How to verify there is turbulence in a flow with a k-epsilon turbulence model

Maveryck · March 7, 2019, 08:14

Good day every one,

I have implemented a k-epsilon turbulence model for an incompressible flow solver, and now I want to verify that the results I get correspond to a turbulence regime. By solving the k equation I have the kinetic energy, and the rate of dissipation from the epsilon equation. With those values how could I stimate if the flow is turbulent?

Thanks a lot in advance!

sbaffini · March 7, 2019, 08:56

Quote:

Originally Posted by Maveryck

Good day every one,

I have implemented a k-epsilon turbulence model for an incompressible flow solver, and now I want to verify that the results I get correspond to a turbulence regime. By solving the k equation I have the kinetic energy, and the rate of dissipation from the epsilon equation. With those values how could I stimate if the flow is turbulent?

Thanks a lot in advance!

The resulting turbulent viscosity ratio is always a good measure. Yet, in principle, assuming proper bc are given, a RANS turbulent model is not supposed to give non-turbulent results and they are only valid in a fully turbulent regime.

However, you should compare with other k-epsilon implementation if you want to validate the code.

FMDenaro · March 7, 2019, 10:21

As wrote by Paolo, there is no sense in asking if the turbulence exists after you get a RANS solution. It is exactly the opposite, if you have a fully developed turbulence problem in energy equilibrium then the RANS formulation could provide reasonable statistical results.

LuckyTran · March 7, 2019, 15:09

There is formally no way to tell. The only hint is the eddy/turbulent viscosity which should be 0 for laminar flow and non-zero for turbulent flow. But chances are, you will always have non-zero turbulent viscosity because you are using a turbulence model which assumes the flow is already fully turbulent.

In other words, you should assume that the solution corresponds to the fully turbulent result even if this solution is not accurate. As a corollary, if the real flow is supposed to be laminar, then your simulation corresponds to what the flow would look like if it were turbulent.

VSass · March 8, 2019, 03:22

The most reliable indicator is the insight of the physics of each specific case you are working on. Also, validation of your results on simple cases from the literature.

March 7, 2019, 08:14	How to verify there is turbulence in a flow with a k-epsilon turbulence model	#1
Maveryck New Member Maveryck Andres Join Date: Mar 2019 Posts: 6 Rep Power: 7	Good day every one, I have implemented a k-epsilon turbulence model for an incompressible flow solver, and now I want to verify that the results I get correspond to a turbulence regime. By solving the k equation I have the kinetic energy, and the rate of dissipation from the epsilon equation. With those values how could I stimate if the flow is turbulent? Thanks a lot in advance!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Low Re Turbulence model VS Low Re flow and Turbulence Modelling background	pchoopanya	Main CFD Forum	4	May 10, 2018 11:05
Choice Turbulence Model Flow around Airfoil	kayla1994	Main CFD Forum	3	May 10, 2018 04:28
Overflow Error in Multiphase Modelling with Two Continuous Fluids	ashtonJ	CFX	6	August 11, 2014 14:32
Turbulence model for Laminar flow	Terry	Main CFD Forum	8	August 14, 2010 05:45
validating turbulence model on flow around a car	Pedro	CFX	1	February 20, 2008 16:32

March 7, 2019, 10:21		#3
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,768 Rep Power: 71	As wrote by Paolo, there is no sense in asking if the turbulence exists after you get a RANS solution. It is exactly the opposite, if you have a fully developed turbulence problem in energy equilibrium then the RANS formulation could provide reasonable statistical results.

March 7, 2019, 15:09		#4
LuckyTran Senior Member Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,674 Rep Power: 66	There is formally no way to tell. The only hint is the eddy/turbulent viscosity which should be 0 for laminar flow and non-zero for turbulent flow. But chances are, you will always have non-zero turbulent viscosity because you are using a turbulence model which assumes the flow is already fully turbulent. In other words, you should assume that the solution corresponds to the fully turbulent result even if this solution is not accurate. As a corollary, if the real flow is supposed to be laminar, then your simulation corresponds to what the flow would look like if it were turbulent.

March 8, 2019, 03:22		#5
VSass Member VS Join Date: Nov 2012 Posts: 86 Rep Power: 13	The most reliable indicator is the insight of the physics of each specific case you are working on. Also, validation of your results on simple cases from the literature.