Hello,

since the beginning of the year I am working with CFX (previously I used Fluent) and I have a question concerning the mass flow BC. I set up a problem with velocity components as inlet BC and mass flow at outlet. The solution converged quite well (RMS residual->1.e-05) but the mass imbalance was 12%! By setting the same BC in Fluent the static pressure at the outlet was itterated in order to meet the mass flow and the imbalance was ok. Has anybody experienced the same problem? The imbalance (in CFX) reduced to 0.01% by setting the static pressure at the outlet, however I was confused that the previously used BC gave that bad result with regard to mass imbalance. The problem is subsonic. I calculated the flow between turbine exit and combustion chamber inlet thus no rotating domains were included.

Setting a velocity inlet and a mass flow outlet is an unbounded solution, and a poor choice for boundary conditions. You could have an infinite number of pressure/density fields that could result in the solution. You should use a pressure boundary somewhere, as you discovered.

Dear Michele,

In addition to Johnny's comments, the mass flow (numerical integral) at the inlet may not necessarily be the same mass flow at the outlet; therefore, there is no way the solver can enforce continuity..

Opaque.

Hello Johnny, hello Opaque,

thank you very much for your answers. Today everthing is clearer. I thought the solver would set the "reference pressure" at the inlet and change the relative pressure at the outlet to meet the desired mass flow. I think I should read a bit more carefully what is written in the help manual. To be honest I feel a bit uncomfortable for having posed such a "stupid" question. However your comments were a big help. Thanx.

Michele