Skip to content

Acoustic Wave Propagration of a Cylindrical Wave

Problem Definition

All the files in this tutorial can be downloaded here.

This example shows acoustic wave propagation in a transient analysis in 2D. Waves are excited by a sin-burst signal on a circular cylinder. They propagate outwards into a 1/4 domain (2 symmetries have been exploited).

Sketch of the domain

    _________________________
   |*************************|
   |*************************|
   |********** air **********|
   |*************************|
   |*************************|
^  ...***********************|
|  .......*******************|
R  .........*****************|
|  ..........****************|
v  ...........***************|
               ‾‾‾‾‾‾‾‾‾‾‾‾‾‾
    <----R--->

Sketch of the domain

Meshing

Workflow (see the file run.sh):

  1. Mesh the domain, see Cubit file domain.jou
  2. The journal file domain.jou from (1.) is used as an input.
  3. An ANSYS-cdb mesh file domain.cdb is created.
  4. You can also save the created geometry as domain.cub5 to open it directly on Cubit (optional).
  5. Look into the journal file domain.jou to see the Cubit commands to create the mesh.
  6. These commands could also have been run interactively in Cubit.

The file domain.cdb was created this way.

Simulation with CFS

Use an XML-editor (e.g. oXygen or eclipse) to define the simulation input for CFS.

For all the cases, we have the following generic procedure:

The input file (transient.xml) is the simulation input. In the file (mat_acou.xml) the material porperties are defined.

To start the computation run the following command in the terminal 

cfs -p transient.xml job
where job can be any name you choose for the simulation.

CFS will write some output on the terminal, and produce the two files * job.info.xml, which contains some details about the run, and * job.cfs in the results_hdf5 directory, which you can view with ParaView.

Postprocessing

  1. collect all sensor-array files to one by using combineSensorArray.py
  2. look at the field result in ParaView (load pipeline from post.pvsm)

Further Suggestions

Think about the following questions:

  • Time of flight: how much time does the signal take from the excitation to a microphone location in the center of the domain?
  • Amplitude decay: A wave with a pressure amplitude of 1 is excited, why is the amplitude in the center of the domain smaller?
  • What happens when the wave reaches the boundary of the domain? Apply absorbing boundary conditions (ABC) and compare the results.
  • How coarse can the mesh be? How large can the timestep get? Conduct a convergence study ...