Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

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.

PML Scaling factor in COMSOL 4.3

Esrom Aschenaki Kifle

Please login with a confirmed email address before reporting spam

Hi,

I am doing a 3D simulation with COMSOL 4.3 to find the field distribution and the mode of a waveguide where all the cladding, the core and the substrate have different refractive index. I have implemented a PML at the end of my waveguide to avoid any back reflection in to the waveguide and any standing wave patterns which might arise due to this reflection.

But I am still having some reflection from my PML which means it is not optimized. I have tried to manually scale down at the ‘’PML scaling factor’’ option for each of the interface as:

wavelength/nguide/PML_length , wavelength/nsubstrate/PML_length and wavelength/nsubstrate/PML_length

but it didn't help. Can anyone suggest how to properly scale the PML in COMSOL 4.3 for optimized performance? Thank you very much for any help and suggestions.

esrom

8 Replies Last Post 11 mar 2013, 07:30 GMT-4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 feb 2013, 01:58 GMT-5
Hi

do you have some "air / vacuum around your wave-guide?
I do remember from previous simulations that the PML worked better when they surrounded a all air domain

--
Good luck
Ivar
Hi do you have some "air / vacuum around your wave-guide? I do remember from previous simulations that the PML worked better when they surrounded a all air domain -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 28 feb 2013, 13:48 GMT-5
3 things are important:
1. enough mesh elements in the PML
2. same material for PML and the domain it touches
3. change PML type to General if you want to change PML scaling factor and other parameters manually
3 things are important: 1. enough mesh elements in the PML 2. same material for PML and the domain it touches 3. change PML type to General if you want to change PML scaling factor and other parameters manually

Esrom Aschenaki Kifle

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 1 mar 2013, 10:53 GMT-5
Hi,

Thank you for the suggestions.

I haven’t used any air/ vacuum my system is a waveguide Silicone (n=3.4) as core between silica (n=1.45) and water (n=1.33) as claddings.

I split the PML in to 3 for each material in my system so that the PML and the materials match.
I have used a sweep mesh for the PML from the interface of the Materials and the PML with 10 elements. But I am still getting some back reflections?

When you said change the PML to general for manual scaling is it in version 4.3 ? I didn’t see an option to do that in the definition node.

Best regards,
Hi, Thank you for the suggestions. I haven’t used any air/ vacuum my system is a waveguide Silicone (n=3.4) as core between silica (n=1.45) and water (n=1.33) as claddings. I split the PML in to 3 for each material in my system so that the PML and the materials match. I have used a sweep mesh for the PML from the interface of the Materials and the PML with 10 elements. But I am still getting some back reflections? When you said change the PML to general for manual scaling is it in version 4.3 ? I didn’t see an option to do that in the definition node. Best regards,

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 1 mar 2013, 15:05 GMT-5

When you said change the PML to general for manual scaling is it in version 4.3 ? I didn’t see an option to do that in the definition node.


It's in the settings for PML. Geometric Settings, Type: General
[QUOTE] When you said change the PML to general for manual scaling is it in version 4.3 ? I didn’t see an option to do that in the definition node. [/QUOTE] It's in the settings for PML. Geometric Settings, Type: General

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2 mar 2013, 03:11 GMT-5
Hi

then I have a question, if we have different materials, and we split up the PML accordingly to have the "same material" what about oblique vaves hitting the interface in the PML region, would that scatter ?

What is best, set all air around and PML air, or remove the air for large INF domains, and split the PML and give material properties ?

--
Have fun COMSOLing
Ivar
Hi then I have a question, if we have different materials, and we split up the PML accordingly to have the "same material" what about oblique vaves hitting the interface in the PML region, would that scatter ? What is best, set all air around and PML air, or remove the air for large INF domains, and split the PML and give material properties ? -- Have fun COMSOLing Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 4 mar 2013, 10:39 GMT-5
Hi,


then I have a qusion, if we have different mateials, and we split up the PML accordingly to have the "same material" what about oblique vaves hitting the interface in the PML region, would that scatter ?


I haven't seen the back-scattering at such interfaces. Sometimes there is a "surface wave" going along this interface. But seems like scattering BC outside of PML takes care of it.


What is best, set all air around and PML air, or remove the air for large INF domains, and split the PML and give material properties ?


I never set up air all around.For the simulations I do, it won't represent the correct structure. But I guess it should work in antenna-type simulations just fine.
Hi, [QUOTE] then I have a qusion, if we have different mateials, and we split up the PML accordingly to have the "same material" what about oblique vaves hitting the interface in the PML region, would that scatter ? [/QUOTE] I haven't seen the back-scattering at such interfaces. Sometimes there is a "surface wave" going along this interface. But seems like scattering BC outside of PML takes care of it. [QUOTE] What is best, set all air around and PML air, or remove the air for large INF domains, and split the PML and give material properties ? [/QUOTE] I never set up air all around.For the simulations I do, it won't represent the correct structure. But I guess it should work in antenna-type simulations just fine.

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 4 mar 2013, 14:38 GMT-5
Hi

Thanks for your comments ;)

--
Good luck
Ivar
Hi Thanks for your comments ;) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11 mar 2013, 07:30 GMT-4
Dear All


thanks allot for these information, i also simulate the silicon wave guide: brillouin scattering ( interaction between the pump and Stoke waves), i used the PML to prevent the reflection in the wave guide, but still i can not get the scattered wave because i do not know how to simulate the Stoke wave and get the acoustic mode. I used the RF electromagnetic and elastic wave modules.

please any suggestions and if you want i will attach my Comsol file.


Best Regards
Dear All thanks allot for these information, i also simulate the silicon wave guide: brillouin scattering ( interaction between the pump and Stoke waves), i used the PML to prevent the reflection in the wave guide, but still i can not get the scattered wave because i do not know how to simulate the Stoke wave and get the acoustic mode. I used the RF electromagnetic and elastic wave modules. please any suggestions and if you want i will attach my Comsol file. Best Regards

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.