XANSYS Message: 92973 [Go back to message list]
[bookmark on del.icio.us]
No rating yet
Rate item:

Subject: Re: [WB] Advice on modelling buckling of a pipe
Author: Y Y Tan
Date: 2008-04-23 16:58:05

Hi Martin & Mally,

First of all, thank you for responding so quickly to my post.

I'm not in such a good state now as I am rushing against my dissertation
deadline so please pardon me if my questions sound a little desperate as this
is my only source of advice.

I'm a little confused between what both of you mentioned. Martin said that
smaller time-step better convergence but Mally said that smaller time step
might lead to instabilities due to that particular time step hitting an
unstable solution. Can someone explain further? and also no one mentioned about
mesh size, does it mean that mesh (supposed an adequate mesh) is not important
in getting a converged solution?

Anyway I'm using the auto time stepping feature and I've tried having the
minimum substeps changed from 100 to around 300 and they still fail to converge
at about 90% of the loading.

Martin, I have employed the stabilize command before and I tried it with the
recommended constant energy stabilization method with the value equal to 1e-4
up to maybe 0.1 and it has not really been better. Once there was this really
weird thing that happen, I was using the stabilize command and the structure
buckled and when i remove the stabilize command the structure did not buckle at
all, doesnt make sense to me at all because the stabalize command was supposed
to damp the buckling not cause the buckling.

Martin, I've also checked the ds.dat file amd yea the Solsh 190 was used without
any changes to the keyopts meaning its using the pure displacement element
formulation. Does anyone think that this is affecting my results?

One last question, I applied a pertubation load to guide the structure to buckle
and say I wanted to remove this load once the structure starts to buckle. Can I
use a 2 load step analysis where in one load step I apply both the perturbation
load and the external pressure to cause the structure to buckle and in the
second load step I disable this perturbation load? Would this be a non-static
analysis?

Sorry for the long post guys. Hope you all can help me, just feel that my FEA
ability is really amateurish. Have been doing this for almost 2 months and
still cant get it to work. T_T. Again would like to thank you all in advance.


Best regards,
Tan Yi Yong
The University of Sheffield








Quoting "Mally, Carl" :

> >One thing I dun really understand is, how does time step size and mesh
> size affect the convergence of the solution. I always thought that time
> step size and mesh size only affects the accuracy of the solution and
> shouldn't affect the solution's convergence. Please could anyone help me
> clarify this.<
>
> Time step affects convergence in post buckling particularly if one of
> your time steps is at a point where the solution is unstable such as
> when the part is snapping through. There sometimes fewer timesteps
> will allow convergence because the solution will jump over the actual
> snap through to a stable solution on the other side. Generally in non
> linear analysi the opposite is true as Martin said.
>
> Carl Mally
> Senior Applications Engineer
> Centro Inc.
> 950 North Bend Drive
> North Liberty, IA 52317
>
>
>
>
>
> --------------------------------------------------------------------
> NOTICE: This e-mail (including attachments) is covered by the Electronics
> Communications Privacy Act, 18 U.S.C. 2510-2521, thus is confidential and may
> be legally privileged. The contents of this e-mail should not be forwarded
> to anyone without consent of the sender. If you are not the intended
> recipient, you are hereby notified that your retention, copying or any
> distribution of this communication is strictly prohibited. Please reply to
> the sender that you have received this e-mail message in error, then delete
> it. Thank you.
> ^--------------------------------------------------------
> | XANSYS - www.xansys.org |
> | The Discussion List for users of ANSYS, Inc. Software |
> | Hosted by PADT - www.padtinc.com |
> ^--------------------------------------------------------
>


^--------------------------------------------------------
| XANSYS - www.xansys.org |
| The Discussion List for users of ANSYS, Inc. Software |
| Hosted by PADT - www.padtinc.com |
^--------------------------------------------------------


Posts possibly associated with message #92973AuthorDateScore
92973Re: [WB] Advice on modelling buckling of a pipeY Y Tan 2008/04/23 
92980Re: [WB] Advice on modelling buckling of a pipeCarl Mally2008/04/23