cyl.axis

[Christophe Geuzaine]

Samuel Kvasnica wrote:
> 
> Christophe Geuzaine wrote:
> 
> > OK. Could you try to put "Dimension 2" as an option in the
> > "PostProcessing" field? (It forces the search algorithm to look after
> > two dimensional elements.)
> 
> YES, this really helps, but I wouldn't find it out without your help. MANY THANKS !!!

OK, I found the bug in the search algorithm. In future versions, the
effect of "Dimension 2" will be the default.

> 
> Well, there's still that problem with smoothness. Could you give me some link to
> effective smoothing algoritms that retain divB=0 ? I was trying to compute divB, use
> Poisson
> solver on  grid to compute potential from divB and substract it from B, but results were
> not
> very satisfactory.

If you just want a smooth solution on a curve (I suppose it is the one
on the symmetry axis), you can use gnuplot. I use it regularly to
present smoothed results (when I _know_ that due to the approximation
spaces my criterion on accuracy is fullfiled, and that the audience
really wants to see smooth curves...). Given a file 'test' containing my
data, I use something like

gnuplot> plot 'test' smooth cspline
gnuplot> set term table
gnuplot> set output 'new'
gnuplot> replot

which will output the new points in the file 'new'.

My best,

-- 
Christophe Geuzaine

Tel: +32-(0)4-366.37.10    mailto:Christophe.Geuzaine at ulg.ac.be
Fax: +32-(0)4-366.29.10    http://www.geuz.org