Conceptual doubt: If velocity is zero at the wall...
 

Dear all,

maybe you can help me solving this doubt:

If the velocity of a fluid is zero in the wall, why does the wall end by being in contact with the fluid?

The fact is that I am simulating how a fluid moves near a wall, so I am simulating a very little region near the wall. I am using a velocity inlet with zero speed in the wall with very little velocities when being around 300nm away. But this leads to the fact that I need around 2 seconds for the fluid going through 900 nm, when in real life it is a matter of miliseconds.

The velocity profile is extracted from a macro simulation, and because of this little zone being very near to the wall the inlet profile has very low velocity values.

http://s7.postimg.org/ecebowosb/conditions.jpg

If I haven't been clear in some point, please ask me.

sorry, could you better formulate the question?
If you are working on laminar flat plate you have everywhere an analytical solution, apart the location x=0 that remains a singular point

Dear FMDenaro,

wonder a simulation like the image I posted. Which is nothing more than a submodelling of a macro simulation.

Since the velocity is zero in the wall, the velocity is very little in my submodeling inlet (it is only 300nm height). This leads to the fact that the fluid goes very slow through my model. This is not coherent with the macro simulation, because such a little region is filled with fluid very fast.

Should I try to explain it another way?

Still not clear for me...
1) considering nano-fluidic physics, other effects are relevant that are disregarded in classic fluid dynamics
2) assuming to disregard any effect and using the classical fluid dynamic solution for a flat plate, you have a well-know solution valid from y=0 up to y=+Inf, that is covering micro and macro-regions

Dear FMDenaro,

I will explain you the whole context.

I am tyring to simulate how a fluid fills a nano-cavity.

It is impossible to mesh the full model with a mesh with nanodetail and refining strategy gives bad results because it is not possible to have a mesh with mm size and nm size.

Because of this, I'm carrying out a submodeling:

I have carried out a macro simulation of the fluid filling my macro cavity.

From this result, I interpolate the velocity of the fluid near the wall and carry out a simulation at the nanoscale with the geometry I have posted before.

It turns out, that for filling the nano-geometry I need a lot of time, because I am using a velocity inlet with very low values, which are the values extracted from the macro simulation.

This result doesn't fit the reality, because such a little region is fastly filled! But how can it be fastly filled if the velocity of the fluid near the wall is very low?

if you want to study the time-dependent filling in a cavity, you must consider this region covered by two different fluids, taking into account the relative compressibility (In the nano-fluidic you must also consider other effects).
Are you using the incompressible model? this would lead to an instantaneous propagation of the fluid velocity everywhere..

I am using VOF simulation (polymer/air).

Related to compressibility, both polymer and air have a fixed density, so I understand I am not calculating any compressibility effect. Am I wrong?

By the way, could you please tell me which other effects should I consider?

I am not an expert in nano-fluidic but I know that some effects can be relevant such as capilarity, wall-rugosity, etc...

for example, http://www.sciencedirect.com/science...90072911000603

FMDenaro,

A nanofluid is a fluid containing nanometer-sized particles, called nanoparticles. I am not simulating a nanofluid, but a normal fluid going through nano-scale cavities :)

However, I still doesn't understand what is wrong in my submodeling approach... Do you see something conceptually wrong?

if you are within a classical fluid mechanics hypothesis and you are working on a flow over a straight wall with laminar condition, why don't you check your flow solution with an analytical one? particle are simply traced in a one-way model?

I am sure someone more expert than me can suggest an idea for your model..

I'll check so.

What do you mean with "particle are simply traced in a one-way model?"?

In the theory of classical fluid dynamics, the fluid velocity at a wall is that of the wall itself, known as the no-slip condition, a direct consequence of the fluid viscosity [ write the definition of shear stress at walls and you will see that by yourself ].
In some specific conditions, you can introduce slip in your analysis so read the following thesis http://arxiv.org/pdf/1101.4421.pdf
and you can google for [slip velocity of a fluid at a wall]
good luck

Maybe I am not clear enough...

Wonder the flat plate model.

The velocity at the wall is always zero, then:

Why does the fluid touch the flat?

I'm thinking in a transient analysis.

again, in the classical fluid mechanics, within the macroscopic continuum hypothesis, one introduces a model for an average velocity v(x). At the wall, such continuos function has the requisite to match with the velocity of the wall.

In the limit of the continuous model (f.m.p. about 70 nm) you cannot use the continuous model but you have to come back to the particle (statistical) description of the medium.

As I already wrote, if you work in nanofluidic (order of f.m.p.) you have to consider if the continuous model is suitable for your study!