SIROCCO, SIlant ROtors by aCoustiC Optimisation

1
SIROCCO, SIlant ROtors by aCoustiC Optimisation

• Participants
• Energy Research Centre of the Netherlands, ECN
  (Gerard Schepers and Toine Curvers)
• National Aerospace Laboratory, NLR(Stefan
  Oerlemans)
• University of Stuttgart, Ustutt(Kurt Braun,
  Andreas Herrig, Thorsten Lutz, Werner Wuerz)
• Gamesa Aeólica, Gamesa(Beatriz Mendez López,
  Alvaro Matesanz)
• GE Wind Energy/Global Research, GE (Rainer
  Arelt, Thierry Maeder)
• Funded by
• EU 5th Framework
• SenterNOVEM (ECN/NLR)

2
OUTLINE

• OBJECTIVE
• PROJECT SET-UP (Period, participants, tasks)
• MAIN RESULTS
• RESULTS FROM PHASE 1 ACOUSTIC ARRAY MEASUREMENTS
• RESULTS FROM PHASE 2 DESIGN AND VALIDATION OF
  ACOUSTIC AIRFOIL DESIGN
• CONCLUSIONS/FUTURE WORK

3
SIROCCO, OBJECTIVE

• To develop tip-airfoils (r/R gt0.75) by which
  aerodynamic noise of wind turbines can be reduced
  significantly without loss in aerodynamic
  performance
• Focus is on reduction of trailing edge noise!
• Background Previous EU project DATA
• Noise reduction of 3-6dB(A) on model wind
  turbine in the DNW wind tunnel
• Trailing edge noise dominant

4
SIROCCO, OBJECTIVE Ctd.

• Two baseline turbines
• Gamesa (G58, D58 m, 850 kW, at Zaragoza)
• GE (2.3 MW, D94 m, at ECN test field)

5
OUTLINE

• OBJECTIVE
• PROJECT SET-UP (Period, participants, tasks)
• MAIN RESULTS
• RESULTS FROM PHASE 1 ACOUSTIC ARRAY MEASUREMENTS
• RESULTS FROM PHASE 2 DESIGN AND VALIDATION OF
  ACOUSTIC AIRFOIL DESIGN
• CONCLUSIONS/FUTURE WORK

6
Sirocco Project period

• January 1st 2003 until February 28th 2007
• GE joined project in May 2005

7
SIROCCO, 4 phases

• Start-up phase Is acoustic behaviour of baseline
  turbines as expected, i.e. is trailing edge noise
  dominant?
• Acoustic array measurements Location and
  quantification of noise sources on wind turbine
  blade
• 2D design phase Design and test acoustically
  optimised airfoils
• Development of aero-acoustic design method
• Supported by 2D wind tunnel measurements
• 3) 3D design and manufacturing
• 4) Final validation in the field Compare
  acoustic and aerodynamic behaviour of optimised
  turbine and reference turbine

February 2006
8
SIROCCO Participants, main role

• Energy Research Centre of the Netherlands, ECN
• Coordination, design consultancy, aerodynamic
  field measurements
• National Aerospace Laboratory, NLR
• Acoustic measurements Field and wind tunnel (2D)
• University of Stuttgart
• Development design methodology for acoustically
  optimised airfoils
• Design of acoustically optimised airfoils
• Validation Wind tunnel (2D) Aerodynamic and
  acoustic
• Gamesa
• Design of blades with optimised airfoils and
  manufacturing
• GE Wind Energy
• Design of blades with optimised airfoils and
  manufacturing

9
OUTLINE

• OBJECTIVE
• PROJECT SET-UP (Period, participants, tasks)
• MAIN RESULTS
• RESULTS FROM PHASE 1 ACOUSTIC ARRAY MEASUREMENTS
• RESULTS FROM PHASE 2 DESIGN AND VALIDATION OF
  ACOUSTIC AIRFOIL DESIGN
• CONCLUSIONS /FUTURE WORK

10
Is trailing edge noise the dominant noise
source?

• Answer yes as derived from NLRs acoustic array
  measurements on GE2.3 turbine/G58 turbine
• Assume monopole source at scan point
• Location and quantification of noise sources on
  rotating wind turbine blades

GE
Scan points
Microphone array
G58
Delaysum
Sound rays
11
Acoustic array measurements Some observations
GE

• Turbine noise dominated by rotor blades
• Noise radiated from outer part of blades (but not
  the very tip)
• Practically all blade noise (emitted to ground)
  produced during downward movement

G58
12
Dominance of down-going blades

• The down-going blades are dominant for all
  frequencies and all measurements
• Indication for trailing edge noise to be dominant
• Can be explained by combination of
• Convective amplification
• Trailing edge noise directivity

sin2(?/2)
13
Noise sources on individual G58 blades (clean,
untreated, tripped)

• Typical source plots for individual blades
• Rotating focus plane for each blade
• Averaged over downward part of one rotation

12 dB
0 dB

• Tripped blade significantly noisier than other
  two
• Indication for TE-noise to be dominant

14
TE-noise dominant in calculations, but howgood
are the calculations?

• Spin-off Use measurements to validate wind
  turbine noise prediction code SILANT

15
SILANT

• Originally developed by Stork Product
  Engineering, NLR and TNO ECN
• Divide blade in a number of blade elements
• Calculate noise spectrum per element
• Inflow noise Amiet and Lowson
• Trailing edge noise Brooks Pope and Marcolini
• dp and ds at trailing edge from XFOIL and
  blade element momentum model
• Sum over elements

16
SILANT/meas (noise vs. power)
17
Measured spectrum vs SILANT spectrum Total
noise, trailing edge noise, inflow noise
18
Indications for TE-noise to be dominant

• Practically all blade noise (emitted to the
  ground) produced during downward movement
  Directivity of trailing edge noise
• Tripped blade significantly noisier than other
  two Tripping influences trailing
  edge noise
• Calculated results show TE noise to be dominant
  above inflow noise
• Blade noise levels scale with 5th power of local
  speed Dependancy of trailing edge
  noise
• Broadband TE noise is the dominant noise
  source

19
OUTLINE

• OBJECTIVE
• PROJECT SET-UP (Period, participants, tasks)
• MAIN RESULTS
• RESULTS FROM PHASE 1 ACOUSTIC ARRAY MEASUREMENTS
• RESULTS FROM PHASE 2 DESIGN AND VALIDATION OF
  ACOUSTIC AIRFOIL DESIGN
• CONCLUSIONS/FUTURE WORK

20
Main results Phase 2 2D design phase

• Design philosophy Modify boundary layer at
  trailing edge
• Improved noise airfoil prediction code coupled to
  an aerodynamic airfoil prediction code and
  optimizer
• Requirements/constraints of manufacturers
  implemented
• Aerodynamics Acoustics Mean
  boundary layer profile turbulent properties at
  trailing edge
• Using boundary layer wind tunnel measurements on
  airfoil with Variable Trailing Edge (VTE)

21
ORIGINAL LINK BETWEEN AERODYNAMIC/AERO-ACOUSTIC
MODELLING

• Mean boundary layer profile assumed to be a Coles
  profile
• Derived from integral boundary-layer parameters
• Turbulence properties
• v2 ,Kt derived from mean boundary layer profile
  with mixing length approach
• Vertical integral length scale (?2) from mixing
  length scale and scaling law
• Equilibrium approach!!
• Anisotropy is constant factor

22
DESIGN METHOD IMPROVEMENTS MADE DURING SIROCCO
PROJECT

• History and anisotropy effects important for
  flow regions with acceleration/deceleration
• EDDYBL FD boundary-layer code with stress-?
  model (Wilcox)
• Vertical integral length scale ?2 is not
  provided by available turbulence models
• adequate scaling laws required
• Directly found from ?2 measurements in the wind
  tunnel

23
Aerodynamic/Acoustic validation in windtunnel

• Acoustic validation 1) AWB tunnel (NLR) Array
  technique
• Open jet
• 1.2x0.8 m2
• 2) LWT tunnel (USTUTT) CPV technique

• Aerodynamic validation
• LWT tunnel of USTUTT
• closed test section
• 0.73x2.73 m2

Re1.6 106 Clean/tripped airfoils
CPV Coherent Particle Velocimetry Measurement
of particle velocitieswith hot-wire at pressure
and suction side of airfoil TE, instead
of (microphone) measurement of pressure
fluctuations
LWT
AWB
24
GE
Gamesa
Drag polars
Noise polars
25
OUTLINE

• OBJECTIVE
• PROJECT SET-UP (Period, participants, tasks)
• MAIN RESULTS
• RESULTS FROM PHASE 1 ACOUSTIC ARRAY MEASUREMENTS
• RESULTS FROM PHASE 2 DESIGN AND VALIDATION OF
  ACOUSTIC AIRFOIL DESIGN
• CONCLUSIONS/FUTURE WORK

26
Conclusions/Future work

• Acoustic array measurements very successfully
  applied on full scale turbine Trailing edge
  noise is dominant
• Design methodology for acoustically optimised
  airfoil has been improved and validated
  successfully in wind tunnel
• New airfoils
• Noise reduction 1-1.5 dB and 2.5-2.9 dB
• Aerodynamic characteristics hardly changed
• Blade design/manufacturing underway
• Full scale validation with hybrid rotor in April
  2006/Autumn 2006