[MekTek]

Hi all,

I am performing analysis on a radial turbine stator+rotor configuration. My issue is that the stator CFD analysis results does not match with the design conditions and values obtained from CFTurbo.

This is my approach to the work:

The model was generated from CFTurbo for the required input pressure, rotor speed, mass flow rate and temperature and output pressure and temperature.

The mesh is generated on turbogrid and analysis is done on CFX using turbo mode.

The CFD simulation of the stator is done with the below boundary configurations:

Configuration 1:
Inlet: Total pressure and temperature, Outlet: static pressure-->results in lower mass flow rate (roughly one-third of the design value).

Configuration 2:
Inlet: Mass flow and temperature, Outlet: Static pressure-->results in the total pressure at inlet and outlet increasing thrice the design values and high mach numbers (around 2)

Configuration 3:
Inlet: Total pressure and temperature, Outlet: mass flow outlet-->results in the simulation exiting due to errors.

What I have done so far to resolve:

1. Tried verifying the nozzle width from CFTurbo with manual calculations-my manual calculations suggest a bigger nozzle width. When implemented, it doesn't work too.

2. Increased and decreased the number of blades-very limited success when blades decreased, but issue persisted with increased number of blades.

3. For a lower mass flow rate (half the design value), the simulation does work without throwing up errors. But, the design mass flow rate is a requirement and hence cannot be lowered.

4. I have tried using both ideal and real gas models (Peng-Robinson) (ideal in cfx and real in fluent as the model values are pre-specified) but both result in the same issue.

Simulation settings:

1. My choice of gas model- ideal or real (Peng-Robinson)
2. Choice of turbulence mode (using k-epsilon)

3. Choice of heat transfer -isentropic expansion is required in the nozzle (I am setting the walls as adiabatic-is this right?)

Any pointers in resolving this is greatly appreciated.

[Stel]

Did you try a simulation with constant fluid properties (no equation of state)? Sometimes starting from scratch with real or even the ideal EOS makes the solver to diverge very easily. Try to make a first simulation with constant fluid properties (for instance at inlet P and T) and with Heat Transfer disabled. See how the model behaves, check if the pressure rise value makes sense. Than use this flow field as an initial guess for a simulation with the ideal or real gas model and the heat transfer enabled.