Predicting Wind Turbine Blade Loads using Vorticity Transport and RANS Methodologies

1
Predicting Wind Turbine Blade Loads
usingVorticity Transport and RANS Methodologies
2
Vorticity Transport and RANS Models
Vorticity Transport Model
RANS

• Reynolds-Averaged Navier-Stokes equations are
  solved in finite volume form on a hybrid grid

• Solves unsteady vorticity transport equation

• Inviscid flux terms are solved by flux-difference
  splitting viscous terms by a central-difference
  method

• Flow assumed fully turbulent Spalart-Allmaras
  model used

• Solved in finite-volume form on a structured
  Cartesian mesh

• Simulations are steady, mostly using one blade
  with periodic boundary conditions

• Lifting-line blade aerodynamic model, trailed and
  shed vorticity is accounted for using the source
  term, S

• Numerical diffusion of vorticity is limited by
  using a weighted average flux method wake
  structure is preserved

3
Blade Aerodynamic Loading
Normal Force Coefficient, Cn

• Wind Turbine Model
• NREL Phase VI rotor
• Rigid
• 2 blades
• Radius 5.029m
• S809 aerofoil
• Rotor speed 72rpm (constant)
• Blade pitch 3

r/R
Wind 7 m/s
Tangential Force Coefficient, Ct
r/R
4
Blade Aerodynamic Loading
Cn
Cn
r/R
r/R
Wind 10 m/s
Wind 25 m/s
Ct
Ct
r/R
r/R
5
Distribution of Pressure Coefficient
Wind 10 m/s
Wind 25 m/s
6
Suction Surface Streamlines
7
Wake Structure
Radial Location (r/R)
VTM
Axial Location (z/R)
RANS
Vortex Age deg
Rotor
8
Conclusion

• VTM can be run using coarse aerodynamic
  discretisation to give efficient performance
  predictions
• VTM simulations using a finer discretisation of
  the wake have revealed the subtle characteristics
  of the vortex filaments and the changes in wake
  structure that result from natural vortex
  instability
• RANS provides accurate predictions of the
  aerodynamic loading on the blades, and the
  velocity and pressures fields that surround them.
  The computational cost is high, though.
• In the short term the VTM and RANS methods can
  be used to inform the results provided by each
  other
• In the longer term improvements in computational
  resources may allow hybrid VTM-RANS schemes to be
  used