Core Loss Calculation

Numerical methods and mathematical models of Elmer
vencels
Posts: 63
Joined: 20 Sep 2016, 17:05
Antispam: Yes
Location: Latvia
Contact:

Re: Core Loss Calculation

Post by vencels »

It was written to work in 3D, not sure about 2D.
Do you have a 3D test?
mr_monkey
Posts: 17
Joined: 20 Jan 2023, 18:41
Antispam: Yes

Re: Core Loss Calculation

Post by mr_monkey »

No, sorry.
mr_monkey
Posts: 17
Joined: 20 Jan 2023, 18:41
Antispam: Yes

Re: Core Loss Calculation

Post by mr_monkey »

I got this working properly now. Turns out this was a rather silly mistake.

I accidentally changed the Non Linear System Convergence Tolereance of the Magnetodynamics solver from 1e-06 to 1e-02 in the sif file that I used for the investigations. This caused some weird high frequency oscillations in the magnetic flux. These high frequency oscillations caused the core loss to go through the roof when increasing the Fourier Series Components.

The results are now within 5 % of the ANSYS Maxwell model I am comparing against, which seems reasonable to me.

Thanks everybody for your help.
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Core Loss Calculation

Post by raback »

Hi

That's great! Devil often lies in the detail.

It would be really nice if you could share some actual comparison, for example in this repo:
https://github.com/ElmerCSC/elmer-elmag

How many components did you end up using? They are sort of married with the time resolution. The more components you want the finer timestepping is needed.

-Peter
mr_monkey
Posts: 17
Joined: 20 Jan 2023, 18:41
Antispam: Yes

Re: Core Loss Calculation

Post by mr_monkey »

Hi Peter,

I ended up using 24 components since the machine I am modelling has 24 slots. I had a look at how increasing the number of components impacts the core loss (view attachment), and this seems a reasonable value.

I used 60 time steps per electrical period. Is there a rule of thumb for how many time steps/period are required to pick up higher harmonics? I could just do a sensitivity study on that as well I suppose.

I will add the example to the repo you suggested.

Thanks
Kevin
Attachments
loss_comp.png
loss_comp.png (33.97 KiB) Viewed 362 times
Post Reply