[barathwaj]

Hey I am doing CFD analysis of thermostat valve in ANSYS FLUENT . my pipe OD is 53mm . my Reynolds number is around 200000 . so my flow has a high turbulence. My fluent is water. Water temperature is 90 deg celcius. I ve incorporated the properties in materials database
Setup:
No energy equation. I use velocity inlet c=1.16m/s.pressure outlet. Realisable ke model. When I solve this I m getting an error in epsilon . can anyone help me out?? Thank you

[fluid23]

Its probably a mesh issue near your wall, but you posted it in the wrong forum. A moderator is going to move this to the Fluent user forum. Try looking there for help.

[barathwaj]

Thanks . I did my mesh in ICEM. I just used fine and highly smooth mesh. Can you tell me what exactly I must do??

[fluid23]

Why no energy equation? Not convinced Newton was right?

Without more information I would just be guessing, and my best guess is a mesh issue. Create a post in the Fluent forum and someone will help.

http://www.cfd-online.com/Forums/fluent/

[barathwaj]

Lol . If I used energy equation I m getting divergence in temperature. Anyways the tempature across the entire domain is a constant . thats why I didn't on it

[fluid23]

It could be that your turbulence model isn't appropriate. Try k-w with sst.

Just out of curiosity, what kind of wall y+ values are you seeing?

[fluid23]

Also, are you using a coupled or segregeated solver? Sometimes if you have convergence issues with a coupled approach, switching to segregeted will get you moving again. However, there will be some loss of accuracy.

[barathwaj]

Oh thanks I am using a segregated solver only .what do mean by saying wall y+ values

[fluid23]

If you don't know what your wall y+ values are then you really have no idea if your mesh is ok by the wall or not. y+ is (put simply) a measure of how well suited your near wall cells are for boundary layer calculations. They are a product of your turbulence model's near wall approximations. I bet your issue lies here.

Look at your turbulence model, see what wall treatment you use. There will be near wall, far wall or all wall (which is a mix of the previous 2). Each of these will require different y+ values for accuracy. You will want to talk to someone in the Fluent forum who can give you giudance on what y+ values to target for your model. Unfortunately, even though the math is basically the same from software to software I have found that you rarely get a concensus on what y+ should be. I use star ccm+ usually with k-e or k-w models with all wall treatment. When I do that I try to keep y+ betwen 1 and 5 or 30 and 60 based on the recommendation of their support engineers.

[fluid23]

You also might try a coupled solver just to see if that is any more reliable.

[fluid23]

Oh, and use the energy equation. It's sort of a fundamental aspect of CFD. You have a viscous turbulent flow that has kinetic energy and dissipation. That energy has to be transformed from mechanical to thermal somehow, even if your temperature gradients are very low.

[barathwaj]

Actually my Temperature gradient is almost zero Thts why i turned it off. Can you tell me in detail what type of mesh I need to perform in order to eliminate this problem

[barathwaj]

actually this is my model and i ve placed it in a pipe . my l/d ratio is 150 . I ve come to know that splitting the volume and meshing it will decrease the y+ values near my walll. But i dont know how to do it in ICEM is there any related tutorials available.

[fluid23]

There are all types of mesh that will work. It is not an issue of type, but of quality. I am not familar with ICEM so you will need to talk to someone in the appropriate forum who knows your software.

I use star-ccm+ which has generic polyhedrals and poly prisms that are very robust, but that option isn't available everywhere. What you really want to do is make sure that whatever mesh you use, that you conform to proper mesh practices. There are all kinds of tips, tricks and rules of thumb that you can use. I don't have time to get into all that. You should do some research on the wiki for this site. They have lots and lots of good information on this and other topics.

[fluid23]

and y+ is very important, you can't ignore it.

[bshkoj]

Hi. can every one help?
I want to simulate weir using ansys cfx code and i used RNG turbulence model. but the results are not good and i can not get the ventilation region in water profile.