Hello everybody,
I have previously posted on a problem I was trying to simulate about multiphase flows
(Multiphase Flow LS - Air Bubble in Water Cylinder - Conservative forms fails) and I have tried working out the problem, but with small results. Therefore I decided to change approach and I started from the "electrocoalescence" model from the comsol model library.
My aim is to use electric forces to actuate bubbles, without any pressure driven flow. Therefore I started making changes to the original coalescence model. First I wanted to check what happens when we shrink the dimensions to the ones I would like to use in my device. I deactivated the electrostatic module, to look only at the two phases flow, I reduced the inlet velocity and I changed from two small bubbles to only a single larger bubble.
Then I compute the surface integral of the volume fraction of the fluid in the bubble for the time dependent simulation and I realize, mass is not conserved. So I try running the conservative form of the simulation, and again mass is not conserved (see graph). The original bubble is swept away (as it should by the inlet flow) and at a certain time it should disappear through the outlet and the volume fraction should go to zero, but instead, I have bubbles of fluid 1 entering through the inlet, when I have defined the volume fraction at the inlet to be 100% fluid 2. The volume fraction of fluid 1 is increasing in time. Anyone has a guess where my mistake is?
I am using version 4.2a. Thanks a lot for any comment on the model that I attach together with the results plots. The pressure point constraint was added only, because I have seen that it stabilizes the solution giving a reference (absolute) constraint.
Ah, the property solnum mistake is just due to the wrong multiple of the time step used as a stop for the simulation.
Thanks again,
Stefano
I have previously posted on a problem I was trying to simulate about multiphase flows
(Multiphase Flow LS - Air Bubble in Water Cylinder - Conservative forms fails) and I have tried working out the problem, but with small results. Therefore I decided to change approach and I started from the "electrocoalescence" model from the comsol model library.
My aim is to use electric forces to actuate bubbles, without any pressure driven flow. Therefore I started making changes to the original coalescence model. First I wanted to check what happens when we shrink the dimensions to the ones I would like to use in my device. I deactivated the electrostatic module, to look only at the two phases flow, I reduced the inlet velocity and I changed from two small bubbles to only a single larger bubble.
Then I compute the surface integral of the volume fraction of the fluid in the bubble for the time dependent simulation and I realize, mass is not conserved. So I try running the conservative form of the simulation, and again mass is not conserved (see graph). The original bubble is swept away (as it should by the inlet flow) and at a certain time it should disappear through the outlet and the volume fraction should go to zero, but instead, I have bubbles of fluid 1 entering through the inlet, when I have defined the volume fraction at the inlet to be 100% fluid 2. The volume fraction of fluid 1 is increasing in time. Anyone has a guess where my mistake is?
I am using version 4.2a. Thanks a lot for any comment on the model that I attach together with the results plots. The pressure point constraint was added only, because I have seen that it stabilizes the solution giving a reference (absolute) constraint.
Ah, the property solnum mistake is just due to the wrong multiple of the time step used as a stop for the simulation.
Thanks again,
Stefano