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.
how to excite a plane wave in comsol
Posted May 5, 2011, 12:07 p.m. EDT 11 Replies
Please login with a confirmed email address before reporting spam
Guys:
Anybody know how to excite a plane wave excitation in COMSOL? Really appreciate it
Best
Meng
Anybody know how to excite a plane wave excitation in COMSOL? Really appreciate it
Best
Meng
11 Replies Last Post Aug 29, 2016, 11:27 a.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I believe the easiest is to set it in the Ex,Ey,Ez definition in the main node of the RF physics. Or to look at hte different RF models proposed in the Model library, it depends slightly what you want to do
--
Good luck
Ivar
I believe the easiest is to set it in the Ex,Ey,Ez definition in the main node of the RF physics. Or to look at hte different RF models proposed in the Model library, it depends slightly what you want to do
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Do you mean the Ex,Ey,Ez under the Background section?
I am confused on what a Background plane wave actually is; what is a background wave physically?
--
JB
I am confused on what a Background plane wave actually is; what is a background wave physically?
--
JB
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi,
You can excite your system with a plane wave using a "port" and within the port you can either define your electric field or magnetic field and the propagation constant.
BR
Mohamed
You can excite your system with a plane wave using a "port" and within the port you can either define your electric field or magnetic field and the propagation constant.
BR
Mohamed
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi Ivar,
I still have trouble on defining a plane wave in COMSOL 2D model. I define the incident plane wave in PORT on the left boundary. And PMLs are used for top, bottom and right boundaries. However, the simulation result shows that the plane wave suffers much distortion as propagating even in the uniform space.
I am wondering if COMSOL regards my definition of plane wave as a finite aperture, so that I fail to get the perfect plane wave.
Can you explain the plane wave distortion in the attachment, or offer me any advice of plane wave excitation?
thanks,
Shengli
I still have trouble on defining a plane wave in COMSOL 2D model. I define the incident plane wave in PORT on the left boundary. And PMLs are used for top, bottom and right boundaries. However, the simulation result shows that the plane wave suffers much distortion as propagating even in the uniform space.
I am wondering if COMSOL regards my definition of plane wave as a finite aperture, so that I fail to get the perfect plane wave.
Can you explain the plane wave distortion in the attachment, or offer me any advice of plane wave excitation?
thanks,
Shengli
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
If your objective is to generate a plane wave for the purpose of studying scattering from a finite-extent object, use the scattered field formulation. For example, In a 2D problem, set for Electromagnetic Waves, Frequency Domain, under the Settings tab, you will see "Solve for:" and the default setting is "Full Field." Change that to "Scattered Field."
After you do that, you can set the "Background electric field". Here is where you will specify your plane wave. You will see boxes in which you can specify the x, y, and z components of Eb. Let's consider a specific example. For a plane wave traveling in the +x direction, with wavenumber k0 (which Comsol will compute later, from your still-to-be-specified value of the frequency), and with a polarization (i.e., the direction of Eb) aligned along the y direction, you would specify Eb as follows:
Eb, x component: 0
Eb, y component: exp(-j*emw.k0*x)
Eb, z component: 0
I hope now you can get the idea of how to specify other orientations and polarizations of plane waves. Next, let's discuss the boundary conditions. First, specify "scattering boundary conditions" on the surfaces perpendicular to k. Think of these as the surfaces from which the wave is launched and to which the wave is received/dumped. Second, specify either PEC or PMC conditions for the surfaces along which the wave travels tangentially, as appropriate. For the above planewave specific example, if you have a square in 2D, then the vertical boundaries of the square should be scattering bcs, while the horizontal boundaries would be PEC (since Eb is normal to them). This eliminates the distortion in the wave, that was causing trouble for you.
I've attached an animated .gif showing the resulting wave, which is very clean.
After you do that, you can set the "Background electric field". Here is where you will specify your plane wave. You will see boxes in which you can specify the x, y, and z components of Eb. Let's consider a specific example. For a plane wave traveling in the +x direction, with wavenumber k0 (which Comsol will compute later, from your still-to-be-specified value of the frequency), and with a polarization (i.e., the direction of Eb) aligned along the y direction, you would specify Eb as follows:
Eb, x component: 0
Eb, y component: exp(-j*emw.k0*x)
Eb, z component: 0
I hope now you can get the idea of how to specify other orientations and polarizations of plane waves. Next, let's discuss the boundary conditions. First, specify "scattering boundary conditions" on the surfaces perpendicular to k. Think of these as the surfaces from which the wave is launched and to which the wave is received/dumped. Second, specify either PEC or PMC conditions for the surfaces along which the wave travels tangentially, as appropriate. For the above planewave specific example, if you have a square in 2D, then the vertical boundaries of the square should be scattering bcs, while the horizontal boundaries would be PEC (since Eb is normal to them). This eliminates the distortion in the wave, that was causing trouble for you.
I've attached an animated .gif showing the resulting wave, which is very clean.
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi Robert,
Thank you very much for your kind reply. I can get the desired perfect plane wave by defining background incident wave. Also it can be obtained by setting Scattering Boundary Condition for the vertical boundary and PEC (for TM) or PMC (for TE) boundary for horizontal ones.
Now I can excite plane wave in the background and simulate the pulse scattering by frequency sweep technique.
One more question about the ''solve for scattering field'', when the scatters are disabled, the scattering field should be zero, physically. However, the COMSOL result just shows the full field, plane wave. So could you please explain the scattering field defined in COMSOL?
Best,
Shengli
Thank you very much for your kind reply. I can get the desired perfect plane wave by defining background incident wave. Also it can be obtained by setting Scattering Boundary Condition for the vertical boundary and PEC (for TM) or PMC (for TE) boundary for horizontal ones.
Now I can excite plane wave in the background and simulate the pulse scattering by frequency sweep technique.
One more question about the ''solve for scattering field'', when the scatters are disabled, the scattering field should be zero, physically. However, the COMSOL result just shows the full field, plane wave. So could you please explain the scattering field defined in COMSOL?
Best,
Shengli
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi Fan,
The same puzzle. Some confusion about what "scattering field" really is.
And I am wondering how to excite plane waves in a 2D axisymmetric model.
Best regards,
Jiang
The same puzzle. Some confusion about what "scattering field" really is.
And I am wondering how to excite plane waves in a 2D axisymmetric model.
Best regards,
Jiang
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi Jiang,
Please follow Robert's explanation above or refer to the following link (srdjancomsol.weebly.com/port-boundary-condition-2d.html). Hope it is useful for you.
Best,
Shengli
Please follow Robert's explanation above or refer to the following link (srdjancomsol.weebly.com/port-boundary-condition-2d.html). Hope it is useful for you.
Best,
Shengli
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi Shengli,
Thanks for your reply. There are a lot of useful information in the website you gave.
However, little introduction about tips in a 2D AXISYMMETRICAL model is mentioned.
When dealing with a case such as focusing by a zone plate illuminated by plane waves(3D), large memories are needed. So a 2D axisymmetrical model is considered, for the sake of simplicity. In this case, problem is how to excite plane waves in a Cylindrical coordinate system. And that is where I am stuck in.
Regards,
Jiang
Thanks for your reply. There are a lot of useful information in the website you gave.
However, little introduction about tips in a 2D AXISYMMETRICAL model is mentioned.
When dealing with a case such as focusing by a zone plate illuminated by plane waves(3D), large memories are needed. So a 2D axisymmetrical model is considered, for the sake of simplicity. In this case, problem is how to excite plane waves in a Cylindrical coordinate system. And that is where I am stuck in.
Regards,
Jiang
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Plane waves can exist in 2D and 3D space. Strictly speaking, a 2D axi-symmetric space is applicable only to problems that exhibit axial symmetry, both in terms of geometry, material properties, and the EM wave involved. But... "plane" EM waves can never exhibit axial symmetry!
That said, interestingly, you may still be able to model (sort of) a subset of relevant plane-wave scattering phenomena using 2D axi-symmetric FE models. How? Well, there is at least one, and maybe more, Comsol-supplied RF scattering example (I forget which one) where this is done, but (of course) it is not really a full-blown numerical attack, because Comsol cleverly supplemented the 2D axi-sym FE model with an analytic expression (clever!) to account (in the example considered) for the azimuthal dependence part of the computation. So this is a trick (and it is a good, educational, and praiseworthy trick) but it is not a general-purpose approach. If you really want to compute plane wave scattering from a general target, you are going to have to work in 2D or 3D. And frankly, consider yourself lucky if you can get away with a 2D cartesian space to represent your problem accurately.
That said, interestingly, you may still be able to model (sort of) a subset of relevant plane-wave scattering phenomena using 2D axi-symmetric FE models. How? Well, there is at least one, and maybe more, Comsol-supplied RF scattering example (I forget which one) where this is done, but (of course) it is not really a full-blown numerical attack, because Comsol cleverly supplemented the 2D axi-sym FE model with an analytic expression (clever!) to account (in the example considered) for the azimuthal dependence part of the computation. So this is a trick (and it is a good, educational, and praiseworthy trick) but it is not a general-purpose approach. If you really want to compute plane wave scattering from a general target, you are going to have to work in 2D or 3D. And frankly, consider yourself lucky if you can get away with a 2D cartesian space to represent your problem accurately.
Please login with a confirmed email address before reporting spam
Posted:
8 years ago
If your objective is to generate a plane wave for the purpose of studying scattering from a finite-extent object, use the scattered field formulation. For example, In a 2D problem, set for Electromagnetic Waves, Frequency Domain, under the Settings tab, you will see "Solve for:" and the default setting is "Full Field." Change that to "Scattered Field."
After you do that, you can set the "Background electric field". Here is where you will specify your plane wave. You will see boxes in which you can specify the x, y, and z components of Eb. Let's consider a specific example. For a plane wave traveling in the +x direction, with wavenumber k0 (which Comsol will compute later, from your still-to-be-specified value of the frequency), and with a polarization (i.e., the direction of Eb) aligned along the y direction, you would specify Eb as follows:
Eb, x component: 0
Eb, y component: exp(-j*emw.k0*x)
Eb, z component: 0
I hope now you can get the idea of how to specify other orientations and polarizations of plane waves. Next, let's discuss the boundary conditions. First, specify "scattering boundary conditions" on the surfaces perpendicular to k. Think of these as the surfaces from which the wave is launched and to which the wave is received/dumped. Second, specify either PEC or PMC conditions for the surfaces along which the wave travels tangentially, as appropriate. For the above planewave specific example, if you have a square in 2D, then the vertical boundaries of the square should be scattering bcs, while the horizontal boundaries would be PEC (since Eb is normal to them). This eliminates the distortion in the wave, that was causing trouble for you.
I've attached an animated .gif showing the resulting wave, which is very clean.
Hi Robert,
In 2D, I excite plane wave in COMSOL, temw solver. Scattering boundary condition is set for the left and right boundary. At the same time, PMC boundary is set for top and bottom boundary since the plane wave is TE polarized.
Now I want to use COMSOL to simulate pulse propagation in free space. A plane wave pulse (e.g., wavelength=1.0um, pulse width=10fs) field is input from the left boundary and let is propagate for several microns (e.g. d=10um). And a cutline/cutpoint is put at d=10um to grab the complex data.
My question is how to set the time step based on a fixed space step (e.g., dx=wavelength/10) to get an accurate result, like pulse envelope phase, arrival time for pulse peak? Thanks!
Best,
Shengli