Solid elements rotations in FEA

[HA649]

Hello the respected engineering community,
I am currently solving a linear static anlaysis problem in Femap, NASTRAN with the aim of getting rotations of solid elements after structural deformation just like I get that in shell elements with no success. I know the fact that solid elements have only translations in their formulation but as I am able to get shear strains of the solid elements which are based on the rotation of the elements I think there should be a way to get rotations in solid elements once it get deformed. Any guidance will be highly appreciated.

 

[WKTaylor]

Recommend re-posting this to the NASTRAN forum..

Nastran

www.eng-tips.com

 

[3DDave]

I think it is because rotation doesn't make sense with respect to solid elements. The nodes translate; being points they cannot individually rotate. The result does deform, but the deformation is the result of the translation of the nodes. Rotation only makes sense for elements that cannot deform, which would not resolve any stresses or strains.

 

[dcascap]

First of all, you can get the nodal rotations, but there is not really element rotation.

As Dave already mentioned, for solid elements the rotational DOFs at the nodes don't make sense due to the solid formulation as they don't have bending stiffness.

In a nutshell, for plate elements you have plate theory, and you can derive the FE element formulation where you have bending stiffness for the rotation of the nodes. Solids, the rotation of the structure comes from the x,y,z translation of your nodes, and the strains are calculated at the integration points of your element using the shape functions derivatives, which only depends on the node displacements.

 

[rb1957]

are you asking about the rigid body rotation of the element ? how the element CID would rotate (slightly) as the structure deforms ?

 

[HA649]

the strains are calculated at the integration points of your element using the shape functions derivatives, which only depends on the node displacements

Thanks for your response,
I am concerned with these derivates evalutation in Femap, NASTRAN by which the strains are evaluated as these derivates also represent rotation of elements (Shear strain is angular distortion of elements) once deformed in linear static analysis. When I plot R1, R2 & R3 rotations in femap the results become void for solid elements which I think due to inherent absenece of rotational DOF's.

 

[HA649]

are you asking about the rigid body rotation of the element ? how the element CID would rotate (slightly) as the structure deforms ?

Thanks for your response,
I am asking about the rotations of elements used to evaluate the shear strains of solid elements. I wanna calculate them in femap software just like I get those for plate elements.

 

[a27pak]

First of all, you can get the nodal rotations, but there is not really element rotation.

As Dave already mentioned, for solid elements the rotational DOFs at the nodes don't make sense due to the solid formulation as they don't have bending stiffness.

In a nutshell, for plate elements you have plate theory, and you can derive the FE element formulation where you have bending stiffness for the rotation of the nodes. Solids, the rotation of the structure comes from the x,y,z translation of your nodes, and the strains are calculated at the integration points of your element using the shape functions derivatives, which only depends on the node

As per standard element formulations, I understand that solid elements possess only translational degrees of freedom (DOF 1, 2, and 3), unlike shell and beam elements, which also include rotational DOFs.
However, since solid elements do exhibit shear strains, and shear strains are inherently linked to rotational behavior from a theoretical standpoint, it suggests that some form of rotational information is being internally computed during the analysis.

I think the question posted by user is about the reason why there is no provision to extract or output rotational results from solid elements , even if they are derived quantities and not primary DOFs?

If so, I would appreciate guidance on how to access or visualize this data within the software.

 

[3DDave]

Here's how solid elements are solved, without ever calculating a rotation or needing to:

 

[dcascap]

Thanks for your response,
I am concerned with these derivates evalutation in Femap, NASTRAN by which the strains are evaluated as these derivates also represent rotation of elements (Shear strain is angular distortion of elements) once deformed in linear static analysis. When I plot R1, R2 & R3 rotations in femap the results become void for solid elements which I think due to inherent absenece of rotational DOF's.

Yes it is as the nodes are freely to rotate so the stay in 0.
Femap and Nastran only plots the Nodal translations and rotations. There is no element translation or rotation.

As Dave said, you don't really solve rotations to calculate strains, as the strain is mathematically a derivative, and is not really based on an angle, even though they are related in principle.

I've been working with Nastran and Femap for quite a while now, and I don't think you can't post process what you're asking directly. You would need to create a tool for it using the api or in python or other coding language...

May I ask why do you need angles? Because from a solid mechanics perspective it doesn't really make sense for solid elements (you already have the strains which you can relate to most mechanical tensors...)

 

[rb1957]

"I am asking about the rotations of elements used to evaluate the shear strains of solid elements."

Ok, you'll need to deconstruct this logic for us (well, for me at least).

1) why do you think that rotation of the elements causes shear strains ? (I've "cue bid" the answer to this by saying that these rotations are "rigid body motion".)

2) why do you think this strain energy isn't captured in the translational strains, as calculated ?

3) why do you think you've uncovered something that many decades of analysis with 3D elements hasn't realised has been overlooked ?

 

[ThomasH]

I know the fact that solid elements have only translations in their formulation but as I am able to get shear strains of the solid elements which are based on the rotation of the elements I think there should be a way to get rotations in solid elements once it get deformed.

I don't know what exactly you are trying to do. But this may help you a bit on the way.

If you start with a rectangle with two parallel line pairs and four 90-degree angles, that is the undeformed element. If you now change the angles so that two will be larger than 90 degrees and two will be smaller, the opposite sides are still parallel. I would say that this deformation describes the shear, it is not the nodal rotations.

The angles give you the shear and that can be calculated from the translations. No rotations needed.

Please correct me if I am wrong .