How to Prism Layer :)

Sebastian. · November 9, 2024, 07:16

Hey,

i need to do some simulations as a part of an project at university. In my case i work with a rather simple problem: the flow through an orifice flowmeter.

Now i want to create the mesh and as part of it the prism layers. I just read some resources but there are still some questions open so I hope maybe some of you can help me.

Laminar flow:
What i've read, says that i should resolve the boundary layer with ~10 prism cells. But how do i know the thickness of my boundary layer? I read about the Blasius solution and saw some estimates based on it for external flows. Is there some estimate for pipe flows, that i haven't found?

Turbulent flow:
Here I have the some problem with the thickness of the boundary layer. But there is also the question of y+ for my first cell. Are there best paractice defaults for pipe flow or maybe orifice flow?
From literature i get that if I use wall functions for the whole boundary layer I can use y+ ~ 500. If I just want to model the viscous sub-layer I need ~5? Are those correct or applicable for my problem?

Thanks in advance

xisluke · November 12, 2024, 06:14

Hi Sebastian, in my opinion as first guess you can evaluate the boundary layer thickness by using Blasius and and the one-seventh-power-law for the turbulent BL; those formulations are related to a flat plate, but since I do not think your analysis will be computationally costly, in case the correlations show different results you can adjust in a second moment the grid.

The laminar BL is linear, therefore you will not need too much cells, 10 are ok. For the turbulent BL, it really depends also on the turbulence model adopted; if you want to make evaluations in a low-Re fashion, you can choose a k-omega SST model aiming for y+ < 1 (the best thing you can do should be clustering more than one cell within the viscous sub-layer, but the number of cells will increase rapidly) and at least 15-20 (the best could be also 20-25) layers. On the other hand, in high-Re you can adopt wall-functions and generally speaking the first inflation layer cell should be placed in the range between 30 < y+ < 300; in this case you can use a k-epsilon model (you could still adopt SST, since I think it is y+ insensitive in a lot of platforms).

You have also to consider that, from what I understood, you have not a fully developed velocity profile, but an orifice flowmeter where the BL thickness will not be constant along the orifice length.

November 9, 2024, 07:16	How to Prism Layer :)	#1
Sebastian. New Member Sebastian Thaler Join Date: Jun 2024 Location: Germany Posts: 4 Rep Power: 2	Hey, i need to do some simulations as a part of an project at university. In my case i work with a rather simple problem: the flow through an orifice flowmeter. Now i want to create the mesh and as part of it the prism layers. I just read some resources but there are still some questions open so I hope maybe some of you can help me. Laminar flow: What i've read, says that i should resolve the boundary layer with ~10 prism cells. But how do i know the thickness of my boundary layer? I read about the Blasius solution and saw some estimates based on it for external flows. Is there some estimate for pipe flows, that i haven't found? Turbulent flow: Here I have the some problem with the thickness of the boundary layer. But there is also the question of y+ for my first cell. Are there best paractice defaults for pipe flow or maybe orifice flow? From literature i get that if I use wall functions for the whole boundary layer I can use y+ ~ 500. If I just want to model the viscous sub-layer I need ~5? Are those correct or applicable for my problem? Thanks in advance

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[snappyHexMesh] Prism layer collapsing in refinement transition	rubenparedes	OpenFOAM Meshing & Mesh Conversion	8	May 13, 2021 11:55
[snappyHexMesh] layer not added	Rasmusiwersen	OpenFOAM Meshing & Mesh Conversion	1	January 2, 2020 10:43
Questions about Polyhedral and Prism layer meshing	niablis	STAR-CCM+	3	September 23, 2016 01:32
Trimmer & prism layer	peresah	STAR-CCM+	6	October 6, 2009 15:39
errors	Fahad	Main CFD Forum	0	March 23, 2004 14:20

November 12, 2024, 06:14		#2
xisluke Member Luca Join Date: Mar 2021 Location: Italy Posts: 57 Rep Power: 6	Hi Sebastian, in my opinion as first guess you can evaluate the boundary layer thickness by using Blasius and and the one-seventh-power-law for the turbulent BL; those formulations are related to a flat plate, but since I do not think your analysis will be computationally costly, in case the correlations show different results you can adjust in a second moment the grid. The laminar BL is linear, therefore you will not need too much cells, 10 are ok. For the turbulent BL, it really depends also on the turbulence model adopted; if you want to make evaluations in a low-Re fashion, you can choose a k-omega SST model aiming for y+ < 1 (the best thing you can do should be clustering more than one cell within the viscous sub-layer, but the number of cells will increase rapidly) and at least 15-20 (the best could be also 20-25) layers. On the other hand, in high-Re you can adopt wall-functions and generally speaking the first inflation layer cell should be placed in the range between 30 < y+ < 300; in this case you can use a k-epsilon model (you could still adopt SST, since I think it is y+ insensitive in a lot of platforms). You have also to consider that, from what I understood, you have not a fully developed velocity profile, but an orifice flowmeter where the BL thickness will not be constant along the orifice length.