ViscoBoundFactor

Hello!

Can anyone explain the visco bound factor and show in which equation it appears, in DualSPHYsics? I cannot figure it out from the documentation.

My understanding is that it acts as some kind of roughness factor for fixed, moving and floating objects - I just want to understand how it acts physically / appears mathematically.

Kind regards

Comments

  • Hi @Asalih3d,

    Have you looked here ?


    Also, how to use it and calibrate it is describe here:

    Barreiro A, Domínguez JM, Crespo AJC, González-Jorge H, Roca D, Gómez-Gesteira M. 2014. Integration of UAV photogrammetry and SPH modelling of fluids to study runoff on real terrains. PLoS ONE, 9(11): e111031. doi:10.1371/journal.pone.0111031.

    Best regards


  • Hi, thanks for the answer @TPouzol

    I have read through the paper and I understand that in their approach they had a Manning number to go after, which seems like a good way to do it for different materials. To be completely sure you linked this equation;

    My understanding is that "alpha" is the ViscoBoundFactor - is this correct?

    And does this means that this factor is only available for artificial viscocity? If I change it while using LES SPS then it should have no effect what so ever?

    Kind regards

  • Hi,

    Yeah the coeffcient is only available for artificial viscosity.

    Alpha represent the energy dissipation when 2 part are interacting. But there is 2 different alphas: 1) alphaFF for fluid fluid ineraction, 2) alphaBF for boudary fuid ineraction.

    In .xml, you provide alphaFF

    <parameter key="Visco" value="0.025" comment="Viscosity value" />

    But not alphaBF directly, rather you give a coefficient such as AlphaBF = ViscoBoundFactor * alphaFF

    <parameter key="ViscoBoundFactor" value="3.5" comment="Multiply viscosity value with boundary (default=1)" />

    hope I'm clear

    Kind regards

  • Thanks for clarifying @TPouzol. If this is the case I fear that there unfortunately is a bug in the code, I am using the following parameters

    And I get the result:

    Which is clearly wrong since the top wall (yellow) is moving with 1 m/s in x-direction. If I change the visco bound factor to 1 I see;

    Which is much more accurate. Nothing is changed between the two cases except this ViscoBoundFactor. Just tagging @Alex as well. It seems to me this term should have any effect on LES approach, but maybe there is something missing. Both pictures were taken at the same time step, time step 27.

    Kind regards

  • Hi all,

    I've been playing around with these two formulations for simulating flows in long, open channels (100 m, dp = 1 m). I've noticed that the L-SPS config appears to behave inviscidly regarding the boundary interaction. I've tried changing ViscoBound with no results. Additionally, I observed that boundary interactions do occur when the flow direction is aligned with the gravitational force resulting in some flow inhibition at the bottom, could someone confirm this?

    Also, is the L-SPS method recommended for large scale wave simulations (600 m long tank) at low resolutions (1 m)? I've read that the L-SPS model is not recommended for near-hydrostatic simulations

  • @Asalih3d do you mean that the ViscoBoundFactor should or should not have an impact in the Lam+SPS visco treatment?

    I'm currently modeling flow across a circular cylinder in 2D with Lam+SPS (based on example #5 in mDBC) and I'm performing a sensitivity analysis on the VBfactor. Although it seems relatively small, I am noticing a difference in the vorticity values in the wake of the flow for constant Reynolds numbers with differing VBfactor values. I am interested in modifying surface roughness to observe its impact on flow and VBFactor is the only input I've found that could represent surface/boundary roughness.

    @TPouzol , is it in fact a bug that changing VBFactor has an impact in Lam+SPS? I'm using v5.0 which seems to be after this thread since it's in v4.4.

    The following figures show the same simulation (Lam+SPS, alpha = 10^-6 for water, and Reynolds number = 50800) at time step 42, where VBFactor equals 1, 5, and 10 respectively.

    Thank you for any pointers you can provide!

    Sandra

  • Hi Sandra

    I am assuming the ViscoBoundFactor is implemented in Lam+SPS the same way it is in Artificial viscosity.

    Regards

  • @Sandra according to the math shown on the wiki it should not have an influence when using SPS-LES.

    The good thing is though, that it does have an influence! Which is really important for being able to add some tiny bit of friction in the simulation. In this case the friction acts as a damping of the numerical system. Though that is almost the only thing you can use it for.

    Perhaps the DualSPHysics Team wants to clarify this @Alex would know better who to tag

    Kind regards

Sign In or Register to comment.