Discussion Forum

Hi

When COMSOL operates in Assembly mode, all common boundaries (that are not inside a specific union) will be doubled. By default COMSOL creates automatically identity "pairs" of these common boundaries, you can also select conact pairs, instead, and have COMSOL to cut up common parts of overlapping boundaries that might not necesarily be of the same size.

I believe your issue is that either you have a circle and a bloc just touching, or not even touchng, in which case one should not ask COMSOL to define these boundaries, as it's at best a common "point" and that is not a "boundary" in 2D, it must be a line, or there are no common parts. Then you need to define the contact pairs manually.

One thing to consider, keep the contact pair boundaries of reasonable size, and ensure that the normal to each boundary, scanned over the boundary intercepts the other contact boundary. Check also the doc carefully concenrning tips on how to mesh and to select master slave contacts

--
Good luck
Ivar

Hi Dear Ivar,

Thanks. I got the point. I have manually defined the contact and therefore I will uncheck the create pair in this part. I had the feeling that it works better.

Yours,
Marjan

Hi Dear Ivar,

Thanks. I got the point. I have manually defined the contact and therefore I will uncheck the create pair in this part. I had the feeling that it works better.

Yours,
Marjan

Hi

Note that if you have a clearly defined common boundary, in assembly mode, then using the default identity or contact pair boundary generation saves a lot of time, so mostly they works well with the default settings selected

--
Good luck
Ivar

Hi Dear Ivar,

They are just in contact in 1 point. A sphere on cylinder. I think I am fine with unchecking that.

Yours,
Marjan

Hi

Indeed you should then select the contact manually, but even then you might have some issues with the mesh and the solver. You might need to start with a little gap, test it out

--
Good luck
Ivar

Hi again Ivar,

From speed point you are really right. I again checked and selected contact pair.

Yours,
Marjan

Hi Dear Nagi,

I have a hard spherical indenter, indenting into my polymer film in presence of a boundary load which is laplace pressure and in opposite direction ( The boundary load and pressing are in opposite directions). It is a sphere and flat surface with liquid bridge in between. The liquid bridge produces a pressure on surface that can be mathematically described as a boundary load.

At some point for soft material between 100 MPa - 40 MPa, the indentation is too much and the polymer film behaves like pudding! I mean indentation and restoring force responce are not reasonable. I checked contact elements, contact area, penalty factor, discritization and I could not understand why. It depends very much on this opposite boundary load. For example for the polymer film with 100 MPa, the polymer behaves like a pudding when the laplace pressure is 29.3 MPa and for 40 MPa it already starts when Laplace pressure is only 15 MPa. Can you comment on that?

I ma using weak constraint and inner iteration to calculate reaction forces. I am also using linear elastic material physics.

Thanks a lot in advance,
Marjan

Hi Marjan,

It is hard to tell from the text description without any visual aids. It may be that you are using the linear elastic material model way past its applicable range of deformations. Consider elastic-plastic or viscoelastic model instead. Also check that the magnitude of the Laplace pressure is not unrealistically large since the assumptions involved break down at some point.

Nagi Elabbasi
Veryst Engineering

Hi Dear Nagi,

I have some experimental data under pressure controlled conditions. The meniscus radius of ethanol as what kelvin equation predicts changes from 0.22 nm to 4.5 nm while changing relative vapor pressure from 10 % to 90% and the laplace pressure changes as well. I have some Atomic force microscopy data and the simulation explains the them very well. However it has a limit which I have explained to you. The manuscript of this work is almost prepared and I do not know how to fix or explain that limit.

The real material property is hyper elastic but still with linear elastic material, I expect to get some results. I have defined adhesion force = restoring force- laplace force and I expect that at some indentation the sign becomes inverse (restoring force becomes higher than laplace force) but it does not happen even for 1200 nm indentation on 100 MPa film at 50% relative humidity.for 60%, the change in sign happens at 300 nm which is resonable. However the meniscus radius in 60% is 1.01 nm and in 50%, 0.57 nm.

Just image the force I need to insert my sphere into polymer in comparison with laplace pressure. That is how I have the feeling that it behaves like a pudding.

By the way, Would you please tell me which conditions may break? I would send the model to you tomorrow :) probably as a private message. It would be great if you can give me feedbacks by taking a look at the model.Thanks again :).

--
Marjan