[Hah01]

I'm attempting to compare the effects of surface dimples i.e a golf ball to a smooth sphere for a university assignment. I thought I would start off validating my domain size and mesh by computing the drag coefficient at a low Reynolds number of 100. However, I can't seem to get a value any where close to the graphs of Cd vs Re that I have found on the web (https://www.grc.nasa.gov/www/k-12/ai...ragsphere.html)

The most common Force Coefficient I am getting is about 0.3, from looking at graphs this should be about 1.0.

Problem definition: Diameter = 0.04267, the sphere is situated 5D from the inlet and 25D from the outlet and at the midpoint of the height and width of the domain, both of which have a length of 10D.

From browsing these forums for similar problems to my own, I see mention of y+ treatment (which I do not fully understand) and for my turbulent sims I have set minimum wall distance to 0.005 which I got from an online calculator.

Many people on this forum have recommended turbulent models. Is this due to laminar models being unable to model the stable vortices behind the sphere at this Reynolds number (100)?

Please could someone recommend physics models and mesh sizes that would fit this problem? Apologies for being vague about my current set up but I have tried loads of different combinations with varying convergence and can't seem to get close.

[underrelaxed]

I realize this doesn't immediately address your question, but can we see a picture of your mesh?

[Hah01]

Thanks for replying. I just refined the mesh again, I am going to run another sim and will post the results along with that of the mesh once it is finished.

[Hah01]

Originally Posted by underrelaxed

I realize this doesn't immediately address your question, but can we see a picture of your mesh?

I cant seem to get my image hosting approved by the mods let me see if this works



Base = 0.02
10 prism layers at 33% of base, layer stretching 1.2
Custom surface mesh on sphere at 10% minimum 25% target
and The volume control block which is set at 80% of base

[Hah01]

Looks like I was uploading wrong before. Here is a close up. I'm using a tetrahedral mesher with surface remesher

[Hah01]

... and one more which shows the whole domain. Boundary conditions are slip walls for top and bottom, and symmetry planes for side walls. Flow is in positive x direction from inlet to outlet.


Here are the results from the two steady tests I did, blue is Laminar, orange is Turbulent (RANS Spalart-Allmaras standard):

[acalado]

Have you checked what are the parameters used to non-dimensionalize the Force coefficient? (area, reference velocity... )

[Hah01]

Originally Posted by acalado

Have you checked what are the parameters used to non-dimensionalize the Force coefficient? (area, reference velocity... )

I've doubled checked these and it looks like I've been using the wrong frontal area the whole time! This has improved my results loads I am now getting them close to the 0.7 mark which isn't correct but it is a lot closer, thank you for pointing out my schoolboy error, I was getting way too hooked up on the physics models.

[ping]

sad to see someone using a tet mesh - serious star users abandoned these a decade ago


Sent from my iPad using CFD Online Forum mobile app

[Qkarl]

Hi, can I know the outcome of this simulation? what turbulence model did you used and the mesh setting? Thanks in advanced