surface tension - liquid gas code

Comments

• JOJO

  January 2020

  

  function: KerInteractionForcesFluidMP in GPU version
• olivliv

  January 2020

  thanks @JOJO

  found it here

  

  but im still in doubt about what value should i choose to simulate an air-water flow.

  for the two-phase flow liquid-gas tutorials, the only one where i saw the use of this surfcoef was the last one, 06_Surface Tension, but it actually deals with a liquid-liquid (water-heavy water) case and it uses a surfcoef=0.5.

  from Hu and Adams 2006 paper ("A multiphase SPH method for macroscopic and mesoscopic flows"), I saw a numerical example of a droplet of water surrounded by air, where they used surfcoef=1. im using this value, but i´m not sure if there are other things to take into account for this choice.
• athmokos

  January 2020 edited January 2020

  Hello,

  the surfcoef coefficient is the physical surface tension coefficient (usually denoted by γ or σ) with units of N/m. Within the code is denoted as surfc1 and is called by

         surfc1=phasedata[iphase].PhaseSurf;

  It is only used for calculation of the surface tension force.

  The physical value for water at surface temperature is 0.072N/m. I have not had issues using this value. If there are large instabilities in the interface you can try a larger value like Hu & Adams, but bear in mind that their multi-phase and viscosity model were also different.
• olivliv

  January 2020

  @athmokos thanks for the explanation.

  in the xml file we edit < surfcoef value="0"/> instead of surfc1. in order to work with 0.072N/m as surfc1 what value should i use for surfcoef? or should i add <surfc> instead of <surfcoef> in the xml? in the code i could find both variables.

  i read in the guide that dualsphysics surface tension formulation was based on CSF model and SPH implementation of Hu and Adams (2006), thats why i went to look further the paper, but i dont know if they refer to surfcoef or surfc1.

  thanks in advance
• olivliv

  January 2020

  

  got this from caseconfig at jsph.cpp

  
  

  is it surfcoef=surfc1?