[samiam1000]

Dear Foamers,

I am gonna try to figur out how to set a simple simulation over a diffuser.

Here, you can download my case: I have run it, but the solution (after 100000 steps, pretty close to convergence) seems to be laminar.

Do you think that I am using the right turbulence model? Or, to you, there is something wrong with my simulation?

Do the results make sens, to you?

Thanks a lot,
Samuele.

PS: the simulation has is ready to be seen with paraFoam. Don't run either ./Allrun or ./Allclean!

[Bernhard]

Based on what do you conclude that this result is laminar?

Your bottom wall has a non-zero fixedValue bc. Are you sure you can also apply wallfunctions there, as it is not a wall? Maybe you can try slip there (I think that is what you actually want to do), with zero gradient for the other variables.

[samiam1000]

Originally Posted by Bernhard

Based on what do you conclude that this result is laminar?

Just looking at the solution!

Originally Posted by Bernhard

Your bottom wall has a non-zero fixedValue bc. Are you sure you can also apply wallfunctions there, as it is not a wall? Maybe you can try slip there (I think that is what you actually want to do), with zero gradient for the other variables.

I am trying this: thanks a lot!

[bscphil]

Originally Posted by samiam1000

Dear Foamers,

Do you think that I am using the right turbulence model? Or, to you, there is something wrong with my simulation?

Do the results make sens, to you?

Thanks a lot,
Samuele.

PS: the simulation has is ready to be seen with paraFoam. Don't run either ./Allrun or ./Allclean!

Check your y+ values for right usage of wall functions in the k-omega-SST y+ should be approx 30.

For modelling recirculating zones due to adverse pressure gradient in the diffuser, i think: you have to use a so called Low-Re turbulence model and a finer mesh at the walls with y+ values approx 1.

Your "bottom" is no wall, i think?, just use patch-type with slip-bc or symmetric-bc ?? !!

[icaroaccordi]

Hi Samuele,
We can not judge if there is transition using a k-omega SST or another high-Re model (wall function). High-Re are for completely turbulent flow. If your flow are completely developed I'm quite sure that you have a turbulent flow due the Reynolds of the flow. If the inlet of your geometry is a entrance and you have turbulence intensity great than 1% and the turbulent flow start before the diffuser you can try a low-Re model. I could recommend Launder-Sharma. It is not good enough but it is a beginning. Any way after the beginning of the diffuser you are going to have a turbulent flow due the transition trigged by laminar detachment. To deal with this kind of transition or natural transition the only model available in OF is kklOmega. To be worse the model in OF have some errors due some mistypes in the based paper.
Conclusion: pray to flow be turbulent and follow Phillip’s recommendation.

Best regards

[Madeliz]

Hi, I'm working on a project that involves wind generation with a diffuser augmented wind turbine. I would like to know how can I improve the y+ value on the diffuser. It presents a value of 533 aprox on the y+.

[bscphil]

To improve your y+ values use a strong clustered (finer) mesh in the wall normal direction near the walls and decrease the height of your first wall-cell, just use a cell-to-cell expansion ration in wall normal direction of 1.1-1.2!

[Madeliz]

I'm working with an inflation, first layer thickness of 1e-4. I'm new on this where can I cange the cell-to-cell expansion ration in wall normal direction?

[bscphil]

Hi Madelaine,
what do you mean with "inflation"? What meshing-tool do you use?

Please give us more detailed information about your problems, then you'd have to get more help...

[Madeliz]

Hi! I'm working with CFX with k-epsilon moldel on a wind turbine design, I have great results on the rotor but when I'm working with the diffuser It presents high values of y+ . I'm using a 1e-5 inflation value and a Reynolds about 2.4e-5

[bscphil]

Medelaine, this isn't the forum for questions about CFX, just use the Ansys forum.

Nevertheless, i'm not able to give you some tips about inflation and Reynolds (what do you mean with Reynolds, the Reynolds number, no !?) Some pictures of your mesh might help, is it unstructered or structured?

[samiam1000]

Hi all: actually, the bottom wall is a wall, which moves in the x direction.

Also, in the next days I'll work on the turbulence models.

Thanks a lot,
Samuele.

[Costas]

Hi guys, im trying to have an inflation layer around my 3D diffuser but Workbench is not able to give me that mesh. Has anyone used a method that was successful for something similar?