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.

Induced voltages in saddle coils excited by two simple Helmholtz coils - simulation ends in physical impossible results

Please login with a confirmed email address before reporting spam

Dear Sir or Madam,

I already tried to get help for a similar problem in another thread (www.comsol.com/community/forums/general/thread/87301/) but I think more people will see this discussion if I start a new one; especially some new physical impossible effects occured.

I use COMSOL to simulate induced voltages in saddle shaped detection coils (this is the reason why I have to use COMSOL - it is the only real 3D simulation software for which I have a license at the moment). Nirmal Paudel (COMSOL employee) was kind enough to point me to the secret of the electrical conductivity of air which apparently must not be smaller than 1e-1 S/m in my frequency domain study. Otherwise the biconjugate gradient solver does not converge - for whatever reason (I would be very grateful if someone is able to explain me why since the real electrical conductivity of air is in the range of 1e-15 S/m). Like I already described in my referenced post (see above) I was not able to understand the simulation results of my splitted Helmholtz coil model because a symmetric model and a symmetric magnetic field must theoretically end in identical induced voltages in symmetric detection coils - but they're different. I reduced my model to not-splitted Helmholtz coils and at least both induced voltages fit theoretical considerations with respect to the curve trend U = f(freq) but not to the range of values I think. Hereupon I implemented a new (simple and reduced) model (see attached image: coils.png) and the induced voltage in the saddle coils changes its direction over the given frequency range freq = 1kHz up to 25kHz (see the other attached image: induced_voltages_U_freq.png). I will attach my *.mph file and I hope that someone is that kind that she or he will take the time to look into my files. I would be so grateful if someone could show me my big mistakes.

Thank you for your effort in advance,
Sebastian Draack


7 Replies Last Post 24 août 2015, 05:24 UTC−4
Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 20 août 2015, 13:18 UTC−4
Sebastian,

try to add 'Gauge Fixing for A-field' on all domains and check if it improves the result. It is in most cases required in mf and 3D.

Cheers
Edgar

--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Sebastian, try to add 'Gauge Fixing for A-field' on all domains and check if it improves the result. It is in most cases required in mf and 3D. Cheers Edgar -- Edgar J. Kaiser emPhys Physical Technology http://www.emphys.com

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 21 août 2015, 05:19 UTC−4
Hey Edgar,

thank you for your promising response. Unfortunately this was not the solution of the problem. I added a "Gauge Fixing for A-Field" item to my "Magnetic Fields" physics and selected all domains to it. The result is the same even if the solver needs some more time to solve the problem.

I hope anyone else has any suggestions. I'm looking forward to solve the problem with your help.
Sebastian
Hey Edgar, thank you for your promising response. Unfortunately this was not the solution of the problem. I added a "Gauge Fixing for A-Field" item to my "Magnetic Fields" physics and selected all domains to it. The result is the same even if the solver needs some more time to solve the problem. I hope anyone else has any suggestions. I'm looking forward to solve the problem with your help. Sebastian

Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 21 août 2015, 05:26 UTC−4
Strange: I tried your model and it looked pretty different and the results looked more realistic here. E.g. a linear dependence of induced voltage and frequency.
--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Strange: I tried your model and it looked pretty different and the results looked more realistic here. E.g. a linear dependence of induced voltage and frequency. -- Edgar J. Kaiser emPhys Physical Technology http://www.emphys.com

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 21 août 2015, 08:52 UTC−4
Thank you for your use, Edgar. I tried to simulate a second time and the results do not change. Is it important to follow a specific order of the "Magnetic Fields" item steps? I attached my model with the gauge fixing for A-field. Maybe you can see any differences between your modifications and mine? Like you see in my attached image I only determine a linear dependence of induced voltage and frequency between 10kHz and 25kHz.

Thank you for your effort again,
Sebastian
Thank you for your use, Edgar. I tried to simulate a second time and the results do not change. Is it important to follow a specific order of the "Magnetic Fields" item steps? I attached my model with the gauge fixing for A-field. Maybe you can see any differences between your modifications and mine? Like you see in my attached image I only determine a linear dependence of induced voltage and frequency between 10kHz and 25kHz. Thank you for your effort again, Sebastian


Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 21 août 2015, 09:28 UTC−4

this is what In get when I run your model in 5.1.
--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
this is what In get when I run your model in 5.1. -- Edgar J. Kaiser emPhys Physical Technology http://www.emphys.com


Please login with a confirmed email address before reporting spam

Posted: 9 years ago 24 août 2015, 04:25 UTC−4
Hey Edgar,

thank you for your reply. You're right - it's very strange that you get another result in Comsol 5.1 but the result must be wrong, too. Since the model is completely symmetrical, both induced voltages should be identical. Did you change the frequency range? Otherwise it would be strange that you do not get any results for freq = 25kHz. I suppose the indcued voltage should be in the range of U = 1 to 100 volts at a H-field with a magnetic flux density of about B = 15mT - so that's curious, too.
Hey Edgar, thank you for your reply. You're right - it's very strange that you get another result in Comsol 5.1 but the result must be wrong, too. Since the model is completely symmetrical, both induced voltages should be identical. Did you change the frequency range? Otherwise it would be strange that you do not get any results for freq = 25kHz. I suppose the indcued voltage should be in the range of U = 1 to 100 volts at a H-field with a magnetic flux density of about B = 15mT - so that's curious, too.

Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 24 août 2015, 05:24 UTC−4
Yes I only ran 1, 11 and 21 kHz to save time. The deviation between the two coils can be a meshing isssue.

--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Yes I only ran 1, 11 and 21 kHz to save time. The deviation between the two coils can be a meshing isssue. -- Edgar J. Kaiser emPhys Physical Technology http://www.emphys.com

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.