Question about inviscid model in CFD software

Adorkable · November 25, 2022, 22:20

Hi, hello everyone，I have a strange question.
When can the inviscid model (not potential model) be applied in CFD software? In other words, how to judge the inviscid model and turbulent model has little influence on the numerical results?
Thank you everyone!

aerosayan · November 26, 2022, 04:37

When the simulation you're trying to do, doesn't need viscosity to give accurate results.

For example: Hypersonic flows can ignore viscosity.

Don't worry too much. Learn how to solve with viscosity. The problem you're trying to solve will most likely state if you need viscosity or not.

agustinvo · November 26, 2022, 04:47

If you are doing RANS, the turbulent stress tensor is there because of the fluctuating velocities in the advection term. Should we neglect as well the turbulence modelling?

FMDenaro · November 26, 2022, 04:57

Quote:

Originally Posted by Adorkable

Hi, hello everyone，I have a strange question.
When can the inviscid model (not potential model) be applied in CFD software? In other words, how to judge the inviscid model and turbulent model has little influence on the numerical results?
Thank you everyone!

Yes, strange question…
Inviscid fluid model (Euler equations) is used prevalently for potential flows and shock-waves dominated problems.
Few applications can be found for LES.

kinematic_presser · November 27, 2022, 06:27

If you non-dimensionalize the Navier-Stokes equations (https://en.wikipedia.org/wiki/Non-di...okes_equations) you will see that the coefficient in front of the viscous can be written as 1/Re . (The article shows the incompressible NS, but this is true for compressible NS as well).

Therefore in order to ignore the effects of viscosity the Reynolds number should be quite large, probably a million at the very least.

FMDenaro · November 27, 2022, 06:46

Just to address some further comments.

Assuming ideal flows (no viscosity, no conducibility) you disregard any type of dissipation of kinetic energy. That means you have a model for the turbulence where the energy cascate is never terminated by dissipation.

In no way that can be simulated on a finite grid since the theoretical Kolmogorov lenght is exactly zero and you can never approach a DNS.

Therefore, in any simulation you have to terminate numerically the energy cascate.

Furthermore, since you have no dissipation, you cannot suppose a statistical energy equilibrium between production and dissipation of kinetic energy to introduce a turbulence model for the closure of the equations.
In this sense, RANS is not suitable as formulation.

The Euler equations are mainly solved for flows with shock by disregarding the turbulence effects.

https://link.springer.com/chapter/10.1007/11545989_1

November 25, 2022, 22:20	Question about inviscid model in CFD software	#1
Adorkable New Member Zheng Xiaoluan Join Date: Dec 2018 Posts: 2 Rep Power: 0	Hi, hello everyone，I have a strange question. When can the inviscid model (not potential model) be applied in CFD software? In other words, how to judge the inviscid model and turbulent model has little influence on the numerical results? Thank you everyone!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Zone model: CFD or analythic model?	bhigahAshish	Main CFD Forum	0	July 24, 2020 03:34
CFD model validation with experimental data	Alsumri	ANSYS	6	November 14, 2019 11:14
Validation of CFD model	flow_CH	Main CFD Forum	4	November 4, 2019 10:30
General question about SST turbulence model	Richard Renaud	CFD Freelancers	1	December 19, 2015 04:13
Overflow Error in Multiphase Modelling with Two Continuous Fluids	ashtonJ	CFX	6	August 11, 2014 14:32

November 26, 2022, 04:37		#2
aerosayan Senior Member Sayan Bhattacharjee Join Date: Mar 2020 Posts: 495 Rep Power: 8	When the simulation you're trying to do, doesn't need viscosity to give accurate results. For example: Hypersonic flows can ignore viscosity. Don't worry too much. Learn how to solve with viscosity. The problem you're trying to solve will most likely state if you need viscosity or not.

November 26, 2022, 04:47		#3
agustinvo Senior Member Agustín Villa Join Date: Apr 2013 Location: Alcorcón Posts: 313 Rep Power: 15	If you are doing RANS, the turbulent stress tensor is there because of the fluctuating velocities in the advection term. Should we neglect as well the turbulence modelling?

November 27, 2022, 06:27		#5
kinematic_presser New Member K Join Date: Oct 2022 Location: Madison, WI Posts: 19 Rep Power: 3	If you non-dimensionalize the Navier-Stokes equations (https://en.wikipedia.org/wiki/Non-di...okes_equations) you will see that the coefficient in front of the viscous can be written as 1/Re . (The article shows the incompressible NS, but this is true for compressible NS as well). Therefore in order to ignore the effects of viscosity the Reynolds number should be quite large, probably a million at the very least.

November 27, 2022, 06:46		#6
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,769 Rep Power: 71	Just to address some further comments. Assuming ideal flows (no viscosity, no conducibility) you disregard any type of dissipation of kinetic energy. That means you have a model for the turbulence where the energy cascate is never terminated by dissipation. In no way that can be simulated on a finite grid since the theoretical Kolmogorov lenght is exactly zero and you can never approach a DNS. Therefore, in any simulation you have to terminate numerically the energy cascate. Furthermore, since you have no dissipation, you cannot suppose a statistical energy equilibrium between production and dissipation of kinetic energy to introduce a turbulence model for the closure of the equations. In this sense, RANS is not suitable as formulation. The Euler equations are mainly solved for flows with shock by disregarding the turbulence effects. https://link.springer.com/chapter/10.1007/11545989_1