[akaobi]

Hello CFD online fraternity, I'm having difficulty with my model here and hope you will offer me your kind help.

I have a filter model of dimensions 52μmx52μmx8μm. The fibers are 1μm in diameter and structure's porosity is 0.86

I ran a CFD-DEM simulation to represent the theory of dynamic particle loading in a real air filter.
In theory, with dynamic filter loading, pressure drop progressively increases and same happens with particle capture efficiency.
My results for efficiency correlate well with theory, however, pressure drop is not; whereby its graph rises initially just as the flow is beginning and then stays constant for the rest of the process.
I set pressure drop from reports; setting it higher at the entrance and lower at the exit.

Help will be highly appreciated!
Thank you all
Kindly check the attached pictures

OBED

[acalado]

Do you have transient boundary conditions?

Have you tried setting velocity inlet and pressure outlet, and then looking at pressure drop?

[akaobi]

Originally Posted by acalado

Do you have transient boundary conditions?

Have you tried setting velocity inlet and pressure outlet, and then looking at pressure drop?

Hello Acalado, thanks for the response.
Sure I set the inlet to velocity and outlet to pressure
In addition, I chose the IMPLICIT UNSTEADY solver for laminar flow (my superficial velocity is 0.5),
I'm using time-steps of 0.001 and 10 iterations per time step

Thanks for your time and help

[fluid23]

A couple things...

1. Do you have the solid phase (particles) modeled properly? The correct mass, etc...? What you describe seems like the fluid forces acting on the particles (and vice versa) are not resulting in momentum loss (i.e. drag) which causes the pressure change.

2. If so, it could also be how you set up your report. It sounds like you put these on the boundaries themselves.... this is rarely a good idea. I would suggest extruding your inlet and outlet then defining planes where you had the inlet and outlet previously (or inset some planes away from your boundary). You may not be able to use the pressure drop report since this (I think) requires an region interface. However, you can easily re-create this yourself with two surface or mass flow averaged pressure reports and a user defined field function/report.

[akaobi]

Hello Matt, thanks a lot for those expert ideas.
You have really given me a number of new dimensions to consider.

From your explanation, I'm worried it may be the particles because I kept the default settings for density(1100) and Poisson coefficient (0.45) only slightly increasing the Young's Modulus (20MPa).
I'm gonna set realistic particle parameters I see if that can offer me a better simulation.


ALSO,
There are two options for particle size specification (particle size or mass), here, I opted for particle radius in all my simulations. Do you think changing to particle mass instead would change the result in this case?
I too will try to get familiar with the effect of using either of these two from the USERGUIDE.


About the pressure drop in the REPORTS not being so reliable,
I actually set up my model in the way you described by having an extension on both the inlet side and outlet. (Kindly check out the attached picture)
For pressure drop report (which apparently could not give me the expected result), I set the inlet as the high pressure side and the outlet as the low pressure side.

When the above could not work, I then set up two DERIVED PARTS (PLANES) One just near the upstream side of the filter and the other, right at the downstream end of the filter and on each of these I set up a SURFACE AVERAGE report with pressure as the function after which I took a difference of these two
Still, it seemed to give an almost similar trend to pressure drop.
Thanks a lot
Best regards