Hi all,
I’m trying to set up a simulation with multiple airfoil tables in FAST 8.10. I have 2 polar databases for Reynolds number of 10^6 and 10^7. However, the performance results given by Fast are higher than expected.
For comparison, I have removed the Reynolds interpolation and made the simulation run with only one airfoil table. I got the curves below for power coefficient at a fixed rotor speed of 10rpm.
I don’t understand why the curve with interpolation is above the 10^7 Reynolds number curve. It should be located between the curves without Reynolds interpolation.
I have read in the AeroDyn user’s manual that there is no extrapolation for Reynolds tables… Has somebody got an explanation for this ?
Best regards,
Aurelien
Dear Aurelien,
Can you plot both airfoil polars (Cl and Cd at Re=10^6 and 10^7) for comparison?
Best regards,
Dear Jason,
Thanks for your fast reply, I have plotted airfoil polars for the airfoil element of 20.4% thickness.
The polar curves for all 18 blade elements are in this excel file.
Airfoil polars.xlsx (377 KB)
Thanks for your help,
Regards,
Aurelien
Dear Aurelien,
I don’t see anything odd between your airfoil polars at 10^6 and 10^7 Reynolds number. (However, the airfoil polars at 44.4% thickness and 48.1% thickness look a bit odd.) I don’t typically use AeroDyn’s Reynolds number interpolation, but I’m not aware of any problems with it either.
What Reynolds number is the interpolated AeroDyn result based on?
Best regards,
Dear Jason,
The polars had been calculated using X-foil and Airfoil prep to extend range of angle of attack to [-180; 180°]. Does it answer to your question ? Because I’m not sure I have understood the meaning of it.
I wonder if the problem with interpolation may be due to induction factor calculation. Does Fast take into account Reynolds interpolation when calculating the induction factor ? Or is the Reynolds interpolation only use for polar coefficient values ?
Thanks for your answers,
Best regards,
Aurelien
Dear Aurelien,
We dug into the source code and found a bug. It turns out that in AeroDyn v14.03.01 within FAST v8.10.00:
- The Reynolds number is being set to zero for the induction factor (wake) calculation. (But because AeroDyn doesn’t extrapolate the airfoil data, this means that the airfoil data at the smallest Reynolds number will always be used in the induction calculation.)
- After the induction factors are calculated, the Reynolds number is being set correctly (based on the local relative wind speed, including induction) and is used to interpolate the airfoil data for the aerodynamic load calculation. (AeroDyn computes the loads after calculating the induction.)
This bug may have been around since AeroDyn v13 in FAST v7.
A quick fix to this bug is to change line 1053 of AeroDyn.f90 from:
O%AirFoil%MulTabLoc = u%MulTabLoc(IElement,IBlade)
to:
IF (.not. p%Reynolds) O%AirFoil%MulTabLoc = u%MulTabLoc(IElement,IBlade)
This will prevent the Reynolds number from being overwritten to zero every time step; however, the Reynolds number used in the induction factor calculation will still be one time step behind that used for the load calculation. Making the Reynolds number consistent between the induction and load calculations will require more changes to AeroDyn (and as far as we can tell, AeroDyn has never done that correctly). We will include the quick fix in the next release of FAST/AeroDyn. In the meantime, you want to make the change yourself, recompile FAST, and see if it resolves your problem.
Best regards,
Thanks a lot for your help Jason ! I’m actually trying to recompile FAST but i’m facing some troubles with it.
I have written a post describing my compilation problems in the appropriate topic. When I achieve to recompile FAST with your correction to the code, I will tell you if it resolves my problem.
Best regards,
Aurelien
