[Fluidpower]

Hey there,
I have a familiar problem with air-bubbles in water. But all threads nor the WWW helped me to fix it.
I use a 18mm x 30mm 2D rectangle (made with spaceclaim) with air bubbles of 0.2-3mm diameter. The mesh has an element size of 0.05mm with Quads. The top is marked as outlet. Since a bubble rising is hard to detect already (my problem) I use the biggest diameter of 3mm here.
In short terms my model:
Vof, implicit body force (“level set” on and off: not the key for my problem), k-epsilon realizable, transient, pressure based, gravity, phase1 (water), phase2 (air), surface tension 73 dynes/cm ^-1 (mNm/m^-1), pressure outlet, operation conditions-operating density 1.225kg/m^3, Scheme-PISO, Pressure-PRESTO, Momentum- Second order upwind.
In the center I patch a region-circle with a radius of 1.5mm (diameter = 3mm) as phase2 -> air
Time step size 0.01s
1. Picture: Just the pressure level without a patched bubble. It stays the same with a patched circle as it should. The pressure difference is the same as calculated manually.
2. Picture: Time = 0.05s
3. Picture: Time =1s
According to my research (“A single air bubble rise in water”, M.T.Islam) and the book by R.Clift a al,1978 the terminal velocity is surely reached after around 0.3s and for a 3mm bubble diameter ~ 0.22m/s.
I transferred the data with Calculations activities to CFD Post. Unfortunately the selected volume fraction doesn´t appear as variable option but it did in other simulations before and the problem was also there.
In my pictures it is quite obvious that I have a fundamental problem somewhere. I have spent so much time with it already that I miss the wood for the trees. If somebody is there to help I would appreciate it. With a pm I can also send you a word document with all my steps in pictures.

Thank you in advance for your time and help!

[Fluidpower]

Picture numer 2

[Fluidpower]

Picture number 3

[sfn]

interesting.. a lot of things may influence the result. You should test other multiphase models, turbulence models, solution methods and find out which case gave you the most accurate results. You may try to generate new mesh with smaller cells, reduce the timestep, but I think it will not work. Nevertheless if I were you, I would use DPM to model air bubbles. It provided quite accurate results when I used it to calculte terminal velocity of mineral particles

[mvoss]

Did you activate gravity?!


Sent from my iPhone using CFD Online Forum mobile app

[sfn]

yes of course

[Fluidpower]

Exactly the gravity is included and correct. It´s also visible in the first picture where the gravity is part of the pressure level calculation (delta_pressure=roh*gravity*hight)

I tried the DPM before but I got an error message that I still have to figure out. I combined it with the VOF model. Otherwise I wasn´t able to patch the air bubble inside nor to set a surface tension. I doubt that the model would provide me all the information as the shape change of the bubble for instance. I understand just DPM as a pure tracking model, isn´t it?! But I honestly haven´t worked with DPM in detail yet. I´ll try more variations and settings.

I tried already different solvers. Indeed there are differences but not as big as it would even nearly compensate my discrepancy. In regards to other simulations I found online and made myself I think I can neglect a mesh problem or time step problem as you assumed already. With my current background knowledge it wouldn´t explain the huge difference.

More hints and advices are welcome

[Fluidpower]

Anyone else with any idea where I could have screwed it up?!

[mvoss]

Shouldn't the reference density be arround 998?


Sent from my iPhone using CFD Online Forum mobile app