different results between move mesh and importing offset CAD geometry

DRQDRQ
Hi,
I've run a couple of analyses of a item that is rotating, but rotating off centre. I've tried two ways of applying this eccentricity and seem to get two very different sets of results.
The geometry is stacked cylinders, rotating about an axis parallel to the axis of the cylinders but offset by a small amount.
The model is the full geometry.
The only load is "centrifugal force". I thought this would calculate a force at each node based on planar distance from origin and mass of element(s) involved, in a direction directly away from the defined axis through the origin.

case 1
The first way I tried was to offset the geometry in CAD before importing it to Mecway.

case 2
The second way wat to import a CAD model that is centered, create a mesh and then use:
"Mesh tools -> Move/copy..."
to apply the same offset as case 1.

However I get very different results.

I've checked the centre of mass for both cases
"Tools -> Center of mass"
and it appears pretty close and shows the offset I was expected. The differences only be at many decimal places.

Am I doing something stupid?

Comments

  • JohnMJohnM
    That does sound a little strange. I would expect the meshes to be identical, but does the offset affect that somehow? Are the messages identical?
  • disladisla
    edited March 2023
    The Center of mass was a nice check ...Additionally:

    -Same Density/Mass?
    -Same Orientation in 3D Space?

    I would perform a fast Modal vibration analysis to see if they both agree. Request 12 modes and pay attention also to the first six and how small they are (<<1e-3).

    EDITED ¿How do those differences relate to the overall dimensions of the model.? Maybe it is a big diffrence between models but without any importance respect the model dimensions. 10E-8mm & 10E-10mm but in a model of 10mm size.
  • DRQDRQ
    edited March 2023
    JohnM
    12:34PM
    That does sound a little strange. I would expect the meshes to be identical, but does the offset affect that somehow? Are the messages identical?

    I hadn't thought to look.
    There is a difference in the quantities, but not a great deal.
    For example:
    Elements: 201727 vs 201439
    Nodes: 306363 vs 306107
    (figures for offset CAD vs figures for moved mesh)
    I guess since the mesher is starting with slightly different geometry then a slight difference is to be expected.
  • DRQDRQ
    disla
    1:32PM
    The Center of mass was a nice check ...Additionally:

    -Same Density/Mass?
    -Same Orientation in 3D Space?

    I would perform a fast Modal vibration analysis to see if they both agree. Request 12 modes and pay attention also to the first six and how small they are (

    Done, and I think that has pointed me toward what I was doing wrong, perhaps not in the way you thought it would though. Will put up another comment once I'm a bit more certain.
  • DRQDRQ
    It looks like I'd made the same error in both cases but that the error caused different results, or at least the results I looked at were different. And understandably so.

    I had failed to put a constraint in to eliminate the rotational degree of freedom, about the axis that the item will rotate about. A silly mistake. I overlooked it because that's the axis it will be rotating about, so it didn't even cross my mind that it's need to be stopped from rotating.

    Normally I do such things with sectors and use frictionless or fixed supports to impose cyclic repetition / symmetry but this time I was doing the whole geometry (because of the offset).

    I'm now re-running all my models to check and so far there is much better correlation between the ones with imported offset CAD and those where the offset was applied by moving the mesh.

    Many thanks you all who helped with this, and that includes those who looked at it and didn't post because they hadn't considered that I'd be so stupid as to overlook something so simple!
  • prop_designprop_design
    Not sure if this is what you are experiencing. However, a long time ago, I noticed Netgen yields different meshes based on where the part is located. I brought this up to Victor. As far as I know, that bug still exists.
  • DRQDRQ
    prop_design
    4:45PM
    Not sure if this is what you are experiencing. However, a long time ago, I noticed Netgen yields different meshes based on where the part is located. I brought this up to Victor. As far as I know, that bug still exists.


    I think that is occurring but not what was causing me the problem. This would give the different element/node counts mentioned above.

    For me, the mesh being different when the part is in a different place makes sense, and isn't a problem.
Sign In or Register to comment.

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In Register