Axial induction factor in Dynamic inflow model

Wind & Water Rotor Aerodynamics
1

Dear NWTC Team,

I am a master student currently investigating dynamic inflow cases for my thesis by pitching the blades and simulating gusts in FASTv7. The implemented wind turbine is a scaled-down model developed at ForWind for wind tunnel experiments based on the generic NREL 5MW. When looking at the rotor aerodynamics, we are firstly focusing on the axial induction factors throughout the blade.

Here below are the main aerodynamic characteristics we are selecting in AeroDyn:

BEDDOES					StallMod
NO_CM					UseCm
SWIRL					IndModel 
0.005					AToler
PRANDt					TLModel
PRANdt					HLModel

In order to compare between dynamic and steady-state aerodynamic simulations, the InfModel flag would be correspondingly set as either DYNIN or EQUIL.

The first figure below illustrates the axial induction factors (“aa [-]”) throughout the blade (“r/R [-]” dimensionless blade segment radii) for design operation conditions. The second and third graphs show the axial induction factors for a dynamic inflow situation (blade pitching) at the segment located at, respectively, the 40% and 60% of the blade radius. All of them are obtained in the three following situations:

*  “aa EQUIL”: these are the axial induction factors obtained directly as output ([i]elm.AxInd[/i]) when setting the EQUIL inflow model.
  • “aa DYNIN”: it represents the axial induction factors obtained directly as output (elm.AxInd) when setting the DYNIN inflow model.
  • “reconstructed aa DYNIN based on Phi and C_N”: this is the axial induction factor calculated based on a reverve BEM approach using the element normal coefficient (elm.CNorm) and the local inflow angle (phi) at each blade element.

So, on the one hand, the axial induction factors throughtout the blade for dynamic inflow at steady-state wind conditions shows a curve tendency we cannot really understand. On the other hand, the value of these factors when calculating based on a reverse BEM approach with the outputs from simulating with the dynamic inflow model (“reconstructed aa DYNIN based on Phi and C_N”), tend somehow to those obtained from simulating in steady-state (“aa EQUIL”).
Hence, our main questions would be:

  1. Why do we see this behaviour on the axial induction factors when using the dynamic model?
  2. How would you recommend us to proceed with the comparisons?

Thanks in advance!

Best Regards,
Olga Belenda Álvarez

aa_rR.png
aa_rR.png778×583 9.73 KB

aa_0c4_time.png
aa_0c4_time.png778×583 8.58 KB

aa_0c6_time.png
aa_0c6_time.png778×583 8.53 KB

2

Dear Olga,

I’m not sure I really understand how you are deriving the axial induction factor from C_N and phi, so, I can’t really comment on that.

Regarding the difference between EQUIL and DYNIN in AeroDyn v13 within FAST v7, in EQUIL, the axial induction factor is calculated locally at each blade analysis node. But in DYNIN, the induced flow at the rotor is expressed as Fourier series in the radial and azimuthal directions with only 10 flow states considered, so, the induction may have similar trends as EQUIL as a function of span, but the induction will be much smoother from DYNIN (which is what I see from your results).

Best regards,