How dangerous are wind turbines in cold climate regions? Can we do something about it?
Wind Turbines in Cold Climate: Fluid Mechanics, Ice ...
Transcript of Wind Turbines in Cold Climate: Fluid Mechanics, Ice ...
Wind Turbines in Cold Climate: Fluid Mechanics, Ice Accretion and Terrain EffectsM Ronnfors, R-Z Szasz, J Revstedt, Lund University
Work packages
•WP1: Detailed simulations of ice accretionon a single blade•WP2: Studies of how the acoustic sourcesare affected by ice accretion•WP3: Studies of terrain effects•WP4: Studies of noise generation from several turbines
Goals•Develop tool to model simultaneously flow and ice accretion
–Efficient (relative)–Flexible
»Avoid/fewer model coefficients»Complex/moving geometries
–Combine with other modules»Performance»Noise
uAdt
dM φααα 321=
Wind turbine blades vs. aircraft wings
• Programs for ice accretion on aircraft wings– Not optimal for wind turbines
• Differences in – Incompressible vs compressible– Subsonic vs supersonic– High AOA vs low AOA– Small chords vs big chords
• Need better tools to predict ice on wind turbine blades
Source: pixabay.com
Methods ice accretion
• 1. In-house code– LPT (Lagrangian Particle tracking)– LES (Large Eddy Simulation)– Immersed Boundary Method
» Move surface with source terms, same mesh• 2. OpenFoam (OpenSource, free CFD software)
– Makkonen (Euler-Euler approach)– LPT (Lagrangian Particle tracking)– LES (Large Eddy Simulation)– Dynamic mesh
» Move mesh with surface
• Compare programs and models
Methods ice accretion
• Method 2 OpenFoam– Incompressible N-S– LES– FVM (2:nd order)
» Slower– Hexahedral grid
» Refined around the wing
• Method 1 In-house code– Incompressible N-S– LES– Finite differences (3:rd, 4:th)
» Faster– Equidistant cartesian grid
» No refinement around wing
Method 1 In-house code
Set-upParameter Value
Profile NACA 63415
Angle of attack 3∘, 9∘
LWC 0.37 g/m3
MVD 27.6 μm
Vrel 18.7 m/s
Re 2.49e5
Time 10.6 min
Mass of ice 24 g
Change the surface shape
• Every n:th time step– Can be extrapolated in time:
Vice= Vice* time– Trapped air can be accounted for
Ice shapes
Velocity fluctuationsRough iceClean airfoil Smooth ice
3◦
9◦
Comparing ice shapes, 3∘, method 1 and 2
Noise computations
•Hybrid-method (Lighthill)
•Advantages–Dedicated solvers for flow & acoustics–Acoustic sources can be iterated–Possibility of different
»Mesh»Computed physical time
ji
ij
xxT
ptp
c ∂∂
∂=∇−
∂∂ 2
22
2
2 ''1
Vortical structures (λ2)
Long structure
Small vortices
Acoustic sources
RMS Lighthill source
•Larger amplitude sources on pressure side
Acoustic density fluctuations
Acoustic presssure spectra
Some tones damped by ice
Acoustic presssure spectra
Continued work•Other icing conditions
–Add heat transfer•Acoustics
–Account for monopoles and dipoles as well
•Single and multiple turbines•Landscape/ground effects
Acknowledgements
• Financing: Energimyndigheten
• Computational resource: SNIC, Lunarc, Lund University