Learning Forum

[Kilian Nuss]

Hi Ansys Comunity, 

 

My questions regard a internal laminar flow simulation in a pipe with a constant temperature at the wall.

The velocity of the fluid is around 0.01m/s and has a high dynamic viscosity of 100mPas. Reynolds number does not exceed over 100.

The main goal is to determine the heat transfer coefficient and the temperature profil.

 

After reading articles about resolving turbulent flow my issue is to decided which solver I should use and what y+ value is recommended. 

So could you please give me any advice for this instance?

 

Furthermore, how can i check if I resolved the heat transfer/temperature profil appropriately analougous to turbulent flows and their boundary layers.

 

Thank you in advance.

[SRP]

Hi,

For selection of model, please refer to the theory guide: 4.18. Near-Wall Treatments for Wall-Bounded Turbulent Flows (ansys.com)

I suggest to create an additional variable for the eddy viscosity ratio. Then by plotting this variable on a suitable plane, and superimposing mesh in the near-wall region, we can visualize the boundary layer resolution.

Hope you find this useful

Thank you

[Rob]

Mesh resolution studies can help here too: keep refining the mesh until it stops changing. Also, don't forget to check you resolve any flow separation and bulk flow phenomena. Too many fall foul of focussing on the near wall region to the detriment of the rest of the flow domain so miss what's going on. 

[Kilian Nuss]

so you would advice me to use a turbulent flow model instead of the laminar? even though the Re<100?

 

thank you Rob for the advice I am going to keep it in mind when I am doing the mesh refinement.

[Rob]

Generally no. It can be worth running one case laminar & again turbulent to see if there's a significant difference. Reason being if you're borderline transitional due to flow conditions (not Re - that's a guide) it can show up in the turbulent case. 

Note, depending on conditions transition flow in a pipe can occur at Re 800-1000   The undergrad experiment when I was at Uni gave very odd results every 20mins: that coincided with a bus parking outside....