Learning Forum

[mmiah]

Hello,

I am using LSDYNA to model (3D) appropriate interactions between various parts of a floating fluid-solid system as shown in the attached graphic. A rigid body with an air void is floating in water and the air is trapped inside the void space of the floating object. Both water and air are modeled as ALE multi material group. Can you advice how I should define the following contact interactions?

1. water and tank - I am using LAGRANGE_IN_SOLID with PFAC=0.001. Is it appropriate?
2. water and the float - same as above
3. water and entrained air - EULER_IN_EULER
4. floating object and entrained air - This is needed to gravity initialize the system as the weight of the floating body will compress the entrained air. Also, there will be a vertical seismic excitation on the system which will cause interactions among various components. I am not sure what types of interaction I should use. I would greatly appreciate any suggestions.

Thank you,
Mamun

[Ian Do]

Hello Mamun,

Coupling is too involved to explain via this medium! I suggest you try to review my ALE class in ALH (ANSYS Learning Hub). It's difficult, but I'll give it a shot here:

Q1. water and tank - I am using LAGRANGE_IN_SOLID with PFAC=0.001. Is it appropriate?

A1: Always start with the default - this works for many cases - only change its value (PFAC) by 10-30% at a time, only change 1 value at a time.  

Q2. water and the float - same as above

A2: ?

Q3. water and entrained air - EULER_IN_EULER

A3: No need - Think of the ALE computational domain or mesh as containing 2 mats. If 1 mat moves into an elm, the other has to be pushed out. This is automatic. 

Q4. floating object and entrained air - This is needed to gravity initialize the system as the weight of the floating body will compress the entrained air. Also, there will be a vertical seismic excitation on the system which will cause interactions among various components. I am not sure what types of interaction I should use. I would greatly appreciate any suggestions.

A4: I would suggest you learn S-ALE approach, newer, better, simpler approach, but this is complex and cannot be taught in this forum. It will require a class! 

Since your questions are set in the traditional ALE approach, I will answer the way you ask.

• define LAG part, H1, say, (& it's better to make it deformable, ex; MAT_ELASTIC ...)
• define the total ALE domain/mesh = H2 = background mesh = outside air MMG3
• use volume filling KW to fill in the water space (MMG1, using a simple box) and the trapped air (MMG2, using the segment-set of the internal surface of the cavity, call this segment-set1)
• define coupling card 1: MCOUP==> MMG1 is coupled to H1 (CONSTRAINED_LAGRANGE_IN_SOLID ...)
• define coupling card 2: MCOUP==> MMG2 is coupled to the segment-set1

I hope this helps.

Ian Do

 

[mmiah]

Dear Ian,

Thank you so much for providing your expert suggestions. Let me implement those and get back to you. I will also look into the S-ALE approach.

Regards,
Mamun

[mmiah]

Hi Ian,

I have tried to follow your instructions on volume filling KW. My LS-PrePost (v4.10.3) shows this KW is available only for S-ALE. So I converted my model from ALE to S-ALE using STRUCTURED_MESH KW. Then I defined MMG1, MMG2, and the coupling cards as you suggested. When I save and re-run my model I get the following error messages (multiple of them):

Error 10267 (KEY+267)

in converting external to internal
ELE-ID 61912 N7 68934

Also, for volume filling should I use a box that is equal to the size of the cavity and the segment-set that consists of 5 internal cavity surfaces?

Thanks again,
Mamun

 

 

 

 

 

[Ian Do]

Hi Maman,

I m away to a conference. Will be back tomorrow..make sure you offset the beginning nid & eid to values above the max already occupied by the lag sstructures.

vol fill:

use segment set oc box that fits the volume you want to fill. 

[mmiah]

Hi Ian,

I left the NID and EID values as defaults which set them automatically above their maximum numbers. Is it okay with you to take a look at the k file? That way we can reduce lots of back and forth communications and I can post the debugging finding here afterward for the community.

Regards,
Mamun

[Ian Do]

Hi Mamun,

Pls first try it with the latest beta exe. Pls contact our support for the proper link to it. And yes, if the latest exe gives the same error, you can send the model to me. FYI: I am under a lot of pressure for the next week or more. Thanks!

[mmiah]

Hi Ian, 

I have a question about the Volume Filling KW. Does it calculate the volume changes occupied by the individual fluid inside a cavity? Because the problem I am trying to model includes volume space changes with time by air and water due to dynamic seismic loading.

[Ian Do]

Hi Mamun,

If your ALE mesh is fine enough to resolve the motion of the LAG structure, and if you represent the trapped air properly and distinguish it from the air outside, then as your structure moves, the ALE elements will be updated to contain the proper volume fraction of whatever AMMG (ALE multi-material groups) that you have defined in the domain accordingly. This means that if the LAG structure moves 20 cm , say in the x dir, you will need enough ALE elements to span that 20 cm so that the AMMG displacements can be calculated and the effects reported. How many elements? That is up to your engineering judgment.

Ian