[hitzhwan]

Originally Posted by vinerm

If you are modeling two-phase flow using multiphase model then you need to ensure that you have at least 4-6 layers within the thin film thickness that you expect for the liquid phase. If the simulation is single phase, then you do not need any inflation at all.

Are you sure the 4-6 layers is enough，but I think it is so few that cannot capture the flow boundary layer for the film，which has quite a strong velocity gradient.

[vinerm]

You mentioned that the flow is laminar. Then the gradient cannot be high. High gradient implies turbulent flow. So, if the flow is really laminar and single-phase, you do not require any inflation layers at all, not even 4-6.

[hitzhwan]

Originally Posted by vinerm

You mentioned that the flow is laminar. Then the gradient cannot be high. High gradient implies turbulent flow. So, if the flow is really laminar and single-phase, you do not require any inflation layers at all, not even 4-6.

Do you have the theory which can verify your opinion？

[vinerm]

Just pick up any fluid dynamics book that explains boundary layer and you will understand. For a laminar flow through a duct, the thickness of boundary layer is equal to radius while the thickness of boundary layer for a turbulent flow through the same duct is much smaller. Gradient increases as thickness reduces. For a laminar flow, velocity varies from 0 to maximum over r while for turbulent flow same variation occurs over a fraction of r. Which one in your opinion requires finer mesh? For the same reason, is not even available as a field variable in laminar flows, since it is not important.

[hitzhwan]

Originally Posted by vinerm

Just pick up any fluid dynamics book that explains boundary layer and you will understand. For a laminar flow through a duct, the thickness of boundary layer is equal to radius while the thickness of boundary layer for a turbulent flow through the same duct is much smaller. Gradient increases as thickness reduces. For a laminar flow, velocity varies from 0 to maximum over r while for turbulent flow same variation occurs over a fraction of r. Which one in your opinion requires finer mesh? For the same reason, is not even available as a field variable in laminar flows, since it is not important.

The thickness of boundary layer is equal to radius ？ What is the radius? I have found many books about fluid dynamics， but there are no results， do you have any books recommend for me？ Thank you very much.

[vinerm]

Radius implies radius of the duct through which fluid is flowing. The concept of boundary layer does not exist in undergrad texts. Analytical solution is usually based on similarity analysis and you can find it many books, such as, Boundary Layer Theory by Schlichting or in Convection Heat Transfer by Adrian Bejan. It is also called Blasius solution. You can find such work online as well. And it is not very time-taking and can even be derived within a post here. If you are mathematically oriented, then you can go through An Introduction to Fluid Dynamics by G.K.Batchelor, but only if you comfortable with mathematical analysis.

[hitzhwan]

Originally Posted by vinerm

Radius implies radius of the duct through which fluid is flowing. The concept of boundary layer does not exist in undergrad texts. Analytical solution is usually based on similarity analysis and you can find it many books, such as, Boundary Layer Theory by Schlichting or in Convection Heat Transfer by Adrian Bejan. It is also called Blasius solution. You can find such work online as well. And it is not very time-taking and can even be derived within a post here. If you are mathematically oriented, then you can go through An Introduction to Fluid Dynamics by G.K.Batchelor, but only if you comfortable with mathematical analysis.

Thank you for your answer. I have another question:I have set the use of manufacturable value in the DOE， but I find it does not generate the design values after undating DOE?

[vinerm]

I am afraid I do not have any idea about WB or DoEs.

[hitzhwan]

Originally Posted by vinerm

I am afraid I do not have any idea about WB or DoEs.

It does not matter. I have another problem, how should we set the residuals in the fluent if we have a transient flow? Is 1e-3 enough ? Or should we change it it under 1e-6 or more? Or we only observe the monitor variables?