Ideas on MagNet Toolkit

[minjiechen]

One recommendation - we can make the core loss prediction function reading both time sequence B(t), and descriptive language, "sine", "trig", "trapz", with specified duty ratios, something like this:

something like this -
Pv = f ("N87", "sine", f, Bdc, T, "paderborn") - reading descriptive language for waveforms
Pv = f ("N49", "trig", D, Bdc, T, "sydney") - reading descriptive language for waveforms
Pv = f ("N27", B(t), T, "princeton") - reading B(t) for waveforms

perhaps B(t) should be compatible with arbitrary number of steps.
we will internally reformat/generate B(t) for a particular resolution (say 1024) based on the information we receive.

[wkirgsn]

Great ideas, i will implement that.
Note that the arbitrary b sequence length is in fact already implemented in the Paderborn model, as it needs 1024 samples per period.
Likewise, other models will certainly have such resampling/Interpolation, too.

[wkirgsn]

Hi Minjie, is there a Princeton model that we could add to the Toolkit?

[minjiechen]

Not sure, ccing Shukai/Haoran ~ Bests, Minjie ------------------ Minjie Chen Princeton University Zoom: https://princeton.zoom.us/j/6177687869

…

On Tue, Feb 25, 2025 at 11:15 PM Wilhelm Kirchgässner < ***@***.***> wrote: Hi Minjie, is there a Princeton model that we could add to the Toolkit? — Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB7OHY4BIUNZG3JB4RG4RX32RU5V7AVCNFSM6AAAAABX4HXXDKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTEMZSGA3DQNA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[ehavugi]

related add PrincetonUniversity/magnet#125