Rigid Body Modes with Parabolic Mesh?

[Anonymous]

Simple 2-part assembly.  1 part fully-constrained, the other part has no constraints or bonding (by intention to test rigid body modes).

Works fine with linear mesh - first 6 modes are <=0.  Setting the mesh to parabolic results in fatal error E5001.

[Anonymous]

Simple 2-part assembly.  1 part fully-constrained, the other part has no constraints or bonding (by intention to test rigid body modes).

Works fine with linear mesh - first 6 modes are <=0.  Setting the mesh to parabolic results in fatal error E5001.

[John_Holtz]

Hi Mark,

I will take a look at the model to see what I can figure out.

[John_Holtz]

Hi Mark,

I will take a look at the model to see what I can figure out.

[Anonymous]

Thanks, John.  Marwan just emailed me and confirmed it is an issue.

[Anonymous]

Thanks, John.  Marwan just emailed me and confirmed it is an issue.

[John_Holtz]

Just to explain the model to other readers, the assembly consists of a cantilever beam (fully fixed on one end) and a block. As Mark explained, the block is not connected to the beam -- it is completely free. Therefore, a modal analysis should detect 6 rigid body modes at a frequency of 0. Mode 7 should be the first vibration mode of the beam, mode 8 would be the second vibration mode, and so on.

Here are a few things I found.

1. The error would occur when using parabolic elements at some mesh sizes but not other mesh sizes.
2. When the analysis would run (such as at a mesh size of 0.1), some of the six rigid-body modes were skipped. Modes 1-4 were the rigid body motion of the block. Mode 5 was the "first" vibration mode of the beam.

The work around is to set the "Modal Setup > Extraction Method" to "Subspace Iteration". When this is done, the analysis runs successfully with the parabolic elements.

(edited. I noticed that the description of "thing 2" was incomplete )

[John_Holtz]

Just to explain the model to other readers, the assembly consists of a cantilever beam (fully fixed on one end) and a block. As Mark explained, the block is not connected to the beam -- it is completely free. Therefore, a modal analysis should detect 6 rigid body modes at a frequency of 0. Mode 7 should be the first vibration mode of the beam, mode 8 would be the second vibration mode, and so on.

Here are a few things I found.

1. The error would occur when using parabolic elements at some mesh sizes but not other mesh sizes.
2. When the analysis would run (such as at a mesh size of 0.1), some of the six rigid-body modes were skipped. Modes 1-4 were the rigid body motion of the block. Mode 5 was the "first" vibration mode of the beam.

The work around is to set the "Modal Setup > Extraction Method" to "Subspace Iteration". When this is done, the analysis runs successfully with the parabolic elements.

(edited. I noticed that the description of "thing 2" was incomplete )

[Anonymous]

Should subspace be the default method then if the parameter RIGIDBODYMODES is set to AUTO or FORCED?

[Anonymous]

Should subspace be the default method then if the parameter RIGIDBODYMODES is set to AUTO or FORCED?

[John_Holtz]

Hi Mark,

I just wanted to make sure the work-around of using the subspace iteration worked for you (and more importantly, for the real model). If so, I will mark this as "solved".

As for the program defaults, I think the developers will attempt to fix whatever the problem is with the default solver rather than using the subspace method.

[John_Holtz]

Hi Mark,

I just wanted to make sure the work-around of using the subspace iteration worked for you (and more importantly, for the real model). If so, I will mark this as "solved".

As for the program defaults, I think the developers will attempt to fix whatever the problem is with the default solver rather than using the subspace method.

[jayesh.shinde]

Hi @John_Holtz,

I am facing similar issue in linear static analysis. I am doing analysis for Pipe (midsurface) with flange (solid). Is there setting in linear static with parabolic mesh?

Thanks, Jayesh

[jayesh.shinde]

Hi @John_Holtz,

I am facing similar issue in linear static analysis. I am doing analysis for Pipe (midsurface) with flange (solid). Is there setting in linear static with parabolic mesh?

Thanks, Jayesh

[John_Holtz]

Hi @jayesh.shinde 

I think the only similarity is the error message. Linear Static versus Normal Modes are completely different analysis types, and the error message is caused by completely different reasons. See my reply to your post https://forums.autodesk.com/t5/inventor-nastran-forum/offset-surface-of-pipe-to-solid-pipe-flange-co....

John

[John_Holtz]

Hi @jayesh.shinde 

I think the only similarity is the error message. Linear Static versus Normal Modes are completely different analysis types, and the error message is caused by completely different reasons. See my reply to your post https://forums.autodesk.com/t5/inventor-nastran-forum/offset-surface-of-pipe-to-solid-pipe-flange-co....

John