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 The information provided may be out of date.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.

Electromagnetic field distribution of a cylindrical cavity

Please login with a confirmed email address before reporting spam

Firstly, Many thanks for your interest of reading my questions. I hope you could give me your generous advise

I used the physics setting listed as below

*Module:'Electromagnetic Wave(emw)'
*I use 2.45GHz TE10 mode microwave as 'Port'
*Displacement Field:'Relative permittivity'
*Why do 'h_bend_waveguide_2d/3d'?models use 'refractive index' as the 'displacement field'?
*Study:'Frequency Domain'
*I can not use other frequencies except 2.45GHz in the experiment equipment

to simulate a Nitrogen-filled cylindrical cavity's electromagnetic field distribution.

I am not familiar with the physical properties value of the nitrogen and stainless-steel boundary of the cavity.
I just input some simple value into the model I set up. For example,
* Electric conductivity = 0.24
* Relative permittivity = 1.0 for Nitrogen
* Relative permeability = 1.0 for Nitrogen
Maybe I missed the setting of the boundary condition for the stainless steel cavity.(I don't know how to do it or do it appropriately.)

I ran the model and only got zero electrical field distribution inside the cavity( the electrical field is non-zero only at
the input of the regular hexahedron). Why? Can you help me out with this? Many thanks!!!!

Model description:
*a large cylinder cavity A (boundary material:stainless steel, filled with Nitrogen gas)

*a small cylinder B (located in the center of the cavity A, material:stainless steel,
it will hold some material to be heated, the height of cylinder B can be changed )
*a regular hexahedron (2.45GHz TE10 mode microwave will be input at the end of it, acting as wave guide)

I attach the model file 'cylinder cavity_emw.mph' with this post. Many thanks for your generous advise!


6 Replies Last Post 11 sept. 2011, 20:30 UTC−4

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 juil. 2011, 20:42 UTC−4
Please help me.
no one did this before?
I am studying the userguide of RF module again.
I hope I could got some hint.
Many thanks, anyway.
Please help me. no one did this before? I am studying the userguide of RF module again. I hope I could got some hint. Many thanks, anyway.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 juil. 2011, 11:49 UTC−4
Try setting the conductivity to 0. I don't think that nitrogen is conductive at all. When I did this I was able to see results in the Efield plots.

Also, your mesh is set up a little confusingly. Underneath the free tet you have, you added another global condition which is not really helping you that much. I would tell you that you do not need that size feature. I would suggest adding another size feature right below the global one. See my file attached. I added more mesh elements to the port BC. Reason is that if you try and calculate s-parameters you want the port meshed finer. You may have to make it finer than what I have as well.

Your version of s/w is in another language, as I get gibberish when I read the file. I am able to figure out what you did but it would be confusing to a new to COMSOL person.

I hope none of this leads you astray.
Try setting the conductivity to 0. I don't think that nitrogen is conductive at all. When I did this I was able to see results in the Efield plots. Also, your mesh is set up a little confusingly. Underneath the free tet you have, you added another global condition which is not really helping you that much. I would tell you that you do not need that size feature. I would suggest adding another size feature right below the global one. See my file attached. I added more mesh elements to the port BC. Reason is that if you try and calculate s-parameters you want the port meshed finer. You may have to make it finer than what I have as well. Your version of s/w is in another language, as I get gibberish when I read the file. I am able to figure out what you did but it would be confusing to a new to COMSOL person. I hope none of this leads you astray.


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 juil. 2011, 20:48 UTC−4
Dennis, Many thanks for your response and advise.
I used the Japanese version when I set up this model.
Now I changed the language into English and changed the conductivity of the Nitrogen(yes, it should not be 0).
I referenced the mesh setting in the user guide.
I am totally green with COMSOL and trying to learn it.
Looking forward for you further reply.
Best regards
Dennis, Many thanks for your response and advise. I used the Japanese version when I set up this model. Now I changed the language into English and changed the conductivity of the Nitrogen(yes, it should not be 0). I referenced the mesh setting in the user guide. I am totally green with COMSOL and trying to learn it. Looking forward for you further reply. Best regards


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26 juil. 2011, 12:08 UTC−4
Lets start with the meshing part. When you first add a free tet node to a model, you will get two nodes that come up. One is a size node and one is the free tet node. The first size node controls the overall mesh fineness/density. The first free tet node will apply to the the whole geometry unless you change it. You can have multiple free tet nodes in a simulation but that would be a case where you have several domains or areas where you are meshing them separately.

In your case since the overall geometry is simple, you can just use one free tet for the whole geometry. With that said, if you allow the program to mesh with just those two nodes, you will get a mesh density controlled by the first size node. What you did in your model was that you added a second size node underneath the first free tet. In that size node, you allowed it to control the entire geometry. What this means is that you have that first size node controlling the entire geometry and second size node underneath the free tet node controlling the geometry and they appear to be kind of fighting each other. What happens is that the first size node wins out. You did not see an error with meshing because the program was able to do so but it also did not tell you why you were doing it less efficiently. With the way you sent the model, the number of mesh elements shown by COMSOL is 9925. If I disable the size node underneath the free tet node, I then get 9998. The nodes are fighting each other and not improving your mesh. When I posted my first file, I had changed the size node underneath the tet node to be a boundary only, specifically where you had the port for applying RF. In general, when you apply extra size nodes below a free tet node, you are looking to make a particular domain/boundary/edge have a finer or coarser mesh than what the overall mesh will be. Do you understand this?

Meshing is not as simple as it appears. You only get better at it when you do it a lot. You need to work several examples and maybe take some classes that COMSOL offers.

As far as the actual simulation goes, you will need to find out what the conductivity of Nitrogen really is. When I had conductivity set to 0, there were fields present in the output plots. When it goes slightly above 0, the fields change and decrease. I have not taken them time to really look into the problem as I can't spend the time at the moment. I don't think Nitrogen would be very conductive. Since it is in our atmosphere, it is a good insulator so I would expect the conductivity to be low. Maybe somebody else reading this has more insight. I think your problem is set up right but there may be finer details that I do not know about. If I get a moment to really look at this, I will but maybe others can help out.

BTW, what is it that you are trying to do here? Are you trying to measure input impedance of your structure? Determine return loss? Other?

Lets start with the meshing part. When you first add a free tet node to a model, you will get two nodes that come up. One is a size node and one is the free tet node. The first size node controls the overall mesh fineness/density. The first free tet node will apply to the the whole geometry unless you change it. You can have multiple free tet nodes in a simulation but that would be a case where you have several domains or areas where you are meshing them separately. In your case since the overall geometry is simple, you can just use one free tet for the whole geometry. With that said, if you allow the program to mesh with just those two nodes, you will get a mesh density controlled by the first size node. What you did in your model was that you added a second size node underneath the first free tet. In that size node, you allowed it to control the entire geometry. What this means is that you have that first size node controlling the entire geometry and second size node underneath the free tet node controlling the geometry and they appear to be kind of fighting each other. What happens is that the first size node wins out. You did not see an error with meshing because the program was able to do so but it also did not tell you why you were doing it less efficiently. With the way you sent the model, the number of mesh elements shown by COMSOL is 9925. If I disable the size node underneath the free tet node, I then get 9998. The nodes are fighting each other and not improving your mesh. When I posted my first file, I had changed the size node underneath the tet node to be a boundary only, specifically where you had the port for applying RF. In general, when you apply extra size nodes below a free tet node, you are looking to make a particular domain/boundary/edge have a finer or coarser mesh than what the overall mesh will be. Do you understand this? Meshing is not as simple as it appears. You only get better at it when you do it a lot. You need to work several examples and maybe take some classes that COMSOL offers. As far as the actual simulation goes, you will need to find out what the conductivity of Nitrogen really is. When I had conductivity set to 0, there were fields present in the output plots. When it goes slightly above 0, the fields change and decrease. I have not taken them time to really look into the problem as I can't spend the time at the moment. I don't think Nitrogen would be very conductive. Since it is in our atmosphere, it is a good insulator so I would expect the conductivity to be low. Maybe somebody else reading this has more insight. I think your problem is set up right but there may be finer details that I do not know about. If I get a moment to really look at this, I will but maybe others can help out. BTW, what is it that you are trying to do here? Are you trying to measure input impedance of your structure? Determine return loss? Other?

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 juil. 2011, 02:21 UTC−4
Many thanks, Dennis.

I have just experienced the COMSOL Multiphysics 4.1 for nearly one month and actually I was trying to simulated the microwave heating at first (we have a practical equipment here in Japan). You know, I am totally green with the COMSOL, so I thought it could be a little bit difficult for me to realize that. I stepped back and tried to simulation the electromagnetic field distribution in our cylindrical cavity connected with a hexahedron waveguide(TE10 mode, 2.45GHz-fixed frequency).

Unfortunately, I just had almost 0 electrical field inside the cavity. So I summarized my problem and posted here.

Maybe I should keep on hoping to realize the microwave heating simulation. Anyway, I should learn harder from the user guide and some posts in this forum.

I also have studied the microwave_oven model in the COMSOL. I don't know why did they set the microwave port using defined Port Mode Setting instead of by the rectangular Port Type setting. Only because it is a 1/2 model? How to figure out the formula of the Port Mode Setting if I also use 1/2 model in my application.

Thank you very much.
Many thanks, Dennis. I have just experienced the COMSOL Multiphysics 4.1 for nearly one month and actually I was trying to simulated the microwave heating at first (we have a practical equipment here in Japan). You know, I am totally green with the COMSOL, so I thought it could be a little bit difficult for me to realize that. I stepped back and tried to simulation the electromagnetic field distribution in our cylindrical cavity connected with a hexahedron waveguide(TE10 mode, 2.45GHz-fixed frequency). Unfortunately, I just had almost 0 electrical field inside the cavity. So I summarized my problem and posted here. Maybe I should keep on hoping to realize the microwave heating simulation. Anyway, I should learn harder from the user guide and some posts in this forum. I also have studied the microwave_oven model in the COMSOL. I don't know why did they set the microwave port using defined Port Mode Setting instead of by the rectangular Port Type setting. Only because it is a 1/2 model? How to figure out the formula of the Port Mode Setting if I also use 1/2 model in my application. Thank you very much.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11 sept. 2011, 20:30 UTC−4
Many thanks for all your kind help in advance.

These days I kept reading some papers and revising my models according to your
suggestions and the microwave-heating model in COMSOL.

I set up a 1/2 model

Metal wall of the microwave chamber,
TE1.0 port,
Ferrous powder block-I don't believe that there is some suitable physics or properties for powder
Ceramic mold
Upper stainless steel block-it's a load cell for press the powder
Lower stainless cylinder-for holding the powder block and ceramic mold

and used the 'Selection' in COMSOL for latter definitions of the material properties,physics,and
boundary conditions.

I also attached my model set up in COMSOL4.2 in this email, I hope you could open it and please
have a look if you have time.I deleted the solution to reduce the size of the file.Please run it again
and study the 'hot spot' in results.

Now, let me talk the problem I want to solve as below

Material properties(especially, the permittivity and permeability)

according to some trial run
Permeability Permittivity
Iron(Heated sample) 1.0 14-20*j
Al2O3(Ceramic Mould) 1.0 changing(real value and complex value)

I got different results showing that ceramic mold got heated instead of the iron heated sample, and different 'hot spot' patterns.

I changed the permittivity of the Al2O3(Ceramic Mold) such as 0.096,0.96,9.6,20 and 140(for real value)
as well as 9.6-1.2*j,0.96-0.12*j,0.96-0.12*j,96-1.2*j,96-12*j,140-0.1j,140-0.01*j and 140-0.001j(for complex value)

Actually, I really do not have a concise idea of these dielectric properties value. Would you give me some advises or some references about the permittivity and permeability of the iron powder and some ceramics(Al2O3,AlN,SiN).Yes, I knew these dielectric properties got changing at different temperatures as well. Is there any tricky setting in the environment of COMSOL that I could put some function into it? For example, the permittivity : epsilonr =f(T).

Please help me.
Best Regards
Many thanks for all your kind help in advance. These days I kept reading some papers and revising my models according to your suggestions and the microwave-heating model in COMSOL. I set up a 1/2 model Metal wall of the microwave chamber, TE1.0 port, Ferrous powder block-I don't believe that there is some suitable physics or properties for powder Ceramic mold Upper stainless steel block-it's a load cell for press the powder Lower stainless cylinder-for holding the powder block and ceramic mold and used the 'Selection' in COMSOL for latter definitions of the material properties,physics,and boundary conditions. I also attached my model set up in COMSOL4.2 in this email, I hope you could open it and please have a look if you have time.I deleted the solution to reduce the size of the file.Please run it again and study the 'hot spot' in results. Now, let me talk the problem I want to solve as below Material properties(especially, the permittivity and permeability) according to some trial run Permeability Permittivity Iron(Heated sample) 1.0 14-20*j Al2O3(Ceramic Mould) 1.0 changing(real value and complex value) I got different results showing that ceramic mold got heated instead of the iron heated sample, and different 'hot spot' patterns. I changed the permittivity of the Al2O3(Ceramic Mold) such as 0.096,0.96,9.6,20 and 140(for real value) as well as 9.6-1.2*j,0.96-0.12*j,0.96-0.12*j,96-1.2*j,96-12*j,140-0.1j,140-0.01*j and 140-0.001j(for complex value) Actually, I really do not have a concise idea of these dielectric properties value. Would you give me some advises or some references about the permittivity and permeability of the iron powder and some ceramics(Al2O3,AlN,SiN).Yes, I knew these dielectric properties got changing at different temperatures as well. Is there any tricky setting in the environment of COMSOL that I could put some function into it? For example, the permittivity : epsilonr =f(T). Please help me. 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.