Fluids

Fluids

Vapor-liquid interface penetrates into liquid

    • Z.PAN-2
      Subscriber

      Dear Fluent professionals:

      I need some advice. I'm modeling the water evaporation using the transient species mass transfer model (Raoult's law). There are 2 phases. The primary phase contains liquid water, and the second phase contains water vapor and air.

      The problem is that the vapor keeps penetrating into the liquid. What I expect to see is the range of the interface, where the vapor and liquid water exist in the same cells, maintains a certain thickness. However, it seems like the interface keeps broadening downwards. The figure shows the mole fraction of vapor and liquid water after certain time steps. The height of the liquid water is almost the same from the beginning. Above the water is pure air at the beginning. There is only one pressure outlet on top, the rests are stationary walls with 0 heat flux and no inlet. The setup is given below. Any suggestion is welcomed!

      Meshing - have tried both quadrilateral (all rectangles), triangles (best fit)
      Transient - time step (have tried 5e-05 ~ 1e-10 s)
      Mixture model - Slip velocity, Implicit body force, Sharp/dispersed (zonal discretization)
      Phase interaction - continuum surface stress/force (tried both), wall adhesion, surface tension coefficient 0.072, drag coefficient (schiller-naumann)
      Species mass transfer - saturation pressure (piecewise linear), From phase (ranz-marshall), To-phase (zero-resistance)
      Interfacial area - ia-gradient
      laminar or k-epsilon model - enhanced wall treatment (thermal effect), turbulence damping

    • Rob
      Ansys Employee

      What's the scale? I'd expect some vapour diffusion into the liquid, but also check convergence & mesh resolution as a lack of numerical accuracy will enhance that. 

    • Z.PAN-2
      Subscriber

      Could it be the setting of the model?
      It is 2 cm wide and 7 cm high.
      Convergence seems fine. Every residual reduces to below 1e-03 in every time step.

    • Rob
      Ansys Employee

      If you've resolved the geometry & interface that's not a problem. I meant the contour scale, I assume 0-1 but it's not displayed. 

Viewing 3 reply threads
  • You must be logged in to reply to this topic.