Inductance extraction for parallel wires

 Posts: 12
 Joined: 11 Sep 2018, 01:05
 Antispam: Yes
Inductance extraction for parallel wires
Geometry:
Two parallel wires inside a cube.
Each wire is defined as a physical surface (1 and 2)
There is one physical volume ( Cube, two wires).
I'm trying to find the inductance of two parallel wires.
I created a sif file based on elmer's examples on "MGDynamics" solver.
Can someone explain how I should define boundaries for my project and how I should use the MagnetoDynamics 3D solver to calculate inductance. Links would be very much appreciated.
I'm attaching my sif and geometry file
Two parallel wires inside a cube.
Each wire is defined as a physical surface (1 and 2)
There is one physical volume ( Cube, two wires).
I'm trying to find the inductance of two parallel wires.
I created a sif file based on elmer's examples on "MGDynamics" solver.
Can someone explain how I should define boundaries for my project and how I should use the MagnetoDynamics 3D solver to calculate inductance. Links would be very much appreciated.
I'm attaching my sif and geometry file
 Attachments

 pwire.sif
 (3.31 KiB) Downloaded 64 times

 Site Admin
 Posts: 3420
 Joined: 22 Aug 2009, 11:57
 Antispam: Yes
 Location: Espoo, Finland
 Contact:
Re: Inductance extraction for parallel wires
Hi
For this particular case how about computing the inductance from the fact that the total energy is, E=LI^2/2. At least the CalcFields does automatically compute the energy.
Peter
For this particular case how about computing the inductance from the fact that the total energy is, E=LI^2/2. At least the CalcFields does automatically compute the energy.
Peter

 Posts: 12
 Joined: 11 Sep 2018, 01:05
 Antispam: Yes
Re: Inductance extraction for parallel wires
Hi Peter,
Thanks for your help. When I run the provided MGDynamicsSteady.sif inside /home/arguest/Desktop/elmerfem/fem/tests/mgdyn_steady/, I could see the Magnetic field energy. I viewed the mesh inside mgdyn_steady : it was a simple cube with six physical surfaces and one volume.
I made my own cube with 6 physical surfaces and one physical volume in gmsh. I converted the .msh to mesh.* using Elmergrid. I modified boundaries for my cube in MGDYnamicsSteady.sif ran it on my mesh.I'm not getting the magnetic field energy (It's displayed as 0.0000).
I'm attaching all the required files. It would be really helpful if someone could point out my trivial error.
Thanks for your help. When I run the provided MGDynamicsSteady.sif inside /home/arguest/Desktop/elmerfem/fem/tests/mgdyn_steady/, I could see the Magnetic field energy. I viewed the mesh inside mgdyn_steady : it was a simple cube with six physical surfaces and one volume.
I made my own cube with 6 physical surfaces and one physical volume in gmsh. I converted the .msh to mesh.* using Elmergrid. I modified boundaries for my cube in MGDYnamicsSteady.sif ran it on my mesh.I'm not getting the magnetic field energy (It's displayed as 0.0000).
I'm attaching all the required files. It would be really helpful if someone could point out my trivial error.
Re: Inductance extraction for parallel wires
Hi,
you forgot to define the body number in the body section. Try this:
Regards
Martin
you forgot to define the body number in the body section. Try this:
Code: Select all
Body 1
Equation = 1
Target Bodies(1) = 7
Material = 1
Body Force = 1
Initial Condition = 1
End
Martin

 Posts: 12
 Joined: 11 Sep 2018, 01:05
 Antispam: Yes
Re: Inductance extraction for parallel wires
Hi,
I was able to calculate magnetic field energy and flux average
SaveScalars: 1: res: eddy current power 0.000000000000E+000
SaveScalars: 2: res: magnetic field energy 3.981562920553E031
SaveScalars: 3: res: magnetic flux average 4.566833176772E021
SaveScalars: 4: res: magnetic flux density average 3.044555451182E009
SaveScalars: 5: res: magnetic flux area 1.500000000000E012
Energy/Flux should be sensitive to changes in current density so that Inductance(L) remains constant.
However I'm getting the same energy values for different current density resulting in different inductance values. Is this the normal behavior or my procedure to calculate magnetic energy/flux wrong? Can someone help me with this?
(I'm calculating inductance as L= flux/ Current. Current = Current Density * Area.
Area = 0.25um*um, Current Density= 4 nA/um*um => Current = 2nA)
Thanks.
I was able to calculate magnetic field energy and flux average
SaveScalars: 1: res: eddy current power 0.000000000000E+000
SaveScalars: 2: res: magnetic field energy 3.981562920553E031
SaveScalars: 3: res: magnetic flux average 4.566833176772E021
SaveScalars: 4: res: magnetic flux density average 3.044555451182E009
SaveScalars: 5: res: magnetic flux area 1.500000000000E012
Energy/Flux should be sensitive to changes in current density so that Inductance(L) remains constant.
However I'm getting the same energy values for different current density resulting in different inductance values. Is this the normal behavior or my procedure to calculate magnetic energy/flux wrong? Can someone help me with this?
(I'm calculating inductance as L= flux/ Current. Current = Current Density * Area.
Area = 0.25um*um, Current Density= 4 nA/um*um => Current = 2nA)
Code: Select all
Boundary Condition 1
Target Boundaries(3) = 2 3 4
P {e} = Real 0
P {e} 1 = Variable coordinate
Real MATC "x=tx(0);y=tx(1); 3*y^2/2"
P {e} 2 = Variable coordinate
Real MATC "x=tx(0);y=tx(1); x^2/2"
P {e} 3 = Real 0
Electric Current Density = Real 4e9
Magnetic Flux Average = Logical True
End
Re: Inductance extraction for parallel wires
Hmm, your sif file looks very complex for this simulation. There are different exitations
 body force with electric potential
 current density on the boundary
 predefined vector potential on boundary
It seems your simulation is overdetermined. I would like to propose following:
Remove predefined vector potential on the boundary (set it to 0 on every surface), remove electric potential, and set only current density in the body force section.
Regards
Martin
 body force with electric potential
 current density on the boundary
 predefined vector potential on boundary
It seems your simulation is overdetermined. I would like to propose following:
Remove predefined vector potential on the boundary (set it to 0 on every surface), remove electric potential, and set only current density in the body force section.
Regards
Martin

 Posts: 12
 Joined: 11 Sep 2018, 01:05
 Antispam: Yes
Re: Inductance extraction for parallel wires
Hi Guys,
[1] I found this line under fem/tests/mgdyn_steady_wire2/ ( boundary stimulation) => AV = Real 5.542e5 which is tuned for 1 Ampere.
How do I calculate AV value to specify input current. Ex: 1nA => AV = ?, 1uA => AV = ?

[2] Also under fem/tests/mgdyn_steady_wire/ which is for body force stimulation, I found this
Is coilPot a parameter ? If so how can I input it [or] is it a value specified in some file.
How does "Electric Potential" value equates to current flowing in this case.
Thanks.
[1] I found this line under fem/tests/mgdyn_steady_wire2/ ( boundary stimulation) => AV = Real 5.542e5 which is tuned for 1 Ampere.
Code: Select all
Boundary Condition 3
Name = "WireEnd"
Target Boundaries(1) = 3
Coil End = Logical True
AV {e} 1 = Real 0.0
AV {e} 2 = Real 0.0
! Tuned to obtain current of 1 A
AV = Real 5.542e5
End
How do I calculate AV value to specify input current. Ex: 1nA => AV = ?, 1uA => AV = ?

[2] Also under fem/tests/mgdyn_steady_wire/ which is for body force stimulation, I found this
Code: Select all
Body Force 1
Electric Potential = Equals "CoilPot"
End
How does "Electric Potential" value equates to current flowing in this case.
Thanks.

 Site Admin
 Posts: 3420
 Joined: 22 Aug 2009, 11:57
 Antispam: Yes
 Location: Espoo, Finland
 Contact:
Re: Inductance extraction for parallel wires
Hi
Well, you can 1st try with 1V and then see the current you get and scale accordingly. There may even be some automated scaling but this was probably done manually.
The "CoilPot" is a potential computed by another solver.
The difference of these two is that in the 1st the {A,V} is solved as a strongly coupled system. For higher frequencies this will make the current concentrate on the surface. When the potential is precomputed we enforce a hierarchical solution which eliminates the effect of magnetic field to current.
Peter
Well, you can 1st try with 1V and then see the current you get and scale accordingly. There may even be some automated scaling but this was probably done manually.
The "CoilPot" is a potential computed by another solver.
The difference of these two is that in the 1st the {A,V} is solved as a strongly coupled system. For higher frequencies this will make the current concentrate on the surface. When the potential is precomputed we enforce a hierarchical solution which eliminates the effect of magnetic field to current.
Peter

 Posts: 12
 Joined: 11 Sep 2018, 01:05
 Antispam: Yes
Re: Inductance extraction for parallel wires
Hi,
I'm getting consistent inductance values for different current densities. However, this happens only if the current densities are powers of 2(J = 1, 2, 4, 8, etc.,). I'm not getting consistent inductance values when J= 1.1, 3, 6 etc., ( E= 0.5*L*(I*I) => L= (2*E)/(I*I) )
I'm attaching my sif file and results file
Also I want to try with Electric Potential. So If I set the Potential in the body force section, is it possible to view the current flowing through my geometry. If so, how?

Same issue for setting AV variable for boundary stimulation. If set a certain AV value, is it possible to view the current through the geometry. If so, how?

By current through the geometry, I mean the total current flowing from metal start to metal end. Not individual element current values which "Saveline" solver dumps to f.dat.
Thanks.
I'm getting consistent inductance values for different current densities. However, this happens only if the current densities are powers of 2(J = 1, 2, 4, 8, etc.,). I'm not getting consistent inductance values when J= 1.1, 3, 6 etc., ( E= 0.5*L*(I*I) => L= (2*E)/(I*I) )
I'm attaching my sif file and results file
Code: Select all
[b]My body force section:[/b]
Body Force 1
Current Density 1 = 0
Current Density 2 = 0
Current Density 3 = 1.2e3
! Electric Potential = Real 0.01
End

Same issue for setting AV variable for boundary stimulation. If set a certain AV value, is it possible to view the current through the geometry. If so, how?

By current through the geometry, I mean the total current flowing from metal start to metal end. Not individual element current values which "Saveline" solver dumps to f.dat.
Thanks.

 Posts: 12
 Joined: 11 Sep 2018, 01:05
 Antispam: Yes