boundary conditions - Latest Discussions - COMSOL Forums

One basic and fundamental question about boundary conditions

Fri, 10 Feb 2012 09:40:36 +0000

I feel confused about the introduction of Lagrange multipliers in generalized Neumann boundary condition. Why do we need to have extra dependent variables? I checked relevant items in documentation, and found the following description:

"When Dirichlet boundary conditions are introduced, the finite element algorithm makes a corresponding change to the Neumann boundary conditions so that the resulting model becomes solvable."

It sounds like there exists certain inconsistency within Neumann and Dirichlet BC's, making the equation system unsolvable.

If anyone knows about these details, could u please explain it? Thanks a lot! I really feel unsafe to run simulations with doubts in underlying systems.

free finite prescribed displacement

Mon, 16 Jan 2012 21:43:46 +0000

Evening everyone,

Is there a way of modelling a finite free displacement along a direction, as in the case of components attached via bolts with axial clearance?

For example, to allow a max 0.001m free displacement in the "x" direction I've tried to apply a prescribed displacement u0 of 0.001*(u>=0.001) but it appears not to work.

Many thanks for any suggestion.
George

Best practice for conditionaly switched boundary condition (simulations with multiple steps)

Tue, 22 Nov 2011 10:03:58 +0000

Hello!

I am trying to model the coupled heat and mass transfer of a refrigeration cycle. The model includes distinct steps with different boundary conditions. Switching times cannot be fixed but depend on physical quantities (e.g. pressure at a certain point). Now, I am wondering what is the best practice to implement these conditional switches in the boundary conditions. I tryed different approaches documented in the forum, knowledge base or conference paper that all have drawbacks. I feel like this topic has not been fully discussed, especially for version 4.2 (see http://www.comsol.com/community/forums/general/thread/6746/).

The simplest approach would be hard switching (BC1*(p<p1) + BC2*(p>=p1)). This leads to long solving times due to unsteadiness of the BC or (in my case) to infinity switching back and forth (negative feedback between BC and switch condition).
This approach can be further refined with a) smoothed step functions (http://www.comsol.com/support/knowledgebase/905/) or b) a complicated procedure that uses a sub domain to get "one-way-only" step by integrating a positive pulse of a switch variable defined by a differential equation (http://www.comsol.com/conference2011/usa/abstract/id/10337/lam_abstract.pdf).
Nevertheless, even with all these laborious refinements a well converging model needs manual adjustment (smoothening etc.) when ever parameters of the model are changed. This disqualifies the approach for parametric sweep, hence this is exactly what I want to use the model for. Furthermore, solving times are long, as the solver wants to calculate the moment of switching with all precision, though these moments are not interesting for me. (Simply spoken: the approach seems improvised, a bit von-hinten-durch-die-Brust-ins-Auge)

A second approach, that has sparsely been mentioned in some threats in this forum (first link above): a "save solution" in between the steps and a (automatic?) restart (of a second solver?) with the new BC and the solutions from the end of the step before as new initial values. I still have not found how to implement this in 4.2 (neither by searching here nor by trying on my own). This seems promising and more straight forward. However, I am not sure if post-processing becomes a issue (haven't yet been able to try;).

A third approach includes a matlab frame, calling the model with different BC for each step. The problem: no matlab available. (I remember some one writing that matlab is not necessary anymore with Comsol 4 ...)

Is there a way to implement approach 2 in 4.2, or are there even better solutions to this, I guess, not unusual problem?

Thank you!

Boudary conditions for 2D modelling of superconductors with in plane currents

Tue, 11 Oct 2011 07:18:17 +0000

Hi

I am trying to model the current distribution along a superconducting cable. I am using the general PDE module so I can get away from the problem of circular dependency of conductivity and the current by calculating the magnetic field distribution and then calculating the current from that. This works well with the current perdendicular to the modelled plane but I am now trying to model the current in plane either as a slab or as a wire using 2D-axial symmetry. With the current in plane I either get the error message "may have reached a singularity" or "recoverable error i n residual, but solver could not recover".

Has anybody successfully built a similar model? What boundary conditions could I use to allow the solver to converge?

Thanks.

Using the Interpolation Function to Pass Boundary Conditions

Tue, 19 Jul 2011 18:09:14 +0000

I am a new COMSOL user, and I was hoping to understand how to properly use the interpolation function to transfer boundary conditions between models. I am working on a microfluidic model which is too large to mesh and solve in a single step, and so I need to break the model into segments, solve them individually, and pass the solution values from the outlet of model 1 to the inlet of model 2. (etc.)

After creating a global interpolation function, I set the inlet velocity values of a second model (with the exact same geometric dimensions) to the outlet values exported from the first model. The model runs and solves, but the solution is unexpected for several reasons:

1) The average velocity over the outlet surface of model 1 is more than 1000x greater than the average velocity over the inlet surface of model 2.

2) The highest velocity magnitude for model 2 was in the lower left corner of the conduit, not centered as in model 1.

A very simple model and the data file used for interpolation are both attached (the excel file was saved as a tab-separated text file). The data file was creating by exporting the velocity data from the outlet surface of model 1.

Any advice would be tremendously appreciated.

Thank You!

Electric Field Question

Thu, 31 Mar 2011 07:19:39 +0000

Hello,

I have been creating models of electric fields due to objects in air - I usually use a cube or sphere around the object to which i assign the material AIR.

My question is regarding boundary conditions: What is the most physically realistic boundary condition to apply to the outer edge of the AIR in my models? Is is Zero Charge, or Ground or something else?

I have attempted using the infinite elements condition but have always failed there.

Weak form boundary - details

Fri, 07 Jan 2011 11:40:49 +0000

Hi all,

I come back here because I have some question about Weak form
Boundary.

I have a boundary integral equation int_{boundary)u(s)=-u(s).
I use the Weak form boundary to find the unknown u(s) definined at
the surface of the geometry.

Brievely, how does comsol do to solve it by using FEM.
I think that comsol just expand the solution u in a shape functions set
then solve the equation system obtained for each node of the mesh.
I am right ? How does play the test function ?
Someone can explain to me the different steps
used by comsol to solve a problem using
Weak form boundary ?

Thank you very much !

Sullivan

2D Navier-Stokes w/ moving interior boundary

Mon, 13 Sep 2010 18:49:28 +0000

Hello all,

I'm trying to simulate a 2D actuator pump that is a rectangle with a moving interior boundary which moves like a propagating wave constrained at one end. Water fills the two subdomains partitioned by the interior boundary. My goal is to solve the navier-stokes equations and investigate the pressure at the constrained end of the moving interior boundary. My method henceforth has been to try to solve for the mesh velocity w/ appropriate boundary conditions and then solve for the ns variables (u,v) using the stored mesh velocities as input for the ns boundary conditions on the interior boundary.
I have been able to successfully calculate the moving mesh velocities, but when I go to solve for the ns variables, I get an error message relating that the solver "failed to find consistent initial boundary conditions" and that the last time step did not converge. These obstacles, along with many inverted mesh warnings, are frustrating me. Does anyone know of a better way to solve this model/ have suggestions for how to correct these errors?
Thanks,
Robert Polackwich

Plotting integrals of two boundaries

Wed, 14 Jul 2010 21:18:29 +0000

I'd like to plot the integrals of two boundaries divided by each other. Is there a way to specify the boundary number within an expression?

Thanks!

Change boundary conditions during a transient solution

Thu, 24 Jun 2010 06:29:16 +0000

Hello,

in an acousitc transient simulation, I would like to change a boundary condition during the simulation. I would to impose on this boundary:
1. a pressure condition during the 100 first time steps
2. an impedance boundary condition during the next time steps

Is it possible? With the solver sequence? How to do it?

Thank you. Pascal

How to distinguish between inner and outer bounadry conditions.

Tue, 08 Jun 2010 10:24:43 +0000

Hello!

I have created a MATLAB script to set up 2D and 3D RF models.
To assign the correct boundary conditions I would need to be able to distinguish between inner and outer boundaries.

Is there any way to find out which index in fem.appl.bnd.ind corresponds to an inner and which to an outer boundary?

There should be a way, because Comsol in GUI mode is able to distinguish between the boundaries, however, I couldn't find a solution yet.

Thanks alot for your ideas!!!!!

Clemens

sx_smsld in boundary conditions

Tue, 16 Feb 2010 14:16:53 +0000

Hello everyone,

does anybody know how to apply sx_smps expression in periodic boundary conditions???
It works with displasement u but not with stress...

I have a 3D cube (with sphere cutting inside) whitch is periodic in x and y directions and i need to calculate stiffness of this structure...

in help I found examples of periodic boundary conditions only for displasement but I need to apply pressure or stress..

please help me with this..
best regards,
Aleksandr

"boundary integration" question

Wed, 20 Jan 2010 17:02:40 +0000

Hi all, I would like to make sure my understanding of 'Boundary Integration' in Postprocessing is correct.

Assuming we are doing boundary integration of a pressure expression, P, with unit [Pa], and we can get the result with unit [N/m].

And I think we should get the same result if we use the method below.
1. Find the pressure P [Pa] from Domain Plot Parameter -> Line/Extrusion;
2. Find the interval size [m] depending on the mesh quality from Domain Plot Parameter -> Line/Extrusion;
3. Find the force density [N/m] in each mesh interval;
4. Finally, sum all the force density up.

Is that how COMSOL does the boundary integration numerically? Thanks.

C.M.S.

Changing Boundary Conditions with Time

Sun, 30 Aug 2009 11:59:55 +0000

Hi,

I am implementing a model which implements electro-osmotic switching. I have a "Y" shaped channel and after so many time steps I need to make one of the outlets change from having an electric potential (boundary condition - electric potential) to having the boundary condition of electric insulation.

Ideally the method should be able to switch back and forth between the two. It is important that this transition happens seemlessly in a single simulation, as I need to monitor the switching time and velocity profile.

I have looked at a scripting approach, such as:

fem.sol=femtime(fem, ...
'init',init, ...

but I am confused about the 'tlist',[t_st:0.1:350] parameter! Would I have to first simulate from 0:0.1:0.1 and save the solution with the boundary condition set in one of the conditions previously described, and then run 0.1:0.1:0.2 .... and so on ....?

What is the neatest way of doing this?

Regards,
Tom