Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Turbulent two-phase flow

Please login with a confirmed email address before reporting spam

Hi all,

I am modelling a turbulent two-phase flow (phase-field) over a dike profile with different roughness sections. However I am having some troubles. Currently I am using an initial interface thickness of 0.07 metres. After approximately 4 seconds al the water dissapeared, which means there is loss of mass. When I decrease the thickness of the initial interface, and also decreasing the size of the mesh, I keep getting the error: 'non linear solver did not converge'. Besides this, the calculation time increases significantly when decreasing the mesh size and thickness of the initial interface.

Does anyone have a suggestion how I can change my model so it is capable to model the turbulent flow over the dike profile, with conservation of mass. I have to do a lot of simulations and therefore it is important that the calculation time will not become extremely large (days).

Thank you in advance.


With kind regards,

Anouk Bomers


1 Reply Last Post 2015年3月11日 GMT-4 09:49
Luke Gritter Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 2015年3月11日 GMT-4 09:49
Anouk,

Your boundary condition at the top of the water domain is inconsistent with your initial conditions. I recommend adding an air domain above the water domain so that you solve for the air/water interface in the Phase Initialization step. This will provide a better quality solution at time zero and increase the stability.

Mass loss is a common problem with phase field and level set analyses. A finer mesh will typically improve the performance in this regard, and adjusting the mobility tuning parameter can also sometimes help.

Due to the requirement of tracking the phase boundaries and the dynamics of a time-dependent fluid analysis, phase field and level set models often require a long time to solve. For this model, however, I would think you should be able to solve it in hours rather than days, assuming you're running it on a decent computer.

--
Luke Gritter
AltaSim Technologies
Anouk, Your boundary condition at the top of the water domain is inconsistent with your initial conditions. I recommend adding an air domain above the water domain so that you solve for the air/water interface in the Phase Initialization step. This will provide a better quality solution at time zero and increase the stability. Mass loss is a common problem with phase field and level set analyses. A finer mesh will typically improve the performance in this regard, and adjusting the mobility tuning parameter can also sometimes help. Due to the requirement of tracking the phase boundaries and the dynamics of a time-dependent fluid analysis, phase field and level set models often require a long time to solve. For this model, however, I would think you should be able to solve it in hours rather than days, assuming you're running it on a decent computer. -- Luke Gritter AltaSim Technologies

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.