Questions regarding EME simulation and importing source fields into EME and FDTD

Ciel16Ciel16 Member Posts: 1

Hello All,

I am currently trying to simulate an AMMI structure. The total length of the structure is about 5500 um, and it's hard to simulate by using varFDTD. The structure can be seen below:

Due to the antisymmetric design, I found it is impossible to simulate it only by using EME which shows:

I have tried both methods: rotate the input port or use a bent waveguide. And neither worked as I expect from varFDTD (Fig3):

My first question is if there is a way to simulate this structure in EME?

Then I tried to simulate the different components individually that I simulate input & output parts by FDTD and the long-distance propagation by EME. I saved the field information from the FDTD simulation, and import it to port 1. The resulted simulation from EME looked fine for me:

However, the E-field intensity is totally different from what I obtained from varFDTD:

This also happens when I import the field profiles from FDTD to EME:

My second question is how this happens and will this affect the simulation results?

My third problem is when I tried to import the field from EME to FDTD, FDTD cannot perform my desired simulation that the light cannot propagate to the correct directions:

The E-filed from EME (in this case the magnitude for both E/H-field graphs are the same):

The resulted simulation graph from FDTD:

The light only propagates in the forward direction. Although I felt it might be the problem of losing the H-field profile, I checked the data by using Matlab and it seemed there were no missing components. 

My steps for the simulations are:

FDTD

  1. Create the FDTD simulation profile. Using “frequency-domain field and power monitor” to generate the 2D X-normal E and H fields.
  2. After finishing the simulation, I clicked the monitor and send all results & rawdata to script. Using command

EM1 = rectilineardataset("EM fields",x,y,z);

EM1.addattribute("E",Ex,Ey,Ez);

EM1.addattribute("H",Hx,Hy,Hz);

matlabsave("testfields.mat",EM1);

EME

  1. Import the profile to port 1 of the EME simulation file. Set port 2 to be the fundamental mode or TE or TM or user select (which I tried all of them and it seemed this did not affect). Add an “EME profile monitor” to generate data.
  2. After finishing the simulation, send the field profile from monitor_2 to the script, and use command:

data=getresult("monitor_2", "field profile");

matlabsave("testfields2.mat",data);

         FDTD (I did this step again because EME cannot handle the angle)

  1. Generate another simulation file for the end section, using an import source for the collected filed information from EME.

Then I the light could not propagate correctly all the time. I also attached the EME & FDTD & field data.

Thank you for taking the time to read this.

Sincerely,

Keru

Answers

  • kjohnsonkjohnson Posts: 254Ansys Employee

    Hello @Ciel16 ,

    Sorry for the delay in replying. This will be a very challenging device to simulate. FDTD/varFDTD shouldn't be used to simulate the long waveguide section due to the artificial dispersion introduced by the discretization. It might be possible to simulate the entire device with the EME solver, though. The EME solver can simulate propagation in angled waveguides, but it will require a large number of modes and cells along the angled sections. The issue with your EME simulations might be that you're not using enough modes, I would first try increasing the number of modes you're using, then using a mode convergence test to see if you're using enough modes.

Sign In or Register to comment.