Compare two simulations with same friction Reynolds and different bulk Reynolds

Moreza7 · March 3, 2020, 11:15

Hi,
I have simulated a duct flow using LES with Bulk Reynolds number = 2800 and friction Reynolds number=160.
There's a DNS data with bulk Reynolds number = 2500 and friction Reynolds number = 160.

Can I compare these two results or they should have the same Bulk Reynolds number? Can I scale the results?

sbaffini · March 3, 2020, 12:04

Well, you can compare them, but only one of them is probably correct, if any. Guess which one I'm referring to?

However, it is typical for LES simulations of wall bounded turbulent flows to produce higher mean velocities for given pressure drops, when low order methods (<3) are used. In contrast, higher order methods tend to produce lower mean velocities.

The culprit here is that, ideally LES is a model and DNS is correct. Your model is failing here. Is it important? It depends from what you actually want to achieve, what is the scope of your analysis. If the result is too off you probably have to refine your grid at the wall (probably spanwise).

If you scale the results you simply obtain the correct Re number but now with the wrong friction Re number. There is no escape from this.

FMDenaro · March 3, 2020, 12:11

As addressed by Paolo, fixing the pressure driving force for a prescribed Re_tau number can produce quite different results in terms of the bulk Re number.

For this reason I suggest to get a solution without any SGS model to assess the effect of this latter on the averaged velocity.

LuckyTran · March 3, 2020, 12:21

I just want to add that (assuming constant viscosity), for fully developed flow in a duct, the bulk Reynolds number and friction Reynolds number are proportional to one another via the friction factor.

One way to interpret this is, if you target one (either the bulk or friction Reynolds number) and you also get the right friction factor, then you will match the other one. That is, if you match one of the Reynolds numbers with the right pressure drop, you also match the other Reynolds number.

So if you try to target the same friction Reynolds number, the reason you don't also have the same bulk Reynolds number is because your solution is slightly different for whatever reason. It's not an error per se, just that your model produces a slightly different result.

March 3, 2020, 11:15	Compare two simulations with same friction Reynolds and different bulk Reynolds	#1
Moreza7 Senior Member Join Date: Jan 2018 Posts: 121 Rep Power: 8	Hi, I have simulated a duct flow using LES with Bulk Reynolds number = 2800 and friction Reynolds number=160. There's a DNS data with bulk Reynolds number = 2500 and friction Reynolds number = 160. Can I compare these two results or they should have the same Bulk Reynolds number? Can I scale the results?

March 3, 2020, 12:04		#2
sbaffini Senior Member Paolo Lampitella Join Date: Mar 2009 Location: Italy Posts: 2,195 Blog Entries: 29 Rep Power: 39	Well, you can compare them, but only one of them is probably correct, if any. Guess which one I'm referring to? However, it is typical for LES simulations of wall bounded turbulent flows to produce higher mean velocities for given pressure drops, when low order methods (<3) are used. In contrast, higher order methods tend to produce lower mean velocities. The culprit here is that, ideally LES is a model and DNS is correct. Your model is failing here. Is it important? It depends from what you actually want to achieve, what is the scope of your analysis. If the result is too off you probably have to refine your grid at the wall (probably spanwise). If you scale the results you simply obtain the correct Re number but now with the wrong friction Re number. There is no escape from this. FMDenaro, pcosta and Moreza7 like this.

March 3, 2020, 12:11		#3
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,896 Rep Power: 73	As addressed by Paolo, fixing the pressure driving force for a prescribed Re_tau number can produce quite different results in terms of the bulk Re number. For this reason I suggest to get a solution without any SGS model to assess the effect of this latter on the averaged velocity. Moreza7 likes this.

March 3, 2020, 12:21		#4
LuckyTran Senior Member Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,761 Rep Power: 66	I just want to add that (assuming constant viscosity), for fully developed flow in a duct, the bulk Reynolds number and friction Reynolds number are proportional to one another via the friction factor. One way to interpret this is, if you target one (either the bulk or friction Reynolds number) and you also get the right friction factor, then you will match the other one. That is, if you match one of the Reynolds numbers with the right pressure drop, you also match the other Reynolds number. So if you try to target the same friction Reynolds number, the reason you don't also have the same bulk Reynolds number is because your solution is slightly different for whatever reason. It's not an error per se, just that your model produces a slightly different result. sbaffini and Moreza7 like this.