3.5a - Latest Discussions - COMSOL Forums

Airfoil in Air using FSI

Wed, 16 May 2012 22:31:14 +0000

Hi,

I am trying to compute an FSI simulation of an airfoil with a constant wind flowing from the right to the left with slip wall condition. The model will converge if the Re number is low (less than 10^5). As I increase my Re, the model will not converge. In additional, my inlet ramp flow velocity will need to be increase very slowly before it reaches constant. I am wondering if there is something wrong in my setup? Or, can FSI be used for high Re.

I am wondering if anybody can point me to the right direction.

Thanks a million in advance!

Ting

What is P_chms?

Wed, 16 May 2012 09:25:04 +0000

I noticed this variable in some equations auto-generated when importing parameters from the Chemical Reaction Engineering Library. If I have an application mode that has pressure as a dependent variable (say General Laminar Flow) are "p" and "P_chms" the same thing?

Simple Acoustic

Tue, 15 May 2012 07:53:45 +0000

Hello,

I am trying to model a 2D acoustic time harmonic square (10x10[m]; left bottom corner at (0,0)).
At the top the pressure is 101325[Pa] and now I just wanted to get started with the gravity force as a kind of a load, but how can I implementate it to get a continous rise of the pressure?

The 2nd step should be to create a point-source which is located at (0,0). But this is as well unknown for me yet.

Thanks for any hints on this.
Nikolaus

error 2100

Mon, 14 May 2012 07:50:11 +0000

Hi,
I'm a comsol 3.5a user.I want analyse a cantilever composed of several layers.I draw 4 layers on my system and,when I add an other layer, error 2100 :'internal error in geometry decomposition' appears. I don't know why!
If someone can help me.

Thanks

Constitutive relations (coupling E and H)

Sat, 12 May 2012 18:54:44 +0000

I am trying to simulate a medium where the constitutive relations are given in a more general form, as follows:

D=epsilon*E + alpha*B
H=(1/mu)*B + beta*E

I do not know how to implement that in comsol 3.5a.
Can somebody help me with this?
Thanks
Ben

Average Temperature in a Subdomain

Thu, 10 May 2012 20:18:24 +0000

Hi,

I'm running a multiphysics simulation in COMSOL (3.5a), using Conduction and Convection and Darcy's Law. There are three subdomains of interest to me, and I am trying to obtain the average temperature in each, as well as in all 3. However, the results do not make sense.

To obtain the average temperature, I measure the volume using the geometric properties button (agree with hand calculations), then select the subdomain, and under 'post processing', select 'Subdomain Integration'. The expression I use is 'T/vol' where 'vol' is the volume of the subdomain in question. These results do not make sense - the subdomain that ought to be the 'hottest' both logically and based on the temperature slices is in fact the coldest based on this calculation method.

Anyone else have any problems with this method or see some misbehavior?

Thanks,

Nick

Postplot calling an external matlab function --> no plot

Thu, 10 May 2012 18:29:06 +0000

Hello, (this was sent to COMSOL support, too.)

I'm having an issue in version 3.5a using postplot function: in matlab, when postplot tries to surface plot a function that is defined as an external matlab function, a figure window opens but no plot appears. There, I notice that matlab uses 100% CPU resource. What could be the problem?

Here is a specific scenario that I tested:

- In a 2D electrostatics setup, make some region (subdomain) between two constant-potential boundaries have a spatially-varying permittivity defined as an external matlab function.
- Solve for the potential V (done in 'comsol matlab' environment).
- Save the project as a .m file
- In matlab, load the .m file
- Try making these plots:

1. Plot the solved potential V via

h=postplot(fem,'tridata','V');

A correct plot of potential distribution appears.

2. Plot the permittivity via

h=postplot(fem,'tridata','epsilonr_emes');

Matlab goes to work (100% CPU usage), but no plot appears (indefinite wait).

This behavior was observed in both linux and windows versions.

What could be the problem?

Thank you in advance,

Do-Hoon

Question about the microring simulations.

Sun, 06 May 2012 00:29:35 +0000

Hi,

I didn't mention about my email address which is as follows:
photon.ktlee@gmail.com

Thanks so much.

Best regards,

Kitae

Question about the microring simulation.

Sun, 06 May 2012 00:17:45 +0000

Hi Jiawei,

I am trying to run some simulations to see the optical field distribution of the microring resonator.
But, I couldn't get any reasonable results yet from the calculation.
The PDE weak form, subdomain module was employed for the simulation.

When I saw your discussion, it seems that you already finished some parts of the microring simulation, right?
What kinds of module did you use for that?

If you don't mind, could you please send the .mph file of the microring to examine it by myself?

Thanks a lot for your kind consideration in advance.

Best wishes,

Kyu-Tae

The order of coordinates in interpolation function

Thu, 03 May 2012 17:48:34 +0000

Hi,

I'm trying to define an interpolation function in COMSOL 3.5a using txt file. My data structure is as follow:
% x y z Data
x1 y1 z1 D1
x2 y2 z2 D2
...

It's a 3-D problem, so there are 3 coordinate values. What should be the order of those coordinates? Shall I store them up in ascending order? How about arbitrary order?

Some of the x values are the same. If an ascending order of x-coordinate is required, then how to deal with such a problem?

Thank you very much

Dan

use of planes

Thu, 03 May 2012 12:49:51 +0000

Under what circumstances 2D plane stress and 2D plane strain modes should be used,
respectively?

comsol post processing

Thu, 03 May 2012 12:48:13 +0000

Please explain the following postprocessing plot modes: surface, contour,
boundary. anyone there please

boundary conditions of applying an alternative current

Mon, 30 Apr 2012 10:07:03 +0000

Hello everyone,

i need to simulate a method which uses AC signal in a relative low frequency of 128 Hz. I have run the simulation of the same device (test Wenner) with DC giving not bad results, but now i need to simulate with AC, as it really is, hoping that it will give similar results to that of DC (since it works in low frequencies). So, i changed my module and i went to AC/DC module, on order to use the module Electric currents of quasi statics, Electrics as you have proposed to me in an earlier question. The thing is that i create points that they have to inject on to a block an alternative current, i.e. sinusoidal. However, when i go to the point boundaries i do not get the option for currents but only potentials.

If instead of the points i create surfaces i go to the surface boundaries and i write an expression for the inward current flow : 0.01sin(8t) for exemple, will it comsol understand automatically the form of the signal or should i also change the solver parameter to time dependent and adjust the time stepping etc? (on the same window i am using the time harmonic currents and not the transient electric currents, is that right?)

Thank you in advance!!!

Marilena

FSI+Rotating mechinery

Sat, 28 Apr 2012 06:28:01 +0000

hello, I am ready to simulate the strength of strired tank of baffle. I used the Rotating Mechinery model and the plan strain model. when I initial the model, the comsol 3.5 gives error:Messages:
Failed to evaluate expression
- Expression: cos(-2*pi*rpm*t)*X-sin(-2*pi*rpm*t)*Y

Failed to evaluate variable
- Variable: t
- Geometry: 1
- Boundaries: 14,15,16,17,18,19,20,21,22,23, ...

Can anyone help me fix this error?

Thank you!

Solution not converged

Thu, 26 Apr 2012 15:17:20 +0000

Hi, all

Recently I'm encountering convergence problems. Physically the problem has decent solution. So I assume the problem lies in numerics. However it does not seem easy to find the reason for the divergence...

The complication of the problem may be one boundary condition. It's a model coupling an RF module (2D, solving Hz) and a PDE. In the PDE module, a boundary condition has electric fields in the expression. For some direct solver, and a particular value of a parameter in the model, it can converge. When the parameter takes other values, it always diverge, leading to non-physical solution with several singularly large values.

For some simple iteration methods, divergence may be avoided by tuning certain parameters. Similarly can we change something in the solver settings to secure convergence?

Thanks a lot!

reaction rate for physical absorption

Wed, 01 Feb 2012 08:32:23 +0000

I am using Comsol Multiphysics 3.5a to model a physical absorption. I chose the che chemical engineering module/ mass transport/ convection and diffusion. I want to know how to set the reaction rate in Sumdomian settings for physical absorption? Hope any professional friends can give me some help. Thans a lot!

Question about example- "Plasmonic wire grating"

Fri, 15 Jul 2011 14:38:25 +0000

Dear all:
I have a question about the example "Plasmonic wire grating".
In this model, I saw that the Floquet boundary condition applied at the left and right boundary is that kF=kax (wave vector), which is equal to emw.k0*sin(alpha). However, we know that the wave vector depends on the wavelength, and wavelength depends on the refractive index of the medium. In that example, the lower sub-domain is a dielectric with refractive index of nb=1.2. So in the periodic boundary settings, why don't we set the kF=kax*nb for the lower sub-domain?
This question has bugged me for some time. I really appreciate for your idea!

Extruding to 3d

Wed, 29 Jun 2011 18:00:15 +0000

Hi,
2 questions:

1) I am creating a ring in 2d, and it seems perfectly circular, but when I extrude it to 3d, it suddenly gets slightly jagged sides and becomes a sort of polygon. I tried to fillet it in 2d and then extrude, but it makes no difference, the 3d still has edges. Anyone know how to fix this?

2) Is there any way to put a 2d face on a 3d object? For example, on a cylinder, I would like to put a small circle on its top face to model a laser's heat deposition. I think this would let me impose heat transfer conditions on the boundary. Other suggestions are also welcome.
Thanks

How to calculate Extinction efficiency in comsol35a?

Fri, 06 May 2011 06:59:12 +0000

Dear Users,
I am new to comsol. I have tried the example of a gold nanosphere in RF module. I have obtained directly the scattering cross section, extinction cross section and absorption cross section in comsol. Now my question is that, how to find the extinction efficiency, scattering efficiency and absorption efficiency?

Thanks in advance.

Best regards,
Cecil

Determine Arrenhius damage integral for a solution

Wed, 12 Aug 2009 18:05:07 +0000

I am trying to apply an Arrenhius damage integral DURING the solution of an electrosurgical problem using the Bioheat Transfer Module along with the Conductive Media DC module (T and V are the dependent variables) . The integral is of the form:

Adi = int(A*exp(-E/(R*T),t)

where Adi is the Arrenhius damage integral, int stands for integral and A, E, and R are constants. T is the nodal tissue temperature and t is time.

Is there a way to apply this into the Global Equations by taking the derivative of Adi? Thus making the equation:

Adit = A*exp(-E/(R*T)

I tried this, but received an error that T could not be solved for. Do I need to do something in order for Global Equations to access my dependent variables?

Thanks for any help...I am also interested in any other way that anybody can think of!