Welcome to the forums. Please post in English or French.

You are not logged in.

#1 2021-06-21 20:40:10

aep93
Member
Registered: 2020-02-21
Posts: 82

Attach point mass in a FSI problem

Hello,
I have a Fluid Solid interaction model in code_aster that contains 3D elements + 3D_fluid elements + FSI elements. I want to add a point mass to one of the nodes of FSI surface. How is this possible?
I tried to do it with M_T_D_N elements, however, it seems that the point mass does not change the results. Therefore, I am wondering whether I am doing the right thing.
Any help is appreciated.

Offline

#2 2021-06-27 12:40:06

CLF
Member
From: Sweden
Registered: 2009-01-27
Posts: 133
Website

Re: Attach point mass in a FSI problem

Hi

Are you dealing with fluid flow - structure interaction or, do you refer to a vibroacoustic analysis?

CA cannot do flow-structure, but is quite capable wrt vibroacoustics.

/C

Offline

#3 2021-06-27 19:24:08

aep93
Member
Registered: 2020-02-21
Posts: 82

Re: Attach point mass in a FSI problem

Thanks for your response CLF. I am dealing with vibroacoustics. I just want to know whether it is possible to use M_T_D_N point mass elements on one of the nodes of an FSI surface (FLUI_STRU elements)?

Last edited by aep93 (2021-06-27 19:24:34)

Offline

#4 2021-06-29 21:54:48

CLF
Member
From: Sweden
Registered: 2009-01-27
Posts: 133
Website

Re: Attach point mass in a FSI problem

Hi

No. You cannot.

Structure elements can only connect to other structure elements.

/Claes

Offline

#5 2021-06-29 23:18:30

aep93
Member
Registered: 2020-02-21
Posts: 82

Re: Attach point mass in a FSI problem

Thanks for your response CLF. However, I assume that FLUI_STRU elements are also considered as structural elements. I am saying this because these elements have degrees of freedom of DX, DY, DZ, and PHI as mentioned in following documentation:

u3.14.02

Am I missing something?
Thanks

Offline

#6 2021-07-01 15:22:17

CLF
Member
From: Sweden
Registered: 2009-01-27
Posts: 133
Website

Re: Attach point mass in a FSI problem

Hi

No, FLUI_STRU are Interface elements.

/Claes

Last edited by CLF (2021-07-01 15:22:49)

Offline

#7 2021-07-01 16:28:58

aep93
Member
Registered: 2020-02-21
Posts: 82

Re: Attach point mass in a FSI problem

Thanks for your response CLF. I think it is an important point. Is it also mentioned somewhere in the documentation that we cannot attach DIS_T elements to FLUI_STRU elements?
How can I attach a point mass to FLUI_STRU elements then?

Also one other point:
consider a node that is on the FLUI_STRU surface. This means that this node belongs to three elements: a structural element, a fluid element, and a flui_stru element that actually connects the structure and fluid. It means that structural elements still exist there. I am wondering why it is not possible to assign a point mass using dis_t elements to that point.

Last edited by aep93 (2021-07-02 02:28:00)

Offline

#8 2021-07-02 15:38:48

CLF
Member
From: Sweden
Registered: 2009-01-27
Posts: 133
Website

Re: Attach point mass in a FSI problem

Hi

You should not attach other than matching structure/acoustic elements to a FLUI_STRU interface element.

To clarify, a 2D shell connects to the 2D FLUI_STRU, which in turn connects to the 2D Face of an acoustic 3D element. This is an element matching.

You connect beams and lumped elements to the structure shell nodes.

/Claes

Last edited by CLF (2021-07-02 21:17:43)

Offline

#9 2021-07-06 06:28:38

aep93
Member
Registered: 2020-02-21
Posts: 82

Re: Attach point mass in a FSI problem

Thanks for your response CLF. Do you have any suggestions on how to attach a point mass to a volumetric solid structure that is in contact with a fluid then?
Maybe I should attach the mass to one of the nodes of the volumetric solid just below the Flui_stru surface so that it is not on the Flui_stru surface anymore. Do you think it is a good suggestion?
Thanks in advance.

Offline

#10 2021-07-08 06:20:12

aep93
Member
Registered: 2020-02-21
Posts: 82

Re: Attach point mass in a FSI problem

Hello,
I tried to create a sample case for this question and I have attached the code here. I have attached a M_T_D_N (dis_t) element to a surface that has FLUI_STRU elements and the code runs without any errors (returns 0). I am wondering why there are no errors if it is not possible to attach a dis_t element to a FSI surface?
Should I consider the results of this simulation to be wrong because of attaching a dis_t element to a FLUI_STRU element?


Attachments:
codes.zip, Size: 159.56 KiB, Downloads: 25

Offline

#11 2021-07-18 08:50:20

CLF
Member
From: Sweden
Registered: 2009-01-27
Posts: 133
Website

Re: Attach point mass in a FSI problem

Hi

At present, I have no way to run your problem, so I can just look at your model definition.

The topology in CA looks like this.

Acoustic element, eg 3D with 2D Face : connects to : FLUI_STRU : connects to : 2D structural element or a 2D Face of a 3D structural element. Differently stated, CA uses an element matching procedure.

Therefore, there must be a 1:1:1 2DFace:FLUI_STRU:2D structural element/face matching, usually achieved by copying element groups.

To connect a 0D element like a mass or spring, it is connected to a structural element node.

Looking at your AFFE_MODELE, you define a FLUI_STRU group called FSI, a 3D element group called solid, but no 2D Face (e.g. named solid_FSI) of these elements, i.e. you leave CA clueless on which structural Face elements to connect to your FSI group unless you define also 2D faces of your 3D solid.

The 3D_FLUIDE has fluid, walls and bottom groups defined. I assume walls and bottom are 2D Face elements. However, these groups do not seem to be identical with the FSI elements group, as one is used as velocity excitation and the other is fitted with an acoustic impedance boundary condition. It seems that you need to define a 2D Face group of the 3D_FLUIDE elements with elements that are identical to the FSI group (e.g. named fluid_FSI).

The simplest way to define an interface, usually is to copy a structure or acoustic group and define it as the FLUI_STRU elements using CA element copying commands, i.e. it can be done in the *.comm file and does not have to happen in the pre-processor.

So, all in all,
1) if you define 2D face elements of the 3D structure elements that are 'wet'
2) 2D face elements of the 3D_Fluide that meet structure
3) The 2D Face elements are identical to your FLUI_STRU elements
4) You add these definitions to your AFFE_MODELE and Material definitions
, then there is a chance that your script might execute as CA then understands where to connect the structure domain to the acoustic domain.

Last - a piece of advice - it usually pays off to have very clear group names when using CA.

So, application of strict define naming rules for your models make it easier for you to see if a step in the model definition is wrong or forgotten.

Hope this helps
Claes

Last edited by CLF (2021-07-18 16:41:28)

Offline

#12 2021-07-19 20:37:44

aep93
Member
Registered: 2020-02-21
Posts: 82

Re: Attach point mass in a FSI problem

Thanks for your explanations CLF. I have a question regarding the solid_FSI and fluid_FSI elements that you mentioned. My mesh is conformal (elements share nodes in their contact region) and the FSI is the surface between fluid and solid elements. So, if I want to define some face elements for the fluid to solid part, it will be exactly the same face as my current FSI face. As far as I understood, the same method was used in test case fdnv100a. Can you please let me know whether I understood what you mean correctly or not?
Thanks in advance.

Offline

#13 2021-07-20 01:03:46

CLF
Member
From: Sweden
Registered: 2009-01-27
Posts: 133
Website

Re: Attach point mass in a FSI problem

Hi

I hope I get this right (from memory).

Identical meshes, different nodes for acoustic/structure.

The FLUI_STRU can be a copy of either group, ie I think it can share nodes from either group, but I am not quite sure wrt this.

Suggestion - start with something simple, eg a box of solid 3D acoustic elements. Use one side as Face, copy this face mesh for a FLUI_STRU and for a Shell. You can make the shell simply supported and excite a node on the shell with a force.

Yes, you should find example cases. (I haven't looked them up).

/C

Last edited by CLF (2021-07-20 01:05:18)

Offline

#14 2021-07-21 16:56:42

aep93
Member
Registered: 2020-02-21
Posts: 82

Re: Attach point mass in a FSI problem

Thanks for your response, I will try that. However, I have one question: it seems that in the test cases (such as fdnv100a that I have attached here), the workflow is not similar to what you mentioned. I mean in this case, they only define 3D solid and fluid elements and then define FSI elements without having a surface for fluid or solid. Can you please look at this test case that I have attached?
Thanks in advance.


Attachments:
codes2.zip, Size: 355.71 KiB, Downloads: 12

Offline

#15 2021-07-21 19:23:59

CLF
Member
From: Sweden
Registered: 2009-01-27
Posts: 133
Website

Re: Attach point mass in a FSI problem

Hi

MODELE=AFFE_MODELE(MAILLAGE=MAILLA2,
                   AFFE=(_F(GROUP_MA='POI1',
                            PHENOMENE='MECANIQUE',
                            MODELISATION='DIS_T',),
                           _F(GROUP_MA=('SURFLIBR',),
                              PHENOMENE='MECANIQUE',
                              MODELISATION='2D_FLUI_PESA',),
                         _F(GROUP_MA=('FLUID1','FLUID2','FACELATE',),
                            PHENOMENE='MECANIQUE',
                            MODELISATION='3D_FLUIDE',),
                         _F(GROUP_MA=('IFLUSTRU',),
                            PHENOMENE='MECANIQUE',
                            MODELISATION='FLUI_STRU',),
                         _F(GROUP_MA=('PLAQUE',),
                            PHENOMENE='MECANIQUE',
                            MODELISATION='3D',),
                         _F(GROUP_MA=('FONDS',),
                            PHENOMENE='MECANIQUE',
                            MODELISATION='DKT',),),);

You need to check documentation, but my guess is

'FLUIDx' : 'IFLUSTRU' : 'FONDS'

'DKT' ('FONDS') is a shell element. Might be 'FLUID1' or 'FLUID2' that is a Face element

The above demonstrate the usefulness of having a strict naming convention in CA.

Have fun
C

Last edited by CLF (2021-07-21 19:25:37)

Offline

#16 2021-07-22 21:53:23

aep93
Member
Registered: 2020-02-21
Posts: 82

Re: Attach point mass in a FSI problem

Thanks for your response CLF. I checked the mesh of the fdnv100a case study again. Based on the mesh, the 'FLUID1','FLUID2' are all volumetric elements and therefore. Also, 'SURFLIBR' and 'FACELATE' are surface elements but they are not connected to 'IFLUSTRU' that is related to FSI elements. Thus, it seems that there is no fluid face that is connected with the FSI surface.
Regarding solid elements, DKT is a shell element and is connected to a part FSI surface, however, other parts of FSI surface in the model that are only connected to a solid volume (name of group of solid volume is PLAQUE) and there are no solid faces between the FSI surface and solid elements.
This is why I am saying the workflow is here not similar to what you mentioned.
Unfortunately, the details of the names of groups do not exist in the documentation. But you can find them in the mesh file of this study that is attached here.


Attachments:
codes2.zip, Size: 355.71 KiB, Downloads: 9

Offline

#17 2021-07-27 14:59:30

CLF
Member
From: Sweden
Registered: 2009-01-27
Posts: 133
Website

Re: Attach point mass in a FSI problem

Hi

If you follow my workflow, CA works.

I suggest that you first give it a try.

/C

Offline