multi-support excitation

In this example, we illustrate how to perform response history analysis with multiple support excitations.

The model can be downloaded here.

The multiple support excitation refers to different prescribed motions at various supports. The general discussion of linear elastic systems, as presented in typical textbooks (see for example Chopra's textbook) , is somehow not suitable to numerical implementation.

The implementation in suanPan converts whatever support excitations applied to prescribed displacements and the displacement controlled algorithm MPDC is automatically enabled to solve the system.

Thus, support excitations in forms of displacement, velocity and acceleration are supported.

A Simple Model

Here we first create a simple cantilever beam with additional point mass at the free end. The basic geometry related commands are stored in file geometry.sp.

# geometry.sp

node 1 0 0
node 2 0 1

material Elastic1D 1 100 .1

section Rectangle2D 1 12 1 1

element B21 1 1 2 1 6

mass 2 2 10 1

hdf5recorder 1 Node U1 1 2

amplitude Tabular 1 EZ

The analysis setting consists of a dynamic step with Newmark integrator. Those settings are stored in setting.sp file.

# setting.sp

step dynamic 1 30
set ini_step_size 1E-2
set fixed_step_size true

integrator Newmark 1

converger RelIncreDisp 1 1E-10 4 1

analyze

save recorder 1

exit

Apply Excitation On All Nodes

To apply acceleration excitation on all nodes, the boundary conditions need to be assigned in a normal, conventional way so that no rigid body motion is allowed.

file geometry.sp

fix 1 P 1

# on DoF 1 node 2
acceleration 2 1 -.2 1 2

file setting.sp

It shall be noted $-\ddot{u}_g$ is applied.

Apply Excitation On Support

To apply acceleration excitation on support, the corresponding DoF should not be constrained. Thus, rigid body motions are expected.

file geometry.sp

fix 1 2 1
fix 2 3 1

# on DoF 1 node 1
supportacceleration 2 1 .2 1 1

file setting.sp

Now it is ready to perform the analysis.

Result

Here the displacement of the free end is plotted.

displacement history

It shall be noted that since rigid body motions are present when excitations are applied to the supports, the relative displacement needs to be obtained by computing the difference between displacements of target DoF and support.

All time integration methods are implemented in such a way that both ways of applying excitations would yield the same results if they are equivalent theoretically. This is done by applying algorithm-aware conversion among displacement, velocity and acceleration excitations.

Implicit vs Explicit

Implicit integration methods are different from explicit methods in the sense that different quantities are used as the base quantity. In implicit methods, the base quantity is displacement, while in explicit methods, the base quantity is acceleration.

Thus, with implicit methods, both supportvelocity and supportacceleration will be converted to prescribed displacement. However, with explicit methods, supportacceleration will be converted to prescribed acceleration, while supportvelocity cannot be used. This is because typically there is no way to convert velocity to acceleration in explicit methods.

The following is a validation of several integration methods.

support motion validation

The presented results can be reproduced via this script.