Skip to content
Snippets Groups Projects
Commit c114731a authored by François Henrotte's avatar François Henrotte
Browse files

relecture du tuto et amelioration des explications

parent d787223b
No related branches found
No related tags found
No related merge requests found
Pipeline #
......@@ -13,21 +13,23 @@
------------------------------------------------------------------- */
/*
A thing GetDP is pretty good at is the management of global (non-local) basis functions.
Finite element expansions typically associate basis functions to individual nodes
or edges in the mesh. But consider the situation where a scalar field is set to be uniform
over a region of the problem (Say a floating potential electrode
in an Electrostatics problem, to fix the idea).
By factorizing the identical nodal value "v_electrode",
A thing GetDP is pretty good at is the management of global (non-local) basis
functions. Finite element expansions typically associate basis functions to
individual nodes or edges in the mesh. But consider the situation
where a scalar field is set to be uniform over a region of the problem
(Say a floating potential electrode in an Electrostatics problem,
to fix the idea). By factorizing the identical nodal value "v_electrode",
a global (non-local) basis function "BF_electrode" is obtained as factor
which is the sum of the shape functions of all the nodes in the electrode region.
This basis function "BF_electrode"
- is a continuous function
- is equal to 1 at the nodes of the electrode region, and to 0 at all other nodes
- decreases from 1 to 0 over the one element thick layer of outside finite elements
immediately in contact with the electrode region.
One such glabal basis function can be associated with each electrode in the system,
so that the finite element expansion of the electric scalar potential reads:
which is the sum of the shape functions of all the nodes in the electrode
region. This basis function "BF_electrode"
- is a continuous function, scalar in this case,
- is equal to 1 at the nodes of the electrode region, and to 0
at all other nodes
- decreases from 1 to 0 over the one element thick layer of outside
finite elements immediately in contact with the electrode region.
One such glabal basis function can be associated with each electrode
in the system, so that the finite element expansion of the electric
scalar potential reads:
v = Sum_k sn_k vn_k + Sum_electrode v_electrode BF_electrode
......@@ -36,10 +38,10 @@
- to reduce the number of unknowns
- to compute efficiently the electrode charges "Q_electrode",
which are precisely the energy duals of the global "v_electrode" quantities
- to deal with floating potentials, which are the computed electrode potential
when the electrode charge is imposed
- to provide output quantities (charges, armature voltages, capacitances, ...)
that can be immediately used in a external circuit.
- to deal with floating potentials, which are the computed electrode
potential when the electrode charge is imposed
- to provide output quantities (charges, armature voltages,
capacitances, ...) that can be immediately used in a external circuit.
*/
......@@ -100,9 +102,10 @@ Constraint {
{ Name SetGlobalPotential; Type Assign;
Case {
/* Define the imposed potential regionwise on the different parts of "Electrodes_Ele".
No voltage imposed to the Microstrip electrode
when the "Fixed charge" option is enabled ( MicrostripTypeBC = true ). */
/* Define the imposed potential regionwise on the different parts of
"Electrodes_Ele". No voltage imposed to the Microstrip electrode
when the "Fixed charge" option is enabled
( MicrostripTypeBC = true ). */
{ Region Ground; Value 0; }
If( !MicrostripTypeBC )
{ Region Microstrip; Value MicrostripValueBC; }
......@@ -128,15 +131,28 @@ Group{
FunctionSpace {
/* The magic in the treatment of global quantitities by GetDP is in the fact
that nearly all the work is done at the level of the FunctionSpace definition.
The finite element expansion is
that nearly all the work is done at the level of the FunctionSpace
definition. The finite element expansion of "v" is
v = Sum_k sn_k vn_k + Sum_electrode v_electrode BF_electrode
The sum_k runs over all nodes except those of the electrode regions.
"sf" stands for "BF_electrode"
"vf" stands for "v_electrode"
with the the sum_k running over all nodes except those of the electrode
regions. This is exactly what one finds in the FunctionSpace definotion
below with "sf" standing for "BF_electrode" and "vf" for "v_electrode".
The global quantities are also be attributed a more explicit
and meaningful name. Moreover the "AssociatedWith" statement manifests
the fact that the global potential of an electrode is the (electrostatic)
energy dual of the electric charge carried by that electrode.
By checking the "Display global basis functions" checkbox and running
the model, you can take a look on how the two "BF_electrode" basis
functions in this model look like.
Constraints can then be set on either component of the FunctionSpace.
Besides the usual Dirichlet boundary condition conditions, which is
left here for the sake of completeness but is not used in this model,
there is the possibility to fix either the GlobalPotential or the
ArmatureCharge of each indidual electrode (not both, of course).
When the ArmatureCharge is fixed, the computed GlobalPotential computed
for that electrode is the so-called floating potential.
*/
{ Name Hgrad_v_Ele; Type Form0;
......@@ -186,8 +202,9 @@ Integration {
Formulation {
/* Only minor changes in the formulation.
The global quantities are declared in the "Quantity{}" section,
and a "GlobalTerm" is added that triggers the assembly of an additional equation
per electrode in the system to compute the charge Q_electrode
and a "GlobalTerm" is added that triggers the assembly
of an additional equation per electrode in the system
to compute the charge Q_electrode accordint to:
Q_electrode = (-epsr[] Grad v, Grad BF_electrode)_Vol_Dielectric_Ele
*/
......@@ -239,7 +256,7 @@ PostProcessing {
}
}
}
// next lines only needed to display global basis functions in PostProcessing
// next lines only needed to display global BF in PostProcessing
{ Name BFGround; Value { Term { [ BF {vf} ]; In Dom_Hgrad_v_Ele;
SubRegion Ground; Jacobian Vol; } } }
{ Name BFMicrostrip; Value { Term { [ BF {vf} ]; In Dom_Hgrad_v_Ele;
......@@ -249,12 +266,11 @@ PostProcessing {
}
}
/* Various output results are generated.
They are both displayed in the graphical user interface, and stored in disk files.
In particular, global quantities related results are stored in the "output.txt" file.
There is a user option to display the global basis functions of the two electrodes.
Another option allows the user to chose to not overwrite
the "output.txt" file when running a new simulation. */
/* Various output results are generated, which are both displayed
in the graphical user interface, and stored in disk files.
In particular, global quantities related results are stored
in the "output.txt" file. A user option allows to chose
to not overwrite the "output.txt" file when running a new simulation. */
PostOperation {
{ Name Map; NameOfPostProcessing EleSta_v;
......@@ -267,7 +283,8 @@ PostOperation {
"View[l].ShowElement = 1;"],
File "BFGround.opt", LastTimeStepOnly] ;
Print[ BFMicrostrip, OnElementsOf Dom_Hgrad_v_Ele, File "BFMicrostrip.pos" ];
Print[ BFMicrostrip, OnElementsOf Dom_Hgrad_v_Ele,
File "BFMicrostrip.pos" ];
Echo[ StrCat["l=PostProcessing.NbViews-1;",
"View[l].IntervalsType = 1;",
"View[l].NbIso = 40;",
......@@ -294,7 +311,8 @@ PostOperation {
Print[ C, OnRegion Microstrip, File > "output.txt", Color "AliceBlue",
Format Table, SendToServer "Output/Global/Capacitance [F]" ];
Echo[ "Electrostatic energy [J]:", Format Table, File > "output.txt"] ;
Print[ energy, OnRegion Vol_Dielectric_Ele, File > "output.txt", Color "AliceBlue",
Print[ energy, OnRegion Vol_Dielectric_Ele, File > "output.txt",
Color "AliceBlue",
Format Table, SendToServer "Output/Global/Energy [J]" ];
}
}
......
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment