[CIG-SHORT] Stable time step for Quasi-static simulations

[Charles Williams]

Hi Josimar,

1.  For elastic materials, there is no restriction on the stable time step size for implicit (quasi-static) problems.  It is set to the maximum size of a float.  For other materials (e.g., viscoelastic), it is generally related to a fraction of the relaxation time.

2.  You can save all of the time steps by setting the skip value to 0 (see many of the examples in examples/3d/hex8, beginning with step04.cfg).

3.  The number of iterations for convergence may be related to the time step size, but the time step size restrictions are for reasons of accuracy, not convergence.  For a purely elastic problem, convergence is unlikely to be affected by time step size, although it may be affected by the load increment over the given time.

Cheers,
Charles


> On 30/04/2017, at 1:14 PM, Josimar Alves da Silva <jsilva.mit at gmail.com> wrote:
> 
> Dear PyLith developers,
> 
> I am going through the PyLith manual (version 2.2.0) but I have a question regarding the stable time step computation that is mentioned on page 63 and 64.
> 
> It is not clear for me how the stable time step is computed for the Quasi-static case. For the dynamic case it is clear. 
> 
> Assume a simulation that is quasi-static and all the materials are elastic, some questions that I have are:
> 
> 1) How is the stable time step computed ? Does it depend on the grid size ?
> 
> 2) I am under the impression that if the user request a time step that is larger than the stable one, then PyLith reverts it to the stable dt. Is it possible to save simulation results at every time step, including the stable ones, and not only the time steps requested on the .cfg file ?
> 
> 3) Does the number of iterations for convergence is related to the stable time step or to the user requested time step ? Example: if I run a simulation from time 0 to 100 years in time steps of 1 year; and then another simulation from 0 to 100 year in time step of 50 years. If the stable time step is the same then the number of iterations would  be the same for converge (if everything else if keep the same) ? 
> 
> 
> thank you in advance,
> Josimar 
> _______________________________________________
> CIG-SHORT mailing list
> CIG-SHORT at geodynamics.org
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-short

Charles Williams I Geodynamic Modeler
GNS Science I Te Pῡ Ao
1 Fairway Drive, Avalon 5010, PO Box 30368, Lower Hutt 5040, New Zealand
Ph 0064-4-570-4566 I Mob 0064-22-350-7326 I Fax 0064-4-570-4600
http://www.gns.cri.nz/ <http://www.gns.cri.nz/> I Email: C.Williams at gns.cri.nz <mailto:your.email at gns.cri.nz>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/cig-short/attachments/20170501/bdbe462d/attachment.html>