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.
Optimization Electromagnetic Device
Posted 5 juil. 2010, 08:24 UTC−4 3 Replies
Please login with a confirmed email address before reporting spam
Hello!
I need help for an electromagnetic problem- a solenoid who has one main coil and another 2 correcting coils.
I want to optimize the length of the 2 correcting coils and the distance between them, so that the magnetic field be as uniform as possible.
The objective function is f(S,L)=(Bmax-Bmin)/2, where Bmax and Bmin are maximum and minimum of magnetic induction within the controlled subregion and B0 is the induction at the centre of the compound solenoid.
And I need to find the minimum f(S,L).
No constrains are imposed on the latter except the fact that S,L must be positive.
How do I putt that constrains or the objective function in Comsol?
Please help!
I need help for an electromagnetic problem- a solenoid who has one main coil and another 2 correcting coils.
I want to optimize the length of the 2 correcting coils and the distance between them, so that the magnetic field be as uniform as possible.
The objective function is f(S,L)=(Bmax-Bmin)/2, where Bmax and Bmin are maximum and minimum of magnetic induction within the controlled subregion and B0 is the induction at the centre of the compound solenoid.
And I need to find the minimum f(S,L).
No constrains are imposed on the latter except the fact that S,L must be positive.
How do I putt that constrains or the objective function in Comsol?
Please help!
Attachments:
3 Replies Last Post 12 juil. 2010, 04:25 UTC−4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi ,
I am not sure what is the problem but I think you need to have physic node next to optimization node, included as well. Additionally I would use deformed mesh to if you are running V3.5 but if you have V4.0 I heard (thou not yet tried) the geometry optimization is possible simply by creating geometry with variables (instead of the number you can say that your rectangle is variable rect_a which you can use after in you optimization.)
KR
Nenad
I am not sure what is the problem but I think you need to have physic node next to optimization node, included as well. Additionally I would use deformed mesh to if you are running V3.5 but if you have V4.0 I heard (thou not yet tried) the geometry optimization is possible simply by creating geometry with variables (instead of the number you can say that your rectangle is variable rect_a which you can use after in you optimization.)
KR
Nenad
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I believe first you might restrain your model to the first quadrant and use a horizontal "antisymmetry" boundary on Y=0. It will speed up the calculations.
Then you need to solve the B value, so either you set up a full FEM model with an "air" interface (quarter of circle), perhaps add a spherical PML (perfectly matched layer) and integrate your B value to get a nice optimisation error function, or you add just the equations and solve the B value as 1D PDE case along Z.
In V4 you can also define the position of the coil as a variable, and have COMSOL reconstruct the geometry as it solves.
hope this helps
Good luck
Ivar
I believe first you might restrain your model to the first quadrant and use a horizontal "antisymmetry" boundary on Y=0. It will speed up the calculations.
Then you need to solve the B value, so either you set up a full FEM model with an "air" interface (quarter of circle), perhaps add a spherical PML (perfectly matched layer) and integrate your B value to get a nice optimisation error function, or you add just the equations and solve the B value as 1D PDE case along Z.
In V4 you can also define the position of the coil as a variable, and have COMSOL reconstruct the geometry as it solves.
hope this helps
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thanks a lot!