Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Modelling coils as current carrying edges
Posted 29.01.2011, 16:12 GMT-5 Geometry Version 4.0a 5 Replies
Please login with a confirmed email address before reporting spam
Hello fellow COMSOL users,
for my electromagnetic DC simulation I need a lot of coils. Hence I want to model them as lines in order to avoid long calculation times. I have experimented a bit in 2D with current carrying edges and everything works perfect.
However, if I turn now to 3D, I am having trouble defining the direction of the current.
I am modeling the coil as a circle on a 2d workplane, coerce it to a curve and then I am embedding it into the 3D geometry. If I now want to define the current on the edges, the coil splits into 4 edges and I have to define a current for each segment. However, there is no way to tell COMSOL which direction the current flows! I just can define a scalar value. Is there any way to specify the direction? If I give each segment the same value, the field turns out garbage :(
Or maybe I am doing something wrong with the modeling method? I am on 3.5a btw, but I also have 4.1 available.
Thanks in advance for your help!
for my electromagnetic DC simulation I need a lot of coils. Hence I want to model them as lines in order to avoid long calculation times. I have experimented a bit in 2D with current carrying edges and everything works perfect.
However, if I turn now to 3D, I am having trouble defining the direction of the current.
I am modeling the coil as a circle on a 2d workplane, coerce it to a curve and then I am embedding it into the 3D geometry. If I now want to define the current on the edges, the coil splits into 4 edges and I have to define a current for each segment. However, there is no way to tell COMSOL which direction the current flows! I just can define a scalar value. Is there any way to specify the direction? If I give each segment the same value, the field turns out garbage :(
Or maybe I am doing something wrong with the modeling method? I am on 3.5a btw, but I also have 4.1 available.
Thanks in advance for your help!
5 Replies Last Post 31.03.2011, 21:02 GMT-4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I believe our are better off in v4.1 as there you have several new coil types (multiturn etc) for the BC's.
Then in 3D its tricky because you need to respect the current flow globally (all closed loop, in the wrst case via the external boundaries). Otherwise you are violating the physics and the PDE do not solve correctly albeit, if not respecting this, you might get nice rainbow views, but they are wrong !
I always try to stay in 2D as far as possible, or 2D axi (unfortunatly in 2D axi the Lorentz velocity terms are not yet defined as a BC, you need to write out the equations fully by your own, good exercice ;)
Be sure you patch to latest update as there are a few errors in the ACDC equations in the early v4 versions
--
Good luck
Ivar
I believe our are better off in v4.1 as there you have several new coil types (multiturn etc) for the BC's.
Then in 3D its tricky because you need to respect the current flow globally (all closed loop, in the wrst case via the external boundaries). Otherwise you are violating the physics and the PDE do not solve correctly albeit, if not respecting this, you might get nice rainbow views, but they are wrong !
I always try to stay in 2D as far as possible, or 2D axi (unfortunatly in 2D axi the Lorentz velocity terms are not yet defined as a BC, you need to write out the equations fully by your own, good exercice ;)
Be sure you patch to latest update as there are a few errors in the ACDC equations in the early v4 versions
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
I have experienced the same problem defining currents on edges.
(I also decided to model coils as edges, because the true 3D coil with finite wire diameter would not mesh.)
I have found that in my model of a coil (13 line segments per turn, and 14 turns) half of the segments have a direction opposite to the other half. I could not find much in the documentation about a direction property of line geometry.
I noticed the problem when I applied the same current to all segments, and then on one side of a plane cutting the coil that the magnetization vector mirrored that from the other side of the plane.
My solution was to manually select all the segments making up one side and define the current using a negative sign.
So half of my segments have current on a line = I, the other half have current on a line = -I
Now I'm trying to figure out how to apply complex permeability to a domain in my model. I have noticed that 4.0a has this ability, but I want to use an expression for u' and u" that depends on frequency...
(I also decided to model coils as edges, because the true 3D coil with finite wire diameter would not mesh.)
I have found that in my model of a coil (13 line segments per turn, and 14 turns) half of the segments have a direction opposite to the other half. I could not find much in the documentation about a direction property of line geometry.
I noticed the problem when I applied the same current to all segments, and then on one side of a plane cutting the coil that the magnetization vector mirrored that from the other side of the plane.
My solution was to manually select all the segments making up one side and define the current using a negative sign.
So half of my segments have current on a line = I, the other half have current on a line = -I
Now I'm trying to figure out how to apply complex permeability to a domain in my model. I have noticed that 4.0a has this ability, but I want to use an expression for u' and u" that depends on frequency...
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
there are a few thing with current on edges:
1) you have no skin effect only an uniform currents (no densities as its over an area =0[m^2])
2) you must, manually, loop your current along the closed loop to respect the edge direction / current sign.
If you turn on geometry view EDGES you will see that all edges have an arrow defining their direction (the length variable "s" varies from 0 to 1 in 2D along each edge, warning not always so in 3D as there are s1 and s2). This is depending in which quadrant where the edge is located and is so for internal geometrical reasons.
This is NOT the easiest way for an edge representation of the current as you must define 2 edge currents with opposite sign and loop yourself (select which edges go where)
This is no bug but a feature ;)
It's just like that one have to know it and do the job. It's rather quickly done once you know about it.
There are a several cases when edge direction are important and must be sorted out by hand. IN PARTICULAR FOR INTERIOR BOUNDARIES you are warned ;)
--
Good luck
Ivar
there are a few thing with current on edges:
1) you have no skin effect only an uniform currents (no densities as its over an area =0[m^2])
2) you must, manually, loop your current along the closed loop to respect the edge direction / current sign.
If you turn on geometry view EDGES you will see that all edges have an arrow defining their direction (the length variable "s" varies from 0 to 1 in 2D along each edge, warning not always so in 3D as there are s1 and s2). This is depending in which quadrant where the edge is located and is so for internal geometrical reasons.
This is NOT the easiest way for an edge representation of the current as you must define 2 edge currents with opposite sign and loop yourself (select which edges go where)
This is no bug but a feature ;)
It's just like that one have to know it and do the job. It's rather quickly done once you know about it.
There are a several cases when edge direction are important and must be sorted out by hand. IN PARTICULAR FOR INTERIOR BOUNDARIES you are warned ;)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
If you turn on geometry view EDGES you will see that all edges have an arrow defining their direction (the length variable "s" varies from 0 to 1 in 2D along each edge, warning not always so in 3D as there are s1 and s2). This is depending in which quadrant where the edge is located and is so for internal geometrical reasons.
Hi Ivar, thanks for confirming this.
How can I turn on "geometry view EDGES" in Comsol 4.0a to see the direction arrows?
Thanks!
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Regarding edge directions and defining current on edges:
In Comsol 3.5a I found the above mentioned behaviour of geometry with some edges having direction opposite to others, depending on their global coordinate. (In my model the split was the X-Z plane).
I'm not sure if I used the method Ivar talks about above, but I did an Arrow plot on Edges, and selected Tangent 1 vector from geometry to see what was happening, as shown here:
img4.imageshack.us/img4/5469/35acoil.png
With that same model I observed the same effect in Comsol 4.0a.
However, using a similar model that models the same coil structure with more segments (31 instead of 14) I find the edge direction problem is (strangely) now gone!
To verify this, I did a Arrow Line plot (3D plot Group) and entered the t1x, t1y, t1z variables as in comsol 3.5a. These are not any of the default choices for variables to plot.
Produced this result
img687.imageshack.us/img687/8558/40acoil.png
Posting this just in case anyone else runs into the same problem.
In Comsol 3.5a I found the above mentioned behaviour of geometry with some edges having direction opposite to others, depending on their global coordinate. (In my model the split was the X-Z plane).
I'm not sure if I used the method Ivar talks about above, but I did an Arrow plot on Edges, and selected Tangent 1 vector from geometry to see what was happening, as shown here:
img4.imageshack.us/img4/5469/35acoil.png
With that same model I observed the same effect in Comsol 4.0a.
However, using a similar model that models the same coil structure with more segments (31 instead of 14) I find the edge direction problem is (strangely) now gone!
To verify this, I did a Arrow Line plot (3D plot Group) and entered the t1x, t1y, t1z variables as in comsol 3.5a. These are not any of the default choices for variables to plot.
Produced this result
img687.imageshack.us/img687/8558/40acoil.png
Posting this just in case anyone else runs into the same problem.