IEA 22 MW - Issues with Turbulence

[Barry.Britton]

Hi,

I am modelling the IEA 22MW turbine with a fixed base and running a range of turbulent wind cases similar to Orcina’s setup in IEA 22 MW RWT validation study - Orcina for load case 3.5 (full-field, IEC 61400-1IB ETM). I’m running this setup for mean wind speeds from 3 - 25 m/s to get statistics, for OpenFAST I’ve had no issues apart from 11 m/s & 13 m/s are aborting due to this error message:

Figure 1: Error message from my initial 13 m/s run:

Figure 2: Results from the final 100s of the total 400s simulation, compared to OrcaFlex:

I tried 11 & 13 m/s for a second time in OpenFAST and got a similar result - both failing around 200s without any obvious reasons from looking at the results.

I know that these are the wind speeds near the maximum rotor thrust and flapwise deflections but I can’t quite understand why they are failing when 10,12,14 m/s etc run without any problems.

Figure 3: Full timehistory for 11 m/s:

Figure 4: Full timehistory for 12 m/s:

For other information: the same turbsim .bts files are used in both OpenFAST & OrcaFlex and the simulations are ran with an initial rotor speed of 3 rpm. And the only changes to the setup between speeds is the URef in the turbsim input file.

Can anyone here suggest anything that might be causing these failures?

Thanks,
Barry Britton
University of Galway, Ireland

[Jason.Jonkman]

Dear @Barry.Britton,

Just a couple comments / questions:

• Why are you initializing RotSpeed to 3 rpm for different wind speeds? We generally recommend initializing RotSpeed and BldPitch in ElastoDyn to their expected (mean) values conditioned on the mean hub-height wind speed you are simulating. Not doing so can often confuse the controller (which may not know what state it is in) resulting in unexpected behavior.
• You mention running OpenFAST a “second time and got a similar result”. What are you changing between simulations? I would expect OpenFAST would generate the same result if you run the same simulation multiple times.

Best regards,

[Barry.Britton]

Hi @Jason.Jonkman,

Thanks for your reply!

• The reason I was initializing with 3 rpm was generally because I’m trying to do a timetrace comparison with Flexcom as well as OrcaFlex. Flexcom doesn’t currently have the capability to initialise with a rotor speed - therefore I was attempting to get the fairest comparison (I tried OpenFAST with zero rotor speed and this wouldn’t work - but 3 rpm worked generally with some of the higher wind speeds requiring 7 rpm). But there seems to be no way around this - and I assume for mean/max/min comparison it’s ok to start with different I.C (would you have any comment on this?)

• Yes, that probably wasn’t entirely clear from me, I did a batch run overnight and was slightly concerned that this was a problem with my laptop - so I just ran the same setup again on 11 & 13 m/s because the results looked ok and there wasn’t anything clear that caused the failure. But third time round, I followed your instructions and this worked well, thanks! Here are the results for 13 m/s:
  

  

• These are statistics (final 50 seconds) from the full set of windspeeds 3-25 m/s (ETM), with the mean values plotted as solid lines and max & min as dotted (I haven’t got OrcaFlex to run below 10 m/s). We’re observing a discrepancy in blade pitch and RotAeroFxh around rated speed (11.5 m/s) which is of course impacting out of plane deflection and flapwise root moment. Would you have any comment on this? I can provide more information on the Flexcom setup, it’s generally trying to follow Orcina’s setup in their validation report IEA 22 MW RWT validation study - Orcina
  

  

Thanks!
Barry Britton

[Jason.Jonkman]

Dear @Barry.Britton,

I’m glad to hear that things are looking better now. I don’t think I can speculate what is causing the small differences that are remaining.

Regarding initializing at a small or zero rotor speed, I would expect Flexcom would need a modified controller that supports the start-up transient before normal operating mode. And if you’ve modified the controller for Flexcom, you should be able to use the same controller modification in OpenFAST.

Best regards,