Wind Turbine Dynamica, NREL, 2008

Sandy Butterfield

Workshop on Research NeedsFor

Wind Resource Characterization

January 14, 2008

Wind Turbine Dynamics

22006 Wind Program Peer Review

Outline of Presentation

Design process overview

What have we learned (so far)

What’s working

What’s not

What will it take to meet COE goals


First a Little History

Late 70s – early 80s research prototypes

Demonstrated large turbines could be made

Not economical

MOD-2 (2.5 MW)MOD-5 (3.2 MW)

MOD-12 MW

MOD-0A200 kW

Westinghouse600 kW

WTS-44.2 MW

SNL34m

VAWT


Small Companies Chose Small Turbines

Early 80s wind farms in California

Economics were better

Reliability was poor


Evolution of Commercial U.S. Wind Technology (and Design Process)


Design Process Evolution80s:

– Extreme load design – Minimal testing– No standards

90s:– Extensive structural dynamic load testing– New structural dynamic design tools– Turbulence models ( 1D homogeneous)– Fatigue load dominated design– Standards document design process– Predict, test, tune, evolve design

2008:– Greater investment in:

• Design load accuracy • Turbulence models (Homogeneous, 3D

correlated)• Dynamic coupling • Component development• Controls for load mitigation• Hydrodynamic loading • Environmental characterization• 1000s of Design Load Cases

– Site specific design• Rotor diameter matching to site conditions

(wind)• Site assessment


Importance of Accurate Loads

This is usually a matter of repeated loads or environmental effects (Load, temperature, moisture, etc.)

Material resistance to repeated loads is both sensitive and variable.

A small load uncertainty results in an enormous lifetime uncertainty.

A large margin on the mean life is required to avoid early failures

Number of cycles survived

Inte

nsity

of t

he lo

ad

Logarithmic plot

Uncertainty in Load

Uncertainty in Lifetime


High-ReliabilitySystems

Accurate Loads -Design Requirements

Reduced Failure Rates Improved O&M

Inflow Characterization is Critical forHigh-Reliability Systems


Design Approach

Optimize Performance– Aerodynamic efficiency– Maximize swept area

• Site specific

Estimate Loads– Turbulent inflow– Aerodynamics (steady &

unsteady)– Structural dynamics


Aerodynamics

Must reconcile wake and local aerodynamics– Blade element/momentum– Dynamic inflow– Lifting line theory

Airfoil/blade geometry characteristics

Time variant applied forces

Integrate forces to power curve

Power/Rayleigh probability wind distribution

Energy estimatesLocal Blade Aero

Wake Aero


First Maximize Rotor Efficiency

High tip speed ratio rotors = high efficiency & low solidity (blade area/swept area)

Increasing noise


Performance: Maximize Area

13i wV V=

For Maximum Power:

316 127 2 wP AVρ⎛ ⎞= ⎜ ⎟

⎝ ⎠

The Betz Limit


Typical 5 MW Power and Thrust

Site Wind Probability

Density

Power Curve from a Specific

Turbine

Thrust Curve

from Turbine

Site Specific Energy

Estimates

Site Specific Life Time

Load Matrices


Measured Electrical Output of a Wind Turbine

Power

Power Standard Deviation


Dynamic Loads

Mean tower base bending loads decrease in high winds

Fatigue equivalent loads increase

Energy available decreases in higher winds


Turbulence Drives Turbine Dynamics

Estimating Loads (over 20 year life)


Turbulence models3 components

Based on von Karmon isotropic spectrum

Ten minute simulations

Spatial coherence models

Turbulence intensity set by IEC Design Class

Tuned to site specific turbulence intensity data for site suitability assessment

Looking down from above

turbinerotor

flow

Eddy Vorticity Field Associated with a Fully Turbulent Inflow

Neil Kelley 2005


Energy Spectrum of Wind Speed Fluctuation in the Atmosphere

Design Wind Modeling

Turbulence ModelForecasting Models

Wind Waves

Swell Waves


Deterministic Wind Models

Simple models of extreme events

Alternative to extreme turbulence model

Specifies gust characteristics

Combined gusts with direction changes

Facilitates analysis of unfavorable phasing between control system events and gusts 0

10

20

30

40

-5 0 5 10

Time, t (s)

ED

C W

ind

dire

ctio

n ch

ange

, θ(t

) (d

eg)

IEC 61400-1 ed3 (ECD)


IEC 61400-1 Onshore Turbine Design Classes

Table 1 - Basic parameters for wind turbine classes[1]

Wind Turbine Class I II III S

Vref (m/s) 50 42,5 37.5 Values

A Iref (-) 0,16 Specified

B Iref (-) 0,14 by the

C Iref (-) 0,12 Designer

In Table 1, the parameter values apply at hub height and Vref is the reference wind speed average over 10 minutes,•A designates the category for higher turbulence characteristics,B designates the category for medium turbulence characteristics,C designates the category for lower turbulence characteristics andIref is the expected value of the turbulence intensity[2] at 15 m/s.


Coupled Aero-elastic/Hydro-elastic Design Codes

AeroDynTurbSim

HydroDyn

FAST &ADAMS

Wind TurbineAppliedLoads

ExternalConditions

Soil

Hydro-dynamics

Aero-dynamics

Waves &Currents

Wind-Inflow PowerGeneration

RotorDynamics

Substructure Dynamics

Foundation Dynamics

DrivetrainDynamics

Control System

Soil-Struct.Interaction

Nacelle Dynamics

Tower Dynamics


What's Working Why 98% reported availability Design process, improved design

tools, Standards Rotor performance excellent

(80% of theoretical limit) Steady aero codes, airfoils, testing

CapEx drastically reduced Accurate design tools, load control, quality control

Blade Development Standards (design, test, certify) Product evolution strategy Stretch rotor, control loads Power quality control Power electronics


To meet DOE cost goals

Stop gearbox failures

Need new design strategy

Better site specific characteristics

Evolve design tools

Evolve design process

What's Not Working Why

Gearboxes bearing failures, inaccurate internal loads?

OpEx too high "unscheduled maintenance", low reliability, lack O&M automation

CapEx still too high to DOE goals

lack of fatigue load and deflection control

Rotor stretching strategy hitting limits

tower clearance limit, materials, aeroacoustics limiting tip speed, dynamic

Ludeca, Inc.


Commercial Blades - R2.35

0

5

10

15

20

25

20 30 40 50 60Rotor Radius (m)

Wei

ght (

103 kg

)

Commercial Blade Data

Modeling Results

Modeling Results - R2.9

Rotor Innovations key to Scaling Strategy

Finite ElementComputer Model

Scaling of Rotors


RNA Mass / Swept Area

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

40 60 80 100 120 140

Diameter (m)

Mas

s/sw

ept a

rea

(kg/

m^2

)

WindPact Baselines

WindPact Task#5 Final

NREL Baseline 5MW

GPRA 2005 - 2025 Estimates

RePower 5MW

Enercon 6MW

Vestas 4.5MW

MultiBrid 5MW

GE 3.6MW

Clipper

V80

V90

Siemens

How well has the strategy worked?Can we meet the COE goals?

Offshore Turbines

DOE COE pathway (cents/kwh)4.4 3.9 3.4


What will it take?

Design code enhancements– Dynamic coupling of major components– Steady & unsteady aerodynamics– Aeroacoustics (higher tip speeds, reduced tower shadow signature)

Advanced controls (load reduction, deflection control)

System and subsystem innovation (lower cost, greater reliability)– Rotor (reduced dynamic loads)– Blades (increased flexibility, longer fatigue life)– Drivetrain (greater reliability, lower cost)

Site specific turbulence characterization and linkage between:– Local atmospheric physics– 50m – 200m Inflow turbulence (3D coherent structures?)– Unsteady aerodynamic response– Wake to rotor interactions


Carpe Ventem


Gaps(according to Sandy)

Aerodynamics - More accurate steady & unsteady aero models

Aeroacoustics (limits high speed flexible rotors & downwind option)

Increasing flexibility w/o complexity, cost & failure rates

Accurate prediction of coupled dynamic rotor loads

Greater fidelity between loads codes and component design codes

Greater drive-train reliability while reducing cost and weight.

MIMO Control of turbulence & extreme loads without firm measure of inputs (need robust sensor technology)

More accurate inflow characterization, especially greater than 100m.

Linkage between local atmospheric/turbulence/aerodynamic/wakes


Trends

Lifelong O&M (“unscheduled maintenance” becoming critical)

Lighter rotors, higher tip speeds, more flexible blades (lower loads)

Twist/flap coupling

Drivetrain innovation

Controls for load reduction

Offshore design concepts incorporated into onshore turbines (load control, component placement, design for reliability, condition monitoring)

Onshore COE Cost BreakdownO&M (After Tax)

9%LRC & Lease

Cost10%

Electrical Infrastructure

7%Foundation

3% Misc BOS11%

Turbine60%

Offshore COE Cost Breakdown

LRC & Lease Cost6%

Electrical Infrastructure

12%

Eng/Permits 4%

Support Structure14%

Misc BOS13%

Offshore Warranty

6%

Turbine32%

O&M (After Tax)13%


Can rotor improvements help the rest of the system?WindPact Rotor study shows benefits of:– Controlling tower dynamics– Passive blade load relief through twist/flap coupling– High tip speed/low solidity blades

Need follow up system study– SeaCon Turbine

study– Perform system

optimization– Apply practical

implementation experience


Advanced Drivetrain R&D

Today

Tomorrow

GEC

NPS


45-Meter Fatigue Test

Larger blades becoming more flexible

Design innovations require design verification

Aerodynamic advancements improve performance.

Structural improvements increase fatigue tolerance and reduce dynamic loads.

Single-axis Flap Fatigue Test Using B-REX Test System.

Nov.24.2004

45-meter Blade Root Mount


Horns Rev, Denmark 80 Turbines, 160 MW


Aeroelastic Simulators

Codes integrate : – Turbulent inflow– Aerodynamic forces– Coupled structural dynamics– Controls– Wave loading– Other environmental effects


Structural Dynamics

Tower TorsionBlade FlatwiseDeflection

Tower DeflectionBlade EdgewiseDeflection

Yawing

Rolling

Pitching

Wind

TowerShadow

MassLoads

Non-stationaryAerodynamic Loads

CentrifugalForces

BoundaryLayer

ObliqueInflow

GyroscopicForces

Gust

Blade Torsion

Blade vibrations interact with aerodynamic forces = aeroelasticity

Mode shapes and natural frequencies critical


Floating Offshore Turbine Research Interface of SML to FAST and ADAMS

Measurements(power, loads, etc.)

Aerodynamics(AeroDyn)

StructuralDynamics

(FAST, ADAMS)

Controls(user-defined)

Wind Field(TurbSim, field

exp., etc.)

Actuator Inputs(blade pitch, gen. torque, yaw)

Aerodynamic Loads(lift, drag, pitch mom.)

Blade Motions(blade pitch, element pos. & vel.)

Wind-Inflow

Time Series Loads(forces, moments)

Time Series Motions(defl., vel., accel.)

Output

Moorings(Lines)

Hydrodynamic Loads(added mass, damping)

Platform Motions(defl., vel., accel.)Time-Domain

Hydrodynamics(Motion)

Wave Env.(Motion, field

exp., etc.)

Freq. To Time(Motion)

Wave Spectrum

Wave History

Freq.-DomainHydrodynamics

(Swim)

Added Mass &Damping Matrices

Mooring Loads(restoring)

Platform Pos.


Time series simulations

Nonlinearities require time marching solution approach

– Control system

– Aerodynamics

– Large rotations

Load combinations

Limit ability to simulate life time.

requires extrapolation to life time load spectrum

Extreme conditions simulated and added into the load matrix

1.35 load factor applied to all unfavorable loads estimates.


Turbine Design Evolution

80s: (US dominated market)– US = Light weight/flexible – Euro = Heavy/stiff

90s: (Euro dominated market)– Low speed = low tip noise^5– Heavy/stiff evolved– Lighter/larger rotors– Variable speed– Custom airfoils/tips

2008: (World market)– Dynamically active– Flexible for load shedding– Power quality improvements

Documents

Wind Turbine Dynamica, NREL, 2008