CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   mass flow rate not conserved in turbomachine, interface defined wrong? (

wildli February 9, 2010 16:06

mass flow rate not conserved in turbomachine, interface defined wrong?
Dear all,
I am simulating two counter rotating axial wheel in one cylinder chamber using ANSYS FLUENT 12 and encountering a problem with mass flow rate not conserved.

Basically this is the flow description in this chamber:
flow runs into chamber inlet --> 1st axial wheel --> short volume before second wheel --> 2ed counter rotating axial wheel--> outlet

Boundary condition:
Pressure inlet with total temperature and gauge total pressure defined
Pressure outlet with gauge pressure and back flow temperature define
Moving reference frame to two wheels and stationery condition for other fluid field

Model: Pressure-based, steady-state, ideal gas setting, simple scheme,
mixing plane model for each interface between rotating area and stationery area

Result: Mass flow rate at inlet and outlet not conserved, and there is about 20% difference when they are stable.
Severe reverse happens at each interface

Possible reason: Interface between 1st wheel and short volume is set up as default in mixing plane model (radial mixing plane geometry + area averaging method), I doubt mass averaging method would be the right one for my setting, but reverse flow never dies out at interface and switching into averaging method diverges the result immediately.

Question: would FLUENT not be able to solve my problem? or because of my blade design, this is not a steady state problem so that reverse flow happens which can not allow me continuing the calculation?

If anyone here has similar experience, please help me.

cventura March 24, 2010 20:23

Hi Wild,

Well, my case is a little bit different from yours: I am trying to run a few steady-state simulations regarding a radial-inflow turbine, considering both the stator and rotor using Fluent-ANSYS 12. I am also using a mixing plane (axial mixing plane geometry + area averaging method) to model the stator-rotor interaction.
However, I have been experiencing similar problems and tested many different options to run the cases.
Initialisation is a key factor and therefore I am running my cases in a step-by-step mode increasing its complexity during calculation. I changed the inlet boundary condition for both the rotor and stator from "pressure inlet" to "mass-flow inlet" and used a stationary rotor to start with. By changing the boundary conditions this way Fluent will force mass conservation across the mixing plane (see manual). In a preliminary stage I did not use the mixing plane, so I set fixed reasonable values for both the stator outlet and rotor inlet.
I have also lowered the under-relaxation coefficients to reasonably low values and let the residuals decay a few orders of magnitude between each simulation step.
I am not sure if this can help you since I might not be able to answer your question directly.
These cases were also tested with CFX and the simulations without any problem, so I am not sure if this would be a better option for this type of cases.

johntynanrrp July 8, 2015 11:38

Wildi / Carlos,

Did you find any solutions. I am running a simple model in Star CCM+. I have an inlet plenum going into a fan fluid section and on to an outlet plenum. Moving reference frame on fan fluid is driving the fluid flow. Boundary conditions: Inlet 0 Total Pressure, Outlet 0 Static Pressure, No mass flow specified on boundaries.

When I add a mixing plane at the interface between the fan fluid and the outlet plenum I am seeing a difference in the inlet and outlet mass flow.

For example the mass flow at the inlet is 9.18Kg/s in comparison to 12.57Kg/s at the outlet.


All times are GMT -4. The time now is 20:29.