https://linen.dev logo
Join Discord
Powered by
# les-ras
  • k

    kandelabr

    09/20/2021, 1:09 PM
    if i start with upwind, i can finish with linearUpwind
  • s

    slopezcastano

    09/20/2021, 2:15 PM
    Not so important for those terms
  • s

    slopezcastano

    09/20/2021, 2:16 PM
    I'd suggest to sume limitedLinear 1.0
  • s

    slopezcastano

    09/20/2021, 2:16 PM
    in the worst case, just upwind
  • k

    kandelabr

    09/20/2021, 6:15 PM
    thanks, i'll try. but first, it may be a bad mesh, got to try that as well
  • k

    kandelabr

    09/21/2021, 8:10 PM
    so yes, now that i've fixed the mesh the results are the same with
    upwind
    ,
    linearUpwind
    or whatever.
  • k

    kandelabr

    09/21/2021, 8:21 PM
    limitedLinear
    is a bit oscillatory, though
  • f

    filthyasfk

    11/25/2021, 10:32 PM
    Hello everyone, I'm trying to do some turbulence modelling and I found out that the standard or recommended way to calculate initial values for the internal mesh of the turbulent quantities doesn't converge, I'm now guessing the values but seems a bit random. I'm using Bounding Gauss Linear Upwind Grad(U) scheme for every quantity ( U, omega, k ) and sometimes I get good convergence, but the results show strange phenomena for high Reynolds Numbers, for instance, I'm looking for a trend on wake zone dimensions for a surface mounted bluff body to different Reynolds Numbers but for a semi-cylinder, the trend seems to slow or decrease at some point or oscillate for bigger Reynolds Numbers ( which doesn't make any sense to me), any help with this would be very appreciated, thanks in regards.
  • n

    nikobiele

    11/26/2021, 2:38 AM
    I'm not an expert (I'm just learning OpenFOAM), but have you tried to find an example that uses the k-omega model in the Tutorials? For example: (/usr/lib/openfoam/openfoam2106/tutorials/incompressible/pimpleFoam/RAS/rotatingFanInRoom/). That might give you an idea of things to tweak in order to get better results. ๐Ÿ™‚ You can also browse around here: https://www.openfoam.com/documentation/guides/latest/doc/guide-turbulence-ras-k-omega-sst.html
  • f

    filthyasfk

    11/26/2021, 3:16 AM
    Hello NikoBiele and thank you for ur response, I have tried a lot of different combinations of meshes, discretization schemes, and expressions to estimate k and omega, but the problem is that none of those expressions give good results, so the thing that I did is take note of the bounding values displayed on my screen and use a average as my initial guess ( This makes everything more stable and I reach convergence in a reasonable amount of time ), the big problem is that the principal dimension of the wake zone seems to reduce for a semicilinder as the Reynolds numbers goes up ( this is totally different for my other geometries for the bluff body in which the length of the wake seems to go up as I turn the Reynolds number Up), sorry if I'm not very clear and thank you again for the help provided.
  • f

    filthyasfk

    11/26/2021, 3:17 AM
    also I was using this expressions:
  • f

    filthyasfk

    11/26/2021, 3:17 AM
    $\frac{U{\infty}}{L} < \omega{farfield} < 10 \frac{U_{\infty}}{L}$
  • t

    TeXit

    11/26/2021, 3:17 AM
    ๆฌ  ใ‚’ ้ฃŸในใ‚‹
  • f

    filthyasfk

    11/26/2021, 3:17 AM
    $10^{-5} \frac{ {U{\infty}}^{2}}{Re{L}} < \kappa{farfield} < 0.1 \frac{{U{\infty}}^{2}}{Re_{L}}$
  • t

    TeXit

    11/26/2021, 3:17 AM
    ๆฌ  ใ‚’ ้ฃŸในใ‚‹
  • n

    nikobiele

    11/26/2021, 3:27 AM
    How is you semi cylinder positioned in the flow? I haven't studied semi cylinders myself, but I know that for a full cylinder (not semi) the wake is supposed to get smaller at certain high Reynolds numbers, due to the boundary layer becoming turbulent (as can be seen on the drawing). However, I don't know if the same is true for a semi cylinder. ๐Ÿ™‚
  • f

    filthyasfk

    11/26/2021, 3:53 AM
    For all my geometries it's something like this:
  • f

    filthyasfk

    11/26/2021, 3:53 AM
  • f

    filthyasfk

    11/26/2021, 3:56 AM
    the thing is that, for a cube or a triangular shape the horizontal length of the wake behind the object tends to grow with Reynolds Number, but for the half-cylinder ( or semi I don't know how should it be called hahaha) it tends to go down.
  • n

    nikobiele

    11/26/2021, 3:59 AM
    Which Reynolds numbers do you use?
  • f

    filthyasfk

    11/26/2021, 4:11 AM
    I'm doing from 5e3 to 1e6, about 7 Reynolds numbers between those.
  • n

    nikobiele

    11/26/2021, 4:13 AM
    Well did you see the numbers on the drawing above?? You seem to be right in the range where the wake decrease?
  • f

    filthyasfk

    11/26/2021, 4:19 AM
    Exactly , the thing is that for the other geometries the trend is to go up and not down, so I was thinking that maybe I have some error in my simulations for the half cylinder geometry ( taking into account that the simulations all converge to 10^-6 and my Y+ value is below 50 all the time. ), also I didn't mentioned but I'm doing steady state Kw SST case, thank you for ur help! ๐Ÿ™‚
  • f

    filthyasfk

    11/26/2021, 4:20 AM
    sorry if I'm mistaken I'm also very new to OpenFoam and turbulence modelling in general.
  • n

    nikobiele

    11/26/2021, 4:41 AM
    I don't think it is an error :)) . To me it makes sense that the flow in the wake behaves very differently depending on the degree of separation at the geometry in question. And it is quite natural that squares and triangles result in a larger separation of the flow due to sharp edges, as compared to the smooth surface of a half cylinder.
  • f

    filthyasfk

    11/26/2021, 7:12 PM
    Thank you for ur help NikoBiele ๐Ÿ™‚
  • n

    nikobiele

    11/26/2021, 7:41 PM
    my pleasure, glad I could help. ๐Ÿ™‚
  • d

    Dave M

    11/26/2021, 7:59 PM
    Does anyone know which version(s) of openfoam has the kOmega SST DDES or IDDES model?
  • n

    nikobiele

    11/26/2021, 9:13 PM
    maybe try google first next time ๐Ÿ˜‰ https://www.openfoam.com/documentation/guides/latest/doc/guide-turbulence-des-k-omega-sst-ddes.html
  • d

    Dave M

    11/26/2021, 9:47 PM
    Thanks. I did but targeted a particular version I intend to work with and it didn't come up
1...212223...52Latest