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 The information provided may be out of date.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.

Eddy Currents in a moving cylinder

Please login with a confirmed email address before reporting spam

I'm doing a small project in Comsol 3.5a. I have a cylinder made of a
dielectric in a DC magnetic field. The circular base of the cylinder is
parallel with the direction of the magnetic field (the base is on the X-Y
plane and extruded into the plane Z while the field is along the X axis).

I want the cylinder to rotate a fixed angle along the Z axis in the field
and calculate the power loss due to eddy currents in the field. The eddy
current examples I found were all static with and AC field. I need find a
way to either A. combine the AC/DC module with the structural model (for
the rotation) or B. rotate the magnetic field (over a given time) in the
AC/DC module. I haven't found a tutorial that allows me to rotate the field
or combine modules.

My lab has physically built these parts, so we have experimental values
for the time of rotation inside the magnetic field.

The ultimate goal of this model is so we can calculate eddy currents. I
want to input the rotation time (experimentally measured in the field) and
determine the power dissipated with in various DC fields.

2 Replies Last Post 28 oct. 2011, 02:59 UTC−4
Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 oct. 2011, 19:51 UTC−4
One approach could be to specify the velocity of the medium in the domain of the cylinder, using the quasi-static emqav mode in version 3.5a. [Another *might* be to try the newer "rotating machinery" application mode, but I haven't ever used that so I can't vouch for its applicability in this case.] Anyway, if you take the former approach, don't worry about the apparent contradiction that it is a time-harmonic formulation, but you are looking at a dc case. You can set the sinusoidal frequency to very low, and then just look at the magnitudes of the results (and be sure to compute your power losses using peak values combining appropriate variables, not sinusoidal-time power averages). For rotation, you'll need to specify the local vector field for the velocity of the cylinder as a function of both x and y, in the subdomain specs. E.g., let's say your cylinder rotates a constant myomega radians/sec around the z-axis in a right-handed (thumb along z, fingers point along motion) direction. Then at any point (x,y) in the cylinder, the x-component of v is v_x = -myomega*y and the y-component of v is v_y = myomega*x, while the z-component of v is zero. (Basically, we are simulating your rotating cylinder as if it were a partially-conducting fluid -- which is basically what this computational mode is meant for -- but where the velocity distribution in the fluid just happens to be identical to that of a solid, rotating, cylinder.)
One approach could be to specify the velocity of the medium in the domain of the cylinder, using the quasi-static emqav mode in version 3.5a. [Another *might* be to try the newer "rotating machinery" application mode, but I haven't ever used that so I can't vouch for its applicability in this case.] Anyway, if you take the former approach, don't worry about the apparent contradiction that it is a time-harmonic formulation, but you are looking at a dc case. You can set the sinusoidal frequency to very low, and then just look at the magnitudes of the results (and be sure to compute your power losses using peak values combining appropriate variables, not sinusoidal-time power averages). For rotation, you'll need to specify the local vector field for the velocity of the cylinder as a function of both x and y, in the subdomain specs. E.g., let's say your cylinder rotates a constant myomega radians/sec around the z-axis in a right-handed (thumb along z, fingers point along motion) direction. Then at any point (x,y) in the cylinder, the x-component of v is v_x = -myomega*y and the y-component of v is v_y = myomega*x, while the z-component of v is zero. (Basically, we are simulating your rotating cylinder as if it were a partially-conducting fluid -- which is basically what this computational mode is meant for -- but where the velocity distribution in the fluid just happens to be identical to that of a solid, rotating, cylinder.)

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 28 oct. 2011, 02:59 UTC−4
Hi

for info, I'm beeing studying 4.2a for some 48 hours now, one thing related is the new way to allow for splitting of the complex results into real sin-cos values (new settings check box in the COMPILE Equations solver nodes) And as Robert says to clearly differentiate the rms values and the pp values, this naming of these variables has also changed quite alot in v4.2


--
Good luck
Ivar
Hi for info, I'm beeing studying 4.2a for some 48 hours now, one thing related is the new way to allow for splitting of the complex results into real sin-cos values (new settings check box in the COMPILE Equations solver nodes) And as Robert says to clearly differentiate the rms values and the pp values, this naming of these variables has also changed quite alot in v4.2 -- Good luck Ivar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.