Menu

Discussion Forum

New Discussion
Filter

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Linear Buckling no eigenvalue found error

Posted Oct 29, 2020, 12:38 a.m. EDT Structural & Acoustics Version 5.4 4 Replies

Chanho Jeong
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

I've tried a simple buckling test. But an error occured

-Feature: Eigenvalue Sover 1 (sol1/e1) No eigenvalues found.

The simulation is a really simple one. - Just a cylinder(radius=0.03m , Height=1m) - Structural steel material - Fixed bottom boundary and -10N (z direction) boundary load.

I'm trying the 3D model, Solid Mechanics, Linear Buckling. Can everyone tell me what the problem is? Or is it a Comsol version issue?

Thank you

4 Replies Last Post Nov 5, 2020, 12:12 a.m. EST
Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 4 years ago Oct 29, 2020, 1:41 p.m. EDT

The underlying reason is the symmetry of the structure, which leads to that there is a double eigenvalue with an eigenmode which can bend in an arbitrary direction in the xy-plane.

There are several possible solutions. Three of them are:

  • Select to solve for 2 (or more) eigenvalues instead of 1.
  • Set a manual value (>2) for the size of the Krylov space in the Eigenvalue Solver node.
  • Relax the relative tolerance to 1e-4 in the Eigenvalue Solver node.

The behavior is a bit odd; the solver should be able to handle this case. We will have to look into it.

-------------------
Henrik Sönnerlind
COMSOL
The underlying reason is the symmetry of the structure, which leads to that there is a double eigenvalue with an eigenmode which can bend in an arbitrary direction in the xy-plane. There are several possible solutions. Three of them are: * Select to solve for 2 (or more) eigenvalues instead of 1. * Set a manual value (>2) for the size of the Krylov space in the *Eigenvalue Solver* node. * Relax the relative tolerance to 1e-4 in the *Eigenvalue Solver* node. The behavior is a bit odd; the solver should be able to handle this case. We will have to look into it.
Chanho Jeong
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 4 years ago Oct 30, 2020, 2:05 a.m. EDT

Thaks a lot. By simply solving with 2 eigenvalues, the error dissapeared.

However there is one more issue about the result I can not understand. When I adjust the deformation scale factor to 1 the result is a lot weird and unrealistic. It looks more like 100000 times of the scale factor. It doesn't matter how much small load I apply (-0.000001 N/m^2 for example).

Is there any method i can solve this problem? Thank you.

Thaks a lot. By simply solving with 2 eigenvalues, the error dissapeared. However there is one more issue about the result I can not understand. When I adjust the deformation scale factor to 1 the result is a lot weird and unrealistic. It looks more like 100000 times of the scale factor. It doesn't matter how much small load I apply (-0.000001 N/m^2 for example). Is there any method i can solve this problem? Thank you.
Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 4 years ago Oct 30, 2020, 9:43 a.m. EDT

The amplitude of an eigenmode is per definition arbitrary and has no physical meaning.

You can control how the scaling is done by modifying Scaling of eigenvectors in the settings for Eigenvalue Solver.

-------------------
Henrik Sönnerlind
COMSOL
The amplitude of an eigenmode is per definition arbitrary and has no physical meaning. You can control how the scaling is done by modifying *Scaling of eigenvectors* in the settings for *Eigenvalue Solver*.
Chanho Jeong
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 4 years ago Nov 5, 2020, 12:12 a.m. EST

Thanks a lot. I really appreaciate your help.

So it sounds like I can not expect the exact deformed geometry visually. Anyway, there is still some phenomena I can't understand when performing buckling simulation.

The simulation is still simple. - Version 5.4, 3D model, Solid Mechanics, Linear Buckling. - Just a cylinder(radius=0.03m , Height=1m) - Structural steel material - Fixed bottom boundary - Top boundaty has prescribed displacement -0.1m z direction (x,y are fixed to 0) .

There isn't an error but the result is not acceptable First, checking the z direction deformation (w-Displacement field, Z component), the top boundary hasn't moved at all. Second, maybe becasue of the first reason, total displacement (solid.disp-Total displacement-m) also says that there isn't displacement for the top boundary.

Could you also tell me why this happen or any solution for this problem? Thank you

Thanks a lot. I really appreaciate your help. So it sounds like I can not expect the exact deformed geometry visually. Anyway, there is still some phenomena I can't understand when performing buckling simulation. The simulation is still simple. - Version 5.4, 3D model, Solid Mechanics, Linear Buckling. - Just a cylinder(radius=0.03m , Height=1m) - Structural steel material - Fixed bottom boundary - Top boundaty has prescribed displacement -0.1m z direction (x,y are fixed to 0) . There isn't an error but the result is not acceptable First, checking the z direction deformation (w-Displacement field, Z component), the top boundary hasn't moved at all. Second, maybe becasue of the first reason, total displacement (solid.disp-Total displacement-m) also says that there isn't displacement for the top boundary. Could you also tell me why this happen or any solution for this problem? Thank you

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.

Suggested Content