Hi I'm working on oil distribution analysis. My flow model consists on small passages (diam. 4 mm) and I have to run a steady-state analysis with a commercial Code. Re number is about 12000. Flow enters from a circular tube (Inlet b.cond.) and exit from two ducts (2 Pressure b. cond) 1) I cannot use k-eps high-Re with wall function because I should use a 0.8-1 mm cell at wall in order to have correct y+ (30-100). 2) I tried to use both two layer models and k-eps low-Re/Hybrid Wall Function with a proper fine mesh at wall (15 layers with y+ less than 1). I do not know if the second way is correct or not. I can only say that by using the last two turbulence models 3) I found some strange values for heat transfer coefficients at wall, about 24000-30000 W/m2 K while proper values should be 4000-5000 W/m2 K

Could anyone tell me why I found so high heat trasfer coefficients and if low-Re models are correct or not for such calculations?

Thanks in advance.

Massimo

Have you calculate conjugate heat tranfer ?

No, I have not. Nusselt number is a function of Re number and Pr number so it depends only on flow field and fluid properties (Cp,k, etc.). My calculation is isothermal. Heat transfer coefficient are required in order to be passed, later, to a stress analysis.

It seems you are use Star-CD. The have two 1. Heat transfer coefficient calculated by empirical formule (Nu=f(Re,Pr)) 2. Hcoefficient calculated as qcell=alfacell/(Tbulk-Tcell)

First otimized fo some poblem(i do not know what)

Second work fine.

In conclusion you hawe wrong calculated heat transfer coefficient becouse you do not solfe heattransfer. Do not pay atention to this result about Hcoef, this do not means anything.

Yes, I'm a Star-CD user. I'll try to evaluate Hcoe after having fixed wall temperature.

Thanks Massimo

Hcoef give probable value only if you solve conjugate hettransfer. You mast bulid real solid wall.