[kikimon]

I’m running a 2-D Steady simulation in an LP-turbine. Experimental results under the same conditions show no flow separation, but it should be close (very small Re).

The conversion graphs of my residuals and mass-weighted average Cp at back flow are linear with abrupt periodic spikes. The results during the linear stages of the conversion match with the experiment and the flow is attached. But during the spike stages, shows strong flow separation and therefore don’t match at all the experimental results.

I have used 3 turbulence models, kw, k-kl and Transition SST. I have also varied turbulence inputs and relaxation factors, which seems to only change period of appearance of the spikes.

How do I get it to converge without these sudden spikes? Any idea is very appreciate it.
Thank you very much.

Pablo