FDTD + Mur : Known Stability issues (Please Help People!!!)

[somada141]

Hello everyone,

Since my last post was viewed by 24 people and no one bothered to help in anyway I'll try to clarify my problem.

I am making a 3D FDTD code with Mur ABCs.

When I only use the pure FDTD formulas everythin works fine but when I enable the mur formulas on the boundaries then after a while of normal behavior an abnormal increase takes place

Here is a snapshot :









This is a 3D golden waveguide with two opening on the y axis. I have only enforced Mur at y=0 to make the difference between the Mur and its absence clear.

The boundaries at y=0 is 50 cells away from the wave guide meaning 3wavelengths away at 8GHz which is a great frequency for this waveguide.

Does anyone ANYONE AT ALL know what may have gone wrong? Has anyone seen anything similar .

Concerning the standard FDTD stability issues meaning the D (cell size) and Dt (time step) there are both covered by the standard equation by Allen tavlove. D is 15 times smaller than the wavelength and Dt is produced by D/(sqrt(3)*C).

Please someone help. If someone bothers to help me then I could help someone else with this method and its various shortcomings and problems.

Thanks in advance,
Adam

somada141@yahoo.gr

 

[Winn]

Re: FDTD + Mur : Known Stability issues (Please Help People!

Actually the Mur ABC is one of the easiest ABC, you should program it in such a way that:

// Abc at the beginning of the simulation domain
ex[0] = ex_low_m2;
ex_low_m2 = ex_low_m1;
ex_low_m1 = ex[1];

// Abc at the ending of the simulation domain
ex[nxsize-1] = ex_low_m2;
ex_low_m2 = ex_low_m1;
ex_low_m1 = ex[nxsize-2];

Hope, this helps you !

Regards

 

[somada141]

Re: FDTD + Mur : Known Stability issues (Please Help People!

Actually I found the mistake I was making. It was something of a stupid mistake.

I only posting this in case someone finds it useful:

While updating the inner field components for E with FDTD so that you update the outer field components with MUR you calculate the inner fileds only on the directions where you enforce MUR.

For example for Ex:

We update with FDTD in:

for x in (0,end)
for y in (1,end-1)
for z in (1,end-1)
FDTD

what I did was calculate x for 1,end-1...Hence the instability caused.

Hope this helps someone out there

 

[medaziz]

Re: FDTD + Mur : Known Stability issues (Please Help People!

hi

is there more discussion about this subject

 

[somada141]

Re: FDTD + Mur : Known Stability issues (Please Help People!

no, from the beginning there was practically no interest in this topic from other user. However I have solved my problem so if you have any similar problems or if you have any other questions concerning FDTD & Mur please ask...

 

[medaziz]

Re: FDTD + Mur : Known Stability issues (Please Help People!

hi

i'm working also with FDTDsince 2003 but my interset is focalised to study biomedicals antenna. Sincerly i have usually some FDTD problems.

cordially

 

[somada141]

Re: FDTD + Mur : Known Stability issues (Please Help People!

what sort of problems are you facing with FDTD?

 

[somada141]

THIS MATTER HAS BEEN RESOLVED.CHECK OUT THIS :