Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

K-Epsilon Turbulent Model - Turbulence Damping for paper pulp

Please login with a confirmed email address before reporting spam

Hi folks.

I am modelling paper pulp flow in pipes.

I have a lot of experimental data that I have to simulate and validate.

I am using single phase turbulent k-epsilon model to do so. However, I get pressure drops higher than
expected.
I suspect that the turbulence damping that occurs for these flows is not being implemented.

Does anyone have suggestions or experience with this matter?
I could use a few pointers....

Best Regards,

Rui

9 Replies Last Post 17 nov. 2011, 03:03 UTC−5

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 nov. 2011, 11:18 UTC−5
Hi,

I am not sure what you mean with 'turbulence damping for these flows'; The starting point for these analysis is always the question: what is your Reynolds number?

At any rate a coarse mesh will cause large-than-expected pressure drops, however if the pipes have several change of directions it may be a good idea to use the k-omega model.

Cheers
Hi, I am not sure what you mean with 'turbulence damping for these flows'; The starting point for these analysis is always the question: what is your Reynolds number? At any rate a coarse mesh will cause large-than-expected pressure drops, however if the pipes have several change of directions it may be a good idea to use the k-omega model. Cheers

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 nov. 2011, 11:31 UTC−5
I guess I should have provided more details :)

It's a straight 4 meters pipe, with a paper pulp solution flowing at 6 m/s with constant temperature.

I believe the behavior of this solution is Non-Newtonian.

And typically, the pressure drop is lower than for water, for example, for the same pipe length.
This is due to Turbulence Damping, because it occurs a damp in the turbulence production in the wall.

The mesh is quite fine with boundary layer.

Hope this helps.

Cheers.



Hi,

I am not sure what you mean with 'turbulence damping for these flows'; The starting point for these analysis is always the question: what is your Reynolds number?

At any rate a coarse mesh will cause large-than-expected pressure drops, however if the pipes have several change of directions it may be a good idea to use the k-omega model.

Cheers


I guess I should have provided more details :) It's a straight 4 meters pipe, with a paper pulp solution flowing at 6 m/s with constant temperature. I believe the behavior of this solution is Non-Newtonian. And typically, the pressure drop is lower than for water, for example, for the same pipe length. This is due to Turbulence Damping, because it occurs a damp in the turbulence production in the wall. The mesh is quite fine with boundary layer. Hope this helps. Cheers. [QUOTE] Hi, I am not sure what you mean with 'turbulence damping for these flows'; The starting point for these analysis is always the question: what is your Reynolds number? At any rate a coarse mesh will cause large-than-expected pressure drops, however if the pipes have several change of directions it may be a good idea to use the k-omega model. Cheers [/QUOTE]

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 nov. 2011, 11:54 UTC−5
Hi,

All right, non-Newtonian flow with turbulence? That sounds tough enough, can you provide the Re number for the averaged density and the max and min value?

Cheers
Hi, All right, non-Newtonian flow with turbulence? That sounds tough enough, can you provide the Re number for the averaged density and the max and min value? Cheers

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 nov. 2011, 19:16 UTC−5

if the pipes have several change of directions it may be a good idea to use the k-omega model


Hi Amir,

If you have 2 minutes, could you please indicate a reference where it's advised to use the k-w model instead of k-e for flow in complex geometries?

Thanks!
Francois

[QUOTE] if the pipes have several change of directions it may be a good idea to use the k-omega model [/QUOTE] Hi Amir, If you have 2 minutes, could you please indicate a reference where it's advised to use the k-w model instead of k-e for flow in complex geometries? Thanks! Francois

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16 nov. 2011, 03:56 UTC−5
Hi,

There is quite a large volume of literature about the failure of the k-epsilon and how Menter developed the k-omega for bounded flows. You can start with two papers by Menter:

Menter, "Zonal Two Equation k-ω Turbulence Models for Aerodynamic Flows", AIAA Paper 93-2906, 1993

Menter "Two-Equation Eddy-Viscosity Turbulence Models for Engineering Applications", AIAA Journal, vol. 32, no 8. pp. 1598-1605, 1994


Cheers
Hi, There is quite a large volume of literature about the failure of the k-epsilon and how Menter developed the k-omega for bounded flows. You can start with two papers by Menter: Menter, "Zonal Two Equation k-ω Turbulence Models for Aerodynamic Flows", AIAA Paper 93-2906, 1993 Menter "Two-Equation Eddy-Viscosity Turbulence Models for Engineering Applications", AIAA Journal, vol. 32, no 8. pp. 1598-1605, 1994 Cheers

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16 nov. 2011, 09:44 UTC−5
Hello Amir,

I apologize for the delay but I have been away from my desk.

The Re number for my flow is 45 730 for a pipe diameter with 0.0762 m and 2 meters long. The velocity with 6 m/s.
The density is 1004.77 and the viscosity is a function of the velocity:

ETA=0.0132*V^(-0.1524)

Most of the times I can't get it to converge, and when it does, the pressure drop is 3x larger...

Hope you can give some more insight...

Thank you.

Cheers.




Hi,

All right, non-Newtonian flow with turbulence? That sounds tough enough, can you provide the Re number for the averaged density and the max and min value?

Cheers


Hello Amir, I apologize for the delay but I have been away from my desk. The Re number for my flow is 45 730 for a pipe diameter with 0.0762 m and 2 meters long. The velocity with 6 m/s. The density is 1004.77 and the viscosity is a function of the velocity: ETA=0.0132*V^(-0.1524) Most of the times I can't get it to converge, and when it does, the pressure drop is 3x larger... Hope you can give some more insight... Thank you. Cheers. [QUOTE] Hi, All right, non-Newtonian flow with turbulence? That sounds tough enough, can you provide the Re number for the averaged density and the max and min value? Cheers [/QUOTE]

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16 nov. 2011, 10:00 UTC−5
Hi,

What are the units of your density? Kg/m^3 I think. How about the viscosity? Is it kinematic or dynamic? Did you try to average it to have an idea of the average Re?

It seems the viscosity goes to infinity as V approaches zero, do you have an idea of the actual limits on your viscosity?

Do you have an idea of the length scales of the flow in your pipe?

Cheers
Hi, What are the units of your density? Kg/m^3 I think. How about the viscosity? Is it kinematic or dynamic? Did you try to average it to have an idea of the average Re? It seems the viscosity goes to infinity as V approaches zero, do you have an idea of the actual limits on your viscosity? Do you have an idea of the length scales of the flow in your pipe? Cheers

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16 nov. 2011, 10:16 UTC−5
Yes, those the units for the density.

The viscosity is dynamic.

In principle, the paper pulp behaves like a shear thinning fluid and our experiments tell us that the
range for the dynamic viscosity is [0.0405 - 0.0101]Pa.s, and for higher velocities the dynamic viscosity decreases.

The length scales are : It=0.01*Re^(-1/8) and Lt=0.0005.

Hope this helps.

Rui


Hi,

What are the units of your density? Kg/m^3 I think. How about the viscosity? Is it kinematic or dynamic? Did you try to average it to have an idea of the average Re?

It seems the viscosity goes to infinity as V approaches zero, do you have an idea of the actual limits on your viscosity?

Do you have an idea of the length scales of the flow in your pipe?

Cheers


Yes, those the units for the density. The viscosity is dynamic. In principle, the paper pulp behaves like a shear thinning fluid and our experiments tell us that the range for the dynamic viscosity is [0.0405 - 0.0101]Pa.s, and for higher velocities the dynamic viscosity decreases. The length scales are : It=0.01*Re^(-1/8) and Lt=0.0005. Hope this helps. Rui [QUOTE] Hi, What are the units of your density? Kg/m^3 I think. How about the viscosity? Is it kinematic or dynamic? Did you try to average it to have an idea of the average Re? It seems the viscosity goes to infinity as V approaches zero, do you have an idea of the actual limits on your viscosity? Do you have an idea of the length scales of the flow in your pipe? Cheers [/QUOTE]

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 17 nov. 2011, 03:03 UTC−5
Hi,

Crunching the numbers you gave I get:

Re=rho*V*D/eta=rho*D*V^1.1524/0.0132 where D is a characteristic length and can be thought to be the diameter of the pipe

for a critical value of Re=2000 then:

Vcritical=(2000*0.0132/(rho*D))^(1/1.1524)=0.399m/s

It is quite crude but it tells us that for a characteristic length equal to the diameter of the pipe with a velocity-dependent viscosity your flow becomes laminar as the velocity reaches 0.399 m/s.

So I would say that you need to mix laminar and turbulent in one go, honestly I do not know how that can be done in Comsol.

Cheers
Hi, Crunching the numbers you gave I get: Re=rho*V*D/eta=rho*D*V^1.1524/0.0132 where D is a characteristic length and can be thought to be the diameter of the pipe for a critical value of Re=2000 then: Vcritical=(2000*0.0132/(rho*D))^(1/1.1524)=0.399m/s It is quite crude but it tells us that for a characteristic length equal to the diameter of the pipe with a velocity-dependent viscosity your flow becomes laminar as the velocity reaches 0.399 m/s. So I would say that you need to mix laminar and turbulent in one go, honestly I do not know how that can be done in Comsol. Cheers

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.