STERA3D Technical Manual

STERA 3D
STructural Earthquake Response Analysis 3D
Technical Manual
Version 5.5
Dr. Taiki SAITO
TOYOHOSHI UNIVERSITY OF TECHNOLOGY (TUT), JAPAN
1
UPDATE HISTORY
2007/07/17 STERA_3D Technical Manual Ver.1.0 is uploaded.

Explanation of passive damper element is added.
Explanation of freedom vector is added.
Mistakes in transformation matrix are modified.
Base isolation element is modified.
Multi-spring model for Column is modified.
Masonry element is installed.
Viscous damper element is installed.
Chapter 7 and Chapter 8 are added.
Chapter 9 and Chapter 10 are added.
The definition of shear deformation for passive damper is added.
Explanation of floor element is added.
Connection panel is installed.
The freedom of walls in case they are connected in series is explained in Chapter 4.
The error for calculating yield rotation of nonlinear spring is fixed..
New definition of mass distribution is added in Chapter 5.
Element model for wall panel is added.
The equation of shear strength of wall element is modified.
Chapter for the release of unbalance force is added.
2
Missing part (2.10 Connection Panel) is supplemented. Appendix for Degrading
Tri-linear Slip Model and Modified Bi-linear Model are added.
Definition of effective slab width is added. Also, the Appendix for Degrading Tri-linear
Slip Model is modified. Operator splitting method is added for numerical integration
method.
Modified Bilinear Model for HDRB (High Damping Rubber Bearing) is added for
isolation devices.
Steel elements are added.
Column element with direct input is added.
K-brace (Chevron brace) is added.
SRC members are added for beam, column and wall.
For some isolation devices, strength reduction by dissipated energy is considered.
Hysteresis of Bouc-Wen Model is added for isolator and damper
Definition of External Spring is extended in three directions.
5.5 Eigen value problem, 7.4 Calculation of ground displacement are added.
Formulation of initial stiffness of nonlinear spring is fixed (Eqs. (3-1-34), (3-1-51))
“5.5 Modal analysis” is modified including participation factor, effective mass, etc.
3
INDEX
1. Basic Condition
1.1 Coordinate
2. Constitutive Equation of Elements
2.1 Beam
2.2 Column
2.3 Wall
2.4 Brace
2.5 External Spring
2.6 Base Isolation
2.7 Masonry Wall
2.8 Passive Damper
2.9 Floor Element
2.10 Connection Panel
3. Hysteresis Models of Nonlinear Springs
3.1 Beam
3.1.1 RC Beam
3.1.2 Steel Beam
3.1.3 SRC Beam
Appendix: Detail rule of degrading trilinear slip model
3.2 Column
3.2.1 RC Column
3.2.2 Steel Column
3.2.3 SRC Column
3.3 Wall
3.3.1 RC Wall
3.3.2 Steel Wall (Brace)
3.3.3 SRC Wall (Brace)
3.4 External Spring
3.5 Base Isolation
Appendix:
A-1. Hysteresis of LRB (Lead Rubber Bearing)
A-2. Hysteresis of HDRB (High Damping Rubber Bearing)
A-3. Hysteresis of Lead Damper
A-4. Hysteresis of Elastic Sliding Bearing
A-5. Hysteresis of Bouc-Wen Model
3.6 Masonry Wall
3.7 Passive Damper
4
4. Freedom Vector
4.1 Node freedom
4.2 Freedom vector
4.3 Dependent freedom
4.4 Formulation of independent displacement of the element
5. Equation of Motion
5.1 Mass matrix
5.2 Stiffness matrix
5.3 Input ground acceleration
5.4 Equation of motion
5.5 Modal analysis
5.6 Damping matrix
5.7 Numerical integration method
5.8 Energy
6. Nonlinear Static Push-Over Analysis
6. 1 Lateral distribution of earthquake force
6. 2 Capacity Curve
7. Others
7.1 Decomposition of shear and flexural deformation
7.2 P-D effect
7.3 Unbalance force correction
7.4 Calculation of ground displacement
5
1. Basic Condition
1.1 Coordinate
(1) Global 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
Figure 1-1-1 Global coordinate
Z 3 Z Z
6
X 1 7 X X
4
5 8
Y Y Y
2
(a) lateral and rotational displacement (b) shear displacement
6
(2) Local Coordinate
The local coordinate is defined for each element. The displacement freedoms and force freedoms are named
with subscripts indicating the coordinate direction and node name. For example, the local coordinate of a
beam element in Figure 1-2 is defined to have its x-axis in the same direction of the element axis. Also the
displacement and force freedoms of a beam element are expressed as shown in Figure 1-1-2.
z
u zA u zB
 yA  yB x
A B
Displacement freedoms
y
Q zA Q zB
Local coordinate
M yA M yB
A B
Force freedoms
Figure 1-1-2 Local coordinate of a beam element
7
2. Constitutive Equation of Elements
2.1 Beam
 ' yA
x
M ' yA M ' yB
A B
Nx
 ' yB
elastic element
 ' yA  yA
M ' yA M ' yB M ' yA M ' yB
A B A B
N 'x  ' yB  yB
nonlinear bending springs
l'
 ' yA   ' yA   yA   yA   yA
        M ' yA M ' yB
 ' yB    ' yB    yB    yB  A  yB B
  '   '   0   0 
 x  x    
nonlinear shear springs
Figure 2-1-1 Element model for beam
Force-displacement relationship for elastic element

The relationship between the displacement vector and force vector of the elastic element in Figure 2-1-1 is
expressed as follows:
 l' l' 
 3EI  0 
6 EI y
 ' yA   y
  M ' yA 
   l' l'  
 ' yB    0  M ' yB  (2-1-1)

  '   6 EI y 3EI y
 N ' 
 x
 0 l'  x 
0
 EA 
where, E , I y , A and l ' are the modulus of elasticity, the moment of inertia of the cross-sectional area
along y-axis, the cross-sectional area and the length of the element. The rotational displacement vector of
the nonlinear bending springs is,
 yA   f yA 0   M ' yA 
    (2-1-3)
 yB   0 f yB  M ' yB 
where, f yA and f yB are the flexural stiffness of nonlinear bending springs at both ends of the element.
The rotational displacement vector from the shear deformation of the nonlinear shear spring is,
8
 1 1 
 yA   k sz l ' k sz l '   M ' yA 
  1 
1  M ' yB 
(2-1-2)
 yB  
 k sz l ' k sz l ' 
where, k sz is the shear stiffness of the nonlinear shear spring. Then, the displacement vector of the beam
element is obtained as the sum of the above three displacement vectors.
 ' yA   ' yA   yA   yA   M ' yA 

         
 ' yB    ' yB    yB    yB    f B M ' yB  (2-1-4)
  '   '   0   0   N' 
 x  x      x 
where,
 l' 1 l' 1 
 f yA  3EI  k l '  
6 EI y k sz l '
0 
 y sz

l' 1
 f B    f yB   0 

(2-1-5)
3EI y k sz l '
 l' 
 sym. 
 EA 
[ f B ] is the flexural stiffness matrix of the beam element. By taking the inverse matrix of [ f B ] , the
constitutive equation of the beam element is obtained as,
 M ' yA   ' yA   ' yA 

  1    
 M ' yB    f B   ' yB   k B  ' yB  (2-1-6)
 N'  '  ' 
 x   x  x
where, [k B ] is the stiffness matrix of the beam element.
Including rigid parts and node movement

Including rigid parts and node movement as shown in Figure 2-1-2, the rotational displacement vector is,
 ' yA   yA    u zB   B l ' yB   u zA   Al ' yA 

  ,  
 ' yB   yB    l'
u 
 1 1  1 1   zA 
 yA  l ' u zA   A yA  l ' u zB   B yB   l '  l ' 1   A B  u
 zB 
       (2-1-7)
1 1
 yB  u zA   A yA  u zB   B yB  
1

1
A 
1   B   yA 
 l' l'  l' l'   yB 
 
9
 yB
 ' yA  yA 
 yB
 yA   ' yB
u zB
Z
u zB   B l ' yB
u zA u zA   A l ' yA
A B X
 Al ' l' B l '
Figure 2-1-2 Including rigid parts and node movement
From node axial displacements, relative axial displacement is,

 ' x   xB   xA (2-1-8)
Therefore
 ' yA   ' yA 
 ' yA  1 0 0 0    

 '
 
 0 1 0 0   ' yB   n  ' yB 
 yB     B   (2-1-9)
  '  0 0  1 1   xA    xA 
 x       xB 
 xB 
Combining Equations (2-1-7) and (2-1-9),
u  u zA 
1 1   zA  u 
 ' yA    1  A B 0 0 u zB
l ' l '    zB 
 '   1 1     
 yB    A 1  B 0  yA    B  yA 
   l'  l' 0 (2-1-10)

 xA     yB  yB 
  xB   0 0 0 0 1 0    xA 
 xA
0 0
 0 0 0 1    
 xB   xB 
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,
 ' zA   zA    u yA   Al ' zA   u yB  B l ' zB 

  ,  
 ' zB   zB    l'
10
u 
 1 1   1 1   yA 
 zA  l ' u yA   A zA  l ' u yB  B zB   l ' l'
1  A B  u
 yB 
  1    (2-9-7)
 A 1  B   zA 
1 1 1
 zB  u yA   A zA  u yB  B zB  
 l' l'   l' l'   zB 
 Al ' l' B l'

Y
u yA  zA
 ' zA  u yB
 zB
  zB  ' zB
 zA
Figure 2-1-4 Beam displacement with rigid connection (X-Y plane)
u zA 
 ' yA   1 1  u zB 
1  A B  
 '   l ' l'
 yB   1 1   yA 
 ' zA   A 1  B   yB 
   l' l'  
 ' zB   1

1
1  A B  u yA 
   u 
  xA   l'
1
l'
1   yB 
  xB    A 1  B   zA 
   l' l'  
  xA   1   zB 
    
 xB   1  xA 
 xB 
From global node displacement to element node displacement

Transformation from global node displacements to element node displacements is,
u zA 
u 
 zB   u1 
 yA  u 
 2
   TixB   (2-1-11)
 yB  
 xA  un 
 
 xB 
The component of the transformation matrix, [TixB ] , is discussed in Chapter 4 (Freedom Vector).
11
From global node displacement to element face displacement
Transformation from the global node displacement to the element face displacement is,
u1   u1 
 ' yA  u  u 
   2  2
 '
 yB   n B  B T 
ixB    T 
xB   (2-1-12)
 '    
 x un  un 
In case of Y-direction beam
z Z
y X
x Y
Local coordinate of Y-beam Global coordinate
Figure 2-1-3 Relation between local coordinate and global coordinate
In case of Y-direction beam, the axial direction of the beam element coincides to the Y-axis in the global
coordinate, transformation of the sign of the vector components of the element coordinate is,
x  0 1 0  X 
   
 y   1 0 0  Y  (2-1-13)
z  
 Y  Beam  0 0 1  Z  Global
Therefore
u zA  1  u zA  u zA 
u    u 
 zB 
 1
 0  u zB   zB 
 yA   1   xA   xA 
      s B   (2-1-14)
 yB   1   xB   xB 
 xA   0 1   yA   yA 
      
 xB Y  Beam  1  yB  Global  yB  Global
12
Transformation from the global node displacement to the element node displacement is,
u zA 
u 
 zB   u1 
 xA  u 
   TiyB    2
  (2-1-15)
 xB  
 yA  u n 
 
 yB 
 u1   u1 
 ' yA  u  u 
 
 
 ' yB   n B  B s B  TiyB
 2
 
   T yB
 2
  (2-1-16)
 '   
 x u n  u n 
Constitutive equation
Finally, the constitutive equation of the X-beam is,
 P1   u1  u1 
P  u  u 
 2  2  2
   TxB T
k B T 
xB    K 
xB   (2-1-17)
     
 Pn  un  un 
For Y-beam,
 P1   u1   u1 
P  u  u 
 2
   TyB   T
   
k B  TyB  2   K yB  2
  (2-1-18)
  
 Pn  un  un 
Transformation matrix for nonlinear spring displacement

The nonlinear spring displacement vector is obtained from the element face displacement as,
   
 yA   f yA 0 0  M ' yA   f yA 0 0  ' yA   ' yA 
      
 yB    0 f yB 0 M ' yB    0
 
 
f yB 0k B  ' yB   T pB
 
 ' yB  (2-1-19)
   1 1    1 1   '   ' 
 y  0  N ' x   0  x  x
 k sz l ' k sz l '   k sz l ' k sz l ' 
where,
13
 
 f yA 0 0
 
T 
pB  0 f yB 0k B  (2-1-20)
 1 1 
 0
 k sz l ' k sz l ' 
14
2.2 Column
Element model for column is defined as a line element with nonlinear bending springs at both ends and two
nonlinear shear springs in the middle of the element in x and y directions as shown in Figure 2-2-1.
X-Z plane Y-Z plane
M ' xB M ' xB
M ' yB
M ' yB
B B B
 ' xB
 ' yB
N 'z Z
l' = +
T 'z
 ' yA  ' xA X
A A A
M ' yA Y
M ' yA
M ' xA M ' xA
Figure 2-2-1 Element model for column

In the same way as the beam element, the relationship between the displacement vector and force vector of
the elastic element is,
 l' l' 

 ' yA   3EI y 6 EI y  M ' yA 
  
l '  M ' yB 
in X-Z plane (2-2-1)
 ' yB   l '
 6 EI y 3EI y 

 l' l' 

 ' xA   3EI x 6 EI x   M ' xA 
 
 l '  M ' xB 
in Y-Z plane (2-2-2)
 ' xB   l '
 6 EI x 3EI x 
The axial displacement is,
l'
 ' 'z  N 'z (2-2-3)
EA
The torsion angle by torque force is,
l'
 '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.
15
Force-displacement relationship for nonlinear bending springs
Nonlinear interaction M x  M y  N z is considered in the nonlinear bending springs,
 yA  M ' yA 
N ' zB ,  zB
 
 
 xA   f pA
 
 M ' xA  at end A (2-2-5)
   N'  M ' xB ,  xB
 zA   zA 
M ' yB ,  yB
 yB  M ' yB 
 xB    f pB M ' xB  at end B
   
(2-2-6)
B
   N' 
 zB   zB 
where, [ f pA ] and [ f pB ] are the flexural

stiffness matrices of the nonlinear bending springs.
Therefore, the force-displacement relationship of A
nonlinear bending springs is,
M ' yA ,  yA
 yA   M ' yA 
  M '  M ' xA ,  xA
 xA   xA 
 
 zA   f pA 0   N ' zA 
N ' zA ,  zA
 
 yB   0   
f pB  M ' yB 
 (2-2-7)
 xB   M ' xB  Figure 2-2-2 Nonlinear bending springs

   
 zB   N ' zB 
Rearrange the order of the components of the displacement vector and change the node axial displacements
into the relative axial displacement,
 yA   yA 
 yA  1 0 00 0 0    
  0    xA   xA 
 yB   0 01 0 0
 zA   zA 
 xA   0 1 0 0 0 0    n p     (2-2-8)
  0   yB
0 0 0 1 0    yB 
 xB    xB   xB 
  z  0 0  1 0 0 1    
 zB   zB 
The force-displacement relationship in Equation (2-2-7) is then expressed as,
 yA   M ' yA   M ' yA 
  M '  M ' 
     
 yB   f pA 0   yB   yB 
 xA   n p 
T  
 
n  M '   f p  M ' xA 
f pB  p  xA 
(2-2-9)
   0 M' M ' 
  xB
 xB   xB 
  z   N ' z   N ' z 
16
Force-displacement relationship for nonlinear shear springs
 1 1 
 yA   k sx l ' k sx l '   M ' yA 
  1 
1  M ' yB 
in X-Z plane (2-2-10)
 yB  
 k sx l ' k sx l ' 
 1 1 
 xA   k sy l ' k sy l '   M ' xA 
  1 
1  M ' xB 
 xB  
 k sy l ' k sy l ' 

where, k sx and k sy are the shear stiffness of the nonlinear shear springs.
The displacement vector of the column element is obtained as the sum of the displacement vectors of
elastic element, nonlinear shear springs and nonlinear bending springs,
 ' yA   ' yA   yA   yA   M ' yA 

 '   '      M ' 
 yB   yB   yB   yB   yB 
 'xA   ' xA   xA   xA   M ' 
         f C  xA  (2-2-12)
 'xB   ' xB   xB   xB   M ' xB 
  ' z   ' ' z  z   0   N 'z 
         
  'z    'z elastic element  0 bending spring  0  shear spring  T ' z 
The flexural matrix [ f C ] is;
 l' l' 
 3EI  
6 EI y
 y

 l' 
0
 3EI y 
 
 l' l' 

 f C    3EI x 6 EI x  
l' 
 
 3EI x 
 l' 
 EA 
 l' 
 sym. 
 GI z  elestic element
17
 f p11 f p12 f p13 f p14 f p15 0
 f p 22 f p 23 f p 24 f p 25 0

 f p 33 f p 34 f p 35 0
  
 f p 44 f p 45 0
 f p 55 0
 
 sym. 0 bending spring
 1 1 
 k l' k l' 
 sx sx

 1
0 
 k sx l ' 
 1 1 
  (2-2-13)
 k sy l ' k sy l ' 
 1 
 
 k sy l ' 
 0 
 
 sym. 0 shear spring
By taking the inverse matrix of [ f C ] , the constitutive equation of the column element is obtained as,
 M ' yA   ' yA   ' yA 

M '   '   ' 
 yB   yB   yB 
 M ' xA  1  ' xA 
   ' xA 
    fC     k C   (2-2-14)
 M ' xB   ' xB   ' xB 
 N 'z    'z    'z 
     
 T ' z    ' z    ' z 

Change relative axial displacement and torsion displacement into node displacement,
 ' yA   ' yA 
   ' 
 ' yA  1   ' yB   yB 
 '   1 0   '   ' xA 
 yB     xA   
 ' xA   1   ' xB   ' xB 
     nC   (2-2-15)
 ' xB   1   zA    zA 
  'z   0 1 1    zB    zB 
      
  'z    1 1   zA    zA 
  zB    zB 
Including rigid parts and node movement,
18
u xA  u xA 
 1 1 
 u xB  u 
 l ' 1  A B     xB 
l'    yA 
 ' yA     yA
 '   1 1
A 1  B 
0   yB   
 yB   l ' l'     yB 
 ' xA   1

1
1  A B  u yA  u yA 
   
 u   
 ' xB   l' l'
     C  yB 
 yB  u
  1 1
  zA    A 1  B   xA   xA 
l' l'   
  zB   1  xB 
     xB   
 zA  
 1   zA   zA 
      
 zB   0 1   zB   zB 
  
1  zA   zA 

   
 zB   zB 
(2-2-16)
B u xB
u yB
Bl'
 yB
 xB  ' yB
l'
 'xB
 ' yA Z
 'xA
 yA
 Al ' X
 xA
A u xA
u yA
Y
Figure 2-2-3 Including rigid parts and node movement
19
Transformation from global node displacement to element node displacement is;
u xA 
u 
 xB 
 yA 
 
 yB 
u yA   u1 
  u 
u yB   2
   T 
iC   (2-2-17)
 xA  
 xB  un 
 
 zA 
 zB 
 
 zA 
 
 zB 
The component of the transformation matrix, [TiC ] , is discussed in Chapter 4 (Freedom Vector).

 ' yA 
 ' 
 yB   u1   u1 
 ' xA    u 
u 2   2
   n  T 
iC    T 
C   (2-2-18)
 '
 xB 
C C
  
  'z  u n  u n 
 
  ' z 
Finally, the constitutive equation of the column is;
 P1   u1 
P  u 
 2  2
   K C   (2-2-19)
 
 Pn  un 
where,
K C   TC T kC TC  (2-2-20)
20
The nonlinear spring displacement vector is obtained from Equations (2-2-7), (2-2-10) and (2-2-11),
 
 yA  
  
f 
pA 0 

 xA     M ' yA   M ' yA 
 zA   
 M ' xA  M ' 
       xA 
 yB  
 
0 f    N ' zA    f  N ' zA 
pC  
  M ' yB 
pA (2-2-21)
xB       M ' yB 
 zB   1 1 M '   M ' xB 
   0 0 0 0  xB   
 N'
  y   k sx l ' k sx l '
  zB   N ' zB 
  x   0 1 1
0 0 0
 k sy l ' k sy l ' 
Furthermore, in the same way as Equation (2-2-8),
 M ' yA   M ' yA   M ' yA 

M '  M '  M ' 
 xA   yB   yB 
 N ' zA 
  
  np
T

M ' 
0  xA   n' p    M ' xA 
  (2-2-22)
 M ' yB   M ' xB   M ' xB 
 M ' xB   N 'z   N 'z 
     
 N ' zB   T ' z   T ' z 
Therefore, the nonlinear spring displacement vector is obtained from the element face displacement as,
 yA 
 
 xA   M ' yA   ' yA   ' yA 
 zA  M '   '   ' 
   yB   yB   yB 
 yB     ' xA   ' xA 
  
 ' xA 
   f pC n' p 
M
  
  f pC n' p k C    T pC     (2-2-23)
 xB   M ' xB   ' xB   ' xB 
 zB   N 'z    'z    'z 
       
 y   T ' z    ' z    ' z 
 
 x
21
2.3 Wall
Element model for wall is defined as a line element with nonlinear bending springs at both ends and three
nonlinear shear springs; one is in the middle of the wall panel and others are in the side columns as shown
in Figure 2-3-1.
M ' xB1 M ' yBc M ' xB 2
 ' xB  ' xB Z
 ' yB
l' X
 ' xA  ' yA  ' xA
Y
M ' xA1 M ' yAc M ' xA2
Figure 2-3-1 Element model for wall

In the same way as the beam element, the relationship between the displacement vector and force vector of
 l' l' 

 ' yAc   3EI c 6 EI c   M ' yAc 
  
l '   M ' yBc 
in wall panel (2-3-1)
 ' yBc   l '
 6 EI c 3EI c 
 l' l' 

 ' xA1   3EI 1 6 EI 1   M ' xA1 
  
l '   M ' xB1 
in side column 1 (2-3-2)
 ' xB1   l '
 6 EI 1 3EI 1 
 l' l' 

 ' xA2   3EI 2 6 EI 2   M ' xA2 
  
l '   M ' xB 2 
 ' xB 2   l '
 6 EI 2 3EI 2 
The axial displacement is,
l'
 ' ' zc  N ' zc (2-3-4)
EA
22
Force-displacement relationship for nonlinear bending springs
Nonlinear interaction M x  M y  N z is considered in the nonlinear bending springs,
N ' zBc ,  zBc

M ' yBc ,  yBc
M ' xB1 , xB1 M ' xB 2 ,  xB 2
B
A
M ' xA1 ,  xA1 M ' xA2 ,  xA2
M ' yAc ,  yAc
N zAc ,  zAc
Figure 2-3-2 Nonlinear bending springs
 yAc   M ' yAc 

  M ' 
 xA1 
  
  f pA
 xA1 
  at end A (2-3-5)
xA2  M ' xA2 
 zAc   N ' zAc 
 yBc   M ' yBc 

  M ' 
 xB1 


 
  f pA
 xB1 

M '
 at end B (2-3-6)
 xB 2   xB 2 
 zBc   N ' zBc 
where, [ f pA ] and [ f pB ] are the flexural stiffness matrices of the nonlinear bending springs. Therefore,
the force-displacement relationship of nonlinear bending springs is,
23
 yAc   M ' yAc 
  M ' 
 xA1   xA1 
 xA2   M ' xA2 
 
 
 zAc   f pA

0   N ' zAc 



 yBc   0    
f pB   M ' yBc 
(2-3-7)
 xB1   M ' xB1 

   
 xB 2  M ' xB 2 
   N' 
 zBc   zBc 
Rearrange the order of the components of the displacement vector and change the node axial displacements
into the relative axial displacement,
 yAc   yAc 
 yAc  1    
     xA1   xA1 
 yBc   1    xA2 
xA1   1   xA2   
  zAc   zAc 
  
xB1    1     n p     (2-3-8)
   1   yBc   yBc 
 xA2   
 xB1  xB1 
xB 2   1    
    xB 2  xB 2 
  zc   1 1    
 zBc   zBc 
The force-displacement relationship in Equation (2-3-7) is then expressed as,
 yAc   M ' yAc   M ' yAc 

  M '  M ' 
 yBc   yBc   yBc 
 xA1   M ' xA1   M ' xA1 
 
     
 f pA
 xB1   n p 
0 

T  
   
n p  M ' xB1   f p  M ' xB1  (2-3-9)
   0 f pB  M '  M ' 
 xA 2
  xA2   xA2 
 xB 2  M xB 2 
' M ' xB 2 
   N'   N' 
  zc   zc   zc 
Force-displacement relationship for nonlinear shear springs

 1 1 
 yAc   k scl ' kscl '  M ' yAc 
  
1  M ' yBc 
in wall panel (2-3-10)
 yBc   1
 k scl ' kscl ' 
 1 1 
 xA1   k s1l ' k s1l '   M 'xA1 
  
1  M 'xB1 
 xB1   1
 k s1l ' k s1l ' 
24
 1 1 
 xA2   k s 2l ' k s 2l '   M ' xA2 
  
1  M ' xB 2 
 xB 2   1
 k s 2l ' k s 2l ' 
where, k sc , k s1 and k s 2 are the shear stiffness of the nonlinear shear springs.
 ' yAc   ' yAc   yAc   yAc   M ' yAc 

 '   '      M ' 
 yBc   yBc   yBc   yBc   yBc 
 ' xA1   ' xA1   xA1   xA1   M ' xA1 
         
 ' xB1    ' xB1    xB1    xB1    fW  M ' xB1  (2-3-13)
 '   '      M ' 
 xA2   xA2   xA2   xA2   xA2 
 ' xB 2   ' xB 2   xB 2   xB 2  M ' xB 2 
         
  ' zc   ' ' zc elastic element   zc bending spring  0 shear spring  N ' zc 
The flexural matrix [ f W ] is;
 l' l' 
 3EI  
6 EI c
 c

 l' 
 3EI c 
 l' l' 
  
 3EI 1 6 EI 1 
 l' 
 fW     
 3EI 1 
 l' l' 

 3EI 2 6 EI 2 
 l' 
 sym. 
 3EI 2 
 l' 
 
 EAc  elestic element
 f p11  f p17 
 
    
 f p 71  f p 77 
 bending spring
25
 1 1 
 k l' k scl ' 
 sc 
 1 
 k scl ' 
 1 1 
 
 k s1l ' k s1l ' 
 1  (2-3-14)
 k s1l ' 
 
 1 1 
 k s 2l ' k s 2l ' 
 1 
 sym. 
 k s 2l ' 
 0 shear spring
By taking the inverse matrix of [ fW ] , the constitutive equation of the column element is obtained as,
 M ' yAc   ' yAc   ' yAc 

M '   '   ' 
 yBc   yBc   yBc 
 M ' xA1   ' xA1   ' xA1 
  1    
 M ' xB1    f W   ' xB1   kW  ' xB1  (2-3-15)
M '   '   ' 
 xA 2   xA2   xA2 
 xB 2 
M ' 
 xB 2 
'  ' xB 2 
     
 N ' zc    ' zc    ' zc 

Change relative axial displacement and torsion displacement into node displacement,
 ' yAc   ' yAc 

 ' yAc  1    ' 
 '   1   ' yBc   yBc 
 yBc     '   ' xA1 
 ' xA1   1   xA1   
     ' xB1   ' xB1 
 ' xB1    1   '   nW  '  (2-3-16)
 '   1   xA 2   xA 2 
 xA2   
  ' xB 2   ' xB 2 
 ' xB 2   1     
    '   ' 
  ' zc    1 1  zAc
 
zAc

 ' zBc   ' zBc 
Including rigid parts and node movement,
26
u xAc  u xAc 
u  u 
 1 1   xBc   xBc 
 l ' 1 A B   yAc   yAc 
l'
    
 ' yAc   1 1
A 1 B   yBc   yBc 
 '   l ' l'   u yA1   u yA1 
 yBc   1 1    
 ' xA1    1 A B   u yB1   u yB1 
  l' l'
 ' xB1   1 1     
   A 1 B   xA1    W  xA1 
 ' xA 2   l' l'
1 1
  xB1   xB1 
 ' xB 2    1 A B  u  u yA2 
   l' l'   yA2   
 ' zAc   1 1  u yB 2  u yB 2 
 '    A 1 B    
 zBc   l' l'   xA2  xA2 
 1   
 xB 2   xB 2 

 1    
 zAc   zAc 
 zBc   zBc 
(2-3-17)
Transformation from the center displacements to the node displacements is,
 yc
 z 2   z1
 yc 
 z1  zc  z2 w
 z1   z 2
 zc 
2
w
Figure 2-3-3 Relationship between center and node displacements
1 
u xAc     u xA1   u xA1 
u   1    

 xBc   1 1   zA1   zA1 
 yAc    w w 
 zA2   zA 2 
       
 yBc  
1 1 u
   xB1   u xB1 
 u yA1   w w   zB1   zB1 
   1     
u yB1    zB 2    zB 2 
1 
     u yA1   u 
  DW 
xA1 yA1
  1  

 xB1   u
  yB1  u
 yB1 
1
u yA2     xA1   xA1 
   1    
u yB 2     xB1    xB1 
1
    u  u 
 xA2   1   yA2   yA2 

 xB 2    u  u yB 2 
   1  yB 2   

 zAc   0.5 0.5   xA2   xA2 
 zBc   0.5 0.5
  xB 2   xB 2 
  (2-3-18)
27
Transformation from the global node displacements to the element node displacements is;
 u xA1 
 
 zA1 
 zA2 
 
 u xB1 
 zB1 
 
 zB 2   u1 
 u  u 
 2

yA1
  T 
ixW   (2-3-19)
u yB1  
 xA1  u n 
 
 xB1 
u 
 yA2 
u yB 2 
 
 xA2 
 xB 2 
The component of the transformation matrix, [TixW ] , is discussed in Chapter 4 (Freedom Vector).

 ' yAc 
 ' 
 yBc   u1   u1 
 ' xA1  u  u 
   2  2
 ' xB1   nW  W DW TixW    TxW   (2-3-20)
 '   
 xA 2
 u n  u n 
 ' xB 2 
 
  ' zc 
In case of Y-direction wall
z Z
y X
x Y
Local coordinate of Y-wall Global coordinate
28
In case of Y-direction wall, the wall panel direction coincides to the Y-axis in the global coordinate,
transformation of the sign of the vector components of the element coordinate is,
x  0 1 0  X 
   
 y   1 0 0  Y  (2-3-21)
z  
 Y Wall  0 0 1  Z  Global
Therefore
 u xA1  1   u yA1   u yA1 
   1     
 zA1     zA1   zA1 
 zA2   1   zA2   zA2 
      
 u xB1   1  u yB1  u yB1 
 zB1   1   zB1   zB1 
      
 zB 2   1   zB 2   zB 2 
 u   1   u xA1  u 
 xA1 
    W 
yA1
   
u yB1   1   xB1 
u  u xB1 
 xA1   1     yA1 
     yA1   
 xB1   1   yB1   yB1 
u    u  u xA2 
 1   xA2 
 yA 2
  
u yB 2   1  u xB 2  u xB 2 
     
 1   yA2 
 xA2     yA2 

 xB 2  Y Wall  
1  yB 2  Global  yB 2  Global
(2-3-22)
Transformation from the global node displacement to the element node displacement is;
 u yA1 
 
 zA1 
 zA2 
 
u yB1 
 zB1 
 
 zB 2   u1 
u  u 
 xA1 
  
  TiyW
 2
  (2-3-23)
 u xB1  
 yA1  u n 
 
 yB1 
u xA2 
 
u xB 2 
 
 yA2 
 yB 2 
29
 ' yAc 
 ' 
 yBc   u1   u1 
 ' xA1  u  u 

 '
 xB1 

 nW  W DW  W T  2
ixW 

 
  T yW
 2
  (2-3-24)
 '    
 xA2  u n  u n 
 ' xB 2 
 
  ' zc 
Finally, the constitutive equation of the wall is;
 P1   u1 
P  u 
 2  2
   K xW   (2-3-25)
 
 Pn  u n 
where,
K xW   TxW T kW TxW  (2-3-26)
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)

The nonlinear spring displacement vector is obtained from Equations (2-3-7), (2-3-10)~(2-3-12),
30
 
 yAc  
 f pA 0 

   
 xA1     M ' yAc   M ' yAc 
 xA2     M ' 
     M ' xA1   xA1 
 zAc     M ' xA2   M ' xA2 
 yBc   0 f pA    
     N ' zAc   f
 xB1      M ' yBc 
 
pW
 N ' zAc 
 
       M ' yBc 
 xB 2   1 1   M ' xB1   M ' xB1 
 zBc    M '   
   k sc l ' k sc l '   xB 2  M ' xB 2 
  yc   1 1   N ' zBc   N' 
  x1    zBc 
k s1l ' k s1l ' 
   
 x 2   1 1 
 k s 2l ' k s 2l ' 
(2-3-29)
Furthermore, in the same way as Equation (2-3-8),
 M ' yAc 
M '   M ' yAc 
 xA1  M ' 
 M ' xA2   yBc 
   M ' xA1 
 N ' zAc 

M '
 
  np
T  
 M ' xB1  (2-3-30)
 yBc  M ' 
 M ' xB1   xA2 
  M ' xB 2 
M ' xB 2   N' 
 N'   zc 
 zBc 
Therefore, the nonlinear spring displacement vector is obtained from the element face displacement as,
 yAc 
 
 xA1 
 xA2   M ' yAc   ' yAc   ' yAc 
  M '   '   ' 
 zAc   yBc   yBc   yBc 
 yBc   M ' xA1   ' xA1   ' xA1 
 
  
 xB1   f pW n p
T  
  
 M ' xB1   f pW n p
T
kW  ' xB1   T pW   
 ' xB1 

(2-3-31)
  M '   '   ' 
 xB 2   xA2   xA2   xA2 
 zBc  M ' xB 2   ' xB 2   ' xB 2 
       
 yc   N ' zc   ' zc   ' zc 
  x1 
 
 x 2 
31
2.4 Brace
Element model for Brace is defined as a truss element with a nonlinear axial spring and pin-supported at
both ends as shown in Figure 2-6-1.
3 B 4
N',  '
Z
h
1 2 X
A
w Y
Figure 2-4-1 Element model for brace
Force-displacement relationship
~ ~ ~ ~
f x4 , u x4 f x3 , u x3
4 3
N1 ,  1
x N2, 2
y x
~ ~ y ~ ~
f x1 , u x1 f x2 , u x2
1 2
(Brace 1) (Brace 2)
Figure 2-4-2 Local coordinate
The relationship between axial deformation and axial force of the truss element is,
N1  k1 1 (2-4-1)
N 2  k 2 2 (2-4-2)
Replacing with the nodal force and displacement in local coordinate along the element,
~ ~
N 1   f 1x  f 4 x ,  1  u~4 x  u~1x (2-4-3)
~ ~
N 2   f 2 x  f 3 x ,  1  u~3 x  u~2 x (2-4-4)
32
In a matrix form,
 u~1x 
u~ 
 1y 
~
u1x  u~2 x 
  ~ 
 1   1 0 0 1 u~2 x   1 0 0 0 0 0 1 0 u 2 y 
 u~    0 0  1 0 1 0 0 0 u~   nb u 
~
  (2-4-5)

 2  0  1 1 0  3x    3x 
u~4 x  u~3 y 
~ 
u 4 x 
u~4 y 
 
~
 f x1   1 0 
~  
 f z1   0 0 
~ ~
 f 1x   1 0  f x 2   0  1
~     ~   
 f 2 x   0  1  N 1 
~     f  ~    
~  f z2   0 0   N1 

1 N 2 

T N 
 nb   1  (2-4-6)
 f 3x   0 1 N 2   f x3   0 
N 2 
~
f   1   ~
f   0
 4x  0 0
 ~z 3  
 f x4   1 0
~ 
f   0 0 
 z4  
From Figure 2-4-3, the relation between the nodal forces in local coordinate and those of global coordinate
is,
~
f x1  f x1 cos   f z1 sin 
~ for Brace 1 (2-4-7)
f y1   f x1 sin   f z1 cos 
and
~
f x 2   f x 2 cos   f z 2 sin 
~ for Brace 2 (2-4-8)
f y 2   f x 2 sin   f z 2 cos 
Eq. (2-4-8) can be also obtained from the Eq. (2-4-7) by replacing  by     and using the
formulas sin      sin  , cos      cos  .
33
f z4 ~ ~ f z3
f x4 f x3
4 3
f x4 f x3
h
l  w2  h 2
f z1 ~ f z2
~ f 2x
f x1
1 2
f x2
f x1
f z1 f z2
f z1 cos  f z1 sin 
 f x1 cos  f z 2 sin  
f z 2 cos 
 f x 2 cos 
 f x1 
 f x2
f x1 sin   f x 2 sin 
(Brace 1) (Brace 2)
Figure 2-4-3 Coordinate transformation
In a matrix form,
~
 f x1  c s   f x1   f x1 
~   s c  f  f 
 f z1     z1   z1 
~
f    f x2   f x2 
c s
 ~x 2      
 f z2   s c   f z2  f 
~      Cb  z 2  (2-4-9)
 c s  f x3
 f x3      f x3 
~
f   s c  f z3   f z3 
 ~z 3      
 f x4  c s f x4   f x4 
~  
f   s c   f z 4 Global f 
 z 4  Local   z4 
where
w h
c  cos   , s  sin  
l l
Since C b C b   I ,
T
Cb  is an orthogonal matrix, therefore,
Cb 1  Cb T (2-4-10)
34
In a similar manner, from Figure 2-4-4, the relation between the nodal displacements in local coordinate
and those of global coordinate can be obtained as,
u x1  u~x1 cos   u~z1 sin 
for Brace 1 (2-4-11)
u  u~ sin   u~ cos 
z1 x1 z1
and
u x 2  u~x 2 cos   u~z 2 sin 
for Brace 2 (2-4-12)
u  u~ sin   u~ cos 
z2 x2 z2
Eq. (2-4-12) can be also obtained from the Eq. (2-4-11) by replacing  by     and using the
formulas sin      sin  , cos      cos  .
u~3
u z4 u z3
u~4
4 3
u x4 u x3
h
l  w2  h 2
u~2 uz2
u z1 u~1
1 u x2
u x1 2
w
u z 2  u~x 2 sin   u~z 2 cos 
u~x 2
u z1  u~x1 sin   u~z1 cos 
u x 2  u~x 2 cos   u~z 2 sin 

u~x1 
u~z1

u x1  u~x1 cos   u~z1 sin  u~z 2
Figure 2-4-4 Coordinate transformation
In a matrix form,
 u x1  c  s   u~x1  u~x1 
u  s c   u~   u~ 
 z1     z1   z1 
u x 2   c s  u~x 2  u~x 2 
   ~  ~ 
u z 2   s c  u z 2  T u z 2 
    Cb   ~ 
u     u~x 3 
  x 3

c s
~  u x 3 
u z 3   s c 
 u z3  u~z 3 
     ~ 
u x 4  c  s  u~x 4  u x 4 
  ~
u    u~ 
 z 4  Local  s c  u z 4  Global  z4 
35
The stiffness matrix of brace element is,
N   k~  
 N 1   k1 0   1 
    or (2-4-13)
N 2   0 k 2   2 
Where
   nb u~  nb  Cb T  u  nb Cb u

1
(2-4-14)
 f   Cb 1 ~f   Cb T nb T N  (2-4-15)

 u1 
u 
u  TixBr  2  (2-4-18)

u n 
The component of the transformation matrix, [TixBr ] , is discussed in Chapter 4 (Freedom Vector).

 u1   u1 
u  u 
   nb Cb u  nb Cb TixBr    TxBr  2 
 2
(2-4-19)
 
u n  u n 
Finally, the constitutive equation of the brace is;
 P1   u1 
P  u 
 2  2
   K xBr   (2-4-20)
 
 Pn  u n 
where,
K xBr   TxBr T k Br TxBr  (2-4-21)
36
In case of Y-direction brace
z Z
y X
x Y
In case of Y-direction brace, transformation of the sign of the vector components of the element coordinate
is,
x  0 1 0  X 
   
 y   1 0 0  Y  (2-4-20)
z  
 Y  Beam  0 0 1  Z  Global
Therefore
 u x1  1   u y1  u y1 
u   1 u  u 
 z1     z1   z1 
u x 2   1  u y2
 u y 2 
      
u z 2   1  u z 2  u 
       z2  (2-4-21)
 1  u y3
u x 3     u y 3 
u z 3   1  u z 3  u z 3 
     
u x 4 
 1  u y 4  u y 4 
 
u 
 z 4 Y  Brace  1 u z 4  Global u z 4  Global
 
 u1 
u 
u  TiyBr  2
  (2-4-22)

u n 
 u1 
u 
   T yBr 
 2
 , T   n C T 
yBr b b iyBr (2-4-23)

u n 
37
The constitutive equation of the Y-direction brace is;
 P1   u1 
P  u 
 2

   K yBr 
 2
  (2-4-21)
 
 Pn  u n 
where,
K   T  k T 
yBr yBr
T
Br yBr (2-4-22)
38
In case of K-brace (or Cheveron brace)
3 6 4
1 5 2
3 6 6 4
l'
2
 w
l'    h2
2
1 5 5 2
For the left half part, as we defined before for the ordinary brace, the stiffness equation of brace element is,
 f L  k L uL k L  Cb T nb T k~ L nb Cb  (2-4-15)
where
 f L   f x1 f z1 f x5 f z5 f x3 f z3 f x6 f z6 
T
uL  u x1 u z1 u x 5 u z5 u x3 u z 3 u x6 uz6 
T
k~ k1, L

0 
k 2,L 
, nb   
1 0 0 0 0 1 0
0
0  1 0 1 0 0 0
 0
L
0
c s 
 s c 
 
 c s 
 
Cb   
 s c , c  cos 
w / 2  , s  sin  
h
c s  l' l'
 
 s c 
 c s
 
  s c 
39
For the right half part, in the same way, the stiffness equation of brace element is,
 f R  k R uR k R  Cb T nb T k~ R nb Cb  (2-4-15)
where
 f R   f x 5 f z5 f x2 f z2 f x6 f z6 f x4 f z4
T
uR  u x5 u z5 ux2 uz2 u x6 uz6 ux4 uz4

T
We can express the nodal displacement vector as,
u x1  u x1 
u  u 
 z1   z1 
u x1  u x 2  u x 2 
u     
 z1  u z 2  u z 2 
u x 5  [1] 
 u x3  u x 3 
    u   
u  [1]   z 3   D u z 3 
uL   z5   
u x 3    u x 4  L
[1] u x 4 
u z 3    
[1] u z 4 u z 4 
     
u x 6  u x 5  u x 5 
u  u  u 
 z6   z5   z5 
u x 6  u x 6 
u  u 
 z6   z6 
u x1  u x1 
u  u 
 z1   z1 
u x 5  u x 2  u x 2 
u     
 z5  u z 2  u z 2 
u x 2   [1] 
 u x3  u x 3 
    u   
u  [1]   z 3   D u z 3 
uR   z2   
u x 6   [1] u x 4 
R
u x 4 
u z 6   [1]
 
 uz4 u z 4 
     
u x 4  u x 5  u x 5 
u  u  u 
 z4   z5   z5 
u x 6  u x 6 
u  u 
 z6   z6 
40
We assume the displacements of intermediate nodes, 5 and 6, are calculated from those of end nodes as
follows,
u z 2, 4
u z 5, 6
1 1
u x 5  u x1  u x 2 , u z 5  u z1  u z 2  u z1,3 u x 2, 4
2 2 2,4
1 1 u x 5, 6
u x 6  u x 3  u x 4 , u z 6  u z 3  u z 4  u x1,3
5,6
2 2
1,3
In a matrix form
u x1  u x1 
u  u 
 z1   z1 
u x 5  1 / 2 0 1 / 2 0 0 0 0 0  u x 2  u x 2 
u   0 1/ 2 0 1/ 2 0     
 z5  0 0 0  u z 2  u 
       hCh  z 2 
u x 6   0 0 0 0 1 / 2 0 1 / 2 0  u x 3  u x 3 
u z 6     u z 3 
 0 0 0 0 0 1 / 2 0 1 / 2 u z 3
Local
   
u x 4  u x 4 
u  u 
 z 4  Local  z 4  Local
Therefore,
u x1 
u 
 z1 
u x 2  u x1  u x1 
    u 
u z 2   u z1   z1 
u x 3  u x 2  u x 2 
     
u z 3   I   u z 2  u 
      TCh  z 2 
u x 4  hCh  u x 3  u x 3 
u z 4  u z 3  u z 3 
     
u x 5  u x 4  u x 4 
u  u  u 
 z5   z 4  Local  z 4  Local
u x 6 
u 
 z6 
41
Therefore,
u x1  u x1 
u  u 
 z1   z1 
u x1  u x 2  u x1  u x 5  u x 2  u x1 
u    u  u    u 
 z1  u z 2   z1   z5  u z 2   z1 
u x 5  u x 3  u x 2  u x 2  u x 3  u x 2 
           
u  u  u  u  u  u 
uL   z 5   DL  z 3   DL TCh  z 2  , uR   z 2   DR  z 3   DR TCh  z 2 
u
 x 3 u
 x 4 u
 x 3 u x 6  u x 4  u x 3 
u z 3  u z 4  u z 3  u z 6  u z 4  u z 3 
           
u x 6  u x 5  u x 4  u x 4  u x 5  u x 4 
u  u  u  u  u  u 
 z6   z5   z 4  Local  z4   z5   z 4  Local
u x 6  u x 6 
u  u 
 z6   z6 
Finally the force-displacement relationship of Cheveron brace is,
k L  Cb T nb T k~ L nb Cb 
 f x1  u x1 
f  u 
 z1   z1 
 f x2  u x 2 
   
 f z2 
   T T T T u 

 TCh  DL  k L DL TCh   TCh  DR  k R DR TCh   z 2 
 f x3  u x 3 
 f z3  u z 3 
   
 f x4  u x 4 
f  u 
 z 4  Local  z4 
42
2.5 External Spring
 zB
 yA Z
B
A B A
 xA  xB X
B
Y
 yB A
 zA
Figure 2-5-1 Element model for external spring
Force-displacement relationship for the element

The relationship between the displacement vector and force vector of the elastic element in Figure 2-5-1 is
expressed as follows:
N 'i   k E  'i , i  x, y , z (2-5-1)
 ' x  u xB  u xA
 ' y  u yB  u yA (2-5-2)
 ' z   zB   zA
Therefore
u  u 
 xB   xB 
u  u 
 ' x    1 1 0 0 0 0 yA   n xE  yA  (2-5-3)
u yB  u yB 
 zA   zA 
   
 zB   zB 
u  u 
 xB   xB 
u yA 
 ' y   0 0  1 1 0 0uu yA   n yE     (2-5-4)
 yB  u yB 
 zA   zA 
   
 zB   zB 
43
u  u 
 xB   xB 
u yA  u 
 ' z   0 0 0 0  1 1   n zE  yA  (2-5-5)
u yB  u yB 
 zA   zA 
   
 zB   zB 
u xA 
u 
 xB  u1 
u yA  u 
 2
   TiE   (2-5-6)
u yB  
 zA  u n 
 
 zB 
The component of the transformation matrix, [TiE ] , is discussed in Chapter 4 (Freedom Vector).

 u1  u1 
u  u 
 'i   niE TiE    TE  2 , i  x, y, z
 2
(2-5-7)
 
u n  u n 
The constitutive equation of the external spring is;
 P1   u1 
P  u 
 2  2
   K E   (2-5-3)
 
 Pn  u n 
where,
K E   TE T k E TE  (2-5-4)
44
2.6 Base Isolation
 zB
B Z
u xB
u yB
l X
A u xB
Y
u yA
 zA
Figure 2-6-1 Element model for base isolation
Force-displacement relationship for the element

The relationship between the displacement vector and force vector of the element is expressed as follows:
Q' x 
   k pBI   '' 
x
(2-6-1)
Q' y   y 
Including the axial stiffness,

 Q ' x   k pBI
 
 0   'x   'x 
Q ' y    EA    ' y   k BI   ' y  (2-6-2)
  '   0 l'

   ' z   ' 
 z  z
From node displacements, relative displacements are;
 ' x  u xB  u xA
 ' y  u yB  u yA (2-6-3)
 ' z   zB   zA
Therefore
u  u 
 ' x   1 1  
xB  xB 
    u yA   n u yA 
 
 ' y    1 1  u  BI   (2-6-4)
 '    yB   u yB 
 z   1 1
 zA   zA 
   
 zB   zB 
45
u xA 
u 
 xB   u1 
u yA  u 
 2
   TiBI   (2-6-5)
u yB  
 zA  u n 
 
 zB 
The component of the transformation matrix, [TiBI ] , is discussed in Chapter 4 (Freedom Vector).

u1   u1 
 ' x  u  u 
   2  2
 ' y   n BI TiBI    TBI   (2-6-6)
 '   
 z u n  u n 
The constitutive equation of the Base isolation is;
 P1   u1 
P  u 
 2  2
   K BI   (2-6-7)
 
 Pn  u n 
where,
K BI   TBI T k BI TBI  (2-6-8)
46
2.7 Masonry Wall
Element model for Masonry wall is defined as a line element with a nonlinear shear spring and a vertical
spring in the middle of the wall panel as shown in Figure 2-6-1.
B1 B B2
N ' z1 ,  ' z1 N 'z 2 ,  'z 2

l
Q' xc ,  ' xc
A1 A2
A
Figure 2-7-1 Element model for masonry wall
The relationship between the shear deformation and shear force of the nonlinear shear spring is,
Q' xc  k sx  ' xc (2-7-1)
For axial spring,

N ' z1  k z  ' z 1 , N ' z 2  k z ' z 2 (2-7-2)
In a matrix form,
 Q' xc  k sx 0 0   ' xc   ' xc 

       
 N ' z1    0 kz 0   ' z1   k N  ' z1  (2-7-3)
N '   0 0 k z   ' z 2   ' 
 z2    z2 
Including node movement

The shear angle of the frame with four nodes, A1, A2, B1, B2, is defined as,
 z u x
  (2-7-4)
x z
where,
 z 1   zA2   zA1  zB 2   zB1 

    (2-7-5)
x 2 w w 
47
u z 1  u xB1  u xA1 u xB 2  u xA2 
    (2-7-6)
z 2 l l 
The shear deformation,  ' xc , is then,
 ' xc   l 
l
 zA2   zA1   zB 2   zB1   1 u xB1  u xA1  u xB 2  u xA2  (2-7-7)
2w 2
The axial deformation,  ' z1 ,  ' z 2 , is,
 ' z1   zB1   zA1 ,  'z 2   zB 2   zA2 (2-7-8)
In a matrix form,
 u xA1   u xA1 
   
 zA1   zA1 
 l l l l  u xA2  u xA2 
 ' xc   0.5  0.5 w  0.5 0.5 0.5  0.5 0.5 0.5  
w     
   w w  
  ' z1    0 1 0 0 0 1 0 0   zA 2   DN  zA2 
 u
 '   0 0 0 1 0 0 0 1   xB1  u xB1 
 z2     zB1 
   zB1   
u xB 2  u xB 2 
   
 zB 2   zB 2 
(2-7-9)
 u xA1 
 
 zA1 
u xA2  u1 
  u 
 zA 2   2
   T 
ixN   (2-7-10)
 u xB1  
 zB1  u n 
 
u xB 2 
 
 zB 2 
The component of the transformation matrix, [TixN ] , is discussed in Chapter 4 (Freedom Vector).

48
 u1   u1 
 ' xc  u  u 
   2  2
 '
 z1   D N T 
ixN    T 
xN   (2-7-11)
 '   
 z2  u n  u n 
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,
x  0 1 0  X 
   
 y   1 0 0  Y  (2-7-12)
z  
 Y  Beam  0 0 1  Z  Global
Therefore
 u xA1  1   u yA1   u yA1 

   1     
 zA1     zA1   zA1 
u xA2   1  u yA 2
 u yA2 
      
 zA 2   1   zA2   zA 2 
   
 u yB1   (2-7-13)
u  1 u yB1 
 xB1      
 zB1   1   zB1   zB1 
      
u xB 2  1  u yB 2  u yB 2 
 
   1   zB 2  Global  zB 2  Global
 zB 2 Y Wall 
49
 u yA1 
 
 zA1 
u yA2  u1 
  u 
 zA2 
  
  TiyN
 2
  (2-7-14)
 u yB1  
 zB1  u n 
 
u yB 2 
 
 zB 2 
 u1   u1 
 ' xc  u  u 
 
 ' z1   D N  TiyN   2
 
   T yN
 2
  (2-7-15)
 '   
 z2  u n  u n 
Finally, the constitutive equation of the wall is;
 P1   u1 
P  u 
 2  2
   K xN   (2-7-16)
 
 Pn  u n 
where,
K xN   TxN T k N TxN  (2-7-17)
For Y-wall,
 P1  u1 
P  u 
 2
 
   K yN
 2
  (2-7-18)
 
 Pn  u n 
where,
K   T  k T 
yN yN
T
N yN (2-7-19)
50
2.8 Passive Damper
Element model for passive damper with a shear spring is defined as a line element with a nonlinear shear
spring as shown in Figure 2-8-1.
B1 B B2
l
Q' xc ,  ' xc
A1 A2
A
Figure 2-8-1 Element model for passive damper
The relationship between the shear deformation and shear force of the nonlinear shear spring is,
Q' xc  k sx  ' xc (2-8-1)
Including node movement

The shear angle of the frame with four nodes, A1, A2, B1, B2, is defined as,
 z u x
  (2-8-2)
x z
where,
 z 1   zA2   zA1  zB 2   zB1 

    (2-8-3)
x 2 w w 
u z 1  u xB1  u xA1 u xB 2  u xA2 

    (2-8-4)
z 2 l l 
The shear deformation,  ' xc , is then,
 ' xc   l 
l
 zA2   zA1   zB 2   zB1   1 u xB1  u xA1  u xB 2  u xA2  (2-8-5)
2w 2
51
The axial deformation,  ' z1 ,  ' z 2 , is,
 ' z1   zB1   zA1 ,  'z 2   zB 2   zA2 (2-8-6)
In a matrix form,
 u xA1   u xA1 
   
 zA1   zA1 
 l l l l  u  u xA2 
 ' xc   0.5  0.5 w  0.5 0.5 0.5  0.5 0.5 0.5   xA2 
w  zA2   
   w w  zA2 
  ' z1    0 1 0 0 0 1 0 0   DD 
 u xB1  
u xB1 
 '   0 0 0 1 0 0 0 1    
 z2 
   zB1   zB1 
   
u xB 2  u xB 2 
   
 zB 2   zB 2 
(2-8-7)
 u xA1 
 
 zA1 
u xA2  u1 
  u 
 zA 2   2
   TixD   (2-8-8)
u xB1  
 zB1  un 
 
u xB 2 
 
 zB 2 
The component of the transformation matrix, [TixD ] , is discussed in Chapter 4 (Freedom Vector).

 u1   u1 
 ' xc  u  u 
   2  2
 '
 z1   DD T 
ixD    T 
xD   (2-8-9)
 '   
 z2  un  un 
52
In case of Y-direction damper
z Z
y X
x Y
In case of Y-direction damper, the damper direction coincides to the Y-axis in the global coordinate,
x  0 1 0  X 
   
 y   1 0 0  Y  (2-8-10)
z  
 Y  Beam  0 0 1  Z  Global
Therefore
 u xA1  1   u yA1   u yA1 

   1     
 zA1     zA1   zA1 
u xA2   1  u yA 2
 u yA2 
      
 zA 2   1   zA2   zA 2 
   
 u yB1   (2-8-11)
u  1 u yB1 
 xB1      
 zB1   1   zB1   zB1 
      
u xB 2  1  u yB 2  u yB 2 
 
   1   zB 2  Global  zB 2  Global
 zB 2 Y Wall 
 u yA1 
 
 zA1 
u yA2  u1 
  u 
 zA 2 
  
  TiyD
 2
  (2-8-12)
 u yB1  
 zB1  un 
 
u yB 2 
 
 zB 2 
53
 u1   u1 
 ' xc  u  u 
 
 ' z1   DD  TiyD   2
 
   TyD
 2
  (2-8-13)
 '   
 z2  un  un 
Finally, the constitutive equation of the damper is;
 P1   u1 
P  u 
 2  2
   K 
xD   (2-8-14)
 
 Pn  un 
where,
K xD   TxD T kD TxD  (2-8-15)
For Y-damper,
 P1  u1 
P  u 
 2
   K yD   2
  (2-8-16)
 
 Pn  un 
where,
K   T  k T 
yD yD
T
D yD (2-8-17)
54
Appendix : Calculation of shear component
For “Masonry Wall” and “Passive Damper”, the shear deformation is defined as follows:
1) Shear deformation in one direction
Shear strain is τ = Δl / l ≈θ
2) Shear deformation in two directions
Shear strain is τ = θ1 + θ2 = Δlx / lｙ+Δly / lx
u x u y
If we discuss small element    Eq. (2-7-4) and Eq. (2-8-2)
y x
55
This definition is necessary to remove rotational component. To explain this, suppose there is only
rotational (or bending) deformation,
-θ
From the above definition, shear angle will be

τ= θ + (- θ) = 0
For example, in the upper story of the building under horizontal deformation, the bending
component is dominant and the shear component is small. Therefore, the brace damper
doesn’t work in the upper story.
56
3) In case of damper element
We define the shear angle in one direction as follows:
θ1 θ2
We adopt the average angle,
θ = 1/2 ( θ1 + θ2 )  Eq. (2-7-5) and Eq. (2-8-3)
In the same way, the shear angle in another direction is
θ'2
θ'1
θ’ = 1/2 ( θ’1 + θ’2 )  Eq. (2-7-6) and Eq. (2-8-4)
57
2.9 Floor Element
In the default setting, STERA 3D adopts “rigid floor”. However, elastic deformation of a floor diaphragm
in-plane can be considered by the option menu selecting “flexible floor”. The stiffness matrix of the floor
element is constructed using a two dimensional isoparametric element.
Figure 2-9-1 4-nodes isoparametric element
The stiffness matrix with 4-nodes isoparametric is expressed as,
 P1  u1 
Q  v 
 1  1
 P2  u 2 
   
Q2  v 2 
   K F  
 P3  u 3 
Q3   v3 
   
 P4  u 4 
Q  v 
 4  4
F =K u (2-9-1)
The coordinate transfer function {x, y} is expressed using the interpolation functions as follows:
4
1 1 1 1
x (r , s )   hi (r , s ) xi  (1  r )(1  s ) x1  (1  r )(1  s ) x 2  (1  r )(1  s ) x3  (1  r )(1  s ) x 4
i 1 4 4 4 4
4
1 1 1 1
y (r , s )   hi (r , s ) y i  (1  r )(1  s ) y1  (1  r )(1  s ) y 2  (1  r )(1  s ) y 3  (1  r )(1  s ) y 4
i 1 4 4 4 4
(2-9-2)
58
The deformation function {u, v} is also expressed using the same interpolation functions.
4
1 1 1 1
u (r , s )   hi ( r , s )u i  (1  r )(1  s )u1  (1  r )(1  s )u 2  (1  r )(1  s )u 3  (1  r )(1  s )u 4
i 1 4 4 4 4
4
1 1 1 1
v ( r , s )   hi ( r , s ) v i  (1  r )(1  s )v1  (1  r )(1  s )v 2  (1  r )(1  s )v 3  (1  r )(1  s )v 4
i 1 4 4 4 4
(2-9-3)
Stiffness matrix can be obtained from the “Principle of Virtual Work Method,” which is expressed in the
following form:
  dv  u T F
T
(2-9-4)
V
where,  is a virtual strain vector,  is a stress vector, u is a virtual displacement vector and F is a
load vector, respectively.
In case of the plane problem, the strain  vector is defined as,

 u 
 
  x   x 
   v 
y    (2-9-5)
    y 
 xy   u v 
 y  x 
 
Substituting equation (2-9-3) into equation (2-9-5), the strain vector is calculated from the nodal
displacement vector as,
 u   4
hi 
   
i 1 x
ui 
  x   x   
   v   4
hi 
y      vi 
    y   i 1 y 
 xy   u v   4 hi 4
hi 
 y  x    y u i   x vi 
   i 1 i 1 
 u1 
 
 h1 h2 h3 h4  v1 
 0 0 0 0  u 
 x x x x  2 
 h1 h2 h3 h4  v 2 
 0 0 0 0  u 
 y y y y  3 
 h1 h1 h2 h2 h3 h3 h4 h4  v3 
 y x y x y x y x  
  u4
 
v 
 4
ε= B u (2-9-6)
59
In the plane stress problem, the stress-strain relationship is expressed as,
 x    
  1  0   x 
 y  
E  1 0    (2-9-7)
  1    1   y 
 xy  0 0   xy 
 2 
σ= C ε
Substituting equation (2-9-6) into equation (2-9-7),
σ= C B u (2-9-8)
From the Principle of Virtual Work Method,
 
 Bu  CBu dv  u  B T CBdxdy u  u T F
T

T
(2-9-9)
 
V  V ( x, y ) 
Therefore, the stiffness equation is obtained as,
F  Ku , K   B T CBdv (2-9-10)
V
If we assume the constant thickness of the plate (= t), using the relation dv  tdxdy ,
K t B
T
CBdxdy (2-9-11)
V ( x, y )
Since this integration is defined in x-y coordinate, we must transfer the coordinate into r-s coordinate to use
the numerical integration method. Introducing the Jacobian matrix,
 x y 
 
J   r r  ; Jacobian Matrix (2-9-12)
 x y 
 
 s s 
the above integration is expressed in r-s coordinate as,
1 1
 ( x, y )
K  t   B 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
60
Evaluation of Jacobian Matrix
y   hi x hi 
4 4
 x
  
r    i 1 r
i  r yi 
J   r i 1  (2-9-15)
 x y   4 hi hi 
4
s    
  xi yi 
 s i 1 s i 1 s 
Evaluation of the matrix B
 h1 h2 h3 h4 

 0 0 0 0 
 x x x x 
 h1 h2 h3 h4 
B 0 0 0 0  (2-9-16)
 y y y y 
 h1 h1 h2 h2 h3 h3 h4 h4 
 y x y x y x y x 

h1 h h h
The derivatives , , 4 , 1 , , 4 are calculated as,
x x y y
h1 h1 r h1 s h h r h4 s
  ,  , 4  4  ,
x r x s x x r x s x
h1 h1 r h1 s h h r h4 s
  ,  , 4  4 
y r y s y y r y s y
In a matrix form,
 h1 h2 h3 h4   r s  h1 h2 h3 h4 

    
 x x x x    x x  r r r r 
 h1 h2 h3 h4   r s  h1 h2 h3 h4 
 y
 y y y   y
  y  s s s s


 h1 h2 h3 h4 

 
 J 1  r r r r  (2-9-17)
 h1 h2 h3 h4 
 
 s s s s 
Evaluation of partial derivatives of the interpolation functions
h1 1 h1 1
 (1  s )  (1  r )
r 4 s 4
h2 1 h2 1
  (1  s )  (1  r )
r 4 s 4
, (2-9-18)
h3 1 h3 1
  (1  s )   (1  r )
r 4 s 4
h4 1 h4 1
 (1  s )   (1  s )
r 4 s 4
61
The 3 points Gauss Integration Formula is defined as:
 f (t )dt  0.5556 f (0.7746)  0.8889 f (0)  0.5556 f (0.7746)

1
(2-9-19)
  1 f (t1 )   2 f (t 2 )   3 f (t 3 )
where,  1  0.5556,  2  0.8889,  3  0.5556

t1  0.7746, t 2  0, t 3  0.7746
f(0.7746)
f(t)
f(-0.7746)
f(0)
t
-1 -0.7746 0 +0.7746 +1
The stiffness matrix is then calculated numerically as follows:
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   i 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)
 1  0.5556,  2  0.8889,  3  0.5556

r1  s1  0.7746, r2  s 2  0, r3  s3  0.7746
62
Transformation from global node displacements to element node displacements is,
 u1 
v 
 1
u 2   u1 
  u 
v 2   2
   T 
iF   (2-9-21)
u 3  
 v3  u n 
 
u 4 
v 
 4
The component of the transformation matrix, [TiF ] , is discussed in Chapter 4 (Freedom Vector).
63
2.10 Connection Panel
1) General case
In the default setting, STERA3D assumes the rigid connection zone between column and beam. You can
consider shear deformation of the connection area (we call “connection panel”) by the “Connection
member” menu.
Figure 2-10-1 Connection area
Shear deformation of the connection panel, γ, is defined as shown in Figure 2-10-2.
uC= -0.5γAh θC= -0.5γA
C C
h
0.5γA 0.5γA
A B θB= 0.5γA A 0.5γA vB= -0.5γAw

h B
w w
Figure 2-10-2 Definition of shear deformation
Differences of displacement at node B and C are;
 u B   0   u C   0.5 A h
       
Node B:  v B    0.5 A w , Node C:  vC    0  (2-10-1)
   0.5      0.5 
 B  A   C  A 
64
First we consider nodal movement without shear deformation of the connection panel. As shown in Figure
2-10-3, the displacement at node B and node C will be;
u B   u A  u C  u A   A h
       
Node B: v B   v A   A w , Node C: vC    v A  (2-10-2)
         
 B  A   C  A 
Then, we consider shear deformation of the connection as shown in Figure 2-10-4. By adding Equation
(2-10-1) to (2-10-2), the displacement at node B and node C will be;
  Ah
h
 Aw
A
C w
vA
uA
A B
Figure 2-10-2 Nodal movement without shear deformation of the panel
A B
Figure 2-10-4 Nodal movement with shear deformation of the panel
65
Node B:
u 
u B   u A   0   uA  1 0 0 0  A 
          v A 
v B   v A   A w   0.5 A w  v A   A w  0.5 A w  0 1 w  0.5w  
      0.5     0.5 
 B    A  
 0 0 1
  0.5   A 
A A A
 A 
(2-10-3)
Node C:
u A 
u C  u A   A h  0.5 A h u A   A h  0.5 A h 1 0 h  0.5h  
         v 
 vC    v A    0  vA   0 1 0 0   A 
       0.5     0.5  0 0 1  0.5   A 
 C  A   A   A A     
 A
(2-10-4)
2) Beam element
In case of rigid connection, as described in Equation (2-1-7), the nodal displacement is expressed as,
 ' yA   yA    u zB   B l ' yB   u zA   Al ' yA 

  ,  
 ' yB   yB    l'
u 
 1 1  1 1  zA
  u     u     l'  1  A  B  u 
yA
l'
zA A yA
l'
zB B yB
l'  zB 
   1    (2-10-5)
1 1
 yB  u zA   A yA  u zB   B yB   
1
A 
1   B   yA 
 l' l'  l' l'   yB 
 
 yB
 ' yA  yA 
 yB
 yA   ' yB
u zB
Z
u zB   B l ' yB
u zA u zA   A l ' yA
A B X
 Al ' l' B l '
Figure 2-10-5 Beam displacement with rigid connection
66
If we consider shear deformation of connection panel, from Figure 2-10-6,
 ' yA   yA  0.5 yA    u zB  B l '  yB  0.5 yB   u zA   Al '  yA  0.5 yA 

  ,  
 ' yB   yB  0.5 yB    l'
 1 1 
 yA  l ' u zA   A yA  l ' u zB   B yB  0.5 yA  0.5 A yA  0.5 B  yB 
 
1 1
 yB  u zA   A yA  u zB   B yB  0.5 yB  0.5 A yA  0.5 B  yB 
 l' l' 
u zA 
u 
1 1   zB 
l'  1  A B 0.5  0.5 A  0.5 B   
l '  yA 
    (2-10-6)
1 1
   A 1  B  0.5 A 0.5  0.5B   yB 
l' l'   yA 
 
 yB 
 yB

 ' yA  yA  0.5 yA
 yB  0.5 yB  ' yB

 yA u zB
u zB   B l '  yB  0.5 yB 
u zA u zA   Al '  yA  0.5 yA 
A B
Z
 Al ' l' B l '

X
Figure 2-10-6 Beam displacement with shear deformation of connection panel
67
The transformation matrices for beam element are;
Including connection panel and node movement
u zA  u zA 
u  u 
1 1   zB   zB 
 ' yA   l '  l ' 1   A B 0.5  0.5 A  0.5 B 
 yA   yA 
 '   1 
    
1 
 yB  
   l'  l' A 1  B  0.5 A 0.5  0.5 B      B  yB 
 yB 
  xA     yA
1     yA 
  xB    yB  yB 

 1    
 xA   xA 
   
 xB   xB 
(2-10-10)
u zA 
u 
 zB 
 yA   u1 
  u 
 yB   2
   TixB   (2-10-11)
 yA  
 yB  u n 
 
 xA 
 
 xB 

u1   u1 
 ' yA  u  u 
   2  2
 '
 yB   n B  B T 
ixB    T 
xB   (2-10-12)
 '    
 x un  un 
68
In case of Y-direction beam
x  0 1 0  X 
   
 y   1 0 0  Y  (2-10-13)
z  
 Y  Beam  0 0 1  Z  Global
u zA  1  u zA  u zA 
u   1  u  u 
 zB     zB   zB 
 yA   1   xA   xA 
      
 yB   1    xB 
   s B  
xB
   
  xA  (2-10-14)
 yA  
1
   xA 
 yB   1   xB   xB 
     
 xA 
 1   yA   yA 
 
  1  yB  Global  yB 
 xB Y  Beam    Global
u zA 
u 
 zB 
 xA   u1 
  u 
 xB 
   TiyB    2
  (2-10-15)
 xA  
 xB  u n 
 
 yA 
 yB 
 
 u1   u1 
 ' yA  u  u 
 
 
 ' yB   n B  B s B  TiyB
 2
 
   T yB
 2
  (2-10-16)
 '   
 x u n  u n 
69
3) Column element
In case of rigid connection, as described in Equation (2-2-16), the nodal displacement in X-Z plane is
expressed as,
 ' yA   yA    u xA   Al ' yA   u xB  B l ' yB 

  ,  
 ' yB   yB    l'
u 
 1 1   1 1  xA
  u     u      l ' 1  A  B  u 
yA
l '
xA A yA
l '
xB B yB
l'  xB 
  1    (2-10-17)
 A 1   B   yA 
1 1 1
 yB  u xA   A yA  u xB   B yB  
 l' l'   l' l'   yB 
 
B u xB
Bl '  yB
u xB   B l ' yB  ' yB
 yB

l' Z

 yA
u xA   A l ' yA  ' yA X
 yA
 Al '
Y
A u xA
Figure 2-9-7 Column displacement with rigid connection (X-Z plane)
70
 ' yA   yA  0.5 yA    u xA   Al '  yA  0.5 yA   u xB  B l '  yB  0.5 yB 

  ,  
 ' yB   yB  0.5 yB    l'
 1 1 
 yA  l ' u xA   A yA  l ' u xB   B yB  0.5 yA  0.5 A yA  0.5 B  yB 
 
1 1
 yB  u xA   A yA  u xB   B yB  0.5 yB  0.5 A yA  0.5 B  yB 
 l' l' 
u xA 
u 
 1 1   xB 
 l ' 1  A B  0.5  0.5 A 0.5 B   yA 
 l'
1 1    (2-10-18)
  A 1  B 0.5 A  0.5  0.5 B   yB 
 l' l'   yA 
 
 yB 
B u xB
B l '  yB
u xB   B l '  yB  0.5 yB 
 ' yB
Z  yB  0.5 yB

l'
X

 yA  0.5 yA
 ' yA
u xA   Al '  yA  0.5 yA 
 yA
 Al '
u xA
A
Figure 2-9-8 Column displacement with shear deformation of connection panel (X-Z plane)
71
In the same manner, assuming rigid connection, the nodal displacement of column in Y-Z plane is
expressed as,
 ' xA   xA    u yB   Al ' xB   u yA   Al ' xA 

  ,  
 ' xB   xB    l'
  yA 
u
 1 1  1 1
 xA  l ' u yA   A xA  l ' u yB   B xB   l '  l ' 1   A B  u
 yB 
       (2-10-19)
 A 1  B   xA 
1 1 1 1
 xB  u yA   A xA  u yB   B xB   
 l' l'  l' l'   xB 
Bl ' u yB
u yB   B l ' xB
 xB
 'xB
l'  xB
 Z

 xA u yA   Al ' xA
X
 'xA
 Al '
 xA A Y
u yA
Figure 2-9-9 Column displacement with rigid connection (Y-Z plane)
72
 ' xA   xA  0.5 xA   u yB   B l '  xB  0.5 xB   u yA   Al '  xA  0.5 xA 

  ,  
 ' xB   xB  0.5 xB   l'
 1 1 
 xA  l ' u yA   A xA  u yB   B xB  0.5 xA  0.5 A xA  0.5 B  xB 
l'
 
1 1
 xB  u yA   A xA  u yB   B xB  0.5 xB  0.5 A xA  0.5 B  xB 
 l' l' 
u yA 
u 
1 1   yB 
l'  1  A B  0.5  0.5 A 0.5 B   xA 
 l '
  (2-10-20)
 0.5  0.5 B   xB 
1 1
  A 1  B 0.5 A
l' l'   xA 
 
 xB 
u yB B
 xB Bl '
u yB   B l '  xB  0.5 xB 
 ' xB
  xB  0.5 xB
Z
l'

Y
 xA  0.5 xA
 ' xA
u yA   Al '  xA  0.5 xA 
 xA
 Al '
u yA
A
Figure 2-9-10 Column displacement with shear deformation of connection panel (Y-Z plane)
73
The transformation matrices for column element are;
Including connection panel and node movement
u xA 
u 
 xB 
 yA 
 1 1 1 A B   yB 
 1  A B    
 ' yA   l '   yA 
l' 2 2 2
 '   1 1 A 1 B
 A 1  B     
 yB   l ' l' 2 2 2   yB 
 ' xA   1 1 1 A B  u yA 
    1  A B    u 
 'xB   l' l' 2 2 2
  yB 
  1 1 A 1 B
  zA     A 1  B     xA 
l' l' 2 2 2
  zB    
  
1   xB 
  zA   1   xA 
    
 zB   1   xB 

 1  zA 
 
 zB 
 zA 
 
 zB 
u xA 
u 
 xB 
 yA 
 
 yB 
 yA 
 
 yB 
u 
 yA 
u 
  C  yB 
 xA 
 xB 
 
 xA 
 
 xB 
 zA 
 
 zB 
 zA  (2-10-21)
 
 zB 
74
u xA 
u 
 xB 
 yA 
 
 yB 
 yA 
 
 yB 
u  u1 
 yA  u 
u yB   2
   TiC  
 xA  
 xB  u n 
 
 xA 
 
 xB 
 zA 
 
 zB 
 zA 
  (2-10-22)
 zB 

 ' yA 
 ' 
 yB   u1   u1 
  u 
 ' xA  u 2   2
   n  T 
iC    T 
C   (2-10-23)
 '
 xB 
C C
  
  'z  u n  u n 
 
  ' z 
75
4) Force-displacement relationship for the connection
x
y
Figure 2-9-11 Shear deformation of connection area
The relationship between the displacement vector and force vector of the element is expressed as follows:
M Px  k Px 0   x 
   
k Py   y 
(2-10-24)
M Py   0
where, initial stiffness of connection area is,
k Px  k Py  GV (2-10-25)
where, G is the shear modulus and V is the volume of the connection.

 u1 
 x  u 
 2
   T 
P   (2-10-26)
 y  
u n 
The component of the transformation matrix, [TP ] , is discussed in Chapter 4 (Freedom Vector).
The constitutive equation of the external spring is;
 P1   u1 
P  u 
 2  2
   K P   (2-10-27)
 
 Pn  u n 
where,
K P   TP T k P TP  (2-10-28)
76
3. Hysteresis model of nonlinear springs
Notation
at : Area of rebar in the tension side of the section
As : Total area of rebar in the section
y : Strength of rebar
B : Compression strength of concrete
 wy : Strength of shear reinforcement
D : Depth of the section

d : Effective depth of the section.
b : Width of the beam
j : Distance between the centers of stress in the section (  7 / 8d ).
Ze : Section modulus including the slab effect.

n : Ratio of Young’s modulus (= E s / Ec )
pt : Tensile reinforcement ratio
pw : Shear reinforcement ratio
Ie : Moment of inertia of section considering the slab effect
Mc : Crack moment
My : Yield moment
M/(QD) : Shear span-to-depth ratio
c : Crack rotation of the beam end
y : Yield rotation of the beam end
c : Crack rotation of the nonlinear bending spring
y : Yield rotation of the nonlinear bending spring
k0 : Initial stiffness
ky : Tangential stiffness at the yield point
k y2 : Stiffness after the yield point in the nonlinear bending spring
k y3 : Stiffness after the ultimate point in the nonlinear shear spring
y : Stiffness degradation factor at the yield point
Qc : Crack shear force
Qy : Yield shear force
Qu : Ultimate shear force
xs : Distance between the corner springs in the Multi-spring model
77
c : Crack shear deformation
y : Yield shear deformation
u : Ultimate shear deformation
78
3.1.1 Beam
3.1.1 RC Beam
a) Section properties
S
as
d1
t d1
a1
D
d2 a2 d2
B
B : Width of beam,
D : Height of beam,
S : Effective width of slab,
t : Thickness of slab
d1 : Distance to the center of top main rebars,
d2 : Distance to the center of bottom main rebars,
a1 : Area of top main rebars,
a2 : Area of bottom main rebars
as : Area of rebars in slab
Figure 3-1-1 RC Beam Section
Area of section to calculate axial deformation
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  
 BD g    S  B t  D   g  
t
Ie  
12 12  2  2 
2
 
n E  1a1 d1  g   n E  1a 2 D  d 2  g   n E  1a S  D   g 
2 2 t
(3-1-3)
 2 
where, g is the center of beam section calculated by
BD 2 / 2  ( S  B )t D  t / 2  n E  1a1 d 1  a 2 D  d 2   a S D  t / 2 
g (3-1-4)
AN
79
b) Nonlinear bending spring
A
x
MA MB
A B
B
Nx
elastic element
A A
MA MB MA MB
A B A B
Nx B B

l
 A   A   A   A  A
        MA MB
 B    B   B    B  A B B
     0   0 
 x  x    
Hysteresis model of a nonlinear bending spring is defined as the moment-rotation relationship under the
anti-symmetry loading in Figure 3-1-3. The initial stiffness of the nonlinear spring is supposed to be infinite,
however, in numerical calculation, a large enough value is used for the stiffness.
 
M   M
A B
 
Moment distribution
M M M
My My
6 EI = +
Mc k0  Mc
l
k 0  y k0 k0 kp  
c y   c  y 
Elastic element Nonlinear bending spring
Figure 3-1-3 Moment – rotation relationship at bending spring
80
Crack moment force
For reinforced concrete elements, the crack moment, M c is calculated as,
M c1  0.56  B Z e1 , Z e1  I e / g when tension in top main rebars (3-1-5)
M c 2  0.56  B Z e 2 , Z e 2  I e / D  g  when tension in bottom main rebars (3-1-6)
where,
B : Compression strength of concrete (N/mm2)
Z e1 , Z e 2 : Section modulus
Yield moment force

The yield moment, M y is calculated as,
M y1  0.9a1 y D  d1   0.9a S  y D  t / 2 when tension in top main rebars (3-1-7)
M y 2  0.9a 2 y D  d 2  when tension in bottom main rebars (3-1-8)
where,
y : Strength of rebar (N/mm2)
Yield rotation
The tangential stiffness at the yield point, k y , is obtained from the following equation,:
6 Ec I e
k y   y k0 , k0  (3-1-9)
l
where,
 y is the stiffness degradation factor at the yield point, which is obtained from the following
empirical formulas:
 y  0.043  1.63npt  0.043a / D d / D 2 , a / D  2 (3-1-10)
 y   0.0836  0.159a / D d / D 2 , a / D  2 (3-1-11)

where,

pt  a1  aS  / BD  (when tension in top main rebars)
pt  as  / BD  (when tension in bottom main rebars)
a/D : ≈ Shear span-to-depth ratio (= l /(2 D) )
d : Effective depth
d  D  d1 (when tension in top main rebars)
d  D  d2 (when tension in bottom main rebars)
81
 y is modified in case of tension in top main rebars as
I
 y '   y e0 (3-1-12)
Ie
BD 3
where I e0  : the moment of inertia of square section without slab
12
The yield rotation of the nonlinear bending beam,  y , is then obtained from,
 1 My
 y    1 (3-1-13)
 k
y  0
In general, the relation between the rotation of bending spring and that of nonlinear bending spring is
My
   (3-1-14)
k0
Crack rotation
From Figure 3-1-2, the crack rotation of the nonlinear bending beam,  c , is supposed to be zero value,
however, in STERA_3D program, it is assumed as,
 c  0.001  y (3-1-15)
82
Effective width of slab
Ls
B Lb
Sb
Sb B
Figure 3-1-4 Effective slab area for flexural capacity of beam
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)
where, Lb : Length of beam

D : Height of beam
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:
1) For calculating section are and moment of inertia

S b  0.1 Lb  D
2) For calculating the yield moment, M y , in Equation (3-1-8),
S b   s Ls (3-1-17)
where, Ls : Length of span

s : Effective slab ratio 0.1 ~ 0.5 , the default value is 0.1.
83
Hysteresis model
To consider the difference of the flexural capacity between positive and negative side of the beam, a
degrading tri-linear slip model is developed based on the Takeda Model for the hysteresis model of the
bending springs of the beam.
M M
kp
My My
Mc Mc
ks kr
 
x y m
 
My  My  y  Mm  y
kp    kr    k s    (3-1-18)
   
 y   y  m  m   x  m
Figure 3-1-5 Degrading Tri-linear Slip Model

(α=0.5, β=0.0 and η=0.001 as default values)
The strength degradation under cyclic loading is considered by elongating the target displacement,  m , to
be  ' m as shown in the following Figure:
M
kp
Mm
My
ks kr

n m n y  y m  'm
x
 y 
 ' m  1    m
n
 (3-1-19)
 n m 
Figure 3-1-6 Introducing strength degradation (γ=0.0 as default value)
84
Relationship between curvature and rotation
c
 EI
M M
A B
 M
D

l
Figure 3-1-7 Rotation angle and curvature at beam ends
Let’s think about the relationship between curvature and rotation at the end of a beam.
In the above loading condition, the relationship between moment and rotation is
6 EI
M  (3-1-20)
l
On the other hand, the relationship between moment and curvature is
M
 (3-1-21)
EI
Therefore,
6
  (3-1-22)
l
Assuming the neutral axis is in the middle of the section, the relationship between curvature and
compression strain at the section end is
c
 (3-1-23)
D/2
Therefore, the relationship between rotation and compressive strain is
l l
  c (3-1-24)
6 3D
l
Assuming D  , then
9
  3 c (3-1-25)
If  c reaches 0.003,  is around 0.01 (=1/100).

It corresponds to the yielding rotation of a beam.
85
c) Nonlinear shear spring
Hysteresis model of nonlinear shear spring is defined as the shear force – shear rotation relationship using
an origin-oriented poly-linear model.
A
MA  MB
MA MB Q
A B B l
 A  B   l
l
Q Q
k y 3  0.001k 0
Qy
k y3
Qu

Qc
k 0  GA
k0
c y u 
Figure 3-1-8 Force–deformation relationship of shear spring
Yield shear force
The yield shear force, Q y is calculated as,
 0.053 pt 0.23 ( B  18) 

Qy    0.85 p w   wy b  j (3-1-26)
 M /(QD)  0.12 
where,

Crack shear force

The crack shear force is, Qc , is assumed as,
Qy
Qc  (3-1-27)
3
86
Ultimate shear force
The ultimate shear force is, Qu , is assumed as,
Qu  Qc (3-1-28)
Crack shear deformation

The crack shear deformation is obtained as,
Qc
c  (3-1-29)
GA
Yield shear displacement
The yield shear deformation is assumed as,
1
y  (3-1-30)
250
Ultimate shear displacement
The ultimate shear deformation is assumed as,
1
u  (3-1-31)
100
NOTE)
In STERA_3D, the stiffness after yielding is temporary assumed to be positive to avoid instability in
numerical analysis.
Q
Qu
Qy
k y 3  0.001k 0
Qc
k 0  GA
c y u 
Figure 3-1-9 Stiffness after yielding
87
d) Modification of initial stiffness of nonlinear springs
In numerical calculation, a large dummy value is used for the initial stiffness of the nonlinear spring to
represent rigid condition. This large stiffness may cause an error for estimating the force from the
displacement. One possible way to solve the problem is to reduce the initial stiffness of the nonlinear spring
to be a certain value reasonable for calculation, and on the other hand, increase the stiffness of the elastic
element so that the total initial stiffness of the beam element does not change from the original one. This
idea is proposed by K-N Li (2004) for MS model.
 
M   M
A B
 

M M M
My My
6 EI = +
Mc k0  Mc
l
k 0  y k0 k0 kp  
c y   c  y 
My My  1 M
y  =  + y    1 y
 y k0 k0   k
 y  0
Increase Reduce
stiffness stiffness
M M
My
+ Mc
k  k0 /  1
*
0
k *p  EI / p1
* c*  y* *
My  1M y  1 M
y  = *  +  y*    1  y
 y k0 k0   k
 y  0
Figure 3-1-10 Modification of moment – rotation relationship
88
The idea is realized using flexibility reduction factors,  1  1,  2  1 , in the relationship between the
displacement vector and force vector of the elastic element in Equation (2-1-1) as,
 l' l' 
 1 3EI 
6 EI y
0 
 ' yA   y   M ' yA 
   l' l'  
 ' yB     2 0  M ' yB  (3-1-32)

  '   6 EI y 3EI y
 N ' 
 x l'  x 
 0 0
 EA 
l' l' l' l'

It must be 1  or  1  0.5 and.  2  or  2  0.5 .
3EI y 6 EI y 3EI y 6 EI y
Also the initial flexibility matrix of the nonlinear spring can be expressed as follows, introducing the
parameters, p1 , p 2 to increase the initial flexibility.
 yA   p1 EI 0   M ' yA 
    (3-1-33)
 yB   0 p2 EI  M ' yB 
When p1  0, p 2  0 , it represents the infinite stiffness for rigid condition. Accordingly, the crack and
yield rotation will be modified as,
Mc  1 M
 c*  p1 , y    1  y
*
(3-1-34)
EI   k
 y  0
My
    1 (3-1-35)
k0
Making the modified flexibility matrix to be identical to the original one,
 l' l'   p1 l' l' 

 3EI 
6 EI y
0   EI   1 3EI 
6 EI y
0 
 y   y 
 l'  l' 
0 
p2
 2 0  (3-1-36)
 3EI y  EI 3EI y
   
 sym. l'   l' 
sym.
 EA  original  EA  mod ified
This gives the flexivility reduction factors as:

3 3
1 1 p1 ,  2  1  p 2 (3-1-37)
l' l'
From the conditions  1  0.5 and  2  0.5 ,
l' l'
p1  , p2  (3-1-38)
6 6
l'
K-N Li (2004) calls these parameters, p1 , p 2 , as “plastic zones” and recommends to be p1  p 2  .
10
Them the reduction factors will be  1   2  0.7 .
89
e) Modification of stiffness degradation factor at the yield point
(The following modification of the stiffness degradation factor,  y , is suggested by Prof. Okano at Chiba
University.)
From Equations (3-1-32) and (3-1-34), the yield rotation of the member  y under anti-symmetric loading
condition, M A  M B  M y , is calculated as,
2  1M y  1 M  1 M
y     y      1 y (3-1-39)
  k   k
k0  y  0 y  0
where 1   2   .
The yield rotation  y in Equation (3-1-39) is different from the formula in Figure 3-1-10 since the factor
 is multiplied to only diagonal elements of flexural matrix in Equation (3-1-32).
The stiffness degradation factor is then obtained as,
1  1 
    1 (3-1-40)
 ' y   y 

To realize the designated value of stiffness degradation factor,  y should be modified as,
 1 
y  1  1   (3-1-41)
' 
 y 
For example, to realize the stiffness degradation factor  ' y  0.4, assuming   0.7 , the modified  y is
 1 
y 1   1  0.7   0.357
 0.4 
This modification is done automatically in STERA_3D.
90
f) Modification of considering rigid zone ratio
A beam-column connection can be idealized as a rigid zone. In case of a beam element, the default length
of the rigid zone is set to be a half of the column width, and the nonlinear bending spring of the beam
element is arranged at the position of the column face.
On the other hand, if elastic deformation of the connection is considered by reducing the length of rigid
zone, the position of the nonlinear bending spring will be inside the connection area. In this case, when the
nonlinear bending spring is yielding, the moment value at the position of the column face is smaller than
the yield moment.
Column Column
My
A B
My
dA l dB
A B
d A l' d B
My' My
Figure 3-1-11 Reduction of rigid zone and modification of yield moment
To make the moment at the column face to be the same as yield moment, the yield moment of the nonlinear
bending spring is increased as,
l / 2  (1   )d A
M y ' M y  M y
l/2
(3-1-42)
d
  1  2(1   ) A
l
For example, when l  540cm, d A  30cm,   0.75 ,
  1  2  (0.25)  30 / 540  1.027 (3-1-43)
91
3.1.2 Steel Beam
tf
H
tw
B : Width, H : Height, tw, tf : Thickness
Figure 3-1-12 Steel Beam Section
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)

BH 3  ( B  t w )( H  2t f ) 3
Iy  : along strong axis (3-1-46)
12
3
2t f B  ( H  2t f )t w
3
Iz  : along weak axis (3-1-47)

12
Moment of inertia for torsion
3 3
2 Bt f  ( H  2t f ) t w
J (3-1-48)
3
92
A A
x
MA MB
MA MB A B
A B B
B Nx
Nx
elastic element
l A
 A   A   A  MA MB
      A B
 B    B    B  B
     0 
 x  x  
Hysteresis model of a nonlinear bending spring is defined as the moment-rotation relationship under the
anti-symmetry loading as shown in Figure 3-1-14. The initial stiffness of the nonlinear spring is supposed
to be infinite, however, in numerical calculation, a large enough value is used for the stiffness.
 
M   M
A B
 
Moment distribution
M M M
My My
6 EI = +
k0 
l kp  
k0 k0
y   y 
93
Yield moment force

 1 
M y   Bt f ( H  t f )  t w ( H  2t f ) 2  y (3-1-49)
 4 
where,
y : Strength of steel (N/mm2)
y
My
Figure 3-1-15
Yield rotation
From Figure 3-1-14, the yield rotation of the nonlinear bending beam,  y , is supposed to be zero value,
however, in STERA_3D program, it is assumed as,
 y = 0.001  y (3-1-50)
where
6 EI
 y  M y / k0 , k0 
l
Hysteresis model
A bi-linear model is assumed for the hysteresis model.
My
y
Figure 3-1-16 Hysteresis of steel
94
d) Modification of initial stiffness of nonlinear springs
In numerical calculation, a large dummy value is used for the initial stiffness of the nonlinear spring to
represent rigid condition. This large stiffness may cause an error for estimating the force from the
displacement. One possible way to solve the problem is to reduce the initial stiffness of the nonlinear spring
to be a certain value reasonable for calculation, and on the other hand, increase the stiffness of the elastic
element so that the total initial stiffness of the beam element does not change from the original one. This
idea is proposed by K-N Li (2004) for MS model, and can be used for nonlinear spring model also.
 
M   M
A B
 
M M M
My My
6 EI = +
k0 
l kp  
k0 k0
y   y 
My My
y  =  + y  0
k0 k0
Increase Reduce
stiffness stiffness
M M
My
+
k0*  k0 /  1 k *p  EI / p1
*  y* *
My  1M y My
y  = *  +  y*  1   1 
k0 k0 k0
95
The idea is realized using flexibility reduction factors,  1  1,  2  1 , in the relationship between the
displacement vector and force vector of the elastic element in Equation (2-1-1) as,
 l' l' 
 1 3EI 
6 EI y
0 
 ' yA   y   M ' yA 
   l' l'  
 ' yB     2 0  M ' yB  (3-1-51)

  '   6 EI y 3EI y
 N ' 
 x
 0 l'  x 
0
 EA 
l' l' l' l'

It must be 1  or  1  0.5 and.  2  or  2  0.5 .
3EI y 6 EI y 3EI y 6 EI y
Also the initial flexibility matrix of the nonlinear spring can be expressed as follows, introducing the
parameters, p1 , p 2 to increase the initial flexibility.
 yA   p1 EI 0   M ' yA 
    (3-1-52)
 yB   0 p2 EI  M ' yB 
When p1  0, p 2  0 , it represents the infinite stiffness for rigid condition. Accordingly, the yield
rotation will be modified as,
My
 y*  p1 (3-1-53)
EI
My
    1 (3-1-54)
k0
 l' l'   p1 l' l' 

 3EI 
6 EI y
0   EI   1 3EI 
6 EI y
0 
 y   y 
 l'  l' 
0 
p2
 2 0  (3-1-55)
 3EI y  EI 3EI y
   
 sym. l'   l' 
sym.
 EA  original  EA  mod ified

3 3
1 1 p1 ,  2  1  p 2 (3-1-56)
l' l'
From the conditions  1  0.5 and  2  0.5 ,
l' l'
p1  , p2  (3-1-57)
6 6
l'
K-N Li (2004) calls these parameters, p1 , p 2 , as “plastic zones” and recommends to be p1  p 2  .
10
Then, the reduction factors will be  1   2  0.7 .
96
3.1.3 SRC Beam
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
Figure 3-1-18 SRC Beam Section
AN  BD  S  B t  nE  1a1  a2  a S  a ST ) (3-1-58)
where,
n E  E s / Ec : Ratio of Young’s modulus between steel (Es) and concrete (Ec)
aST  2b1  tw t f  h1 t w :Area of steel

AS  BD (3-1-59)
2 2
BD 3 ( S  B )t 3  D  
 BD g    S  B t  D   g  
t
Ie  
12 12  2  2 
2
nE  1a1 d1  g 2  nE  1a2 D  d 2  g 2  nE  1aS  D  t 

 g 
 2 
3
b1h1  (b1  t w )(h1  2t f ) 3
nE  1 (3-1-60)
12
97
where, g is the center of beam section calculated by
BD 2 / 2  ( S  B)t D  t / 2  nE  1a1d1  a2 D  d 2   aS D  t / 2   aST D / 2 
g
AN
(3-1-61)
Hysteresis model of a nonlinear bending spring is the same as RC beam.
Crack moment force
M c1  0.56  B Z e1 , Z e1  I e / g when tension in top main rebars (3-1-62)
M c 2  0.56  B Z e 2 , Z e 2  I e / D  g  when tension in bottom main rebars (3-1-63)
where,
B : Compression strength of concrete (N/mm2)
Z e1 , Z e 2 : Section modulus
Yield moment force

M y  M y1, 2, RC  M y ,S (3-1-64)
where
M y1, 2, RC : Yield moment of reinforced concrete (3-1-65)
M y1,RC  0.9a1 y D  d1   0.9aS  y D  t / 2 when tension in top main rebars
M y 2,RC  0.9a2 y D  d 2  when tension in bottom main rebars
where,
y : Strength of rebar (N/mm2)
 1 
M y ,S  b1t f (h1  t f )  t w (h1  2t f ) 2  y ,S : Yield moment of steel (3-1-66)
 4 
where,
 y,S : Strength of steel (N/mm2)
98
Appendix: Detail rule of degrading trilinear slip model
In OPTION menu in Beam Editor, you can control the shape of hysteresis loop.
R s : Effective Slab Ratio

As described in Eq. (3-1-8), when tension in slab side, the yield moment of beam , M y1 , is
M y1  0.9a 2 y D  d 2   0.9a s  y D  t / 2
where, a s is the area of rebars in effective width of slab, S b , which is defined as Eq.(3-1-17),
S b   s Ls
 s ( R s in the menu) is the effective slab ratio , the default value is 0.1.
Depending on the effective slab ratio R s , the yiled moment M y and the yield rotaion R y will change together
as shown in the Figure below, since the tangential stiffness at the yield point, K y , is assumed to be the
same.
Ru : Ultimate rotation angle to define the maximum moment before degradation. The default value is 1/50.
K p : The stiffness after the yield rotation angle, R y .
K u : The stiffness after the ultimate rotation angle, Ru .
It can be the negative value to consider strength degradation, however, the default value of the ratio K u / K y is
1/1000 without degradation.
99
M
kp
My
Effective slab ratio Rs
ku
(  s in Eq.(3-1-17))
Stiffness over Ru
(could be negative)
ky
R
Ry Ru
Ultimate rotation angle
(recommended over 1/50)
R1 : stiffness degrading ratio in the trilinear hysteresis is 0.5. (0: no degradation)

R 2 : slip stiffness ratio in the trilinear hysteresis is 0.0 (0: no slip).
R3 : strength degrading ratio in the trilinear hysteresis is 0.0.
Those parameters control the shape of hysteresis loop as descrived in Eqs. (3-1-18) and (3-1-19). That is,
M

Mm My  y
My kr    (  R1 )
 
 y  m
m n y ks kr  Mm  y

k s   
n
 (   R2 )
x y m  'm  m   x  m
 y 
 ' m  1    m
n
 (  R3 )
 n m 
More detail rule in the hysteresis loop is described in the following sections:
100
1. Elastic range
fy(1)
fc(1)
s0(1)
1
s0(2)
fc(2)
Initial stiffness, s0, is calculated from

2 s0(1) = fc(1) / dc(1)
fy(2)
2. From crack point to yield point
sy
4
sc
(drm(1), frm(1))
2
unloading point
3
s0
1
3
towards the maximum point of the other side
(drm(2), frm(2))
unloading point
101
3. Loading on the primary curve after yielding
(drm(1),frm(1))
fu 4
fy
fdm 5
ssd
xdm xd0
xd0 du xdm
(du,fu)
(drm,frm) (du,fu)
ssd fdm
su
5
ssd
4 xdm
(drm(2),frm(2))
If the displacement is less than the degrading point (du, fu),
the unloading point will the target point (drm, frm)
 fy  dy 
The stiffness of unloading. ssd, will be calculated from ssd    , where αis the parameter to control the stiffness
d  drm
 y 
degradation depending on the ductility factor, (drm/dy). The default value of is α = 0.5
(du,fu) (drm,frm)
fdm
If the displacement is over the degrading point,
su
intersection of the LINE1 (degrading line) ssd
and LINE2 (unloading line) will be the target
point. LINE2 LINE1
xdm
If the force is lower than 0.1fy, intersection of

the LINE1 (degrading line) and LINE2 (lower (drm,frm)
boundary) will be the target point.
ssd
LINE1
0.1fy
Lower boundary
LINE2
xdm
102
4. Crossing zero force line
(drm,frm)
4
fy
5
ssd
xd0
Target point of the other side, drm(2), will be increased

according to the ductility factor, (drm(1)/dy(1)), as follows:
 drm(1) 
drm(2)  1   drm(2)
 dy (1) 
where γ is the parameter to control the strength degradation.
The default value is γ=0.0 (no degradation).
(drm(2),frm(2)) (drm(2),frm(2))
srm
(drm,frm)
(dsl,fsl)
6 5
ssl
xd0
ssl
(dsl,fsl) 6
5
srm
(drm(2),frm(2))
103
5. Calculation of slip point
srm
(dsl,fsl)
5
xd0
(dsl,fsl) 6
ssl
srm
After crossing zero force line, the stiffness will be

calculated as:

frm(2) dy
(drm(2),frm(2)) ssl 
drm(2)  xd 0 drm
where β is the parameter to control the slip ratio. For the
default value (β=0.0), it will be no slip and towards the
maximum point.
LINE1 xd0
6
(dsl,fsl) ssl
LINE2
8
srm
The stiffness will change at the intersection of

LINE1 and LINE2. After the intersection point,
(dsl, fsl), Level 8 will be towards the maximum
(drm(2),frm(2)) point.
ssd ssd
7 xd0 srm
ssd 6
ssl
10
8
ssd srm Stiffness unloading from Level 6 (Level 7) and

stiffness unloading from Level 8 (Level 10) is the
same as the unloading stiffness, ssd ( Level 5).
(drm(2),frm(2))
5
104
dslp = s3*xd0/(s3-s2) 4
ssd
5
ssd
xd0
ssl 5
srm 6
11
9 ssd
5 8
7 xd0
6
After crossing zero force line from Lavel 7, Level 9 will be towards the
maximum point. Stiffness unloading from Level 9 (Level 11) is the same as
the stiffness of ssd (Level 5).
5
9
xd0
ssl 5
srm 6
11
5 8
7 xd0
6
105
3.2 Column
3.2.1 RC Column
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
Figure 3-2-1 RC Column Section
AN  BD  n E  1a1  a 2  a c  (3-2-1)
AS  BD /  ,   1.2 (3-2-2)
2
DB 3 B 
Iy   n E  1a c  a1   d1  (3-2-3)
12 2 
2
BD 3 D 
Ix   n E  1a c  a 2   d 2  (3-2-4)
12 2 
106
To consider nonlinear interaction among M x  M y  N z , the nonlinear bending spring at the member
end is constructed from the nonlinear vertical springs arranged in the member section as shown in Figure
3-2-2.
N zB ,  zB
M xB ,  xB z
M yB ,  yB
B y
x  z  i xi
x
yi
i
A
y
M yA ,  yA
M xA ,  xA
N zA ,  zA
Displacement of the i-th nonlinear axial spring is,
 i   z  y i x  xi y (3-2-5)
Equilibrium condition in the nonlinear section is,
M ' y   k i  i xi   k i ( z  yi x  xi y ) x i

i i
M ' x    k i  i yi    k i ( z  yi x  xi y ) y i (3-2-6)

i i
N ' z   k i  i   k i ( z  yi x  xi y )

i i
In a matrix form
 
 M ' y  
k i xi2   k i xi y i  k x   i i
  y 
  i 
y
x   k p  x 
 i i
  
 x  
M ' k y i
2
i   k y   i i (3-2-7)
 N'      
 k  
i i
 z   sym. i
z  z
 i
Therefore
107
 y  M ' y  M ' y 
x   k p   M ' x    f p  M ' x 
  1    
(3-2-8)
   N'   N' 
 z  z  z
For both ends
 yA   M ' yA 
  M ' 
 xA   xA 
 zA   f pA  0   N ' zA 
 
 yB   0  f pB  M ' yB  (3-2-9)
xB   M ' xB 
   
 zB   N ' zB 
Hysteresis model of nonlinear bending spring is defined as the moment-rotation relationship under the
anti-symmetry loading in Figure 3-2-3. The initial stiffness of the nonlinear spring is supposed to be infinite,
however, in numerical calculation, a large enough value is used for the stiffness.
 


 
A
Moment distribution
M
M M M
My My
6 EI = +
Mc k0  Mc
l
k0 k y   y k0 k0 kp  
c y   c  y 
108
ND
M c  0.56  B Z e  (3-2-10)
6
The yield moment, M y is calculated from the following formula under the balance axial force, N b ,
 Nb 
M y  0.8a t y D  0.5N b D 1   (3-2-11)
 bD B 
N b  0.4bD B (3-2-12)
Note that the balance axial force, N b , is used instead of actual axial force, N , in this formula since the
characteristics of nonlinear vertical springs in a section are determined later from the equilibrium condition
under the balance axial force.
The tangential stiffness at the yield point, k y , is obtained from the following equation,:
6 EI
k y  y K0 K0  (3-2-13)
l
where,
 y is the stiffness degradation factor at the yield point, which is obtained from the following
empirical formulas:
 y  0.043  1.63np t  0.043a / D  0.325 b d / D  2 , a / D  2 (3-2-14)
 y   0.0836  0.159a / D  0.169 b d / D 2 , a / D  2 (3-2-15)
where,

p t  a c  a1  / 2 BD  (when tension in x-main rebars)
p t  a c  a 2  / 2 BD  (when tension in y-main rebars)
a/D : ≈ Shear span-to-depth ratio (= l /(2 D) )
d : effective depth
d = D-d1 (when tension in bottom main rebars)
d = D-d2 (when tension in upper main rebars)
k y  y K0
2
 M N b  d  6 EI (3-2-16)
 y   0.043  1.63npt  0.043  0.33   , K0 
 QD bD B  D  l
 1 My
 y    1
K
(3-2-17)
 y  0
109
c) Nonlinear vertical springs
The nonlinear bending spring is constructed from the nonlinear vertical springs arranged in the member
section as shown in Figure 3-2-4. This model is called “Multi-spring model” proposed by S. S. Lai, G. T.
Will and S. Otani (1984) and modified by K-N. Li (1988). The section is devided in 5 areas; where 4 corner
areas have steel springs and concrete springs and the center area has one concrete spring.
The strength and the location of nonlinear springs are obtained from the equilibrium condition under
the balance axial force, N b , in Equation (3-2-8).
A1 1 2
A2
x x
5
3 4 Concrete spring
Steel spring
y y
(a) Original column section (b) Multi-spring model
(tension) (tension)
s fy
c d y
s dy
c f y
(compression) (compression)
(c) Hysteresis of steel spring (d) Hysteresis of concrete spring
Figure 3-2-4 Nonlinear vertical springs
Strength of steel spring

The strength of the steel spring is one-forth of total strength of rebars in the section, i.e.,
As y
s fy  (3-2-18)
4
where,
As : Total area of rebar in the section
y ： Strength of rebar
Strength of concrete spring

As shown in Figure 3-2-5, 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,
110
Nb
c f y1   0.2bD B (3-2-19)
2
Therefore, the area of the corner concrete, A1 , is,
fy
A1 
c
(3-2-20)
0.85 B 
 Nb
s f y  c f y1 My
s f y  c f y1
xs x
 s f y1
 s f y1
Figure 3-2-5 Equilibrium condition in the column section
The area of the center concrete, A2 , is the rest of the area of the section,
A2  bD  4 A1  0 (3-2-21)
The strength of the center concrete spring is then obtained as,
c f y 2  0.85k B A2 (3-2-22)
where, k is the confined effect k  1.3 of the concrete.
Location of vertical springs

The distance between the corner springs, x s , is obtained from the equilibrium condition regarding the
moment force in Figure 3-2-4,
M y  x s 2 s f y  c f y1   x s 2 s f y  0.5 N b  (3-2-23)
Therefore,
My
xs  (3-2-24)
2 s f y  0 .5 N b
Note that M y is calculated from Equation (3-2-7) for the balance axial force, N b .
111
Example)
To verify the efficiency of the Multi-Spring model for the column element, the M-N relationship is
compared between MS-model and Theory using one column element. The column section is shown in the
Figure below:
N=1000kN
300cm
50cm
Figure 3-2-6
Theoretical results of the M-N relationship are obtained from the equilibrium condition as,
if (0  N  N b )
 N 
M y  0.8at  y D  0.5 ND1   (3-2-25)
 bD B 
if ( N b  N  N max )
 N N 
 
M y  0.8at  y D  0.12bD 2 B  max  (3-2-26)
 N max  N b 
where, N b is the balance axial force,

N b  0.4bD B (3-2-27)
and N max is the maximum axial force,
N max  bD B  As y (3-2-28)
112
Firstly, the strengths and locations of vertical springs are calculated as
a t  15 .484 ( cm 2 )  y  1 .1 f y  32 .45 ( kN / cm 2 )  B  2 .4 ( kN / cm 2 )
N b  0.4bD B  2400 (kN ) N max  bD B  As y  6502 ( kN )
s f y  251.2 (kN ) c y1f  1200 (kN ) f

c y2  390 (kN ) x s  30 (cm)
In the range (0  N  N b ) , the Multi-Spring model gives

M y  2 s f y  0 .5 N x s 
M-N relationship
6000
Theory
5000 Multi-Spring
4000
N (kN)
3000
2000
Under-estimate
1000
0
0 10000 20000 30000 40000 50000 60000
M (kN*cm)
Figure 3-2-7 Comparison of M-N relationship
The results of Multi-Spring model give smaller values than theoretical results in the range 0 < N < Nb.
113
K-N. Li (1988) proposed to use the following formulation for deciding the location of vertical springs in
stead of Equation (3-2-24), as follows:
M y0
xs  (3-2-29)
2 s f y  0.5 N 0
where, N 0 is the axial force from the dead loads and the live loads acting on the column ( N 0  N b ),
and M y 0 is the yield moment under the axial force N 0 , that is:
 N0 
M y 0  0.8a t  y D  0.5 N 0 D1   (3-2-30)
 bD B 
For the example column, assuming N 0 = 1000 (kN),
x s  35.8 (cm)
M-N relationship
6000
Theory
5000 Multi-Spring
4000
N (kN)
3000
2000
1000
0
0 20000 40000 60000 80000
M (kN*cm)
Figure 3-2-8 Comparison of M-N relationship
It improves the results of Multi-Spring model in the range 0 < N < Nb.
114
Yield displacement of vertical spring
 N0
 2 fc M y0 Compression
y s dy c f y sf y
dc fc
2 s f y
xs dc c d y s d y
Figure 3-2-9 Equilibrium condition under the axial force N0
From the equilibrium condition under the axial force N 0 as shown in the above Figure, the yield
displacement of the tension side steel spring, s d y , is obtained as follows:
s dy  d c   y xs
fc
dc  sdy
s f y c f y
N 0 2 s f y
fc  (3-2-31)
2
 y xs
sdy 
N 0 2 s f y
1
2 s f y 2c f y
The yield displacement of concrete spring, c d y , is assumed to be the same as that of the steel spring,
c d y s d y (3-2-32)
115
d) Nonlinear shear spring
There are two nonlinear shear springs in x and y directions. Hysteresis model of the nonlinear shear springs
is the same as that in the beam element in Figure 3-1-4.
Yield shear force
 0.053 pt 0.23 ( B  18) 

Qy    0.85 p w   wy  0.1 0 b  j (3-2-33)
 M /(QD)  0.12 
where,

M/(QD) : ≈ Shear span-to-depth ratio (= l /(2 D) )
0 : Axial stress of the column
Crack shear force
Qy
Qc  (3-2-34)
3
The crack shear force is, Qu , is assumed as,
Qu  Qc (3-2-35)

Qc
c  (3-2-36)
GA
1
y  (3-2-37)
250
1
u  (3-2-38)
100
116
Qx
Qy
 Qx y
 Qy
Figure 3-2-10 Nonlinear shear springs in column
117
e) Modification of initial stiffness of nonlinear springs
The same modification can be done for the nonlinear springs of column element as described for those of
beam element by reducing the initial stiffness of the nonlinear spring and increasing the stiffness of the
elastic element as shown in the following figure:
 


 
A
Moment distribution
M
M M M
My My
6 EI = +
Mc k0  Mc
l
k 0  y k0 k0 kp  
c y   c  y 
Increase Reduce
stiffness stiffness
M M
My
+ Mc
k *
0
k *p
 c*  y* 
118
Introducint the concept of “plastic zones”, the initial stiffness of the i-th multi-spring can be expressed as,
Ei Ai
k 0i  (3-2-39)
pz
where Ei : the material young’s modulus, Ai : the spring governed area, and pz : the length of assumed
plastic zone. When p z  0 , it represents the infinite stiffness for rigid condition.
From Equation (3-2-7), when we consider the flexural flexibility in x-z plane, the flexibility matrix for the
nonlinear MS section is,
1
 y   k i
0 xi2 0   pz
 M ' y   
E A x i i
2
i 0
 M ' y 
      (3-2-40)
i i
 z  0 1  k 0i   N ' z   0 p z  Ei A  N ' z 
 i   i 
Also, introducing the flexibility reduction factors,  0  0,  1  0,  2  0 , the flexibility matrix of
 l' l' 
 1 3EI 
6 EI y 
 y

' l' 
 f C     l
2

(3-2-41)
6 EI y 3EI y
 
 l' 
0
 EA 
 
 p z1 l' l' 
 l' l'   1  0
 E A x2 

 3EI  0  3 EI 6 EI y
6 EI y i i i y

 y
 
i

l' l'
 0 
pz2
 2 0 


3EI y 

  i i i
E A x 2
3 EI y 
l'   i 
 sym.  p z1 pz2 l' 
 EA  original  sym.  0
 i Ei A i Ei A EA 
mod ified
(3-2-42)

3 3 1
1  1 p z 1 ,  2  1  p z 2 ,  0  1   p z1  p z 2  (3-2-43)
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 (3-2-44)
119
f) Tri-linear hysteresis for nonlinear springs
The original hysteresis models used for steel and concrete springs are bi-linear types as shown in Figure
3-2-3. To control both the initial stiffness and yield displacement, it is convenient to define tri-linear type
hysteresis.
For the steel spring, the maximum-oriented model is adopted for the hysteresis before yielding, and the
tri-linear model is adopted after yielding as shown in Figure 3-2-12.
 sf y
 sf y
(a) before yielding point (b) after yielding point
s fy
 sf y
E i Ai
k 0i  s
pz
s d *y
Figure 3-2-12 Normal tri-linear model for steel spring
The hysteresis of steel spring has the degradation point at the forces,  sf y and  s f y , where  and 
are the arbitrary parameters   1,   1 . The STERA_3D Program adopts the values as:
  1 / 3,   0.5 (3-2-45)
*
Then, the yield deformation, s d y, may be obtained by Equations (3-2-24) and (3-2-10) considering the
reduction factor .
 y* xs
sdy  (3-2-46)
N 0 2 s f y
1
2 s f y 2c f y
 1 1My
 y*     (3-2-47)
  y   k0
 
120
The hysteresis of concrete spring is also defined as tri-linear hysteresis model as shown in Figure 3-2-13.
After compression yielding, strength degradation is considered by reducing the strength of the target point
in reloading stage.
dc
fc
fy
(a) hysteresis rule after compression crack point
dc
fc
fy
dm dc
fc
fy
(b) hysteresis rule after compression yield point
 dp 
dm'  dm1  r 
 dy  dp
dm’ dm dc
fc
fy
(c) strength degradation rule
Figure 3-2-13 Tri-linear hysteresis model for concrete spring
121
References
1) S. S. Lai, G. T. Will, and S. Otani (1984), “Model for Inelastic Biaxial Bending of Concrete Members,”
Journal of Structural Division, ASCE, Vol. 110, ST1, 1984, pp.2563-2584.
2) K-N. Li (1988), “Nonlinear Earthquake Response of Reinforced Concrete Space Frames,” the
dissertation for the degree of Doctor in University of Tokyo (in Japanese), 1988.12.
3) K-N. Li (2004), CANNY, Technical Manual.
122
3.2.2 Steel Column
tf t2
H t1
tw t1
B B D
B : Width, H : Height, tw, tf, t1, t2, t: Thickness
Figure 3-2-14 Steel Column Section
AN  total area of section (3-2-48)

AS  ( ) (3-2-49)
AS  0.5 AN
Figure 3-2-15 Area of section for shear

1) H section
BH 3  ( B  t w )( H  2t f ) 3
I : along strong axis (3-2-50)
12
3
2t f B  ( H  2t f )t w
3
I : along weak axis (3-2-51)

12
2) Box section
BH 3  ( B  2t1 )( H  2t 2 ) 3
I (3-2-52)
12
123
3) Circle section
I

64
D 4
 D  2t 
4
 (3-2-53)
Moment of inertia for torsion

1) H section
3 3
2 Bt f  ( H  2t f ) t w
J (3-2-54)
3
4) Box section
2t1t 2 ( B  t1 ) 2 ( H  t 2 ) 2
J 2 2
(3-2-55)
BHt1t 2  t1  t1
5) Circle section
J

32
D 4
 D  t 
4
 (3-2-56)
124
3-2-16.
N zB ,  zB
M xB ,  xB z
M yB ,  yB
B y
x  z  i xi
x
yi
i
A
y
M yA ,  yA
M xA ,  xA
N zA ,  zA
 i   z  y i x  xi y (3-2-57)
Equilibrium condition in the nonlinear section is,
M ' y   k i  i xi   k i ( z  yi x  xi y ) x i

i i
M ' x    k i  i yi    k i ( z  yi x  xi y ) y i (3-2-58)

i i
N ' z   k i  i   k i ( z  yi x  xi y )

i i
In a matrix form
 
 M ' y  
k i xi2   k i xi y i  k x   i i
  y 
  i 
y
x   k p  x 
 i i
  
M 'x    k y i
2
i   k y   i i (3-2-59)
 N'      
 k  
i i
 z   sym. i
z  z
 i
Therefore
125
 y  M ' y  M ' y 
x   k p   M ' x    f p  M ' x 
  1    
(3-2-60)
   N'   N' 
 z  z  z
For both ends
 yA   M ' yA 
  M ' 
 xA   xA 
 zA   f pA  0   N ' zA 
 
 yB   0  f pB  M ' yB  (3-2-61)
xB   M ' xB 
   
 zB   N ' zB 
Hysteresis model of nonlinear bending spring is defined as the moment-rotation relationship under the
anti-symmetry loading in Figure 3-2-17. The initial stiffness of the nonlinear spring is supposed to be
infinite, however, in numerical calculation, a large enough value is used for the stiffness.
 


 
A
Moment distribution
M
M M M
My My
6 EI = +
k0 
l kp  
k0 k0
y   y 
126
Yield moment force (full plastic moment)
1) I shape
tf
H
tw
a) When the neutral axis is inside the web, i.e., N < Aw y = t w ( H - 2t f ) y
M y  M y 0  y 0 t w y
2
(3-2-62) Bt f  y
1
t w ( H  2t f ) y
2
where
 1  B
M y 0   Bt f ( H  t f )  t w ( H  2t f ) 2  y
 4 
N M y0
y0 
2t w y H
b) When the neutral axis is inside the flange, i.e., N  Aw y  t w ( H - 2t f ) y
H  H 
M y  B  y 0   y 0  y (3-2-63) H 
2  2  B  y 0  y
where
2 
1  N  N y 
y0  H
2  B y 


N y  2 Bt f  t w H  2t f   y
My
127
2) H shape
a) When the neutral axis is inside the web, i.e., N  Aw y  t w H y
1
M y  M y 0  y 0 H y
2
(3-2-64) Bt f  y
2
1
where t w ( H  2t f ) y
2
1 1 2 
M y 0   B 2 t f  t w ( H  2t f ) y
2 4  H
M y0
N
y0 
2 H y
B
b) When the neutral axis is inside the web, i.e., N  Aw y  t w H y
B  B  1 
M y  2t f   y 0   y 0  y (3-2-65) t f  B  y 0  y
2  2   2 
where
1  N  N y 
y0   B
2  2t f  y 

H
My
 
N y  2 Bt f  t w H  2t f   y
3) Box shape
1
t w ( H  2t f ) y
t2 Bt f  y 2
H M y0
t1
B
B H
a) Moment around x-axis
M y  M y ( I shape by changing t w  2t1 , t f  t 2 ) (3-2-66)
b) Moment around y-axis
M y  M y ( I shape by changing t w  2t 2 , t f  t1 , B  H ) (3-2-67)
128
4) Circle shape
t1
 N 
M y  M y 0 cos  (3-2-68)
 2N 
 y 
where
M y 0  D  t1  t1 y
2
N y   D  t1 t1 y
Yield rotation
The yield rotation is
6 EI
 y  M y / k0 , k0  (3-2-69)
l
129
section as shown in Figure 3-2-18. This model is called “fiber model”. The section is devided in several
areas which have steel springs.
30°
(H-2tf)/5 tf t2
H t1
tw t1
B B
B : Width, H : Height, tw, tf, t1, t2, t: Thickness
(tension)
f y ,i
d y ,i
(compression)
Hysteresis of steel spring
Strength of steel spring

The strength of the i-th steel spring is,
f y ,i  Ai y (3-2-70)
where, Ai : the spring governed area,  y : the strength of steel
Yield displacement of steel spring

The yield displacement of the i-th steel spring is,
d y ,i  f y ,i / k 0i , k 0i  E s Ai (3-2-71)
where Es : the young’s modulus of steel
130
The same modification can be done for the nonlinear springs of column element as described for those of
beam element by reducing the initial stiffness of the nonlinear spring and increasing the stiffness of the
elastic element as shown in the following figure:
 


 
A
Moment distribution
M
M M M
My
6 EI = +
k0 
l kp  
k0 k0
y   y 
Increase Reduce
stiffness stiffness
M M
My
+
k 0*
  y* 
Ei Ai
k 0i  (3-2-72)
pz
131
From Equation (3-2-7), when we consider the flexural flexibility in x-z plane, the flexibility matrix for the
nonlinear MS section is,
1
 y   k i
0 xi2 0   pz
 M ' y   
E A x i i
2
i 0
 M ' y 
      (3-2-73)
i i
 z  0 1  k 0i   N ' z   0 p z  Ei A  N ' z 
 i   i 
Also, introducing the flexibility reduction factors,  0  0,  1  0,  2  0 , the flexibility matrix of
 l' l' 
 1 3EI 
6 EI y 
 y

' l' 
 f C     l
2

(3-2-74)
6 EI y 3EI y
 
 l' 
0
 EA 
 
 p z1 l' l' 
 l' l'   1  0
 E A x2 

 3EI  0  3 EI 6 EI y
6 EI y i i i y

 y
 
i

l' l'
 0 
pz2
 2 0 


3EI y 

  i i i
E A x 2
3 EI y 
l'   i 
 sym.  p z1 pz2 l' 
 EA  original  sym.  0
 i Ei A i Ei A EA 
mod ified
(3-2-75)

3 3 1
1  1 p z 1 ,  2  1  p z 2 ,  0  1   p z1  p z 2  (3-2-76)
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 (3-2-77)
132
3.2.2 Column with direct input
X-Z plane Y-Z plane
M ' xB M ' xB
M ' yB
M ' yB
B B B
 ' xB
 ' yB
N 'z Z
l' = +
T 'z
 ' yA  ' xA X
A A A
M ' yA Y
M ' yA
M ' xA M ' xA
Figure 3-2-20 Element model for column
In case of direct input for Moment-Rotation relationship, we neglect nonlinear interaction among
M x  M y  N z and define the flexural stiffness of nonlinear bending springs in X and Y directions
independently. The rotational displacement vector of the nonlinear bending springs in Equation (3-2-9) will
be
 yA   M ' yA   f yA   M ' yA 
  M '   f xA M ' 
 xA   xA     xA 
 zA   f pA  0   N ' zA   0   N ' zA 
  
 yB   0   
f pB  M ' yB   f yB

 M ' yB 
 (3-2-78)
 xB   M ' xB   f xB   M ' xB 
      
 zB   N ' zB   0  N ' zB 
 ' yA   ' yA   yA   yA   M ' yA 

 '   '      M ' 
 yB   yB   yB   yB   yB 
 'xA   ' xA   xA   xA   M ' 
         f C  xA  (3-2-79)
 'xB   ' xB   xB   xB   M ' xB 
  ' z   ' ' z  z   0   N 'z 
         
  'z    'z elastic element  0 bending spring  0  shear spring  T ' z 
133
The flexural matrix [ f C ] is;
 l' 1 l' 1 
 f yA     
3EI y k sx l ' 6 EI y k sx l '
 
 l' 1 
f xA   0
 3EI y k sx l ' 
 
 l' 1 l' 1 
f yB    
 f C    3EI x k sy l ' 6 EI x k sy l ' 
l' 1 
 f xB   
 3EI x k sy l ' 
 l' 
 
 EA 
 l' 
sym.
 GI z 
(3-2-80)
134
3.2.3 SRC Column
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
Figure 3-2-21 RC Column Section
AN  BD  nE  1a1  a2  ac  aST  (3-2-81)
where,
aST  n f b1  t w  t f  nw h1 t w :Area of steel
n f  2, nw  1 : Type1, Type2,
n f  4, nw  2 : Type3
n f  3, nw  2 : Type4, Type5
AS  BD /  ,   1.2 (3-2-82)
135
DB 3  B 
2

Iy   n E  1 a c  a1   d 1   I S , y  (3-2-83)
12  2  
 
BD 3  D 
2

Ix   n E  1 a c  a 2   d 2   I S , x 
 (3-2-84)
12  2  
 
where
IS : Moment of inertia of steel
I S,x I S,y
II 
1
12

b1h1  b1  t w h1  2t f 
3 3
 IH 
1
12
 3
2t f b1  (h1  2t f )t w
3

Type 1
IH II
Type 2
II  IH II  IH
Type 3
2
II  IH h 
I I  I H  AH  1 
Type 4 2
2
h  II  IH
I I  I H  AH  1 
Type 5 2

Hysteresis model of a nonlinear bending spring is the same as RC beam.
Crack moment force

ND
M c  0.56  B Z e  (3-2-85)
6
136
Yield moment force
M y  M y , RC  M y , S (3-2-86)
where
M y , RC : Yield moment of reinforced concrete
 Nb 
M y , RC  0.8a t  y D  0.5 N b D1   (3-2-87)
 bD B 
M y,S : Yield moment of steel
M y,S , x M y,S , y
M yI M yH
Type 1
M yH M yI
Type 2
M yI  M yH M yI  M yH
Type 3
M yI  M yH M yI  M yT
Type 4
M yI  M yT M yI  M yH
Type 5
 1 
M yI  b1t f (h1  t f )  t w (h1  2t f ) 2  y ,S
 4 
1 2 1 2 
M yH   b1 t f  t w (h1  2t f ) y , S
2 4 
 1 
M yT  b1t f (h1  t f )  t w (h1  t f ) 2  y , S
 2 
b1 M yH h1 b1
M yI M yT
h1 b1 h1
137
3.3 Wall
3.3.1 RC Wall
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
Figure 3-3-1 Wall Section
AN  AN ,C1  AN ,C 2  t l w 2  n E  1a w  (3-3-1)
where,
AN ,C1 , AN ,C 2 : Area of section of side columns for axial deformation

AS  AS ,C1  AS ,C 2  t l w 2 /  ,   1.2 (3-3-2)
where,
AS ,C1 , AS ,C 2 : Area of section of side columns for shear deformation

2 2
l t3 l  l 
I y  I y ,C 1  I y ,C 2  w 2  AN ,C1  w1   AN ,C 2  w1  (3-3-3)
12  2   2 
where,
I y ,C 1 , I y ,C 2 : Moment of inertia of side columns
138
3-3-2.
z
x 2
x1  yc
 zc i
yi x
xi
 i   zc  xi yc in a wall panel

 i   zc  yi x1  xi yc in a side column 1 (3-3-4)
 i   zc  yi x 2  xi yc in a side column 2
N ' zc ,  zc
M ' yc ,  yc
Figure 3-3-3 Equilibrium condition in the wall panel direction
139
In the wall panel direction, all vertical springs in the nonlinear section are assumed to work against the
moment and the axial force. The equilibrium conditions are,
Nc N1 N2
M ' yc   k i  i xi   k i  i xi   k i  i xi
i i i
Nc N1 N2
  k i ( zc  xi yc ) x i   k i ( zc  y i x1  xi yc ) xi   k i ( zc  y i x 2  xi yc ) xi
i i i
 yc 
 
 Nc  N 1 N 2 N1 N2 Nc  N 1 N 2
  x1 
   k i xi2   k i xi y i   k i xi y i i k i xi   
 i i i  x2 
 zc 
(3-3-5)
Nc N1 N2
N ' zc  k  i
i i   k i i   k i i
i i
Nc N1 N2
  k (
i
i zc  x i  yc )   k i ( zc  y i  x1  x i  yc )   k i ( zc  y i  x 2  x i  yc )
i i
 yc 
Nc  N 1 N 2   
 Nc  N 1 N 2 N1 N2
  x1 
   k i xi   k i yi   ki yi i k i   
 i i i  x2 
 zc 
(3-3-6)
where, Nc, N1 and N2 are the number of vertical springs in a wall panel, side column 1 and side column 2,
respectively.
M ' x1 ,  x1 M ' x2 , x2
side column 1 side column 2
Figure 3-3-4 Equilibrium condition in the out of wall direction
140
In the out of wall direction, we establish the equilibrium condition for each side column independently. The
equilibrium condition for the side column 1 is,
N1
M ' x1   k i  i y i
i
N1
  k i ( zc  y i x1  xi yc ) y i
i
 yc 
 
 N1 N1 N1
  x1 
   k i xi y i k y 2
0   k i yi   
  x 2 
i i
 i i i
 zc 
(3-3-7)
Also, for the side column 2,
N2
M ' x 2   k i  i y i
i
N2
  k i ( zc  y i x1  xi yc ) y i
i
 yc 
 
 N2 N2 N2
  x1 
   k i xi y i 0 k y 2
  k i yi   
  x 2 
i i
 i i i
 zc 
(3-3-8)
In a matrix form
 Nc  N 1 N 2 2 N1 N2 Nc  N 1 N 2

  k i xi   k i xi y i   k i xi y i k x  i i
 M ' yc   N 1
i i
N1
i i
N1     yc 
 M '    k i xi y i k y   k i yi   
2 yc
0  
 x1   i   x1 
   x1 
i i
    N2     k p  
i i
N2 N2
M ' x 2     k i xi y i 0 k y 2
  k i yi   x 2   x2 
 N ' zc   i    zc 
i i
i i   zc 
 Nc  N 1 N 2 N1 N2 Nc  N 1 N 2

  k i xi   k i yi   k i yi i k i 
 i i i
(3-3-9)
Therefore
 yc   M ' yc   M ' yc 
  M '  M ' 
 x1 
   kp   1 

x1 
  f p   
 x1 
 (3-3-10)
 x 2  M '
 x2  M ' x 2 
 zc   N ' zc   N ' zc 
For both ends
141
 yAc   M ' yAc 
  M ' 
 xA1   xA1 
 xA2   M ' xA2 
 
 
 zAc   f pA

0   N ' zAc 



 yBc   0    
f pB   M ' yBc 
(3-3-11)
 xB1   M ' xB1 

   
 xB 2  M ' xB 2 
   N' 
 zBc   zBc 
For the out of wall direction, each side columns behave independently in the same way as the column
element. Therefore, we discuss here only the hysteresis model in the wall panel direction. Hysteresis model
of nonlinear bending spring is defined as the moment-rotation relationship under the symmetry loading in
Figure 3-3-5. The initial stiffness of the nonlinear spring is supposed to be infinite, however, in numerical
calculation, a large enough value is used for the stiffness.
M
 


 
A
Moment distribution
M
M M M
My My
2 EI = +
Mc k0  Mc
l
k0 k y   y k0 k0 kp  
c y   c  y 
142
The yield moment, M y is obtained from the equilibrium condition in Figure 3-3-6 as,
M y  a s y l w  0.5a w wy l w  0.5Nl w (3-3-12)
where,
as : Total area of rebar in the side column

y : Strength of rebar in the side column
aw : Total area of vertical rebar in the wall panel
 wy : Strength of rebar in the wall panel
N : Axial load from the dead load
a s y
My
a w wy
lw
Figure 3-3-6 Equilibrium condition under yielding moment
The crack moment, M c is assumed to be,
M c  0.3M y (3-3-13)
The tangential stiffness at the yield point, k y , is obtained from the following equation:
k y  0.2 K 0 (3-3-14)
 1 My
 y    1
K
(3-3-15)
 y  0
where, the stiffness degradation factor,  y , is assumed as,
 y  0.02 (3-3-16)
143
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
(a) Original column section

1 2 6 7
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)
(c) Hysteresis of steel spring (d) Hysteresis of concrete spring
144
Strength of steel spring in wall panel
The strength of the steel spring in the wall panel is one-fifth of total strength of rebars in the section,
a w wy
s fy  (3-3-17)
5
where,
aw : Total area of vertical rebar in the wall panel
 wy : Strength of rebar in the wall panel
Strength of concrete spring in wall panel
The strength of the concrete spring in the wall panel is one-fifth of total strength of concrete in the section,
0.85 A p B
c fy  (3-3-18)
5
where,
Ap : Total area of wall panel section
B ： Compression strength of concrete
Yield displacement of vertical spring in wall panel

The yield displacements of steel and concrete springs in the wall panel are assumed to be the same as those
of the springs in the side columns.
d) Nonlinear shear spring

There are three nonlinear shear springs in x direction in wall panel and y direction in side columns.
Hysteresis model of the nonlinear shear springs is the same as that in the beam element in Figure 3-1-4.
Yield shear force
 0.053 pt 0.23 ( B  18) 

Qy    0.85 p w   wy  0.1 0 b  j (3-3-19)
 M /(QD)  0.12 
where,

M/(QD) : ≈ Shear span-to-depth ratio (= l /( 2 D ) )
0 : Axial stress of the column
145
Crack shear force
Qy
Qc  (3-3-20)
3
The crack shear force is, Qu , is assumed as,
Qu  Qc (3-3-21)

Qc
c  (3-3-22)
GA
1
y  (3-3-23)
250
1
u  (3-3-24)
100
Q y1 Q xc Qy2
 Q y1  Q xc  Qy2
Figure 3-3-8 Nonlinear shear springs in the wall
146
e) Modification of initial stiffness of nonlinear springs
The same modification can be done for the nonlinear springs of wall element as described for those of
beam and column elements by reducing the initial stiffness of the nonlinear spring and increasing the
stiffness of the elastic element as shown in the following figure:
M
 


 
A
Moment distribution
M
M M M
My My
2 EI = +
Mc k0  Mc
l
k0 k y   y k0 k0 kp  
c y   c  y 
Increase Reduce
stiffness stiffness
M M
My
+ Mc
k *
0
k *p
 c*  y* 
147
Ei Ai
k 0i  (3-3-25)
pz
In the same manner of beam and column elements, introducing the flexibility reduction factors,
 0  0,  1  0,  2  0 , the flexibility matrix of the elastic element is,
 l' l' 
 1 3EI 
6 EI c 
 c

 l' 
2
 3EI c 
 l' l' 
 1  
 3EI 1 6 EI 1 
 l' 
 fW    2  (3-3-26)
 3EI 1 
 l' l' 
1 
 3EI 2 6 EI 2 
 l' 
 sym. 2 
 3EI 2 
 l' 
 0 
 EAc 
l'
Also, adopting p z  as discussed for beam and column elements, the reduction factors will be:
10
 1   2  0.7 ,  0  0.8 (3-3-27)
148
3.3.2 Steel Wall (Brace)
a) Axial strength
3 B 4
N',  '
Z
h
1 2 X
A
w Y
Figure 3-4-1 Element model for brace

The axial strength of steel brace is calculated from the following equation.

N c  AN 1  0.4    y ,
2
 for  (3-3-28)
0.6 y
N c  AN , for  (3-3-29)
  2
where
AN : Area of section for axial deformation

y: Strength of steel (N/mm2)
: Slenderness ratio
Le Le
  ,
i I/A
where, Le : effective bucking length = L (in case of pin ends)
: Critical slenderness ratio
 2E

0.6 y Nc
b) Hysteresis model
A bi-linear model is assumed for the hysteresis model.
dy
Figure 3-4-2 Hysteresis of steel
149
3.3.3 SRC Wall (Brace)
b) Nonlinear shear spring
Yield shear force
Qy  Qy , RC  Qy , S (3-3-30)
where
Qy , RC : Yield shear force of reinforced concrete
 0.053 pt 0.23 ( B  18) 
Qy , RC   0.85 pw   wy  0.1 0 b  j (3-2-31)
 M /(QD)  0.12 
Qy , S : Yield shear force of steel
Qy , S  AS  y , S cos R (3-2-32)
where,
AS : Area of steel (mm2)
 y, S : Strength of steel (N/mm2)
R : Angle of steel
150
3.4 External Spring
N 'z ,  'z
Figure 3-5-1 Element model for external spring
a) Nonlinear vertical spring
In STERA_3D, the external spring is attached at the base of the building to express the stiffness and
strength of the foundation of the building. In such a case, hysteresis model of the nonlinear vertical spring
is defined as the axial force – displacement relationship as shown in Figure 3-5-2; where, bilinear skeleton
is defined only in compression side, and the spring has zero stiffness in the tension side assuming that the
building detaches from the ground.
N 'z
(tension)
e d y
 'z
e f y
(compression)
Figure 3-4-2 Hysteresis model of the external spring
Initial stiffness
The initial stiffness of the vertical stiffness can be obtained from the following equation:
k e  a F AF (3-4-1)
where,
aF : Dynamic ground coefficient (kN/m2)

AF : Area of foundation under column or wall element (m2)
151
3.5 Base Isolation
The element model of base isolation consists of shear springs arranged in x-y plane changing its direction
with equal angle interval as shown in Figure 3-5-1. This model is called MSS (Multi-Shear Spring) model
developed by Wada et al.
x x qi , u i
y i x
Figure 3-5-1 Element model of base isolation
a) Nonlinear shear spring
The hysteresis model of each nonlinear shear spring is defined as a bi-linear model as shown in Figure
3-5-2. The force and displacement vectors of i-th shear spring are expressed as,
qi , x  cos  i 
   qi (3-5-1)
q
  
i , y sin  i 
u x 
u i  cos  isin  i   (3-5-2)
u y 
From the relationship, qi  k i u i , the constitutive equation of i-th shear spring is,
qi , x  cos  i  u x   cos 2  i cos  i sin  i  u x 

   ki  cos  i sin  i       (3-5-3)
q i , y   sin  i  u y  cos  i sin  i sin 2  i  u y 
qi
fy ky qi , y
qi
qi , x
k0
ui i x
dy
Figure 3-5-2 Hysteresis model of the shear spring
152
From the sum of all nonlinear shear springs in the element, the constitutive equation of the base isolation
element is,
Q x   N  cos 2  i cos  i sin  i  u x 

     k i     (3-5-4)
Q y   i 1 cos  i sin  i sin 2  i  u y 
where, N is the number of shear springs in an element. In STERA_3D, N=6 is selected.
First and second stiffness
We assume that all nonlinear shear springs in an element have the same stiffness and strength. The initial
stiffness of the base isolation element, K 0 , is obtained from Equation (3-5-4) by substituting
u x  1, u y  0 .
 N 
K 0    cos 2  i k 0 (3-5-5)
 i 1 
Therefore, the initial stiffness of each shear spring is,
K0
k0  N
(3-5-6)
 cos
i 1
2
i
The same relationship is established for the second stiffness after yielding,
Ky
ky  N
(3-5-7)
 cos
i 1
2
i
where, K y and k y are the second stiffness after yielding for the base isolation element and the
nonlinear shear spring, respectively.
Yield shear force

The yield shear force of the base isolation element, Q y , is obtained assuming that all the nonlinear shear
springs reach their yielding points except the spring perpendicular to the loading direction, and the increase
of the force after yielding is negligible (Figure 3-5-3). That is,
 N 
Q y    cos  i  fy (3-5-8)
 i 1 
Therefore, the yield shear force of each shear spring is,
Qy
fy  N
(3-5-9)
 cos
i 1
i
153
fy
fy
Qy i fy
fy
fy
Figure 3-5-3 Assumption of yield shear force
154
Appendix :
A-1. Hysteresis of LRB (Lead Rubber Bearing)
LRB (Lead Rubber Bearing) is composed by rubber layers, steel plates and a lead plug core.
Figure A1-1. Lead Rubber Bearing (from Bridgestone Catalog)
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.
ELASTIC PLASTIC ELASTO – PLASTIC
K2
Fy
Qd
K1
+ Kp = Keq
Qo Kr
Dy Dy Dy
Qd
RUBBER LEAD
Figure A1-2. Bi-linear model
The elastic stiffness, Kr, from the rubber is calculated as,

Ar
K r  Gr (A1-1)
Hr
155
where Gr is the shear modulus of the rubber, Ar is the cross section area of the rubber and Hr is the total height of
the rubber.
The elastic stiffness, Kp, from the lead plug is calculated as,
Ap
K p  Gp (A1-2)
Hp
where Gp is the shear modulus of lead, Ap is the cross section area of lead plug and Hp is the total height of the plug.
The initial elastic stiffness, K1, and the secondary stiffness, K2, of the bi-linear model are then obtained as,
K1  K r  K p
(A1-3)
K2  Kr
The yielding deformation, Dy, is determined from the characteristics of the lead plug. The yielding force, Fy, is
calculated as,
F y  K r  K p D y (A1-4)
2) Modified Bi-linear Model
Hysteresis of a lead rubber bearing has a characteristic of stiffness degrading according to the strain level as shown
in Figure A1-3.
Force, F Skelton curve
Deformation, δ
Figure A1-3. Hysteresis of a lead rubber bearing
The secondary stiffness of a lead rubber bearing, Kd, is expressed as,
K d    C Kd  K r  K p  (A1-5)
where  is a strain ratio (    / H r ) and C Kd   is a modification factor of the secondary stiffness, which
takes into consideration the strain dependency. Also, the intercept force is defined as,
156
Qd    C Qd   p A p (A1-6)
where C Qd   is a modification factor of the yielding shear force and  p is the yielding shear stress of lead.
The force is then expressed by:
F    K d     Qd   (A1-7)
Figure A1-4 Hysteresis loop model of lead rubber bearing
The modification factors, C Kd   and C Qd   , are represented by the following formulas under 15 degrees
Celsius.
0.779 0.43 ,   0.25


C Kd      0.25 , 0.25    1.0 (A1-8)
   0 .12
, 1.0    2.5

 2.036 0.41 ,   0.1


C Qd    1.106 0.145 , 0.1    0.5 (A1-9)
 1 , 0.5  

Under a different temperature, t, the secondary stiffness and the yielding shear force are to be corrected by the
following formulas:
K d t   K d t 0  exp 0.00271 t  t 0  (A1-10)

Qd t   Qd t 0  exp 0.00879 t  t 0  (A1-11)
where t0 = 15 degree Celsius.
The primary stiffness of the lead rubber bearing, Ku, in Figure A1-4 is determined from the secondary stiffness, Kd,
as,
Ku   Kd (A1-12)
where 10    15 .
157
Following the suggestion in the manual of CANNY (K. Li, 2004), the hysteresis rules are:
a) Elastic range
Under the strain level less than γe, the hysteresis is assumed to be linear with the secant stiffness at the strain, γe,
that is:
K 0  Fe /  e (A1-13)
Fe  K d  e   e  Qd  e ,  e   e H r (A1-14)
The value,  e  0.1 , is adopted in STERA3D.
b) Loading on the skeleton curve after elastic range
Under the loading on the skeleton curve after elastic range, tangent stiffness is used to estimate the response at
the next step:
K  dF ( ) / d (A1-15)
Reference:
Response Control and Seismic Isolation of Buildings, Edited by Masahiko Higashino and Shin Okamoto, SPON
PRESS, October 17, 2006.
Canny Technical Manual, Kangning Li, August 2004
158
A-2. Hysteresis of HDRB (High Damping Rubber Bearing)
HDRB (High Damping Rubber Bearing) is composed by rubber layers and steel plates. By adding special
ingredient in the natural rubber, rubber itself demonstrates damping characteristics.
Figure A2-1. High Damping Rubber Bearing (from Bridgestone Catalog)
1) Modified Bi-linear Model
The hysteresis of HRB is defined as a modified bilinear model as shown Figure A2-2.
Figure A2-2. Bi-linear model
The initial stiffness, K 1 , from the rubber is calculated as,
K1    G1   A / H r
u      heq   / 2  u     heq   / 2
G1     Geq   (A2-1)
u      heq   / 2
159
where  : shear strain (    / H r )
Ar : cross section area of the rubber
Hr : total height of the rubber.
Geq   : Equivalent shear modulus

Geq     0  1   2 2   3 3     n n (A2-2)
heq   : Equivalent damping factor
heq     0  1   2 2   3 3     n n (A2-3)
u   : Intercept force
u     0  1   2 2  3 3     n n (A2-4)
Example)
Diameter: 1500
Thickness of rubber: 7.5mm  20 layers  150mm
S 1  49.7
S 2  10.0
2
Nominal compression stress: 10 N / mm
Strain  Coefficient of each order

0 1st 2nd 3rd 4th
G eq   0 .1 ~ 1 .5 1.1503 -2.5382 3.3047 -2.0356 0.4728
( N / mm 2 ) 1.5 ~ 2.5 3.7412 -6.8745 5.1256 -1.6946 0.2092

2.5 ~ 3.0 0.1749 0.0261 0.0071
heq   0 .1 ~ 1 .5 0.135 0.0903 -0.13 0.1067 -0.032
1.5 ~ 2.5 -0.6239 1.5853 -1.1493 0.3627 -0.0427
2.5 ~ 3.0 -0.05016 0.1762 -0.0376
u   0 .1 ~ 1 .5 0.2989
1.5 ~ 3.0 0.3685 -0.0464
160
Hysteresis of a high damping rubber bearing has a characteristic of stiffness degrading according to the strain level
as shown in Figure A2-3.
Force, F Skelton curve
Deformation, δ
Figure A2-3. Hysteresis of a high damping rubber bearing
The secondary stiffness of a lead rubber bearing, K 2 , is expressed as,

K 2    G2   A / H r (A2-5)
G2    1  u  Geq   (A2-6)
The shear force is defined as,

PM    K h   X M (A2-7)
K h    Geq   A / H r (A2-8)
where X M : the maximum deformation
Also intercept force is defined as,
Pd    u   PM   (A2-9)
The hysteresis rules are:
a) Elastic range
Under the strain level less than   0.1 , the hysteresis is assumed to be linear with the secant stiffness at the
strain, that is:
K 0  K h (  0.1) (A2-10)
b) Loading on the skeleton curve after elastic range
Under the loading on the skeleton curve after elastic range, tangent stiffness is used to estimate the response at
the next step:
K  dQ( ) / d (A2-11)
161
2) Consideration of strength reduction by dissipated energy
Reference
1) Takuya Nishimura et al., “Practical Response Evaluation Method for Seismic Isolation System against Long Period
Earthquake Motions - Part2- High-Damping Rubber Bearing and Lead Damper”, AIJ Annual Convention,
Architectural Institute of Japan, 2013, pp.767-768 (in Japanese)
In the above reference, the reduction factors of equivalent stiffness and equivalent damping are proposed as,
C K  0.0073  E V   1.0　　 
E V  10.0 N mm 2  (A2-12)
C K  0.0025  E V   0.952 　　 
E V ＞10.0 N mm 2 

C h  0.0039  E V   1.0 　　E V  10.0 N mm 2  (A2-13)

C h  0.0016  E V   0.977 　　E V ＞10.0 N mm 2 
where E: dissipated energy, V: volume of rubber
To consider the strength reduction by energy dissipation, STERA_3D modifies the equivalent shear modulus and
the equivalent damping factor as,
Geq    C K Geq   0.1 (A2-14)
heq    C h heq   0.1 (A2-15)
162
A-3. Hysteresis of Lead Damper
Reference
1) Takuya Nishimura et al., “Experimental Study on Multi-cyclic Characteristics of Devices for Seismic Isolation
against Long Period Earthquake Motions: Part 7- Lead Damper”, AIJ Annual Convention, Architectural Institute of
Japan, 2011, pp.667-668 (in Japanese)
2) Takuya Nishimura et al., “Study on Multi-cyclic Modeling of Devices and Response Evaluation for Seismic
Isolation against Long Period Earthquake Motions: Part 5-Modeling of Lead Damper and Seismic Response
Analyses”, AIJ Annual Convention, Architectural Institute of Japan, 2012.9, pp.383-384 (in Japanese)
Cast Lead
Steel Flange Plate
Figure A3-1. Lead damper
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
Dissipated energy E (kNm)

Figure A3-2. Relationship between dissipated energy and strength reduction factor
In the above references, from the cyclic loading test of a lead damper with the different horizontal displacement
amplitudes, three line graphs are obtained for the relationship between the dissipated hysteresis energy and the
horizontal strength reduction ratio. The breaking points of the line are proposed as follows to match the test results.
163
a) The first point of strength reduction, R1 , E1 
R1  1.0
0.1  d  0.4 , 195  d ＜0.1, 　892  d ＞0.4 (kNm)
E1  37  2322 d 　　 (A3-1)
b) The second point of strength reduction,  R2 , E 2 

0.1  d  0.4 , 0.680  d ＜0.1, 0.860  d ＞0.4
R2  0.62  0.60 d 　　
E2 　 2,205 (kNm) (A3-2)
c) The third point of strength reduction, R3 , E3 

0.1  d  0.4 , 0.428  d ＜0.1, 0.585  d ＞0.4
R3  0.375  0.525 d 　　
E3 　 8,000 (kNm) (A3-3)
d) The fourth point of strength reduction,  R4 , E 4 

R4  0
0.15  d  0.4 , 9854  d ＜0.15, 　8859  d ＞0.4 (kNm)
E 4 　 9683  2060 d 　　 (A3-4)
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)
where, R : Strength reduction factor

Qd : Initial strength of a lead damper
Force, F
Skelton curve
Deformation, δ
164
A-4. Hysteresis of Elastic Sliding Bearing
Reference
1) Shigeo Minewaki et al., “Study on Multi-cyclic Modeling of Devices and Response Evaluation for Seismic Isolation
against Long Period Earthquake Motions : Part 2-Modeling of Low Friction Bearing and Viscous Damper”, AIJ
Annual Convention, Architectural Institute of Japan, 2012, pp.377-378 (in Japanese)
Figure A4-1. Elastic Sliding Bearing
In the above reference, the dynamic friction coefficient changes according to the temperature of the sliding plate as,
 0  7.5  10 5  T  0.0145 (A4-1)
The change of the friction coefficient is expressed as a function of the increment of temperature as
  0.03  T  1
0.06
 0.03 (A4-2)
On the other hand, the increment of temperature has the following relationship with the dissipated energy
E d (kNmm),
0.9
T  0.00019  E d (A4-3)
Therefore, the dynamic friction coefficient is obtained from the dissipated energy,
   0 T    Ed  (A4-4)
The hysteresis of the elastic sliding bearing is defined as a bilinear model. In STERA_3D, the initial friction
coefficient is temporary assumed as  0  0.029 from the catalog of a manufacture. The strength reduction by
energy dissipation will be expressed as,
Qd   0   E d  Qd 0  0  (A4-5)
where, Qd 0 : Initial strength of an elastic sliding bearing
165
A-5. Hysteresis of Bouc-Wen Model
1 1
0 0
‐3 ‐2 ‐1 0 1 2 3 ‐3 ‐2 ‐1 0 1 2 3
‐1 ‐1
β=0.5, γ=0.5 β=0.1, γ=0.9
Reference
1) Terje Haukaas and Armen Der Kiureghian, “Finite Element Reliability and Sensitivity Methods for
Performance-Based Earthquake Engineering”, PEER 2003/14, APRIL 2004
2) Wen, Y.-K. (1976) “Method for random vibration of hysteretic systems." Journal of Engineering
Mechanics Division, 102(EM2), 249-263.
3) Baber, T. T. and Noori, M. N. (1985). “Random vibration of degrading, pinching systems." Journal
of Engineering Mechanics, 111(8), 1010-1026.
a) Basic formulation
The basic formula of Bouc-Wen model is
f   k 0 x  1    k 0 z (A5-1)
z 

Ax   x z
N 1
z   x z
N
 (A5-2)

where,  ,  , and N are parameters that control the shape of the hysteresis loop, while A,  , and 
are variables that control the material degradation.
From the yield deformation,  y , the parameters  ,  are expressed as,
   0  yN and    0  yN (A5-3)
The model can be written as,
A z
N
 sgnxz     z x
z  x  (A5-4)
 x t
166
This leads to the following expression for the continuum tangent
f z A z
N
 sgnxz    
k   k 0  1    k 0   k 0  1    k 0 (A5-5)
x x 
The evolution of material degradation is governed by the following choice of equations (Baber and
Noori 1985):
A  A0   A e,   1   e,   1    e (A5-6)
where e is defined by the rate equation
e  1    k 0 z x (A5-7)
and A0 ,  A ,  , and   are user-defined parameters.
b) Incremental form for numerical analysis

Incremental form of Eq.(A5-1) is
f ( n 1)   k 0 x( n 1)  1    k 0 z ( n 1) (A5-8)
By a backward Euler solution,
z ( n 1)  z ( n )  t zt n 1 
(A5-9)
x( n 1)  x( n )  t x t n 1 
Applied to Eq. (A5-4),
N
  x( n 1)  x( n )  
A( n 1)  z ( n 1)  sgn z ( n 1)      ( n 1)
  t   x( n 1)  x( n )
z ( n 1)  z ( n )  t (A5-10)
 ( n1) t
where
A( n1)  A0   A e( n1) ,  ( n 1)  1   e( n 1) ,  ( n 1)  1    e( n1) (A5-11)
x( n 1)  x( n )
e( n 1)  e( n )  t 1    k0 z( n 1)
t (A5-12)
 e( n )  1    k0 z( n 1) x(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 
167

f z ( n 1)   z ( n1)  z ( n )  x ( n 1)  x( n )   0 (A5-14)
 ( n1)
N
  A( n 1)  z ( n1)   ( n 1) (A5-15)
   sgnx( n 1)  x( n ) z ( n 1)   (A5-16)
A Newton-Raphson method is applied to solve the nonlinear equation f z ( n 1)  0 ,  

f z ( n 1) 
n 1)  z ( n 1) 
z (new (A5-17)
old
f ' z ( n 1) 
where the prime f ' z ( n 1)   denotes derivative with

respect to z ( n 1) ,
f z ( n1) 
z
new old
z ( n 1) z ( n 1)
Evaluation of the function derivatives is summarized below.

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    k 0 z ( n 1) e' ( n 1)  t 1    k 0
t t
A( n1)  A0   A e( n 1) A' ( n 1)   A e' ( n1)
 ( n 1)  1   e( n 1)  ' ( n1)   e'( n 1)
 ( n 1)  1    e( n 1)  ' ( n 1)    e' ( n 1)
N
  A( n 1)  z ( n1)   ( n 1) '  A' ( n 1)  N z ( n 1)
N 1
 
sgn z ( n 1)   ( n1)
N
 z ( n 1)  ' ( n1)
 '( n 1)   '( n 1)
f z ( n 1)   z ( n 1)  z ( n )  x ( n 1)  x( n )  f ' z( n 1)   1  x  x( n ) 
 ( n1)  2( n 1) ( n 1)
(A5-18)
168
The procedure can now be summarized as follows:
1. While ( z( n1)  z( n1)  tol )

new old
(a) Evaluate function
e( n 1)  e( n )  1    k0 z( n 1) x( n 1)  x( n ) 
A( n1)  A0   A e( n1) ,  ( n 1)  1   e( n 1) ,  ( n 1)  1    e( n1)
   sgnx( n 1)  x( n ) z ( n 1)  
N
  A( n 1)  z ( n1)   ( n 1)

f z( n 1)   z( n 1)  z( n )  x( n 1)  x( n )  (A5-19)
( n 1)
(b) Evaluate function derivatives
x( n1)  x( n )
e' ( n 1)  t 1    k 0
t
A' ( n 1)   A e' ( n1)
 ' ( n1)   e' ( n 1)
 ' ( n 1)    e' ( n 1)
sgnz( n 1)   ( n 1)  z( n 1)  '( n 1)

N 1 N
'  A'( n 1)  N z( n 1)
'( n 1)   '( n 1)

f ' z( n 1)   1  x ( n 1)  x( n )  (A5-20)
 2( n 1)
(c) Obtain trial value in the Newton-Raphson scheme
f z( n1) 
n 1)  z ( n 1) 
z(new (A5-21)
f ' z( n1) 
(d) Update z( n 1)
n 1)  z ( n 1)
z(old and z( n1)  z(new
n 1) (A5-22)
169
c) Tangent stiffness
The tangent stiffness is necessary to compute the nonlinear structural analysis.
From the incremental forms:
f ( n 1)   k 0 x( n 1)  1    k 0 z ( n 1)
N
  x( n 1)  x( n )  
A( n 1)  z ( n 1)  sgn z ( n 1)      ( n 1)
  t   x( n 1)  x( n )
z ( n 1)  z ( n )  t
 ( n1) t
The tangent stiffness is calculated as (T. Haukaas and A. D. Kiureghian, 2004);
f ( n1) z( n1)

k ( n1)    k 0  1    k 0 (A5-23)
x( n1) x( n1)
z( n1) b4
 (A5-24)
x( n1) b5
where
   sgnx( n1)  x( n ) z ( n1)  
N
  A( n 1)  z ( n1)   ( n 1)
b1  1    k 0 z( n1)
b2  1    k 0 x( n1)  x( n ) 
b3 
x ( n1)  x( n ) 
 ( n1)
 
b4  b3 Ab1  b3 z( n1)  b1 
N
x
( n 1)  x( n )  b1 
 2
( n 1)  ( n1)
sgnz( n1)  ( n1)

N 1
b5  1  b3 Ab2  b3 N z( n1)

N
 b3 z( n1)  b2  x ( n 1)  x( n )  b2
(2n1)
170
3.7 Masonry Wall
B1 B B2
N ' z1 ,  ' z1 N 'z 2 ,  'z 2

l
Q' xc ,  ' xc
A1 A2
A
Figure 3-7-1 Element model for masonry wall
a) Nonlinear shear spring
Hysteresis model of the nonlinear shear spring is defined as the poly-linear slip model as shown in Figure
3-7-2.
Q
Q
Qy
k 0
Qc 
Qu
k0
c y u 
Figure 3-7-2 Hysteresis model of the nonlinear shear spring
The characteristic values, Qc , Q y , Qu are obtained based on the formulation described in the reference
(Paulay and Priestley, 1992).
The procedure to obtain the shear strength is shown below:
171
(1) Compression strength of masonry prism
Compression strength of diagonal strut is

R  Z t f 'm (3-7-1)
where,
f 'm : Compression strength of the masonry prism
Z : Width of the diagonal strut (Z = 0.25 d, d is diagonal length)
t : Thickness of wall
The compression strength of the masonry prism ( f ' m ) is determined by the following equation (Paulay and
Priestley, 1992),
f ' cb ( f 'tb f ' j )
f 'm  (3-7-2)
U u ( f 'tb f ' cb )
j
 (3-7-3)
4.1hb
where,
f 'cb : Compressive strength of the brick
f 'tb : Tensile strength of the brick (= 0.1 f ' cb )
f 'j : Compressive strength of the mortar
j : Mortar joint thickness
hb : Height of masonry unit
Uu : Stress non-uniformity coefficient (=1.5)
The shear strength is then obtained as,

Vc  R cos   Z t f ' m cos  (3-7-4)
172
(2) Shear strength by sliding shear failure
The maximum shear stress is obtained from the Mohr-Coulomb criterion:
 f   0   0   0  tan   0 (3-7-5)
where,
0 : Cohesive capacity of the mortar beds (=0.04 f ' m ) (Paulay and Priestly, 1992)
 : Sliding friction coefficient along the bed joint
  0.654  0.000515 f ' j (Chen et.al, 2003)
0 : Compression stress (  W / Aw  R sin  / Aw )
The shear strength is
 W 
V f   f AW   0    AW   0 Aw  W (3-7-6)
 AW 
Substituting V f  R cos  , W  R sin 

where  is an angle subtended by diagonal strut to horizontal plane
R cos    0 Aw  R sin 
R cos  1   tan     0 Aw (3-7-7)
 0 Aw
R cos  
1   tan 
Therefore,
 0 Aw
Vf  (3-7-8)
1   tan 
173
(3) Characteristic values of nonlinear skeleton
The shear resistance, Q y , is calculated to be the minimum value between the shear strength by sliding
shear failure, V f , and the shear strength of diagonal compression failure, Vc , that is,
Q y  min(V f , Vc ) (3-7-9)
The shear displacement at the maximum resistance,  y , is obtained as (Madan et al.,1997),
 'm d m
y  (3-7-10)
cos 
where,
 'm : Compression strain at the maximum compression stress
(  ' m =0.0018, Hossein and Kabeyasawa, 2004)
Initial elastic stiffness is assumed as (Madan et al., 1997)
k 0  2Q y /  y (3-7-11)
From Figure 3-7-2, the shear resistance at crack, Qc , is obtained as,

Q y  k 0  y
Qc  (3-7-12)
1
where,  is the stiffness ratio of the second stiffness and assumed to be 0.2.
Shear displacement at crack is then obtained as,

 c  Qc / k 0 (3-7-13)
Shear resistance and displacement at the ultimate stage are assumed as (Hossein & Kabeyasawa, 2004)
Qu  0.3Q y (3-7-14)
 u  3.5(0.01hm   y ) (3-7-15)
where, hm is the height of masonry wall.
References:
1) T. Pauley, M.J.N. Priestley, 1992, Seismic Design of Reinforced Concrete and Masonry building, JOHN
WILEY & SONS, INC.
2) Hossein Mostafaei, Toshimi Kabeyasawa, 2004, Effect of Infill Walls on the Seismic Response of
Reinforced Concrete Buildings Subjected to the 2003 Bam Earthquake Strong Motion : A Case Study of
Bam Telephone Centre, Bulletin Earthquake Research Institute, The university of Tokyo
3) A. Madan,A.M. Reinhorn, ,J. B. Mandar, R.E. Valles, 1997, Modeling of Masonry Infill Panels for
Structural Analysis, Journal of Structural Division, ASCE, Vol.114, No.8, pp.1827-1849
174
b) Vertical springs
For the moment, the vertical springs of the element model in Figure 3-7-1 are assumed to be elastic springs.
N ' z1  k z  ' z1 , N 'z2  k z 'z2 (3-7-16)
k z  E m (tl w ) / 2 (3-7-17)
where,
Em : Modulus of elasticity of masonry prism (=550 f ' m , FEMA 356, 2000)
t : Thickness of masonry wall
lw : Width of masonry wall
175
3.8 Passive Damper
a) Hysteresis damper
Hysteresis damper is modeled as a shear spring as shown in Figure 3-8-1.
B1 B B2
l
Q' xc ,  ' xc
A1 A2
A
Different types of hysteresis model are prepared for the force-deformation relationship of the spring.
qi qi qi
fy ky ky ky
fy fy
fc k1 fc k1
k0 k0 ui k0 ui
ui
(a) Bi-linear (b) Normal-trilinear (c) Degrading-trilinear
qi
fy
ui
(d) Bouc-Wen
Figure 3-8-2 Hysteresis model of the shear spring
176
b) Viscous damper
Viscous damper is modeled as a shear spring as shown in Figure 3-8-3.
B1 B B2
A1 A2
A
(1) Algorithm for oil damper devise
Figure 3-8-4 shows the Maxwell model with an elastic spring with stiffness, K d , and a dashpot with
damping coefficient, C.
Fij, uij
Node i Node j
Fk, uk Fc, uｃ
Figure 3-8-4 Maxwell model
Since the elastic spring and the dashpot are connected in a series,
Fk  Fc  Fij (3-8-1)
where, Fk : force of the elastic spring
Fc : force of the dashpot
Fij : force between i-j nodes
177
The force of the elastic spring, Fk , is obtained as,
Fk  K d uk  K d (uij  uc ) (3-8-2)
where, u k : relative displacement of the elastic spring
u c : relative displacement of the dashpot
uij : relative displacement between i-j nodes
For an oil damper, the force-velocity relationship of the dashpot is defined as shown in Figure 3-8-5.
Fc
relief point
.
uc
Figure 3-8-5 Dashpot element

The force of the dashpot after the relief point is,
Fc  C 2 u c  Qc (3-8-3)
Substituting Equations (3-8-2) and (3-8-3) into (3-8-1)

K d (u ij  u c )  C 2 u c  Qc (3-8-4)
When the time interval Δt is small enough, the velocity at time t can be expressed as,
u c (t )
u c (t )  (3-8-5)
t
u c (t )  u c (t )  u c (t  t ) (3-8-6)
Substituting above equations into Equation (3-8-4),

 
K d u ij (t )  u c (t  t )  Qc
u c (t )  (3-8-7)
C2
 Kd
t
The algorithm to obtain the force Fij (t ) from u ij (t ) is as follows:

1) Evaluate u c (t ) from Equation (3-8-7)
2) Evaluate u c (t ) from Equation (3-8-6)
3) Evaluate Fij (t ) from Equation (3-8-2)
178
Before the relief point of the dashpot, Equation (3-8-7) will be obtained by changing C2  C1 , Qc  0
as

K d u ij (t )  u c (t  t ) 
u c (t )  (3-8-8)
C1
 Kd
t
When the velocity of the dashpot is over the negative relief point, Equation (3-8-7) will be obtained by
changing Qc  Qc ,
 
K d u ij (t )  u c (t  t )  Qc
u c (t )  (3-8-9)
C2
 Kd
t
In case there is no elastic spring,
Fij, uij
Node i Node j
Fc, uｃ
Figure 3-8-6 Dashpot element without elastic spring
u ij (t )  u c (t )
Fuj  Fc  C 2 u c  Qc
u c (t ) u ij (t )
u c (t )  
t t
Therefore,
u ij (t )
Fij (t )  C 2  Qc (3-8-10)
t
Before the relief point of the dashpot,

u ij (t )
Fij (t )  C1 (3-8-11)
t
When the velocity of the dashpot is over the negative relief point,
u ij (t )
Fij (t )  C 2  Qc (3-8-12)
t
179
(2) Algorithm for viscous damper devise
Figure 3-8-7 shows the Maxwell model with an elastic spring with stiffness, K d , and a dashpot with
damping coefficient, C.
Fij, uij
Node i Node j
Fk, uk Fc, uｃ
Figure 3-8-7 Maxwell model
Since the elastic spring and the dashpot are connected in a series,
Fk  Fc  Fij (3-8-13)
where, Fk : force of the elastic spring
Fc : force of the dashpot
Fij : force between i-j nodes
The force of the elastic spring, Fk , is obtained as,

Fk  K d uk  K d (uij  uc ) (3-8-14)
where, u k : relative displacement of the elastic spring
u c : relative displacement of the dashpot
uij : relative displacement between i-j nodes
For a viscous damper, the force-velocity relationship of the dashpot is defined as shown in Figure 3-8-8,
Figure 3-8-8 Dashpot element

That is,
Fc  C sgn u c (t )  u c (t )

(3-8-15)
180
From Equations (3-8-13) and (3-8-14)
Fij (t )
 u c (t )  u ij (t ) (3-8-16)
Kd
Taking time differential and substituting Equation (3-8-15) give
1/
Fij (t )  Fij (t ) 
Kd

 sgn Fij (t )  
 C


 u ij (t ) (3-8-17)
 
The numerical integration method, Runge-Kutta Method, can be used to solve the Equation (3-8-17).
In general, the solution of the differential equation, y (t )  f ( y , t ) , is obtained by Rungu-Kuttta Method as

follows:
1
y n 1  y n  k 0  2k1  2k 2  k 3  (3-8-18)
6
k 0  f ( y n , t n )t
k1  f ( y n  k 0 / 2, t n  t / 2)t
k 2  f ( y n  k1 / 2, t n  t / 2)t
k 3  f ( y n  k 2 , t n  t )t
Equation (3-8-17) can be written as

1/
  Fij (t )  

 
Fij (t )   u ij (t )  sgn Fij (t )  
 C   d
 K (3-8-19)
    

Applying Runge-Kutta Method gives the following algorithm,
1
Fij (t n 1 )  Fij (t n )  k 0 (t n )  2k1 (t n )  2k 2 (t n )  k 3 (t n )  (3-8-20)
6
1/ 
  Fij (t n )  
 

k 0   u ij (t n )  sgn Fij (t n ) 
 C
 
  K d t
   
 
1/

  Fij (t n )  k 0 / 2  

k1   u ij (t n  t / 2)  sgn Fij (t n )  k 0 / 2 

 
  d
K t
 
C
  

1/ 

  Fij (t n )  k1 / 2  

k 2   u ij (t n  t / 2)  sgn Fij (t n )  k1 / 2 

 
  d
K t
 
C
  

1/ 

  Fij (t n )  k 2  

k 3   u ij (t n  t )  sgn Fij (t n )  k 2 

 
  d
K t
 
C
 
 
181
In this algorithm, it is assumed as,
u ij (t n )  u ij (t n  t )
u ij (t n  t / 2)  (3-8-21)
2
182
4. Freedom Vector
4.1 Node freedom
Each node has six degrees of freedom and the freedom number is defined as shown in Figure 4-1-1.
Z 3 Z Z
6
X 1 7 X X
4
5 8
Y Y Y
2
(a) lateral and rotational displacement (b) shear deformation of connection
4.2 Freedom vector
The freedom vector is defined to indicate the number of all freedoms of the structure, where the restrained
freedom is set to be zero. For the structure in Figure 4-2-1, the freedom vector has zero components for the
fixed nodes (Nodes 1-4) and eight components for other nodes (Nodes 5-8). Therefore, the total number of
freedom of the structure is 8×4 = 32.
183
Node 1  0 
| 
 
Node 4 0
 
1
Node 5  6
  5
8
9
  7 8
Node 6

16 
  1 2
17 
Node 7  

24 3 4
 
25
Node 8 
 
32  shear deformation of connection
Figure 4-2-1 Example of the freedom vector
184
4.3 Dependent freedom
(1) Rigid floor assumption
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
Figure 4-3-1 In-plane and out-of-plane freedom
For example, the in-plane freedoms at the node, A, in Figure 4-3-2 are expressed by the in-plane freedoms
at the center of gravity, G, as follows:
u xA  1 0 l yA  u xG 
    
u yA   0 1  l xA  u yG  (4-3-1)
  0 0 1   zG 
 zA  
G: center of gravity
 zG
G G u xG
l xA
l yA
u yG  zA
A
A u xA
u yA
Figure 4-3-2 Rigid floor assumption
185
For example, in case of the structure in Figure 4-3-3, in addition to the original nodes, a new node for the
center of gravity is defined as “Node 10” in the first floor. Under the rigid floor assumption, the freedom
vector has zero components for the in-plane freedoms at the nodes except the center of gravity. Therefore,
the total number of independent freedom is 23.
0
Node 1-5 :
0
0
0
1
2
Node 6 3
0
4
5
0
0
6
Node 7 7
8
0
9 6 7
10
0 10
0
11 8 9
Node 8 12
13
0
14 1 2
15
0 5
0 3
16 4
17
Node 9 18
0
19
20 shear deformation of connection
21
22
0
0
Node 10 0
23
0
0
Freedom vector
Figure 4-3-3 Example of the freedom vector with rigid floor assumption
186
(2) Including wall element
The wall element model has rigid beams at the top and bottom of the wall, therefore, as shown in Figure
4-3-4, the rotation angles in the wall panel plane,  y1 and  y 2 , are dependent to the vertical
displacements,  z1 and  z 2 . Also, the horizontal displacement in the wall panel plane, u x 2 , is
dependent to the displacement, u x1 . The connection is assumed to be rigid.
 y2
 y1
 z 2   z1
 y1   y 2 
 z1  z2 w
u x1 u x2 u x1  u x 2
Figure 4-3-4 Relationship between node displacements for a wall element (X-wall)
In a matrix form;
 u x1  1 0 0  u x 2 
    
 y1   0  1 / w 1 / w  z1  (4-3-2)
  0  1 / w 1 / w  
 y2    z2 
In case of Y-direction wall, the relationship can be written as;
 x1
u y1  1 0 0  u y 2 
 z1     
 x1   0 1 / w  1 / w  z1  (4-3-3)
  0 1 / w  1 / w  
1  x2    z 2 
 x2
u y1
 z2
2 w
u y2
Figure 4-3-5 Relationship between node displacements for a wall element (Y-wall)
187
For example, in case of the structure in Figure 4-3-6, by eliminating dependent freedoms, the total number
of freedom becomes 17.
0
Node 1-5 :
0
0
0
1
Node 6 2
0
0
0
0
0
0
3
Node 7 4
0
0 6 7
0
0 10
0
0 8 9
5
Node 8 6
7
0 1 2
8
9 5
0
0 3 4
10
Node 9 11
12
0
13
14 shear deformation of connection
15
16
0
0
Node 10 0
17
0
0
Freedom vector
Figure 4-3-6 Example of the freedom vector with a wall element
188
(3) Series of walls
In case several walls are joined together in series, it is assumed that all walls are connected by rigid beams
at the top and bottom. Therefore, as shown in Figure 4-3-7, the rotation angles in the wall panel plane,  y1
and  y 2 , are dependent to the vertical displacements,  z1 and  z 2 . Also, the horizontal displacement in
the wall panel plane, u x 2 , is dependent to the displacement, u x1 . The connection is assumed to be rigid.
 yN
 y2
 y1
 z1  z2  zN
u x1 u x2 u xN
w1
 zN   z1 N 1
 y1   y 2     yN  , L   wi
L k 1
i 1
 zi   z1   yi Li  1  Li / L  z1  Li / L  zN , Li   wi
k 1
u x1  u x 2    u xN
Figure 4-3-7 Series of wall connected by a rigid beam (X-wall)
In a matrix form;
 
 yi   1 / L 1 / L  z1  (4-3-4)
 zN 
 
 zi  1  Li / L Li / L  z1  (4-3-5)
 zN 
189
In case of Y-direction wall, the relationship can be written as;
 x1
 z1
 x2
1
u y1
 z2
2 w1
u y2
 zN
N  z1   zN N 1
 x1   x 2     xN  , L   wi
u yN L k 1
i 1
 zi   z1   xi Li  1  Li / L  z1  Li / L  zN , Li   wi
k 1
u y1  u y 2    u yN
Figure 4-3-8 Series of wall connected by a rigid beam (Y-wall)
In a matrix form;
 
 xi  1 / L  1 / L  z1  (4-3-6)
 zN 
 
 zi  1  Li / L Li / L  z1  (4-3-7)
 zN 
190
(4) Transformation matrix for dependent freedom
In case of rigid floor assumption, Equation (4-3-1) expresses the relationship between dependent freedom
and independent freedom, that is;
u xA  1 0 l yA  u xG 
    
u yA   0 1  l xA  u yG 
  0 0
 zA   1   zG 
Dependent freedom Independent freedom
It can be arranged into the transformation matrix between the freedom vectors of all nodes;
k l m
      
      
    
      
    
i u xA   1 0 l yA   
u yA   0 1  l yA   
    
 zA    u xG  k
   u  l
 xA     yG 
 yA     zG 
     
 zA   0 0 1
  xG 
      yG 
    
      zG  m
       
  
Dependent freedom [TI ] Independent freedom
Since the most components of the transformation matrix, [TI ] , are zero, the components of [TI ] are
remembered using two matrices, [ N I ] and [ FI ] .
 
N I   i k m 0

; Matrix for independent freedom numbers
 
 
FI   i 1 l yA 0
 ; Matrix for transformation components from independent freedoms
 
191
It will reduce the memory size dramatically.
In the same way, for the case of including wall elements, Equation (4-3-2) expresses the relationship
between dependent freedom and independent freedom, that is;
 u x1  1 0 0  u x 2 
    
 y1   0  1 / w 1 / w  y1 
  0  1 / w 1 / w  
 y2    y 2 
Dependent freedom Independent freedom
It can be arranged into the transformation matrix between the freedom vectors of all nodes;
p q r
      
      
    
      
    
 u x1   0 1 0   
      
    
j  y1    1/ w 0 1/ w   y1  p
    
     
 y 2    1/ w 0 1/ w  u x 2  q
     
     
      y 2  r
    
      
   
      
Dependent freedom [TI ] Independent freedom
The components of two matrices, [ N I ] and [ FI ] will be;
 
N I   j  p r 0

; Matrix for independent freedom numbers
 
 
FI   j  1 / w 1 / w 0
 ; Matrix for transformation components from independent freedoms
 
192
Initial conditions of [ N I ] and [ FI ] are:
   
N I   i  i 0 0 , FI   i 1 0 0

   
4.4 Formulation of independent displacement of the element
In Figure 4-4-1 (rigid connection), the element node displacement vector of the beam element between
Node 8 and Node 9 is,
u z8 u z 9  y8  y 9  x8  x 9 T (4-4-1)
Those displacements correspond to the location numbers in the freedom vector as;
u z8 u z 9  y8  y 9  x8  x 9 T  45 51 47 53 43 49T (4-4-2)
1 0
Node 1-5 : :
30 0
31 0
32 0 u z8 u z9
Node 6 33 1
34 2
35 3  x8  x9
36 0 8 9
37 0  y8  y9
38 0
Node 7 39 4
40 5
41 6
42 0
43 0 6 7
44 0
Node 8 45 7 10
46 8
47 9 8 9
48 0
49 0
50 0
Node 9 51 10 1 Z
52 11
53 12
54 0 5
55 13 3 4 X
56 14
57 0
Node 10 58 0 Y
59 0
60 15 (rigid connection)
Freedom vector
Figure 4-4-1 Example of location matrix for beam element
193
From rigid floor assumption, the components of independent matrices, [ N I ] and [ FI ] will be;
     
43 55 60 0
 43 1 l y 8 0
   
   
45 45 0 0 45 1 0 0
   
   
47 47 0 0 47 1 0 0

N I    
, FI     (4-4-3)
49 55 60 0 49 1 l y 9 0
   
   
   
51  51 0 0 51 1 0 0
   
   
53 53 0 0 53 1 0 0
     
From the matrix, [ N I ] , the freedoms of (43) and (49) are replaced to the independent freedoms (55) and
(60). Therefore, the independent location numbers and freedom numbers of the beam element are:
u z8 u z 9  y8  y9  x8  x 9 T
 45 51 47 53 43 49
T
 45 51 47 53 55 60 ; independent location number

T
 u z 8 u z 9  y8  y9 u x10  z10 
T
 5 8 7 10 11 13 ;
T
freedom number
(4-4-4)
The transformation from independent displacements (= global node displacements) to element node
displacements is obtained from the matrix, [ FI ] , as follows:
u z 8  1 0   u z8   u z8 
u   1    u 
 z9    u z9   z9 
 y 8   1    y 8    y 8 
     TixB   (4-4-5)
 y 9   1   y 9   y 9 
 x 8   1 l y 8  u x10  u x10 
      
 x 9  0 1 l y 9   z10   z10 
194
The constitutive equation of the beam element and formulation of global stiffness matrix from element
stiffness matrix are shown below:
5 8 7 10 11 13
 Pz 8  5 k 5,5 k 5,8 k 5, 7 k 5,10 k 5,11 k 5,13   u z 8 
P  8
 k 8 ,8 k 8, 7 k 8,10 k 8,11 k 8,13   u z 9 
 z9   
 M y 8  7  k 7,7 k 7 ,10 k 7 ,11 k 7 ,13    y 8 
    
 M y 9  10  k10,10 k10,11 k10,13   y 9 
 Px10  11  sym. k11,11 k11,13  u x10 
    
M z10  13  k12,12   z10 
Locate element stiffness
Element stiffness matrix
according to the freedom number
1 2 3 4 5 6 7 8 9 10 11 12 13
1  
 
2  
3  
 
4 
5 k 5, 5 k 5, 7 k 5,8 k 5,10 k 5,11 k 5,13 
 
6 
7 k 7,7 k 7 ,8 k 7 ,10 k 7 ,11 k 7 ,13 
 
8 k 8,8 k 8,10 k 8,11 k 8,13 
9  

10  k10,10 k10,11 k10,13 
 
11  sym. k11,11 k11,13 
12  
 
13  k13,13 
Global stiffness matrix
Figure 4-4-2 Formulation of global stiffness matrix
In general, the transformation from independent displacements (= global node displacements) to element
node displacements for the X-beam is described as Equation (2-1-10).
u zA 
u 
 zB  u1 
 yA  u 
 2
   TixB   (2-1-10)
 yB  
 xA  un 
 
 xB 
195
And the constitutive equation of the X-beam is also described in Equation (2-1-16).
 P1  u1 
P  u 
 2  2
   K xB   (2-1-16)
 
 Pn  u n 
Using the same procedure in Figure 4-4-2, the element stiffness matrix is added into the global stiffness
matrix.
196
5. Equation of motion
5.1 Mass matrix
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
G : center of gravity of the floor

Y Y
(1) Influence area of the node (2) Mass and radius of gyration at G
Figure 5-1-1 Mass and radius of gyration at the node
The process to determine the mass based on influence area is as follows:
Step 1. Calculate the slab area (block with cross mark)

Step 2. The area of the block is divided equally to the corner nodes. (Figure 5-1-2)
Step 3. If there is no corner node, the area is divided equally to the all nodes in a floor. (Figure 5-1-3)
197
Figure 5-1-2. Influence area of the node (red)
Figure 5-1-3. Distribution of the rest area
Example) Floor weight = 700kN
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
87.5kN 87.5kN 62.5kN 112.5kN

(a) Same for all nodes (b) Proportional to influence area
Figure 5-1-4 Example of mass distribution
198
In case of rigid floor assumption, in-plane freedoms at the nodes are dependent to the freedoms at the
center of gravity of the floor. Therefore, the mass at the center of gravity, M G , is,
M G  M floor (5-1-2)
The radius of gyration at the center of gravity, IG, is obtained from the following equation:
 
N
I G   M i lix2  liy2 (5-1-3)
i
where, N is the total number of the nodes at the floor. The radius of gyrations at other nodes are,
I i  0, i  1, , N (5-1-4)
The mass matrix is obtained as,
 0 0   
0     
   
u xi  Mi  M i 
   
u yi  Mi  M i 
 zi  Mi  M i 
M      (5-1-5)
 xi  Ii   Ii 
 yi  Ii  I 
   i 
 zi  Ii   Ii 
  0    
   
 0 0    
Since the mass matrix has only diagonal components, those components are saved in one-dimension vector.
For example, the mass vector of the structure in Figure 5-1-5 will be as follows:
M6 
Node 6  0 
 
M7 
 
Node 7  0 
M8 
  6 7
 0 
 
Node 8  0  10
M  8 9
 9
 0 
Node 9  0 
  1 2
M 10 
  5
M 10  3 4
Node 10  I 10 
Figure 5-1-5 Example of mass vector
199
5.2 Stiffness matrix
As shown in Figure 4-4-2, the global stiffness matrix K  is formulated from element stiffness matrices.
5 8 7 10 11 13
 Pz 8  5 k 5,5 k 5,8 k 5, 7 k 5,10 k 5,11 k 5,13   u z 8 
P  8
 k 8 ,8 k 8, 7 k 8,10 k 8,11 k 8,13   u z 9 
 z9   
 M y 8  7  k 7,7 k 7 ,10 k 7 ,11 k 7 ,13    y 8  Example of beam element
    
 M y 9  10  k10,10 k10,11 k10,13   y 9 
 Px10  11  sym. k11,11 k11,13  u x10 
    
M z10  13  k12,12   z10 
Locate element stiffness
Element stiffness matrix
according to the freedom number
1 2 3 4 5 6 7 8 9 10 11 12 13
1  
 
2  
3  
 
4 
5 k 5, 5 k 5, 7 k 5,8 k 5,10 k 5,11 k 5,13 
 
6 
7 k 7,7 k 7 ,8 k 7 ,10 k 7 ,11 k 7 ,13 
 
8 k 8,8 k 8,10 k 8,11 k 8,13 
9  

10  k10,10 k10,11 k10,13 
 
11  sym. k11,11 k11,13 
12  
 
13  k13,13 
Global stiffness matrix
Figure 5-2-1 Formulation of global stiffness matrix
200
5.3 Input ground acceleration
 , Y and Z , in X, Y and Z

Earthquake ground motions are defined as three components acceleration; X 0 0 0
directions. The inertia forces at node i are defined as,
           
           
       
 M i uxi  X 0  uxi  1 0 0 uxi 
  M i uyi  Y0  
        
uyi  0 1 0 
X 0  uyi   X 0 
 
  M i zi  Z0 
  
zi 
  
0

0 
1     
 zi   
  Y0   M     M U  Y0  (5-3-1)
  M    M
  I ixi   xi  0 0 0     xi   Z 
0   Z   0
  I iyi  yi  0 0
0
  
 yi 
        
 I izi 
   zi  0 0 0  zi 
           
       
           
For example, the components of the matrix U  of the structure in Figure 5-4-1 will be as follows:
X 0 Y0 Z0

0 0 1
Node 6 0 0 0

0 0 1
 
Node 7 0 0 0
0 0 1
  6 7
0 0 0
Node 8 0 0 0 10
 
0 0 1 8 9
0 0 0
 
Node 9 0 0 0 1 2
 
1 0 0
5
0 1 0 3 4
 
Node 10 0 0 0
Figure 5-4-1 Components of the matrix U 
201
5.4 Equation of motion
Equilibrium condition of the structure under earthquake ground motion is:
 X 0 
C u  K u  M u  M U  Y0  (5-4-1)
 Z 
 0
Damping force
Restoring force Inertia force
Finally the equation of motion is obtained as:
 X0 
M u  C u  K u  M U  Y0   P (5-4-2)
 Z 
 0
202
5.5 Modal analysis
1) Eigen value problem

The free vibration equilibrium equation without damping is
M u  K u  0 (5-5-1)
where K  is the stiffness matrix and M  is the mass matrix in the form;
m1 0  0
0 m2  0 
M    (5-5-2)
    
 
0 0  mn 
The solution can be postulated to be in the form
u    e it (5-5-3)
where   is a vector of order n,  is a frequency of vibration of the vector   .
Substituting into the equilibrium equation, the generalized eigen problem is obtained as,
K     2 M   (5-5-4)
This eigen problem yields the n eigen solutions  ,  ,  ,  ,,  ,   where the
2
1 1
2
2 2
2
n n
eigen vectors are M-orthonormalized as,
i   M  j    mki ,k j ,k  0

n
T
; i j (5-5-5)
k 1
)
Let’s assume two different set of eigen solutions  ,  ,  ,  .
i
2
i
2
j j
Form Equation (5-5-4),
i   K  j   i   K  j   i 2 i   M  j 

T T T
(5-5-6)
Since K  and M  are the symmetric matrices,
i   K  j    j   K i    j 2  j   M i    j 2 i   M  j 

T T T T
(5-5-7)
Subtracting Equation (5-5-7) from Equation (5-5-6),
 i
2

  j 2 i   M  j   0
T
(5-5-8)
Since i   j , we obtain Equation (5-5-5).

The vector i  is called the i-th mode shape vector, and  i is the corresponding frequency
of vibration.
203
2) Modal decomposition of equilibrium equation
Defining a matrix   whose columns are the eigenvectors and a diagonal matrix   
2
which stores the eigenvalues on its diagonal as,
12 
 

 
2
  1   2    n  ,  
2

2 

(5-5-9)

 2
  n 
We introduce the following transformation on the displacement vector of the equilibrium

equation (5-4-2):
u (t )   q(t ) (5-5-10)
Then,
 M  q  C  q   K  q  P (5-5-11)
Multiplying   ,
T
   M  q    C  q     K  q    P

T T T T
(5-5-12)
where
 m1 
 m2 
   M     M     , m   T  M  
T
(5-5-13)
  i i i
 
 mn 
 k1 
 
k2
               
     K  , k   2m
T T
 K     M    M 
    i i i
 
 kn 
(5-5-14)
A damping matrix that is diagonalized by   is called a classical damping matrix.
c1 
 c2 
  C    C   
 
T
(5-5-15)
 
 
 cn 
where,  M  , C  and  K  are called generalized modal mass, modal damping and modal
stiffness matrix, respectively.
204
Therefore,
 M  q  C  q   K  q     P

T
(5-5-16)
It can be reduced to n- equations of the form
mi qi (t )  ci qi (t )  ki qi (t )  ri (t ) (5-5-17)
 X0 (t ) 
 
where ri (t )  i   P (t )   i   M U   Y0 (t ) 
T T
(5-5-18)
 Z (t ) 
 0 
By setting ci / mi  2hii and ki / mi  i

2
 X0 (t ) 
  
qi (t )  2hii qi (t )  i 2 qi (t )    i   Y0 (t )    i , x X0 (t )   i ,yY0 (t )   i , x Z0 (t )
T
 Z (t ) 
 0 
(5-5-19)
where
i   M U  i   M  U x  U y  U z 

T T
 i   T  
T
  
T
  (5-5-20)
i   M i  i   M i 
T i , x i ,y i ,z
i   M U x,y,z 

T
i , x, y , z  (5-5-21)
i   M i 
T
i , x , y , z is called “participation factor” of i-th mode.
i , x , y , z is the coefficient when you decompose the vector U x ,y,z  into mode vectors as,
U        
n
x ,y,z x, y, z i, x, y, z i  (5-5-22)

i 1
)
   M  ,
T
Multiplying
   M U x,y,z      M   x, y , z    M   x, y , z 

T T
(5-5-23)
Therefore,
    M     M U 
1 T
x, y , z x ,y,z (5-5-24)
It is equivalent to Equation (5-5-21).
205
Equation (5-5-17) can be decomposed again as,
xi (t )  2hii xi (t )  i 2 xi (t )   X0 (t )


y (t )  2h  y (t )   2 y (t )  Y (t )
i i i i i i 0 (5-5-25)
zi (t )  2hii zi (t )  i zi (t )   Z0 (t )
 2
and
qi (t )   i , x xi (t )  i ,y yi (t )  i ,z zi (t ) (5-5-26)
Therefore, the displacement vector is obtained by superposing displacement responses of

single-degree-of-freedom (SDOF) systems in each mode and each direction as,
n n n n
u (t )   q(t )   i  qi (t )   i , x i  xi (t )   i ,y i  yi (t )   i ,z i  zi (t )
i 1 i 1 i 1 i 1
(5-5-27)
 i , x i  is called “participation vector” of i-th mode in x-direction.
3) Effective modal mass

We consider the vibration of the structure with i-th mode participation vector i i  and the
frequency i . The amplitude of the vibration can be expressed as
i (t )  si i  (5-5-28)
where s is a scalar value. The amplitude of the velocity is
i (t )  si i i  (5-5-29)
The kinematic energy of the vibration is then calculated as,

1 1 1
i   M i    si i  i   M i    si i  mi
T 2 T 2
Ei  (5-5-30)
2 2 2
The kinematic energy of SDOF system with an effective mass, M e ,i , under the same
amplitude and the same frequency is
1
 si  M e,i
2
Ee ,i  (5-5-31)
2
Therefore, to be the same kinematic energy, the effective modal mass is defined as,
M e ,i   i 2 mi (5-5-32)
It can be also expressed as,
   M U  m    T M U  i   M U 

T 2
T
M e ,i   i 2 mi   i i T i  i    (5-5-33)
i   M i  i   M i 
i T
206
The sum of effective modal mass is,
n n
 M e,i   i i   M U   U   M U 

T T
(5-5-34)
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 (5-5-10) as,
   M u (t )     M  q(t )   M  q(t )

T T
(5-5-35)
Therefore,
qt 0   M     M ut 0 , qt 0   M     M ut 0

1 T 1 T
(5-5-36)
207
5.6 Damping matrix
In STERA 3D program, the damping matrix is formulated in the following way:
1) Proportional damping
The mass-proportional damping and the stiffness-proportional damping are defined as,
C   a0 M  and C   a1 K  (5-6-1)
where the constants a0 , a1 have units of sec-1 and sec, respectively.
For a system with a mass-proportional damping, the generalized damping for the i-th mode in
Equation (5-5-12) is obtained as,
ci  a0 mi , ci / mi  2hi  i (5-6-2)
Therefore,
a0 1
a0  2hii , hi  (5-6-3)
2 i
Similarly, for a system with a stiffness-proportional damping, the generalized damping for the
i-th mode is,
ci  a1i mi , ci / mi  2hii
2
(5-6-4)
Therefore,
2hi a1
a1  , hi  i (5-6-5)
i 2
hi
C  a0 M C  a1 K
a0 1 a1
hi  hi  i
2 i 2
In STERA_3D, you can select from the two types of stiffness-proportional damping.
One is the proportional damping using the initial stiffness matrix:
C   2h K 0  (5-6-6)
1
208
where, h: damping factor, 1 : circular frequency of the first natural mode, K 0  : the initial stiffness.
Another is the proportional damping using the spontaneous stiffness matrix
C   2h K p  (5-6-7)

where, h: damping factor, 1 : circular frequency of the first natural mode, K p  : the spontaneous
stiffness changing according to the nonlinearity of structural elements.
In the scene of the practical design of Japan, it is common to use the proportional damping using the
spontaneous stiffness matrix.
2) Rayleigh damping
A Rayleigh damping matrix is defined proportional to the mass and the initial stiffness
matrices as,
C   a0 M   a1 K 0  (5-6-8)
The modal damping ratio for the i-th mode is,

a0 1 a1
hi   i (5-6-9)
2 i 2
The coefficients a0 , a1 can be determined from specified damping ratios h1 , h2 modes,
respectively. Expressing Equation (5-6-9) for these two modes in matrix form leads to:
1 1 / 1 1  a0   h1 
   (5-6-10)
2 1 /  2  2   a1  h2 
Solving the above system, we obtain the coefficients a0 , a1 :

21 2 1h2   2 h1 
a0 

1 2   2 2  (5-6-11)
21h2   2 h1 
a1 
 1
2
 2
2

hi
C  a0 M  a1 K
a0 1 a1
hi   i
2 i 2
209
3) Damping matrix with a base isolation building
In an actual design practice for the base isolation buildings, it is common to assume zero viscous damping
for the base isolation floor. For example, in case of the stiffness-proportional damping, the damping
matrix is defined as:
C   2h K upper  (5-6-12)

 
where, K upper : the stiffness matrix consisted with upper structures without base isolation elements.
4) Damping matrix with viscous damper devices

If there are some viscous damper devices in a structure, in addition to the proportional damping matrix, the
global damping matrix formulated from element damping matrices are considered as:
C   C pro   C v  (5-6-13)
 
where, C pro : the proportional damping matrix, C v  : the global damping matrix formulated from
element damping matrices in the same manner of the global stiffness matrix.
210
5.7 Numerical integration method
Two numerical integration methods are prepared; one is the Newmark-β method with incremental
formulation using a step-by-step stiffness matrix, and another one is the Force correction method using a
step-by-step stiffness and a force vector together. In case it is difficult to define the step-by-step stiffness of
the element such as the case of using a viscous damper element, automatically the Force correction method
is selected.
a) Newmark-β method
The incremental formulation for the equation of motion of a structural system is,
M ai   C vi   K d i   f   pi  (5-7-1)
where, M  , C  and K  are the mass, damping and stiffness matrices. d i , vi , ai  and
pi  are the increments of the displacement, velocity, acceleration and external force vectors, that is,
d i   d i 1   d i , vi   vi 1   vi  , ai   ai 1   ai , pi   pi 1   pi  (5-7-2)
f  is the unbalanced force vector in the previous step. Using the Newmark-β method,
vi   ai t   1 ai t  (5-7-3)

2
d i   vi t   1 ai t 2   ai t 2 (5-7-4)

2
From Equation (5-7-4), we obtain
1
ai   d i   1 vi   1 ai  (5-7-5)
 t 2
 t  2
Substituting Equation (5-7-4) into Equation (5-7-3) gives

vi   1
d i   1 vi   1  1 ai t  (5-7-6)
2 t  2  4 
Equations (5-7-5) and (5-7-6) are substituted into the equation of motion, Equation (5-7-2), and we obtain
 1 1 
d i  M   C   K 
  t 
2
2 t  
(5-7-7)
 1  
 p i  M  v i  1 a i   C  1 v i   1  1a i t   f 
  t  2   2  4  
The equation can be rewritten as,
Kˆ  d   pˆ 

i i (5-7-8)
where,
211
Kˆ   K   21t  C    1t  2
M  (5-7-9)
 
pˆ i   pi   M  1
vi   1 ai   C  1 vi    1  1ai t   f  (5-7-10)
  t  2   2  4  
b) Force correction method
The equation of motion of a structural system is,
M a n1   C vn1    f n   f   K d n1   d n   Pn1  (5-7-11)
where, M  , C  and K  are the mass, damping and stiffness matrices. d n1 , vn1  and a n1 
are the displacement, velocity and acceleration vector at time step (n+1).  f n  is the restoring force
vector corresponding to d n , and f  is the unbalanced force vector in the previous step. Pn1  is
the external force vector.
Using the average acceleration method,
d n1   d n   vn t   1 a n   an1 t 2 (5-7-12)

4
vn1   vn   1 a n   a n1 t  (5-7-13)

2
Substituting Equations (6-2-2) and (6-2-3) into (6-2-1),
M a n1   C  vn   1 an   a n1 t    f n   f  

 2 
K  vn t   1 a n   a n1 t 2   Pn1 

 4 
(5-7-14)
Solving for a n1  ,
La n1   Fn  (5-7-15)
where
L  M   1 C t   1 K t 2 (5-7-16)

2 4
Fn   C  vn   1 an t    f n   f   K  vn t   1 a n t 2   Pn1  (5-7-17)
 2   4 
212
M an1  C vn1   f n1  Pn1
from the following Figure,
 f n1   f n   K d n1  d n   f 
f 
 f n1
fn  K 
d n  d n1
an , vn , d n ,  f n , K n , Pn 1
Lan1  Fn 
d n1   d n   vn t   1 a n   a n1 t 2

4
vn1   v n   1 an   a n1 t 

2
K n   K n1 
 f n    f n1 
n = n+1
END
213
c) Operator Splitting method
The Operator Splitting (OS) method is a type of mixed integration method in which stiffness is divided into
linear and nonlinear (Nakashima, 1990). The explicit predictor-corrector method is employed for the
integration associated with the nonlinear stiffness, whereas the unconditionally stable Newmark-β method
is used for the integration associated with linear stiffness. The formulations are described as follows:
The equation of motion of a structural system is,
M a n1   C vn1   K d n1    f n1   K d~n1   Pn1  (5-7-18)
d~  d   v t   14 a t 

n 1 n n n
2
(5-7-19)
d n1   d n   vn t   1 a n   a n1 t 2 (5-7-20)

4
vn1   vn   1 a n   a n1 t  (5-7-21)

2
where, M  , C  and K  are the mass, damping and initial tangential stiffness matrices. d n1 ,
~
 
vn1  and a n1  are the displacement, velocity and acceleration vector at time step (n+1). d n1 is
~
 
the predictor displacement vector,  f n 1  is the restoring force vector corresponding to d n 1 , and
Pn1  is the external force vector.
Force
K d n 1 
Corrector
K d~n1 
 f n1 
Predictor
Displacement
d~ 
n 1 d n 1 
214
Substituting Equations (5-7-19), (5-7,20) , (5-7,21) to (5-7,18),
M a n1   C  vn   1 a n   a n1 t   K  d n   vn t   1 a n   a n1 t 2 
 2   4 
 1 2
  f n 1   K  d n   v n t   a n t    Pn 1 
 4 
(5-7-22)
Solving for a n 1  ,
La n1   Fn  (5-7-23)
where
L  M   1 C t   1 K t 2 (5-7-24)

2 4
Fn   C  vn   1 a n t    f n1   Pn1  (5-7-25)

 2 
The procedure for solving the equation of motion is as follows

Step 1. ~  by Equation (5-7-19).
Calculate the predictor displacement vector d n 1
Step 2. Obtain the restoring force  f n 1  corresponding to d~  in reference to the constitutive model.
n 1
Step 3. Substitute  f n1  to Equation (5-7-25) and solve the acceleration vector a n 1  , and obtain the
corrector displacement vector d n1  from Equations (5-7-20) and (5-7-21).
215
5.8 Energy
a) Equation of energy
As it was mentioned in Equation (5-5-2), the equation of motion is obtained as:
 X 0 
M u  C u  K u  M U  Y0   P (5-8-1)
 Z 
 0
For example, in case of a structure with a rigid floor in Figure 5-8-1, the displacement vector, u, consists
of 15 components (see RED numbers in Figure 5-8-1.)
 u1 
u 
u   2  (5-8-2)
  
u15 
1 4
6 7
2 5
7 15 10
10 3 6
13
8 9
8 11
9 14
12
1 2
5
3 4
Node number Freedom number
Figure 5-7-1 Example of the freedom vector of a structure with a rigid floor
The equation of energy is derived by multiplying the velocity vector, u , and integrating by the time
T
range [0-t]:
t t t t
 u M udt   u C udt   u K udt    u Pdt

T T T T
(5-8-3)
0 0 0 0
216
uT M u  t uT C udt  uT K u   t uT Pdt
2 
0
2 
0
(5-8-4)
WK  WD  WP  WI (5-8-5)
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
If the system is nonlinear, the equation of motion can be expressed as:
 X 0 
M u  C u  Qu, u   M U  Y0   P (5-8-6)
 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 (5-8-7)
where,
WK 
uT M u : Kinematic energy
2
t
W D   u C udt
T
: Damping energy
0
t (5-8-8)
W P   u Qu , u dt : Potential energy
T
0
t
W I    u Pdt : Input energy
T
217
b) Decomposition of potential energy
We can decompose the restoring force vector into the restoring force of each member as,
Qu, u   q1 u, u   q 2 u , u     q n u, u ; n : number of members (5-8-9)
Therefore, the potential energy can be decomposed 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
(5-8-10)
0 0 i 1 i 1  0  i 1
where
t
WP ,i   u qi u , u dt ;
T
potential energy of i-th member (5-8-11)
0
218
6. Nonlinear Static Push-Over Analysis
6. 1 Lateral distribution of earthquake force
The static lateral load representing the earthquake force is applied at the center of gravity in each floor.
There are several formulas to define the load distribution along the height of the building. In “STERA 3D”
program, the following distributions are prepared:
1. Ai 2. Triangular 3. Uniform 4. UBC 5. Mode
(1) Ai distribution
In the “Building Standard Law” in Japan, the design shear force of i-th story, Qi, is defined as,
n
Qi  C i  w j , C i  ZRt Ai C 0 (6-1-1)
j i
where,
Ci : design shear coefficient of i-th story,
wi : weight of i-th story,
Z: seismic zone factor,
Rt : vibration characteristic factor taking into consideration of soil condition,
Ai : lateral distribution of shear force coefficient,
C0 : design base shear coefficient (C0 =0.2 for serviceability limit, C0 =1.0 for safety limit)
If we set, Z=1.0 (Tokyo), Rt=1.0 (stiff soil, a short story building), C0=1.0 (safety design), the design shear
force distribution is simplified as,
n
Qi  Ai  w j (6-1-2)
j i
“Ai” distribution is defined as,
 1  2T
Ai  1   i  (6-1-3)
   1  3T
 i 
where,
n n
 i   w j W , W   w j : the ratio of weight upper than i-th story,
j i j 1
T: the first natural period of a building (=0.02h, h : the building height)
As shown in Figure 6-1-1, the static lateral load is obtained as,
Fn  Qn , Fi  Qi  Qi 1 i  1, , n  1 (6-1-4)
219
w6
F6 = Q6
Q6 = C6 w6
F5 = Q5 – Q6 w5
Q5 = C5 (w5 + w6)
w4
.
. n
. w3 Qi  C i  w j
. i i
. w2 C i  ZRt Ai C 0
.
w1
F1 = Q1 – Q2
Q1 = C1 (w1 + w2 + ・・・+ w6)
Figure 6-1-1 Ai distribution
(2) Triangular distribution

Triangular distribution is defined as:
 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
Figure 6-1-2 Triangular distribution
220
(3) Uniform distribution
Uniform distribution is defined as:
Fi  QB 1 n  (6-1-6)
Fi
Figure 6-1-3 Uniform distribution
(4) UBC distribution

The UBC (Uniform Building Code, 1997) gives the following formula for the calculation of lateral force
distribution:
 n 
Fi  QB  Ft  wi hi w h j

j  (6-1-7)
 j 1 
0.07 TQ B , if T  0.7 sec

Ft   (6-1-8)
 0 , if T  0.7 sec
Ft
Fi
h6
h2
h1
Figure 6-1-4 UBC distribution
221
(4) Mode distribution
Mode distribution is defined as:
 n 
Fi  QB  wi1,i w  j

1, j  (6-1-9)
 j 1 
where,
1,i : component of the first mode distribution in the i-th story
Fi
1,i
Figure 6-1-5 Mode distribution
222
6. 2 Capacity Curve
The Capacity Spectrum Method was proposed by Freeman [1978] as an approximate way to estimate the
maximum response of a structure under an earthquake ground motion. The concept was modified by
Kuramoto et.al [2000] to adopt the distribution of nonlinear story displacement as the first mode shape in
each calculation step. The method was adopted as one of the evaluation procedures in the Building
Standard Law, Japan.
The key concept of the Capacity Spectrum Method is to find out the intersection between the Demand
Spectra (= relationship between Sd (displacement spectra) and Sa (acceleration spectra)) and the Capacity
Curve (= nonlinear push-over curve of an equivalent single-degree-of-freedom system).
1400
1200
Demand Spectra
SA-SD（ h=0.05)
1000
Sa (cm/sec 2)
Demand Spectra
800 SA-SD（ h=0.14)
Capacity Curve
600
400
Performance Point
200
0
0 10 20
Sd (cm)
Figure 6-2-1 Schematic example of the concept of Capacity Spectrum Method
“STERA 3D” provides the menu in the static analysis to show the Capacity Curve based on the following
formula (Kuramoto et.al [2000]):
n n
 mi i2 m  i i
2
S a  QB i 1
2 , Sd  i 1
n
(6-2-1)
 
m 
n
  mi  i  i i
 i 1  i 1
where,
mi : lumped mass in the i-th story
δi : component of the distribution of nonlinear story displacement in the i-th story
223
Sd
Fi δi
≈ M Sa
M : Equivalent SDOF mass
Nonlinear static push-over analysis Capacity Curve of SDOF system
Figure 8-2-2 Capacity Curve of the equivalent SDOF system
As schematically shown in Figure 8-2-2, the step-by-step results of nonlinear push-over analysis is used to
obtain the Capacity Curve of the equivalent SDOF system using Equation (8-2-1).
References
Freeman S. A. (1978), ”Prediction of Response of Concrete Buildings to Severe Earthquake Motion”,
Douglas McHenry International Symposium on Concrete and Concrete Structures, SP-55, American
Concrete Institute, Detroit, Michigan, pp. 589-605.
Kuramoto H., et.al. (200), “Predicting the Earthquake Response of Buildings using Equivalent Single
Degree of Freedom System”, 12th World Conference on Earthquake Engineering (12WCEE), Auckland
New Zealand，2000.2.
224
7. Others
7.1 Decomposition of shear and flexural deformation
a) Equivalent plane for each floor
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
Figure 7-1-1 Equivalent plane
L   z i  axi  by i  c 
2
Minimize
where, i : node number in the floor

a, b, c : parameters of equivalent plane
L L L
Thus,  0,  0, 0
a b c
Parameters, a, b, c are obtained by solving the following linear equation:
  z i xi   x i2  x y  x  a
i i i
  
 z i y i     y  y  b 
2
i i (7-1-1)
  z i   sym. n   c 
  
where,
n: the number of nodes in a floor
b) Decomposition of shear and flexural deformation
A story drift, D, can be divided into shear and flexural components as,
225
D = DS (shear) + DF (flexure) (7-1-2)
Assuming the distribution of floor deformation is expressed by an equivalent plane ( z  ax  by  c ), the
flexural deformation, DF, can be expressed as,
DF = -a H : x-direction (7-1-3)
DF = b H : y-direction (7-1-4)
Note that the coefficient ‘a’ is the negative angle in x-direction.

Then, the shear deformation can be obtained as,
DS = D - DF (7-1-5)
D
DF DS
in x-direction
a
H
z
a z  ax  by  c
DS DF in y-direction
b z
H
z  ax  by  c b
y
Figure 7-1-2 Decomposition of shear and flexural deformation
In STERA 3D, the flexural deformation is calculated taking the average of the bottom floor angle and top
floor angle.
 a i  a i 1 
D Fi  Hi : x-direction (7-1-6)
2
bi  bi 1 
D Fi  Hi : y-direction (7-1-7)
2
226
7.2 P-D effect
Following formulation is suggested in the following book:

James F. Doyle, “Static and Dynamic Analysis of Structures”, Kluwer Academic Publishers, 1991
a) Equilibrium of the beam with an axial load
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
Figure 7-2-1 Equilibrium of small beam segment slightly deformed
Assuming small deflection, the balance of moment on the small segment “Δx” gives
M  V x   F0 v   0 (7-2-1)

Therefore
dM dv
 V  F0 0 (7-2-2)
dx dx
d 2v
From the relationship, M  EI , the governing differential equation for the deflection shape is
dx 2
d 4v d 2v
EI  F0 0 (7-2-3)
dx 4 dx 2
The general solutions are,

for compression loading ( F0  0 ):
v( x)  c1 cos kx  c 2 sin kx  c3 x  c 4 , k 2   F0 / EI , (7-2-4)
for tensile loading ( F0  0 ):
v( x)  c1 cosh kx  c 2 sinh kx  c3 x  c 4 , k 2  F0 / EI (7-2-5)
227
b) Geometric stiffness matrix of the beam with an axial load
We assume that the axial force is constant and compressive. From the general solution, Eq. (7-2-4),
at x  0
dv(0)
v(0)  v1  c1  c 4 ,  1  kc 2  c3 (7-2-6)
dx
Consequently, the deflected shape is
v( x)  c1 cos kx  1  c 2 sin kx  kx   v1  1 x (7-2-7)
Similarly at the end of other node,
v( L)  v 2  c1 cos kL  1  c 2 (sin kL  kL)  v1  1 L (7-2-8)

dv( L)
  2  kc1 sin kL  kc 2 cos kL  1 (7-2-9)
dx
Then, the coefficients, c1 , c 2 , can be arranged as,
(1  C ) (  S )   c1  v1  1 L  v 2 
 (7-2-10)
 S
  (1  C ) c 2   1 L   2 L 
where,
C  cos kL, S  sin kL,   kL (7-2-11)
Solving this equation by Cramer’s rule gives

c1  v1 1  C   1 L( S  C )  v 2 1  C    2 L(  S ) /  (7-2-12)
c 2   v1S  1 L(1  C  S )  v 2S   2 L(C  1) /  (7-2-13)
where
   2  2C  S  (7-2-14)
Now we can rewrite the deflection function in terms of the nodal degrees of freedom. The moment and
shear force distributions can be obtained as
M ( x)  EI
d 2v
dx 2

 EI  k 2 c1 cos kx  k 2 c 2 sin kx  (7-2-15)
d 3v
  EIk 2 1  kc 2 
dv
V ( x)   EI 3
 F0 (7-2-16)
dx dx
228
Calculating nodal loads, V (0)  V1 , M (0)   M 1 , V ( L)  V1 , M ( L)  M 1 , the stiffness matrix is
 V1   2 S L(1  C )   2S L(1  C )   v1 
M    
 1   EI  
2
 L (C  S )  L(1  C )
2
L2 (  S )  1 
(7-2-17)
 V2  L3    2S  L(1  C )  v 2 
    
M 2   sym.  L2 (C  S )  2 
c) Approximation of geometric stiffness matrix
We simplify the geometric stiffness matrix to be linear in the loading F0.

Using the series expansion for the sine and cosine terms, the determinant is,
   2  2C  S 
   
  2  2 1   2 / 2   4 / 24   6 / 720   C      3 / 6   5 / 120    (7-2-18)

  1   / 15   / 12
5 2

also
1 12

 5 1   2 / 15  
 
 (7-2-19)
We now do the expansion on the stiffness terms. For example,

EI  2 2
k11 
L 
3
EI
L
   
12

  EI
 
 S  3  4    3 / 6   5 1   2 / 15    3 12 1   2 / 10  
L

(7-2-20)
Substituting   k 2 L2   F0 L / EI ,
2
k11 
EI
3
12  F0 12  (7-2-21)
L L 10 
In the same manner, we can expand for all the stiffness terms to get the stiffness matrix as
 12 6L  12 6 L   36 3L  36 3L 
 4 L2 2 
 6 L 2 L  F0   4 L2  3L  L2 
k   EI3   (7-2-22)
L  12  6 L  30 L  36  3L 
   
 sym. 4 L2   sym. 4 L2 
We can write as
k   k E   k G  (7-2-23)
where, k E  : the element elastic stiffness, k G  : the element geometric stiffness
229
d) Implementation for beam element
A z
A B
B
uB
x
uA
A B y
Figure 7-2-2 Including node movement
For beam element,
 M A  2 EI 2 1   A  EI 4 L2 2 L2   A 
M        (7-2-24)
 B L 1 2  B  L3 2 L2 4 L2   B 
Including node movement,
u 
1 1  A 
 A   L 1  0 
  
L
    1
A
(7-2-25)
 B  1
0  1 u B 
L L  B
 
 1 1 
 QA   L u 
M  L  1
1 
1  A
0  
 0  4 L
2
2L   L
2
 A   EI  1 
L  A
 QB  L3  1  1  2 L2 
4 L2   1 1  u 
  0  1  B 
 L L L L  B
M B   0 1   
u   12 6 L  u A 
 A  12
 6L 6L  1 6L
 4 L2 2 L2  
1
EI   L
1 
L
0   EI 
 A   4 L2  6 L 2 L2   A 
 3   
L  6 L  6 L   1 0  1 3
1 u B  L 
 12  6 L  u B 
 2   
 2L 4 L2   L L   B   sym. 4 L2   B 
From (7-2-22), the geometric stiffness matrix will be

 36 3L  36 3L 
 4 L2  3L  L2 
k G   F0 
(7-2-26)
30 L  36  3L 
 
 sym. 4 L2 
230
Therefore, the stiffness equation will be
 QA    12 6L  12 6 L   36 3L  36 3L   u A 
M   
 A    EI

 4 L2 2 
 6 L 2 L  F0   4 L2  3L  L2    A 

 Q B   L3  12  6 L  30 L  36  3L   u B 
        
M B    sym. 4 L2   sym. 4 L2    B 
231
e) Implementation for column element
B u xB
u yB
Bl'  yB
 xB  yB
 xB
l'
 yA
 yA
 Al ' X
 xA
A
 xA u xA
u yA Y
Figure 7-2-3 Including node movement
 M yA  2 EI 2 1  yA  EI 4 L2 2 L2   yA 
M         in X-Z plane (7-2-27)
 yB  L 1 2  yB  L3 2 L2 4 L2   yB 
 M xA  2 EI 2 1   xA  EI 4 L2 2 L2   xA 
M         in Y-Z plane (7-2-28)
 xB  L 1 2  xB  L3 2 L2 4 L2   xB 
Including node movement,
u 
 1 1   xA 
 yA   L
1
L
0 
 yA 
    1 1  u  in X-Z plane (7-2-29)
 yB   0 1  xB 
 L L   yB 
 
u 
1 1  yA

 xA   L 1  0  
L  xA 
    1 1   in Y-Z plane (7-2-30)
 xB   0  1 u yB 
L L   xB 
Note that the matrix for node movement in X-Z plane is different from that of beam element. The
force-deformation relationship in X-Z plane is then,
232
 1 1
 Q xA   L  u 
M  L  1 1   xA 
 yA   EI  1 0  4 L
2
2 L2   L
1
L
0 
 yA 
 Q xB  L3  1 1  2 L2 
4 L2   1 1  u 
   0 1  xB 
 M yB   L L   L L   yB 
 0 1   
 6 L  6 L   1 1 u 
 xA  12  6 L  12  6 L  u xA 
0   EI   
 2 2 L2  
EI  4 L  L
1
L  yA    4 L2 6 L 2 L2   yA 
 3  u 
L  6L 6 L   1 0
1 3
1  xB  L 
 12 6 L  u xB 
 2    yB   
 2L 4 L2   L L
   sym. 4 L2   yB 
(7-2-31)
Considering the difference of sign of stiffness matrix in X-Z plane, the geometric stiffness matrix will be
 36  3L  36  3L 
 4 L2 3L  L2 
k xG   F0  in X-Z plane (7-2-32)
30 L  36 3L 
 
 sym. 4 L2 
 36 3L  36 3L 
 4 L2  3L  L2 
  F
k yG  0 
30 L  36  3L 
 
 sym. 4 L2 
Therefore, changing the order of vector component, the force-deformation relationship of column will be
 Q xA  u xA   36  36  3L  3L 0 0 0 0 0 0 0 0 u xA 
Q 
 xB 
u 
 xB 
  36 36
 3L 3L 0 0 0 0 0 0 0 0 u xB 
 M yA   yA    3L 3L 4 L 2
L 2
0 0 0 0 0 0 0 0  yA 
      
M yB   yB    3L 3L  L
2
4L 2
0 0 0 0 0 0 0 0  yB 
 Q yA  u yA   0 0 0 0 36  36 3L 3L 0 0 0 0 u yA 
      
 Q yB  u yB  F0  0 0 0 0  36 36  3L  3L 0 0 0 0 u yB 
   K     
 M xA 

 xA  30 L  0 0 0 0 3L  3L 4 L2  L2 0 0 0 0  xA 

 M xB   xB   0 0 0 0 3L  3L  L2 4 L2 0 0 0 0  xB 
     
 N zA   zA 
 0 0 0 0 0 0 0 0 0 0 0 0  zA 
 
N     0 0 0 0 0 0 0 0 0 0 0 0  zB 
 zB   zB    
 M zA   zA   0 0 0 0 0 0 0 0 0 0 0 0  zA 
M 
 zB 
 
 zB   0 0 0 0 0 0 0 0 0 0 0 0  zB 
233
u xA 
u 
 xB 
 yA 
 
 yB 
u yA 
 
u 
 K   K G  yB  (7-2-34)
 xA 
 xB 
 
 zA 
 
 zB 
 zA 
 
 zB 
where,
 36  36  3L  3L 0 0 0 0 0 0 0 0
  36 36 3L 3L 0 0 0 0 0 0 0 0

 3L 3L 4 L 2
L 2
0 0 0 0 0 0 0 0
 
  3L 3L  L
2 2
4L 0 0 0 0 0 0 0 0
 0 0 0 0 36  36 3L 3L 0 0 0 0
 
0 0 0 0  36 36  3L  3L 0 0 0 0
K G   F0  (7-2-35)
30 L 0 0 0 0 3L  3L 4 L2  L2 0 0 0 0
 
 0 0 0 0 3L  3L  L2 4 L2 0 0 0 0
 0 0 0 0 0 0 0 0 0 0 0 0
 
 0 0 0 0 0 0 0 0 0 0 0 0
 
 0 0 0 0 0 0 0 0 0 0 0 0
 0 0 0 0 0 0 0 0 0 0 0 0
Then, applying translation of Equation (2-2-17), the constitutive equation of the column is;
 P1   u1 
P  u 
 2  2
   K 
C   (7-2-36)
 
 Pn  un 
where,
K C   TC T k C TC   TiC T K G TiC  (7-2-37)
234
7.3 Unbalance force correction
a) Procedure to correct unbalance force

In nonlinear analysis, sudden change of spring stiffness sometimes causes severe error for estimating
element force. For example, estimation of spring force f i 1 is overestimated in Figure 7-2-1 and
“unbalance force” is defined as,
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:
1) Calculate unbalance displacement d from the unbalance force f
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
Figure 7-3-1 Unbalance force
235
b) Unbalance force correction of MS model
For the Multi-spring model (MS model) of Column element, the sum of the unbalance forces of
nonlinear vertical springs in the member section is calculated as:
5 5
N    f i   f c ,i  f s ,i  (7-3-3)
i 1 i 1
where f c ,i : unbalance force of concrete spring,

f s ,i : unbalance force of steel spring
The unbalance displacement is then calculated as:
5 5
D  N  k i  N
i 1
 k
i 1
c ,i  k s ,i  (7-3-4)
where k c ,i : stiffness of concrete spring,

k s ,i : stiffness of steel spring
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
Figure 7-3-2 Unbalance force in MS-model
The same procedure is adopted for the MS model of Wall element.
236
7.4 Calculation of ground displacement
In STERA_3D, the ground displacement is calculated from the ground acceleration data using FFT method
and filtering techniques based on the description in the following reference:
Reference: Yorihiko Osaki, “Introduction of Spectral Analysis of Earthquake Ground Motion”, Kajima
publishing corporation, 1981 (in Japanese)
a) Discrete Fourier Transform

Assume that the acceleration data is collected at an interval, t  T / N and consists of the
N measurement data xm (m  0, 1, 2,, N  1) , where T is the period of the data that
corresponds to the duration time of data. The coefficient of a Fourier series is obtained as:
N 1
1
Ck 
N
x
m 0
m e i ( 2km / N ) k  0, 1, 2,, N  1 (7-4-1)
The inverse discrete Fourier transform is

N 1
xm   C k e i ( 2km / N ) m  0, 1, 2,, N  1 (7-4-2)
k 0
b) Integration of the data in time domain
Assume y m (m  0, 1, 2,  , N  1) is the integration of the discrete data x m in time domain.
The data y m is obtained by the following inverse discrete Fourier transform:
Nt N 1
y m    x m dt  
t  mt
 0

 2 k  0
S k e i ( 2km / N ) m  0, 1, 2,  , N  1 (7-4-3)
where, the coefficients S k are obtained from the coefficients C k as,
2v 0 N / 2 1
ImC k    N  1C 0
S0  2  
Nt k 0 k N
C 0
Sk   1  i cosk N   i C k , S N k  S N k
*
k  1, 2,  , N / 2  1 (7-4-4)
N k
C 0
SN /2  
N
237
The following band pass filter in frequency domain is applied to the coefficient S k .
0 0  f  f L1
 f  f L1 
0.5  0.5 cos   f L1  f  f L 2
 L2
f  f L1 
h f   1 f L2  f  f H1 (7-4-5)
 f  f H1 
0.5  0.5 cos   f H1  f  f H 2
 f H 2  f H1 
0 fH2  f
h f 
f ( Hz)
f L1 f L2 f H1 fH2
STERA_3D adopts the following frequency parameters:

f L1  0.1 (Hz)
f L 2  0.2 (Hz)
f H 1  10 (Hz)
f H 2  20 (Hz)
238
c) Calculation flow
The ground acceleration data is integrated twice to obtain displacement data. Band pass filter is applied
each time of the integration. The flow of calculation is summarized below:
[1] From acceleration data to velocity data
x m (m  0, 1, 2,  , N  1)
FFT Calculate Fourier coefficients of the data
Ck k  0, 1, 2,  , N  1
Eq. (7-4-4) Calculate Fourier coefficients of the data of the integration
Sk k  0, 1, 2,  , N  1
Eq. (7-4-5) Apply band pass filter
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)
[2] From velocity data to displacement data

Repeat the above process again
239

STERA3D Technical Manual

Enviado por

Dados do documento

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

STERA3D Technical Manual

Enviado por

Direitos autorais:

Formatos disponíveis

STERA 3D

STructural Earthquake Response Analysis 3D

Dr. Taiki SAITO

TOYOHOSHI UNIVERSITY OF TECHNOLOGY (TUT), JAPAN

2007/07/17 STERA_3D Technical Manual Ver.1.0 is uploaded.

(1) Global Coordinate

Figure 1-1-1 Global coordinate

Figure 1-1-1 Global coordinate

Figure 1-1-2 Local coordinate of a beam element

Figure 2-1-1 Element model for beam

Force-displacement relationship for elastic element

 ' yA   ' yA   yA   yA   M ' yA 

 M ' yA   ' yA   ' yA 

where, [k B ] is the stiffness matrix of the beam element.

Including rigid parts and node movement

 ' yA   yA    u zB   B l ' yB   u zA   Al ' yA 

Figure 2-1-2 Including rigid parts and node movement

From node axial displacements, relative axial displacement is,

Combining Equations (2-1-7) and (2-1-9),

Out of plane deformation of beam

 ' zA   zA    u yA   Al ' zA   u yB  B l ' zB 

 Al ' l' B l'

Figure 2-1-4 Beam displacement with rigid connection (X-Y plane)

From global node displacement to element node displacement

In case of Y-direction beam

Local coordinate of Y-beam Global coordinate

Figure 2-1-3 Relation between local coordinate and global coordinate

Transformation matrix for nonlinear spring displacement

X-Z plane Y-Z plane

Figure 2-2-1 Element model for column

Force-displacement relationship for elastic element

where, [ f pA ] and [ f pB ] are the flexural

 xB   M ' xB  Figure 2-2-2 Nonlinear bending springs

The force-displacement relationship in Equation (2-2-7) is then expressed as,

 ' yA   ' yA   yA   yA   M ' yA 

The flexural matrix [ f C ] is;

 M ' yA   ' yA   ' yA 

Including rigid parts and node movement

Including rigid parts and node movement,

Figure 2-2-3 Including rigid parts and node movement

From global node displacement to element face displacement

K C   TC T kC TC  (2-2-20)

 M ' yA   M ' yA   M ' yA 

M ' xB1 M ' yBc M ' xB 2

M ' xA1 M ' yAc M ' xA2

Figure 2-3-1 Element model for wall

Force-displacement relationship for elastic element

N ' zBc ,  zBc

Figure 2-3-2 Nonlinear bending springs

 yAc   M ' yAc 

 yBc   M ' yBc 

 xB1   M ' xB1 

The force-displacement relationship in Equation (2-3-7) is then expressed as,

 yAc   M ' yAc   M ' yAc 

Force-displacement relationship for nonlinear shear springs

 ' yAc   ' yAc   yAc   yAc   M ' yAc 

The flexural matrix [ f W ] is;

 M ' yAc   ' yAc   ' yAc 

Including rigid parts and node movement

 ' yAc   ' yAc 

Including rigid parts and node movement,

Figure 2-3-3 Relationship between center and node displacements

From global node displacement to element face displacement

In case of Y-direction wall