Hi I am a new cfx-user and want to simulate the pressure loss of a valve. Maybe someone can help me improving my model...

I just created the mesh in ICEM and imported it to cfx-pre. For the calculation i chose the following boundaries:

Inlet: Mass flow=12kg/s

Opening (as outet): static pressure (entrain)=159 bar

Wall: smooth wall

I didn´t set any other parameters! My problem is now that the results of the calculation (pressure loss) are too low. So maybe someone can give me some hints to improve the result...

I would appreciate any help or tip

Greets Chris

are you sure that's 159 bar?

Yes, its water at 159 bar and 290°C

Why have you kept wall- smooth and I hope your wall has a no slip wall condition on it ? .

I have a no slip condition on my wall. Which parameters should i use for a rough wall when i want to simulate teh flow through a valve. (Housing and pipe consisting of steel)

Hi,

Assuming you have the physics correct the mistake made by almost every new user is to use too coarse a mesh. You need to do a mesh refinement study to determine how fine a mesh is required to capture the flow.

Do this by increasing the mesh density (say reduce all element edge lengths by 25%) and compare. If the pressure loss changes significantly then your mesh is not fine enough and you need to keep refining the mesh until the pressure loss levels out.

Glenn Horrocks

Yes , I agreee with Glenn on mesh density near the wall. You must improve your mesh if its too coarse .

To check your mesh is good ,kindly plot your Yplus at the walls. It should be ideally less than 1 mm, but betweeb 1 to 5 mmm its still ok. If your mesh gives you this results of yplus ,its good cfd results

ALso check what is your required boundary conditions at the outlet. Whats your inlet, velocity or mass flow or something else you have inserted ?

anyways best of luck for your simulation.

Thanks for yor help, i really appreciate it!

But isnt yplus a dimensionless parameter and should be ideally between 20-200? Or what do you mean by: it should be 1mm?

So if i plot yplus-results and have a maximum value of 2500 it means i have to refine the mesh there, is that right??

Hi,

y+ is dimensionless. Pratik incorrectly gave it units of mm.

The y+ you need depends on the turbulence model you are using. If standard wall functions are OK then y+ between 20 and 200 should be OK. If you are integrating to the wall you need y+ about 1. If y+ is in the intermediate range of 1 to 20 the only wall function approach which works is the automatic wall function approach.

Glenn Horrocks

ya, sorry I meant the mesh parameterto be like 1 mm near the wall .