Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Reduce Degree of Freedom antenna array
Posted 2016年2月8日 GMT-5 14:16 RF & Microwave Engineering, Mesh, Studies & Solvers Version 5.0 2 Replies
Please login with a confirmed email address before reporting spam
Hello,
I am using RF module and have a created a model of antenna array, with which I would like to observe scattering of EM waves when penetrating dielectric material (permittivity 50-70)
The antenna array is used to focus radiation in one place. Since antenna array works only in far field, I need a volume at least 1.5-2 wavelengths to observe the scattering of waves in material. This is why I have enclosed antenna array in tall cylinder shape.
When I try to mesh the model, the DOF number is too large for computer to solve (because of large permittivity of material). Are there any ways to reduce DOF's in particular model?
Thank you,
Arturs
I am using RF module and have a created a model of antenna array, with which I would like to observe scattering of EM waves when penetrating dielectric material (permittivity 50-70)
The antenna array is used to focus radiation in one place. Since antenna array works only in far field, I need a volume at least 1.5-2 wavelengths to observe the scattering of waves in material. This is why I have enclosed antenna array in tall cylinder shape.
When I try to mesh the model, the DOF number is too large for computer to solve (because of large permittivity of material). Are there any ways to reduce DOF's in particular model?
Thank you,
Arturs
Attachments:
2 Replies Last Post 2016年2月9日 GMT-5 05:35
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Arturs,
the model you submitted doesn't contain a mesh nor a study. I added a default mesh and a default frequency domain study using the f0 frequency in the parameter list.
It solves right away taking 8 GB and 12 min with the default iterative solver and 26 GB and 10 min with the direct solver. Not too bad . My machine is a 3.5 GHz 4-Core i7 with 64 GB. 3D antenna studies will be difficult with less than 16 GB in your machine.
Geometry wise, your model looks like it could be mirror symmetric. I haven't analyzed wether the BCs respect this symmetry.
Cheers
Edgar
--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
the model you submitted doesn't contain a mesh nor a study. I added a default mesh and a default frequency domain study using the f0 frequency in the parameter list.
It solves right away taking 8 GB and 12 min with the default iterative solver and 26 GB and 10 min with the direct solver. Not too bad . My machine is a 3.5 GHz 4-Core i7 with 64 GB. 3D antenna studies will be difficult with less than 16 GB in your machine.
Geometry wise, your model looks like it could be mirror symmetric. I haven't analyzed wether the BCs respect this symmetry.
Cheers
Edgar
--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi Edgar,
thanks for looking into my model! As I mentioned, I usually put the upper layer of cylinder as a dielectric with high permittivity (see the model attached), which makes the meshing more dense there.
The model given is the simplified version, since I am working not only with 3x3 array.
I will try to find some more information for making the model mirror symmetric, though if I need each antenna to be steered individually, I don't think it would be possible, would it?
Every idea or source of information is welcome,
thanks again!
Arturs
thanks for looking into my model! As I mentioned, I usually put the upper layer of cylinder as a dielectric with high permittivity (see the model attached), which makes the meshing more dense there.
The model given is the simplified version, since I am working not only with 3x3 array.
I will try to find some more information for making the model mirror symmetric, though if I need each antenna to be steered individually, I don't think it would be possible, would it?
Every idea or source of information is welcome,
thanks again!
Arturs
Attachments: