[whuup]

Hi,
I'm using Fluent to calculate multiphase droplet flows. Fluent uses Finite-Volume-Method, right?

As I understand the defintion of FVM, the mass conservation should always be satisfied:

The most compelling feature of the FVM is that the resulting solution satisfies the conservation of quantities such as mass(...). This is exactly satisfied for any control volume as well as for the whole computational domain and for any number of control volumes. Even a coarse grid solution exhibits exact integral balances. http://www.cfd-online.com/Wiki/Finite_volume

In my calculation this is not the case. I calculate multiphase flows through complex 3D geometries and I need a minimum of 6k Iterations to get the difference between the mass-flow at the inlet and the outlet to be smaller than 1%. Until then sometimes up to 20% of the mass goes missing (Yes, the mesh is not ideal but no trash either)

-> How can I have an error in the mass continuity with the finite-volume-method of fluent? OR does fluent use finite-element-method? OR do I missunderstand something about FVM?

any answer is highly appreciated!
thanks whuup

(Details: 5-7M cells, Polyedra-mesh, 5 prism layers, eulerian, realizable k-e, enhanced wall treatment, 1-15% liquid, fully turbulent)