[Getdp] Formulation term to express boundary constraint

Ruth Vazquez Sabariego ruth.sabariego at kuleuven.be
Tue Mar 13 10:31:16 CET 2018

Hi Guillaume,

I finally had a bit of time for running your files.
I’ve notice a problem at the level of the ‘InfiniteDomain’ definition.

You should avoid having physical regions that overlap, I mean, it should be:

// InfiniteDomain       = Region[ INFINITE_DOMAIN          ] ;
InfiniteDomainBottom = Region[ INFINITE_DOMAIN_BOTTOM   ] ;
InfiniteDomainRight  = Region[ INFINITE_DOMAIN_RIGHT    ] ;
InfiniteDomainTop    = Region[ INFINITE_DOMAIN_TOP      ] ;
InfiniteDomainLeft   = Region[ INFINITE_DOMAIN_LEFT     ] ;
InfiniteDomain       = Region[ {InfiniteDomainBottom, InfiniteDomainRight, InfiniteDomainTop,InfiniteDomainLeft} ] ;

GetDP looks for the Physical numbers first and then it goes to the elements.
It could be then that it does not see that the parts are in fact the big region.
Doing just that, the result is different, correct? To check.

Another possible source of a problem is the transformation to infinity that you apply in the region where you impose your particular boundary condition. Not sure that works either.
I would try first without that transformation.


Prof. Ruth V. Sabariego
KU Leuven
Dept. Electrical Engineering ESAT/Electa, EnergyVille

Free software: http://gmsh.info | http://getdp.info | http://onelab.info

On 6 Mar 2018, at 15:37, DILASSER Guillaume <Guillaume.DILASSER at cea.fr<mailto:Guillaume.DILASSER at cea.fr>> wrote:

Dear GetDP users,

I am writing to ask for some help on how to use a Formulation term to enforce a constraint on the value of the unknown field on the boundaries of a modelling domain.

The enclosed files provide an example of what I have been trying to do. I want simulate the response of a superconductor to an external field of the type B(t) = B1*f(t)*ux + B2*g(t)*uy. The functions f and g are different which prevents me from using the usual Constraintblock. Indeed, the use of the TimeFunction statement would multiply the entire B expression, not allowing me to have two separate terms with different time variations.

Therefore, I have been trying to enforce my boundary condition through an additional term in the formulation, taking inspiration from the benchmarks dipole.pro and microstrip.pro. Despite my best efforts, I still have issues with this approach :
•         The field does not properly take the value that I ask for. I imagine this is because the unknowns on the boundary of the modelling domain are the nodal values of Phi (the scalar potential) and I am trying to constrain the values of h (in fact more of dh/dt). I really cannot figure on my own how to constrain the values of Phi in the Formulation.
•         The potential Phi seems to be floating despite the constraints that I set. I do not understand why and it seems to lead to the computation crashing early.

I would really appreciate if someone could advise me, I have spent quite some time on this issue but I really do not grasp what is wrong…

Guillaume DILASSER
Ingénieur de recherche / Research engineer

CEA Saclay
Bât.123 – P. 213c
Point courrier n°11
91191 Gif-sur-Yvette Cedex

guillaume.dilasser at cea.fr<mailto:guillaume.dilasser at cea.fr>

getdp mailing list
getdp at onelab.info<mailto:getdp at onelab.info>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://onelab.info/pipermail/getdp/attachments/20180313/ee1d4472/attachment.html>

More information about the getdp mailing list