wind turbine aeroelastic modeling: basics and cutting edge...

Review ArticleWind Turbine Aeroelastic Modeling Basics and CuttingEdge Trends

Mesfin Belayneh Ageze12 Yefa Hu1 and HuachunWu1

1School of Mechanical and Electrical Engineering Wuhan University of Technology PO Box No 205 Luoshi Road Wuhan China2Department of Mechanical Engineering Woldia University Woldia Ethiopia

Correspondence should be addressed to Mesfin Belayneh Ageze elzmealyahoocom

Received 21 September 2016 Accepted 6 December 2016 Published 14 March 2017

Academic Editor Roger L Davis

Copyright copy 2017 Mesfin Belayneh Ageze et al This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited

The interaction of fluid flow and the structure dynamic of the system is a vital subject formachines operating under their coupling Itis not different for wind turbine either especially as the coupling enhanced formulti-MW turbinewith larger and flexible blades andcomplex control methods and other nonlinearity more comprehensive aeroelastic tools will be required to investigate the realisticphenomena The present paper will overview the aeroelastic tool for wind turbine the efforts done gaps and future directionsindicated One starts with background of the subject presenting a case study to demonstrate the effect of fluid-structure interactionconsidering NREL 5MW blade and a brief comparison of several aeroelastic codes Cutting edge efforts done in the area suchas complex inflow effect of geometric nonlinearity and other stability and smart control issues are addressed and concluded byelaborating the gaps and future direction of aeroelasticity of wind turbine

1 Background

Wind energy is emerging towards a mainstreaming competi-tive and reliable power technology and significant forecasting(such as [1ndash3]) revealed that the progress will continuestrongly So far design trends of wind turbine are towardslonger (refer to Figure 1) and more flexible blades smartrotor and control and offshore application Early designloads on wind turbine were evaluated on the quasi-staticaerodynamic calculation base with structural dynamic eitherignored or included through the use of estimate dynamicmagnification factor This upscaling trend demands a morecomprehensive tool to investigate the complex couplingbetween aerodynamic and structural dynamic of megascalewind turbines that is aeroelasticity The present paper willreview the trend of aeroelastic study of wind turbine andavailable aeroelastic codes and tools and current and futureresearch direction and gaps will be addressed

The review considered a number of original researcharticles review papers summary of reports project com-pletion reports manual and others thoroughly analyzedduring the preparation of this report In the following section

the basis of aeroelastic modeling of wind turbine will bediscussed including aerodynamic structural dynamic andtheir coupling

2 Aeroelasticity

Though the idea of aeroelasticity goes back to 1947 as intro-duced by A R Collar its application for wind turbine designand analysis begins as late as 1976 by Friedmann [4] whoderived a set of coupled flap-lag-torsional equation of motionfor a single blade In later years significant evolution has beenachieved [5ndash8] Aeroelasticity studies the coupling effectsbetween the inertia elastic and aerodynamic forces whichoccur as an elastic body exposed to a fluid flow As the scopeof the study is enhanced to include other relevant phenomenasuch as hydrodynamic for offshore wind turbines thermaleffects or control action one will have to extend the subjectas hydroaeroelasticity aerothermoelasticity or aeroservoe-lasticity respectively Refer to Figure 2 collarrsquos triangle

To model and simulate the aeroelasticity of wind turbineseveral aeroelastic tools and methods are available whichroughly include aerodynamic component to determine the

HindawiInternational Journal of Aerospace EngineeringVolume 2017 Article ID 5263897 15 pageshttpsdoiorg10115520175263897

2 International Journal of Aerospace Engineering

Airbus A380

05 3 5 13 16 2 5 75 1st year of operation

810 rated capacity (MW)45

wing span

Roto

r dia

met

er (m

)

250m Oslash

160m Oslash

126m Oslash126m Oslash

112m Oslash

15m Oslash

100503019997959391898785

80 m

Figure 1 Wind turbine upscaling trend [3]

Aerodynamicforce

Rigid b

ody

aerod

ynam

ics Static

aeroelasticityAeroelasticity

Inertial force Elastic forceMechanicalvibration

Figure 2 Collarrsquos triangle interaction of different forces

wind loads and structural part to calculate the dynamicresponse of structure with time history and spatial distribu-tion of wind as input Figure 3

In the following sections both aerodynamic and struc-tural dynamic component of wind turbine aeroelasticity willbe elaborated including methods to model the correspond-ing phenomena

21 Aerodynamic Aerodynamic part of aeroelastic analysisentitled to determine the aerodynamic loads developeddue to the flow pattern of wind against the wind turbineblade orientation Different models and methods have beendeveloped such as the very common and popular Blade

Element Momentum method lifting line panel and vortexmodels generalized actuator disc models and Navier Stokesbased solvers each theory possesses pros and cons In thissection a brief discussionwill be included about the commonaerodynamic methods

211 Blade Element Momentum (BEM) Method BEM alsocalled strip theory was originally introduced by Glauert [9] itis computationally fast and cheap and provided that reliableaerofoil data is available it will give satisfactory results It isa combination of the simple momentum theory and bladeelement theory [10] and assumes that there is no aerodynamicinteraction between all sections along the rotor and can betreated separately which imply there is no radial flow Referto Figures 4 and 5

The forces on the blades are determined solely by the liftand drag characteristics of the airfoil shape of the blades andit is assumed the flow is incompressible steady state Sub-sequently combining the two expressions for thrustnormalforces and torque frommomentum theory and blade elementtheory the BEM will be derived After some algebraicmanipulation the resulting relations will be ((1)-(2))

119862119897 = 4119865 sin120593 (cos120593 minus 120582119903 sin120593)1205901015840 (sin120593 + 120582119903 cos120593) 1198861015840(1 + 1198861015840) =

12059010158401198621198974119865 cos120593(1)

International Journal of Aerospace Engineering 3

CouplingHydrodynamiccontrol

middot middot middot

BEM

FEM

dFLdFD

dFL

dFL

dFL

dFL

dFD

dFD

dFD

dFD

+

dFL = Cl1

2U2

ＬＦcdr

dFD = Cd1

2U2

ＬＦcdr

InflowZ

UＬＦ

UＬＦ

UＬＦ

UＬＦ

UＬＦ

U (z) =Ulowast

kＦＨ ( z

z0)

Figure 3 Typical aeroelastic tool principle

dr

c

R

r

Ω

Figure 4 Schematic of finite number of elements of the blade (bladeelement theory)

And considering turbulent flow the thrust coefficient can beexpressed with Glauert correction as

119862119879 = 4119886119865 (1 minus 119891119892 sdot 119886) 119891119892 = 1 for 119886 le 0314 (5 minus 3119886) for 119886 gt 03

(2)

Flow direction

Rotatingactuatordisk

Streamtube boundary

U

A1

A2

U(1 minus a)

U(1 minus 2a)

Figure 5 Schematic of momentum theory

where 119862119897 is the lift coefficient 119865 is Prandtlrsquos Tip Loss Cor-rection Factor 120593 is the relative wind angle 120582119903 is the localspeed ratio 1205901015840 is the local blade solidity 1198861015840 and 119886 are theradial and axial induction factor respectively119862119879 is the thrustcoefficient and 119891119892 is Glauert correction factor

Perhaps at this point we can see that the classical BEMtheory is assumed a quasi-staticsteady flow condition but toinvestigate the unsteady aerodynamic effects of wind turbineadditional models have to be included such as dynamicwakeinflow yawtilt model and dynamic stall


100

150

200

250

300

350

400

Roto

r sha

ft to

rque

(kN

m)

50 6030 4010 200Time (s)

BEM with dynamic inflowMeasurement

Figure 6 Comparison betweenmeasured and computed time seriesof the rotor shaft torque for the Tjaereborg machine during a stepinput of the pitch for a wind speed of 87ms [7]

Dynamic InflowThere is a time delay for the wake behind therotor to maintain a steady state condition after a disturbancesuch as sudden change in pitch angle rotor andor windspeed and this phenomenon is called dynamic inflow As thevelocity field is the vectorial sum of the free stream velocityand the induced velocity the dynamic inflow represents thelater one Typical example can be evident from Tjaerborgemachine result presented at Figure 6 (reproduced from [7])For sudden change in pitch angle from 0 to 37 degrees at119905 = 2 s the rotor shaft torque drops from 250 to 150 kNmand it takes around 10 s to settle to the new equilibrium stateTherefore dynamic inflowmodel is required to predicate sucha delay

Under Joule 1 program several investigations have beendone on the effect of dynamic inflow and implementationinto engineering methods [11ndash16] The most accurate modelto represent dynamicwake is unsteady vortexwakemodel butit has computational drawback that makes it less favorable forengineering application following Snel and Schepers [11] whoformulated six different engineeringmodels to determine theeffect of dynamic inflow phenomenon One of these methodsproposed by Sige Oslashye is a filter for the induced velocityconsisting of two first-order differential equations (refer toHansen et al [7] for details) To alleviate the numericaldemand of the existing models [17] proposed a simplifiedmodel which is an approximate modeling of dynamic inflowThe method is placing a lead-lag filter after rotor torqueand thrust calculated from static tables of the power andthrust coefficientsThe filter constants will then vary with theaverage wind speed

Dynamic Stall It is a rapid aerodynamic change that maybring about or delay stall behavior Due to tower shadowyaw or tilt wind shear andor turbulent wind conditionthe boundary begins to separate at the trailing edge andgradually moves upstream with increasing angles of attackthat is dynamic stall Dynamic stall effects occur on time

delay proportional to chord divided with the relative velocityseen at the blade section This phenomenon results in highlytransient forces and results from [18ndash20] revealed alsothe significant effect of dynamic stall more specifically forinstability problems Variety of dynamic stall models havebeen developed such as Gormont and BeddoesndashLeishmanmodel [7 10 20] Considering unsteady loads on wind bladeand the negative influences on the performance and fatiguelife of a turbine introducing a dynamic stall control methodis necessary The control methods can be active control [21]or passive control (such as streamwise vortex generatorsspanwise vortices generated using an elevated wire and acavity to act as a reservoir for the reverse flow accumulation)And these control methods showed significant delay of theonset of dynamic stall by several degrees and reduce theincreased lift and drag forces as well [22]

YawTilt Model During misalignment between rotor normalvector and the incoming wind yawtilt model will redis-tribute the induced velocity so that the induced velocities arehigherwhen a blade is positioned deep in thewake thanwhenit is pointing more upstream Glauert proposed the yawtiltmodel and more detailed discussions are included in [14 15]Hansen et al [7] also discussed a yawtilt model adapted fromhelicopter literatures The typical feature of yawtitle modelis to include the effect of misalignment by increasing anddecreasing the induced velocities on the downstream andupstream part of the rotor disc respectively

212 3D Inviscid Aerodynamic Models These models devel-oped to obtain more detailed description of the three-dimensional flow that develops around a wind turbine withviscous effects neglected Besides maturity to apply as engi-neering tools thesemodels contribute a better understandingof dynamic inflow effect and overall flow development [7 14ndash16] To include the viscous effect several attempts had beenmade in time using viscous-inviscid interaction techniques[7] Detailed discussion about the 3D inviscid models such aslifting line panel and vortex models can be found in [7] andpotential applications and challenges are included

213 CFD Based Models CFD application developed fromaerospace industry which employs potential flow solvers toalleviate their limitation and the use of unsteady Euler solversemerged As the computing power grows the application offull Reynolds Averaged Navier Stokes equations includingviscous effects applied for helicopter rotor computations lateron the full Navier Stokes computations of wind turbine rotoraerodynamics was reported [7]

NS solver originally developed from an aerospace codesolving compressible NS equations intended for high speedaerodynamic in subsonic and transonic regime apparentlythis nature of the code is not compatible for wind turbineapplication because of low Mach numbers around the rootsof the blades As the flowapproaches the incompressible limitit is very difficult to solve the compressible flow equationTwo remedies have been suggested [7] the first one ispreconditioning that changes the eigenvalues of the system


of the compressible flow equations by premultiplying thetime derivative by a matrix The other method is artificialcompressibility method in which an artificial sound speed isintroduced to allow standard compressible solution methodsand schemes to be applied for incompressible flows Thismethod has many merits such as ease of implementationof overlapping grids as the compressible codes and themain limitation is problem to enforce incompressibility intransient computations without the need for a huge amountof subiterations and the problem of determining the opti-mum artificial compressibility parameter The method iswell suited for solving nearly incompressible problems oftenexperienced in connection with wind energy In connectionwith steady state problems the method can be acceleratedusing local time stepping while the method using globaltime stepping still is well suited for transient computa-tions

22 Structural Dynamics Structural component of aeroelas-tic model will determine the dynamic response of the systemfor aerodynamic load and exchange results with aerodynamiccomponent simultaneously The earliest work on dynamicmodeling of wind turbine was by [4] which is the equationof motion of a single blade assumed as an elastic beam withthe root being fixed at the hub and the tip being free Withapplication of Hamiltonrsquos principle and Newtonian methodequations of motions which are valid to second order forlong slender straight homogeneous and isotropic beamsundergoing moderate displacements have been developedby [23] These equations are also validated for several beamproperties and the final equations include different nonlinearstructural and inertial terms which influence the aeroelasticstability and response of hingeless helicopter rotor bladesIn extension to the previous work [24] provides a newset of partial differential equations of motion for a windturbine blade rotating in a gravity field with variable rotorspeed and pitch action Hansen et al [7] employed the twofrequently used approaches (principle of virtual work withmodal shape function and nonlinear beam theory with FEM)to formulate the dynamic structural model of wind turbineAs the flexibility and length of wind turbine are increasingthe capability of the classical beam theory to model thestructural dynamic will not be enough in contrary to the factthat the utilization of more nonlinear beam theory with lessassumption is demanded

Besides the beam theory (linear or nonlinear) and ele-ments (shell or beam elements) to be employed there arethree frequent discretizationmethods tomodel the structuraldynamic in relation to wind turbine that is modal reductionapproach multibody dynamics (MBD) and finite elementsmethods (FEM)

FEM Approach It discretizes the wind turbine system tofinite elements as flexural beam lumpedmasses springs andjoints The methods have the advantage of fewer restrictionsregarding the type of configuration to consider such asgeometrical and material nonlinearity apparently this willresult in a high degree of freedom which will lead to highcomputational effort by extension cost

Modal Approach In this method the deflection of compo-nents such blade tower and support structure is superim-posed from linear combination of some physically realisticmodels typically the lowest eigenmodes such as 1st and2nd flapwise and edgewise modes The deflection of bladesand tower is coupled with a low number of prescribeddiscrete degrees of freedom In contrary to its computationalefficiency this approach has various limitations such as a fixednumber and type of degree of freedom the assumption of lin-earity and inadequacy to handle a certain type of structures

MBDApproach In thismethod the structure is approximatedby a finite number of elements consisting of rigid and flexiblebodies coupled by elastic joints This discretized systemis described with a finite number of ordinary differentialequations This approach combines the merits of both abovemethods since it needs relatively less set of equations ofmotion and nonlinearity is considered In addition thismodel treats nonlinear kinematics efficiently compared toFEM and allows modeling of mechanical system with bothlarge deflection and large rotation

23 Fluid-Structure Coupling The final stage of aeroelasticmodeling is fluid-structure coupling so that the responses ateach model (aerodynamic force and structural deformation)mapped to one another In classical aeroelastic methodsthe fluid and structure interaction is treated separately anduncoupled ignoring the interaction [5] As the computingpower improved several integrated approaches developedincluding inherent fluid-structure coupling As the fidelityrequirement of the analysis increases to include explicitdetails such as turbulence [25] nonlinear composite layeredblades and large deformation [26] application of strong fluid-structure coupling is necessary However on the contrarycomputational efficiency with small compromises on accu-racy is also another route of coupling demands such asintroducing reduced order model as [27] As the choice ofaeroelastic tool is dependent on the application focus areaor operational conditions to be studied accuracy demandedcost time and computational resource available understand-ing the limitations and merits of each tool over the other isvital A few comparisons have been made in this regard aspresented [26 28 29] in the following section a simple casestudy will be presented to compare one- and two-way fluid-structure coupling

Case Study FSI for NREL 5MWBaselineWind Turbine BladeIn the following few paragraphs to demonstrate the effectof coupling choice a simple case study will be discussedconsidering uni- and bidirectional fluid-structure couplingfor half cycle of operation of a blade that is 25 secondsThe target wind turbine blade is NREL 5MW baseline windturbine blade with 615 length and 15 hub radius the aerofoiland chord distribution is based on [30 31] with some minormodification at the tip and root sections Figure 7 Thematerial distribution for leading edge root trailing edgetip and spar caps is set as EP-LT-5500EP-3 composite forthe spar webs it is SaertexEP-3 composite The simulationis carried out on ANSYS that is the fluid flow on ANSYS


Y XZ

Figure 7 NREL 5MW blade 3D model

Fluid domainVelocityinlet Outflow

BladePeriodic boundary

Figure 8 Computational domain

Table 1 Simulation parameters

Parameter ValueBlade lengthhub radius 615m15mRated rotation speed 121 rpm13 radsRated wind speed 114msFlow solver ANSYS FluentTurbulence model Shear stress transport modelStructural solver ANSYS Mechanical

Fluent and the structural model on ANSYS Mechanicalother simulation parameters are included in Table 1 Forunidirectional coupling the ANSYS system coupling featureis employed

The computation domain is limited to be as 13rd of therotor to reduce the computational effort periodic boundarywith 120-degree spacing will be introduced as shown inFigure 8 comprising 251198646 elements for flow solver and 851198644elements for structural solver

Based on the unidirectional and bidirectional couplingthe simulation of the blade done and the velocity pressureand tip deflection were examined as shown in Figures 9ndash11

Comparing the results one can evidence the significantdifference between the two couplings specifically bidirec-tionally coupled simulation produced maximum tip deflec-tion of 12662m while it is 11174m for unidirectional cou-pling moreover the deflection and equivalent stress in uni-directional coupling remain constant after some variationsas compared with its counterpart which is still changing sig-nificantly implying damping difference To finalize althoughthere is enormous amount of computation effort requiredfor strong coupling between fluid and structure to simulatethe full aeroelasticity characteristics of wind turbine with aproper degree of fidelity application of stronger coupling isuncanny

3 Aeroelastic Codes for Wind Turbine

Variety of codes are available to model design and simu-late the aeroelastic characteristic of wind turbine refer to

Table 2 Several aeroelastic modeling codes verifications andliteratures including [32ndash41] are reviewed For offshore windturbine interested readers Passon and Kuhn [39] reviewedcodes which are suitable for such application

In addition to the codes shown in Table 2 several com-mercial and academic institutes developed variety of aeroe-lastic codes such as FOCUS at Stork Product Engineeringthe Stevin Laboratory GAROS at aerodyn-EnergiesystemeGmbH Cp-Lambda at Politecnico di Milano [42] andBHawC at Siemens Wind Power [43] and some general pur-pose programs such as ANSYS ABAQUS and SOLVIA withadd-on packages cosimulation or subroutine programs canbe employed to work with the aeroelasticity of wind turbine

4 Cutting Edge Trends and Gaps

The current energy market demands efficient cost effectivereliable sources as the development of wind turbines withlarger more flexible design (especially torsionally) withcomplex control is inherent Coupling of different phenom-ena and their nonlinear characteristics are escalating thechallenge to alleviate such challenges several researches havebeendone and tools are formulated In this section effort doneto improve and study the aeroelastic characteristic of windturbine system will be reviewed The review is categorizedinto fourmajor areas as complex inflow geometric nonlinear-ity and large blade deflection aeroelastic stability and smartcontrol

41 Complex Inflow Complex terrain will result in extremewind shear and high turbulence intensity and interactionswith large blades and towerwill cause variation of the inducedwind flow as function of blades azimuthal position Hence itis obvious that reliable tools are needed to map the energyproduction and loads expected to improve cost of repair andfatigue life of components Thereafter it has been one area ofinterest of the industry

411 Wind Shear The European UpWind [44 45] projectperformed 3D CFD rotor computation using EllipSys3DNavier Stokes solver to provide new insight about rotoroperation in shear with the aim of improving engineeringmodels The results include the azimuthal variation of rotorloads and inflowvelocity thewake behavior downstream andthe disturbance of the upstream flow due to the rotor loading

412 Tower Shadow Effect As the blades pass the tower thepressure driving them will be weakened so as the instantpower production and the aerodynamic loads creating cyclicimpulsive load on the rotor In general term tower inter-ference can be modeled as anemometer reading [46] CFDsimulation [47 48] or using potential flow method [49]Gomez and Seume [47] evaluated the cyclic load variationdue to tower interference and the results adopted to correctthe prediction of BEM Several investigations had been madeto evaluate their fidelity and [50] simulate wind turbine rotorand tower interaction with wind shear using CFD modeland the result showed this model underpredicted the effectcompared to BEM [46] also extended the effort for various


Table2Com

paris

onof

aeroela

sticcodesfor

windturbine

Nam

eofthe

code

Started

Develo

per

Aerodynamic

mod

elStructuralmod

elDescriptio

n

GHBladed

1996

GarradHassanand

PartnersLtd

BEM

theory

Mod

alapproach

Since1999thistoolextend

edforo

ffsho

reapplicationwith

mon

opile

orgravity

-based

foun

datio

nsTh

elatev

ersio

niscapableo

fmod

elingandanalyzing

both

onshorea

ndoff

shorew

indturbinew

ithvarie

tyof

supp

ortspecification

HAW

C22003ndash200

6

Risoslash

National

Labo

ratory

TechnicalU

niversity

ofDenmark

BEM

theory

Multib

ody

dynamics

Itisas

uccessor

forH

AWEC

tool(w

hich

utilizesF

Emetho

dusingsubstructure

approach

with

Timoshenk

o-beam

elementsforstructuralm

odeling

)HAW

EC2isa

timed

omain

morec

omprehending

toolTh

eaerod

ynam

icmod

elismod

ified

tohand

ledynamicinflo

wdynamicstallskew

inflo

wsheare

ffectso

ntheind

uctio

nandeffectsfro

mlarged

eflectio

n

ADAMSWT

mdashMechanical

Dyn

amicsIncun

der

contractof

NRE

LBE

Mtheory

Multib

ody

dynamics

ADAMSWTisreplaced

with

FAST

-to-ADAMSPreprocessorA

plug-in

AdWiM

oandAe

roDyn

with

A2A

Dinterfa

cecanalso

beintegrated

with

multib

odysoftw

are

ADAMSsolver

tomod

elandsim

ulate

Alcyone

mdashCenterfor

Renewable

Energy

Source

ampNTU

ABE

Mtheory

FEM

NationalTechn

icalUniversity

ofAthens

developedAlcyone(

freew

ake)with

free

wake

panelm

ethodwhich

also

inclu

desa

simulator

ofturbulentw

indfieldstim

edo

mainaeroela

sticanalysisof

thefullw

indturbinec

onfig

uration

and

postp

rocessingof

loadsfor

fatig

ueanalysis

TURB

U2007

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamics

Itisafrequ

ency

domain

lineariz

edaerohydroservoela

sticcode

andthea

ctive

aeroela

sticcontrolcod

e

DUWEC

D1986

TUDelft

BEM

theory

Multib

ody

dynamics

In1993

itwas

mod

ified

tomod

eloff

shorew

indturbinea

ndlatertoinclu

dewave

loads

FAST

mdashOregonState

University

under

contractof

NRE

LBE

Mtheory

Mod

alapproach

Thistoolhasm

uchles

srun

timeIn

1996N

RELhasm

odified

FAST

tousethe

AeroDyn

subrou

tinep

ackage

developedattheU

niversity

ofUtahto

calculatethe

aerodynamicforces

alon

gtheb

lade

FLEX

5mdash

DTU

BEM

theory

Mod

alapproach

Itistim

edom

ainaeroela

sticsim

ulationtoolanduses

relatively

fewer

degree

offre

edom

tomod

eltheturbine

FLEX

LAST

1982

StorkProd

uct

Engineering

BEM

theory

Multib

ody

dynamic

Since1992thiscode

hasb

eenused

asdesig

nandcertificatio

ntool

PHAT

AS

1993

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamic

Todeterm

inethe

nonlineard

ynam

icbehavior

andthec

orrespon

ding

loadso

faho

rizon

tal-a

xiswindturbine(bo

thon

shorea

ndoff

shore)in

timed

omain

TWISTE

R1983

StentecB

VTh

eNetherla

nds

BEM

theory

FEM

Initiallyitisused

tobe

calledFK

Asince

1991itsup

portsscholastic

windfield

simulation

VID

YN1983

Tekn

ikgrup

penAB

Sollentun

aSw

eden

BEM

theory

Mod

alapproach

Itissta

rted

aspartof

thee

valuationprojectsconcerning

twolargeSw

edish

prototypes

Maglarp

andNassuden


(a) (b)

Figure 9 Velocity distribution at 25 seconds Unidirection coupling (a) and bidirectional coupling (b)

(a) (b)

Figure 10 Pressure distribution at 25 seconds Unidirection coupling (a) and bidirectional coupling (b)

wind turbine concepts Zhang et al [49 51] also proposed a3D potential flow model of tower interference for BEM Asthe effect of both wind shear and tower shadow is significanton the power production as well as the loading of the rotorimprovement of the current models and new methods areexpected

413WakeOperation Wind turbines in farmwill be exposedto upwind wake operation which needs better modelingtool to develop better control algorithm adapted for loadreduction in wake Variety of wake models are availabledepending on the fidelity and application required and theeffort and computational resource available The traditionalway to model wake operation is an Equivalent TurbulentMethod [52] that is it takes into account the wake byincreasing the effective turbulence intensity It is based onthe assumption that all load generating mechanisms causingincreased loads in wake operation can be merged into anequivalent value of increased turbulence intensity and isincluded in IEC6400-1 standard for wind turbine safety [53]For extreme response during operation the success of thisapproach depends significantly on the physical mechanismcausing the extremes that is if the physical mechanismcreating increased loads in wake operation differs fromincreased turbulence intensity the resulting extremes mightbe erroneous [54] Other wake models (from lower to higherfidelity resp) are empiricalmodels (eg Parkmodel [55 56])

linearized RANSmodels (eg Eddy viscosity model [57] andFuga model [58 59]) probabilistic and conjugative methods(eg dynamic wake meandering [60 61] and stochasticmodel) nonlinear RANS models (eg 119896-120596 closure withactuator disk line and fully resolved) large eddy simulationmodels (eg dynamic Smagorinsky with actuator disk line)and vortexmethod [62] Power prediction and annual energyproduction tool requires steady and time-averaged wakemodels whereas load calculation requires unsteady andtime accurate and for control strategies both steady andunsteady will be applied The dynamic wake meanderingmodel is more detailed model considering the transversaland vertical dynamics of the wake (ie wake meandering)Thomsen et al [54] compared the load response for a windturbine in wake operation using equivalent turbulent andwake meandering methods and revealed the wake modelconsidered has significant influence for extreme load undernormal operation Ott et al [58] considered three closures asthe ldquosimple closurerdquo using an unperturbed eddy viscosity themixing length closure and the E-120576 closure As comparisonwith wind farm data the ldquosimple closurerdquo showed satisfactoryagreement while mixing length closure and E-120576 closure areunder- and overestimated respectively and for near wakecase allmodels fail Bastankhah andPorte-Agel [63] proposeda new analyticmodel forwind turbinewakesThismodel onlyrequires one parameter to determine the velocity distributionin the wake And the comparison of the high-resolution wind


One-way couplingTwo-way coupling

0

2

4

6

8

10

12

14

Tota

l tip

defl

ectio

n (m

)

05 10 15 20 2500

Time (s)

(a)


0

1

2

3

4

5

6

Max

imum

stre

ss (P

a)

05 10 15 20 2500

Time (s)

times108

(b)

Figure 11 Blade tip deflection (a) and maximum equivalent stress (b) for half cycle of operation

tunnel measurements and the LES results shows that thevelocity profiles obtained with the proposed model are inacceptable agreement with both

There have been different benchmarking and validationresearch for wake models such as [54 61 64ndash67] Thoughthese wake models are developed there are still gaps in thesubject including modeling wake-wake interaction wake-terrain interaction and understanding influence of atmo-spheric stability and nonuniform terrain further more eval-uating these models using yaw control [68] and integratingwith full 3D CFD models

42 Geometric Nonlinearity and Large Blade DeflectionLonger and more flexible blades with mechanical propertiesof high strength and relatively low Youngrsquos modulus (ielower stiffness) will deform significantly Therefore it is clearto include its effect in wind turbine analysis as it has animpact on the overall efficiency of the structure includingaeroelastic stability [69ndash72] Most of the existing commercialcodes use simple linear structural model which might notbe enough to consider large deformationThus it is necessaryto understand the various nonlinear interactions thoroughlyand develop a geometrical nonlinear analysis method forsuch wind turbine blades Different approaches have beenused to deal with large deflection problems such as ellipticintegral formulation numerical integration with iterativeshooting techniques incremental finite element methodincremental finite differences method method of weightedresidual (MWR) and perturbation method [73ndash77]

Larsen et al [72] incorporate three nonlinear approachesto evaluate the effect of including large deflectionThe resultsshowed including the influence of large deflection will reducethe effective rotor area causing a reduction in power output

at low wind speeds and a change in pitch angle setting athigh wind speeds which lead to a higher flapwise mean loadlevel On the contrary no main differences regarding fatigueload levels could be obtained from the load simulations Forstructural behavior an increment in flap frequency is seen asa function of deflection whereas edgewise frequency seems toremain constant Kallesoslashe [70] investigated the effect of bladedeformation on flutter boundaries by comparing naturalmodes of aeroelastic motions of an undeformed blade tothat of a predeformed blade The theoretical analysis showedldquothe flutter instability known from the undeformed bladeis delayed to a higher rotational speed on the other handa new rout to flutter instability appears which has a lowerstability boundary then the original flutter boundary forthe undeformed bladerdquo which imply the significant effect oflarge blade deflectionThe effect of edgewise bending-torsioncoupling on flutter limits of wind turbines is investigated by[69 71] using the aeroelastic mode suggested by [78] andindicated slightly decreased flutter limit on the rotor speeddue to the blade deflection

Yuan and Chen [76] proposed a Variable Step Defor-mation Difference Method (VSDDM) to analyze the non-linear blade structure According to [76] an approximateddeflection equation for moderate large deflection problemsdeveloped from the differential equation of large deflectioncantilever beams (3) using Newton binomial theorem Thismethod possesses the merits of distinct concept ease ofunderstanding rapid convergence speed and simplicity toprogram Analysis based on this method is carried out for200 kW wind turbine blade subjected to extreme wind Theresults revealed that VSDDMprovides an accurate predictionof the blade tip deflection and is effective to solve suchnonprismatic cantilever beams with variable stiffness and


large deflection and subjected to complicated loads

d2ydx2= M(x)

EI[1 + (119889119910119889119909)

2]32

(3)

Besides geometric nonlinearity effect of material nonlin-earity is worth considering [79] Nonlinear effect of largedeflection has a significant effect on power productionloading and also stability more comprehensive study of thesubject and including these nonlinear effects into aeroelasticcodes are expected

43 Aeroelastic Stability In wind turbine instability can bepitch-flap flutter stall induced instability rotor shaft whirlaeromechanical instability andor hydrodynamic interactionbrought on by the ocean currents and surface waves fromoffshore wind turbines Stability is one of the vita designsconstrained of wind turbine as Bir and Jonkman [80] pointedout that future would likely be stability-driven in contrast toloads-driven designs during that time

431 Edgewise Instability Though the shift from stall reg-ulation to pitch control will significantly avoid stall relatedinstability during operation due to the inherent low aerody-namic damping for edgewise model the edgewise instabilityis still a critical problem The experimental evidence ofedgewise instability has been seen in the mid nineties on stallregulated rotors with a diameter of 35ndash40m Hansen et al[7] illustrated the subject matter in detail and explained theearly efforts done typical examples on stability analysis withlinear stability tool HAWCStab are included to elaborate theedgewise instability of wind turbine

Lindenburg and Snel [81] pointed out the reason for edge-wise blade vibration instability as less structural dampingdue to application of carbon fibers more UD (unidirectional)layers vacuum production techniques and a smoother tran-sition from the airfoil-sections to the blade root relativelysmall chord and a decreasing slope of the torque-speedrelation of the generator at full-load

Part of EC Joule III project [82] with objective ofimproving the prediction capability with respect to dynamicloads in stall and stall induced vibration and establishingguidelines to achieve safety margin against stall inducedvibration were one of the early efforts done between 1995and 1998 In contrary to the violent effects of edgewise bladevibration Thomsen et al [83] formulated an experimentalmethod to determine the effective damping for the edgewiseblade mode shape for wind turbines Rasmussen et al [84]used dynamic stall model to analyze and reproduce open airblade section measurements as well as wind tunnel measure-ments The results from wind tunnel experiment revealedthat aerodynamic damping characteristics sensitivity to stallinduced vibrations depends highly on the relative motionof the airfoil in flapwise and edgewise direction and on apossibly coupled pitch variation which is determined by thestructural characteristics of the blade Chaviaropoulos [85]also used differential dynamic stall model and linearizedequation of motion to investigate the combined flaplead-lag

motion characteristic In extension [86] also analyzed andpointed out that thesemodels provided important knowledgeat the qualitative level but also significant uncertainty at thequantitative level

The European project VISCEL (2003 2004) consideredthe stability characteristic of the typical section using anunsteady Navier Stokes treatment of the aerodynamics [8187] another European project DAMPBLADE (2003) made astep to full section of a blade Subsequently several researcheson wind turbine aeroelastic instability had been conductedincluding STABCON [81 88] in which experimental dataare used to cross-validate different methods In later yearsseveral inventions have been recorded such as [89 90]which developed an active stall control method for dampingedgewise oscillations in one ormore blades of a wind turbineThis method works as first detecting if one or more of saidblades oscillates edgewise during operation of said windturbine and substantially cyclically generating a pitch angledifference between at least two of said blades

432 Pitch-Flap Flutter Instability It is a dynamic instabilitycaused by a positive feedback between the bodyrsquos deflectionand aerodynamic force Although this type of aeroelasticinstability is an infant in commercial wind turbines so farhowever as the size of the blades is increasing the flutterspeed decreases due to increasing structural flexibility of theblades and not least the torsional frequency decreases It isa smart way to include a flutter speed calculation in thedesign verification Flutter involves two DOF of the bladetorsion and translation The flutter speed decreases when thefrequency of these twoDOF approaches each otherThe otherdesign parameter for flutter instability is the center of mass inthe blade sections relative to the center of the elastic axis Asthe center of mass moves away from the elastic axis in thedirection of the trailing edge the flutter speed decreases [7]

In [91] the frequency domain techniques developed byTheodorsen adapted to investigate aeroelastic stability of aMW-size blade with andwithout aeroelastic tailoring Resultsindicate that the predicted flutter speed of a MW-sized bladeis slightly greater than twice the operational speed of therotor When a moderate amount of aeroelastic tailoring isadded to the blade a modest decrease (12) in the flutterspeed is observed

44 Smart Rotor and Control

441 Active Load Control Devices Due to complex inflowand turbulence and its dynamic characteristic wind turbineblades are exposed to fatigue loading Several load controlmethods can be employed to modify these aerodynamiccharacteristics of the blades and flow condition by extensionto the aerodynamic forces There are three major categoriesof active load control techniques (i) surface blowingsuction(ii) VGrsquos surface heating plasma and so forth or (iii)changes in section shape (aileron smart materials andmicrotabs) Figure 12

The early progress of the subject matter is reviewed thor-oughly in [92ndash94] Comparison among aerodynamic loadcontrol methods (ie deformable flap microtabs camber


Upper surface tab

Lower surface tab

(a)

098 1 102 104 106094xc

(b)

e

e

h

h

z

z

e

h

z

e

h

zFlow

Flow

Counterrotating Corotating

Vane-type VGs

Wheeler VGs

Wishbone Doublet(c)

Air jet VGAir jet

Vortex

(d)

Figure 12 Active flow control devices (a) Microtab [99] (b) flow pattern after application of Microtab [100] (c) vortex generators [101] and(d) air jet vortex generator [102]

control ormorphed trailing edge active twist boundary layersuctionblowing synthetic jets active vortex generator andplasma actuator) in terms of lift controllability is done by [94]and the result showed that trailing edge flaps camber controland microtabs have very good average and maximum liftcontrol capability Trailing edge flap control is demonstratedas the most efficient control method The change in lift anddrag characteristics as well as the linearity the bandwidthand the simplicity of these concepts makes it attractive fromthe control point of view The other methods have also someunique merits microtabs simplicity bandwidth and smallactuating power needed make it attractive except that itson-off characteristic makes them less efficient for detailedload control further investigation is needed for advance usesActive twist control is rotating the whole span of the bladeabout the blade axis This method in general is feasible butit is expensive results in heavier rotor and consumes morepower whichwillmake it inefficientmethod to reduce fatigueloading

Two researches at SandiaNational Laboratories [95] usingMicrotab concept reported 20ndash32 reduction blade rootflap bending moments and [96] for another procedure that

is increasing the blade and other components size for thesame blade root flap fatigue damage as the baseline rotorby enrolling morphed trailing edge reported 11 incrementin energy capture A smart rotor configuration employinglinear quadratic to control adaptive trialing edge flap wasproposed by [97] and its performancewas evaluated based onaeroelastic simulation of a baseline NREL5MWwind turbinewith the flaps extending along 20 of span using HAWC2code Control algorithm includes frequency weighting todiscourage flap activity at frequencies higher than 05Hz andalso uses periodic disturbance signals described by simplefunctions of the blade azimuthal position to determine periodcomponent of the load

The effects of the adaptive trailing edge flap control arequantified in terms of lifetime fatigue damage equivalentload reduction and it is recorded 10 blade root flapwisemoment reduction including the periodic load anticipationwill improve the result as 138 with the d Sin-Cos con-figuration and 45 with 119889 Wsp Figure 13 Zhang et al[98] also investigate the impact of smart load control usingtrialing edge flap on NREL 5MW and the results showedsignificant reduction onflapwise blade root bendingmoment


Ref no flap d 00

d sinminuscos d Wsp

7000

7500

8000

8500

9000

9500

10000

DEL

Mx

BlR

t (kN

m) (

=10

)m

14 16 18 20 22 2412Mean Wsp (ms)

Figure 13 Fatigue damage equivalent loads D L at the blade rootflapwise bending moment The DEL refers to a 25-year lifetime and10 million equivalent cycles [97]

Furthermore the smart load control altered the nature ofthe flow-blade interactions and changed the in-phased fluid-structure synchronization into much weaker couplings as aresult of fluid-structure damping enhanced

442 Smart Material Actuators Smart materials are mate-rials which possess the capability to sense and actuate ina controlled way in response to variable ambient stimuliActuators for smart load control comprise a vital role In ageneral sense there are two classes of actuators as embeddedand discrete The conventional load control actuators (iehydraulic pneumatic and electrical actuators) are mostlyused in existing wind turbine blade pitch and yaw controlapplications However their inherent demerits includingleakage problems and contamination delay in actuationregular maintenance requirement reduced frequency rangeand exhibiting certain instability weight space and powerrequirement limit them from active smart load controlapplication

The common criteria for active control include lessweight contribution achieving the required deflection beingdynamically responsive at the frequency range of interest lin-ear actuation behavior high resistance to fatigue loads insen-sitivity to oxidation and lightning strikes and limited degra-dation or reduced performance Smart material includesferroelectric materials (piezoelectric electrostrictive andmagnetostrictive) variable rheology materials (electrorhe-ological magnetorheological) and shape memory alloysThough these materials are not yet commercialized severalresearches indicated their feasibility thorough discussionand comparison are presented by [94]

5 Concluding Remarks

The present article reviewed the science of wind turbineaeroelasticity and its trend through time Consideringmarketcompetitiveness and related constraints the design trends aredriven towards multimegawatt large and flexible turbine

utilization of smart rotor control devices more geometricand material nonlinear structure and offshore and complexterrain applications On the contrary this will alter theaeroelastic characteristic and raise numerous system stabilityissues which will demand detailed methods to model andsimulate the system for further optimal outputs In the previ-ous few sections several remedies done have been discussedand the gaps to be addressed can be categorized into the needfor comprehensive aeroelastic tools coupled or hybrid solverand multidisciplinary optimizations

(1) Comprehensive Aeroelastic Tool As the complexity of thewind turbine system is enhancing and number of couplingsystems is increasing the requirement of comprehensiveaeroelastic tool to handle realistic model of the system ismandatory Such requirements include

(i) complex inflow including wind shear(ii) hydrodynamic effects in offshore application(iii) nonlinearity due to large deflection geometric and

material distribution and manufacturing methods(iv) application of smart rotor and control methods and

their coupling to the system

(2) CoupledHybrid Solver Computational efficiency andhigh fidelity output are the two main compromises in com-putational studies As single model only allows achievingeither of the two and due to inherent limitations they possessapplication of hybrid model is canny In fluid flow studyhybrid LESRANS model is a common approach as RANSwill be applied near the wall and LES to the far fielddomain of the flow Its application in wind turbine aeroelasticmodeling will advance the accuracy because of LES andreduce computation effort because of RANS Similarly instructural modeling hybrid model can be applied such asFEM and modal reduction approach

(3) Multidisciplinary OptimizationMost of aeroelastic codesin wind turbine industry are used as a standalone design tooland their application in multidisciplinary optimization ofwind turbine system is not common and at infant stage Mul-tidisciplinary wind turbine system optimization frameworkwill identify the possible aerodynamic structural controland other subsystem configurations to produce minimumcost of energy Such integration will avoid common subop-timal design trend and enhance the competitiveness of windenergy conversion

Conflicts of Interest

The authors declare that there is no conflict of interests

Acknowledgments

This work was financially supported by the China Govern-ment Scholarship ProgramThe first author also would like toacknowledge Wuhan University of Technology for providingholistic assistance in the course of the study


References

[1] International Energy Agency-IEA Long Term Research andDevelopment Needs for Wind Energy for the Time Frame 2000to 2020 International Energy Agency-IEA 2001

[2] IEA Long-Term Research and Development Needs for WindEnergy for the Time Frame 2012 to 2030 International EnergyAgency-IEA 2013

[3] EWEA UpwindmdashDesign Limits and Solutions for Very LargeWind Turbines EWEA (European Wind Energy Association)Brussels Belgium 2011

[4] P P Friedmann ldquoAeroelastic modeling of large wind turbinesrdquoJournal of the American Helicopter Society vol 21 no 4 pp 17ndash27 1976

[5] J GMarshall andM Imregun ldquoA review of aeroelasticitymeth-ods with emphasis on turbomachinery applicationsrdquo Journal ofFluids and Structures vol 10 no 3 pp 237ndash267 1996

[6] A D Garrad ldquoDynamics of wind turbinesrdquo IEE Proceedings APhysical Science Measurement and Instrumentation Manage-ment and Education Reviews vol 130 no 9 pp 523ndash530 1983

[7] M O L Hansen J N Soslashrensen S Voutsinas N Soslashrensen andH A Madsen ldquoState of the art in wind turbine aerodynamicsand aeroelasticityrdquo Progress in Aerospace Sciences vol 42 no 4pp 285ndash330 2006

[8] P Zhang and S Huang ldquoReview of aeroelasticity for windturbine current status research focus and future perspectivesrdquoFrontiers in Energy vol 5 no 4 pp 419ndash434 2011

[9] H Glauert ldquoAirplane propellersrdquo in Aerodynamic Theory W FDurand Ed Dover Publications New York NY USA 1963

[10] J F Manwell J G McGowan and A L Rogers Wind EnergyExplained Theory Design and Application John Wiley amp Sons2nd edition 2009

[11] H Snel and J G Schepers ldquoEngineering moles for dynamicinflowphenomenardquo inProceedings of the EuropeanWindEnergyConference Amsterdam The Netherlands October 1991

[12] S Oslashye ldquoTjaeligreborg wind turbine (Esbjerg) first dynamicinflow measurementrdquo AFM Notat no VK-189 AFM LyngbyDenmark 1991 httporbitdtudkfiles3524739VK189pdf

[13] S Oslashye ldquoTjaeligreborg wind turbine 4 Dynamic inflow measure-mentrdquo AFM Notat no VK-204 AFM Lyngby Denmark 1991httporbitdtudkfiles3524469VK204pdf

[14] H Snel and J G Schepers ldquoJOULE1 joint investigation ofdynamic inflow effects and implementation of an engineeringmethodrdquo Tech Rep ECN-C-94-107 1994

[15] J G Schepers H Snel and G J W Bussel ldquoJOULE2 DynamicInflow Yawed Conditions and Partial Span Pitchrdquo 1995

[16] J G Schepers and H Snel Final Results of the EU Joule ProjectsldquoDynamic Inflowrdquo Netherlands Energy Research FoundationECN Petten The Netherlands 1996

[17] T Knudsen and T Bak ldquoSimple model for describing andestimating wind turbine dynamic inflowrdquo in Proceedings of the1st American Control Conference (ACC rsquo13) June 2013

[18] S Oslashye ldquoDynamic stall-simulated as time lag of separationrdquo inProceedings of the 4th IEA Symposium on the Aerodynamics ofWind Turbines Rome Italy 1991

[19] J W Larsen S R K Nielsen and S Krenk ldquoDynamic stallmodel forwind turbine airfoilsrdquo Journal of Fluids and Structuresvol 23 no 7 pp 959ndash982 2007

[20] J G Leishman and T S Bcddoes ldquoA semi-empirical model fordynamic stallrdquo Journal of the American Helicopter Society vol34 no 3 pp 3ndash17 1989

[21] A Hani Active Control of Dynamic Stall University of Califor-nia Los Angeles Calif USA 1998

[22] A ChoudhryM Arjomandi and R Kelso ldquoMethods to controldynamic stall for wind turbine applicationsrdquo Renewable Energyvol 86 pp 26ndash37 2016

[23] DHHodges and EHDowell ldquoNonlinear equations ofmotionfor the elastic bending and torsion of twisted nonuniform rotorbladesrdquo Tech Rep NASA Washington DC USA 1974

[24] B S Kallesoslashe ldquoEquations of motion for a rotor blade includinggravity pitch action and rotor speed variationsrdquo Wind Energyvol 10 no 3 pp 209ndash230 2007

[25] Y Li A M Castro T Sinokrot W Prescott and P M CarricaldquoCoupledmulti-body dynamics and CFD for wind turbine sim-ulation including explicit wind turbulencerdquo Renewable Energyvol 76 pp 338ndash361 2015

[26] J Kumar and F-HWurm ldquoBi-directional fluid-structure inter-action for large deformation of layered composite propellerbladesrdquo Journal of Fluids and Structures vol 57 pp 32ndash48 2015

[27] F Debrabandere B Tartinville C Hirsch and G CoussementldquoFluid-structure interaction using a modal approachrdquo Journalof Turbomachinery vol 134 no 5 Article ID 051043 6 pages2012

[28] F-K Benra H J Dohmen J Pei S Schuster and B WanldquoA comparison of one-way and two-way coupling methods fornumerical analysis of fluid-structure interactionsrdquo Journal ofAppliedMathematics vol 2011 Article ID 853560 16 pages 2011

[29] Y-B Chen Z-K Wang and G-C Tsai ldquoTwo-way fluid-stru-cture interaction simulation of a micro horizontal axis windturbinerdquo International Journal of Engineering and TechnologyInnovation vol 5 no 1 pp 33ndash44 2015

[30] J Jonkman S Butterfield W Musial and G Scott ldquoDefinitionof a 5-Mw reference wind turbine for offshore system devel-opmentrdquo Tech Rep NRELTP-500-38060 National RenewableEnergy Laboratory-NREL 2009

[31] H J T Kooijman C Lindenburg D Winkelaar and E L vanderHooftDOWEC6MWPre-Design Aero-ElasticModelling ofthe DOWEC 6MWPre-Design in PHATAS ECNWind EnergyPetten The Netherlands 2003

[32] A AhlstromAeroelastic Simulation ofWind Turbine DynamicsKTH-Royal Institute of Technology Stockholm Sweden 2005

[33] T Buhl Research in Aeroelasticity EFP-2007-II Risoslash NationalLaboratory Roskilde Denmark 2009

[34] C Lindenburg Comparison of Phatas Versions and the WindTurbine Module Energy Research Center of the NetherlandsECN 2011

[35] J G Schepers J Heijdra D Foussekis et al VerIficationof European Wind Turbine Design Codes VEWTDC EnergyResearch Center of the Netherlands ECN 2002

[36] J G Schepers J Heijdra K Thomsen et al ldquoVerification ofEuropean wind turbine design codesrdquo in Proceedings of theEuropean Wind Energy Conference (EWEC rsquo01) CopenhagenDenmark 2001

[37] J M Jonkman and M L Buhl Jr FAST Userrsquos Guide NRELGolden Colo USA 2005

[38] T J Larsen and A M Hansen How 2 HAWC2 The UserrsquosManual Risoslash National Laboratory Roskilde Denmark 2007

[39] P Passon andMKuhn State-of-the-Art andDevelopment Needsof Simulation Codes for Offshore Wind Turbines CopenhagenOffshore Wind 2005

[40] F S J Peeringa Aero-Elastic Simulation of Offshore Wind Tur-bines in the Frequency Domain TURBUSea Energy ResearchCenter of the Netherlands ECN 2009


[41] T G Van Engelen ldquoControl design based on aero-hydro-servo-elastic linearmodels fromTURBU (ECN)rdquo in Proceedings of theEuropean Wind Energy Conference and Exhibition (EWEC rsquo07)pp 114ndash140 Milan Italy May 2007

[42] P T D M Lano 2015 httpwwwaeropolimiitsimbottassoPOLI-Windhtm

[43] R Rubak and J T Petersen ldquoMonopile as part of aeroelasticwind turbine simulation coderdquo in Proceedings of the ConferenceCopenhagen Offshore Wind Copenhagen Denmark October2005

[44] N N Soslashrensen and J Johansen ldquoUpWind aerodynamics andaero-elasticity rotor aerodynamics in atmospheric shear flowrdquoin Proceedings of the European Wind Energy Conference ampExhibition Milan Italy May 2007

[45] R Flemming ldquoUpWind aerodynamics and aeroelasticsrdquo inProceedings of the European Wind Energy Conference amp Exhi-bition Brussels Belgium 2007

[46] F Zahle H Aagaard Madsen and N Soslashrensen Evaluationof Tower Shadow Effects on Various Wind Turbine ConceptsDanmarks Tekniske Universitet Risoslash Nationallaboratoriet forBaeligredygtig Energi Roskilde Denmark 2009

[47] A Gomez and J R Seume ldquoLoad pulses on wind turbinestructures caused by tower interferencerdquoWind Engineering vol33 no 6 pp 555ndash570 2009

[48] S-Y Lin and T-H Shieh ldquoStudy of aerodynamical interferencefor a wind turbinerdquo International Communications in Heat andMass Transfer vol 37 no 8 pp 1044ndash1047 2010

[49] P Zhang S Huang T Yang and J Li ldquo3D potential flowmodelof tower interference for upwind wind turbinerdquo Advances inMechanical Engineering vol 2014 Article ID 612453 10 pages2014

[50] F Zahle N N Soslashrensen and H A Madsen The Influence ofWind Shear and Tower Presence on Rotor andWake Aerodynam-ics Using CFD Risoslash National Laboratory Roskilde Denmark2008

[51] P Zhang S Huang T Yang and J Li ldquoResearch on theaeroelastic response of tower effects for great grade windturbinerdquo Journal of Applied Sciences vol 13 no 15 pp 3042ndash3048 2013

[52] S T Frandsen ldquoTurbulence and turbulence-generated struc-tural loading in wind turbine clustersrdquo Risoe-R no 1188(EN)Forskningscenter Risoe Roskilde Denmark 2007

[53] ldquoIEC 6400-1-International Standard Wind turbinesmdashPart 1Design requirementsrdquo International Electrotechnical Commis-sion (IEC) 2005

[54] K Thomsen H A Madsen G C Larsen and T J LarsenldquoComparison of methods for load simulation for wind turbinesoperating in wakerdquo Journal of Physics Conference Series vol 75no 75 2007

[55] N Jensen ANote onWind Generator Interaction Risoslash NationalLaboratory Roskilde Denmark 1983

[56] I Katıc J Hoslashjstrup and N O Jensen ldquoA simple model forcluster efficiencyrdquo in Proceedings of the European Wind EnergyAssociation Conference and Exhibition (EWEC rsquo86) Rome Italy1987

[57] J F Ainslie ldquoCalculating the flowfield in the wake of windturbinesrdquo Journal of Wind Engineering and Industrial Aerody-namics vol 27 no 1ndash3 pp 213ndash224 1988

[58] S Ott J Berg and M Nielsen ldquoLinearised CFD Models forWakes Risoslash-R-1772(EN)rdquo Risoslash National Laboratory RoskildeDenmark 2011

[59] S Ott M Nielsen and K S Hansen ldquoFugamdashvalidating a wakemodel for offshore wind farmsrdquo in Proceedings of the EuropeanWind Energy Association Meeting Dublin Ireland 2013

[60] G C Larsen H A Madsen F Bingol et al DynamicWake Meandering Modeling Risoslash National Laboratory-Risoslash-R-1607(EN) Roskilde Denmark 2007

[61] T J Larsen H A Madsen G C Larsen and K S HansenldquoValidation of the dynamic wake meander model for loads andpower production in the Egmond aan Zee wind farmrdquo WindEnergy vol 16 no 4 pp 605ndash624 2012

[62] L J Vermeer J N Soslashrensen and A Crespo ldquoWind turbinewake aerodynamicsrdquo Progress in Aerospace Sciences vol 39 no6-7 pp 467ndash510 2003

[63] M Bastankhah and F Porte-Agel ldquoA new analytical model forwind-turbine wakesrdquo Renewable Energy vol 70 pp 116ndash1232014

[64] M Gaumond P-E Rethore A Bechmann et al Benchmarkingof Wind Turbine Wake Models in Large Offshore Wind FarmsThe Science of Making Torque from Wind Oldenburg Ger-many 2012

[65] B Schmidt U Smolka S Hartmann and PW Cheng ldquoValida-tion of the dynamic wake meander model with AREVAM5000loadmeasurements at alpha ventusrdquo inProceedings of the EWEAOffshore Frankfurt Germany November 2013

[66] T J Larsen G Larsen H A Madsen and K ThomsenComparison of Design Methods for Turbines in Wake EWECBrussels Belgium 2008

[67] J Annoni P Seiler K Johnson P Fleming and P GebraadldquoEvaluating wake models for wind farm controlrdquo in Proceedingsof the American Control Conference (ACC rsquo14) pp 2517ndash2523IEEE Portland Ore USA June 2014

[68] M J Churchfield ldquoA review of wind turbine wake models andfuture directionsrdquo in Proceedings of the North American WindEnergy Academy Symposium (NAWEA rsquo13) Boulder Colo USA2013

[69] B S Kallesoslashe and M H Hansen ldquoSome effects of large bladedeflections on aeroelastic stabilityrdquo in Proceedings of the 47thAIAA Aerospace Sciences Meeting Including the New HorizonsForum and Aerospace Exposition Orlando Fla USA January2009

[70] B S Kallesoslashe ldquoLarge blade deformations effect on flutterboundariesrdquo in Research in Aeroelasticity EFP-2006 pp 83ndash89Risoslash National Laboratory Roskilde Denmark 2007

[71] M H Hansen and B S Kallesoslashe ldquoSome nonlinear effects onthe flutter speed and blade stabilityrdquo in Research in Aeroelas-ticity EFP-2007 vol Risoslash-R-1649(EN) pp 93ndash105 DanmarksTekniske Universitet Risoslash Nationallaboratoriet for BaeligredygtigEnergi Roskilde Denmark 2008

[72] T J Larsen A M Hansen and T Buhl Aeroelastic Effectsof Large Blade Deflections for Wind Turbines The Science ofMaking Torque fromWind Copenhagen Denmark 2004

[73] MDado and S Al-Sadder ldquoA new technique for large deflectionanalysis of non-prismatic cantilever beamsrdquoMechanics ResearchCommunications vol 32 no 6 pp 692ndash703 2005

[74] JW Larsen and S R K Nielsen ldquoNon-linear dynamics of windturbine wingsrdquo International Journal of Non-Linear Mechanicsvol 41 no 5 pp 629ndash643 2006

[75] A Banerjee B Bhattacharya and A K Mallik ldquoLarge deflec-tion of cantilever beams with geometric non-linearity analyt-ical and numerical approachesrdquo International Journal of Non-Linear Mechanics vol 43 no 5 pp 366ndash376 2008


[76] G Yuan and Y Chen ldquoGeometrical nonlinearity analysisof wind turbine blade subjected to extreme wind loadsrdquo inProceedings of the International Symposium on ComputationalStructural Engineering Shanghai China June 2009

[77] L Chen ldquoAn integral approach for large deflection cantileverbeamsrdquo International Journal of Non-Linear Mechanics vol 45no 3 pp 301ndash305 2010

[78] B S Kallesoslashe J J Thomsen and M H Hansen Aeroservoe-lasticity of Wind Turbines Technical University of DenmarkCopenhagen Denmark 2007

[79] C Athisakul B Phungpaingam G Juntarakong and S Chuc-heepsakul ldquoEffect of material nonlinearity on large deflectionof variable-arc-length beams subjected to uniform self-weightrdquoMathematical Problems in Engineering vol 2012 Article ID345461 9 pages 2012

[80] G Bir and J Jonkman ldquoAeroelastic instabilities of large offshoreand onshore wind turbinesrdquo Journal of Physics ConferenceSeries vol 75 Article ID 012069 2007

[81] C Lindenburg and H Snel ldquoAero-elastic stability analysis toolsfor large wind turbine rotor bladesrdquo in Proceedings of theEuropean Wind Energy Conference Madrid Spain 2003

[82] JThirstrup Petersen H AagaardMadsen A Bjorck et al ldquoPre-diction of dynamic loads and induced vibrations in stallrdquo Risoe-R no 1045(EN) Forskningscenter Risoe Roskilde Denmark1998

[83] K Thomsen J T Petersen E Nim S Oslashye and B PetersenldquoA method for determination of damping for edgewise bladevibrationsrdquoWind Energy vol 3 no 4 pp 233ndash246 2000

[84] F Rasmussen J T Petersen and H A Madsen ldquoDynamic stalland aerodynamic dampingrdquo in Proceedings of the AIAAASMEWind Energy Symposium pp 44ndash51 January 1998

[85] P KChaviaropoulos ldquoFlaplead-lag aeroelastic stability ofwindturbine bladesrdquoWind Energy vol 4 no 4 pp 183ndash200 2001

[86] V A Riziotis S G Voutsinas E S Politis and P KChaviaropoulos ldquoAeroelastic stability of wind turbines theproblem the methods and the issuesrdquo Wind Energy vol 7 no4 pp 373ndash392 2004

[87] P K Chaviaropoulos N N Soerensen M O L Hansen et alldquoViscous and aeroelastic effects on wind turbine blades TheVISCEL project Part II aeroelastic stability investigationsrdquoWind Energy vol 6 no 4 pp 387ndash403 2003

[88] M H Hansen ldquoAeroelastic stability analysis of wind turbinesusing an eigenvalue approachrdquo Wind Energy vol 7 no 2 pp133ndash143 2004

[89] T S B Nielsen and C J Spruce ldquoWind Turbine A Method ForDamping Edgewise Oscillations In One Or More Blades Of AWind Turbine By Changing The Blade Pitch And Use HereofrdquoPatent US20090185901 A1 23 July 2009

[90] T S B Nielsen B J Pedersen and C J Spruce ldquoMethod fordamping edgewise oscillations in one or more blades of a windturbine an active stall controlled wind turbine and use hereofrdquoPatent US8070437 B2 6 December 2011

[91] D W Lobitz ldquoAeroelastic stability predictions for a MW-sizedbladerdquoWind Energy vol 7 no 3 pp 211ndash224 2004

[92] C P van Dam D E Berg and S J Johnson ldquoActive loadcontrol techniques for wind turbinesrdquo Tech Rep SAND2008-4809 TRN US200902565 Sandia National LaboratoriesAlbuquerque NM USA 2008

[93] T K Barlas and G A M van Kuik ldquoState of the art andprospectives of smart rotor control for wind turbinesrdquo Journalof Physics Conference Series vol 75 Article ID 012080 2007

[94] T K Barlas and G A M van Kuik ldquoReview of state of the artin smart rotor control research for wind turbinesrdquo Progress inAerospace Sciences vol 46 no 1 pp 1ndash27 2010

[95] D G Wilson D E Berg M F Barone J C Berg B R Resorand D W Lobitz ldquoActive aerodynamic blade control design forload reduction on large wind turbinesrdquo in Proceedings of the inEuropean Wind Energy Conference amp Exhibition Parc ChanotFrance March 2009

[96] D E Berg D G Wilson M F Barone et al ldquoThe impact ofactive aerodynamic load control on fatigue and energy captureat low wind speed sitesrdquo in Proceedings of the European WindEnergy Conference and Exhibition (EWEC rsquo09) pp 2670ndash2679Marseille France March 2009

[97] L Bergami andNK Poulsen ldquoA smart rotor configurationwithlinear quadratic control of adaptive trailing edge flaps for activeload alleviationrdquoWind Energy vol 18 no 4 pp 625ndash641 2015

[98] M ZhangW Yu and J Xu ldquoAerodynamic physics of smart loadcontrol for wind turbine due to extreme wind shearrdquo RenewableEnergy vol 70 pp 204ndash210 2014

[99] K-C Tsai C-T Pan A M Cooperman S J Johnson and CP van Dam ldquoAn innovative design of a microtab deploymentmechanism for active aerodynamic load controlrdquo Energies vol8 no 6 pp 5885ndash5897 2015

[100] C P Van Dam R Chow J R Zayas and D E Berg ldquoCom-putational investigations of small deploying tabs and flaps foraerodynamic load controlrdquo Journal of Physics Conference Seriesvol 75 2007

[101] J C Lin ldquoReview of research on low-profile vortex generatorsto control boundary-layerrdquo Progress in Aerospace Sciences vol38 no 4-5 pp 389ndash420 2012

[102] S Shun and N A Ahmed ldquoWind turbine performanceimprovements using active flow control techniquesrdquo ProcediaEngineering vol 49 pp 83ndash91 2012

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Active and Passive Electronic Components

Control Scienceand Engineering

Journal of



RotatingMachinery


Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpswwwhindawicom

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics


Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of


Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation



DistributedSensor Networks



Airbus A380

05 3 5 13 16 2 5 75 1st year of operation

810 rated capacity (MW)45

wing span

Roto

r dia

met

er (m

)

250m Oslash

160m Oslash

126m Oslash126m Oslash

112m Oslash

15m Oslash

100503019997959391898785

80 m

Figure 1 Wind turbine upscaling trend [3]

Aerodynamicforce

Rigid b

ody

aerod

ynam

ics Static

aeroelasticityAeroelasticity

Inertial force Elastic forceMechanicalvibration

Figure 2 Collarrsquos triangle interaction of different forces

wind loads and structural part to calculate the dynamicresponse of structure with time history and spatial distribu-tion of wind as input Figure 3

In the following sections both aerodynamic and struc-tural dynamic component of wind turbine aeroelasticity willbe elaborated including methods to model the correspond-ing phenomena

21 Aerodynamic Aerodynamic part of aeroelastic analysisentitled to determine the aerodynamic loads developeddue to the flow pattern of wind against the wind turbineblade orientation Different models and methods have beendeveloped such as the very common and popular Blade

Element Momentum method lifting line panel and vortexmodels generalized actuator disc models and Navier Stokesbased solvers each theory possesses pros and cons In thissection a brief discussionwill be included about the commonaerodynamic methods

211 Blade Element Momentum (BEM) Method BEM alsocalled strip theory was originally introduced by Glauert [9] itis computationally fast and cheap and provided that reliableaerofoil data is available it will give satisfactory results It isa combination of the simple momentum theory and bladeelement theory [10] and assumes that there is no aerodynamicinteraction between all sections along the rotor and can betreated separately which imply there is no radial flow Referto Figures 4 and 5

The forces on the blades are determined solely by the liftand drag characteristics of the airfoil shape of the blades andit is assumed the flow is incompressible steady state Sub-sequently combining the two expressions for thrustnormalforces and torque frommomentum theory and blade elementtheory the BEM will be derived After some algebraicmanipulation the resulting relations will be ((1)-(2))

119862119897 = 4119865 sin120593 (cos120593 minus 120582119903 sin120593)1205901015840 (sin120593 + 120582119903 cos120593) 1198861015840(1 + 1198861015840) =

12059010158401198621198974119865 cos120593(1)




BEM

FEM

dFLdFD

dFL

dFL

dFL

dFL

dFD

dFD

dFD

dFD

+

dFL = Cl1

2U2

ＬＦcdr

dFD = Cd1

2U2

ＬＦcdr

InflowZ

UＬＦ

UＬＦ

UＬＦ

UＬＦ

UＬＦ

U (z) =Ulowast

kＦＨ ( z

z0)


dr

c

R

r

Ω




(2)

Flow direction


Streamtube boundary

U

A1

A2

U(1 minus a)

U(1 minus 2a)





100

150

200

250

300

350

400

Roto

r sha

ft to

rque

(kN

m)

50 6030 4010 200Time (s)





















Y XZ





















Table2Com

paris

onof

aeroela

sticcodesfor

windturbine

Nam

eofthe

code

Started

Develo

per

Aerodynamic

mod

elStructuralmod

elDescriptio

n

GHBladed

1996

GarradHassanand

PartnersLtd

BEM

theory

Mod

alapproach


edforo

ffsho

reapplicationwith

mon

opile

orgravity

-based

foun

datio

nsTh

elatev

ersio

niscapableo

fmod

elingandanalyzing

both

onshorea

ndoff

shorew

indturbinew

ithvarie

tyof

supp

ortspecification

HAW

C22003ndash200

6

Risoslash

National

Labo

ratory

TechnicalU

niversity

ofDenmark

BEM

theory

Multib

ody

dynamics

Itisas

uccessor

forH

AWEC

tool(w

hich

utilizesF

Emetho

dusingsubstructure

approach

with

Timoshenk

o-beam


odeling

)HAW

EC2isa

timed

omain

morec

omprehending

toolTh

eaerod

ynam

icmod

elismod

ified

tohand

ledynamicinflo

wdynamicstallskew

inflo

wsheare

ffectso

ntheind

uctio

nandeffectsfro

mlarged

eflectio

n

ADAMSWT

mdashMechanical

Dyn

amicsIncun

der

contractof

NRE

LBE

Mtheory

Multib

ody

dynamics

ADAMSWTisreplaced

with

FAST


plug-in

AdWiM

oandAe

roDyn

with

A2A

Dinterfa

cecanalso

beintegrated

with

multib

odysoftw

are

ADAMSsolver

tomod

elandsim

ulate

Alcyone

mdashCenterfor

Renewable

Energy

Source

ampNTU

ABE

Mtheory

FEM

NationalTechn

icalUniversity

ofAthens

developedAlcyone(

freew

ake)with

free

wake

panelm

ethodwhich

also

inclu

desa

simulator

ofturbulentw

indfieldstim

edo

mainaeroela

sticanalysisof

thefullw

indturbinec

onfig

uration

and

postp

rocessingof

loadsfor

fatig

ueanalysis

TURB

U2007

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamics

Itisafrequ

ency

domain

lineariz

edaerohydroservoela

sticcode

andthea

ctive

aeroela

sticcontrolcod

e

DUWEC

D1986

TUDelft

BEM

theory

Multib

ody

dynamics

In1993

itwas

mod

ified

tomod

eloff

shorew

indturbinea

ndlatertoinclu

dewave

loads

FAST

mdashOregonState

University

under

contractof

NRE

LBE

Mtheory

Mod

alapproach

Thistoolhasm

uchles

srun

timeIn

1996N

RELhasm

odified

FAST

tousethe

AeroDyn

subrou

tinep

ackage

developedattheU

niversity

ofUtahto

calculatethe

aerodynamicforces

alon

gtheb

lade

FLEX

5mdash

DTU

BEM

theory

Mod

alapproach

Itistim

edom

ainaeroela

sticsim

ulationtoolanduses

relatively

fewer

degree

offre

edom

tomod

eltheturbine

FLEX

LAST

1982

StorkProd

uct

Engineering

BEM

theory

Multib

ody

dynamic

Since1992thiscode

hasb

eenused

asdesig

nandcertificatio

ntool

PHAT

AS

1993

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamic

Todeterm

inethe

nonlineard

ynam

icbehavior

andthec

orrespon

ding

loadso

faho

rizon

tal-a

xiswindturbine(bo

thon

shorea

ndoff

shore)in

timed

omain

TWISTE

R1983

StentecB

VTh

eNetherla

nds

BEM

theory

FEM

Initiallyitisused

tobe

calledFK

Asince

1991itsup

portsscholastic

windfield

simulation

VID

YN1983

Tekn

ikgrup

penAB

Sollentun

aSw

eden

BEM

theory

Mod

alapproach

Itissta

rted

aspartof

thee


twolargeSw

edish

prototypes

Maglarp

andNassuden


(a) (b)


(a) (b)







0

2

4

6

8

10

12

14

Tota

l tip

defl

ectio

n (m

)

05 10 15 20 2500

Time (s)

(a)


0

1

2

3

4

5

6

Max

imum

stre

ss (P

a)

05 10 15 20 2500

Time (s)

times108

(b)










d2ydx2= M(x)

EI[1 + (119889119910119889119909)

2]32

(3)














Upper surface tab

Lower surface tab

(a)

098 1 102 104 106094xc

(b)

e

e

h

h

z

z

e

h

z

e

h

zFlow

Flow


Vane-type VGs

Wheeler VGs

Wishbone Doublet(c)

Air jet VGAir jet

Vortex

(d)







Ref no flap d 00

d sinminuscos d Wsp

7000

7500

8000

8500

9000

9500

10000

DEL

Mx

BlR

t (kN

m) (

=10

)m

14 16 18 20 22 2412Mean Wsp (ms)
















Acknowledgments



References











































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation












BEM

FEM

dFLdFD

dFL

dFL

dFL

dFL

dFD

dFD

dFD

dFD

+

dFL = Cl1

2U2

ＬＦcdr

dFD = Cd1

2U2

ＬＦcdr

InflowZ

UＬＦ

UＬＦ

UＬＦ

UＬＦ

UＬＦ

U (z) =Ulowast

kＦＨ ( z

z0)


dr

c

R

r

Ω




(2)

Flow direction


Streamtube boundary

U

A1

A2

U(1 minus a)

U(1 minus 2a)





100

150

200

250

300

350

400

Roto

r sha

ft to

rque

(kN

m)

50 6030 4010 200Time (s)





















Y XZ





















Table2Com

paris

onof

aeroela

sticcodesfor

windturbine

Nam

eofthe

code

Started

Develo

per

Aerodynamic

mod

elStructuralmod

elDescriptio

n

GHBladed

1996

GarradHassanand

PartnersLtd

BEM

theory

Mod

alapproach


edforo

ffsho

reapplicationwith

mon

opile

orgravity

-based

foun

datio

nsTh

elatev

ersio

niscapableo

fmod

elingandanalyzing

both

onshorea

ndoff

shorew

indturbinew

ithvarie

tyof

supp

ortspecification

HAW

C22003ndash200

6

Risoslash

National

Labo

ratory

TechnicalU

niversity

ofDenmark

BEM

theory

Multib

ody

dynamics

Itisas

uccessor

forH

AWEC

tool(w

hich

utilizesF

Emetho

dusingsubstructure

approach

with

Timoshenk

o-beam


odeling

)HAW

EC2isa

timed

omain

morec

omprehending

toolTh

eaerod

ynam

icmod

elismod

ified

tohand

ledynamicinflo

wdynamicstallskew

inflo

wsheare

ffectso

ntheind

uctio

nandeffectsfro

mlarged

eflectio

n

ADAMSWT

mdashMechanical

Dyn

amicsIncun

der

contractof

NRE

LBE

Mtheory

Multib

ody

dynamics

ADAMSWTisreplaced

with

FAST


plug-in

AdWiM

oandAe

roDyn

with

A2A

Dinterfa

cecanalso

beintegrated

with

multib

odysoftw

are

ADAMSsolver

tomod

elandsim

ulate

Alcyone

mdashCenterfor

Renewable

Energy

Source

ampNTU

ABE

Mtheory

FEM

NationalTechn

icalUniversity

ofAthens

developedAlcyone(

freew

ake)with

free

wake

panelm

ethodwhich

also

inclu

desa

simulator

ofturbulentw

indfieldstim

edo

mainaeroela

sticanalysisof

thefullw

indturbinec

onfig

uration

and

postp

rocessingof

loadsfor

fatig

ueanalysis

TURB

U2007

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamics

Itisafrequ

ency

domain

lineariz

edaerohydroservoela

sticcode

andthea

ctive

aeroela

sticcontrolcod

e

DUWEC

D1986

TUDelft

BEM

theory

Multib

ody

dynamics

In1993

itwas

mod

ified

tomod

eloff

shorew

indturbinea

ndlatertoinclu

dewave

loads

FAST

mdashOregonState

University

under

contractof

NRE

LBE

Mtheory

Mod

alapproach

Thistoolhasm

uchles

srun

timeIn

1996N

RELhasm

odified

FAST

tousethe

AeroDyn

subrou

tinep

ackage

developedattheU

niversity

ofUtahto

calculatethe

aerodynamicforces

alon

gtheb

lade

FLEX

5mdash

DTU

BEM

theory

Mod

alapproach

Itistim

edom

ainaeroela

sticsim

ulationtoolanduses

relatively

fewer

degree

offre

edom

tomod

eltheturbine

FLEX

LAST

1982

StorkProd

uct

Engineering

BEM

theory

Multib

ody

dynamic

Since1992thiscode

hasb

eenused

asdesig

nandcertificatio

ntool

PHAT

AS

1993

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamic

Todeterm

inethe

nonlineard

ynam

icbehavior

andthec

orrespon

ding

loadso

faho

rizon

tal-a

xiswindturbine(bo

thon

shorea

ndoff

shore)in

timed

omain

TWISTE

R1983

StentecB

VTh

eNetherla

nds

BEM

theory

FEM

Initiallyitisused

tobe

calledFK

Asince

1991itsup

portsscholastic

windfield

simulation

VID

YN1983

Tekn

ikgrup

penAB

Sollentun

aSw

eden

BEM

theory

Mod

alapproach

Itissta

rted

aspartof

thee


twolargeSw

edish

prototypes

Maglarp

andNassuden


(a) (b)


(a) (b)







0

2

4

6

8

10

12

14

Tota

l tip

defl

ectio

n (m

)

05 10 15 20 2500

Time (s)

(a)


0

1

2

3

4

5

6

Max

imum

stre

ss (P

a)

05 10 15 20 2500

Time (s)

times108

(b)










d2ydx2= M(x)

EI[1 + (119889119910119889119909)

2]32

(3)














Upper surface tab

Lower surface tab

(a)

098 1 102 104 106094xc

(b)

e

e

h

h

z

z

e

h

z

e

h

zFlow

Flow


Vane-type VGs

Wheeler VGs

Wishbone Doublet(c)

Air jet VGAir jet

Vortex

(d)







Ref no flap d 00

d sinminuscos d Wsp

7000

7500

8000

8500

9000

9500

10000

DEL

Mx

BlR

t (kN

m) (

=10

)m

14 16 18 20 22 2412Mean Wsp (ms)
















Acknowledgments



References











































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










100

150

200

250

300

350

400

Roto

r sha

ft to

rque

(kN

m)

50 6030 4010 200Time (s)





















Y XZ





















Table2Com

paris

onof

aeroela

sticcodesfor

windturbine

Nam

eofthe

code

Started

Develo

per

Aerodynamic

mod

elStructuralmod

elDescriptio

n

GHBladed

1996

GarradHassanand

PartnersLtd

BEM

theory

Mod

alapproach


edforo

ffsho

reapplicationwith

mon

opile

orgravity

-based

foun

datio

nsTh

elatev

ersio

niscapableo

fmod

elingandanalyzing

both

onshorea

ndoff

shorew

indturbinew

ithvarie

tyof

supp

ortspecification

HAW

C22003ndash200

6

Risoslash

National

Labo

ratory

TechnicalU

niversity

ofDenmark

BEM

theory

Multib

ody

dynamics

Itisas

uccessor

forH

AWEC

tool(w

hich

utilizesF

Emetho

dusingsubstructure

approach

with

Timoshenk

o-beam


odeling

)HAW

EC2isa

timed

omain

morec

omprehending

toolTh

eaerod

ynam

icmod

elismod

ified

tohand

ledynamicinflo

wdynamicstallskew

inflo

wsheare

ffectso

ntheind

uctio

nandeffectsfro

mlarged

eflectio

n

ADAMSWT

mdashMechanical

Dyn

amicsIncun

der

contractof

NRE

LBE

Mtheory

Multib

ody

dynamics

ADAMSWTisreplaced

with

FAST


plug-in

AdWiM

oandAe

roDyn

with

A2A

Dinterfa

cecanalso

beintegrated

with

multib

odysoftw

are

ADAMSsolver

tomod

elandsim

ulate

Alcyone

mdashCenterfor

Renewable

Energy

Source

ampNTU

ABE

Mtheory

FEM

NationalTechn

icalUniversity

ofAthens

developedAlcyone(

freew

ake)with

free

wake

panelm

ethodwhich

also

inclu

desa

simulator

ofturbulentw

indfieldstim

edo

mainaeroela

sticanalysisof

thefullw

indturbinec

onfig

uration

and

postp

rocessingof

loadsfor

fatig

ueanalysis

TURB

U2007

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamics

Itisafrequ

ency

domain

lineariz

edaerohydroservoela

sticcode

andthea

ctive

aeroela

sticcontrolcod

e

DUWEC

D1986

TUDelft

BEM

theory

Multib

ody

dynamics

In1993

itwas

mod

ified

tomod

eloff

shorew

indturbinea

ndlatertoinclu

dewave

loads

FAST

mdashOregonState

University

under

contractof

NRE

LBE

Mtheory

Mod

alapproach

Thistoolhasm

uchles

srun

timeIn

1996N

RELhasm

odified

FAST

tousethe

AeroDyn

subrou

tinep

ackage

developedattheU

niversity

ofUtahto

calculatethe

aerodynamicforces

alon

gtheb

lade

FLEX

5mdash

DTU

BEM

theory

Mod

alapproach

Itistim

edom

ainaeroela

sticsim

ulationtoolanduses

relatively

fewer

degree

offre

edom

tomod

eltheturbine

FLEX

LAST

1982

StorkProd

uct

Engineering

BEM

theory

Multib

ody

dynamic

Since1992thiscode

hasb

eenused

asdesig

nandcertificatio

ntool

PHAT

AS

1993

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamic

Todeterm

inethe

nonlineard

ynam

icbehavior

andthec

orrespon

ding

loadso

faho

rizon

tal-a

xiswindturbine(bo

thon

shorea

ndoff

shore)in

timed

omain

TWISTE

R1983

StentecB

VTh

eNetherla

nds

BEM

theory

FEM

Initiallyitisused

tobe

calledFK

Asince

1991itsup

portsscholastic

windfield

simulation

VID

YN1983

Tekn

ikgrup

penAB

Sollentun

aSw

eden

BEM

theory

Mod

alapproach

Itissta

rted

aspartof

thee


twolargeSw

edish

prototypes

Maglarp

andNassuden


(a) (b)


(a) (b)







0

2

4

6

8

10

12

14

Tota

l tip

defl

ectio

n (m

)

05 10 15 20 2500

Time (s)

(a)


0

1

2

3

4

5

6

Max

imum

stre

ss (P

a)

05 10 15 20 2500

Time (s)

times108

(b)










d2ydx2= M(x)

EI[1 + (119889119910119889119909)

2]32

(3)














Upper surface tab

Lower surface tab

(a)

098 1 102 104 106094xc

(b)

e

e

h

h

z

z

e

h

z

e

h

zFlow

Flow


Vane-type VGs

Wheeler VGs

Wishbone Doublet(c)

Air jet VGAir jet

Vortex

(d)







Ref no flap d 00

d sinminuscos d Wsp

7000

7500

8000

8500

9000

9500

10000

DEL

Mx

BlR

t (kN

m) (

=10

)m

14 16 18 20 22 2412Mean Wsp (ms)
















Acknowledgments



References











































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation



















Y XZ





















Table2Com

paris

onof

aeroela

sticcodesfor

windturbine

Nam

eofthe

code

Started

Develo

per

Aerodynamic

mod

elStructuralmod

elDescriptio

n

GHBladed

1996

GarradHassanand

PartnersLtd

BEM

theory

Mod

alapproach


edforo

ffsho

reapplicationwith

mon

opile

orgravity

-based

foun

datio

nsTh

elatev

ersio

niscapableo

fmod

elingandanalyzing

both

onshorea

ndoff

shorew

indturbinew

ithvarie

tyof

supp

ortspecification

HAW

C22003ndash200

6

Risoslash

National

Labo

ratory

TechnicalU

niversity

ofDenmark

BEM

theory

Multib

ody

dynamics

Itisas

uccessor

forH

AWEC

tool(w

hich

utilizesF

Emetho

dusingsubstructure

approach

with

Timoshenk

o-beam


odeling

)HAW

EC2isa

timed

omain

morec

omprehending

toolTh

eaerod

ynam

icmod

elismod

ified

tohand

ledynamicinflo

wdynamicstallskew

inflo

wsheare

ffectso

ntheind

uctio

nandeffectsfro

mlarged

eflectio

n

ADAMSWT

mdashMechanical

Dyn

amicsIncun

der

contractof

NRE

LBE

Mtheory

Multib

ody

dynamics

ADAMSWTisreplaced

with

FAST


plug-in

AdWiM

oandAe

roDyn

with

A2A

Dinterfa

cecanalso

beintegrated

with

multib

odysoftw

are

ADAMSsolver

tomod

elandsim

ulate

Alcyone

mdashCenterfor

Renewable

Energy

Source

ampNTU

ABE

Mtheory

FEM

NationalTechn

icalUniversity

ofAthens

developedAlcyone(

freew

ake)with

free

wake

panelm

ethodwhich

also

inclu

desa

simulator

ofturbulentw

indfieldstim

edo

mainaeroela

sticanalysisof

thefullw

indturbinec

onfig

uration

and

postp

rocessingof

loadsfor

fatig

ueanalysis

TURB

U2007

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamics

Itisafrequ

ency

domain

lineariz

edaerohydroservoela

sticcode

andthea

ctive

aeroela

sticcontrolcod

e

DUWEC

D1986

TUDelft

BEM

theory

Multib

ody

dynamics

In1993

itwas

mod

ified

tomod

eloff

shorew

indturbinea

ndlatertoinclu

dewave

loads

FAST

mdashOregonState

University

under

contractof

NRE

LBE

Mtheory

Mod

alapproach

Thistoolhasm

uchles

srun

timeIn

1996N

RELhasm

odified

FAST

tousethe

AeroDyn

subrou

tinep

ackage

developedattheU

niversity

ofUtahto

calculatethe

aerodynamicforces

alon

gtheb

lade

FLEX

5mdash

DTU

BEM

theory

Mod

alapproach

Itistim

edom

ainaeroela

sticsim

ulationtoolanduses

relatively

fewer

degree

offre

edom

tomod

eltheturbine

FLEX

LAST

1982

StorkProd

uct

Engineering

BEM

theory

Multib

ody

dynamic

Since1992thiscode

hasb

eenused

asdesig

nandcertificatio

ntool

PHAT

AS

1993

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamic

Todeterm

inethe

nonlineard

ynam

icbehavior

andthec

orrespon

ding

loadso

faho

rizon

tal-a

xiswindturbine(bo

thon

shorea

ndoff

shore)in

timed

omain

TWISTE

R1983

StentecB

VTh

eNetherla

nds

BEM

theory

FEM

Initiallyitisused

tobe

calledFK

Asince

1991itsup

portsscholastic

windfield

simulation

VID

YN1983

Tekn

ikgrup

penAB

Sollentun

aSw

eden

BEM

theory

Mod

alapproach

Itissta

rted

aspartof

thee


twolargeSw

edish

prototypes

Maglarp

andNassuden


(a) (b)


(a) (b)







0

2

4

6

8

10

12

14

Tota

l tip

defl

ectio

n (m

)

05 10 15 20 2500

Time (s)

(a)


0

1

2

3

4

5

6

Max

imum

stre

ss (P

a)

05 10 15 20 2500

Time (s)

times108

(b)










d2ydx2= M(x)

EI[1 + (119889119910119889119909)

2]32

(3)














Upper surface tab

Lower surface tab

(a)

098 1 102 104 106094xc

(b)

e

e

h

h

z

z

e

h

z

e

h

zFlow

Flow


Vane-type VGs

Wheeler VGs

Wishbone Doublet(c)

Air jet VGAir jet

Vortex

(d)







Ref no flap d 00

d sinminuscos d Wsp

7000

7500

8000

8500

9000

9500

10000

DEL

Mx

BlR

t (kN

m) (

=10

)m

14 16 18 20 22 2412Mean Wsp (ms)
















Acknowledgments



References











































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










Table2Com

paris

onof

aeroela

sticcodesfor

windturbine

Nam

eofthe

code

Started

Develo

per

Aerodynamic

mod

elStructuralmod

elDescriptio

n

GHBladed

1996

GarradHassanand

PartnersLtd

BEM

theory

Mod

alapproach


edforo

ffsho

reapplicationwith

mon

opile

orgravity

-based

foun

datio

nsTh

elatev

ersio

niscapableo

fmod

elingandanalyzing

both

onshorea

ndoff

shorew

indturbinew

ithvarie

tyof

supp

ortspecification

HAW

C22003ndash200

6

Risoslash

National

Labo

ratory

TechnicalU

niversity

ofDenmark

BEM

theory

Multib

ody

dynamics

Itisas

uccessor

forH

AWEC

tool(w

hich

utilizesF

Emetho

dusingsubstructure

approach

with

Timoshenk

o-beam


odeling

)HAW

EC2isa

timed

omain

morec

omprehending

toolTh

eaerod

ynam

icmod

elismod

ified

tohand

ledynamicinflo

wdynamicstallskew

inflo

wsheare

ffectso

ntheind

uctio

nandeffectsfro

mlarged

eflectio

n

ADAMSWT

mdashMechanical

Dyn

amicsIncun

der

contractof

NRE

LBE

Mtheory

Multib

ody

dynamics

ADAMSWTisreplaced

with

FAST


plug-in

AdWiM

oandAe

roDyn

with

A2A

Dinterfa

cecanalso

beintegrated

with

multib

odysoftw

are

ADAMSsolver

tomod

elandsim

ulate

Alcyone

mdashCenterfor

Renewable

Energy

Source

ampNTU

ABE

Mtheory

FEM

NationalTechn

icalUniversity

ofAthens

developedAlcyone(

freew

ake)with

free

wake

panelm

ethodwhich

also

inclu

desa

simulator

ofturbulentw

indfieldstim

edo

mainaeroela

sticanalysisof

thefullw

indturbinec

onfig

uration

and

postp

rocessingof

loadsfor

fatig

ueanalysis

TURB

U2007

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamics

Itisafrequ

ency

domain

lineariz

edaerohydroservoela

sticcode

andthea

ctive

aeroela

sticcontrolcod

e

DUWEC

D1986

TUDelft

BEM

theory

Multib

ody

dynamics

In1993

itwas

mod

ified

tomod

eloff

shorew

indturbinea

ndlatertoinclu

dewave

loads

FAST

mdashOregonState

University

under

contractof

NRE

LBE

Mtheory

Mod

alapproach

Thistoolhasm

uchles

srun

timeIn

1996N

RELhasm

odified

FAST

tousethe

AeroDyn

subrou

tinep

ackage

developedattheU

niversity

ofUtahto

calculatethe

aerodynamicforces

alon

gtheb

lade

FLEX

5mdash

DTU

BEM

theory

Mod

alapproach

Itistim

edom

ainaeroela

sticsim

ulationtoolanduses

relatively

fewer

degree

offre

edom

tomod

eltheturbine

FLEX

LAST

1982

StorkProd

uct

Engineering

BEM

theory

Multib

ody

dynamic

Since1992thiscode

hasb

eenused

asdesig

nandcertificatio

ntool

PHAT

AS

1993

ECNof

the

Netherla

nds

BEM

theory

Multib

ody

dynamic

Todeterm

inethe

nonlineard

ynam

icbehavior

andthec

orrespon

ding

loadso

faho

rizon

tal-a

xiswindturbine(bo

thon

shorea

ndoff

shore)in

timed

omain

TWISTE

R1983

StentecB

VTh

eNetherla

nds

BEM

theory

FEM

Initiallyitisused

tobe

calledFK

Asince

1991itsup

portsscholastic

windfield

simulation

VID

YN1983

Tekn

ikgrup

penAB

Sollentun

aSw

eden

BEM

theory

Mod

alapproach

Itissta

rted

aspartof

thee


twolargeSw

edish

prototypes

Maglarp

andNassuden


(a) (b)


(a) (b)







0

2

4

6

8

10

12

14

Tota

l tip

defl

ectio

n (m

)

05 10 15 20 2500

Time (s)

(a)


0

1

2

3

4

5

6

Max

imum

stre

ss (P

a)

05 10 15 20 2500

Time (s)

times108

(b)










d2ydx2= M(x)

EI[1 + (119889119910119889119909)

2]32

(3)














Upper surface tab

Lower surface tab

(a)

098 1 102 104 106094xc

(b)

e

e

h

h

z

z

e

h

z

e

h

zFlow

Flow


Vane-type VGs

Wheeler VGs

Wishbone Doublet(c)

Air jet VGAir jet

Vortex

(d)







Ref no flap d 00

d sinminuscos d Wsp

7000

7500

8000

8500

9000

9500

10000

DEL

Mx

BlR

t (kN

m) (

=10

)m

14 16 18 20 22 2412Mean Wsp (ms)
















Acknowledgments



References











































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










(a) (b)


(a) (b)







0

2

4

6

8

10

12

14

Tota

l tip

defl

ectio

n (m

)

05 10 15 20 2500

Time (s)

(a)


0

1

2

3

4

5

6

Max

imum

stre

ss (P

a)

05 10 15 20 2500

Time (s)

times108

(b)










d2ydx2= M(x)

EI[1 + (119889119910119889119909)

2]32

(3)














Upper surface tab

Lower surface tab

(a)

098 1 102 104 106094xc

(b)

e

e

h

h

z

z

e

h

z

e

h

zFlow

Flow


Vane-type VGs

Wheeler VGs

Wishbone Doublet(c)

Air jet VGAir jet

Vortex

(d)







Ref no flap d 00

d sinminuscos d Wsp

7000

7500

8000

8500

9000

9500

10000

DEL

Mx

BlR

t (kN

m) (

=10

)m

14 16 18 20 22 2412Mean Wsp (ms)
















Acknowledgments



References











































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











0

2

4

6

8

10

12

14

Tota

l tip

defl

ectio

n (m

)

05 10 15 20 2500

Time (s)

(a)


0

1

2

3

4

5

6

Max

imum

stre

ss (P

a)

05 10 15 20 2500

Time (s)

times108

(b)










d2ydx2= M(x)

EI[1 + (119889119910119889119909)

2]32

(3)














Upper surface tab

Lower surface tab

(a)

098 1 102 104 106094xc

(b)

e

e

h

h

z

z

e

h

z

e

h

zFlow

Flow


Vane-type VGs

Wheeler VGs

Wishbone Doublet(c)

Air jet VGAir jet

Vortex

(d)







Ref no flap d 00

d sinminuscos d Wsp

7000

7500

8000

8500

9000

9500

10000

DEL

Mx

BlR

t (kN

m) (

=10

)m

14 16 18 20 22 2412Mean Wsp (ms)
















Acknowledgments



References











































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











d2ydx2= M(x)

EI[1 + (119889119910119889119909)

2]32

(3)














Upper surface tab

Lower surface tab

(a)

098 1 102 104 106094xc

(b)

e

e

h

h

z

z

e

h

z

e

h

zFlow

Flow


Vane-type VGs

Wheeler VGs

Wishbone Doublet(c)

Air jet VGAir jet

Vortex

(d)







Ref no flap d 00

d sinminuscos d Wsp

7000

7500

8000

8500

9000

9500

10000

DEL

Mx

BlR

t (kN

m) (

=10

)m

14 16 18 20 22 2412Mean Wsp (ms)
















Acknowledgments



References











































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










Upper surface tab

Lower surface tab

(a)

098 1 102 104 106094xc

(b)

e

e

h

h

z

z

e

h

z

e

h

zFlow

Flow


Vane-type VGs

Wheeler VGs

Wishbone Doublet(c)

Air jet VGAir jet

Vortex

(d)







Ref no flap d 00

d sinminuscos d Wsp

7000

7500

8000

8500

9000

9500

10000

DEL

Mx

BlR

t (kN

m) (

=10

)m

14 16 18 20 22 2412Mean Wsp (ms)
















Acknowledgments



References











































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










Ref no flap d 00

d sinminuscos d Wsp

7000

7500

8000

8500

9000

9500

10000

DEL

Mx

BlR

t (kN

m) (

=10

)m

14 16 18 20 22 2412Mean Wsp (ms)
















Acknowledgments



References











































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










References











































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation









wind turbine aeroelastic modeling: basics and cutting edge...

Documents