[Getdp] Solution precision

[Uri Cohen]

I use gmsh and getdp to solve a problem in electrostatics, calculating
electric field on several meshes which differ only slightly from one
another. Each of those meshes is generated and then the electric field on
it with the same boundary conditions is solved using independent run of
getdp.
I expected the electric field in specific location of the meshes to change
smoothly across the different meshes but I see a non-smooth, non-monotonic
change.

I thought it may be because I don't control the precision in field
calculation and those may represent numerical errors. Is there a way to
increase the precision of the calculation?

Any other ideas on how to recover the smooth change I expect?

Thanks in advance,
            Uri Cohen,
            Hebrew University of Jerusalem
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.geuz.org/pipermail/getdp/attachments/20141007/d722de2c/attachment.html>