Learning Forum

[Amir]

Hello everyone,

Description of the problem:
I am doing a 2D simulation on two droplets on a solid wall, the radius of droplets are 1mm and 0.5mm, the contact angle is 90, gravity were included (-y direction, i.e. to-down), however because of the size of droplets its influence is negligible. No other body force enforces on the droplets. 
For two cases I am doing the simulation,
1)    isothermal condition,
2)    evaporation on the bottom hot wall with 5 deg higher than the saturation temperature of the droplet.

Mesh:
My mesh is fine enough to capture the droplets and in x direction is purposefully uniform.
 
Boundary conditions:
The other three boundary conditions (up, left and right) are pressure outlet. 

Numerical setup:
I am using 
•    Simple pressure velocity coupling,
•    explicit formulation of VOF model 
•    second order discretisation for momentum and also for the case energy equation when is on, 
•    Geo-reconstruct for VOF, 
•    Presto for pressure ,
•    Implicit body force option were activated.
•    For the evaporation case, I am using lee model with different evaporation coeff.

Results and questions:
For the first simulation (without energy equation, without evaporation and in isothermal condition): 
•    droplet moves (oscillates) a little horizontally/vertically, which is tangible in the animation, which I don’t understand why?
For the second case (evaporation)
•    droplets move a lot and in some cases even coalescence, while there is no temperature gradient on the wall, contact angle gradient, and mesh density in x direction, so it should not happen :(

Would you please some experts help me with this problem which is counterintuitive.
Thank you in advance.

For the first simulation (isothermal), the volume fraction contour in sequence,

For the secondsimulation (evaporation), the volume fraction contour in sequence,

[Rob]

Gravity is acting down, but the wall angle is 90 degrees. That may cause some issues. What time step are you using?

[Amir]

thank you for your reply. I tried hydrophilic surface also e.g.,45 deg, but still droplet is not stable and moves.

the time step I chose is 1e-5

[Rob]

Try refining the mesh near the wall and lowering the time step. How well converged were the time steps? 

[Amir]

I took the adaptive time stepping with global courant number<0.2 for this case, fewer than 25 iteractions to be converged.

ok thanks, I thought the mesh size is ok, let me try a finer mesh near wall, I hope this problem will be solved

 

[Rob]

Turn off adaptive time stepping too. If the droplets are (mostly) stationary it's going to create some interesting numerical issues in the time stepping calculator. 

[Amir]

ok thanks alot, so you mean becasue the base of calcuation of time step in adaptive is based on cell’s speed,

when droplet is staionary evaporating and the numerically challeging phenomenon of mass transfer happens which need less time step to be captured,

but becasue of the low speed of the domain cells, wrongly the time step increase, and the physics is missed

am I right?

 

[Rob]

More or less. When the physics is complex the adaptive time step adjustment can happen a time step or so after things have gone wrong - the function is great for "simple" systems where there are no sudden changes in a result. Here, it's going to struggle.