Re:Re:Re: <GEBT>, Converge err or

[Wenbin Yu]

Your original matrix is too ill-conditioned. The largest diagonal term with respect to smallest diagonal term is around 10^15 difference, up to the precision of the double precision. It will be a big challenging for convergence. Yes, double check the sectional analysis.


At 09:04 AM 1/18/2013, ì—†ì Œ wrote:

I checked the flexibility matrix.


Hlighted value is big value that I found.


     3.9565881219E-09    9.8010153627E-28    1.9179001045E-28   -9.5467383501E-29    8.4225928148E-18   -1.1990515532E-17

     9.8010153627E-28    6.5687024949E+02    1.6855752329E-01    5.5852821358E-12   -1.5154855432E-29    6.4307944888E-29

     1.9179001045E-28    1.6855752329E-01    1.2964923629E+02    2.4258851929E-13   -2.7741367658E-30    1.2820833120E-29

    -9.5467383501E-29    5.5852821358E-12    2.4258851929E-13    4.6951083495E-10    8.3932393413E-31   -7.3319786749E-30

     8.4225928148E-18   -1.5154855432E-29   -2.7741367658E-30    8.3932393413E-31    9.1714686830E-13    2.0226671596E-17

    -1.1990515532E-17    6.4307944888E-29    1.2820833120E-29   -7.3319786749E-30    2.0226671596E-17    1.0005281286E-11


I replaced new matrix below


     1.6338077951E-08    2.2352284827E-25    4.4333576268E-25    8.7627445715E-27    1.1898588887E-12   -2.1294335695E-12

     2.2352284827E-25    1.4270316109E-07    4.0630548095E-08    6.3106458140E-10   -2.7030908016E-26   -2.8960442769E-27

     4.4333576268E-25    4.0630548095E-08    1.1241927922E-06    2.0516037737E-09   -3.7625137825E-26   -7.0715804338E-27

     8.7627445715E-27    6.3106458140E-10    2.0516037737E-09    2.3472928930E-10   -7.4786930313E-28   -1.2898765435E-28

     1.1898588887E-12   -2.7030908016E-26   -3.7625137825E-26   -7.4786930313E-28    1.5833238933E-10    9.5761751329E-13

    -2.1294335695E-12   -2.8960442769E-27   -7.0715804338E-27   -1.2898765435E-28    9.5761751329E-13    4.9698634620E-12



So GEBT finished sucessfully.


It may be numerical problem. And my sectional analysis is fault.


Thanks.



Sincerely yours.


 

------ Original Message ------

Date: Friday, Jan 18, 2013 03:50:47 PM

From: "ì—†ì Œ" <bulb...@empas.com>

To: "Qi Wang" <aaron....@gmail.com>

Subject: Re:Re: , Converge error

Hi

Thank you for feedback.

I will update input file to hifi-comp and I will check matrix.

Jun Hwan.

 

------ Original Message ------

Date: Friday, Jan 18, 2013 02:23:05 PM

From: "Qi Wang" <aaron....@gmail.com>

To: <bulb...@empal.com>

Cc: <hifi...@googlegroups.com>

Subject: Re: , Converge error

Hi,

It seems like there is a numerical problem in the Newton-Raphson iteration. I checked with the input file in the attachment, and replace the flexibility matrix with another one and GEBT works well. Please double check the flexibility matrix for your problem. Or you can attach a problem description and email to this group.

Qi

On Thu, Jan 17, 2013 at 8:01 PM, Wenbin Yu < wenbiny...@gmail.com> wrote:

Can u help answer this question?

Begin forwarded message:

From: ì—†ì Œ <bulb...@empas.com>

Date: January 16, 2013, 1:15:20 AM MST

To: Wenbin Yu <wenb...@usu.edu>

Subject: , Converge error

Dear professor

In Hi-fi comp, I don't uploading file. It may be my computer problem.

My question is below.

------------------------------------------------------------------------------------------

Test model is multi-node and elements.

Error message is that "The solution does not converge after the maximum number of iterations"

In this case, is there any option of GEBT ?

Thanks

------------------------------------------

Jang, Jun Hwan

Mobile  :  010-7941-9331

E-mail  :  bulb...@empal.com

------------------------------------------