I want to do a coupled structural-acoustic analysis in abaqus, in which a cantilever beam vibrates freely due to initial conditions, while immersed in an unbounded air domain. I seek to obtain the acoustic pressure at selected nodes of the air domain as a function of time (I want to obtain the sound wave). To accomplish this I created two models. In the first model, I create the beam and preload it (I create the initial conditions). In the second model, I add the air domain to the first model, couple the air and beam parts with a tie constraint, release the preload applied on the structure in model 1 (thus letting it vibrate freely) and calculate the acoustic pressure at certain nodes in air domain. The problem I encounter is that when I view the results of the second model, there’s a gap between the air’s and beam’s tied surfaces, as shown in figure gap.png attached. My questions are:

1.      is this the right way to couple the structure to the air part when doing an import procedure (or should I include the air part in the static step, and import it too)? If not, what is the right way?

2.      How do I remove the gap, and why is it there? I think it has something to do with slave node adjustment. The data file says that 915 slave nodes were tied, 689 were adjusted (the adjustment vectors are insignificantly small), and 226 were not adjusted. Additionally, from the visualization module, displaying the TIEDSTATUS variable shows that all slave nodes were tied, so I don’t understand why there’s a gap.

3.      I calculated the acoustic pressure at nodes on opposite sides and equal distance from the beam (set MIC from job ASI_X_IMPD_1 in .cae file), are those pressures correct even though there’s a gap? They are apparently a mirror of each other (see MIC_POR.png attached).

I upload the .cae file, which is referenced in what follows, for your convenience. The gap can be seen in the deformed shape at increment = 0, step time = 0 of job ASI_X_IMPD_1. If you can't download the .cae file, there should be a little square where the file should be. Right click on the square and hit save link as.

The modeling steps I took in each model are described below:

First model: (named S_PRELOAD)

-        Create beam part and place boundary condition (encastre) to fix one end (set = FIXED).

-        Create static, general step (named STATIC)

-        Apply displacement boundary condition to the free end of the beam (set = BEAMTIP) in static step (in this way, I apply the initial condition).

-        Create job S_PRELOAD_1 and run it. This calculates the beam initial configuration.

Second model (named ASI_X_IMPD): I created this model by editing a copy of the previous model.

-        Copy previous model

-        Create a new part (the air part), and in assembly module place it so it surrounds the beam part.

-        Add nonreflective boundary conditions to the outer surface of the air part (to model the infinite medium)

-        Tie the surfaces of the two parts in contact with each other using tie constraint.

-        Replace static step with explicit dynamics step, and erase the displacement boundary condition defined in the previous model (so the beam can vibrate freely).

-        Import state of the beam part at the last step/frame of job S_PRELOAD_1 (I don’t update reference configuration).

-        Request history output to get the soundwave (the acoustic pressure for set = MIC)

-        Create analysis ASI_X_IMPD_1 and run it.

As a side note, I created job ASI_X_IMPD_1_Y where the only change I made from job ASI_X_IMPD_1 was that when importing results from job S_PRELOAD_1, I updated the reference configuration. However, when viewing the results, some slave nodes are not tied, and some of the ones tied are not tied correctly.

Also, for job ASI_X_IMPD_1 I requested the U2 variable as history output for the sets AIR_U2 (node in air part) and BEAM_U2 (node in beam part). These nodes should have the same spatial coordinates if the gap didn’t exist. The U2 variables are plotted in U2.png. As can be seen, both curves are equal but offset by a distance presumably equal to the gap size at the location of the two nodes.

Thanks for your time, I appreciate any help. Sorry if this was too long but I wanted to make myself understood because this is very important.