[Getdp] example of fixed point iteration in GetDP

Frederic Trillaud ftrillaudp at gmail.com
Tue Jul 15 12:12:22 CEST 2014


Thanks

Best,


-- 
----------------------
Frederic Trillaud, PhD
Assistant professor (Investigador titular A, T.C.)
Coordinación Eléctrica y Computación
Instituto de Ingeniería
Universidad Nacional Autónoma de México (UNAM)
Circuito Escolar, Edificio 12, cub. 209
UNAM, C.U., Coyoacán, D.F. Mexico 04510
Ph.: (+52) 5556233600 ext. 8829
Em.: ftrillaudp at pumas.ii.unam.mx


On mar, 2014-07-15 at 11:08 +0200, Christophe Geuzaine wrote:
> On 15 Jul 2014, at 11:04, Ruth Vazquez Sabariego <Ruth.Sabariego at esat.kuleuven.be> wrote:
> 
> > I see. The demo I am talking about is indeed in the GetDP part: http://www.geuz.org/getdp/
> > 
> > I have a svn release and compile myself on a Mac. 
> > Just checked the precompiled version and it is there. 
> > In the Mac version, it is in the directory 
> > share/doc/getdp/demos
> > 
> 
> Or cf. http://onelab.info/wiki/Inductor : if you set Flag_NL_Newton_Raphson to 0 you will get a fixed point iteration.
> 
> 
> > Ruth
> > 
> > 
> > 
> > On 15 Jul 2014, at 10:54, Frederic Trillaud <ftrillaudp at gmail.com> wrote:
> > 
> >> Dear Ruth,
> >> 
> >> Thank you for your quick reply. I have the oneLab bundle Gmsh-Getdp
> >> installed on my linux box. However, it does not come with examples. Is
> >> the one mentioned can be found on the Getdp website or I should install
> >> Getdp to access it?
> >> 
> >> 
> >> Best,
> >> 
> >> 
> >> Frederic
> >> 
> >> 
> >> On mar, 2014-07-15 at 07:23 +0000, Ruth Vazquez Sabariego wrote:
> >>> Hi Frederic,
> >>> 
> >>> 
> >>> You can inspire yourself from the magnetostatic demo you get when
> >>> downloading GetDP.
> >>> 
> >>> 
> >>> There is a complete Newton-Raphson example, i.e. a term for completing
> >>> the Jacobian matrix is included (JacNL).
> >>> You can just remove or comment it for getting a fixed point iteration.
> >>> The resolution does not change.
> >>> 
> >>> 
> >>> HTH,
> >>> Ruth
> >>> 
> >>> 
> >>> 
> >>> 
> >>> 
> >>> 
> >>> 
> >>> 
> >>> 
> >>> On 14 Jul 2014, at 18:29, Frederic Trillaud Pighi
> >>> <ftrillaudp at pumas.iingen.unam.mx> wrote:
> >>> 
> >>>> Dear all,
> >>>> 
> >>>> 
> >>>> 
> >>>> I have a non-linear magnetodynamics problem that involves a fixed
> >>>> point iteration. Would someone have an example to share or some
> >>>> recommendation how to implement the algorithm?
> >>>> 
> >>>> 
> >>>> 
> >>>> The problem is the following (A formulation, sig: electrical
> >>>> conductivity, Je: applied current):
> >>>> 
> >>>> 
> >>>> 
> >>>> Curl[nu*Curl(A)]+sig*dA/dt = Je
> >>>> 
> >>>> 
> >>>> 
> >>>> with,
> >>>> 
> >>>> 
> >>>> 
> >>>> sig = f(-dA/dt+Je/sig])
> >>>> 
> >>>> 
> >>>> 
> >>>> 
> >>>> Best,
> >>>> 
> >>>> 
> >>>> 
> >>>> 
> >>>> Frederic
> >>>> 
> >>>> _______________________________________________
> >>>> getdp mailing list
> >>>> getdp at geuz.org
> >>>> http://www.geuz.org/mailman/listinfo/getdp
> >>> 
> >>> —
> >>> Prof. Ruth V. Sabariego
> >>> KU Leuven  
> >>> Dept. Electrical Engineering ESAT/Electa
> >>> http://www.esat.kuleuven.be/electa
> >>> 
> >>> 
> >>> 
> >>> 
> >>> 
> >>> 
> >>> 
> >>> 
> >>> _______________________________________________
> >>> getdp mailing list
> >>> getdp at geuz.org
> >>> http://www.geuz.org/mailman/listinfo/getdp
> >> 
> > 
> > —
> > Prof. Ruth V. Sabariego
> > KU Leuven  
> > Dept. Electrical Engineering ESAT/Electa
> > http://www.esat.kuleuven.be/electa
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > _______________________________________________
> > getdp mailing list
> > getdp at geuz.org
> > http://www.geuz.org/mailman/listinfo/getdp
>