UnsteadyAerodynamicsOfFOWT:
Firma convenzione
TowardExperimentalValidationOfEquivalent
Politecnico
di
Milano
e
Veneranda
Fabbrica
Lumped‐elementModels
del Duomo di Milano
IlmasBayati,LucaBernini,AlbertoZasso
Aula Magna – Rettorato
Mercoledì
27 maggio 2015
Department ofMechanical
Engineering
2DoFSetup
• Polimi WT 2011 Test of Vestas V52 (may 2011)
• Surge imposed motion
• 1/25 geometric blade scale D=2.1m Ω=2.5Hz
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
2DoFSetup
experimental Session
•
•
Steady configuration
Unsteady configuration:
Surge andPitchSinewaves
• Wind/NoWind
measuremenents
• Inertial Forces subtraction
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
2DoFSetup
Previous Experimental Session
• Aerodynamic Histeresis registered
• Dimension andshape ofcycle depends
ondynamic conditions
• Does FAST/Aerodyn predict this
behaviour?
Example ofhisteretic cycles:
aerodynamic pitch moment
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
Modelling
ImposedMotionInFAST
UserInputFile
• PlatformDof:Surge andPitch
HydroDyn.dat
• AdditionalDampingandStiffnessmatrices
FAST8customversion
Definitionofacontrolforceatthetower’sbase
togetanimposedmotion
FORCE =K_add ∙(
− )+C_add∙(
−
)
K_add andC_add :
parameters oftheoscillator
f>10*fimposed motion
h=1
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
HydroDyn_input.f90
HydroDyn_Types.f90
HydroDyn.f90
K_add andC_add
(seeFAST7/Seismicmodule)
FAST8Simulations
DYNIN
Generalized Dynamic Wake(GDW)=acceleration potential method
•
Does not require iterativeprocess
o Moregeneralpressuredistribution accross rotor
o Fully nonlinear implementation:
turbulence andspatial variation ofinflow
o Inherent modelling of:
 Dynamic wake effect
 Tip losses
 Skewed wake
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
Accountingfor
timelag inthe
induced velocity
created byvorticity
shed fromblades and
convected downstream
Nomeclature
Experimental/Numerical test scheme
D
Thrust (T)
Wind (V)
Variable definition
 surge motion displacement
 surge motion velocity
 Apparent wind
0,5
Ω /2
Surge Amp (SGamp)
Surge Freq (SGfreq)
tsr
/
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
 Thrust coefficient
4
 Nominal tip speed ratio
 Effective tip speed ratio
Experimental vs Numerical results
Static thrust
Experimental
blade
are
geometrically scaled.
Polar data for model scale are
difficult to be determined with
high accuracy.
Different slope at high TSR
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
Experimental vs Numerical results
• Agreement Experimental & Numerical  Dissipative Hysteresis cycles • Disagreement in the time delay value i.e.  amplitude of hysteresis cycles
Dynamic results
• Surge motion frequency 0.4Hz
• Various amplitude
Exp SGfreq=0.4Hz
0.1
Num SGfreq=0.4Hz
0.1
0.095
0.09
0.09
0.085
0.08
Ct
Ct
0.08
0.07
0.075
0.07
0.06
0.065
0.06
0.05
0.055
0.04
2.5
3
3.5
4
0.05
2.5
TSR
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
3
3.5
TSR
4
Experimental vs Numerical results
Dynamic results
• Surge motion frequency 0.6Hz
• Various amplitude
• Agreement Experimental & Numerical 
Larger Hysteresis cycles • GDW underestimates the hysteresis effects
Num SGfreq=0.6Hz
0.1
0.095
0.09
0.085
Ct
0.08
0.075
0.07
0.065
0.06
0.055
0.05
2.5
3
3.5
TSR
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
4
Lumped-element model:
advantages
• SS model aero (control, integrated with hydro)
• Different parameters (wt verification) relationship with wind/sea states (nominal condition for simulations)
• Large wind farms control Ct
Rheological model
tsr
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
tsr(t)
Proposed lumped model
Objective:
Model the unsteady turbine response to platform motion
Maxwell + Voigt
r2
Ct
k1
Ct_unsteady
k2
tsr
Ct=Ct_static(TSR)+Ct_unsteady(tsr)
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
r1
tsr(t)
tsr oscillation due to surge motion
Proposed lumped model
Good results for turbine thrust unsteady modelling both for numerical and experimental data.
Parameters identification is required
Example of experimental data prediction 0.05
0.005
Example of numerical data prediction 0
0
-0.005
Exp Ct
lumped model Ct
Num Ct
lumep model Ct
-0.01
-0.05
Ct
Ct
-0.015
-0.02
-0.1
-0.025
-0.03
-0.15
-0.035
-0.2
0
10
20
30
40
50
60
-0.04
10
15
time (sec)
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
20
25
30
time (sec)
35
40
45
50
Lumped model identification
Lumped model parameter
Are identified via quadratic
error minimization for
each nominal TSR working
condition.
Model parameters function
of reduced frequency and
amplitude of the surge
motion
f rid
SGfreq  D / 2

V
(tsr )
Amptsr 
2
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
LIFES 50+ Project
2011 PoliMi tests
2016‐17 PoliMi
LIFES50+ tests
1st LIFES50+ deliverable for Polimi
is the validation of steady/unsteady AeroDyn for FOWT
The 2011 Polimi wind tunnel tests were used as preliminary set of data for the
numerical and experimental comparison
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
LIFES50+Anovelhybridrealtimeapproach
(Hardware‐In‐The‐Loop)
Aerodynamics (real)+Hydrodinamics (computed)
Force Input through controlled cables
• Lessconstrictivescalingissues
• Exploitingtheadvantagesofeachtestfacility
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
Aerodynamics
(Computed)+
Hydrodinamics (Real)
LIFES50+
Aeroelastic Model Blade Design: DTU 10 MW
PoliMi &DTU Airfoil Characterization
LowRe Wind Tunnel
Selig SD7032 Airfoil
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
LIFES50+
Aeroelastic Model Blade Design: DTU 10 MW
Aerodynamic Scaling
chord geom scaled
LIFES50+ DEV1 MODEL
0,15
0,10
0,05
0,00
0 000 0 000 0 000 0 000 0 000 0 001 0 001 0 001 0 001 0 001 0 001
Different airfoil used in WT
Model chord is different from the geometrically scaled.
Account for polar Reynolds dependency
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
LIFES50+
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering
Design and verification Tools
Since LIFES50+ will be a multidisciplinary project (Aero, Hydro, Structural, Control,…). Advanced simulation tools both for design and verification are actually under implementation at Polimi.
• Fast (aero‐servo‐hydro)
• Adams (Multibody)
• AdWimo (AeroDyn+Adams)
A design support multibody tool for assessing the dynamic capabilities of a wind tunnel 6DoF/HIL setup. Belloli‐Giberti‐Fiore DeepWind Poster
I.Bayati,L.Bernini,A.Zasso – Department ofMechanical Engineering