[sniak]

Hello all,
I am trying to simulate a simple squeeze film problem. Based on the theory described here:
On the slow dynamics of near-field acoustically
levitated objects under High excitation
frequencies

The problem is described using Reynold's Equation for a thin fluid film, similar to what is done in thin film barrings.
diagram.png

I am trying to start by simulating a very simplified problem as described in the attached diagram.
Right now i have a closed "piston" (100 microns height, 10 mm wide)
I am trying to input a sinusoidal velocity input (using a UDF i attached as well) on the bottom part named "plate". the rest are defined as walls.

I am getting the UDF to work (i think), but i don't get the results i want,
I am trying to use "ideal gas" and "Sutherland" model for viscosity.
using viscosity laminar (as assumed by Reynolds equation).

The UDF:

#include "udf.h"

DEFINE_PROFILE(sinu_velocity_faster, thread, position)
{
face_t f;
real t = CURRENT_TIME;

begin_f_loop(f, thread)
{
F_PROFILE(f, thread, position) = 1.7593*sin(20000.0*2.0*3.14*t);
}
end_f_loop(f, thread)
}

piston_sinus_faster.c

As for the solution. it seems that if i choose a second order or first order transient solution either grows to large or to little.

Anyone here maybe has some experience with similar problems ?
Are there any tips as to what solver to use ? or other physical setup ?
Maybe instead of defining the sides as walls i need to define them as pressure outlets ?

Thank your for reading so far.

diagram.png

piston_sinus_faster.c

[LuckyTran]

It seems you have a working UDF but if you have trouble a simple transient PROFILE may suffice.

For pressure outlets and unsteady compressible simulations in general, you need to pay attention to the acoustic boundary conditions. A pressure outlet is acoustically reflecting. Look into the non-reflecting and wave transparent options, they are only available if you have selected a transient simulation, these options are hidden for steady simulations.

In regards to pressure outlet vs walls. Depending on how much of a cycle you plan to simulate, you may have trouble with the pressure outlet when there is reversed flow. Of course you need a pressure outlet to get the correct induced flow (the viscous part of the squeeze film). If you have all walls, you have mostly compressible effects since the flow is purely kinematic and no boundary layer develops. I would go with whatever works for now until you get comfortable.

As for models, just unsteady laminar + energy with full ideal gas law should be fine.

[sniak]

Thank you for the answer !

It seems you have a working UDF but if you have trouble a simple transient PROFILE may suffice.

Thanks, good to know i got it right.
Regarding the simple transient profile, i am not sure it will suffice, in order to capture the phenomena, i must have a sinusoidal profile, can i do that with a simple transient PROFILE ?

For pressure outlets and unsteady compressible simulations in general, you need to pay attention to the acoustic boundary conditions. A pressure outlet is acoustically reflecting. Look into the non-reflecting and wave transparent options, they are only available if you have selected a transient simulation, these options are hidden for steady simulations.

I did try to put a non-reflecting boundary condition, but it still did'nt help my solution.

My question is,
If i want to simulate the system to be just two plates which are placed in regular room condition, with walls or anything reflecting far enough to be neglected, is putting pressure outlet the right thing ? walls will suffice in this case ?
or is there anything else that can simulate this ?

Thanks again for the support.

[LuckyTran]

Originally Posted by sniak

Thanks, good to know i got it right.
Regarding the simple transient profile, i am not sure it will suffice, in order to capture the phenomena, i must have a sinusoidal profile, can i do that with a simple transient PROFILE ?

With a PROFILE you can specify the value at different times and you just need to specify enough to get the sinusoidal.

Originally Posted by sniak

My question is,
If i want to simulate the system to be just two plates which are placed in regular room condition, with walls or anything reflecting far enough to be neglected, is putting pressure outlet the right thing ? walls will suffice in this case ?
or is there anything else that can simulate this ?

I'm not sure you would get a boundary layer to develop with walls. I doubt it. But have you seen it?

My point was that, both walls and pressure outlets (without non-reflecting boundaries) are acoustically reflecting. The wall is a hard reflection of all frequencies, and the pressure outlet is a type of low-pass filter.

Since acoustic waves is exactly what makes the compressible squeeze film different than a liquid squeeze film, I merely point that out. I don't know how large a magnitude it is, but it seems you've found no issue. But qualitatively... For those acoustic reflections to be negligible they would need to be "non-resonant" at a very different frequency than the driving frequency. It is unlikely the acoustic waves will be damped, because you would need many wavelengths in the lateral direction.