[michaelhoo]

When checking the CFX results, I found the interface between runner and draft tube shows different mass flow rate and other interfaces show the same.The error is around two percent.
This is stable simulation and only water involved.
Why is that?

[ghorrocks]

Is the simulation fully converged?

How are you calculating the mass flow rate on both sides of the interface? Does it match the mass flow rate calculations shown in the output file?

[michaelhoo]

It is fully converged.
Interface for computing mas flow rate: massFlow()@runner tube 1,massFlow()@runner tube 2.In one case, the mass fow rate for runner tube 1 is 101 kg/s equa to the inlet while the mass fow rate for runner tube 2 is 103 kg/s equal to the ouelet.The flow is also from the runner to draft tube.

[michaelhoo]

For the inlet ,set mass flow rate.
For the outlet ,static pressure is given.

[ghorrocks]

How tight is the imbalances?

Are you computing the massFlow()@XXX in the post processor or the solver? The post processor does not get the full integration point data, just the nodal data so is not as accurate as the solver.

I presume there is no pitch change across the interface.

[michaelhoo]

Thank you for your reply.
It's sure that the imbalance is below 1e-3. For one model, I have many cases and all have the issue about inconsitent flow rate in the same nterface.
I get the mass fow rate by using CFD-POST.
How can get accurate result from the solver? You mean monitoring the mass flow rate when the solver is running?
The pitch change is automatic.

[michaelhoo]

[Opaque]

Originally Posted by michaelhoo

What is the ratio between the flows on both sides of the interface? If you solution is reasonably converged, I guess is the same value as the pitch ratio or its inverse.

[ghorrocks]

Note for convergence the residuals and imbalances are different things. For this issue you want to make sure the imbalance is small. So is your imbalances under 1E-3 or the residual under 1E-3? An imbalance under 1E-3 is quite tight convergence but residuals under 1E-3 is loose convergence and I would expect you to have an imbalance of a few percent if this is the case.

To calculate values in the solver you make a monitor point, set to the expression you want to evaluate. You can see the evaluated values in Solver Manager. Then the expression will be evaluated with the full accuracy of all integration points rather than just nodal values.

[michaelhoo]

The ratio of mesh area is 1.02.And the the flow ratio is also 1.02.The pitch change is set at automatic.The interface are all 360-degree rotating surfaces. Is I should set the pitch change to 360?

[michaelhoo]

Thank you.
Yeah, perhaps I should monitor some phsical variables in the solver although I know it is nneded.

[michaelhoo]

Domain Interface Name : guide runner

Discretization type = GGI
Intersection type = Partitioner
Non-overlap area fraction on side 1 = 0.00E+00
Non-overlap area fraction on side 2 = 4.91E-06
Pitch ratio ( area side 1 / area side 2 ) = 1.000
Pitch angle for side 1 [degrees] = 360.001
Pitch angle for side 2 [degrees] = 360.000

Domain Interface Name : runner tube

Discretization type = GGI
Intersection type = Partitioner
Non-overlap area fraction on side 1 = 3.03E-03
Non-overlap area fraction on side 2 = 2.08E-01
Pitch ratio ( area side 1 / area side 2 ) = 0.978
Pitch angle for side 1 [degrees] = 360.000
Pitch angle for side 2 [degrees] = 360.000

[Opaque]

The final, converged, flow ratio across the interface will always be pitch ratio (conservation of xxx).

If you are concerned the automatic pitch ratio is not exactly 1 (since it is 360 on both sides), you can manually set the pitch ratio using the specified angles option. Then, your pitch ratio should be exact, and the ratio of the flows will match that value

[michaelhoo]

yes. It's okay since I set the specified angle.