Acoustic-structural interaction in Abaqus

[alishst]

Dear all,

I'm trying to model a liquid storage tank using abaqus. The tank is anchored at the bottom and I've modeled the liquid (water) with acoustic elements. I am trying to verify my model by comparing it to results of previous research and experimental studies so I need to extract certain natural periods of the system. Now I'm facing two problems:

1) I have applied a free surface boundary condition here but it seems to have no impact on the analysis whatsoever. Even if I remove this interaction, the results don't seem to be changing at all.

2) Secondly, I'm not sure how to obtain modal values for certain modes. Such as first/second convective modes or first/second impulsive modes. Like, how would I know that which mode is second impulsive mode?

I'm relatively new to abaqus and I'd be grateful if you could look through my model and see what I should do about that. Here's the link:

https://files.engineering.com/getfile.aspx?folder=f766f43e-56bd-49d6-adf0-420a68a6f9d2&file=test.cae

 

[Erik Panos Kostson]

As per suggestions you seem to have applied the free surface as recommended by the abaqus manual (unfort. I have only developed code for some long range acoustics using abaqus as input and did not look on sloshing, but wave propagation). So I am not sure if anything else needs addition.

To verify your model against pure sloshing modes, do this:
If you make the walls very thick (say 50 mm or something), thus very stiff, then one can analytically calculate the pure sloshing modes in a tank (see Blevins for equations or google sloshing and tank or something), so do that and compare that to your model. Once that is done then you know it is at least working for stiff walls and with little fsi interaction (pure sloshing of water in a rigid tank). In this case also you do not need the application of hydro pressure on the walls since we assume that it is stiff, thus no stress stiffening effects are present and not accounted for since one is only looking at sloshing modes, and not coupled fsi modes.

Below is a simple example of a rigid water tank of a D:0.17 m and water H:0.1 m done as you set it up in abaqus, and a comparison to Strand7, giving the same results and as predcited by theory for the first sloshing mode (~ 2.3 Hz).

 

[alishst]

Thank you Erik! But there's one thing I didn't quite understand in regards with your suggestions. In the last paragraph, were you implying that the assumptions I used in my Abaqus model were in agreement with the Strand7 model you created?

 

[Erik Panos Kostson]

Yes, I modelled the example above as I have suggested thus free surface as per abaqus manual, and tie constraint for FSI, stiff tank walls, and that gave me the same answers as in Strand7 and theory, which is to be expected. So what I suggests is to do some examples like this where you know the answers (there is analytical equation for first sloshing modes in rigid cyl. tanks), just validation really which should always be done on FEA. That is it really.

 

[alishst]

Thank you again! I have found one example (for a flexible tank this time). I am trying to match my model with the results of that example but my problem is that Abaqus can yield so many modes and I am not sure which one to look at and compare with the example's results.

 

[Erik Panos Kostson]

I am not familiar with these terms. From a quick search (see here:

https://www.researchgate.net/figure...tive-mode-T-935-sec-b-second_fig3_228871006),

the mode I am calling first sloshing mode seems to be the first convective mode. Look on this site, and then look at the mode shapes/displacement distribution in your results (like I shown in the images above), and try to identify these modes.

 

[rafiq-fani]

Dear Ali, I have faced with this problem too, I'm modeling a liquid storage tank under seismic loads via Abaqus software and I need to obtain impulsive and convective modes from modal analysis, but I couldn't find a certain procedure. If you have found a way I would be grateful if you could help me.