[dold80]

Hello everyone.

I encountered a convergence problem when I tried to do a simulation of a turbine.

When I conducted a simulation at the best efficiency point, the user point (pressure on the blade surface) converge well without fluctuations.
But at another design point (maybe off-design point), the user point fluctuates as attached figure although it is a steady-state simulation.

My model is a hydro turbine with a full model, and I am using the MRF method. The frozen rotor was selected as an interface model.

To resolve this problem, I tried to set physical timescale with 1/ω, and auto timescale with a very small value (0.01), but it does not work. (The attached figure is the auto timescale with 0.04)

Are there any solutions or suggestions to solve this problem?

I look forward to your reply.

Thank you very much.

[Gert-Jan]

Are you trying to solve a steady state?

If you ask CFX to find a steady state which is not there, then you ask the wrong question. You should ask CFX to look for a transient solution.

[Stel]

Off-design conditions may be inherently transient in nature (especially when flow rates below the design condition are considered). Your response is not necessarily wrong. When stall cells develop inside the hydraulic channels, for example, velocity and pressure fluctuations will happen. Since the CFX Solver for Steady State simulations is based on a Pseudo Transient approach, this unsteadiness will be reflected in the solution.

Check your solution in CFD Post. Do you see lots of flow recirculation?

Cases like this should be solved as transient if you are really interested on those transient phenomena, or use a Steady State solution anyway and try to average global values if this is what you are interested in.