Has anyone used this tool to model fluid flow with a free surface? I'm interested in modeling a falling stream of water. If anyone has seen the use of a different multiphysics tool I'd love to see that too. Thanks.

Administrator

There are a couple of mscript model examples available implementing the levelset method to demonstrate how one technically can do twophase flow with immersed interfaces:
https://github.com/precisesimulation/featoolmultiphysics/blob/master/examples/ex_multiphase1.m https://github.com/precisesimulation/featoolmultiphysics/blob/master/examples/ex_multiphase2.m https://github.com/precisesimulation/featoolmultiphysics/blob/master/examples/ex_multiphase3.m Note that redistancing/reinitialization of the levelset contour has not been implemented, so this method will be unstable for longer simulations (as the level set curve will deviate from a pure distance function). Freesurface simulations is also available with the external OpenFOAM CFD solver, but a compatible interface in FEATool has not yet been designed. 
Within the ex_multiphase3.m example, what is the physical meaning of the additional terms in the expression for miu:
'+2e2*',num2str(opt.hmax),'*sqrt(u^2+v^2+1e2) I am assuming it is related to surface tension, but I can't determine the origin of the 2e2 value. I also can't understand the units this 2e2 term would have to be to result in a viscosity (N/m) after the multiplications. The surface tension between water and air I've found as 72.8 mN/m at 20°C. Also, why is there a fundamentally different approach in the way ex_multiphase2.m handles the interface utilizing a smoothed delta function and force terms to handle the surface tension at the interface. I would assume the dam break problem could be modified to use this same bubble approach? 
Administrator

This is just a term for numerical stabilization of convective terms which is necessary for the simulation to converge. The constant in this case is just chosen a bit arbitrarily for the simulation to work as it is just a quick example. (Models only available in script format are considered experimental, and when "ready" will be made available in the GUI interface as well. In these models level set reinitialization is not present, so the level set field will eventually get too distorted and the simulations will fail). The approach is the same for both models in that a smoothed heaviside function is used to model the discontinuous density and viscosity fields. However, surface tension is not included in this (multiphase3) model as the dam break problem is such a large scale that surface effects would be negligible, at least at this mesh resolution, in contrast to the smaller scale bubble simulations. But yes, you could include a similar surface tension term as well if you find it contributes to your application. 
Free forum by Nabble  Edit this page 