3d - Latest Discussions - COMSOL Forums

Convergence issue when simulating harmonic propagation with assemblies (3D)

Mon, 14 May 2012 18:34:16 +0000

Hi,
I want to model harmonic propagation through a 3D multilayer. We use the rf module of comsol 3.5. We enclose the geometry, which is essentially a quarter of an inverted square-based pyramid, with PML at the top, and use perfect magnetic and electric conductor on each pair of parallel side boundaries. As we make the wave to enter through an inner boundary we create a pair and thus, we use assemblies for that. The problem is the slow convergence, in case it happens to converge. We follow the instructions of the rf user guide regarding copying the meshes at each side of the pair in order to avoid using the Vanka preconditioner. We checked the rest of the model by removing the top PML and making the wave to enter through the top boundary, and the model is solved within a decent period of time. Can you help me with some tips to speed up convergence in case we use assemblies?
Thanks a lot.

A problem occure when building subdomain in AC/DC 3D

Mon, 12 Dec 2011 21:41:18 +0000

Hi everyone:

I am trying to build a self-designed transformer using the 3D AC/DC COMSOL. Because the whole thing contains both large scale and small scale, I built two subdomains to generate the meshes. Now the meshes works but it seems that the subdomain cannot transfer any flux to the other domains. I tried to modify the bondary conditions but it does not work. The software cannot solve the domains out of one subdomain. I think there must be something wrong with the bondary, and it is not continous. I wonder what I can do to make it continous? Thank you very much!

convergence speed

Fri, 09 Dec 2011 17:36:54 +0000

Hello,

In some previous FEA software I looked at, there was an option to limit the degrees of freedom to speed up solution times. In a 3D model for example, one could only look at movement along the y axis, ignoring the x and z. I am wondering if there is a way to do this in COMSOL, and if so, would it be useful to increase convergence speed?

On a related note, are there any ideas on what convergence speed should I expect from a 3D model of laminar flow interacting with a solid, very simple geometry, on 8GIG 64 bit Windows 7? I realize it's hard to say, but some of my models have been running for days, so I'm wondering if I should expect convergence in days or in hours?

Thanks,
Anton

Optical Fiber_3D Mode Analysis

Tue, 20 Sep 2011 06:23:57 +0000

Hi,

I'm studying the design for the optical waveguides. I wanted to obtain the results of 3D modal analysis of the optical fiber by using COMSOL (Model wizard - 3D - RF Module - EMW). And I selected the frequency domain in the study step. In the Geometry, the optical fiber was drawn, I set up the index of refraction of the core and clad in the Materials.
I made 2 Ports, and set those to the both side of the core of the optical fiber. And after that I formed the Mesh with Free Tetrahedral 1, fixed the frequency range from 3e14 to 4e14(the space is 0.1e14), and compute the Study.
(Referred to the h_bend_waveguide_3d of the model library, I made Parametric Sweep.)

Here, my question is as follows.

1. Why doesn't figure which the electric field shows normally come out properly?
2. The port seems to have to be set up in the optical fiber as the Coaxial,Why is the error generated if it does with the Coaxial? (tarnally, If it does with the Rectangular in the Type of Port, there is no error. Of course,but figure of the electric field came out wrong.)
3. Is the mistake that I selected to choose the frequency domain in the study step? In that case, Do I have to select the Boundary Mode Analysis or other thing?

The COMSOL seems to be so difficult in understanding. Please,a advice is asked I for.
(I attached the mph file which I made in the version 4.0a of the COMSOL)

Integrating two RF application modes

Sun, 21 Feb 2010 10:22:11 +0000

I want to simulate a 3D waveguide geometry using 3D transient wave propagation mode of RF module. The input field at one of the far cross-sections is an eigen-mode of the waveguide.

I create a multiphysics models with both 3D Electromagnetic waves transient wave propagation and 2D perpendicular waves application modes included. The latter is used to define the cross-section of the waveguide which is then simply extruded into a 3D object for 3D transient wave propagation.

After solving the eigen-mode using 2D perpendicular waves analysis, I want to put in the generated field distribution in one cross-sectional boundary of the extruded 3D waveguides. But the problem is how to access the 2D field distribution in this extruded 3D geometry?

Intuitively, I tried defining Extrusion coupling variable in the 2D geometry and selected the desired boundary of 3D geometry as destination. But this doesn't work The variable is not recognized in the scope of 3D geometry.

I will really appreciate if someone could inform me on how to make do with this thing as I am really exhausted searching through manuals in vain. Please let me know if I need to clarify the question or settings more.

Thanks in anticipation.