[AS_Aero]

Hi All

I am doing a turbine simulation with cross flow. I am monitoring the Power value in Transient rotor stator case and the simulation ran for 100000 time step with RPM of 400 and 10m/s and 20blades. Its a quasi2D with 2mm thickness in Z direction. My power monitor is not stable yet although my residual is below 1e-4 and I had a time step of 1e-6 and the courant number max was 0.46 and RMS is 0.01.
So should I run it longer and see or is it normal to have a flucutating power ? And the value is not matching with the RANS results as well. I might have to average and see but not sure how I can average it in CFDpost for the monitor point values !!

Kindly help me to resolve this issue.

[Opaque]

In a transient rotor-stator simulation, the solution in most cases converges to a "steady-periodic flow". In a transient simulation, you are supposed to converge your coefficient loops to a certain degree; therefore, the residual plot is not a measure of "time convergence" only iteration within the timestep convergence.

Most important is what timestep you are using relative to the circumferential mesh spacing at the rotor-stator interface. How many stator faces does the rotor move across in a given timestep? If it moves too many each timestep, it is not small enough yet to capture the interaction.

Your solution is definitely not converged, and no conclusions can be made yet.

[AS_Aero]

Hi Opaque

To be honest I dint understand your explanation completly

What does this ''steady -periodic flow'' means and converge my coefficient loops to a certain degree ?
Regarding the timestep also I dint understand. Could you explain it in a basic manner since am not expert in tubomachinery simulation.

[ghorrocks]

steady periodic flow means that the flow is transient as each blade passing changes the flow field, but each blade passing is the same so it is a repeating pattern.

If you have not conducted a mesh size, time step size and convergence criteria sensitivity analysis your results could be completely wrong and give totally misleading results. You need to check your simulation is accurate before attempting to explain the results.

[AS_Aero]

I did the mesh size study in RANS and I am using that resolution for the transient as well. And in RANS its converging as well. I have given Residual value of 1e-4 and the results are converged. And when I have a huge recirculation for sure in transient simulation convergence will be a huge problem. Time step size I have to find out. But I am using 1e-6 as my time step for URANS and my Courant number is less than 1. So I thought its good enough. And as a monitor I am just monitoring the Power from the rotor.

[ghorrocks]

This thread is now heading in the same direction as your other thread (Negative value for torque).

Unless there is something unique you wish to discuss here, please post further discussion on this topic on the other thread. We don't want multiple threads on the same topic.