Choosing turbulence length scale
 

Hello,
I am confused in calculating the turbulence length scale. My problem is "compressible developing flow in a diffusing duct". As boundary conditions demand turbulence length scale or hydraulic diameter alongwith turbulence intensity. Ansys help 18.1 (Fluent Userguide 6.3.2.1.4) says for wall bounded flows it is 0.4*BL thickness. Are the wall bounded flows termed as developing flows? Or for developing flows need some other way to compute turbulence parameters?

Take the fraction of the boundary layer thickness.

It's not really about developing or fully developed. Wall bounded flows are flows where boundary layers develop on walls. There are also free shear flows, where boundary layers develop between two flows of different momenta. Kinematic blocking by the wall limits some characteristics that the flow can have and that's why it gets its own categories.

Thank you.

One more thing please, Can we use same turbulence parameters for developing flows as are used for fully developed flows? As per my research the calculations exist only for fully developed flows.

If you find that your problem is sensitive to inlet turbulence BC's, then you need to do a parametric study to find out the sensitivity. Frankly, it's nearly impossible to know the correct turbulence BC's at the inlet for any particular problem. Most people are just aiming to get their inlet turbulence BC's to within the correct order of magnitude.

Let's say you use a k-epsilon model. What you need, is actually the boundary condition for k and epsilon. But most people don't think of turbulence this way, instead they like turbulence intensity and length scale. So Fluent has implemented some formulas (for a very specific flow scenario) that allows the user to input intensity and length scale, and Fluent will back calculate the value of k and epsilon. The real question is whether you know k and epsilon in a developing BL for your specific problem? Most people don't have a k-meter or an epsilon-meter to know the answer. All you can do is take your best guess.

Furthermore, the k and epsilon BC's needed are non-uniform over an entire surface (i.e. a field variable). It's not one number. Most people are applying one constant value over the entire boundary. This is (from a theoretical standpoint) completely wrong and idiotic. But practically, no one really has time to pull our their k-meter and epsilon-meter and probe the entire inlet.

thank you.
Yes, i need to do a study to check the effect of turbulence parameters. this is what i learned from this forum, ansys user manual and research articles: I have hydraulic dia at inlet and will use intensity values (<= 1% for developing flows) unless I get reasonable / correct results.