numerical coding of a dispersive media

[ranatel]

Hi every body,
Can you help me with numerical coding of a dispersive media? specially in FDTD codes?
thanx

 

[byteptr]

May be you talk about dielectrics and physics models such as Debye ?

 

[fatma1000]

i have been made such code for ultra short pulse propagation in nonlinear optical fiber. in this code i took into my account the effect of material and waveguide dispersion. the material model i used is the Loranzian model. the results can be found in the papers

1. M. Nady, M. B. El_Mashade, A. M. Attiya , " Analysis of Ultra-Short Pulse Propagation in a Non-Linear Optical Fiber by Using BOR-FDTD Technique ", 25th National Radio Science Conference (NRSC2008), March, 18-20, 2008, Faculty of Engineering, Tanta Univ. Egypt
2. M. Nady, A. M. Attiya, M. B. El_Mashade , " Analysis of Nonlinear Fiber Bragg Grating Using BOR-FDTD with Direct Integration Method", 25th National Radio Science Conference (NRSC2008), March, 18-20, 2008, Faculty of Engineering, Tanta Univ. Egypt

if you intersted i will upload it for you

thanks

 

[rlepkowicz]

fatma1000

Could you please upload those papers. I would be very interested.

Thank you

 

[fatma1000]

this is the first paper

if y intersted on the second one that is about fiber Bragg grating i will ubload it

 

[fatma1000]

y can find the two papers in this link




thanks

 

[MusicMan]

hi... i have a question regarding the use of PML. what are the maximum conductivities that you use in your PML,(both for electric and magnetic)? i ask this because i am having a problem when i implement the condition sigma_magnetic/mu=sigma_electric/epsilon. my update coefficient for the magnetic fields become negative=(....
pls.. pls... i hope you can answer my question

 

[iyami]

It is not entirely clear what kind of pml you are refering to.
PML for dispersive media or standard PML?
In the latter case you are probably running into a problem
which shows up in the regular update equations as well if
you have high conductivities.

Possible solutions

1) use forward differences
2) use max(0,c) instead of c for the update coefficient (of course that means that some pml layers
will be pointless)
3) compute the maximum conductivity which keeps the coefficient >= 0
4) use non-linear discretizations

I think 2 and 3 are easiest and sufficient for many cases.