reflectarray simulation using CST MWS

Update:
I tried to use the Integral Equation solver, it works but takes tooooo long time.

panou said:

Dear all,

I am trying to simulate a reflectarray using CST MICROWAVE STUDIO. The reflectarray is 18*18 circular array with a radius about 4'', 13GHz center frequency. I am using variable-sized patches in this reflectarray. Currently I am using a simple waveguide as the feed, not using the horn feed. But when I tried to use the transient solver, an error occured:

Memory resources exhausted while trying to allocate matrices.
Error in calculating slover matrix.

Maybe I need to use a more powerful computer for the huge matrices? Anyone has simulated reflectarray using CST? Any help? Thanks!


Best,
Panou

Click to expand...

CST MWS is not suitable for large structure. It is based on Finite integration Technique and uses volume discretization. I think, you should use ADS Momentum of FEKO Software. They are based on the Moment Method. Also, if you are familiar with Ansoft HFSS, so you can use Floque boundary condition to overcome this problem.

Thanks, I will try FEKO.

I am using variable-size patches in the reflectarray, it's not periodic unit cell. Could I still use Floque boundary? I think the Floque boundary only apply for array whose elements are all the same.

hamidrezakarami said:

CST MWS is not suitable for large structure. It is based on Finite integration Technique and uses volume discretization. I think, you should use ADS Momentum of FEKO Software. They are based on the Moment Method. Also, if you are familiar with Ansoft HFSS, so you can use Floque boundary condition to overcome this problem.

Click to expand...

panou said:

Thanks, I will try FEKO.

I am using variable-size patches in the reflectarray, it's not periodic unit cell. Could I still use Floque boundary? I think the Floque boundary only apply for array whose elements are all the same.

Click to expand...

Hi panou,

I have tried floquet boundary with variable size element in HFSS, it were able to generate a phase curve. Then, the phase curve was used to design a small reflectarray in HFSS and it can scan the main beam accurately as we predicted in calculations. Funny, when i import the model to CST and simulate the same reflectrray in transient solver, it showed no beam scanning and just give a specular reflected beam. Still figuring out what was the reasons behind it.

hamidrezakarami said:

CST MWS is not suitable for large structure. It is based on Finite integration Technique and uses volume discretization. I think, you should use ADS Momentum of FEKO Software. They are based on the Moment Method. Also, if you are familiar with Ansoft HFSS, so you can use Floque boundary condition to overcome this problem.

Click to expand...

no,floquet mode took long time in HFSS and we simply use normal incident model(waveguide model)
For panou,for my experiences in CST,i try to run 31 x 31 elements
using time solver,it good and ur array can be simulated,try to use E walls and H walla in ur array and i hope u have much memory,4 giga is good for u
good luck

Thanks you guys.

I will try different solutions.....