esa / NIDN

Neural Inverse Design of Nanostructures
GNU General Public License v3.0
37 stars 6 forks source link

Open questions in regard to waveregime solver #38

Closed gomezzz closed 2 years ago

gomezzz commented 2 years ago

Feature

Desired Behavior / Functionality

Please identify literature / answers to the following questions in relation to the FDTD solver.

Not urgent but long-term interest.

What Needs to Be Done

torbjornstoro commented 2 years ago

Regarding runtime, the following variables is of importance:

torbjornstoro commented 2 years ago

The Courant number is proportional to the number of time steps per space step. More precisely S_c = c * delta t/delta x. The upper limit is 1 in 1D, 1/sqrt(2) in 2D and 1/sqrt(3) in 3D. It is set automaticly by the fdtd to the upper limit. Changing this to a lower value will ???

torbjornstoro commented 2 years ago

I think the relative magneic permeability can be assumed to be 1 for most cases, assuming thetemperatures in consideration is fairly high (well above room temperature). The relative magnetic permeability is defined as 1+ magnetic suspectibility. Magnetic materials can be classified into paramagnets, diamagnets, ferromagnets ferrimagnets and antiferromagnets. For paramagnets, ferromagnets, ferrimagnets and antiferromagnets, the suspectibility is proportional to 1/T or 1/T^2 (for the three latter this is only the case above a certain temperature, which usually is relatively low). Thus the susceptibility approaches zero and the permeability can be approximated to 1. For diamagnets, the susceptibility is constant, but usualy really low, so this magnetic permeability can also be approximated to 1.

gomezzz commented 2 years ago

@torbjornstoro check boxes where known