Title: Dynamic Response Analysis of
1Chapter 2 Dynamic Response Analysis of Bridge
Structures
2008 Course
Kazuhiko Kawashima
Tokyo Institute of Technology
2Uniqueness of Bridges for Structural Analysis
Bridges are
- long in the longitudinal direction, and
structural - properties and soil condition are not the same
along the - axial axis
- consists of many structural types (deck bridges,
arch - bridges, cable stayed bridges, suspension
bridges, .) - have various shapes (straight, skewed, curved,
- separation into several segments, and
combination of - those types)
- have various heights (short bridges and high
bridges) - are made of various materials (RC, PC, steel,
- composites)
3Types of Analysis
4Idealization of a Bridge
- Idealization of a superstructure substructures
- Idealization of foundations
- Idealization of soil response
- Multiple excitation effect (out of coherent GM)
5Analytical Model of Superstructures
6Analytical Model of Substructures
7Idealization of Substructures
Soil springs
83D FEM Idealization
9Plastic Deformation of Columns
10Idealization of Bridges
Stiffness Idealization
Total stiffness matrix
Element stiffness matrix
Time dependent stiffness
Mass Idealization
Total mass matrix Element mass matrix
11Viscous damping is not the only one mechanism of
producing damping force of bridges, but there are
various sources energy dissipation mechanism
which contribute to damping
12Why is viscous damping force assumed in most
analyses?
- Closed form solution can be easily available by
assuming viscous damping force, however it cannot
be obtained in other expressions.
13Why is viscous damping force assumed in most
analyses? (continued)
- Damping ratio can be a common value which can be
used in a wide range of application. - It is known based on the past studies that
viscous damping force can represent the
structural response with reasonable accuracy if
other type of energy dissipation mechanism is
appropriately idealized in terms of viscous
damping force.
14How other types of energy dissipation are
idealized in terms of viscous damping force?
Equivalent damping force
15Idealization of Damping Force
- There are various sources which contribute to
energy - dissipation.
- It is common to idealize the energy dissipation
- in terms of the viscous damping.
- Since the valuation of element damping matrix is
generally difficult, the system damping ratio is
often assumed by the Rayleigh damping as
(2.5)
16Orthogonarity condition
17(No Transcript)
18This assumption sometimes results in a
problem, because
- Solution becomes unstable sometimes
- Does not capture the fact that inelastic
response of - structural members dissipate energy which results
in an - increase of damping ratio
19How can we determine the modal damping ratios by
assigning damping ratios of each structural
components?
- Theoretically, damping ratio is defined for a
SDOF system. If we can assume the oscillation of
each structural component as a SDOF system, it
may be possible to assign a damping ratio for
each structural component.
20How can we determine the modal damping ratios by
assigning damping ratios of each structural
components? (continued)
- There is not a single method which is exact and
easy for implementation for design purpose. - Following empirical methods are frequently used
- Strain energy proportional method
- Kinematic energy proportional method
21Strain Energy Proportional Method
22Because
Strain energy of m-th element for k-th mode is
Therefore, the total energy dissipation of the
system is
23(No Transcript)
24Kinematic Energy Proportional Damping Ratio
25Which is better for determining modal damping
ratios between the strain energy proportional
method and kinematic energy proportional method?
- Damping ratios of the structural components where
large strain energy is developed are emphasized
in the strain energy proportional method.
Plastic deformation of columns
Plastic deformation of foundations soils
- Strain energy proportional method is better in a
- system in which hysteretic energy dissipation is
predominant
26Which is better for determining modal damping
ratios between the strain energy proportional
method and kinematic energy proportional method?
- Damping ratios of the structural components with
larger kinematic energy are emphasized in the
kinematic energy proportional method.
- Kinematic energy proportional method is better in
a - system in which hysteretic energy dissipation is
less significant
27Equations of Motion
Single-degree-of-freedom oscillator
28Equations of Motion (continued)
Multi-degree-of-freedom system
29Equations of Motion (continued)
Multiple Excitation
free nodal points
non-zero support displacements
30quasi-static displacements
(2.11)
dynamic displacements
From definition,
31The equation of motions can be written by
enlarging the mass, damping and stiffness
matrices as well as the dynamic load vector to
account for the nb support displacements
(2.12)
The equations of motion associated with n free
nodal point displacement become
(2.13)
32Substitution of Eq. (2.11) into Eq. (2.13) yields
By definition of the quasi-static displacement
(2.15)
0
small compared to inertia force
(2.14)
33From
(2.16)
(2.17)
nb x 1
n x n
n x nb
34Multiple Support Excitation
(2.17)
(2.18)
(2.19)
35Rigid Support Excitation
n x n
n x 3
3 x 1
36Linear Analysis
a) Natural Mode Shapes and Natural Frequencies
37Linear Analysis (continued)
where,
38where,
Only assumption
39Solving equation of motion for a SDOF system
Time History Analysis
Direct integration method
Determine
Determine force, stress and strain
40Response Spectral Method
Equation of motion for the i-th generalized
coordinate is
Instead of solving the equation, we can have the
peak qi using displacement response spectrum as
(2.31)
41Then the maximum displacement of the ith mode can
be written as
(2.32)
Therefore, the peak displacement u can be
obtained by taking root of square summation of
each mode
(2.33)
Forces of the i-th mode
42Nonlinear Dynamic Response Analysis
Equations of motion in the incremental form
(2.47)
43Nonlinear Dynamic Response Analysis (continued)
Newmarks generalized acceleration method
44Newmarks generalized acceleration method
(continued)
Constant Acceleration Method
45Newmarks generalized acceleration method
(continued)
Linear Acceleration Method
46Newmarks generalized acceleration method
(continued)
(2.50)
constant acceleration method
linear acceleration method
47Newmarks generalized acceleration method
(continued)
(2.51)
linear acceleration method
constant acceleration method
48Newmarks generalized acceleration method
(continued)
(2.47)
(2.51)
(2.52)
where,
(2.53)
(2.54)
49Accuracy of Computed Responses
Overshooting
50Accuracy of Computed Responses (continued)
Computed restoring force
Unbalance force
51Accuracy of Computed Responses (continued)
. . .
If error gt tolerance,
52Accuracy of Computed Responses (continued)
If error gt tolerance,
- re-compute using a smaller time increment of
numerical integration-This is always effective to
improve the stability and accuracy of solution. - add unbalanced forces to the incremental forces
at the next time step - use a numerical iteration
53Add unbalanced forces to the incremental forces
at the next time step
Unbalance Force at time t
Incremental Equations of Motion
Add the unbalance force to the incremental
external force
54Add unbalanced forces to the incremental forces
at the next time step (continued)
This method is effective when unloading and
reloading are important
55Numerical Iteration for the Equilibrium of
Equations of Motion
56Computer Soft-wares for Dynamic Response Analysis
General Purpose Soft-wares
- Multi-purposes
- Well maintenance
- Some kind of consensus for the results
- Not so easy to modify
- User routines may be included depending on
programs
57Computer Soft-wares for Dynamic Response
Analysis (continued)
Hand-made
- Easy to include problem-oriented routines
- Difficult for maintenance for the use of other
groups and long-term maintenance - Few consensus for the results
- Open-base forum for source code
- Prepare well documented manual and example
problems
58Structures of Computer Programs for Dynamic
Response Analysis
Input structural shape (coordinate) properties
Input ground motions
Form time invariant structural properties such as
mass matrix, stiffness of elastic elements
At time t
Form time dependent properties such as stiffness
of nonlinear element
Solve
(2.52)
59Check the accuracy by Eq. (2.62) or similar forms
If the accuracy is not enough, use small smaller
time increment, or add unbalance force to the
next incremental force, or iteration
Store the responses on a file
Repeat until the end of ground motion