XANSYS Message: 38793 [Go back to message list]
[bookmark on del.icio.us]
No rating yet
Rate item:

Subject: Re: Superelement query
Author: Sheldon Imaoka
Date: 2002-06-13 12:31:00

Hi Raghavan,

Based on what you wrote, it seems like ANSYS is behaving as
expected. When you mentioned that, in the second
situation, 'coordinates coordinates of the generated nodes 1 to 5 are
replaced by those of the superelement master nodes,' this seems to
indicate that you are using same node numbering in different
locations (non-coincident nodes). ANSYS does not let one define
two 'Node 1' nodes since each node number has to be unique.

Also, are you using bottom-up or top-down substructuring? I was not
sure how you ensured correct node numbering in both cases if you were
doing bottom-up substructuring.

Lastly, regarding your BEAM188 vs BEAM44 difference, I wouldn't be
able to comment more without knowing the details of the structure and
why you expect the answers to show a certain trend.

Regards,
Sheldon

--- In xansys@y..., "Kotur, Srinivasan Raghavan (Export)"
wrote:
> Hello all,

> I have an ANSYS job which has two element types - PLANE82 and
MATRIX50
> (Superelements). Master nodes of the superelements are numbered 1
to 5.
> .Performance of this model is dependent on the order in which the
two types
> of elements are generated.

> If the superelements are generated first and plane82 later (using
AMESH) the
> job gets executed correctly. The generated nodes are replaced by
the master
> nodes of the superelement wherever they are within the specified
tolerance
> limit and the numbers of the master nodes are retained. Please note
that no
> explicit NUMMRG command is given.

> OTOH if PLANE82 elements are generated first and the superelements
later ,
> coordinates of the generated nodes 1 to 5 are replaced by those of
the
> superelement master nodes. This results in element distortion and
topology
> errors.

> Can someone please explain the reason for this diabolical behavior.
Is it an
> ANSYS bug or an inevitable programming necessity.

> Thanks in advance .. Cheers to all .. raghavan (INFOTECH
(India) / Pratt
> & Whitney USA).


Posts possibly associated with message #38793AuthorDateScore
38568Superelement queryK.S. Raghavan2002/06/06 
38573Re: Superelement queryRoberto Porto2002/06/06 
38575Re: Superelement queryK.S. Raghavan2002/06/06 
38578Re: Superelement queryChristopher Wright1969/12/31 
38599Re: Superelement queryK.S. Raghavan2002/06/07 
38604Re: Superelement queryChristopher Wright1969/12/31 
38611Re: Superelement queryRoberto Porto2002/06/07 
38793Re: Superelement querySheldon Imaoka2002/06/13 
38956Re: Superelement queryK.S. Raghavan2002/06/19 
38958Re: Superelement querySheldon Imaoka2002/06/19