Try creating an explicit reference axis at that edge. The simulation environment may not let you make the same axes selections as the geometry modeler.
note, this situation is fine for regular symmetry, circular should be used in cases where you have a tangential force
>>circular should be used in cases where you have a tangential force<<
Jared -- But it **did** work after I created an explicit axis of rotation.
Why do you say the above. I thought it was designed for any slice (no matter how convoluted) that can be patterned into the whole. -- John Willett
circular or more properly cyclic symmetry is for boundaries that are not flat. Nothing normal to the face (or the equivalent for 6 dof elements for planes of symmetry) will work for any flat planar boundary. Cyclic symmetry is used where they are curved or otherwise shaped. The condition is that the displacements on both surfaces are matched. In the later releases they changed it, for reasons that are perplexing for me, so that all the information is available in the post processing to make the reflected plots and they changed the naming to circular symmetry. They take a lot of their input from users that are not super familiar with FEA/structural analysis which can lead to some simplifications but often ends up suboptimal in my view at any rate. This being a case in point. They should have just had you reflect the plots in the post processing and left the symmetry as it was, in my humble opinion.
Bill wrote, "circular or more properly cyclic symmetry is for boundaries that are not flat... Cyclic symmetry is used where they are curved or otherwise shaped."
Jared wrote, "but again, in this case it is symmetry, best to use symmetry, not circular symmetry."
Attilio wrote, "I agree that cyclic, or any symmetry, for that matter should be well understood by the user before implementing,.. When in doubt initially model the full component, or some representative full model, understand the results, and then take advantage of symmetry."
Thanks to all for your comments. However I'm now more confused than ever:
Both Bill and Jared assert that circular (cyclic) symmetry does not apply to plane surfaces, but
1) It worked in the above example and has been verified by comparison to the full disk, as Attilio recommended. (Admittedly this is not a full test, as there are no tangential forces or displacements, only axial and, to a small extent, radial.)
2) SolidWorks Help implies (without explicitly stating) that Circular Symmetry will work for planar cuts:
The faces on the cut section. The faces should be similar in shape and area and can be non-planar. (In fact, the flywheel example shown there apparently has planar cuts.)
- and -
To set up your model for applying circular symmetry restraints, you create a cut (or an assembly cut feature) to produce a section of the model that can be repeated in a cyclical pattern about a central axis of revolution. A valid cut section produces a complete part, when it is repeated with an equal spacing circular pattern feature, and there are no gaps or interferences between the sections. You cannot apply a circular symmetry restraint to an invalid section.
3) Jared suggests that I should use ordinary symmetry in this case; but my cuts are at 60 deg, not 90, and SW will not allow ordinary symmetry planes that are not perpendicular (which makes sense to me).
4) Both Bill and Attilio rightly bemoan the fact that SW Help does not seem to explicitly define Circular Symmetry (nor ordinary Symmetry for that matter). I believe Bill has it right ("The condition is that the displacements on both surfaces are matched"). I'm not sure how the user would simulate this with explicit constraints, since the conditions are not local; but I'm pretty sure it is not the same as ordinary symmetry. I believe ordinary Symmetry requires displacement components normal to the symmetry planes to be zero and in-plane displacements to be matched. (I think this **could** be simulated explicitly with local restraints by zeroing the normal displacements and leaving the in-plane displacements free.) Not so? -- John Willett
cyclic will work it is just over kill but I can understand the confusion as it is now how you do it in Sim - use circular. Cyclic is a more general case. sounds like you have it more less figured out now.
i can't tell what part of that is a statement and what part is a question.
this might be something you want to speak to your reseller with.
1. setup the BCs manually, it works with any geometry, apply normal to restraint to the face that is cut
2. use cyclic symmetry in situations where you have cyclic load (patterned, tangential load)
3. use the symmetry BC when you have a quarter or half model, the restriction is so that it can pattern the results and has limits
>>this might be something you want to speak to your reseller with.<<
Thanks, Jared, for your patience and counsel. This might be the best suggestions of all. I'll see if they can find out for me exactly what are the SW definitions of the Symmetry and Circular Symmetry fixtures. I'll report back if I get anything useful. Cheers! -- John Willett
it is well defined in the help and also the swx kb
but I think having a phone conversation with someone vs text will be helpful in this case
I have a feeling you might be overthinking it a bit.
>>I'll report back if I get anything useful.<<
I think I understand now. Here's a summary. Please call out any errors or omissions:
1) I finally stumbled on a nice description of Symmetry constraints for Shells at http://help.solidworks.com/2015/english/SolidWorks/cworks/c_Symmetry_Restraints_for_Shell_Models.htm?id=a2f32dba1e9942359e6dee8e93aa498c#Pg0.
This link makes it clear that there are one displacement (normal to the symmetry plane) and two rotations (parallel to the symmetry plane) that must be set to zero. (The other motions are free.) "De-generalizing" this to solids (which have no rotational degrees for freedom) suggests that only the normal displacement is constrained. This assumption is now confirmed by Tech Support at my VAR.
The importance of the above information is that it allows analysis of non-perpendicular planes of symmetry (where the Symmetry constraint cannot be applied), such as an acute-angle "pie slice" of a circularly symmetric model with no tangential loads. A Roller/Slider fixture will do this conveniently for a solid model. For a shell model it's a little more difficult, as described above, but readily accomplished through Advanced Fixtures/Use Reference Geometry.
2) Tech Support at my VAR also confirmed my suspicion that there is no such work around for the non-local constraints of Circular (Cyclic) Symmetry. For a "pie slice" of a shell model (to which Circular Symmetry cannot be applied for unknown reasons) that does have tangential loads it is necessary either to convert to a solid model and then apply Circular Symmetry or to analyze the full circular shell model.
Case closed (unless anyone sees a problem). -- John Willett
looks good. i don't think i realized that you were stuck with shells.
but yes, there is no way to manually create cyclic symmetry with solids or shells
the software doesn't know what kind of load you're applying
so it assumes you know the appropriate BC. just like it will let you apply symmetry even if it isn't valid.
the only reason i mention this is i remember an issue a long ways back where you couldn't interchange them, could be a non issue. but again, in this case it is symmetry, best to use symmetry, not circular symmetry. you can do it manually or with the specific bC for it.
I agree that cyclic, or any symmetry, for that matter should be well understood by the user before implementing, There are many boundary conditions any FEA software will let you mimic that may not reflect design intent.
When in doubt initially model the full component, or some representative full model, understand the results, and then take advantage of symmetry. Then again I find myself doubting symmetry BCs every time I use them but that approach has served me well.