[pflaume]

Dear all,


I'm currently trying to conduct compressible flow simulations where I cannot apply the Stokes hypothesis. Hence I need some values for bulk viscosities (or "volume viscosites") for different materials.



So I was wondering if there were any thermophysical data tables / databases including the bulk visc. JANAF-tables for example do not seem have them.

There are some papers and (semi-reliable) websites providing values for the bulk viscosity for a few materials but I'm really looking for a real database.


If there is not such a thing, how do you solve this problem when you simulate compressible flow?


Cheers,
pflaume

[LuckyTran]

Might be my personal opinion, but I don't think there is any reliable database for the second viscosity. From what I know, the community is quite split on what is the second viscosity. Stoke's hypothesis suggests -2/3mu, but there are measurements yielding positive values. It also is supposedly a frequency dependent property, meaning it's not a thermodynamic property purely dependent on the state of the substance.

Originally Posted by pflaume

If there is not such a thing, how do you solve this problem when you simulate compressible flow?

Assume away the problem by using the Stoke's hypothesis.


Btw, if you don't already have the bulk viscosity for your substance, why are you pursuing such a case? How do you know the Stoke's hypothesis doesn't hold if you don't already have the bulk viscosity?

[pflaume]

Thanks for your reply!


That's bad to hear My goal is eventually to simulate some reacting flows, using slightly different equations to model the problem than used in literature. I'm still far away from that, currently trying to simulate a simple compressible flow consisting of one phase (air, but I'm open for any other material ).


I've always read that the Stokes hypothesis holds only true for one-atomic gases and/or for incompressible flow only, hence I thought that it might not be a good idea to ignore the bulk viscosity

But I mean, is this common practice? Do people really apply the Stokes hypothesis even when they are simulating compressible flow?


Thanks and cheers,
pflaume

[sbaffini]

There is no CFD code I am aware of that asks for the bulk viscosity.

[pflaume]

That's a bummer. Looks like everone actually does assume the problem away somehow... Btw, are there other commonly used modells or is everybody only using Stokes hypothesis?

[LuckyTran]

Assuming the Stoke's hypothesis is true does not mean the fluid is incompresssible. You still have plenty of compressibility effects. The bulk viscosity specifically deals with a resistance of the fluid from being compressed and occurs for any substance once the flow time-scales are as short as the molecular time-scales. Note that for substances that have really long molecular time-scales (e.g. molecules with vibration models) we can measure second viscosities, but second viscosities of air are pervasive.

To determine whether it is important or not to consider the bulk viscosity, I challenge you to name a particular fluid flow phenomenon you are trying to simulate or model that specifically requires bulk viscosity to be manifest and cannot be described by any other theory that assumes the Stoke's hypothesis.

Originally Posted by pflaume

But I mean, is this common practice? Do people really apply the Stokes hypothesis even when they are simulating compressible flow?

Is it common for people to use a fluid particle model even when the underlying matter is composed of a large number of much smaller particles each with their own independent momentum and kinetic energy?

Welcome to the world of modeling where we like to make assumptions!


There aren't even good measurements of the second viscosity to even begin formulating a model. The only reason the JANAF tables even exist was due to the Space Race. And those properties are much simpler.


Btw, non-Newtonian behavior (non-linear stress-strain relationships) we do have models for, and these also are rarely applied except where they're absolutely needed. Real gas models are also rarely applied (it's always ideal gas, ideal gas, ideal gas) despite having such models. We have a long way to go before we can apply a second viscosity model to N-S.