Cell size (H) calculation in 2D

RCBE

Dear all,

I noticed that there are different results for the cell size depending on the working plane (XZ, XY,..).
Settings:
dp = 0.0001
coefh = 1.5
according to the formula (2D) H = coefh * sqrt(2) *dp =0.000212. (XZ-plane → ok)

now I set up the project for the XY-plane instead of XZ. The result for H is 0.000260. Why?? For 3D it would be ok because of H = coefh*sqrt(3)*dp = 0.0002598.
And this caused other problems. In the end the simulation is not working in the XY-plane, only in the XZ-plane.

files:
https://www.dropbox.com/sh/lyf3zvwtbam3pt9/AAB_tB7PLvAvoHKf5kRkhxbna?dl=0

This case is based on the one used in the 'Pipe flow with floating particles - error after dp decrease' - discussion
http://dual.sphysics.org/vanilla/discussion/791/pipe-flow-with-floating-particles-error-after-dp-decrease#Item_1

Thanks!

Alex

The 2D simulations are only designed to work in the plane XZ since all forces in Y are forced to be zero. So there is no sense to create a XY case with our code.

You can always create XZ cases and then change gravity if you need that.

Regards

RCBE

Thank you Alex. I forgot about that.

NWRichmond

I've noticed that in the 2D case examples (such as CaseWavemaker2D and CaseDambreakVal2D, the gravity is set in the Z direction (z="-9.81") yet the cases are in the XZ plane. What is the reason for this?

Alex

We use the XZ plane... so that gravity is in -Z direction!
Our 2D examples are just a cut of a 3D case!

NWRichmond

Thanks, Alex! It makes sense now that the 2D examples are a cut of a 3D case along the XZ plane. I've been trying to mimic an XY plane, but I started having problems when changing the gravity.

To get around this: since I was already using an external forces file to accelerate particles in the X direction, I simply kept the gravity as -9.81 m/s^2 in the z-direction but added a constant +9.81 m/s^2 to the "LinearAccZ" column of my external forces file.

It seems to work as expected (judging by the forces output for the sides of a flume that runs along the x-axis; the z-forces are roughly equal in magnitude but opposite in direction), but if there's something wrong with doing it this way, I'd be glad to know.