ∆t (1 − α ) k 0
∆t
A' ( n +1) = −δ A e' ( n +1)
f (z( n+1) )
z(new
n +1) = z ( n +1) − (A5-21)
f ' (z( n+1) )
z(old
n +1) = z ( n +1) and z( n+1) = z(new
n +1) (A5-22)
207
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) β sgn z ( n +1) + γ ν ( n +1)
∆t x( n +1) − x( n )
z ( n +1) = z ( n ) + ∆t
η ( n +1) ∆t
∂f ( n+1) ∂z ( n+1)
k ( n+1) = = α k 0 + (1 − α ) k 0 (A5-23)
∂x( n+1) ∂x( n+1)
∂z( n+1) b4
= (A5-24)
∂x( n+1) b5
where
N
Φ = A( n +1) − z ( n +1) Ψ ν ( n +1)
b1 = (1 − α ) k 0 z ( n+1)
b2 = (1 − α ) k 0 (x( n+1) − x( n ) )
b3 =
(x ( n +1) − x( n ) )
η ( n+1)
Φ Φ
b4 = −b3δ Ab1 − b3 z( n+1) Ψδν b1 −
N
(x ( n +1) − x( n ) )δη b1 +
η 2
( n +1) η ( n+1)
208
3.6 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
3-6-2.
Q
Q
Qy
Qc αk 0
γ
Qu
k0
γc γy γu γ
The characteristic values, Qc , Q y , Qu are obtained based on the formulation described in the reference
(Paulay and Priestley, 1992).
209
The procedure to obtain the shear strength is shown below:
The compression strength of the masonry prism ( f ' m ) is determined by the following equation (Paulay and
Priestley, 1992),
f ' cb ( f 'tb +αf ' j )
f 'm = (3-6-2)
U u ( f 'tb +αf ' cb )
j
α= (3-6-3)
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)
210
(2) Shear strength by sliding shear failure
τ f = τ 0 + µσ 0 = τ 0 + tan φ σ 0 (3-6-5)
where,
τ0 : Cohesive capacity of the mortar beds (=0.04 f ' m ) (Paulay and Priestly, 1992)
µ : Sliding friction coefficient along the bed joint
µ = 0.654 + 0.000515 f ' j (Chen et.al, 2003)
σ0 : Compression stress ( = W / Aw = R sin θ / Aw )
W
V f = τ f AW = τ 0 + µ AW = τ 0 Aw + µW (3-6-6)
AW
Therefore,
τ 0 Aw
Vf = (3-6-8)
1 − µ tan θ
211
(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 ) (3-6-9)
ε 'm d m
γy = (3-6-10)
cos θ
where,
ε 'm : Compression strain at the maximum compression stress
( ε ' m =0.0018, Hossein and Kabeyasawa, 2004)
Initial elastic stiffness is assumed as (Madan et al., 1997)
k 0 = 2Q y / γ y (3-6-11)
Shear resistance and displacement at the ultimate stage are assumed as (Hossein & Kabeyasawa, 2004)
Qu = 0.3Q y (3-6-14)
γ u = 3.5(0.01hm − γ y ) (3-6-15)
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
212
b) Vertical springs
For the moment, the vertical springs of the element model in Figure 3-6-1 are assumed to be elastic springs.
k z = E m (tl w ) / 2 (3-6-17)
where,
Em : Modulus of elasticity of masonry prism (=550 f ' m , FEMA 356, 2000)
t : Thickness of masonry wall
lw : Width of masonry wall
213
3.7 Passive Damper
a) Hysteresis damper
B1 B B2
l
Q' xc , ε ' xc
A1 A2
A
Different types of hysteresis model are prepared for the force-deformation relationship of the spring.
k2
k2
k1 k1 − k2
x x + x
xy xm =
x x x
= +
214
(3) Degrading Tri-linear model
p My
P
n θm kr
ks θs θm
p θy p p
My
p pθm
= µα , k y
kr k y = = , µ
p θy pθ y
β Mm
= m µ ,
ks k= km n
M
n m
( nθ m − pθ s )
x x x
( β 0.5,
= = γ 0.5,
= N 10 ) ( β 0.5,
= = γ 0.5,
= N 2) ( β 0.1,
= = γ 0.9,
= N 2)
k2
k1 x
xy xm
215
b) Viscous damper
B1 B B2
A1 A2
A
Figure 3-7-4 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 (3-7-1)
where, Fk : force of the elastic spring
Fc : force of the dashpot
Fij : force between i-j nodes
216
The force of the elastic spring, Fk , is obtained as,
Fk = K d uk = K d (uij − uc ) (3-7-2)
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 3-7-5.
Fc
relief point
.
uc
∆u c (t )
u c (t ) = (3-7-5)
∆t
∆u c (t ) = u c (t ) − u c (t − ∆t ) (3-7-6)
217
Before the relief point of the dashpot, Equation (3-7-7) will be obtained by changing C2 → C1 , Qc = 0
as
(
K d u ij (t ) − u c (t − ∆t ) )
∆u c (t ) = (3-7-8)
C1
+ Kd
∆t
When the velocity of the dashpot is over the negative relief point, Equation (3-7-7) will be obtained by
changing Qc → −Qc ,
( )
K d u ij (t ) − u c (t − ∆t ) + Qc
∆u c (t ) = (3-7-9)
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 (3-7-10)
∆t
218
(2) Algorithm for viscous damper devise
Figure 3-7-7 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 (3-7-13)
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 3-7-8,
219
From Equations (3-7-13) and (3-7-14)
Fij (t )
+ u c (t ) = u ij (t ) (3-7-16)
Kd
Taking time differential and substituting Equation (3-7-15) give
1/ α
Fij (t ) Fij (t )
Kd
(
+ sgn Fij (t ) )
C
= u ij (t ) (3-7-17)
The numerical integration method, Runge-Kutta Method, can be used to solve the Equation (3-7-17).
1
Fij (t n +1 ) = Fij (t n ) + (k 0 (t n ) + 2k1 (t n ) + 2k 2 (t n ) + k 3 (t n ) ) (3-7-20)
6
1/ α
Fij (t n )
(
k 0 = u ij (t n ) − sgn Fij (t n ) )
C d
K ∆t
1/ α
Fij (t n ) + k 0 / 2
(
k1 = u ij (t n + ∆t / 2) − sgn Fij (t n ) + k 0 / 2
) C
d
K ∆t
1/ α
Fij (t n ) + k1 / 2
(
k 2 = u ij (t n + ∆t / 2) − sgn Fij (t n ) + k1 / 2
)C
d
K ∆t
1/ α
Fij (t n ) + k 2
(
k 3 = u ij (t n + ∆t ) − sgn Fij (t n ) + k 2 )
C
K d ∆t
220
In this algorithm, it is assumed as,
u ij (t n ) + u ij (t n + ∆t )
u ij (t n + ∆t / 2) = (3-7-21)
2
Fij, uij
Node i Node j
Fc, uc
u ij (t ) = u c (t ) (3-7-22)
∆u c (t ) ∆u ij (t )
u c (t ) = = (3-7-24)
∆t ∆t
Therefore,
α
∆uij (t ) ∆uij (t )
Fuj ( t ) = C sgn (3-7-25)
∆t ∆t
221
3.8 Ground Spring
a) When building and foundation on ground is subjected to an earthquake excitation, the system can be
divided into two parts: b-1) building and foundation with interaction forces and b-2) ground with zero-mass
foundation subjected to the reaction of interaction forces and an earthquake excitation, which can be
divided further into c-1) zero-mass foundation subjected to an earthquake excitation (kinematic
interaction) and c-2) zero-mass foundation subjected to the reaction of interaction forces (inertia
interaction).
G
c-2) Inertia interaction
222
In case of c-2), the force-displacement relationship is written as,
PG K H K HR uG
= (3-8-1)
M G K HS K R θG
where PG , M G are sway and rocking forces corresponding to the interaction forces between the
superstructure (building-foundation) and the ground, uG , θG are sway and rocking displacements. This
stiffness matrix is called “dynamic impedance matrix”.
If we neglect the coupling between sway and rocking degrees of freedom, the dynamic impedance matrix is
evaluated separately from the d-1) sway impedance K H and d-2) rocking impedance K R as follows:
PG K H 0 uG
= (3-8-2)
M G 0 K R θG
d-1) Sway
PG , uG
PG , uG
G
d-2) Rocking
M G , θG
PG K H K HR uG G
=
M G K HS K R θG
M=
G K R ⋅θG
223
This corresponds to the Sway-Rocking model as shown below:
M G , θG M G , θG
PG , uG PG , uG
KH
G G
KR
KH
KR
It is important to note that the input ground motion to an embedded foundation is smaller than the input
ground motion in the free field due to the influence of the embedding of the foundation. This effect is called
“kinematic interaction”.
224
3.8.2 Cone model to calculate the static stiffness
The cone model is proposed by Wolf [1994] for determining the dynamic stiffness of a foundation on the
ground. The foundation is assumed as an equivalent rigid cylinder and only vertically incident shear wave
is considered. In case of the stratified ground, a simplified formulation is proposed by IIba et.al. [2000]
without considering the reflection and refraction coefficients at the boundary of the soil layer to obtain the
static stiffness. The following formulation is adopted in the STERA_3D software.
Reference:
1) John P Wolf, Foundation Vibration Analysis Using Simple Physical Models, Prentice Hall, 1994
2) Iiba M., Miura K and Koyamada K, "Simplified Method for Static Soil Stiffness of Surface Foundation",
Proceedings of AIJ Annual Meeting, 303-304, AIJ, 2000. (in Japanese)
a) Sway spring
Consider a semi-infinite cone whose area increases in the depth direction. First, we show the calculation
method of the horizontal ground spring (sway spring) for the rectangular foundation 2b × 2c (ground
surface foundation or embedded foundation). The equivalent radius of a circle having the same area is
bc
obtained as r0 = 2 .
π
2r0
The forces of the minute portion at the distance z from the apex of the cone are:
・Shear force at the upper surface
∂u
=Q π=
r 2Gγ π r 2G (3-8-3)
∂z
・Shear force at the lower surface
2 2
dQ dz ∂ ∂u dz ∂u ∂ 2u
Q+ π 1 + r G u + dz =
dz = π 1 + r 2G + 2 dz
dz z ∂z ∂z z ∂z ∂z
(3-8-4)
Considering the static case ignoring the inertial force acting on the minute part, from the balancing of
forces,
225
dQ
Q + 0
dz − Q =
dz
2
dz 2 ∂u ∂ 2u ∂u
π 1 + r G + 2 dz − π r G
2
0
=
z ∂z ∂z ∂z
2
dz ∂u ∂ 2u ∂u ∂ 2u dz dz 2 ∂u ∂ 2u
1 + + dz − = dz + 2 + + dz= 0
z ∂z ∂z 2 ∂z ∂z 2 z z ∂z ∂z 2
Ignoring high-order small amount terms
∂ 2u 2 ∂u
+ 0
= (3-8-5)
∂z 2 z ∂z
Assuming that the displacement on the ground surface is U and the displacement at the depth d is 0 as
boundary conditions,
A A
U =+ B, 0 =+ B (3-8-7)
l d
From this, the coefficient A is
A=
(l + d ) l U (3-8-8)
d
Let Q0 be the shear force of the ground surface
∂u A l+d
π r02G
Q0 = = π r02G − 2 =
− π r02G U (3-8-9)
∂z l ld
Therefore, the horizontal spring K H on the ground surface is
Q l+d
KH = π r02G
− 0 = (3-8-10)
U ld
Assuming that d is infinite,
π r02G
KH = (3-8-11)
l
The horizontal spring of the circular rigid foundation on semi-infinite uniform ground is obtained
theoretically from the following formula.
8Gr0
KH = (3-8-12)
2 −ν
If the two springs are set to be equal, the distance l from the apex of the cone to the ground surface is
obtained as follows:
226
8Gr0 π r02G
= →=l
( 2 −ν ) π r (3-8-13)
0
2 −ν l 8
In case of the stratified ground, consider a truncated cone of thickness di from the i-th layer of stratified
ground and zi be the coordinate of the bottom of the i-th layer. The radius of the truncated cone ri at
depth zi is then calculated as follows from the geometric relationship.
zi
ri = r0 (3-8-14)
z0
The horizontal spring on the upper surface of this truncated cone is
2
z +d z zi π r02G1 Gi zi zi −1
=K π=
r Gi i −1 i π i −1 r0 Gi=
i 2
(3-8-15)
zi −1 ( zi − zi −1 ) z0 G1 z0 ( zi − zi −1 )
H i −1
zi −1di z0
The horizontal spring K hb at the base bottom position is obtained as a synthetic spring in which
horizontal springs of each layer are connected in series.
1 n −1
1
=∑ i (3-8-16)
K hb i =0 K H
However, in the bottom layer,
π r02G1 Gn zn zn −1 π r02G1 Gn zn −1
=K n −1
→ ( zn → ∞ ) (3-8-17)
G1 z0 ( zn − zn −1 )
H
z0 z0 G1 z0
G zi zi −1 G z
αi = i ( i = 1, 2, , n − 1) , α n = n n−1
G1 z0 ( zi − zi −1 ) G1 z0
2 −ν 1
z0 = π r0
8
227
b) Rocking spring
Rotational spring can be obtained as follows, similar to the method for determining horizontal spring. For
the rectangular foundation 2b × 2c (ground surface foundation or embedded foundation, 2b is the
length in rotational direction), the equivalent radius of a circle having the same moment of inertia is
( 2b ) ( 2c ) .
3
obtained as rr 0 = 4
3π
π
) The moment of inertia of a circle Ic = rr40
4
( 2b ) ( 2c )
3
2r0
∂M
M+ dz
∂z
The forces of the minute portion at the distance z from the apex of the cone are:
・Moment at the upper surface
∂θ π r 4 ∂θ
M=
− EI E r0
= (3-8-19)
∂z 4 ∂z
・Moment at the lower surface
4 4
dM π dz ∂ ∂θ π dz 4 ∂θ ∂ 2θ
M+ − 1 + rr 0 E θ +
dz = − 1 + rr 0 E
dz = + 2 dz
dz 4 z ∂z ∂z 4 z ∂z ∂z
(3-8-20)
Considering the static case ignoring the inertial force acting on the minute part, from the balancing of
forces,
dM
M + 0
dz − M =
dz
4
dz 4 ∂θ ∂ 2θ rr40 ∂u
π
− 1 + rr 0 E + 2 dz + E 0
=
4 z ∂z ∂z 4 ∂z
Ignoring high-order small amount terms
∂ 2θ 4 ∂θ
+ 0
= (3-8-21)
∂z 2 z ∂z
228
A
θ
= +B (3-8-22)
z3
where A and B as undetermined coefficients.
Assuming that the rotational displacement on the ground surface is Θ and the displacement at the depth d
is 0 as boundary conditions,
A A
Θ= + B, 0= +B (3-8-23)
( lr + d )
3
lr3
( lr + d ) lr 3 Θ
3
=A (3-8-24)
( l r + d ) − lr 3
3
3 ( lr + d )
3
π r 4 3 A π rr40
π rr40
∂u
M0 =
− E − r0 E − 4 =
= E Θ (3-8-25)
4 ∂z 4 lr 4 {
( lr + d ) − lr3 lr
3
}
Therefore, the rotational spring K R on the ground surface is
3 ( lr + d )
3
M 0 π rr40
K
= = E (3-8-26)
R
Θ 4 { ( lr + d ) − lr3 lr
3
}
Assuming that d is infinite,
3π rr40 E
KR = (3-8-27)
4lr
The horizontal spring of the circular rigid foundation on semi-infinite uniform ground is obtained
theoretically from the following formula.
8Grr30
KR = (3-8-28)
3 (1 −ν )
If the two springs are set to be equal, the distance lr from the apex of the cone to the ground surface is
obtained as follows:
8Grr30
=
3π rr40 E 3π rr40 E
= 2G (1 +ν ) → =
lr
9 1 −ν 2
π rr 0
( ) (3-8-29)
3 (1 −ν ) 4lr 4lr 16
229
In case of the stratified ground, consider a truncated cone of thickness di from the i-th layer of stratified
ground and zri be the coordinate of the bottom of the i-th layer. The radius of the truncated cone rri at
depth zri is then calculated as follows from the geometric relationship.
zri
rri = rr 0 (3-8-30)
zr 0
The rotational spring on the upper surface of this truncated cone is
3 ( zri −1 + di )
3
i π rri4−1 3π rr40 E1 Ei zri3 zri3 −1
K = Ei 3 3 (3-8-31)
{ } ( )
R
4 ( zri −1 + di ) − zri3 −1 zri −1 4 zr 0 E1 zr 0 zri − zri −1
3 3
The rotational spring K rb at the base bottom position is obtained as a synthetic spring in which rotational
springs of each layer are connected in series.
1 n −1
1
=∑ i (3-8-32)
K rb i =0 K R
However, in the bottom layer,
zr 0 =
(
9 1 −ν 12 )πr
r0
16
230
3.8.3 Embedded foundation
K hb , K rb K he , K re
In case of embedded spread foundation, the resistances at the side of the foundation K he , K re can be
expected in addition to the resistances K hb , K rb at the base of the foundation. That is,
K
= h K hb + K he
(3-8-35)
K
= r K rb + K re
where
De Ghe
K he = ξ he K hb (3-8-36)
r0 Ghb
D De Ghe
3
=K re ξ re K rb 2.3 + 0.58
e
(3-8-37)
rr 0 rr 0 Ghb
m
∑G H i i
( 2 −ν ) K hb
=Ghe =
i =1
m
, Ghb (3-8-38)
8r0
∑ Hi
i =1
De is the depth of the foundation. ξ he and ξ re are the earth pressure reduction coefficients of
horizontal and rotational directions at the side of the foundation and they are set to 0.5 when considering
only the side receiving the reaction force from ground at the time of the earthquake. m is the number of
soil layers from the surface to the bottom at the side the foundation where the earth pressure acts. ν is the
average Poisson's ratio of the ground under the foundation base. The damping at the embedded part is not
considered.
231
3.8.4 Radiation damping
The static stiffness obtained by the cone model alone can not express the radiation damping that the
energy of ground shaking spreads to a distance.
To evaluate the radiation damping, we consider a semi-infinite earth column with the same area of the
foundation where a shear wave travels downward when the foundation sways harmonically in a horizontal
direction.
The wave travels in the earth column can be expressed as the solution of the wave equation.
∂ 2u 2
2 ∂ u G
= V=
s , Vs (3-8-39)
∂t 2 ∂2 z ρ
where G is the shear modulus of the soil, ρ is the density of the soil, and Vs is the shear wave
velocity.
ipt
When the foundation sways harmonically as ue , the solution of the wave equation is
u ( z , t ) = ue
ip ( t − z /Vs )
(3-8-40)
π r4
where I= is the second moment of inertia for a circular foundation with the radius r
4
3.4
η= is the coefficient for vertical wave velocity, where ν is the Poisson’s ratio
π (1 −ν )
C = ρVA
G, ν , ρ
232
In case of the stratified ground, we can use the following formula for the radiation damping
CH = ρeVe A (3-8-44)
3.4
CR = ρe (ηeVe ) I , ηe = (3-8-45)
π (1 −ν e )
where ρe is the average density, Ve is the average shear wave velocity and ν e is the average shear
modulus defined by the weighted average by depth of layers under the basement as
n n n
∑ ρi d i ∑Vi di ∑ν d i i
ρe = i =1
n
, Ve = i =1
n
, νe = i =1
n
(3-8-46)
∑d
i =1
i ∑d
i =1
i ∑d
i =1
i
The damping effect of the soil material can be considered by setting the shear modulus to the following
complex shear modulus.
G * G (1 + 2ih )
= (3-8-47)
where h is the damping factor of the soil. As a result, the dynamic stiffness obtained from the cone model
becomes also complex value as,
Furthermore, the damping coefficient is obtained from the imaginary part of the complex stiffness under the
periodic vibration of the circular frequency ω.
( K + iωC ) x =
( K + iK ') x =
K (1 + 2ih ) x
K ' 2hK
Therefore, C
= = (3-8-49)
ω ω
233
3.8.6 Impedance matrix
It is known that radiation damping is likely to occur in a frequency band higher than the dominant
frequency of the ground ( f G ), and the effect is greater at higher frequencies. Therefore, the damping is
evaluated separately for a lower frequency side and a higher frequency side than the dominant frequency.
b) In case of f > f G (ω > ωG ) for Sway spring and f > 2 f G (ω > 2ωG ) for Rocking spring
In a matrix form
234
3.8.7 Pile foundation
Now we discuss the Sway and Rocking springs for the foundation with piles.
1) Sway
PG , uG
G
PG , uG
Inertia interaction
M G , θG
PG , uG
G
2) Rocking
M G , θG
KV = EAα
( ) ( ),
k B 1 + e −2α L + EAα 1 − e −2α L
α=
kS
(3-8-58)
kB (1 − e ) + EAα (1 + e )
−2α L −2α L
EA
where,
E : Young’s Modulus of the pile, A : Area of the pile, L : Length of the pile
kS : Vertical spring of the soil surrounding the pile, k B : Vertical spring at the bottom of the pile
235
)
P0
u0
Pz
uz
uz
kS u z dz dz
u z + du z
Pz + dPz
uL
PL
a-1) Equilibrium condition of the vertical forces in a pile
The equilibrium condition of the vertical forces in a small segment is
dP k=
= S u z dz 0 (3-8-59)
The axial strain in the segment is obtained as
du z P
= − z (3-8-60)
dz EA
Therefore
dP d 2u z
= EA= kS uz (3-8-61)
dz dz 2
The solution of this second order differential equation is
kS
u z c1eα z + c2 e −α z , α =
= (3-8-62)
EA
Also
= (
Pz EAα c2 e −α z − c1eα z ) (3-8-63)
uL c1eα L + c2 e −α L
= (3-8-65)
236
EAα uL − P0 e −α L EAα uL + P0 eα L
c1 = , c2 = (3-8-66)
(
EAα eα L + e −α L ) (
EAα eα L + e −α L )
The force at the bottom of the pile PL is
2 K P P0
PL = (3-8-68)
KP e ( αL
+e −α L
) + EAα eα L − e −α L ( )
and
2 P0
uL = (3-8-69)
KP e ( αL
+e −α L
) + EAα eα L − e −α L ( )
The displacement at the head of the pile is
u0= c1 + c2 (3-8-68)
Therefore, the stiffness of the vertical spring at the head of the pile is
K=
P0 EAα ( c2 − c1 )
= =
EAα P0 eα L + P0 e −α L ( )
0
u0 c1 + c2 2 EAα uL − P0 e −α L + P0 eα L
=
(
EAα eα L + e −α L )
2
2 EAα + eα L − e −α L ( )
kB e ( αL
+e −α L
) αL
+ EAα e − e −α L
( ) (3-8-69)
EAα e( αL
+e −α L
){k ( e + e ) + EAα ( e − e )}
B
αL −α L αL −α L
=
(
4 EAα + eα L − e −α L ){k ( e + e ) + EAα ( e − e )}
B
αL −α L αL −α L
= EAα
kB (e αL
+ e −α L ) + EAα ( e − e ) αL −α L
kB (e αL
− e −α L ) + EAα ( e + e ) αL −α L
237
a-2) Vertical spring of the soil surrounding the pile
The vertical spring of the soil surrounding the pile k S is obtained as the friction resistance of soil
surrounding the soil (Randolph and Wroth, 1978).
(a) Concentric cylinder around loaded pile (b) Stresses in soil element
Reference: Randolph M.F and Wroth C.P, “Analysis and deformation of vertically loaded piles”, Journal of
Geotechnical Engineering 104(12): 1465-1487. 1978.
From the equilibrium condition of vertical forces
∂τ ∂σ y dr dr
τ + ( r + dr ) d θ dy − τ rd θ dy + σ y + dy r + dθ dr − σ y r + dθ dr =
0
∂r ∂y 2 2
(3-8-70)
Neglecting higher order
∂ (τ r ) ∂σ y
+r 0
= (3-8-71)
∂r ∂y
Assuming the stress change along the depth ∂σ y ∂y is negligible, the second term will be zero. Then,
∂ (τ r )
=0 (3-8-72)
∂r
Integrating from the pile radius r0 to r ,
r
∫ d (τ r ) =⇒
r0
0 τ ( r ) r − τ ( r0 ) r0 =
0
τ ( r0 ) r0 τ 0 r0
τ (r )
= = (3-8-73)
r r
Assuming the deformation along the radius du is smaller than the deformation along the depth dw , the
shear stain is
∂u ∂w dw τ ( r ) τ r
γ= + ≈ = = 00 (3-8-74)
∂z ∂r dr G ( r ) rG ( r )
238
The vertical shear deformation is obtained by integrating from r0 to rm ,
rm 1 τ 0 r0 rm
= 0 r0 ∫r
wS τ= dr ln (3-8-75)
0 rG G r0
Randolf and Worth proposed the following empirical formula for the radius rm
2π G
=P (=
2π r0 )τ 0 wS (3-8-77)
ln ( rm r0 )
Therefore, the vertical spring of the soil surrounding the pile k S is
2π Ge
kS = , rm 2.5 L (1 −ν e )
= (3-8-78)
ln ( rm r0 )
where,
1 n 1 n
Ge = ∑ ii
L i =1
( G d ) : average shear modulus, ν e = ∑ (ν i di ) : average Poisson ratio
L i =1
239
b) Horizontal stiffness of a single pile
H
x y
p ( x)
The flexural deformation of the infinite pile under horizontal load at the top of the pile is
d4y
EI + p ( x) =
0 (3-8-80)
dx 4
Assuming
p ( x ) = kh By (3-8-81)
H = −Q ( 0 ) (3-8-85)
Therefore,
Q (0)d3y H H
=
− 3 −4C1β 3 =
= − C1 = (3-8-86)
EI dx x =0
EI 4 β 3 EI
240
The horizontal deformation of the pile is
H
=y e − β x ( sin β x + cos β x ) (3-8-87)
4 β 3 EI
The deformation of the pile head is
H
y= − (3-8-88)
4 β 3 EI
Therefore, the horizontal stiffness is
( 4 EI ) ( kh B )
1/4 3/4
= β 3 EI
K h 4= (3-8-89)
Francis (1964) proposed the following formula for the horizontal ground spring per unit length of a
single pile:
1/12
1.3ES ES B 4
k=
fS k=h B (3-8-90)
1 −ν S 2 EP I P
where
EP : Young’s modulus of a pile, I P : Moment of inertia of a pile
ES : Young’s modulus of soil, ν P : Poisson ratio of soil
This formula is based on the study by Biot (1937) with respect to the ground spring against bending of
an infinite beam on ground and is modified by Visic (1961). Francis extended this concept to the pile
considered that there is ground on both sides of the beam and doubled the ground stiffness.
Reference:
1) Francis A. J, Analysis of Pile Groups with Flexural Resistance, Journal of the Soil Mechanics and
Foundations Division, 1964, Vol. 90, Issue 3, Pg. 1-32
2) Biot, M. A. Bending of an infinite beam on an elastic foundation. J. Appl. Mech., 1937, 4, 1, Al-A7
3) Vesic A.B, Bending of beams resting on isotropic elastic solid, Journal of the Engineering Mechanics
Division, 1961, Vol. 87, Issue 2, Pg. 35-54
241
b-2) Horizontal damping of a single pile
Gazetas proposed the following formula for the horizontal damping per unit length of a single pile:
Reference: Gazetas, G. and Dobry, R, Horizontal Response of Piles in Layered Soils, J. Geo tech. Engrg.
Div.,ASCE, Vol.110, pp.20-40, 1984
= (
k ' fSi 0.5 k fS ( i −1) H i −1 + k fSi H i ) (3-8-92)
= (
c 'gSi 0.5 cgS ( i −1) H i −1 + cgSi H i ) (3-8-93)
242
c) Impedance of group piles
In case of group piles, the impedance of the foundation can not be obtained from the simple addition
of the impedances of individual piles because of the interaction of piles.
The horizontal stiffness of group piles is obtained from the horizontal stiffness of a single pile as,
K HG = N P β H K HS (3-8-94)
where
direction,
−0.54
( N y 2)
−0.74( S B )
−0.43
−0.59( S B )
β Hx = 0.4 ( S B ) ( N x 2)
0.3
(3-8-95)
where
K HS = ( 4 EP I P ) kS 3/4
1/4
(3-8-96)
where
N P β H ( 4 EP I=
P) ( 4 N P EP I P )
1/4 1/4
= P β H K HS
K HG N= k S 3/4 kG 3/4 (3-8-97)
kG = N P β H 4/3 k S (3-8-98)
243
For the horizontal damping, the group effect is assumed negligible, and the horizontal damping of
where
cHG : damping coefficient of group piles, cHS : damping coefficient of a single pile
In evaluating the horizontal ground stiffness of the group pile K HG in layered ground, it is necessary
to determine the value of the stiffness coefficient kG which represents the average stiffness coefficient in
P
u0 = u ( 0 )
ui
ui +1
where kGi = N P β H
4/3
k Si : horizontal stiffness of group pile at i-th layer from Eq.(3-8-89)
Step. 2 The flexural deformation of a pile under the horizontal load P at the top is approximated by
P kG
=u e − β x ( sin β x + cos β x ) , β=4 (3-8-100)
4 N P EP I P β 3 4 EI
K HG 2 =
∑k u
Gi i
(3-8-101)
u0
Step. 3 Update kG as
kG = ( K HG 2 ) ( 4 N P EP I P )
4/3 1/4
(3-8-102)
244
c-2) Group effect in horizontal direction (damping)
The damping effect of the soil material is considered as
where hGi is the damping factor of the soil in i-th layer. The horizontal damping at the top of group piles
can be calculated by,
hHG =
∑h u
Gi i
(3-8-104)
u0
In the same way, the horizontal radiation damping at the top of group piles can be calculated by
cHG =
∑c u
Gi i
(3-8-105)
u0
where cGi = N p cSi
The group effect in rotational direction is assumed negligible and the coefficient of group effect is
one. Therefore, the rotational stiffness is calculated from the vertical stiffness of individual pile as
m
K RGx = ∑ KVi yi2 : around x-axis (3-8-106)
im
=1
K RGy = ∑ KVi xi2 : around x-axis (3-8-107)
i =1
where
xi , yi : distance from the center of ration in x, y directions
245
3.8.8 Equivalent period and damping factor considering soil structure interaction
a) Equivalent period
δB δS δB δR δS δB
F m F m F
m
K B , CB H
K S , CS
K R , CR
Force and deformation
δB = F KB δ S = F KS θR H ( M KR )
δ R H=
=
δ δ=
= B F KB δ = δ B + δ S = F (1 K B +=
1 KS ) ( FH K R ) F ( K R H 2 )
H=
δ = δB + δS + δR
(
= F 1 KB +1 KS +1 ( KR H 2 ) )
Stiffness
K = KB 1
K= 1
1 KB +1 KS K=
1 KB +1 KS +1 ( KR H 2 )
m m mH 2
TB = 2π TS = 2π TR = 2π
KB KS KR
m m m m m m m mH 2
= π
T 2= TB = π
T 2= 2π + T 2π
= = 2π + +
K K KB KS K KB KS KR
= TB2 + TS2 = TB2 + TS2 + TR2
246
b) Equivalent damping
This is the case to define the damping force to be independent to the frequency of excitation.
This type of damping is called “material damping”.
2 2 2
K K K T T T
h= hB + hS + h = B hB + S hS + R hR
2 R
KB KS KR H T T T
247
b-2) Equivalent damping for viscous damping
Force including damping force is
C K
F = Cδ + K δ = m δ + δ = m 2hωδ + ω 2δ
m m
( )
For a harmonic excitation δ = aeipt
p
m (ω 2 + 2hω pi ) δ =
F= mω 2 1 + 2h i δ
ω
This is the case to define the damping force to be dependent to the frequency of excitation.
This type of damping is called “viscous damping”.
1 p 1 p 1 p 1 p
2
1 − 2h i = 2
1 − 2 h i + 2
1 − 2 h i + 2
1 − 2 h i
ω mωB
B S R
mω ωB mωS ωS mωR ωR
p K p K p K p
h = hB + hS + h
2 R
ω K B ω B K S ωS K R H ωR
In case of the resonance frequency, p =ω
3 3 3
ω ω ω
h= hB + hS + hR
ωB ωS ωR
or
3 3 3
T T T
h = B hB + S hS + R hR
T T T
248
4. Freedom Vector
Each node has six degrees of freedom and the freedom number is defined as shown in the figure below.
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 the figure below, 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.
Node 1 0
|
Node 4 0
1
Node 5
5 6
8
9
7 8
Node 6
16
1 2
17
Node 7
24 3 4
25
Node 8
32 shear deformation of connection
249
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)
θ 0 0
zA 1 θ zG
G: center of gravity
θ zG
G G u xG
l xA
l yA
u yG θ zA
A
A u xA
u yA
Figure 4-3-2 Rigid floor assumption
250
In case of the structure in the figure below, in addition to the original nodes, new nodes for the center of
gravity are defined as “Node 5” and “Node 10”. 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
251
(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, θ y1 and θ y 2 , are dependent to the vertical
displacements, δ z1 and δ 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.
θ y2
θ y1
δ z 2 − δ z1
θ y1 = θ y 2 =
δ z2 w
δ z1
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
θ y1 = 0 − 1 / w 1 / w δ z1 (4-3-2)
θ 0 − 1 / w 1 / w δ
y2 z 2
θ x1
u y1 1 0 0 u y 2
δ z1
θ x1 = 0 1 / w − 1 / w δ z1 (4-3-3)
θ 0 1 / w − 1 / w δ
1 x2 z2
θ x2
u y1
δ z2
2 w
u y2
Figure 4-3-5 Relationship between node displacements for a wall element (Y-wall)
252
For example, in case of the structure in the figure below, 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
253
(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, θ y1
and θ y 2 , are dependent to the vertical displacements, δ z1 and δ 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.
θ yN
θ y2
θ y1
δ z1 δ z2 δ zN
u x1 u x2 u xN
w1
δ zN − δ z1 N −1
θ y1 = θ y 2 = = θ yN = , L = ∑ wi
L k =1
i −1
δ zi = δ z1 + θ yi Li = (1 − Li / L )δ z1 + (Li / L )δ zN , Li = ∑ wi
k =1
u x1 = u x 2 = = u xN
In a matrix form;
δ
θ yi = [− 1 / L 1 / L ] z1 (4-3-4)
δ zN
δ
δ zi = [1 − Li / L Li / L ] z1 (4-3-5)
δ zN
254
In case of Y-direction wall, the relationship can be written as;
θ x1
δ z1
θ x2
1
u y1
δ z2
2 w1
u y2
δ zN
N δ z1 − δ zN N −1
θ x1 = θ x 2 = = θ xN = , L = ∑ wi
u yN L k =1
i −1
δ zi = δ z1 − θ xi Li = (1 − Li / L )δ z1 + (Li / L )δ zN , Li = ∑ wi
k =1
u y1 = u y 2 = = u yN
In a matrix form;
δ
θ xi = [1 / L − 1 / L ] z1 (4-3-6)
δ zN
δ
δ zi = [1 − Li / L Li / L ] z1 (4-3-7)
δ zN
255
(4) Including ground springs
In case there are ground springs (sway and rocking springs) at the basement of the building, the floor
diaphragm of the basement is assumed to be rigid for both in-plane and out-of-plane deformation and the
freedoms other than sway and rocking freedoms are restricted at the center of gravity.
1 4
5
2
(a) Sway freedoms (b) Rocking freedoms
In case of the structure in the Figure below, by eliminating dependent freedoms, the total number of
freedom becomes 21.
256
0
Node 1-4 :
0
1
2
0
Node 5 3
4
0
0
0
0
0
5
Node 6 6
0
0 6 7
0
0 10
0
0 8 9
7
Node 7 8
0
0 1 2
0
0 5
0
0 3 4
9
Node 8 10
11
0
12
13
0
0
14
15
Node 9 16
0
17
18
19
20
0
0
Node 10 0
21
0
0
Freedom vector
257
4.4 Transformation matrix of 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
θ 0 0
zA 1 θ 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
δ zA u xG k
= u l
θ xA yG
θ yA δ zG
θ
θ zA 0 0 1
xG
θ yG
θ 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
258
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
θ y1 = 0 − 1 / w 1 / w δ y1
θ 0 − 1 / w 1 / w δ
y2 y2
It can be arranged into the transformation matrix between the freedom vectors of all nodes;
p q r
u x1 0 1 0
j θ y1 − 1/ w 0 1/ w δ y1 p
=
θ y 2 − 1/ w 0 1/ w u x 2 q
δ 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
259
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 θ y8 θ y9 δ x8 δ 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 δ x8 δ x9
36 0 8 9
37 0 θ y8 θ 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
260
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 θ y8 θ y9 δ x8 δ x 9 }T
⇒ {45 51 47 53 43 49}
T
⇒ {u z 8 u x10 θ z10 }
T
u z 9 θ y8 θ y9
⇒ {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
θ y 8 1 θ y 8 θ y 8
= = [TixB ] (4-4-5)
θ y 9 1 θ y 9 θ y 9
δ x8 1 l y 8 u x10 u x10
δ x 9 0 1 l y 9 θ z10 θ z10
261
4.5 Stiffness matrix corresponding to independent degrees of freedom
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 θ y 8
=
M y 9 10 k10,10 k10,11 k10,13 θ y 9
Px10 11 sym. k11,11 k11,13 u x10
M z10 13 k12,12 θ 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 (2-1-11).
262
u zA
u
zB u1
θ yA u
2
= [TixB ] (2-1-11)
θ yB
δ xA un
δ xB
And the constitutive equation of the X-beam is also described in Equation (2-1-16).
P1 u1
P u
2 2
= [K ]
xB (2-1-17)
Pn u n
Using the same procedure in Figure 4-5-1, the element stiffness matrix is added into the global stiffness
matrix.
Mass is assigned in each node. The inertia force at the node will be also transformed according to the
transformation of the variables. Here, the rotational inertia at each node is ignored.
MG
PxG
G
mA PxA
A PyG
PyA
263
Under the rigid floor assumption, the in-plane freedoms at the nodes in a floor are represented by the
freedoms at the center of gravity of the same floor. Therefore,
G: center of gravity
M zG
G G PxG
l xA
l yA
PyG
A
A PxA
PyA
On the other hand, the inertia force at the center of gravity is calculated as,
Therefore,
PxG 1 0 0 PxA mA 0 0 uxA
0
0 PyA = − [TA ] 0 uyA
T
PyG = 0 1 mA
1 0 0 θzA
M zG l yA −lxA 0 0
(4-5-4)
mA 0 0 uxG m 0 l yA mA u
T A
xG
− [TA ] 0 mA 0 [TA ] uyG =
= − 0 mA −lxA mA uyG
θ
0 0 0
θ zG 2 2
l yA mA −lxA mA lxA + l yA mA zG ( )
If we ignore the off-diagonal components,
PxG m 0 0 u
A xG
P
yG = − 0 mA 0 uyG (4-5-5)
M zG 0 0 2
mA lxA ( 2
+ l yA ) θ zG
Taking the sum of the inertia force from the all nodes at the same floor,
PxG mG 0 0 uxG N N
0 m u , m =
P
yG =
− G 0 yG G ∑i i G ∑i mi lix2 + liy2
m , I = (4-5-6) ( )
M 0
0 I G θzG
zG
264
(2) Including rigid beam
The wall element model has rigid beams at the top and bottom of the wall, and the horizontal displacement
in the wall panel plane, u x 2 , is dependent to the displacement, u x1 .
u x1 1 0 u x1 u
δ z1 δ z2 =
= [T ] u x1
u x1 u x2 u x 2 1 0 u x 2 x2
P
=x1 ( m1 + m2 ) ux1
Px1 Px 2 Px1
m1 m2 m1 + m2
1 2 1 2
Pz1 Pz 2 Pz1 Pz 2
m1 m2 m1 m2
1 2 1 2
265
(3) Series of rigid beams
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.
δ z1 δ zi δ zN
u x1 u xi u xN
Li
L
The all horizontal displacements at the nodes are dependent to the horizontal displacement of
the first node, u x1 .
u x=
1 u x=
2 = u xN
Also, the vertical displacement at the middle node δ zi is dependent to the vertical displacements of
the nodes at both ends, δ z1 δ zN .
Li Li
δ zi =
1 − δ z1 + L δ zN
L
266
The vertical mass is
N L
Pz1 ∑ 1 − i
= mi δ z1
i =1 L
N L
PzN = ∑ i mi δ zN
i =1 L
m1 mi mN
1 i N 1 N N
Li
L
267
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
268
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
269
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 rotational inertia at the center of gravity, IG, along the z-axis 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 rotational inertia 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
δ zi Mi M i
[M ] = ⇒ (5-1-5)
θ xi Ii Ii
θ yi Ii I
i
θ 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
270
In case a complete rigid floor such as a foundation slab for the ground springs, we need to calculate the
rotational inertia at the center of gravity along each axis.
b
X
Y a
If the mass is located at each node, as already mentioned, the rotational inertia at the center of gravity, IG,
along the Z-axis is obtained as
X
=IZ ρ r dV ∑ M ( l
∫=
2
i ix
2
+ liy 2 ) (5-1-9) Mi
liy
G
IG
lix MG
271
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 θ y 8 Example of beam element
=
M y 9 10 k10,10 k10,11 k10,13 θ y 9
Px10 11 sym. k11,11 k11,13 u x10
M z10 13 k12,12 θ 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
272
5.3 Modal analysis
where [K ] is the stiffness matrix and [M ] is the mass matrix in the form;
m1 0 0
0 m2 0
[M ] = (5-3-2)
0 0 mn
Substituting into the equilibrium equation, the generalized eigen problem is obtained as,
This eigen problem yields the n eigen solutions (ω , {φ }), (ω , {φ }),, (ω , {φ }) where the
2
1 1
2
2 2
2
n n
eigen vectors are M-orthonormalized as,
n
]{φ j } ∑ mkφ=
{φi } [ M=
T
i ,kφ j ,k 0 ; i≠ j (5-3-5)
k =1
)
Let’s assume two different set of eigen solutions (ω ,{φ }) , (ω ,{φ }) .
i
2
i
2
j j
{φi } [ K ]{φ j } {=
φ } ([ K ]{φ }) {φ j } [ M ]{φi } ω j 2 {φi } [ M ]{φ j }
T T T T
= j i ω j 2= (5-3-7)
(ω i
2
)
− ω j 2 {φi } [ M ]{φ j } =
T
0 (5-3-8)
273
2) Modal decomposition of equilibrium equation
Defining a matrix [Φ ] whose columns are the eigenvectors and a diagonal matrix Ω [ ]
2
ω12
ω 2
[Φ ] = [{φ1 } {φ 2 } {φ n }] , [ ]
Ω =
2
2
(5-3-9)
2
ω n
{u (t )} = [Φ ]{q(t )} (5-3-10)
Then,
Multiplying [Φ ] ,
T
where
m1
m2
[Φ ] [ M ][Φ=] M= , = {φi } [ M ]{φi }
T T
mi (5-3-13)
mn
k1
k2
[ ] [ ][ ] [ ][ ] [ ][ ] [ ] M = K , k = ω 2m
T T
Φ K Φ = Ω Φ M Φ = Ω =
i i i
kn
(5-3-14)
c1
c2
[Φ ] [C ][Φ=] C=
T
(5-3-15)
cn
where, M , C and K are called generalized modal mass, modal damping and modal
stiffness matrix, respectively.
274
Therefore,
[Φ ] {P}
M {q} + C {q} + K {q} =
T
(5-3-16)
mi qi (t ) + ci qi (t ) + ki qi (t ) =
ri (t ) (5-3-17)
X0 (t )
where ri (t ) = {φi } { P (t )} = − {φi } [ M ][U ] Y0 (t )
T T
(5-3-18)
Z (t )
0
X0 (t )
T
− {βi } Y0 (t ) =
qi (t ) + 2hiωi qi (t ) + ωi 2 qi (t ) = − {βi , x X0 (t ) + βi ,yY0 (t ) + βi , x Z0 (t )}
Z (t )
0
(5-3-19)
where
{β βi ,y βi ,z }
T
{β i }
T
= = (5-3-20)
{φi } [ M ]{φi } {φi } [ M ]{φi }
T T i,x
βi , x , y , z = (5-3-21)
{φi } [ M ]{φi }
T
βi , x , y , z is the coefficient when you decompose the vector {U x ,y,z } into mode vectors as,
n
{U } =
[ Φ ]{β } =
x ,y,z ∑ β {φ } x, y, z i, x, y, z i (5-3-22)
i =1
)
[Φ ] [ M ] ,
T
Multiplying
[Φ ] [ M ]{U x,y,z } =
[Φ ] [ M ][Φ ]{β x, y , z } =
M {β x , y , z }
T T
(5-3-23)
Therefore,
{=
β } [Φ ] [ M ]{U x,y,z }
−1 T
x, y, z M (5-3-24)
275
Equation (5-3-17) can be decomposed again as,
− X0 (t )
xi (t ) + 2hiωi xi (t ) + ωi 2 xi (t ) =
y (t ) + 2h ω y (t ) + ω 2 y (t ) =
i i i i i −Y (t )
i 0 (5-3-25)
2
− Z0 (t )
zi (t ) + 2hiωi zi (t ) + ωi zi (t ) =
and
qi (t ) = βi , x xi (t ) + βi ,y yi (t ) + βi ,z zi (t ) (5-3-26)
n n n n
[Φ ]{q(t )} =
{u (t )} = ∑ {φi } qi (t ) =
∑ βi, x {φi } xi (t ) + ∑ βi,y {φi } yi (t ) + ∑ βi,z {φi } zi (t )
=i 1 =i 1 =i 1 =i 1
(5-3-27)
M e ,i = βi 2 mi (5-3-32)
({φi } [ M ]{U })
T 2
{φi } [ M ]{U= }m
T
βi {φi } [ M ]=
{U }
T
M e ,i β=
= 2
i mi βi (5-3-33)
{φi } [ M ]{φi } {φi } [ M ]{φi }
T i T
276
The sum of effective modal mass is,
n n
∑ M e ,i βi {φi } [ M ]{U }
∑= {U } [ M ]{U }
T T
= (5-3-34)
=i 1 =i 1
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 (5-3-10) as,
[Φ ] [ M ][Φ ]{q(t )} =
[Φ ] [ M ]{u (t )} = M {q (t )}
T T
(5-3-35)
Therefore,
−1 −1
{q}t M [ Φ ] [ M ]{u} {q} M [ Φ ] [ M ]{u}
T T
=
0=
=t 0 t 0=t 0
, = (5-3-36)
277
5.4 Damping matrix
1) Proportional damping
The mass-proportional damping and the stiffness-proportional damping are defined as,
[C ] = a0 [M ] and [C ] = a1 [K ] (5-4-1)
For a system with a mass-proportional damping, the generalized damping for the i-th mode in
Equation (5-4-1) is obtained as,
ci = a0 mi , ci / mi = 2hi ω i (5-4-2)
Therefore,
a0 1
a0 = 2hiωi , hi = (5-4-3)
2 ωi
Similarly, for a system with a stiffness-proportional damping, the generalized damping for the
i-th mode is,
2
ci = a1ωi mi , ci / mi = 2hiωi (5-4-4)
Therefore,
2hi a1
a1 = , hi = ωi (5-4-5)
ωi 2
hi
C = a0 M C = a1 K
a0 1 a1
hi = hi = ωi
2 ωi 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 ] (5-4-6)
ω1
278
where, h: damping factor, ω1 : circular frequency of the first natural mode, [K 0 ] : the initial stiffness.
[C ] = 2h [K p ] (5-4-7)
ω
where, h: damping factor, ω1 : 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 ] (5-4-8)
1 1 / ω1 ω1 a0 h1
= (5-4-10)
2 1 / ω 2 ω 2 a1 h2
279
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 horizontal components of the base isolation floor. For example, in case of the stiffness-proportional
damping, the damping matrix is defined as:
2h
=[C ] K ( + K
ω upper BI ,V
) (5-4-12)
where,
[K upper ] : the stiffness matrix consisted with upper structures without base isolation elements,
K BI ,V : the stiffness matrix of base isolation elements for vertical components.
[C ] = [C pro ] + [C v ] (5-4-13)
[ ]
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.
280
5.5 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 δzi + Z0
= −[M ]
δzi
− [M ]
0
0
1
δ zi
Y0 = −[M ] − [M ][U ] Y0 (5-5-1)
− I iθxi
θ xi 0 0 0 θ xi Z
0 Z 0
− I iθyi θyi
0
0 0 θ yi
− I iθzi
θ zi 0 0 0 θ zi
For example, the components of the matrix [U ] of the structure in Figure 5-5-1 will be as follows:
281
Equilibrium condition of the structure under earthquake ground motion is:
X 0
[C ]{u} + [K ]{u} = −[M ]{u} − [M ][U ] Y0 (5-5-2)
Z
0
Damping force
X0
[M ]{u} + [C ]{u} + [K ]{u} = −[M ][U ] Y0 = {P} (5-5-3)
Z
0
282
5.6 External force by vibrator
A vibrator is assumed to be located at the center of gravity at a certain floor. The external forces from the
vibrator are denoted as Fx , Fy in X and Y directions.
Fx 1 0
Fy 0 1
0 0 0 Fx F
= = [V ] Fx (5-6-1)
0 0 0 Fy y
0 0 0
0 0 0
For example, the components of the matrix [V ] of the structure in Figure 5-6-1 will be as follows:
Fx Fy
0 0
Node 6 0 0
0 0
Node 7 0 0
0 0
6 7
0 0
Node 8 0 0 10
0 0 8 9
0 0
Node 9 0 0 1 2
1 0
5
0 1 3 4
Node 10 0 0
283
Equilibrium condition of the structure under vibrator force is:
F
−[ M ]{u} + [U ]
[C ]{u} + [ K ]{u} = F
x
(5-6-2)
y
Damping force
Inertia force
Restoring force
External force
F
[ M ]{u} + [C ]{u} + [ K ]{u}= [U ] Fx = {P} (5-6-3)
y
284
5.7 External force by wind
A wind force is assumed to be applied at the center of gravity at each floor with the constant distribution
along the height of the building. The external forces at i-th floor from the wind are denoted as
hi , x Fx ( t ) , hi , y Fy ( t ) , hr , y M z ( t ) in X, Y horizontal directions and Z rotational direction.
Fy ( t )
Fx ( t ) M z (t )
hi F ( t )
hx ,i Fx ( t ) hx ,i 0 0
hy ,i Fy ( t ) Fx ( t ) 0 hy , i 0
0 0 0 0
=
0
[W
= ] Fy ( t ) , [W ] 0 0 0
(5-7-1)
0 M z ( t ) 0 0 0
hr ,i M z ( t ) 0 0 hr ,i
285
For example, the components of the matrix [W ] of the structure in Figure 5-7-1 will be as follows:
Fx Fy Mz
0 0 0
Node 6 0
0 0
0 0 0
Node 7
0 0 0
0 0 0 7
6
0 0 0
Node 8 10
0 0 0
8 9
0 0 0
0 0 0
Node 9 1 2
0 0 0
hx ,1 0 0 5
0 3 4
hy ,1 0
Node 10
0 0 hr ,1
286
Equilibrium condition of the structure under wind force is:
Fx
[C ]{u} + [ K ]{u} =
−[ M ]{u} + [W ] Fy (5-7-2)
W
z
Damping force
Inertia force
Restoring force
External force
Fx
[ M ]{u} + [C ]{u} + [ K ]{u}= [W ] Fy = {P} (5-7-3)
W
z
287
5.8 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, the Operator Splitting method is selected.
where, [M ] , [C ] and [K ] are the mass, damping and stiffness matrices. {d } , {v} , {a} and { p}
are the displacement, velocity, acceleration, and external force vectors.
where, {∆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 } (5-8-3)
In case of a system with hysteresis nonlinearity, the equation of motion can be described as,
[ M ]{a} + [C ]{v} + { f ( d )} =
{ p} (5-8-4)
where f ( d ) is the force as a nonlinear function of the displacement {d } . The incremental form can be,
{∆f ( d )} =[ K ]{∆d }
i i i (5-8-6)
Finally, the equation of motion in incremental form is the same as Equation (5-8-2), that is
288
f
∆f
f k ( t ) ∆d
∆=
d
∆d
In the initial condition, the building will deform under the gravity load, i.e., the dead and live loads. It can
be analyzed by solving the following equation,
X0
− [ M ][U ] Y0
[ M ]{a} + [C ]{v} + [ K ]{d } = (5-8-26)
Z − g
0
where g is the gravity acceleration. When the gravitational acceleration is initially applied, the response
may fluctuate in the beginning. Therefore, it is better to apply the static gravity force { f0 } instead of
acceleration as,
X0 0
[ M ]{a} + [C ]{v} + [ K ]{d } =
− [ M ][U ] Y0 + { f 0 } , { f 0 } = − [ M ][U ] 0 (5-8-27)
Z g
0
and set the initial displacement as {d } = {d0 } , where {d0 } is the solution of
[ K ] {d 0 } = { f 0 }
The incremental form will be
∆X0
[ M ]{∆ai } + [C ]{∆vi } + [ Ki ]{∆di } =
− [ M ][U ] ∆Y0
∆Z
0
289
b) 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 } (5-8-2)
1 1
{∆d i } [M ] + [C ] + [K ]
β (∆t )
2
2β (∆t )
(5-8-9)
1
= {∆p i }+ [M ] {vi }+ 1 {a i } + [C ] 1 {vi }+ 1 − 1{a i }(∆t ) + {∆f }
β (∆t ) 2β 2β 4β
290
The equation can be rewritten as,
where,
1
{∆pˆ i } = {∆pi } + [M ] {vi } + 1 {ai } + [C ] 1 {vi } + 1 − 1{ai }(∆t ) + {∆f }
β (∆t ) 2β 2β 4β
(5-8-12)
291
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:
1
{d }= {di } + {vi }( ∆t ) +
− β {ai }( ∆t )
2
i +1 (5-8-15)
2
Equation (5-8-13) can be written as
{di+1} + β {ai+1}( ∆t )
{di +1} =
2
(5-8-16)
Therefore
{ai +1}=
1
β ( ∆t )
2 ({d i +1 } − {di +1})=
1
β ( ∆t )
2 {∆di +1} (5-8-17)
where
{ f=
( d )} [ K ]{d } − {∆f }
i +1 n +1 (5-8-21)
where
{∆f }
= [ K ]{dn+1} − { f ( dn+1 )} (5-8-22)
292
In this formulation, [K ] is the initial stiffness matrix.
Force
[ K ]{di +1}
∆f
{ f ( d )}
i +1 Corrector
[ K ]{di +1}
∆f
{ f ( d )}
i +1 Predictor
[K ] Displacement
{ }
di +1 {di +1}
{ f ( d=)}
i +1 ( ) { ( )}
K di +1 {∆di +1} + f di +1
(5-8-23)
( )
where K di +1 is a predictor stiffness.
The predictor stiffness is not necessary to be the initial stiffness and if the predictor stiffness is close to the
tangent stiffness, the corrector force is more accurate. It is known that if the predictor stiffness is larger than
the tangent stiffness, the OS method is unconditionally stable.
Force
Corrector
{ f ( d )}
i +1
( )
K di +1
{ f ( d )}
i +1 Predictor
Displacement
{ }
di +1 {di +1}
In STERA_3D, the predictor stiffness is created from the initial stiffness or tangent stiffness
if available.
293
Substituting the above equations into the equation of motion,
1 1 1
[ ]
M {∆d i +1 + [ C ]
} 2 β ( ∆t ) {∆di +1} + {vn } + {an }( ∆t )
β ( ∆t )
2
2
( ) { ( )} =
+ K dn +1 {∆di +1} + f di +1 {p } i +1
Kˆ {∆di +1} =
{ pˆ } (5-8-25)
where
1 1
( )
Kˆ = K dn +1 +
2 β ( ∆t ) [ ] β ( ∆t )2 [ ]
C + M (5-8-26)
Under seismic excitation and gravity load, the equation of motion will be,
X0,n +1
[ M ]{an+1} + [C ]{vn+1} + { f ( di +1 )} =
− [ M ][U ] Y0,n +1 (5-8-27)
Z
0,n +1 − g
0
[ K ]{d0 } = − [ M ][U ] 0 (5-8-29)
g
294
5.9 Energy
a) Equation of energy
X 0
[M ]{u} + [C ]{u} + [K ]{u} = −[M ][U ] Y0 = {P} (5-9-1)
Z
0
For example, in case of a structure with a rigid floor in Figure 5-9-1, the displacement vector, {u}, consists
of 15 components (see RED numbers in Figure 5-9-1.)
u1
u
{u} = 2 (5-9-2)
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 5-9-1 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
295
{u}T [M ]{u} + t {u}T [C ]{u}dt + {u}T [K ]{u} = − t {u}T {P}dt
2 ∫
0
2 ∫
0
(5-9-4)
WK + WD + WP = WI (5-9-5)
where,
WK =
{u}T [M ]{u} : Kinematic energy
2
t
WD = ∫ {u} [C ]{u}dt : Damping energy
T
{u} [K ]{u}
T
WP = : Potential energy
2
t
WI = − ∫ {u} {P}dt : Input energy
T
X 0
[M ]{u} + [C ]{u} + Q(u, u ) = −[M ][U ] Y0 = {P} (5-9-6)
Z
0
where, Q(u , u ) is the nonlinear restoring force vector. Then, the equation of energy can be derived as;
WK + WD + WP = WI (5-9-7)
where,
WK =
{u}T [M ]{u} : Kinematic energy
2
t
W D = ∫ {u} [C ]{u}dt
T
: Damping energy
0
t (5-9-8)
WP = ∫ {u} Q(u , u )dt : Potential energy
T
0
t
WI = − ∫ {u} {P}dt
T : Input energy
0
296
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} Q(u , u )dt = ∫ {u} ∑ qi (u , u )dt = ∑ ∫ {u} qi (u , u )dt =∑ WP ,i
T T
(5-9-10)
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 (5-9-11)
0
297
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. ASCE 6. 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 + − αi (6-1-3)
α 1 + 3T
i
where,
n n
α 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)
298
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
299
(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
0 , if T ≤ 0.7 sec
Ft = (6-1-8)
0.07 TQB , if T > 0.7 sec
Ft
Fi
h6
h2
h1
300
(5) ASCE distribution
The IBC (International Building Code) in the U.S. refers to the ASCE 7 “Seismic Design Requirements for
Building Structures” which gives the following formula for the calculation of lateral force distribution:
n
Fi = wi hik ∑w h
j =1
j
k
j (6-1-9)
1 , if T < 0.5sec
k= (T − 0.5 ) / 2 , if 0.5sec < T < 2.5sec (6-1-10)
2 , if T > 2.5sec
k=2 k=1
Fi
h6
h2
h1
301
(6) Mode distribution
Mode distribution is defined as:
n
Fi = QB wiφ1,i ∑w φ j
1, j (6-1-11)
j =1
where,
Fi
φ1,i
302
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
Demand Spectra
Sa (cm/sec 2)
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
∑ miδ i2 ∑m δ i i
2
S a = QB i =1
2 , Sd = i =1
n
(6-2-1)
n
∑ mi δ i ∑m δ 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
303
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.
304
7. Lumped Mass Model
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 (7-1-1)
∑ z i x i ∑ x i2 ∑ x y ∑ x a
i i i
∑ z i y i = ∑ y ∑ y b
2
i i (7-1-3)
∑ z i sym. n c
where,
n : the number of nodes in a floor
305
a-2) Equivalent plane for each floor from potential energy
The equivalent plane ( z = ax + by + c )is obtained from the vertical potential energy distribution by the
least square method:
= N ( ax + by + c )
Nz
N i zi x
∑ ( N z − N ( ax + by + c ) )
2
Minimize L
= i i i i i (7-1-4)
∑ N i zi xi ∑ N i xi2 ∑ N x y ∑ N x a
i i i i i
∑ N i zi yi = ∑ N y ∑ N y b
i
2
i i i (7-1-6)
∑ N i zi sym. ∑ N c
i
where,
n : the number of nodes in a floor
At this moment, STERA_3D adopts the formulation a-1), since it is easier to implement.
306
b) Decomposition of shear and flexural deformation
A story drift, D, can be divided into shear and flexural components as,
DS= D − DF (7-1-10)
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
307
7.2 Lumped mass model with shear and flexural stiffness
Under the external lateral forces, Fi (i = 1, 2,3) , the shear force and moment of each story are expressed
as below.
F3
θ3
EI 3
h3 Q3 = F3
F2 M 3 F=
= 3 h3 Q3 h3 , θ 2
EI 2
h2 Q=
2 F3 + F2
F1 M 2 = F2 h2 + F3 ( h2 + h3 ) = Q2 h2 + Q3 h3 , θ1
h1 EI1
Q1 = F3 + F2 + F1
In general
N
M i = ∑ Qj hj (7-1-11)
j =i
308
From the beam theory
MB
δ
B
τB θB = τ B + R
R
R
τA
θA =τ A + R M A 2 EI 2 1 τ A 2 EI 2 1 θ A − R
= =
M B h 1 2 τ B h 1 2 θ B − R
A
(7-1-13)
MA
M i +1 M i +1
hi EI i hi EI i
Mi −M i
M A = − M i , M B = M i +1 , τ A = θi −1 , τ B = θi −1 + ∆θi , EI = EI i , h = hi
− M i 2 EI i 2 1 θi −1 2 EI i 3θi −1 + ∆θi
= (7-1-14)
M i +1 hi 1 2 θi −1 + ∆θi hi 3θi −1 + 2∆θi
309
The increment of rotational deformation ∆θi is the difference of floor angle. Therefore,
∆θ1 =
θ1
(7-1-16)
∆θi = θi − θi −1 , i = 2, , n + 1
From the beam theory, the flexural deformation is
h3 h
δ= Q + (θ A + θ B ) (7-1-17)
12 EI 2
M + MB 6 EI δ
) Q =
− A − 2 {(θ A + θ B ) − 2 R} , R =
=
h h h
Therefore, the flexural deformation of i-th story is obtained as
h13 h
DF 1
= Q1 + 1 θ1
12 EI1 2
(7-1-18)
hi 3 h
DFi = Qi + i (θi −1 + θi ) , i = 2, , n + 1
12 EI i 2
The shear deformation is then calculated substituting the flexural deformation from the story drift as
D=
Si Di − DFi (7-1-19)
Under the nonlinear push over analysis, it is generally assumed that the flexural component is elastic and
only the shear component is considered as nonlinear.
310
Therefore, the lumped mass model is obtained from the following procedure:
From the next step, we use the same flexural stiffness obtained previously.
6) Calculate increment of angle
hi
∆θi
= ( M i +1 + M i ) (7-1-24)
2 EI i
7) Calculate flexural angle of each floor
i i
a=
i ∑ ∆θk =1
k or b=
i ∑ ∆θ
k =1
k (7-1-25)
311
b) Nonlinear flexural model
To consider nonlinear flexural component, the model to separate shear deformation and bending
deformation is used.
Reference) Akira Wada, et. Al. “Response Control Design of Buildings”, Maruzen (in Japanese), 1998
δs
uB B
δB
ks θB
θA
h φ
uA kb
δA A
The shear and moment force acting at the center springs are
M
= b bφ
k= kb (θ B − θ A ) for nonlinear bending spring (7-1-29)
nε n
N n k=
= kn (δ B − δ A ) for linear axial spring (7-1-30)
In a matrix form,
Qs k s 0 0 δ s δ s
=
=M b 0 kb 0 φb [ kW ] φb (7-1-31)
N 0 0 kn ε
n ε n n
312
F3
θ3
N 3 = W3
h3 Q3 = F3
F2 M 3 F=
= 3 h3 Q3 h3 , θ 2
N=
2 W3 + W2
h2 Q=
2 F3 + F2
F1 M 2 = F2 h2 + F3 ( h2 + h3 ) = Q2 h2 + Q3 h3 , θ1
N1 = W3 + W2 + W1
h1 Q1 = F3 + F2 + F1
By substituting
−M i , M B =
MA = M i +1 , QA ==
Qi , QB Qi +1 , N A ==
N i , N B N i +1 ,
=θ A θi −=
1, θB θ=
i, 1,
u A ui −= i, δA
uB u= δ=
i , δB δ=
i +1 , h hi
the lumped mass model is obtained from the following procedure from the push-over analysis.
and the vertical location at the center of gravity ( xci , yci ) as zci = axci + byci + c .
313
kni = N i ε ni (7-1-37)
6) Calculate moment at each floor
N
M i = η ∑ Qj hj (7-1-38)
j =i
Mi
To make the number of digits equivalent to a shear force, the nonlinear bending moment is divided by the
height of the story, on the other hand, the angle is multiplied by the height of the story to obtain the
deformation. Therefore, the force-deformation of the nonlinear bending spring is transformed as
Mb kb
Equivalent shear force Qb = Qb
= = δ b kbsδ b
h h2
Equivalent story drift δ b = hφ
kb
=
Equivalent stiffness Qb bsδ b ,
k= kbs
h2
After finding the tri-linear model for Qb − δ b relationship, it is returned to M b − φ relationship as,
δb
= b h,
M b Q= φ , kb = kbs h
2
314
c) Trilinear modeling of push-over curve
From the push over results up to the ultimate deformation (for example, up to 1/50 drift ratio), the
relationship between the story drift (shear δ s , bending δ b ) and the shear force (shear Qs , bending
Qb ) of each story is transformed into a tri-linear skeleton.
P3 (D3, Q3)
P2 (D2, Q2)
K3
K2
P1 (D1, Q1)
K1
315
< Case 2 >
When the last stiffness is large (for example, tangent stiffness > 0.1 K1 (initial stiffness))
P1(D1, Q1)
Find initial stiffness K1
Find Q1 that is the force when the tangent stiffness becomes 0.8K1 and determine D1 = Q1/K1
P2(D2, Q2)
The last point is P2.
K2 is the stiffness between P1 and P2
P3(D3, Q3)
D3 = 2×D2
K3 = K2
P3(D3, Q3)
P2(D2, Q2)
K2
K1
316
< Case 3 >
When the last stiffness is small (for example, tangent stiffness < 0.1 K1 (initial stiffness))
P1(D1, Q1)
Find initial stiffness K1
Find Q1 that is the force when the tangent stiffness becomes 0.8K1 and determine D1 = Q1/K1
P2(D2, Q2)
P2 is decided to be the same energy between push-over analysis and the model up to P3.
P3(D3, Q3)
P3 is the last point of push-over analysis
K3 is the tangent stiffness at P3
P3(Q3,D3)
P2(Q2,D2)
K3 = tangent K
K2
0.8K0
P1(Q1,D1)
K1
317
8. 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
318
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 , = φ1 = kc 2 + c3 (7-2-6)
dx
Consequently, the deflected shape is
(1 − C ) (ξ − S ) c1 v1 + φ1 L − v 2
= (7-2-10)
ξS
ξ (1 − C ) c 2 φ1 L − φ 2 L
where,
C = cos kL, S = sin kL, ξ = 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
d 2v
M ( x) = EI
dx 2
[
= EI − k 2 c1 cos kx − k 2 c 2 sin kx ] (7-2-15)
d 3v dv
V ( x) = − EI 3
+ F0 = − EIk 2 [φ1 − kc 2 ] (7-2-16)
dx dx
319
Calculating nodal loads, V (0) = −V1 , M (0) = − M 1 , V ( L) = V1 , M ( L) = M 1 , the stiffness matrix is
V1 ξ 2 S ξL(1 − C ) − ξ 2S ξL(1 − C ) v1
M
1 = EI ξ
2
− L (ξC − S ) − ξL(1 − C )
2
L2 (ξ − S ) φ1
(7-2-17)
V2 L3 ∆ ξ 2S − ξL(1 − C ) v 2
M 2 sym. − L2 (ξC − S ) φ 2
∆ = ξ (2 − 2C − ξS )
[ ( ) (
≈ ξ 2 − 2 1 − ξ 2 / 2 + ξ 4 / 24 − ξ 6 / 720 + C − ξ ξ − ξ 3 / 6 + ξ 5 / 120 − )] (7-2-18)
[
≈ ξ 1 − ξ / 15 + / 12
5 2
]
also
1 12
[
= 5 1 + ξ 2 / 15 +
∆ ξ
] (7-2-19)
Substituting ξ
2
= k 2 L2 = − F0 L / EI ,
EI F0 12
k11 = [12 ] + (7-2-21)
L3 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)
320
d) Implementation for beam element
θA z
τA τB
θB
uB
x
uA
A B y
M A 2 EI 2 1 τ A EI 4 L2 2 L2 τ A
M = L 1 2 τ = 3 2 (7-2-24)
B B L 2 L 4 L2 τ B
u
1 1 A
τ A L 1 − 0 θ
L A
τ = 1 1 (7-2-25)
B 0 − 1 u B
L L θB
1 1
QA 1 1 u
M
L L 1 −
A
0 θ
0 4 L
2 2
A = EI 1 2L L L A
QB L3 − 1 − 1 2 L2
4 L2 1 1 u
0 − 1 B
M B L L L L θB
0 1
6L 6L 1 1 u
A 12 6L − 12 6 L u A
4 L2 2 L2 1 − 0 θ EI 4 L2 − 6 L 2 L2 θ A
EI L L A =
= 3
L − 6 L − 6 L 1 0 − 1
1 u B L
3
12 − 6 L u B
2
2L 4 L2 L L θB
sym. 4 L2 θ B
321
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 θ A
+
Q B L3 12 − 6 L 30 L 36 − 3L u B
M B sym. 4 L2 sym. 4 L2 θ B
322
e) Implementation for column element
B u xB
u yB
λB l ' τ yB
τ xB θ yB
θ xB
l'
θ yA
τ yA
λ Al ' X
θ xA
A u xA
τ xA
u yA Y
M yA 2 EI 2 1 τ yA EI 4 L2 2 L2 τ yA
M = = in X-Z plane (7-2-27)
yB L 1 2 τ yB L3 2 L2 4 L2 τ yB
M xA 2 EI 2 1 τ xA EI 4 L2 2 L2 τ xA
M = = in Y-Z plane (7-2-28)
xB L 1 2 τ xB L3 2 L2 4 L2 τ xB
1 1 u
xA
τ yA − 1 0 θ
L L yA
τ = 1 1 u in X-Z plane (7-2-29)
yB − 0 1 xB
L L θ yB
u
1 1 yA
τ xA L 1 − 0 θ
L
xA
τ = 1 1
in Y-Z plane (7-2-30)
xB 0 − 1 u yB
L L θ 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,
323
Q xA 1 1
u
M − L −
L 1 1 xA
2 L2 − 1 0 θ
yA = EI 1 0 4 L
2
L L yA
Q xB L3 1 1 2 L2
4 L2 − 1 1 u
0 1 xB
M yB L L L L θ yB
0 1
− 6 L − 6 L 1 1 u 12 − 6 L − 12 − 6 L u xA
xA
2 2 L2 −
EI 4 L 1 0 θ EI 4 L2 6 L 2 L2 θ yA
= 3 L L yA =
u
L 6L 6 L − 1 0
1 3
1 xB L
12 6 L u xB
2
2L 4 L2 L L θ yB
sym. 4 L2 θ 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 u − 36 36 3L 3L 0 0 0 0 0 0 0 0 u xB
xB xB
M yA θ yA − 3L 3L 4 L 2
−L 2
0 0 0 0 0 0 0 0 θ yA
θ yB 0 θ yB
2 2
M yB − 3L 3L − L 4L 0 0 0 0 0 0 0
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
θ xA 30 L 0 0 0 0 3L − 3L 4 L2 − L2 0 0 0 0 θ xA
M xB θ xB 0 0 0 0 3L − 3L − L2 4 L2 0 0 0 0 θ xB
N zA δ zA
0 0 0 0 0 0 0 0 0 0 0 0 δ zA
N δ 0 0 0 0 0 0 0 0 0 0 0 0 δ zB
zB zB
M zA θ zA 0 0 0 0 0 0 0 0 0 0 0 0 θ zA
M
zB
θ
zB 0 0 0 0 0 0 0 0 0 0 0 0 θ zB
324
u xA
u
xB
θ yA
θ yB
u yA
u yB
= [[K ] + [K G ]] (7-2-34)
θ xA
θ xB
δ zA
δ
zB
θ zA
θ
zB
where,
36 − 36 − 3L − 3L 0 0 0 0 0 0 00
− 36 36 3L 3L 0 0 0 0 0 0 0
0
2 2
− 3L 3L 4 L −L 0 0 0 0 0 0 00
2 2
− 3L 3L − L 4L 0 0 0 0 0 0 00
0 0 0 0 36 − 36 3L 3L 0 0 00
0 0 0 0 − 36 36 − 3L − 3L 0 0 00
[K G ] = F0 (7-2-35)
30 L 0 0 0 0 3L − 3L 4 L2 − L2 0 0 00
0 0 0 0 3L − 3L − L2 4 L2 0 0 00
0 0 0 0 0 0 0 0 0 0 00
0 0 0 0 0 0 0 0 0 0 00
0 0 0 0 0 0 0 0 0 0 00
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,
325
9. Unbalance force correction
∆f = f i +1 − f i′+1 (7-3-1)
where, f i′+1 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 i′+1
i
fi ∆d
k
d
di d i +1
326
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
327
10. 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)
1 N −1
Ck =
N
∑x
m =0
m e −i ( 2πkm / N ) k = 0, 1, 2,, N − 1 (7-4-1)
N∆t N −1
y m = ∫ x m dt =
t = m∆t
0
∑
2π k = 0
S k e i ( 2πkm / N ) m = 0, 1, 2, , N − 1 (7-4-3)
2πv 0 N / 2 −1
Im(C k ) π ( N − 1)C 0
S0 = −2 ∑ +
N∆t k =0 k N
πC 0
Sk = [− 1 + i cos(πk N )] − i C k *
, S N −k = S N −k k = 1, 2, , N / 2 − 1 (7-4-4)
N k
πC 0
SN /2 = −
N
328
The following band pass filter (Butterworth filter) in frequency domain is applied to the
coefficient S k .
GB ( f ) = GL ( f )GH ( f ) (7-4-5)
( f fL )
2n
GL ( f ) = (7-4-6)
1+ ( f fL )
2n
1
GH ( f ) = (7-4-7)
1+ ( f fH )
2n
1.2 1.2
N=3
1 1
N=5
0.8 0.8
N = 10
0.6 0.6
N=3
0.2 N = 10 0.2
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
GL ( f ) GH ( f )
329
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)
330
11. Damage Index
STERA_3D adopts the following damage index, so called Park and Ang damage index, to evaluate the
structural damage under earthquake.
δ Eh µm E
D =+
m
β =+ β h (11-1)
δu Q yδ u µu Q yδ u
where
Qy
δy δM δ u = µu δ y
Eh = ∫ dE
331
The cumulative ductility factor is the ratio of the cumulative dissipated energy defined as
Eh
η= (11-2)
Q yδ y
The damage index can be rewritten as
µm η
D
= +β (11-3)
µu µu
Figure 11-2 Ultimate ductility factor and failure mode (Park and Ang (1985))
In case of the flexural failure, the value is greater than 10. Therefore, in STERA_3D,
µu = 15
is adopted for the nonlinear flexural springs at both ends of the reinforced concrete beams and columns.
332
Parameter β
The parameter β represents the effect of cyclic loading on damage. According to Park and Ang (1985), β is
calculated as,
l ρ
β=
−0.447 + 0.073 + 0.24n0 + 0.314 pt × 0.7
w
(11-4)
d
where
333
3) RC Wall
There is not much study about the damage index of RC shear walls. Therefore, STERA_3D adopts
arbitrary values for µu and β .
STERA_3D adopts
µu = 15
β = 0.05
for the nonlinear flexural springs at both ends of the reinforced concrete wall.
Also
µu = 8
β = 0.1
is adopted for the shear spring of the reinforced concrete wall.
The damage index for a part of a structure, such as individual story and for the entire structure, can be
evaluated as the weighting average of damage indices of structural elements in the part.
n
D part = ∑ wi Di (11-6)
i
where
The weighting factor wi can be based on the dissipated hysteretic energy of each element as,
Eh ,i
wi = n
(11-7)
∑E
i
h ,i
334
11.2 Damage Index of Steel Members
Reference:
- Study on Seismic Performance for Super-High-Rise Steel Buildings against Long-Period Earthquake
Ground Motions, Building Research Institute, Building Research Data, No. 160, 2014,7 (in Japanese)
where
For the low cycle fatigue with the cyclic plastic deformation, the relationship between the strain amplitude
∆ε i and the number of cycles to fracture N i is expressed by the Mason-Coffin equation as,
∆ε i ( % ) =
C ⋅ Ni − β (11-9)
or
1 1
−
∆ε i β C β
=N i = (11-10)
C ∆ε i
1 1
−
µi C β β
=N i = (11-11)
C µi
According to Figure 11-4 in the report “Study on Seismic Performance for Super-High-Rise Steel Buildings
against Long-Period Earthquake Ground Motions” (BRI, 2014),
C = 4 10
β =1 3
335
Figure 11-4 Fatigue curve for different connection types of steel beams (BRI, 2014)
In this method, the number of cycles ni accumulated at strain level ∆ε i (or ductility factor µi ) must be
336
Appendix) Rain-flow method
Reference:
- RAINFLOW CYCLE COUNTING IN FATIGUE ANALYSIS, Tom Irvine, 2018
The Rain-flow algorithm is the method for counting fatigue cycles from a time history.
337
b) Damage index based on the maximum response
Since it is an intensive work to record the time history of strain (or ductility factor) for all beams and
calculate the damage index using the rain-flow method, a practical method is proposed using the maximum
ductility factor and the cumulative ductility factor (BRI, 2014).
δy δ m = µmδ y
E0 ≈ 4 ( µm − 1) Qyδ y
The energy dissipation per cycle with the deformation of the maximum ductility µm is
E0 ≈ 4 ( µm − 1) Qyδ y (11-13)
Eh η Q yδ y η
N= = = (11-14)
E0 4 ( µm − 1) Qyδ y 4 ( µm − 1)
e
338
2) BRB (Buckling Restrained Brace)
Reference:
- Bucking-Restrained Braces and Applications, Edited by T. Takeuchi and A. Wada, JSSI, 2017
The Mason-Coffin equation for the relationship between the strain amplitude ∆ε i and the number of
∆ε i ( % ) =
C ⋅ Ni − β (11-18)
For the BRB (buckling restrained brace) damper, Takeuchi et al. (2008), proposed the following formulas,
∆ε i ( % ) =
0.5 ⋅ N i −0.14 ( ∆ε ( % ) < 0.1% )
i
Figure 11-3 Relationship between strain and number of cycles to facture (Takeuchi et al. (1985))
339
By combining with Coffin-Manson equations,
340