Mesh - Latest Discussions - COMSOL Forums

Ideal mesh for a regular grid export

Fri, 18 May 2012 10:09:42 +0000

Dear Comsol community,

I have a simple 3D electrostatic configuration of two separated point charges in free space. What I want to do is solving the poison-equation and exporting the data in small cubic region between the point charges on a regular grid.

What is the optimal mesh for exporting on a regular grid with about 201^3 points?

Since the error of the FE-method is minimal (vanishing) on the nodes of the mesh-elements I think it would be perfect to have a regular mesh with nodes exactly on the points of the regular grid. I found no way to achieve this yet and I would be thankful for any kind of help giving me a good exported result.

Mesh

Wed, 16 May 2012 06:26:13 +0000

Hi. I have a problem with moving mesh too. Please help me with materials. Where I can find something about it?

can COMSOL set the type of elements?

Tue, 15 May 2012 03:55:52 +0000

I use the ABAQUS before, in that software we can set the type of the element just like 8-node plain strain element. I wander whether the COMSOL can set such thing and how?
Thank you

Advanced Meshing Fuel Cell

Fri, 11 May 2012 09:31:20 +0000

Hello. I'm an undergraduate research student working on a simulation of a fuel cell. I'm a beginner at Comsol and the physics of fuel cells are complex so I've been trying to use the high temperature pemfc tutorial as a guide and fitting that setup to my model. I've had some success and a lot of failure. It seems that most of my failure is a result of the meshing process. I've tried several techniques and contacted customer support several times but, have had no luck in getting the most important model to work. If I break the model down, it will work but, I lose the integrity of the design. In other words, the results are next to useless.

So I'm attaching my file to see if anyone can give me any advice or ideas on what to try next. Any tips, anything.

Thank you for taking the time to look.

This model has the physics I will use included but, I haven't setup any conditions. I'm just trying to get a good mesh.

moving mesh

Wed, 09 May 2012 12:59:29 +0000

Hello,
I have to model the deformation in an soft object pressed by a rigid punch.
Which option would capture the contraction in moving mesh physics: deformed geometry or moving mesh?
Morevover, is moving mesh physics necessary in this model?

Thanks
Massimo

Air flow through elbow channel

Wed, 09 May 2012 08:38:40 +0000

Hi all!

I have problem with simulation of flow of air through the elbow channel. I have the box 500x50x80mm with the fi 50mm inlet from 50x80mm side and whole base (500x50) is an outlet with grill which should ensure flow only in normal direction (see attached model).

The task I am facing is to obtain the pressure distribution at the outlet.

I think my problem is connected with proper meshing of the problem and setting the boundary conditions, but still I cannot obtain any solution.

The usual errors I get are that the limit of Newtonian iterations have been reached.

Could you help me with making this model to run?

Best regards,
Michał Radziwon

moving mesh on 2D surface embedded in 3-space

Tue, 08 May 2012 12:27:34 +0000

I am trying to make a reaction-diffusion model on a growing cylinder.

When I choose 3D model with General Form PDE and Moving Mesh, I can create a cylindrical surface but it does not appear as a "domain" in the Moving Mesh --> Free Deformation options.

Attached is very simple example.

I see that Moving Mesh works in strictly 2D domains (when starting with a 2D model) and also with fully volumetric 3D domains (when starting with a 3D model). Is there a way to using Moving Mesh on a 2D domain embedded in three space?

If not, what is the best alternative? Must I create a "thin" shell that still has non-zero volume and run that mesh? This leads to problems with very small meshing lengths along the "thin" direction.

Other ideas?

Thanks in advance for any suggestions!
John

electro-thermal analysis of micro-heeater

Mon, 07 May 2012 04:07:32 +0000

sir,when i tried to simulate it it is showing some errror, and i could not find the error .can u please help me to solve this structre..

coupling of ALE and laminar flow

Sun, 06 May 2012 09:47:50 +0000

Hi
I'm trying to model the effect of a moving object (wall) on the flow. I was able to move my walls as I wanted without the coupling, but when I added the laminar flow to the model it all went wrong. My question is how can I do this properly. I tried to give the moving walls the "moving wall" boundary conditions and I put the velocity of the wall (as I gave in the moving mesh section) and it gives me an error:

Nonlinear solver did not converge.
Time : 0
Last time step is not converged.

What am I doing wrong?

Thanks a lot
Hemi

iterative mesh

Fri, 04 May 2012 04:03:44 +0000

hi,
I tried to simulate a pressure reflection at acoustic pressure model.
when I build the mesh definitions i got "out of memory" error
I know there is an option to do iterative mesh / or to define an iterative solution that automatically considering
an iterative mesh until the wanted mesh.

HOW I DO THIS AT PRESSURE ACOUSTICS?

thank you

How to form Jacobian

Thu, 03 May 2012 18:34:50 +0000

I have measurement nodes in my geometry,I want to form a Jacobian matrix which has change in measurement value with respect to change in coefficient of pde at individual node position.

Please help me solving the problem.

Mesh query regarding open boundary FSI simulation

Thu, 03 May 2012 12:02:03 +0000

I've got a model in FSI which consists of two inlets to a central reservoir with a membrane on top.
When I use normal mesh settings, the mesh quality at the inlets is horrible.
Am I forgetting a boundary condition? Or is there something I can do regarding the mesh?
I apply simple open boundary conditions to the inlets.
See attached picture, I've filtered to show only the worst mesh quality,

any idea how to improve this mesh?

can comsol support multilayer shell?

Thu, 03 May 2012 05:31:22 +0000

Dear all,

I am trying to model a membrane using shell element because thickness is quite thin compared to length and width. But the membrane is lamated with three to five different materials and thickness at different location. It seems impossible to use 3D solid model that will lead too many mesh elements. Do anyone know how to do multilayer shell in comsol?

STL Geometry & Meshing

Mon, 30 Apr 2012 13:30:18 +0000

I have been working on importing a geometry from an STL file; it's a thing 3-D object, where the length and width are on the order of 10^5 m and the thickness is 10^3 m. I originally had problems where the finalized geometry contained strange overlaps that came from some interpolation within COMSOL. I have finally figured out how to import without these problems.* Now I have another concern.

The intent of this is to build a Java file that I can run within a Fortran model with a variety of geometries, so I ideally want something that is very general and not gui-based. The import geometries will always have a top surface, a bottom surface, and some side faces that are perpendicular to the y-axis.

If you look at the model, you can see that the imported geometry still contains all of the faces defined within the stl file. I cannot get rids of this segmentation, which is the problem now for meshing and for setting boundary conditions.

Is there a way to combine all of these segments together? They are all treated as one domain within the geometry. I have tried converting to solid and a variety of other processes dealing with the geometry. But it seems that when I go to mesh, the mesh is over each small unit and really takes quite a bit of time to get going.

Second question, is there a way to write in the face number when I create the geometry. Since I want something that can be generalized, but I will always know the number of faces/facets, numbering them would allow me to easily group them into the surface necessary for the boundary conditions that I want to apply.

Unfortunately, my file is too large. I'll try to figure this out so that i can attach it. Probably clearing the mesh will solve that problem.

*Just for your information, I have attempted a variety of other methods to bring in this geometry, including an interpolation function (from a file) with a block & parametric surfaces for the geometry. I never got this to work successfully to make something 3D, though it worked for either the top or bottom just fine (see the mention of the interpolation in the first paragraph). I've also gone through many, many permutations of stl files, and this seems to be the best.

Calculating volume

Fri, 27 Apr 2012 13:38:40 +0000

Hello:

How do I calculate volume of an object in COMSOL?
Please see attached file for a simple example.
I have a sphere which I expand using ALE. I want to know the volume change, i.e., volume at the beginning and volume after expansion.
Thanks
sushant

High pressure around rotating domain boundary

Fri, 13 Apr 2012 10:10:27 +0000

Hi, and thanks for reading.

I am modelling an impeller in a turbulent vessel (2d for now). I have an ALE moving mesh rotating with the impeller. When I compute the model the areas of highest pressure and velocity are at the boundary between the moving and the fixed mesh - regardless how large I set the rotating domain, this pattern always follows it.

Has anyone got any advice on what I am doing wrong? The file is 180MB, but if it would help I can attach it.

Many thanks,
Andrew

Eliiminate inverted elements

Tue, 21 Feb 2012 07:36:18 +0000

Hi, I have a 3D problem with lots of terrible aspect ratios in the geometry, which seems to encourage COMSOL to create a few inverted elements.

When I run solver, it says it takes care of these by using linear elements, whatever that means.....

Could this be done at the meshing stage - to have some intermediate processing stage which checks the geometry for any mistakes before I then go and run solvers? E.g. at the moment I'm running Laplace equation, but then next it'll be advection, etc. Would be nice to get the mesh cleaned up before that.

Regards, john

Out of Memory while creating mesh!!

Wed, 11 Jan 2012 02:37:36 +0000

Hi,
I am trying to create a 2D model, wherein I import geometry from AutoCad.
The Problem is that the geometry is so huge, that I cannot form mesh on it, even with a very high element size.

Now, I cannot use symmetry BC, because it isnt symmetric bout any axis and I cannot scale it or redesign it.

So is there any way, I can solve this?

Is there a way to increase the memory allotment of comsol?

Will running the program on a supercomputer cluster help solve the problem?

Thank you,

-Rohan.

scripting import and interpolation of solution

Sat, 06 Aug 2011 13:40:23 +0000

Hi all,

I'm attempting to automate a series of time-dependent 3D simulations, in which the geometry changes from one simulation to the next, but the final distribution of the dependent variable from one simulation should be the initial condition for the next one.

The geometry is set by a matlab script, but due to some memory issues in COMSOL (v 4.1, I haven't rewritten the code sonce 4.2 to try whether they've been solved) the only way to convert the matlab-made geometry to COMSOL is piece by piece, exiting and entering COMSOL after each bit, so as not to encounter the memory problem. Once the geometry is assembled, it is saved as a mphbin file.

Now here's my problem: I'd like to take this saved geometry, mesh it, and interpolate the previous solution onto its mesh. I cannot find a way, however, to import the solution exported by a previous COMSOL simulation. I could perhaps simply read a solution report to a set of matlab variables (coordinates and data) and try write a function using matlab's interp3 to set the initial values from that as f(x,y,z). I'm unclear, however, as to the proper way to set the outputs up so that the they are in the right shape to be used as initial values.

Alternatively, I could try to open up the model with the solution, import the geometry and try mesh it, then interpolate the previous solution to create the initial values with mphinterp (by the way, I have the same issue as with interp3 here: I'm not sure how to reshape the mphinterp outputs so they can be used as initial values), and solve on the new mesh. I'm not sure which of these approaches, if either, works with the way COMSOL is conceived.

Does anyone know of a successful approach to this or a similar problem?

Thanks,
Martin

calculating average element size

Thu, 17 Feb 2011 23:29:22 +0000

Hi,

I am trying to characteise the effect of mesh quality and element size on the solution of a model. I have generated a standard and several refined meshes using an external software (Simpleware) and imported them to Comsol.

I am using q = meshqual(fem.mesh); to calculate the mesh quality.

Is there a similar command or calculate the average element size?

I have tried a subdomain plot of "h" the element size, exporting the plot data to Matlab and calculating the average. However I have noticed the number of elements in the output file is alot more than the real number of elements.

I am not sure how the export plot data works in this case? Is it exporting the element size at each vertex?

Thanks

Amr