Cervenka Consulting Forums — Use iteration with lowest error (Problem Data in GiD)

Re: Use iteration with lowest error (Problem Data in GiD)

null@example.com (dpryl) — Wed, 22 May 2019 15:21:41 +0000

Yes.

Re: Use iteration with lowest error (Problem Data in GiD)

null@example.com (pavlo) — Wed, 22 May 2019 06:04:38 +0000

Dear Dobromil,

There is the following information in ATENA Input File Format:

For n>0 and the iterating process within the
current step does not yield a converged solution,
then the solution is reverted to the best converged
iteration based on the convergence criteria n1, n2,
...
For n=0 the use of best iteration is reset to not
using best_iteration feature.
If divergence step's (or iteration's) stop criteria
are met, the current step is marked as nonconverged.
When this option is combined with
STEP_LOAD_REDUCTION_ALLOWANCE n,
then the iteration is reverted only when (nnumber
of attempts to revert the current step)=0.
By default n=0, i.e. this feature is N/A and v=1. ,
i.e. the step is marked as not converged step.

If I understood correctly, I can write in "Use best iteration for criterion" numbers from 1 to 4 combining them with single space. Where 1 to 4 are Displacement error, Residual error, Abs. Res. Err., Energy Err. Am I right?

Re: Use iteration with lowest error (Problem Data in GiD)

null@example.com (dpryl) — Tue, 21 May 2019 15:03:31 +0000

Dear Pavlo, if I remember correctly, there should be at least some basic information in the hep bubbles in the dialog? For more information, please see the ATENA Input File Format manual - options USE_BEST_ITERATION_FOR_CRITERIA and BEST_ITERATION_MIN_ID (Table 7: &SET command parameters).

Use iteration with lowest error (Problem Data in GiD)

null@example.com (pavlo) — Tue, 21 May 2019 14:43:08 +0000

Hello!
what should I write in space "Use best iteration for criterion"? It is "2 3 4" by default. What does it mean?
I remember that it was explained that here I can define how many last iterations can be considered to select. This is why I assume that only simple number can be defined rather than list of numbers...