why PML layer is backed by PEC layer in FDTD simulation

Hi everyone,

I want to see why in simulation in FDTD method, PML layer is backed by PEC?

Hi ali.
this is done,to truncate simulation domain.
in fact,composition of PML and PEC in the end,simulate a open structure.i.e,you don't have any reflected wave,ideally.
many applications, such as scattering and radiation problems, require the boundaries simulated as open space.
you can refer to Taflove book.
best regards...
ghasem

There is an argument about applying ABC or PEC to PML exterior. Conclusion is if PML is optimized, ABC will make very small improvement over PEC since the PML already decayed field to almost zero before reaching its back wall. However, if PML is not well designed or not working at its optimal mode. ABC will compensate.

ghasem_008 said:

Hi ali.
this is done,to truncate simulation domain.
in fact,composition of PML and PEC in the end,simulate a open structure.i.e,you don't have any reflected wave,ideally.
many applications, such as scattering and radiation problems, require the boundaries simulated as open space.
you can refer to Taflove book.
best regards...
ghasem

Click to expand...

Hi,

Thank you for your good description, Actually I have been long working with softwares, but now involving with codes, thats why I was wondering why it is so, because in software that I have been working we only define PML and that is all.

anyhow, my another question is, I have seen also if the structure is symmetric in some dimension they take the PML layer thickness zero, why is that? but they keep PEC .

- - - Updated - - -

rodgerwxh said:

There is an argument about applying ABC or PEC to PML exterior. Conclusion is if PML is optimized, ABC will make very small improvement over PEC since the PML already decayed field to almost zero before reaching its back wall. However, if PML is not well designed or not working at its optimal mode. ABC will compensate.

Click to expand...

Thanks a million,

That is it, actually I read a paper discussing as you say, your last sentence helped me a lot. Actually I have been long working with softwares, but now involving with codes, thats why I was wondering why it is so, because in software that I have been working we only define PML and that is all.

anyhow, my another question is, I have seen also if the structure is symmetric in some dimension they take the PML layer thickness zero, why is that? but they keep PEC .

PEC can also be used as an image. In EM a theory called imaging theory will mirror the EM field as symmetric. That is why they do not need PML but keep PEC. check here for a demonstration of image theory (https://www.youtube.com/watch?v=e69DcWUi7N0)

Hi ali

reading this paper helps you to understand the PML as a very good absorbing boundary condition.

Three-Dimensional Perfectly Matched Layer for the Absorption of
Electromagnetic Waves by JEAN-PIERRE BERENGER
**broken link removed**
good luck

**broken link removed**, this is another major contributor to PML research.