Jump to content
Dimon

Mesh Density of Bulky Object at 72GHz

Recommended Posts

Hello,

 

I need to calculate RCS of large cube (8x2x1.5 m) at 70 GHz. Unfortunatelly, FEKO solver triangulates my object in too dense grid. Is there any way to reduce total simulation time? Are there any settings which I can play with to optimize my simulation? I used the most coarse settings of the meshgrid in my simulation. But it still works to loooooong....

 

BR, Dimon

Share this post


Link to post
Share on other sites

Hello Dimon,

 

What method are you currently using? Are you using MoM or MLFMM? FEKO has a wide variety of solution methods and these have also been hybridised in many cases. This is a huge problem to solve - it is 1868 wavelengths in the largest dimension of your cube. If you are talking about mono-static RCS, large element physical optics (LE-PO) is an ideal solution. PO is less accurate in the shadow region and thus, depending on your requirements, it might lot be suitable for a full bi-static RCS calculation. The other high frequency (assymptotic) solution method that you should consider is ray launching geometrical optics (RL-GO). It will generate very few elements, but remember that still needs many rays at 70 GHz.

 

By my rough calculation, at 70 GHz, you are going to need roughly 500 000 000 triangles for MoM/MLFMM. For RCS calculations, you could even go even more coarse (I have seen good results at lambda / 4 and even lower in some cases), but even then the required number of elements will probably bee too high and thus I would try to use one of the asymptotic methods, but what you can use depends on the required calculation (mono-static or bistatic RCS) and the desired accuracy.

Share this post


Link to post
Share on other sites
3 hours ago, JIF said:

Hello Dimon,

 

What method are you currently using? Are you using MoM or MLFMM? FEKO has a wide variety of solution methods and these have also been hybridised in many cases. This is a huge problem to solve - it is 1868 wavelengths in the largest dimension of your cube. If you are talking about mono-static RCS, large element physical optics (LE-PO) is an ideal solution. PO is less accurate in the shadow region and thus, depending on your requirements, it might lot be suitable for a full bi-static RCS calculation. The other high frequency (assymptotic) solution method that you should consider is ray launching geometrical optics (RL-GO). It will generate very few elements, but remember that still needs many rays at 70 GHz.

 

By my rough calculation, at 70 GHz, you are going to need roughly 500 000 000 triangles for MoM/MLFMM. For RCS calculations, you could even go even more coarse (I have seen good results at lambda / 4 and even lower in some cases), but even then the required number of elements will probably bee too high and thus I would try to use one of the asymptotic methods, but what you can use depends on the required calculation (mono-static or bistatic RCS) and the desired accuracy.

 

Thank you for your answer. Yes you' re right about asymptotic method, and it would be fine for me in terms of accuracy. But I thought that FEKO selects  the model automatically depending on wavelength/object size ratio. If it doesn't then where can I manually select a certain asymptotic model? Could you advise where the setting is located in Feko interface? 

Share this post


Link to post
Share on other sites
1 hour ago, Dimon said:

But I thought that FEKO selects  the model automatically depending on wavelength/object size ratio. If it doesn't then where can I manually select a certain asymptotic model?

FEKO does not currently select (or suggest) solution methods for the model. It is something that we are working on, but as you can imagine, it is not so easy to derive the best solution method by simply looking at the model (it is reasonably easy for some of our experts, but writing the code to do it is not so easy). Once the user has selected the solution method, FEKO (CADFEKO) can automatically mesh correctly for the chosen solution method.

 

The settings that you are looking for are on faces. Go to the Details tree (after selecting the cuboid) and select all the Faces. Right click on the selected faces and click on Properties in the menu. This will show the propoerties dialog for the selected faces. Now go the the Solution tab. In the combo box, select "Large element PO - full ray tracing" for LE-PO or "RL-GO" and click on OK to apply the settings. Then re-mesh and solve the model.

 

I would suggest that you try both LE-PO and RL-GO and compare the results. For your model, LE-PO should be faster. If you have multiple reflections in your model (you don't), I would rather go for RL-GO. PO does support multiple interactions, but it quickly becomes very slow when more than one interaction is taken into account.

Share this post


Link to post
Share on other sites

Hello JIF,

 

Thank you for your help! It works! I got the idea of mesh grid optimization in FEKO.

 

Well, actually you're right about complexity to select the right simulation methods for the model. But if I would point a goal criterion before calculation, then it might help. For example "highest accuracy", "fast computational time" or "balanced mode". Also, if you have a choice to select a trade-off between them in terms of number, then it might be useful as well. For example accuracy - 0.7, simulation time - 0.3. And of course the numerical criterion may also concern an optimization of mesh, wedges, edges etc...

 

BR,

Dimon

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
You are posting as a guest. If you have an account, please sign in.
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...