Hi,

I want to investigate on fluid flow in straight square microchannel with dimension 5micron wide and high and length 1mm. When im choose laminar equation in CFX-pre solver the results shows laminar flow(parabolic flow). Therefore, here i want to use Turbulent Equation model but im not sure which is the best model that can cooperate with laminar equation. Preliminary, i choose BSL method with turbulence scale as " zero gradient" but the results does not shows any significant differences with laminar equation. Am i wrong? Would anyone give best way to solve this problems.

Really appreciate you assistant much. Thanks in advance.

rgrds, Simon

Hi,

What is the Re number of the flow? For a feature of that size it is bound to be very low and therefore a laminar flow. What makes you think a turbulence model is required?

Glenn Horrocks

Dear Glenn,

Thanks for your reply and interest. Initially, the Re number is 0.05, later i increaed up to 2700 to qualify whether any turbulent flow occurs but not. I simulate straight channel as a preliminary tests for (lamianr&turbulent) equations.Actually, i have to simulate for various bend channels and i suspecting there might be turbulent flow. So that, i cant use laminar equation to justify this case and i need correct setting for turbulent model in cfx-solver in case the flow is turbulent.

Hi,

If your flow is transitional (significant laminar and turbulent regions) then you have a very difficult model on your hands. If your highest Re is 2700 then it is a very low Re turbulent flow so the turbulent effects are likely to be small - I would still think laminar is the correct approach.

You could consider the turbulent transition model but it has been tuned for external airfoil flows so it would need to be carefully validated for internal flows like this.

Alternately you could use the k-w model as that can degenerate into a laminar flow when k=0. Note that this is not a physically valid approach as there is nothing in the k-w model to model turbulence transition effects, it is just that when k=0 the k-w model gives a laminar flow.

I recommend the first approach as the most likely to be suitable. Whatever you choose it will need to be carefully validated.

Glenn Horrocks