Dear Rui,

Does changing 'L' to a floating point value (40.0 instead of 40)
improve the first script?

Andy


On Fri, Mar 7, 2014 at 7:48 PM, rui tu <[email protected]> wrote:

> Dear FiPy group,
>
> I am trying to solve a very simple Cahn-Hilliard equation in FiPy, where I
> am using a scaling parameter, L, to adjust the dimensions of the domain.
>  When I set the length, L = 40, the equation cannot be solved (test1.py).
>
>  However, t if I set L = 0.4 e-5.*Magnification
>
> where
>
>         Magnification = 1.0e7
>
> which numerically is ALSO equal to 40, the equation becomes
> solvable(test2.py). I am baffled by this, and cannot understand why one
> simulation works and the other one doesn't.
>
> By typing,
>
>         diff test1.py test2.py
>
> in a shell terminal you can find:
>
>     16c16,17
>     < L = 40 # dimension of simulation domain
>     ---
>     > Magnification = 1.0e7 # magnification
>     > L = 0.4e-5*Magnification # dimension of simulation domain
>
> which proves that the both have the same L value. The only difference is
> how it is being inputed. I am attaching both scripts, test1.py and test2.py
>
> I would appreciate your help.
>
> Thanks,
> Rui
>
> _______________________________________________
> fipy mailing list
> [email protected]
> http://www.ctcms.nist.gov/fipy
>   [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
>
>


-- 
--------------------
Andrew Reeve
School of Earth and Climate Sciences
University of Maine
Orono, ME 04469
_______________________________________________
fipy mailing list
[email protected]
http://www.ctcms.nist.gov/fipy
  [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]

Reply via email to