Modeling simple K-epsilon turbulent flow in annulus

122 views
Skip to first unread message

sohaib malik

unread,
Jun 26, 2021, 12:43:41 PM6/26/21
to Nek5000
Dear Neks,

I am trying to model turbulent flow in an annulus with inner diameter of 0.01905m and outer diameter of 0.03810m and Hydraulic Diameter for annulus=Dh=Do-Di=0.01905m. Height of annulus heated region is 0.6m. The geometry and boundary conditions are reflected in the attached picture.

 Temperature at inlet = t-inlet=83, velocity at inlet= v-inlet=0.248 m/s , Heat flux at inner side =q=1.1535e5 w/m2, 

Fluid Properties at (0.8 MPa, T=83oC),

dynamic viscosity (nu)=3.4e-4, Density=970 kg/m3, Conductivity (k)=0.6695 w/m-k, Cp=4198 J/kg-k,

Reynolds number based on the inlet average velocity and hydraulic diameter comes out to be 1.4*10^4. When I plot the temperature and velocity profile, they show smooth straight lines, that I feel obvious that since I didn't use any turbulent model and just increased Reynolds number it doesn't model turbulent flow.

Now, I am going to model turbulent flow. I am thinking about using simplest K-epsilon turbulent model. Please help me which Nek example should I follow that would be easy to modify based on my example. Thanking you in advance. 

Sohaib

temperature600.png
annulusgeometry.png
velocityh600.png

sohaib malik

unread,
Jun 28, 2021, 3:53:00 PM6/28/21
to Nek5000
Hi,
I am sorry to post it again. I would highly appreciate any guidance on what direction to go for my case.
Best Regards,
Sohaib

lbca...@vcu.edu

unread,
Jun 28, 2021, 6:37:15 PM6/28/21
to Nek5000
Sohaib,

There isn't currently a k-e turbulence model (there is some work with k-w and a few other variants (SST)). Are you wanting to do RANS? Or were you trying to do LES? You likely aren't seeing turbulent flow since you are either unresolved (go to a higher lx1) and/or haven't "kicked" the system into turbulence. One trick for the later is run it at a higher Reynolds number and then drop back down after it's tripped.

sohaib malik

unread,
Jun 28, 2021, 7:11:19 PM6/28/21
to lbca...@vcu.edu, Nek5000
Thank you so much for your helpful reply. 
I don't want to do LES, would be ok to do RANS. I got your first point about lx1 and the way how to do it. But, can you please help and elaborate on the second point how to do this?  "One trick for the later is run it at a higher Reynolds number and then drop back down after it's tripped".
Thank you so much.
sohaib

--
You received this message because you are subscribed to a topic in the Google Groups "Nek5000" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/nek5000/DSm6-EZASEs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to nek5000+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nek5000/7e7e71f2-a628-4544-9302-69a29ae34789n%40googlegroups.com.

lbca...@vcu.edu

unread,
Jun 30, 2021, 12:35:37 PM6/30/21
to Nek5000
If you just want to do RANS, the "kick" isn't needed. You just need to look for the k-w turbulence model example. Not sure what the procedure for implementing that is.
Reply all
Reply to author
Forward
0 new messages