The Sandia Legacy VAWT Research Program...2016/09/07 · VAWT model Delft University of Technology 6 Challenges in Offshore VAWTS 7-9 Sept 2016 Sandia Legacy VAWT Program (continued)

Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company,

for the United States Department of Energy’s National Nuclear Security Administration

under contract DE-AC04-94AL85000.

The Sandia Legacy VAWT Research Program

Dale BergWind Energy Consultant

Formerly

Wind & Water Power Technologies Sandia National Laboratories

[email protected](+1) 505-235-6392

Scientific and Technological ChallengesIn Offshore VAWTs

7-9 September 2016Delft, NL

DelftUniversity ofTechnology

Challenges in Offshore VAWTS 7-9 Sept 20162

Acknowledgements

Members of the Wind group between ’73 and ‘95

• Sandia National Laboratories, Albuquerque, NM.

Matrix support from other groups

• Sandia National Laboratories, Albuquerque, NM.

Contractors to the Wind group



Outline

U.S. Federal Wind Program

Sandia National Labs VAWT R&D program – legacy work• Technology development – 17-m

technology transfer to industry

• VAWT-specific blade design

• Large machine/34-m turbine technology transfer to industry

FloWind Experience

VAWT Review

FloWind cost-reduction efforts

Summary



U.S. Federal Wind Program Started under National Science Foundation in ’74 (response to

’73 oil crisis), placed under new Department of Energy in ‘77

Early program consisted of Large Wind (HAWTs), Small Wind (HAWTs) and Innovative Systems (VAWTs and many others)

Goal of Large Wind was to drive large machine technology –fund building of large numbers of machines• MOD series of HAWT machines

• Planned VAWT machine, MOD 6V

Approach changed in ’81 - funding of large number of machines was replaced by industry “carrots” of federal and state tax credits



Sandia Legacy VAWT Program NRC of Canada “Re-invented” Darrieus concept in late ’60s

(original patents date back to ’25-’31)

SNL acquired basic information from NRC in ‘73

Built original turbine in ’74 (placed on rooftop – soon moved)• 5-m diameter

• crude blades (airfoil – NACA 0012 – only in center section)

• spun up to speed by hand

• stopped with manual disk brake

• Cpmax = 0.27 (3 blades)

Refit with new blades in ’77• aluminum extrusion

• NACA 0015,troposkein shape

• Cpmax = 0.39 (3 blades) 5-m Turbine with

aluminum blades

Original SNL

VAWT model



Sandia Legacy VAWT Program(continued)

Built 2nd Machine in ‘77• 17-m diameter, helicopter technology blades (NACA 0012)

• 2 or 3 blades

• deep airfoil struts


Refit with new blades in ’79• extruded aluminum

• NACA 0015 profile

• no struts


2-m wind tunnel model

Darrieus configuration

All required guy cables

17-m with NACA 0012 Blades

17-m with NACA 0015 Blades




Structural and aero codes developed to explain experimental results/predict performance of proposed turbines

5-m relatively stiff, so effects of gravity and centrifugal stiffening not readily apparent

VAWTDYN developed to model 17-m structural measurements

NASTRAN-based codes quickly followed• FEVD – free vibration code

• FFEVD – forced vibration code

Veers developed 3-d turbulence simulation code• turbulence expands the range of influence of the turbine “per-rev”

excitation frequencies

• Lobitz showed that this explained the MOD2 driveshaft failures (’84)




Aerodynamic codes began with the single streamtube code developed by Templin at NRC

DARTER (multiple streamtube) quickly followed to give force distribution

PAREP (curve fitting) predicted new machine performance, based on measured performance of existing machines

Strickland provided vortex-element VDART codes in ’82• quickly modified to include Reynolds number variation along the blade

• added dynamic stall model (Gormont) in ’83

Paraschivoiu provided Double-Multiple Streamtube (DMST) in ’83• modified into the Sandia COSUDIS, SIDIF and SLICEIT codes



Sandia Legacy VAWT ProgramTechnology Transfer

Alcoa Low-Cost effort (’79 – ’85)• DOE funding, Sandia technical support

• built 5 17-m turbines (commercial prototypes)

• Rocky Flats, CO; Martha’s Vineyard, MA; Block Island, RI, Culebra, PR

Federally-funded development of large VAWT (part of NASA MOD program) started in ’80, but terminated in ’81 (MOD 6V)

Remainder of Tech Transfer efforts were limited to Sandia support of private industry efforts (enticed by state and federal tax credits and California SO-4 contracts)• CA SO-4: utility companies committed to long-term purchase of wind

energy at high prices to hedge against anticipated oil price increases



Sandia Legacy VAWT ProgramTechnology Transfer (continued)

Alcoa built a few “123-82” machines (’82-’83)• based on “Low-Cost” design

• 3-bladed Darrieus with guy cables

• extruded aluminum blades

• Southern California Edison machine failed on start-up due to controller problem

• quit the wind business

Forecast Industries (later VAWTPower)• based on “Low-Cost” design

• 2-bladed Darrieus with guy cables


• only 15-20 built

• structural resonance problems from the start; quickly went bankrupt




FloWind• based on “Low-Cost” design (17-m and 19-m versions)

• 2-bladed Darrieus design with guy cables


• built over 520; 511 sited in two California wind areas (San Gorgonio Pass and Altamont Pass) (’83 – ’85)

• largest and most successful wind turbines available in the mid ’80s

• one of the largest wind companies of that era

• fatigue of the blades (common to most HAWTs as well) led to wide-spread blade issues starting in ’86 or so

didn’t understand importance of fatigue

didn’t understand fatigue properties of aluminum

didn’t understand wind loading environment

FloWind 17-m Turbines




Paraphrasing the words of Paul Gipe, well-known critic of VAWT technology*:

“By the end of 1985, FloWind….had installed 95 MW of its signature product (equivalent to 15,000 Quiet Revolution’s 5-meter).”

“At their most productive, FloWind’s fleet generated 100 million kilowatt hours (per year) – enough electricity for nearly 20,000 California homes.”

“FloWind’s 17-meter model was rated at 142 kW at a wind speed of 38 mph (17 m/s), their 19-meter model was rated at 250 kW at a wind speed of 44 mph (20 m/s).”

“For comparison, a conventional wind turbine 18 meters in diameter would typically be rated at 100 kW and a 21-meter turbine would be rated at 150 kW. Thus, the FloWind turbines were overrated in comparison to their peer by at least 50%.”*Paul Gipe, 2009, “Wind Energy Basics Revised: A Guide to Home- and Community-scale Wind Energy Systems”




FloWind (continued)• some machines survived for years in a very, very aggressive site known

as Horned Toad Ridge several different makes of HAWTs had quickly failed there

very high sustained, highly turbulent winds

energy capture for some of these VAWTs was significantly higher than any HAWTs were achieving

FloWind 19-m

Turbines

• 511 of these machines continued to operate through 1997 (output dropped significantly after ’93 as more and more problems occurred)



Sandia Legacy VAWT ProgramBlade Design

All early turbines (HAWTs and VAWTs) used NACA airfoils

Sandia optimization studies (Kadlec) revealed changing airfoilcharacteristics could yield lower COE• lower and wider drag bucket

• sharp stall

• reduced sensitivity to roughness

Worked with Prof. Gregorek at Ohio State to design Sandia symmetric airfoils (’82 time frame)• similar to previous NACA natural laminar flow (NLF) designs

• SNL 0015/47, 0018/50, 0021/50

• extrusion method of blade fabrication makes NLF achievable

NLF Profile



Sandia Legacy VAWT ProgramLarge Machines

Sandia started the MW-size MOD 6V in ’80 • DOE effort to drive large machine technology

• Sandia built up large staff

• cancelled when tax credits implemented

Sandia started work on 34-m (500 kW) in ‘84• not a commercial prototype, but a research machine

• create experimental database

• demonstrate new technology

• validate codes

Sandia MOD 6V Staff

Artist’s Sketch of 34-m



Sandia Legacy VAWT Program34-m Turbine

34-m Design Features• variable-chord blade profile

tailored to aero/load environment

NACA 0021 at root

SNL 18/50 for rest of blade

step changes in chord length

multiple extrusions required for each of 5 sections

no blade/tower struts

• variable-speed drive train (28-38 rpm) evaluate structural codes

evaluate true variable-speed operation

improve safety by enabling limitation of power production

• heavily instrumented 128 channels of continuously-recorded data

Blade Details

NACA 0021 Blade Section



Sandia Legacy VAWT Program34-m Turbine (continued)

Rotor• diameter 34 m

• height 50 m

• speed 28 to 38 rpm

• swept area 955 m2

• solidity 0.13

Performance• 500 kW at 37.5 rpm, 12.5 m/s

• survival wind speed 67 m/s

Blade Chord• 1.22m at root

• 0.91m at equator

• 1.07m transition 34-m Turbine


Sandia Legacy VAWT Program34-m Turbine (continued)

Structural Response Aero performance



Sandia Legacy VAWT ProgramLarge Machines Tech Transfer

Commercialized version of 34-m (Point Design)• Sandia effort – reduce cost of 34-m technology

• smaller tower

• single-speed transmission

• Invited companies to partner with Sandia to produce new generation of VAWTs

• FloWind won the DOE cost-share contract



FloWind Experience FloWind first-gen VAWTs demonstrated good aero

performance

Fatigue problems created high maintenance costs• sustainable for several yeas because of high SO-4 prices

Initial machine design issues (easy to see in retrospect)

- “Lessons learned”• did not understand the fatigue-critical nature of wind turbines

sudden change in blade stiffness at outer end of blade mount/deep strut

straight line of bolts used to attach blades to struts/mounts (“crack here”)

• did not understand high-cycle fatigue characteristics of 6063 T5 aluminum

not determined until early ’90s

• poor quality control on material (and blade profile)



Review of VAWT Loading Even with steady wind, VAWT angle

of attack (and lift) varies dramatically with each rotation • dynamic stall encountered at higher

wind speeds

Importance of turbulent wind not appreciated until late ’80s• addition of turbulence adds relatively

small perturbation to already unsteady loads

Long-term fatigue properties of materials not known until recently• aircraft requires 106 loading cycles

• wind requires 109 loading cycles



Common Negative Perceptions about VAWTs (1980s)

VAWTs are fatigue prone due to inherent cyclic aero loading• Many FloWind machines started to experience fatigue problems in mid

’80s (6063-T5 aluminum has poor fatigue properties)

VAWTs produce less energy than HAWTs of similar rating• Early VAWTs didn’t experience power regulation at higher winds (output

continued to increase as winds increased) due to use of NACA airfoils

• rated at high wind speeds (17-m rated as 142 kW at 17m/s, 19-m rated as 250 kW at 20m/s) – most machines are rated at 12-13m/s!

VAWTs have lower efficiencies than HAWTs



Common Negative Perceptions about VAWTs (continued)

Blades are much more expensive than HAWT blades for comparable turbine output• VAWT blade twice as long as HAWT blade for given swept area

VAWTs produce oscillating torque• major problem for drive trains



Reality Check on Negative Perceptions of VAWTs

HAWTs were also experiencing many fatigue failures (and yaw drive failures)

Turbulent winds have small impact on VAWT fatigue loading• small addition to cyclic loads resulting from steady-wind aerodynamic

loading

Turbulent winds have large impact on HAWT fatigue loading• major addition to constant loads resulting from steady-wind aerodynamic

loading

• MOD2 driveshaft failures (all 3 machines at Goodnoe Hills, WA) due to omission of wind turbulence in analysis (Lobitz ’84)

VAWT-specific airfoils result in improved power regulation• turbine ratings better defined today



Reality Check on Negative Perceptions of VAWTs (continued)

VAWT blades cheaper per unit length• supported both ends (Darrieus)

• lack of twist/taper

VAWT torque oscillations easily damped out

VAWT peak performance comparable to HAWT peak performance• SNL 17-m turbine Cpmax measured at 0.45 in ’77 (comparable to any HAWT

of that period)



Disadvantages/Advantages of VAWTs

Disadvantages of early-generation VAWTs• blades experience oscillating aerodynamic loads

torque output is not uniform

• can’t capture higher winds typically found at greater heights – difficult to place them on tall towers

• must be spun up to operating speed (no ability to pitch blades to capture low winds)

requires bi-directional gearbox

• guy cables result in large “footprint” cantilever design could eliminate this issue

• entire machine was supported by a single bearing



Advantages/Disadvantages of VAWTs (continued)

Advantages of VAWTs • blades do not experience oscillating gravity-induced loads

extremely large machines are theoretically possible

• blades are attached at both ends and are primarily loaded in tension (Darrieus-type)

• no need to yaw blades into the wind no loss in efficiency due to wind direction vertical shear

• drive train at ground level (low overturning moment, easy to access)

• quieter than HAWTs?



FloWind Cost Reduction Efforts Extended height/diameter (EHD) development

• FloWind effort (cost share with DOE) with Sandia/NREL technical support (FloWind won the cost-share Point Design contract)

• replacement for fatigue-plagued 17-m and 19-m turbines

• development of new generation of larger VAWTs

FloWind goals• use composite blades to eliminate the fatigue problem

no available source for composite blades of that size

long-term fatigue characteristics not considered a problem

• use existing towers/generators decrease cost

increase energy capture

• use Sandia NLF airfoil to improve performance

• improve reliability of braking system and other

high-maintenance items



FloWind Cost Reduction Efforts(continued)

Extended height/diameter (EHD) development• 3 blades to reduce tower/blade dynamics issues

• extended height to capture more energy on same base (EHD)

• utilized pultrusion technology bend into place

▪ significant bending stresses in blades

▪ unable to assemble on ground

deep struts to stiffen/control dynamics

• developed pultrusion process for blades unable to find contractor to produce blades (new design)

purchased pultrusion machine/mastered process

SNL 21/50 had laminar separation bubble – high drag

▪ disappointing performance – 15% low

first blades with new blade profile buckled when bent

redesigned blades with new blade profile successful

▪ performance improved, but still 10% low



FloWind Cost Reduction Efforts(continued)

Extended height/diameter (EHD) development• initial performance disappointing

deep, unfaired struts

• fairing on struts improved performance by about 5%

• built two 3-bladed prototype machines (’93 – ‘96)

• declared bankruptcy in ‘96 development of new machine took too long/cost too much

SO-4 contracts expired

high maintenance costs on existing fleet – not competitive

couldn’t afford to replace existing fleet

No other U.S. company has attempted large-scale VAWT development• negative perceptions of VAWTs


Sandia Legacy VAWT Program

Much of this information has been taken from a Sandia report published in early 2012 – SAND2012-0304



Summary SNL research developed/validated tools for VAWT design

FloWind efforts demonstrated VAWTs were reliable technology

Fatigue failures increased maintenance cost• lack of knowledge regarding importance of fatigue, long-term fatigue

behavior of materials

FloWind cost reduction efforts failed• main problem was fatigue of aluminum blades

• limited time window for redesign/refurbish effort expiring SO-4 contracts

• attempted to modify complete turbine design

• encountered numerous problems with blade design/fabrication

• experienced time/cost overruns

FloWind declared bankruptcy; machines sold, eventually replaced with HAWTs


Thank You

Questions?



Other VAWT Programs(80’s Time Frame)

ECN (’85)• Pioneer I (15-m Darrieus with cantilever tower)

Sir Robert McAlpine & Sons (VAWT Ltd) (’85 – ’92)• H-turbines (100, 130 and 500 kW) with cantilever towers

• 260 (100 kW) H turbine

• 450 (130 kW) reefing H turbine (Musgrove)

• 850 (500 kW) H turbine capital cost too high to compete with HAWTs

blade-to-support arm joint very, very expensive

DAF Indal (’74 – ’80s)• 50 kW and 6400 (250/500 kW)

• Darrieus with guyed towers

(a) (b)

Schematic Diagram of the Musgrove

“Arrowhead” Rotor: (a) Full

Extension and (b) Reefed

“H” Turbine



Impact of Legacy Work on Current VAWT Efforts

Creation of a body of knowledge/expertize on VAWT technology• encourage renewed interest in technology

Initiation of wind turbine-specific airfoil design efforts• set the stage for recent efforts at NREL, TUDelft and elsewhere

• demonstrated leading-edge roughness sensitivity is a major concern

Demonstration of the benefits of blade taper• tailored the blade chord to the aero/loads environment

Demonstration of successful control of torque ripple with flexible drivetrain couplings• torque ripple is not a major problem

Reduction of braking load through regenerative braking



Impact of Legacy Work on Current VAWT Efforts (continued)

Development/validation of structural response codes

Development of vortex methods for aerodynamic performance/loads prediction

Demonstration of system optimization (aero/structures/

manufacturability) benefits during design/build of new VAWT

Demonstration of benefits of full variable-speed drive train

Development of modal test techniques

Development of high-cycle fatigue property database (109)• accelerated coupon testing

Demonstration of potential aero performance penalty associated with use of struts



Impact of Legacy Work on Current VAWT Efforts (continued)

Identified several technical barriers to successful VAWT development• need for reliable /redundant braking system

• crucial role of system resonances

• crucial role of fatigue loading

material properties

Documents

The Sandia Legacy VAWT Research Program...2016/09/07 · VAWT model Delft University of Technology 6 Challenges in Offshore VAWTS 7-9 Sept 2016 Sandia Legacy VAWT Program (continued)