frequency shift in FDTD simulation

[salma]

geomerty discretization error cst

Hi all

What is the possible resoan for frequency shift between exact (analytic) and FDTD simulation?

Best
Salma

 

[RFSimulator]

fdtd frequency shifted

Two possible reasons:

1. Staircase error (geoemtry)
2. Mesh Dispersion Error (free space)

In either case the error should get smaller if you use a denser mesh..

F.

 

[yek]

frequency shifts after em simulation

Hi

I have the same problem, but can you explain that what you mean by Staircase error ?

I mean that how can this error make the problem?

thanks alot

 

[RFSimulator]

fdtd dispersion shift frequency

Hi,

in a "brute force" FDTD codes the geometry is discretized by using a staircase mesh. There is a difference between the simulated and the "true" geometry.

Consider a round patch antenna. If you discretize a round patch by a rectangular mesh the patch area will change -> shift in resonance frequency.

Some Time Domain Codes (such as CST) use additional techniques to correct for this error.

F.

 

[mitchphysics]

w.google.com

Maybe the cofficency depends on the fequencey~

 

[milli]

Where can i find more details on Staircase error....

and one more small query..
is this error comes into action when we simulate a electromagnetic signal in 2-D space....

i am asking this question becoz when i simulated.. i found energy changes on a circular wavefront in a lossless media...
i think that the reason for the energy change is this Staircase error..

help me in this case...

thanks in advance..

 

[fugalee]

Where can i find more details on Staircase error....

and one more small query..
is this error comes into action when we simulate a electromagnetic signal in 2-D space....

i am asking this question becoz when i simulated.. i found energy changes on a circular wavefront in a lossless media...
i think that the reason for the energy change is this Staircase error..

help me in this case...

thanks in advance..

Any book on FDTD would discuss this issue.

In fact, this is not a big problem in FDTD. The finer the mesh is, the better the results will be.

So, it's easy to identify whether you encontering this problem--Make your mesh finer, and compare the results.

There are also some novel Conformal FDTD technique(such as triangular mesh, PBA) to solve this, but are not very easy to be implemented in programs.