Can turbulence help measure liquid viscosity more accurately?

[chaksome]

Hi, all. Recently I plan to participate in an physics competition which focuses on viscosity measurements of liquid~I’ve investigted some kinds of methods such as capillary viscometers, Rotational viscometers, Vibrating Viscometers, co-flowing laminar viscometers and so on.

I hope to find a more innovative approach to measure the viscosity, thus I think about the effect of viscosity in turbulence. What I expect to do now is setting up a coaxial-flowing device, where the inner liquid is test fluid and the outer one is reference liquid. We can observe the turbulence if we inject the inner fluid into the outer fluid.

Could you please give me some advice on it? >w<Which parameter should I measure if I want to get the viscosity of the test liquid? Or Which lectures or papers may I refer to. Of course, you’re welcomed to give me some other suggestions if you think this method is unpractical or not good enough~ Thanks a lot!

 

[chaksome]

http://www.tsfp-conference.org/proceedings/2019/276.pdf
Here is one of the reference I found. I am going to set up a device similar to the device in it.

 

[Arjan82]

Do you want to measure viscosity itself? Or the effect of viscosity on turbulence (or, closely related, on mixing, as in your referred paper)? Those are two different things. The co-axial setup you are referring to is not suitable to measure viscosity itself.

 

[chaksome]

Do you want to measure viscosity itself? Or the effect of viscosity on turbulence (or, closely related, on mixing, as in your referred paper)? Those are two different things. The co-axial setup you are referring to is not suitable to measure viscosity itself.

That‘s the point, I am trying to measure the viscosity itself. So I should upgrade the setup if I choose that way.
Maybe it is hard to calculate it theoretically and maybe measure viscosity in that way is meaninglessSo I should think twice before I make a decision
Could you please detail the reason why it is not suitable or if there is a more suitable method. Thanks!

 

[Arjan82]

Uhm, are you suggesting to compute the viscosity based on the turbulent behavior? In that respect "hard to calculate it theoretically" is quite the understatement... There is a Nobel price and a million dollars waiting for you if you manage to do that! Consider Feynman's words, who called it the oldest unsolved problem in classical physics, on the topic:

"
Finally, there is a physical problem that is common to many fields, that is very old, and that has not been solved. [..] Nobody in physics has really been able to analyze it mathematically satisfactorily in spite of its importance to the sister sciences. It is the analysis of circulating or turbulent fluids.
"

And Horace Lamb, a well known 19th century mathematical physicist, as an old man quipped that when he gets to heaven, he would ask God two questions: "Why relativity? and why turbulence?" He then said he is optimistic about getting a good answer to the first question.

Anyway, I'm curious as to how you were thinking to approach this

 

[boneh3ad]

I don't understand why, if the idea is to use fluid flow to measure viscosity, you don't just compare the laminar velocity profile in a pipe or a Blasius boundary layer to theory. Why introduce turbulence to a problem like this?

 

[Chestermiller]

I strongly agree with @Arjan82. Using a turbulent flow to determine viscosity is a loser of an idea, and is destined to failure. Turbulence is chaotic motion, with velocities varying at high frequencies temporally. So you don't even know what velocities and velocity gradients you have at anyone time. Even the average velocities won't give you good measures of viscosity because turbulent momentum transport swamps what you are looking for. All you end up with then is the "eddy viscosity" which is much greater than the actual fluid viscosity.

If you want to do an interesting viscosity measurement using laminar flow with time variations, you might consider a forced oscillatory experiment. But, please, for your sake, lose the idea of using a turbulent flow.

 

[chaksome]

Uhm, are you suggesting to compute the viscosity based on the turbulent behavior? In that respect "hard to calculate it theoretically" is quite the understatement... There is a Nobel price and a million dollars waiting for you if you manage to do that! Consider Feynman's words, who called it the oldest unsolved problem in classical physics, on the topic:

"
Finally, there is a physical problem that is common to many fields, that is very old, and that has not been solved. [..] Nobody in physics has really been able to analyze it mathematically satisfactorily in spite of its importance to the sister sciences. It is the analysis of circulating or turbulent fluids.
"

And Horace Lamb, a well known 19th century mathematical physicist, as an old man quipped that when he gets to heaven, he would ask God two questions: "Why relativity? and why turbulence?" He then said he is optimistic about getting a good answer to the first question.

Anyway, I'm curious as to how you were thinking to approach this

Thanks for your introduction（and sorry for late reply) I am not going to do such thing that may bring me Nobel Prize. And maybe what I am trying to do is not giving a complete explanation to the relation between viscosity and turbulence. I am trying to get some empirical formula(such as the relation between the viscosity and the break-up length or the primary break-up frequency as the reference did in(26)(27)(28)).
https://www.sciencedirect.com/science/article/pii/S0301932215301956
But now, as what you say, I find that I’m just making trouble for me if I try that way. I can just use the typical methods of laminar viscosity measurement and make some optimization～（such as microfluid co-flowing methods and forced oscillation methods as @Chestermiller mentions before.
Thanks again!

 

[Arjan82]

Note that with that empirical relation you are not measuring viscosity directly. You rely on someone else to establish an empirical relationship that you then try to exploit. I would read very carefully in the competition rules if that is even allowed. And if it is, note that you can only 'measure' the viscosity within the validity of the empirical relation, which usually does not capture the entire range of possible (and reasonable) viscosities. Extrapolation is almost never correct when turbulence is considered, since its behavior is highly nonlinear.

Also, I suspect that using turbulence to measure viscosity in the way you mention is going to give you a significant error bar.