Loading...
The URL can be used to link to this page
Your browser does not support the video tag.
Home
My WebLink
About
AWEA Wind Resource & Project Energy Assessment Workshop, September 2007
The AWEA Edge — Your Leading Resource for Wind Energy Education AWEA Wind Resource & Project Energy Assessment Workshop September 18 - 19, 2007 Portland, Oregon American Wind Energy Association * www.awea.org/events * 202.383.2512 Marriott. PORTLAND DOWNTOWN WATERFRONT —— 1401 SW Naito Parkway, Portland, OR 97201 503-226-7600 * portlandmarriott.com Matriott. PORTLAND DOWNTOWN WATERFRONT AMERICAN WinD ENERGY ASSOCIATION GRASSROOTS PROGRAM The American Wind Energy Association (AWEA) is active on Capitol Hill advocating for the wind energy industry's legislative priorities. While a strong lobbying team in Washington is important, the power of any industry resides in the activism of its members. You vote for your representatives, and it's your voice and your experiences that can help guide their decisions - if you make your voice heard. AWEA’s new Grassroots Program helps industry members become more politically active and advocate for AWEA's legislative agenda. By joining the Grassroots Program, AWEA will supply you with the tools to successfully advocate the industry's legislative agenda, including the extension of the Production Tax Credit, the creation of a National Renewables Portfolio Standard, and the creation of a Small Wind Systems Investment Tax Credit. Check out AWEA's website, www.awea.org, for the latest information on other legislative priorities. BECOME A GRASSROOTS PROGRAM PARTICIPANT As a Grassroots Program Participant, you pledge to advocate the wind industry's legislative agenda to your elected officials by: ‘ = Responding to AWEA's emailed Legislative Action Alerts and making every effort to personalize the provided sample letter. Legislative Action Alert ing letters regarding important wind issues snarour Members of Congress J Action Website, attend a project dedication. GRASSROOTS PROGRAM: CONGRESSIONAL TOURS Invite one of your Members of Congress to tour your wind project, facility, or office. These tours create relationships and educate Congressional leaders about the wind energy industry and its potential contribution to jobs and economic development. This event will give you and your staff the chance to develop a closer relationship with your representative. AWEA can help your company organize the tour. (Left) Sen. James Inhofe (R-OK) and his advisor tour Bergey Windpower's small wind turbine manufacturing facility with Grassroots Program Participant Mike Bergey. GRASSROOTS PROGRAM: PROJECT DEDICATIONS Wind farm project dedications or groundbreaking ceremonies are ideal occasions to show your elected Officials the positive impact the wind energy industry has in their state or district. Don't miss the opportunity to invite your Members of Congress or their district staff to your event. AWEA can help your company by offering support, whether it is hand delivering the dedication invitation to a Member of Congress or following up with appropriate Congressional, federal agency, or state government staff. GET THE TOOLS TO MAKE A DIFFERENCE Grassroots Program Participants will receive the following materials to assist in advocating the wind energy industry's legislative agenda with AWEA: = The monthly Grassroots Program e-Newsletter. = Periodic emails with important websites and town hall meetings in your Congressional district. _A Congressional District Brochure with information Join AWEA's GRASSROOTS PROGRAM TODAY To join, fill out this form and drop it in the mail. No postage is necessary. YOUR CONTACT INFORMATION: Name: Company: Address (NO P.O. Boxes please): City: State: Zip Code +4: Phone: Fax: Email: FEDERAL CONTACTS: 0 | would like to participate in the Grassroots Program but 1 do not know my federal legislators yet. 0 I know my U.S. Senator(s) and/or Representative. O I know my U.S. Senator's and/or Representative's staff member(s). NAME OF SENATOR(S) AND/OR REPRESENTATIVE: CO Do Not Know O Know Slightly OD Know Well Name oF STAFF MEMBER(S): 0 Do Not Know O Know Slightly O Know Well NATURE OF YOUR RELATIONSHIP: (Relative, close friend, campaign work, fellow alumni/ae, business associate, etc.) How OFTEN DO YOU COMMUNICATE WITH THIS PERSON? times/year ADDITIONAL COMMENTS: CONGRESSIONAL TOUR PROGRAM: O Yes, my company would like to participate in the Congressional Tour Program. Please contact me. City/State of Tour: Preferred Dates: - PROJECT DEDICATION PROGRAM: =e a Tien, iy ooeectiad CURT Ta iesmeeANin Vs Dinaet Dedication Program. Please contact me. City/State of Dedication: Dates (if known): cL Td MN LS HLbT TOIT LILI-LL00@ Od ‘NOLONIHSVM NOILVIOOSSV ADYUNG GNIM NVOIMGNV I a 52 |: Zz 3 g22 |: Zoos ]5 eso. gos] eP a] Sn2 i Sa]: ao 2 za B 3 5 a ES 3|° 18 =] 0 els [2 ae 2/28 >1s cle g e| | < A x 2p <=. oh Q : a ' a ue SaeeS od Qa sgees SP BRE 5 ro AMERICAN WinD ENERGY ASSOCIAI The American Wind Energy Association (A national association for the U.S. wind energy AWEA is dedicated to advancing the use of cle hs renewable wind energy - one of the fastest-gro energy technologies in the world. s TAKE ACTIO Wind energy is a reality today. The U.S. wind industry installed almost 3,000 megawatts (MW) wind power in 2006, or enough to power the eq of 800,000 homes. This record amount will help skyrocketing home heating and electric bills by reducing the demand for natural gas and other American Wind Energy Association GRASSROOTS PROGRAM President Bush has stated that wind energy can provide as much as 20% of the nation's electricity. W forward-looking, steady policies in place to allow businesses to plan for strategic growth, wind energy ~ can provide a significant portion of the U.S. energy portfolio, revitalize farms and rural communities, reduce volatility in natural gas prices, and strengthen the security of the U.S. electricity supply. Wind Energy Works for America's Economy, Environment and Energy Security LEGISLATIVE DEPARTMENT STAFF Jaime Steve, Director Legislative Affairs Jon Chase, Deputy Director Legislative Affairs Bree Raum, Grassroots Outreach Coordinator Ron Stimmel, Small Wind Advocate Richard Mark Blanche John Jon Barbara Dianne Wayne Ken Christopher Joseph Thomas Joseph Bill Mel Johnny Saxby Daniel Daniel Tom Charles Mike Larry Richard Barack Evan Richard Pat Sam Mitch Jim Mary David Edward John Barbara Benjamin Olympia Susan Debbie 110th Senate Murkowski Stevens Sessions Shelby Pryor Lincoln McCain Kyl Boxer Feinstein oe Allard Salazar Dodd Lieberman Carper Biden Nelson Martinez lsakson Chambliss Akaka - Inouye Harkin Grassley Crapo Craig Durbin Obama =o Bayh a. Lugar Roberts Brownback McConnell Bunning Landrieu ~~ Vitter Kennedy Kerry Mikulski Cardin Snowe Collins Stabenow (202) 224-6665 (202) 224-3004 (202) 224-4124 (202) 224-5744 (202) 224-2353 ~ (202) 224-4843 (202) 224-2235 (202) 224-4521 (202) 224-3553 (202) 224-3841 _ (202) 224-5941 (202) 224-5852 (202) 224-2823 (202) 224-4041 (202) 224-2441 (202) 224-5042 (202) 224-5274 (202) 224-3041 (202) 224-3643 (202) 224-3521 (202) 224-6361 (202) 224-3934 (202) 224-3254 (202) 224-3744 (202) 224-6142 (202) 224-2752 (202) 224-2152 (202) 224-2854 (202) 224-5623 (202) 224-4814 (202) 224-4774 (202) 224-6521 (202) 224-2541 (202) 224-4343 (202) 224-5824 (202) 224-4623 (202) 224-4543 (202) 224-2742 (202) 224-4654 (202) 224-4524 (202) 224-5344 (202) 224-2523 (202) 224-4822 Chuck Kleeschulte Karina Waller Stephen Boyd Ryan Welch Derrick Freeman Todd Wooten Becky Jensen Lucy Murfitt Vacant Jim Folger Mandi McKinley Steve Black Mark Wenzel David Mcintosh Tom Lawler Lisa Borin Susie Perez Quinn Brydon Ross Mike Quiello Camila Knowles Tulsi Gabbard Tamayo. . Marie Blanco Eldon Boes Kurt Kovarik “= Craig Ferguson Corey McDaniel Jessica Lenard Todd Atkinson Chris Murray Steve Koerner Joel Leftwich Riley Scott Allison Thompson Bill Beaver Elizabeth Craddock Garret Graves Ron Carlton Heather Zichal Brigid Kolish Emily Work-Dembowski Ginny Worrest David Hunter Chris Adamo MI [MN |MN MO MO MO MS MS MT MT NC NC ND ND NE NE NH NH NJ NJ NM NM NV NV NY NY OH OH OK OK OR OR PA PA RI RI Sc sc SD SD TN T™N TX TX UT UT VA VA WA WA Wl Wi wv wv WY WY Sen. |Sen. Sen. |Sen. Sen. Sen. |Sen. Sen. Sen. Sen. Sen. Sen. |Sen. Sen. |Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. 'Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. |Sen. ‘Sen. |Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. Sen. |Carl Norm |Amy Christopher |Claire Christopher Trent |Thad Max Jon Richard Elizabeth Kent Byron Ben Chuck John Judd Robert Frank Pete Jeff John Harry Charles Hillary George Sherrod Tom James Ron Gordon Arlen Robert Sheldon Jack Jim Lindsey John Tim Bob Lamar Kay Bailey John Robert Orrin John James Bernard Patrick Maria Patty Herbert Russ Robert John John Michael |Levin Coleman |Klobuchar Bond |McCaskill Bond | Lott Cochran |Baucus Tester Burr Dole Conrad Dorgan Nelson Hagel Sununu Gregg Menendez Lautenberg Domenici Bingaman Ensign Reid Schumer Clinton Voinovich Brown Coburn Inhofe Wyden Smith Specter Casey Whitehouse Reed DeMint Graham Thune Johnson Corker Alexander Hutchison Cornyn Bennett Hatch Warner Webb Sanders Leahy Cantwell Murray Kohl |Feingold Byrd Rockefeller Barrasso Enzi (202) 224-6221 \(202) 224-5641 |(202) 224-3244 |(202) 224-5721 |(202) 224-6154 |(202) 224-5721 (202) 224-6253 (202) 224-5054 (202) 224-2651 (202) 224-2644 (202) 224-3154 (202) 224-6342 (202) 224-2043 (202) 224-2551 (202) 224-6551 (202) 224-4224 (202) 224-2841 (202) 224-3324 (202) 224-4744 (202) 224-3224 (202) 224-6621 (202) 224-5521 (202) 224-6244 (202) 224-3542 (202) 224-6542 (202) 224-4451 (202) 224-3353 (202) 224-2315 (202) 224-5754 (202) 224-4721 (202) 224-5244 (202) 224-3753 (202) 224-4254 (202) 224-6324 (202) 224-2921 (202) 224-4642 (202) 224-6121 (202) 224-5972 (202) 224-2321 (202) 224-5842 (202) 224-3344 (202) 224-4944 (202) 224-5922 (202) 224-2934 (202) 224-5444 (202) 224-5251 (202) 224-2023 (202) 224-4024 (202) 224-5141 (202) 224-4242 (202) 224-3441 (202) 224-2621 (202) 224-5653 (202) 224-5323 (202) 224-3954 (202) 224-6472 (202) 224-6441 (202) 224-3424 Mary Louise Wagner |Tony Eberhard |Hilary Bolea |Michael Collins [Nichole Distefano | John Stoody |Beth Spivey {Emily Brunini |Paul Wilkins |Matt Jennings |Jon Pierpan |Arjun Mody |Mark Schipper |Nathan Hill |Jonathan Coppess ‘Tara Rothschild |Grant Bosse |Nancy Perkins |Megan Bartley Daniel Rosenberg | Energy Committee Jonathan Epstein Alexis Bayer Chris Miller Christine Parker Dan Utech Lauri Hettinger Vacant Brian Treat Dan Barron Dave Berick Valerie West Mary Beth Laverghetta Vacant | Vacant Kristen Sarri Chris Socha Matt Rimkunas Vacant |Matthew Thornblad ‘Ashley Horning Jessica Holliday Jamie Moore |Spencer Chambers Tyler Owens J.J. Brown Tack Richardson Trent Bauserman |Jessica Maher ‘Brian Baenig Amit Ronen |Carrie Desmond |Chad Metzler |Kelly Reed Paul Gay | John Richards Colin Hayes Chris Tomassi eee ee awea ™ american wind merican_Wind Znergy Association TL -£ = energy association AWEA WIND RESOURCE AND PROJECT ENERGY ASSESSMENT WORKSHOP Portland, Oregon September 18 — 19, 2007 TABLE OF CONTENTS Program Agenda Sponsor Information Attendee List Speaker Information Introduction Overviews Actual Performance Wind Speeds . Wind Measurement 10. Data Analysis 11. Flow Modeling 12. Technical Loss Estimates OONAARWNS All the presentations for which AWEA received permission from speakers will be posted online after the workshop. The link will be provided only to the attendees with an online survey via email 1-2 weeks after the workshop. Upcoming AWEA Educational Events: AWEA Wind Power Finance & Investment Workshop October 10 - 11, 2007, New York, NY AWEA Wind Energy Fall Symposium 2007 November 1 - 2, 2007, Carlsbad, CA AWEA Wind Power Asset Management Workshop January 17 — 18, 2008, San Diego, CA AWEA Wind Project Siting Workshop February 14 — 15, 2008, Austin, TX AWEA and CanWEA Wind & Transmission Workshop March 18 — 19, 2008, Detroit, MI WINDPOWER 2008 Conference & Exhibition June 1-4, 2008, Houston, TX For more information on upcoming AWEA events, visit http://www.awea.org/events. DD Wa iia ae PROGRAM AGENDA “Ze __Aitercansyima cnersscaion” P-ROGRAM AGENDA Wind Resource and Project Energy Assessment Workshop Program Chairperson: Robert Z. Poore, President, Global Energy Concepts, LLC MONDAY, SEPTEMBER 17 1:00 pm — 5:00 pm Optional Pre-conference Seminar: Introduction to Wind Energy Location: Oregon Ballroom 1:00 pm — 1:30 pm Overview of the Wind Industry Jeff Anthony, Manager — Utility Programs and Policy, American Wind Energy Association 1:30 pm — 2:00 pm Technology Overview Bruce Bailey, Ph.D., President, AWS Truewind LLC 2:00 pm — 2:30 pm Project Economics Overview William Sutherland, Vice President — Project Finance, Manulife Financial Corporation 2:30 pm — 3:00 pm Break Location: Oregon Ballroom Lobby 3:00 pm — 3:30 pm Siting Overview Dana Peck, Project Manager, Horizon Wind Energy 3:30 pm — 4:00 pm Transmission and Integration Overview Jeff Anthony, Manager — Utility Programs and Policy, American Wind Energy Association 4:00 pm — 4:45 pm Wind Resource Analysis Overview Holly Hughes, Project Engineer, Global Energy Concepts, LLC 4:45 pm — 5:00 pm Wrap-up Session Jeff Anthony, Manager — Utility Programs and Policy, American Wind Energy Association TUESDAY, SEPTEMBER 18 7:30 am Registration Opens Location: Oregon Ballroom Lobby Continental Breakfast / Exhibits Location: Oregon Ballroom, Salons A - D 8:30 am — 8:45 am Welcome & Introduction Comments on AWEA Anti-Trust Guidelines Jeff Anthony, Manager — Utility Programs and Policy, American Wind Energy Association Introduction to Wind Resource and Project Energy Assessment Program Chair: Robert Z. Poore, President, Global Energy Concepts, LLC 8:45 am — 9:45 am Overviews of Current Practices — Part 1 Session Chair: Andy Oliver, Ph.D., Vice President of Technologies & Resource Analysis, RES America Inc. Data - What We Need, Why We Need It and How We Measure It e Wind speed and its distribution (to combine with power curve) e Wind direction (need for topo model / wake model) e Shear (measurement and extrapolation) e Temperature (density, low high temp shutdown losses) e Pressure e Turbulence (turbine certification, wake model) e Towers e Loggers Speaker: John Vanden Bosche, Principal Engineer, Chinook Wind The Role of Wind Flow Modeling and Mapping e Importance of atmospheric models in energy production assessments e Types of atmospheric models: mesoscale and microscale e Examples of model applications Speaker: Michael Brower, Chief Technical Officer, AWS Truewind LLC 9:45 am — 10:15 am Break / Exhibits Location: Oregon Ballroom, Salons A - D 10:15 am — 11:15 am Overviews of Current Practices — Part 2 Session Chair: Andy Oliver, Ph.D., Vice President of Technologies & Resource Analysis, RES America Data Analysis - Multiple Approaches to Evaluating Identical Data e Data processing & quality control - issues and approaches e Predicting the long term wind climate e Extrapolating wind Speeds to hub height Speaker: Andy Oliver, Ph.D., Vice President of Technologies & Resource Analysis, RES America Inc. Energy Yield Calculations e Turbine power curves and wind speed frequency distributions e Energy calculations for a single turbine and extension to multiple turbines in a project e Uncertainty in the windfarm energy yield calculation expressed as Psp and Pos. Speaker: Thomas Hiester, Vice President, Development, Noble Environmental Power LLC 11:15 am — 12:15 pm How Well Have They Done? An assessment of the actual performance of wind power projects compared with pre-construction predictions. Session Chair: William Sutherland, Vice President — Project Finance, Manulife Financial Corporation Speakers: Andrew Tindal, Director, Garrad Hassan & Partners Ltd Steve Jones, Director of Utility & Investor Services, Global Energy Concepts, LLC Panelists: Joel Spenadel, Executive Director, JP Morgan William Marder, Vice President, Fortis Capital Corporation 12:15 pm - 1:30 pm Lunch Location: Oregon Ballroom, Salon E 1:30 pm — 2:45 pm Recent Studies of Wind Speeds in North America Session Chair: Richard Simon, Director, Consulting Meteorologist, V-Bar LLC Speakers: Dennis Elliott, Principal Scientist, National Renewable Energy Laboratory (NREL) Jeffrey Freedman, Ph.D., Research Scientist, AWS Truewind LLC Pascal Storck, President, 3TIER Richard Simon, Director, Consulting Meteorologist, V-Bar LLC 2:45 pm — 3:15 pm Break / Exhibits Location: Oregon Ballroom, Salons A - D 3:15 pm — 5:30 pm Identifying and Reducing Wind Measurement Bias and Uncertainty Session Co-Chairs: Owen Clay, Director of Engineering, NRG Systems Patrick Quinlan, Director, Wind Systems Business Development, Second Wind Inc. Sensors e Definition of anemometer uncertainty e Uncertainty as it relates to the anemometer geometry e Role calibration plays in reducing uncertainty e Bias errors associated with calibration e Ultrasonic anemometer uncertainty Speaker: Bruce Bailey, Ph.D., President, AWS Truewind LLC Towers and Mounting e Sensor orientation, sensor spacing, boom lengths, boom offsets, multiple sensors, and redundant sensors e Tower shadowing affect, and top level sensor configurations e QC of data based on configuration effects Speaker: Dave Baker, President, Phoenix Engineering Inc. Siting e Atmospheric stability and its impact on terrain effects e Placement of met towers to reduce uncertainty e Understanding modeling problem areas and terrain effects Speaker: Ron Nierenberg, Consulting Meteorologist SODAR - Taking the Mystery Out of SODAR e Overview of technology, strengths, limitations, differences from in-situ observations e Discussion of quality assurance and quality control procedures e Examples of performance as a function of noise, humidity, stability, rain, etc. Speaker: Gennaro Crescenti, Manager, Meteorology, PPM Energy Inc. LIDAR - Highlighting Uncertainty with LIDAR e Wind shear and wind veer profiling e Wind flow model validation e Power performance assessment Speaker: Peter Clive, M.Inst.P., Renewable Energy Consultant, SgurrEnergy Ltd Overall Systems View Additional Panelist: Marc LeBlanc, Director, Garrad Hassan Canada Inc. 5:30 pm - 5:45 pm Wrap-Up Session Program Chair: Robert Z. Poore, President, Global Energy Concepts, LLC 5:45 pm — 6:00 pm AWEA Legislative & Grassroots Program: What We Are Doing in Washington, D.C. e AWEA's Legislative Priorities e AWEA's Grassroots Program o Congressional Tour Program o Project Dedication Program o Company Employee Program Speaker: Bree Raum, Manager of Grassroots Advocacy & WindPAC, American Wind Energy Association Networking Reception Location: Oregon Ballroom, Salons A— D & Lobby Immediately following the program for an hour Sponsored by NRG Systems, Inc. WEDNESDAY, SEPTEMBER 19 7:30 am Registration Opens Location: Oregon Ballroom Lobby Continental Breakfast / Exhibits Location: Oregon Ballroom, Salons A - D 8:30 am — 8:45 am Daily Update Program Chairperson: Robert Z. Poore, President, Global Energy Concepts, LLC 8:45 am — 10:15 am Identifying and Reducing Data Analysis Bias and Uncertainty Session Chair: Matt Hendrickson, Director, Energy Assessment, Horizon Wind Energy Introduction & Data Processing e Data processing objectives e Systems design e Things to watch out for Speaker: Matt Hendrickson, Director, Energy Assessment, Horizon Wind Energy Uncertainty in Correlations and Long-Term Adjustments e Quantifying uncertainty associated with period of record data e Identifying appropriate reference stations e Methods for assessing relationships between datasets Speaker: Gordon Randall, Senior Technical Analyst, Global Energy Concepts, LLC The Long-Term Wind Resource: Comparing Data Sources and Techniques for Predicting the Performance of Wind Farms e Atmospheric data sources e MCP methods e Reducing bias and uncertainty Speakers: Robert Conzemius, Ph.D., Senior Atmospheric Scientist, WindLogics Inc. Shear Extrapolation Speakers: David Matson, Director, Consulting Meteorologist, V-Bar LLC Allen Becker, Director, Consulting Meteorologist, V-Bar LLC 10:15 am — 10:45 am Break / Exhibits Location: Oregon Ballroom, Salons A - D 10:45 am — 12:15 pm Identifying and Reducing Flow Modeling Bias and Uncertainty Session Chair: John Wade, John Wade Wind Consultant LLC A New and Objective Empirical Model of Wind Flow Over Terrain e Parameterization of terrain elevation variation e How the wind responds to terrain variations — the data tells us e The development of the model and model results Speaker: Jack Kline, Consulting Meteorologist, RAM Associates Computational Fluid Dynamics (CFD) Speaker: Mike Zulauf, PPM Energy Inc. Mesoscale and Downscaling Techniques Speaker:Justin Sharp, Manager, Wind Asset Management Meteorology, PPM Energy Inc. 12:15 pm - 1:30 pm Lunch Location: Oregon Ballroom, Salon E 1:30 pm - 2:45 pm Identifying and Reducing Bias and Uncertainty in Technical Loss Estimates -Part 1 Session Chair: Clint Johnson, Manager, Wind Farm Projects Group, Garrad Hassan America Inc. Typical Losses — A Survey of Practitioners Panel e Brief overview of all relevant loss factors e Reach consensus among panelists on how loss factors should be defined e List range of typical values of loss factors Panelists: Daniel Bernadett, Chief Engineer, AWS Truewind LLC Dave Baker, President, and Rob Istchenko, Phoenix Engineering Inc. Steve Jones, Director of Utility & Investor Services, Global Energy Concepts, LLC Wake Losses - State of the Art and Validation Studies e Summarize various wake loss modeling techniques e Discuss/quantify areas of potential errors (large wind farms, tight arrays, low turbulence) e Review literature and new techniques for addressing above challenges e Provide recommendations for solutions to above challenges Speaker: Hans Ejsing Jorgensen, Seniorforsker, Ris@ National Laboratory 2:45 pm — 3:15 pm Break / Exhibits Location: Oregon Ballroom, Salons A - D 3:15 pm - 4:45 pm Identifying and Reducing Bias and Uncertainty in Technical Loss Estimates — Part 2 Session Chair: Clint Johnson, Manager, Wind Farm Projects Group, Garrad Hassan America Inc. Power Curves: The Effect of Environmental Conditions e Theoretical and measured effect of co Air density o Turbulence intensity o Wind shear co Blade fouling Speaker: Saskia Honhoff, Ir., Power Performance Engineer, GE Energy Availability Losses e Summary of experience to date (typical values) e Energy-weighted availability Other Losses e Environmental effects — how to predict: oc Icing o Temperature related losses e Other relevant loss factors Panelists: Jesper Christensen, Wind & Site Engineer, Vestas Americas Mike Kelly, Director of Operations Service, Horizon Wind Energy Phil Stiles, Senior Product Engineering Manager, Acciona Energy North America Mark Young, Project Engineer, Global Energy Concepts, LLC 4:45 pm — 5:00 pm Wrap-up Session Program Chairperson: Robert Z. Poore, President, Global Energy Concepts, LLC SPONSOR INFORMATION uopjewsojul sosuods AWEA Wind Resource and Project Energy Assessment Workshop Sponsor Contact Details GIGA-WATT SPONSORS Max SYSTrenm s NRG Systems, Inc: NRG Systems is the global leader in the development and manufacture of complete wind measurement systems and turbine control sensors. Contact: Larry Jacobs 110 Riggs Road Hinesburg, VT 05461 (802) 482-2255 Itj{@nrgsystems.com www.nrgsystems.com a Energy GE Energy: GE Energy is one of the world’s leading wind turbine suppliers. With wind manufacturing and assembly facilities in Germany, Spain, China, Canada and the United States, our current product portfolio includes wind turbines with rated capacities ranging from 1.5 to 3.6 megawatts and support services ranging from development assistance to operation and maintenance. Contact: Kristin Schwarz Marketing Commun. Manager 1 River Road Schenectady, NY 12345 (518) 385-7343 Kristin.schwarz@ge.com www.ge-energy.com/wind & FRIEDRICH, LLP Michael Best & Friedrich LLP: Michael Best is a Midwest based law firm that provides the sophisticated legal support that renewable energy developers and investors need to succeed, including cutting-edge advice and world-class service at Midwest rates. MICHAEL BEST Contact: John D. Wilson, Partner One South Pinckney St., Suite 700 Madison, WI 53703-4257 (608) 257-3501 jdwilson@michaelbest.com www.michaelbest.com Information listed as provided by sponsoring companies. GOLD MEDIA SPONSORS RENEWABLE ENERGY Renewable Energy Access: offers business and Contact: Dan Harper Individuals comprehensive Daily and Weekly e-news, Customer Development Podcasts and marketplace services including: 375 Jaffrey Road brand-focused banner and e-newsletter advertising, Peterborough, NH 03458 online product and company directories, job postings, (603) 924-4405, 211 events calendar, interactive news commentary, dan@renewableenergyaccess.com technology basics and much more. www.RenewableEnergyAccess.com north american WiINDSPOWER www.nawindpower.com Contact: June Han North American Windpower: In print and online, PO Box 2180 North America’s leading wind information resource. Waterbury, CT 06722 (203) 262-4670, 226 han@nawindpower.com www.nawindpower.com Information listed as provided by sponsoring companies. MEGA-WATT SPONSORS <2 BABCOCK & BROWN Babcock & Brown: Babcock & Brown is a global investment and advisory firm specializing in asset oriented financing, investment and management, and one of the largest producers of wind energy in the world. Contact: Hunter Armistead 2 Harrison Street, 6” Floor San Francisco, CA 94105 (415) 618-3334 Hunter.armistead@babcockbrown.com www.babcockbrown.com CHADBOURNE & PARKE LLP Chadbourne & Parke LLP: Chadbourne & Parke LLP is one of the world's leading law firms serving the renewable energy industry, providing legal support in development, tax structuring, finance, equipment supply, construction, permitting and legislative matters. Contact: Edward W. Zaelke, Managing Partner 350 S. Grand Avenue, Suite 3300 Los Angeles, CA 90071 (213) 892-2012 ezaelke@chadbourne.com www.chadbourne.com ATTORNEYS AT LAW Stoel Rives LLP: Stoel Rives LLP is a leading law firm serving wind and other renewable energy developers and investors in projects throughout the United States, Canada and overseas with lawyers involved in various aspects of wind projects totaling more than 5,000 MW in the United States alone. Contact: Robert Van Brocklin, Partner 900 SW Fifth Street, Suite 2600 Portland, OR 97204 (503) 224-3380 rdvanbrocklin@stoel.com www.stoel.com Vestas No.1 in Modern Energy Vestas Americas: Vestas is the world leader in wind technology, with a core business comprised of the development, manufacture, sale, marketing, and maintenance of wind power systems that use wind energy to generate electricity. Contact: 1881 SW Naito Parkway, Suite 100 Portland, OR 97201 (503) 327-2000 vestas-americas@vestas.com www.vestas.com Information listed as provided by sponsoring companies. o Shermco Industries One Line. One Company. Company. Shermco Industries: Shermco Industries is the leading field service and generator repair facility in North America. Contact: Scott Meador 2425 E. Pioneer Drive Irving, TX 75061 (972) 793-5523 smeador@shermco.com www.shermco.com > enXco ss An EDF EN Company enXco, Inc.: enXco, an EDE EN Company, is a wind industry leader in development, operations and maintenance, and asset management for over 20 years. Contact: Donna Lotz PO Box 581043 N. Palm Springs, CA 92258 (760) 329-1437 donnal@enxco.com www.enxco.com Airtricity Airtricity, Inc.: Airtricity is a world leading renewable energy company developing and operating wind farms in the United States, Canada, Ireland, Scotland, England, and Wales. Contact: Nicole Herbert, Business Facilities Manager 401 N. Michigan Avenue, Suite 3020 Chicago, IL 60611 (312) 923-9463 nherbert@airtricity.com www..airtricity.com bp alternativenergy” ‘5 BP Alternative Energy NA Inc.: BP Alternative Energy provides low and zero-carbon electricity from solar, wind, hydrogen and gas-fired power generation. Contact: 700 Louisiana Street, 33 Floor Houston, TX 77002 bpalternativenergy@bp.com www.bpalternativenergy.com DICKSTEINSHAPIROuw- Dickstein Shapiro LLP: Dickstein Shapiro LLP is a multi- service law firm with more than 400 attorneys in Washington, DC, New York, and Los Angeles, representing clients that include more than 100 of the fortune 500 companies, as well as start-up ventures and entrepreneurs, multinational corporations, leading financial institutions, charitable organizations, and government officials. Contact: Larry F. Eisenstat Partner and Energy Practice Leader 1825 Eye Street NW Washington, DC 20006-5403 (202) 420-2224 eisenstat@dicksteinshapiro.com www.dicksteinshapiro.com Information listed as provided by sponsoring companies. on Zaz oma TPI Composites, Inc.: TP! Composites is a leading manufacturer of large-scale composite structures for the wind energy, transportation and military vehicle markets. Contact: Steve Lockard, President & CEO 373 Market Street Warren, RI 02885 (915) 494-6162 slockard@tpicomposites.com www.tpicomposites.com Suzlon Wind Energy Corporation: Suzlon is one of the world’s largest suppliers of wind turbines and has the strongest market capitalization, which has allowed the multi-national company to dramatically expand its installed capacity in the U.S., China, Australia and Europe. Contact: 8750 W. Bryn Mawr Avenue, Suite 720 Chicago, IL 60631 (773) 328-0588 info@suzlon-usa.com www.suzlon.com '*’ SECONDWIND KILO-WATT SPONSOR Second Wind Inc.: Since 1980, Second Wind has been providing products to wind energy developers, consulting meteorologists, and government agencies for wind prospecting and wind farm monitoring. Contact: Pat Quinlan 366 Summer Street Somerville, MA 12144 (617) 776-8520 pat@secondwind.com www.secondwind.com Information listed as provided by sponsoring companies. AWEA Wind Resource & Project Energy Assessment Workshop Exhibitor Contact Details ASC atmospheric systems corporation Atmospheric Systems Corporation Dr. Kenneth Underwood 24900 Anza Drive, Unit D Santa Clarita, CA 91355 Phone: (661) 294-9621 Email: Ken@minisodar.com Web: www.minisodar.com Dialight Dialight FAA Lights Jerry Ehlers 315 Cedar Ridge Drive Lake Villa, IL 60045 Phone: (847) 945-9911 Email: jehlers@dialight.com Web: www.dialight.com CAMPBELL’ SCIENTIFIC, INC. Campbell Scientific, Inc. Emilie Stewart 815 West 1800 North Logan, UT 84321 Phone: (435) 753-2342 Email: estewart@campbellsci.com Web: www.campbellsci.com Garrad Hassan America, Inc. Taylor Greer 11770 Bernardo Plaza Ct., Suite 209 San Diego, CA 92128 Phone: (858) 451-7013 Email: taylor.greer@garradhassan.com Web: www.garradhassan.com CERMAK CG ie 8 PETERKA PETERSEN CPP, Inc. Rick Damiani 1415 Blue Spruce Drive Fort Collins, CO 80524 Phone: (970) 221-3371 Email: rdamiani@cppwind.com Web: www.cppwind.com - meteorology & dynamics Meteodyn Aurélien Chantelot 75, bd Alexandre Oyon 72100 Le Mans, France Phone: +33 (0) 243 862 124 Email: aurelien.chantelot@meteodyn.com Web: www.meteodyn.com Information listed as provided by exhibiting companies. AWEA Wind Resource & Project Energy Assessment Workshop Exhibitor Contact Details NRG SYSTEMS NRG Systems Larry Jacobs 110 Riggs Road Hinesburg, VT 05461 Phone: (802) 482-2255 Email: Iti}@nrgsystems.com Web: www.nrgsystems.com waa] PRECISION =| WIND Precision Wind — an Energy Unlimited Company Peder M. Hansen 840 N. Telshor Suite A-415 Las Cruces, NM 88011 Phone: (402) 319-4649 Email: phansen@eui-tech.com Web: www.precisionwind.com natural power Natural Power Consultants Ltd. Mr. lan Locker The Green House, Forrest Estate Dalry, Castle Douglas Scotland, DG7 3XS Phone: +44 1644 430008 Email: ian.locker@naturalpower.com Web: www.naturalpower.com SECONDWIND Second Wind Inc. Pat Quinlan 366 Summer Street Somerville, MA 12144 Phone: (617) 776-8520 Email: pat@secondwind.com Web: www.secondwind.com proenis consulting USA Phoenix Consulting USA G. McDougall #103, 2710 — 3 Avenue NE Calgary, Alberta Canada T2A 2L5 Phone: (403) 248-9463 Email: mcdougallgc@phoenixengg.com Web: www.phoenixengg.com WIDACE WindSim AS Tine Volstad Jarlso 3124 Tonsberg, Norway Phone: +47 33 38 18 00 Email: tine@windsim.com Web: www.windsim.com Information listed as provided by exhibiting companies. ATTENDEE List 3S! eepueyy AWEA Wind Resource and Project Assessment Workshop PRE-REGISTRATION LIST (as of September 10, 2007) FULL NAME Greg Adams James Adams Ryan Adams Mark Ahlstrom Marie Ammerman Jeff Anthony _ Cory Arce Jeff Armbruster Bradford Axel Natalie Babij Bruce H. Bailey David Baker Nancy Baker David Balfrey Philip Barbour Michael D. Barnett Andrew Becker Glen Benson Daniel Bernadett Jean-Marc Bernier Ty Bettis Chris Bevil Chris Boeckman Peter Bower David Boyer Dennis Bradley Joseph Bradley Kenny Braun Katy Briggs Will Brilliant Steve Brink Mark Britton Grant Brohm Allan Broide Michael C. Brower Geoff Brown Lester Brown Chief Executive Officer Wind Resource Analyst President _ Senior Policy Analyst Chief Engineer _Senior Project Manager Operations Manager _ Project Engineer Sales Director Energy Analyst Information listed as provided on registration forms. TITLE Consultant Director of Business Development _ COMPANY \Chermac Energy Corporation _|AWS Truewind, LLC -/Garrad Hassan America, Inc. |WindLogics Inc. ,PPM Energy, Inc. [American W Wind Energy Association (AWEA) Meteorologist Manager - Utility Programs and Policy __ Project Coordinator __ Global Energy Concepts, LLC V Resou! UPC > Wind Management, LLC _ Shareholder “Stokes Lawrence P.S. Data Technician [Global Energy Concepts, LLC President |AWS Truewind, LLC _Phoenix Contact Inc. - USA “Public Power Council eee fH Staff Scientist [Acciona Wind Energy USA, LLC ‘Oregon State University - Mechanical Engineering Department Senior Principal Professional Kleinfelder |Second Wind Inc. AWS Truewind, LLC AWS Truewind, LLC AAT Inc. Portland General Electric Company Soe Puget Sound Energy ores \Global Energy Concepts, LLC |Ledcor Technical Services jenXco |Sealed Air Corporation ‘Windward Ways II LLC University of Washington \Vestas Americas _Global Energy Concepts, LLC _ Wind System Sales Senior Engineer GIS Technician Sales Manager Student Associate | ___Mazama Capital Management |PAR Electrical Contractors __WindLogics Inc. NW Area Manager Mesoamerica Energy “AWS Truewind, LLC \Garrad Hassan America, Inc. |Madison Valley Renewable Energ Business Developer Principal Partner Page 1 of 8 [LOCATION |Edmond, OK, USA |Milwaukee, WI, USA I _ |Seattle, WA, USA ‘Chicago, IL, USA Topeka, KS, USA |Albany, NY, USA Albany, NY, USA Portland, OR, USA |Seattle, WA, USA ‘Seattle, WA, USA Seattle, WA, USA ‘Albany, NY, USA [Albany, | NY,USA ‘San Diego, CA, USA [Saint Paul, MN, USA __ Portland, OR, USA Seattle, WA, USA _Newton, MA, USA ‘Seattle, WA, USA ‘Albany, NY, USA Middletown, PA, USA___ Portland, OR, USA Corvallis, OR, USA Somerville, MA, USA Gaspe, PQ, Canada Bellevue, WA, USA Mercer Island, WA, USA _ Portland, OR, USA ‘Woodinville, WA, USA Portland, OR, USA ‘Seattle, WA, USA ‘New York, NY, USA Portland, OR, USA _ Boring, OR, USA (Saint Paul, MN, USA ‘San Antonio De Belen, Costa Rica Portland, OR, USA [Los Angeles, CA, USA AWEA Wind Resource and Project Assessment Workshop PRE-REGISTRATION LIST (as of September 10, 2007) Information listed as provided on registration forms. FULL NAME Stefanie Brown Kelly Browne Mike Burghart Alex Byrne = Robert Campbell Alex Canta Jeff Carlson, Sr. Marni Carroll Alicia Cashman Ben Chambers Aurelien Chantelot MinCheng Jesper Christensen Steve Clark Owen Clay Peter Clive Kimberly Comstock Robert Conzemius Lance Cooke Graig Cooper Wayne Coste Alan Cowan David Cozzens Gennaro Crescenti Michael D. Cutbirth Martina Dabo Rick Damiani __ David S. Danner Danielle Davidian ____Manager TITLE ; Deputy Director Conference & Meetings Development Engineer _ Project Engineer __ _GIS Specialist Associate Business Developer Acting Site Assessment Manager Site Assessment Engineer Wind Energy Manager Wind & Site Engineer Director of Engineering Technical Development Officer Senior Scientist _Engineering Manager _Renewable Energy Program Manager Supervisor, Geographical Infor Western District Manager Meteorologist ‘President Wind Program Manager Sr. Engineer Technical Engineer Global Energy Concepts, LLC enXco ScottishPower - PPM ienXco COMPANY American Wind Energy Association (AWEA) Wind Prospect Group Ridgeline Energy, LLC Phoenix Engineering Inc. enXco _METEODYN _Southern California Edison Vestas Americas [NRG Systems NRG Systems, Inc. SgurrEnergy Ltd Global Energy Concepts, LLC WindLogics Inc. IETS Engineering Services _Competitive Power Ventures, Inc. (CPV) ISO New England Energy Trust of Oregon Sequoia Energy Inc. _Deublin Company PPM Energy, Inc. Champlin WindPower, LLC Alaska Energy Authority _Cermak Peterka Petersen, Inc. _ a _Westcon - Wind Turbine Too! Division RES America - Renewable Energy Systems Mike Delles Kleinfelder Brant DeMuth oeens | Z ______Mazama Capital Management EE Alex Depillis GIS Specialist EcoEnergy LLC Daniel Deslauriers President Aqua-Terra Maritime Nicolas Deve Marketing and Sales Engineer _LEOSPHERE Bjorn Doskeland J | LLL _____Windland, Inc. Los Mohit Dua Wind Resource Manager UPC Wind Management, LLC Rachel Dwarzski _Global Energy Concepts, LLC Tim Dwyer | Todd Eagleston Director, Renewables Development Nevada Power Company Las Vegas, NV, USA Pac £8 ‘Halifax, NS, Canada Escondido, CA, USA Portland, OR, USA _Rosemead, CA, USA San Francisco, CA, USA __ _ Fort Collins, CO, USA Austin, TX, USA Portland, OR, USA Gulfport, MS, USA ‘Boise, ID, USA LOCATION Los Angeles, CA, USA Seattle, WA, USA Seattle, WA, USA Calgary, AB, Canada Madison, WI, USA Escondido, CA, USA Escondid, CA, USA Le Mans, France Portland, OR, USA Hinesburg, VT, USA Hinesburg, VT, USA Glasgow, Scotland ‘Seattle, WA, USA Saint Paul, MN, USA Ridgefield, WA, USA Holyoke, MA, USA ‘Portland, OR, USA Winnipeg, MB, Canada Portland, OR, USA Portland, OR, USA Santa Barbara, CA, USA Anchorage, AK, USA Portland, OR, USA Boise, ID, USA Paris, France Newton, MA, USA Seattle, WA, USA Salt Lake City, UT, USA AWEA Wind Resource and Project Assessment Workshop PRE-REGISTRATION LIST (as of September 10, 2007) Information listed as provided on registration forms. FULL NAME Joseph Ebersbach Jerry Ehlers ul Mr. Dennis L. Elliott Sakura Emerine I Kenton Epard Qing Fang He Morgan Farley-Chrust William Fetzer Sean Finnerty Frank Flottemesch Andrew B. Flynn Erik Foley Padriac Fowler Cameron Fredkin Jeffrey Freedman Kenny French David Frost Jake Frye Liahna Fuller-Noon John Gaglioti Dave Gardner Taylor Geer Jeff Gessert Danny Gessman Pravin Ghantiwala Laura Goodfellow Jessie Gourlie Are R. Gravdahl_ Mark Green Errol Halberg Stuart Hall Richard Hamilton Jackie Hanberg Michael Hanratty Peder M. Hansen Dan Harris John Harris Susan Haupt _ Matthew Hazard ‘Principal Scientist _ Project Director Manager _ TITLE Director, Power Division Education Manager Student Director, Business Development Senior Vice President Technical Leader Client Service Manager Research Scientist Electric Coordinator VP, Strategy & Development Project Engineer Project Coordinator " Distict Sale Manager TET ‘Data Technician Area Manager Electrical Engineer Sales Project Developer COMPANY Henkels & McCoy _Dialight ¢ Corporation a National Renewable Energy Laboratory (NREL) American Wind Energy Association (AWEA) AirStream Energy, LLC ‘AES Wind Generation [Optical Air Data Systems _Competitive Power Ventures, Inc. (CPV), \Tacke Windenergie GmbH (Kleinfelder _Saint Francis University Renewable Energy Center Ls Power Group AWS Truewind, LEC Charles Messina Plumbing & Electric 'Precision Wind - An Energy Unlimited Company |Global Energy Concepts, LLC _Global Energy Concepts, LLC Deublin Company |Garrad Hassan America, Inc. \Global Energy Concepts, LLC |Henkels & McCoy |Henkels & McCoy, Inc. NRG Systems, Inc. [Third Planet Windpower, LLC __|WindSim AS ~_\Natural Power Consultants Ltd. ‘Phoenix Consulting USA, Inc. Natural Power Consultants Ltd. |Mountain Wind Energy LLC _ ‘Logistics _Development Consultant Marketing Director Business Development Manager Chair, Environmental Engineering ‘ODOT NEPA Coordinator |Global Energy Concepts, LLC __|New v Energy _ au _ [Energy Unlimited, Inc. _|Flash Technology PT Saint Francis University Renewable Energy Center _|Oregon | Department of Fish & Wildlife Page 3 of 8 _ Loretto, PA, USA _|San Diego, CA, USA LOCATION San Dimas, CA, USA Lake Villa, WA, USA |Golden, CO, USA Washington, DC, USA _ ‘Denver, CO, USA ___San Diego, CA, USA ‘Salt Lake City, UT, USA Manassas, VA,USA | ‘Braintree, MA, USA (Salzbergen, Germany Golden, CO, USA Lubbock, TX, USA Chesterfield, MO, USA Albany, NY, USA Dover, DE, USA ‘Portland, OR, USA Seattle, WA, USA Seattle, WA, USA Santa Monica, CA, USA Portland, OR, USA Seattle, WA, USA ‘Portland, OR, USA San Dimas, CA, USA Hinesburg, VT, USA Cheyenne, WY, USA Tonsberg, Norway Vancouver, BC, Canada _ ‘Chestermere, AB, Canada _ DUMFRIESHIRE, United Kingdom iReno,NV,USA ‘Seattle, WA, USA ‘Austin, TX, USA _ Omaha, _ NE, USA |Franklin, STN; USA) ‘Loretto, PA, USA (Bend, OR,USA Washougal, WA, USA AWEA Wind Resource and Project Assessment Workshop PRE-REGISTRATION LIST (as of September 10, 2007) Information listed as provided on registration forms. FULL NAME Jordan Hemaidan Matthew Hendrickson _ Steven J. Herzog Tom Hiester Ormand Hilderbrand Saskia Honhoff_ Tim Hoopes Bobby Houck Greg Howard Cheng-Hu Hu Holly Hughes Tim Hughes Steven Hunter Brian Hurley Stephanie Hylen Zack L. Irons Larry Jacobs Pramod Jain Blatchford James Raymond Janssen James Jensen Clint Johnson _ Jessica Johnson Erin Johnston | Mose Joiner Bruce Jones Steve Jones Steve Jones Hans Ejsony Jorgeusen Ivona Kaczynski Michael Kelly Robert Kelly llya Khripko Kitty Kilmer Rebecca King Jack Kline Vijayaut Kumar Marc Leblanc Design Engineer Vice President, Development | President Ar. ‘Data Technician TITLE Senior Appraiser Mechanical Engineer Project Information Manager Project Engineer _Meteorologist _ Research Meteorologist Technical Services Manager President Sales Engineer Chief Technology Officer Sr. Policy Issues Rep Project Manager Asst. Project Manager __Program Coordinator - Renewable Energy _ _Program Manager, Renewable Energy (President Director of Utility and Investor Services Director, Midwest Region PhD, Senior Researcher ject Coordinator ‘Director of Operations Service Commercialization and Distribution ‘Director, Portfolio Analysis student Consulting Meterorologist Wind Resource Assessment Analy. _Garrad Hassan America, Inc. __ _|Energy Trust of Oregon ‘COMPANY Michael Best and Friedrich LLP (Horizon Wind Energy ‘ASD - Department of the Interior Noble Environmental Power, LLC ‘Oregon Trail Wind Farm LLC \GE Energy [NRG Systems, Inc. Generation Energy, Inc. : ‘AES Wind Generation Vestas Wind Systems A/S |Global Energy Concepts, LLC 'PPM Energy, Inc. ‘Bureau of Reclamation _ Airtricity ‘Global Energy Concepts, LLC Vermont Environmental Research Assoc NRG Systems, Inc. |Wind Energy Consulting and Contracting Inc. Caiso Wisconsin Public Service Corp. Alaska Energy Authority _PPM Energy, Inc. |Henkels & McCoy, Inc. Functional Systems, LLC Global Energy Concepts jenXco Riso National Laboratory Global Energy Concepts, LLC Horizon Wind Energy Precision Wind An Energy Unlimited Company Edison Mission Energy (EME) (CH2M Hil RAM Associates Fremantle Energy Holdings LLC Pac £8 ‘Houston, TX, USA ‘Walla Walla, WA, USA Leesburg, VA, USA _ Randers, Denmark ‘Denver, CO, USA ‘Seattle, WA, USA Folsom, CA, USA Green Bay, WI, USA ‘Portland, OR, USA ‘San Dimas, CA, USA | LOCATION Madison, MI, USA Portland, OR, USA Essex, CT, USA = Salzbergen, Germany Hinesburg, VT, USA San Diego, CA, USA Seattle, WA, USA Portland, OR, USA Clonard, Sandyford, Dublin, Ireland Waterbury Center, VT, USA Hinesburg, VT, USA Jacksonville, FL, USA Anchorage, AK, USA Portland, OR, USA Portland, OR, USA Portland, OR, USA Seattle, WA, USA Minneapolis, MN, USA Roskilde, Denmark Seattle, WA, USA Houston, TX, USA West Conshohocken, PA, USA Boston, MA, USA Afton, MN, USA \Bellevue, WA, USA Brentwood, CA, USA ‘Austin, TX, USA Garrad Hassan & Partners Ltd. Bristol, United Kingdom AWEA Wind Resource and Project Assessment Workshop PRE-REGISTRATION LIST (as of September 10, 2007) Information listed as provided on registration forms. FULL NAME David LeMieux Geoffrey Lewis lan Locker Abby Lunstrum John Lyons Verene Lystad Bruce MacLennan Randy T. Manion Christina Marcello Trent Markell Michael Markus — Valerie Marquis Marina Martin-Tretto Julie Massey _ Beth Mast David Matson Patrick McAllister Christopher McAloon Elizabeth McCoy Scott McDonald Matt Mcloughlin Barton N. Merle-Smith Charles Messina Phil Metcalfe Matthew Meyers. Brent Miller John Miller Thomas Mills Chris Moehrl James Moorman — Pedro Mulero Barry Neal Murrianna Neessen Jim Newcomb __ Ron Nierenberg Chris Nuckols Andrew Oliver Calvin Olson _ Karen Olson Director, ZephiR Data Technician |Principal ‘Client Ser Services Director _Chief Meteorologist Director of Marketing and Sale National Sales Manager ‘Sales Engineer manager Wind Resource Engineer ,VP Technology _ TITLE ‘Energy Engineer Data Technician Field Technician Manager, Renewable Resource Pr. Test Engineer Data Technician Project. Coordinator Wind Resource Developer V.P. Technical Services Meteorology Manager Account Manager-Resource Assess. President _National Sales Manager Atmospheric Scientist Data Analyst Wind Site Engineer Meteorologist Consulting Meteorologist Sales Director _ Global Energy Concepts, LLC Global Energy Concepts, LLC _ _|NRG Systems, Inc. ‘COMPANY Montana Dept. of Environmental Quality Natural Power Consultants Ltd. Global Energy Concepts, LLC _Global Energy Concepts, LLC Western Area Power Administration |GE Energy Financial Services ‘Harris Group, I Inc. |AWS Truewind, LLC |Global Energy Concepts, LLC |Global Energy Concepts, LLC Iberdrola |Windots / V-Bar, LLC jenXco |Noble Environmental Power, LLC McCoy Alternative Power Reports [PPM Energy, Inc. _ Second Wind Inc. (Charles Massina Plumbing & Electric [Rohn Products _ |Airtricity ‘Clipper Windpower, Inc. |CH2M Hill Vestas Americas [Westwood Professional Services, Inc. _|CMC/ESP UTILITY PRODUCTS _ |PPM Energy, Inc. [RUD Chain, Inc. ‘SummitWindLLC _ |Ron Nierenberg [Navitas Energy, Inc. _ [RES America - Renewable Energy Systems -|Fremantle Energy ‘Holdings LLC |WindLogics Inc. Page 5 of 8 LOCATION Helena, MT, USA — [Ann Arbor, MI, USA __ Castle Douglas, Scotland, \United Kingdom |Seattle, WA, USA Seattle, WA, USA Seattle, WA,USA ‘New York, NY, USA Lakewood, CO, USA (Stamford, CT, USA ‘Denver, r, CO, USA Albany, NY, USA |Seattle, WA, USA Seattle, WA, USA — Seattle, WA, USA _ Radnor, PA, USA Pleasant Hill, CA, USA Escondido, CA, USA Essex, CT, USA (Salt Lake City, UT, USA ‘Clifton Park, NY, USA Somerville, MA, USA [Hinesburg, VT, U USA, Dover, DE, USA Peoria, IL, USA ‘Austin, TX, USA ‘Carpinteria, CA, USA Boston, MA, USA Portland, OR, USA Eden Prairie, MN, USA ‘Hamilton, OH, USA__ Portland, OR, USA ____ Kailua-Kona, HI, USA dgerton, MO, USA |Fort Collins, CO, USA ‘Camas, WA, USA Minneapolis, MN, USA _ Austin, TX, USA ‘Austin, TX, USA Saint Paul, MN, USA AWEA Wind Resource and Project Assessment Workshop PRE-REGISTRATION LIST (as of September 10, 2007) Information listed as provided on registration forms. FULL NAME Lucille Olszewski Barbara O'Neill Kirsten Orwig Elizabeth Osborne Steven Ostrowski, Jr. Marta Otermin Robert Owen, Jr. Kyle Owens" Tom Ozanich Paul Panagapka Remy Parmentier Peter Pawlowski Dana Peck Goncalo Pereira Pedro Rebecca Perry Elesha Peterson Carr Franco Petrucci Douglas L. Pfeister Rohit Poddar Robert Z. Poore Ram Poudel Greg Poulos Bill Prentice Patrick Quinlan Gordon Randall Bree A. Raum Michael Rea MattReed Meghan Reha Derek Reiber Mike Resca Mr. Kevin Rhodes Scott Richardson Nicholas Rigas Justin Robinson Zack Robinson Jeffery Roche Kevin Romuld TITLE _Meteorologist Director, Central Region _ Data Technician President ‘Consulting Eng & Meteorologist ; Project Coordinator _Owner/President _Manager _Project Manager Data Technician Workshop & Meetings Coordinator Meteorologist Senior Vice President/New Jers Chairman President Energy Assessment Specialist _ Meterologist_ Sr. Counsel _Director, Wind Systems Business __ _ Technical Analyst Manager of Grassroots Advocacy and _ WindPAC ‘Site Acquisition Specialist _ Data Technician ‘Manager Applications Engineer | Project Coordinator Vice President ____Clipper Windpower, Inc. (COMPANY ‘Blackwater Environmental Services LLC ‘Cermak Peterka Petersen Global Energy Concepts, LLC _LotusAutomation Acciona Wind Energy USA, LLC Global Energy Concepts, LLC 7 Madison Valley Renewable Energy LLC Rigarus Construction Inc. |LEOSPHERE _Competitive Power Ventures, Inc. (CPV) ‘Horizon Wind Energy Vestas Wind Systems A/S ‘Global Energy Concepts, LLC American Wind Energy Association (AWEA) Bluewater Wind VIPO Energy Resources Ltd. Global Energy Concepts, LLC Horizon Wind Energy Energy Northwest (Second Wind Inc. ‘Global Energy Concepts, LLC American Wind Energy Association (AWEA) Global Energy Concepts, LLC Vestas Americas _Competitive Power Ventures, Inc. (CPV) _Campbell Scientific, Inc. ‘S.C. Institute for Energy Studies (SCIES) Campbell Scientific, Inc Global Energy Concepts, LLC Vestas Americas ‘National Wind, LLC Pac £8 ‘Seattle, WA, USA LOCATION Pittsburgh, PA, USA Denver, CO, USA Fort Collins, CO, USA Seattle, WA, USA Vancouver, WA, USA (Chicago, IL, USA Middleton, WI, USA | ‘Seattle, WA, USA Los Angeles, CA, USA Elmira, ON, Canada Paris, France __ Braintree, MA, USA Portland, OR, USA Ringkobing, Denmark Seattle, WA, USA Washington, DC, USA Notre-Dame-lle-Perro, PQ, Canada ‘Hoboken, NJ, USA Pittsburgh, PA, USA Seattle, WA, USA Houston, TX, USA Denver, CO, USA Richland, WA, USA Somerville, MA, USA Washington, DC, USA Vancouver, WA, USA Portland, OR, USA ‘Seattle, WA, USA Portland, OR, USA Braintree, MA, USA Logan, UT, USA State College, PA, USA Westminster, SC, USA Logan, UT, USA Seattle, WA, USA Portland, OR, USA Grand Forks, ND, USA AWEA Wind Resource and Project Assessment Workshop PRE-REGISTRATION LIST (as of September 10, 2007) Information listed as provided on registration forms. FULL NAME TITLE COMPANY LOCATION Magdalena Rucker Senior Meteorologist Sea Breeze Power Corp. Vancouver, BC, Canada Jim Salmon |President paket E \Zephyr N North eee PEPE ere Burlington, ON, Canada__ Ron Sanderson Business Development Manager _ENMAX Vancouver, BC, Canada Adam Sandler _ |EIT A ee eee eae ett ____|Murray River, PE, Canada Walter Sass President & CEO Second Wind Inc. ‘Somerville, MA, USA Stephanie Savage | Pee eee [Entegrity Wind Systems Inc ____ Boulder, CO, USA Bruce Sellers Controls Engineer |Sustainable Automation, Inc. ‘Boulder, CO, USA Justin Sharp Manager, Wind Asset Management ____|PPMEnergy, Inc, Boat ____Portland,OR, USA Michael Sheriff. Strategic Planner |OGE Energy Corp. Oklahoma City, OK, USA David Simkins Marketing & Sales |NRG Systems, Inc. : |Hinesburg, VT, USA Rich Simon V-Bar, LLC / Windots, LLC Sausalito, CA, USA Rob Smart Vice President, Marketing&Co \Distributed Energy Systems +H ‘Barret, VT, USA Robert Sperhac Vice President _JPMorgan Capital Corporation \Chicago, IL, USA James Stalker Director & CTO Precision Wind An Energy Unlimited Company Las Cruces, NM, USA Sherri Stanton Vice President Tower Foundations, Inc. Livingston, MT, USA Phil Stiles Senior Product Engineering Man. Acciona Wind Energy USA, LLC \Chicago, IL, USA Kirk Story | Beaverton, OR, USA D. James Suit ‘Bozeman, MT, USA William Sutherland VP-Project Finance ‘Manulife Financial Corporation Toronto, ON, Canada Regina Sweet _ | He A Pee Beer _Garrad Hassan America, Inc. San Diego, CA, A, USA Frere Josef Tadich Project Engineer DNV fistlerup) copenhagen | Denmark Doug Taylor | Sept _ John DeereCredit Johnston, IA, USA John Thomas Vice President ‘The Cinnabar Companies ‘Tulsa, OK, USA Andrew Tindal _ _Director eee Pee ee _|Garrad ‘Hassan &PartnersLtd. (Bristol, United Kingdom Curtis Torrey Lab Technician ~ |Global Energy Concepts, LLC Seattle, WA, USA Jeremy Traurig [PPM Energy, Inc. Portland, OR, USA Dave Trofimenkoff Wind Resource Engineer (TransAlta Calgary, AB, Canada Kenneth H. Underwood Sr. Program Manager |Atmospheric Systems Corp. Valencia, CA, USA Javier Urbina AES Abilene, TX, USA Jaimes Valdez _Project Coordinator _____|Global Energy Concepts, LLC Seattle, WA, USA John Vanden Bosche Principal Engineer Chinook Wind ‘Everson, WA, USA Dave VanLuvanee _ Project Engineer _ SPEEEEPEg Eee Global Energy Concepts, LLC ae S352 Seattle, WA,USA John E. Wade John Wade Wind Consultant LLC Portland, OR, USA Melanie Wagner Data Technician _ eee Global Energy Concepts, LLC Seattle, WA, USA Stel N. Walker |Oregon State University - Mechanical Engineering Corvallis, OR, USA |Department | Kevin Walter |TradeWind Energy, LLC Lenexa, KS, USA Sandra Watts Seattle, WA, USA Stephen Weiner Environmental Manager 'West Coast Environmental and Engineering Woodland Hills, CA, USA Page 7 of 8 AWEA Wind Resource and Project Assessment Workshop PRE-REGISTRATION LIST (as of September 10, 2007) Information listed as provided on registration forms. FULL NAME TITLE COMPANY LOCATION Scott W. White | JW Prairie Windpower LLC Lawrence, KS, USA Scott Williams _ : (Project Manager - : (Puget SoundEnergy Bellevue, WA, USA John Wilson Attorney Michael Best & Friedrich LLP Madison, WI, USA Thomas Wind _ President Wind Utility Consulting Jefferson, IA, USA | Justin Wolfe __WARS Analyst ‘Airtricity ; ‘Austin, TX, USA DanWood ———_Consultant Ridgeline ee ______ Maple Falls, WA, USA Bassil Youakim Assistant Vice President GE Energy Financial Services Stamford, CT, USA Mark Young __ _ (ProjectEngineer Global Energy Concepts, LLC _ ___ Seattle, WA, USA Peter Young Development Manager Wind Capital Group ; : Madison, WI, USA Rana Zucchi Senior Technical Analyst Global Energy Concepts, LLC Seattle, WA, USA Pag £8 SPEAKER INFORMATION uONeUWOjU] JeyeedS SPEAKER BIOGRAPHIES [as of August 27, 2007] Jeff Anthony Manager of Utility Programs and Policy American Wind Energy Association 1101 14TH Street NW, 12th Floor Washington, DC 20005 Phone: (414) 967-5950 Fax: (202) 383-2505 Email: janthony@awea.org Jeff Anthony joined AWEA in March 2007. He is responsible for supporting utilities in their efforts to integrate and adopt wind power as a mainstream generation technology. He works with individual utilities across the U.S. as they expand their use of wind power, to aid them in their understanding of wind’s benefits, and to help them address integration and other implementation issues. Prior to joining AWEA, Jeff worked at Wisconsin Electric for 19 years, as the Manager of Renewable Energy Strategy. In that capacity, he was responsible for growing and accelerating the utility’s efforts in the renewable energy area and launched a number of new wind, solar, and biomass initiatives that led to recognition as one of the leading utilities in the country in terms of renewable energy adoption and advancement. Mr. Anthony graduated from Purdue University with a B.S. in Nuclear Engineering and received an Executive MBA from Northwestern University. Bruce Bailey President AWS TrueWind LLC 255 Fuller Road Suite 274 Albany, NY 12203-3640 Phone: (518) 437-8655 Fax: (518) 437-8659 Email: bbailey@awstruewind.com Bruce Bailey is the President of AWS Truewind, one of the world’s leading meteorological and engineering consulting firms serving the renewable energy industry. He specializes in wind resource assessment, siting, wind mapping, wind project planning, and short-term wind forecasting. Dr. Bailey and his firm are advising on over 10,000 MW of wind development throughout North America and abroad. Dr. Bailey has been active in wind energy for 30 years and has degrees in Meteorology and Engineering Management. He is also a Certified Consulting Meteorologist. Before forming AWS Truewind over 20 years ago, Dr. Bailey was a tenured research associate at the Atmospheric Sciences Research Center at the State University of New York at Albany. He has published or presented over 100 papers and is very engaged in training and outreach activities. Dr. Bailey works on behalf of most of North America’s leading energy development companies as well as for several state and federal agencies and major utility companies. He leads a staff of over 30 professionals, and his firm has conducted work in over 40 countries. Dr. Bailey is a past Board member for the American Wind Energy Association and chaired its Offshore Committee. David Baker President Phoenix Engineering Inc. 161 Lakeside Greens Drive Chestermere, Alberta Canada T1X 1B9 Phone: (403) 248-9463 Fax: (403) 272-1071 Email: bakerdr@phoenixengg.com David Baker began performing wind resource assessments in 1981 for the Alberta Research Council. He has installed meteorological equipment throughout Alberta and performed one of the first solar and wind resource assessments of the province of Alberta. Since 1985 Mr. Baker has owned his own consulting firm, Phoenix Engineering Inc., which provided wind resource assessments and wind facility design for private developers. During this period, Phoenix Engineering Inc. has been responsible for over 500 wind resource assessments, including the design and yield prediction of over 200 wind facilities. Mr. Baker's firm has designed approximately 40% of the operating facilities in Canada. Allen J. Becker Director V-Bar, LLC 6105 Thorncrest Drive Greendale, WI 53129 Phone: (414) 423-0206 Fax: (414) 423-0218 Email: ajlbecker@v-bar.net Allen Becker is a Director of V-Bar, LLC, (formerly Windots) and had a college and university teaching background prior to joining the V-Bar team in 1996. He is an expert in data base management and processes data from hundreds of meteorological towers monthly. Mr. Becker also specializes in internet climate research, data acquisition, site visits, turbine micrositing, power performance testing, and wind resource report preparation. Daniel W. Bernadett Chief Engineer AWS Truewind LLC 463 New Karner Road Albany, NY 12205 Phone: (518) 213-0044 x1006 Fax: (518) 213-0045 dbernadett@awstruewind.com Dan Bernadett has been working in wind energy for over 20 years. He has been with AWS Truewind for over 14 years and presently serves as Chief Engineer, where his role includes engineering review of all projects for which AWS Truewind provides energy production estimates (currently over 5000 MW per year). Mr. Bernadett is the project manager in charge of planning for construction of four wind farms in Ontario totaling 500 MW. He has a Master's degree in Mechanical Engineering and a Bachelor's degree in Aeronautical Engineering from the University of California, Davis, and is a Registered Professional Engineer in New York State. Michael Brower Chief Technical Officer AWS Truewind 463 New Karner Road Albany, NY 12205 Phone: (518) 213-0044 Fax: (518) 213-0045 Email: mbrower@awstruewind.com Michael Brower is Chief Technical Officer of AWS Truewind, one of North America’s leading wind energy consulting firms. His background is in physics; he holds a PhD from Harvard University and a Bachelors of Science from the Massachusetts Institute of Technology. Dr. Brower has been involved in many aspects of atmospheric modeling, including the development of AWS Truewind’s MesoMap and SiteWind systems. For several years, Dr. Brower headed AWS Truewind’s Department of Meteorology and Modeling Services, in which role he wrote the company’s Manual of Procedures for Energy Production Estimation. Dr. Brower now focuses primarily on research and development and quality assurance. Jesper Christensen Siting and Verification Engineer Vestas — American Wind Technology, Inc. 1881 SW Naito Parkway, Suite 100 Portland, OR 97201 Phone: (503) 327-2146 Fax: (503) 327-2001 Email: jchristensen@vestas.com Jesper Christensen is a Siting and Verification Engineer with Vestas — American Wind Technology, Inc., and has been with Vestas since 2001. His professional experience includes wind resource and energy yield assessment, micro siting, turbine load analysis, and measurement and verification of turbine power curves and sound yields. Mr. Christensen earned a Master of Science from Aalborg University, Denmark in 2002. Owen Clay Director of Engineering NRG Systems 110 Riggs Road Hinesburg, VT 05461 Phone: (802) 482-2255 Fax: (802) 264-1934 Email: owen@nrgsystems.com Owen Clay is responsible for the management of NRG Systems' Engineering Department, including new product development, manufacturing engineering support, and technology implementation companywide. He has 25 years prior experience in large commercial aircraft systems development at Goodrich Corporation. Mr. Clay Graduated in 1980 from the University of Vermont with a Bachelor of Science in Mechanical Engineering. Peter Clive Renewable Energy Consultant SgurrEnergy Ltd 79 Coplaw Street Glasgow G42 7JG United Kingdom Phone: +44 (0) 141 433 4675 Fax: +44 (0) 141 433 4647 Email: peter.clive@squrrenergy.com Peter Clive is a Renewable Energy Consultant with SgurrEnergy, where he has worked since completing a Ph.D. in Nuclear Physics in 2002. Dr. Clive’s data acquisition and analysis roles led to an interest in remote sensing, resulting in his participation in the IEA Expert Group drafting guidelines for SODAR and LIDAR; his inclusion in the technical working group of the UK’s first industry led LIDAR evaluation study; and the development of SgurrEnergy’s own LIDAR capabilities. In addition to his work with SgurrEnergy, Dr. Clive is a founding director of the company Counting Thoughts which recently collaborated on the AMEE software engine powering the UK government's carbon footprint calculator. Perhaps his most unusual exploit of recent years was repeating the 1774 “Schiehallion” experiment to determine the mass of the Earth, which was featured on the Discovery Channel and the BBC. Robert Conzemius Senior Atmospheric Scientist WindLogics, Inc 201 NW 4" Street Grand Rapids, MN, 55744 Phone: (651) 556-4285 Fax: (651) 556-4210 Email: bobc@windlogics.com Robert Conzemius is a senior atmospheric scientist for WindLogics, Inc. He received his master’s degree in meteorology from M.1.T. in 1990 and his Ph.D. in meteorology from the University of Oklahoma in 2004. He is an expert in boundary layer meteorology and is working to improve the representation of boundary layer physics for wind energy applications of atmospheric models. Dr. Conzemius has authored a half dozen refereed scientific journal articles and over a dozen conference presentations on the topics of boundary layer and mesoscale meteorology. Jerry Crescenti Manager, Meteorology PPM Energy 1125 NW Couch Street, Suite 700 Portland, OR 97209 Phone: (503) 796-6997 Fax: (503) 796-6907 Email: jerry.crescenti menergy.com Jerry Crescenti is a Lead/Senior Meteorologist and Manager with PPM Energy. He has extensive experience in meteorological observations and instrumentation and has authored or coauthored over 60 scientific publications. Prior to joining PPM Energy in April 2006, Mr. Crescenti held positions as a meteorologist with FPL Energy from 2003 to 2006; a research meteorologist in NOAA's Air Resources Laboratory from 1991 to 2003; and a research assistant at the Woods Hole Oceanographic Institution from 1988 to 1991. Mr. Crescenti has a B.S. in Earth Science from Southern Connecticut State University and a M.S. in Meteorology from Florida State University. Hans Ejsing Jorgensen Seniorforsker Wind Energy Department Ris@ National Laboratory Technical University of Denmark — DTU Building 118, P.O. Box 49 DK-4000 Roskilde, Denmark Phone: +45 4677 5034 Fax: +45 4677 5083 Email: hans.e.joergensen@risoe.dk Dennis Elliott Principal Scientist National Renewable Energy Laboratory 1617 Cole Boulevard Mail Stop 3811 Golden, CO 80401-3305 Phone: (303) 384-6935 Fax: (303) 384-6901 Email: dennis _elliott@nrel.gov Dennis Elliott is a principal scientist at the U.S. Department of Energy's (DOE) National Renewable Energy Laboratory (NREL), where he works primarily in the areas of wind resource characterization including projects for the development and validation of updated wind resource maps. He has been involved in wind characteristics research for more than two decades, and he has authored more than 70 publications including popular documents such as the "Wind Energy Resource Atlas of the United States". Prior to joining NREL in 1994, Mr. Elliott was at DOE's Pacific Northwest National Laboratory where he worked primarily in the wind energy program. Jeffrey Freedman Research Scientist AWS Truewind, LLC 463 New Karner Road Albany, NY 12205 Phone: 518-213-0044 Fax: 518-213-0045 Email: jfreedman@awstruewind.com In February 2007, Jeffrey Freedman joined AWS Truewind, LLC as a Research Scientist. In this key role he helps to further develop AWS Truewind’s research and development efforts with emphasis on improving observational and modeling techniques for wind energy assessment. Dr. Freedman’s primary research focuses on boundary layer meteorology. Most recently Dr. Freedman has been co-principal investigator on the National Science Foundation sponsored Hudson Valley Ambient Meteorology Study which addresses wind circulations in the Hudson Valley of New York. In addition, Dr. Freedman is engaged with the DOE NREL sponsored project, Development of Atmospheric Profiling and Environmental Modeling Techniques for Offshore Wind Energy, studying the offshore wind environment south of Long Island, NY. Dr. Freedman possesses both academic and professional degrees in meteorology and oceanography, atmospheric science and law. Prior to earning his Ph.D. at the University at Albany, State University of New York, Dr. Freedman was Assistant Counsel to the New York City Department of Environmental Protection. Dr. Freedman is a Certified Consulting Meteorologist whose breadth of experience includes seven years as principal for the meteorological consulting firm Atmospheric Information Services and EnviroLaw, an environmental and regulatory consulting firm. Matt Hendrickson Director, Energy Assessment Horizon Wind Energy 808 Travis Street, Suite 700 Houston, TX 77002 Phone: (713) 265-0380 Fax: (713) 265-0365 Email: Matthew.Hendrickson@horizonwind.com Matthew Hendrickson is Director of Energy Assessment for Horizon Wind Energy. Prior to joining Horizon, in 2003, Mr. Hendrickson was general manager of one of Houston’s top fine dining restaurants. He received his BS in Electrical Engineering from University of Houston in 2002. Tom Hiester Vice President, Development Noble Environmental Power 8 Railroad Avenue, Second Floor, Suite 8 Essex, CT 06426 Phone: (860) 581-5010 Fax: (860) 526-4404 Email: hiestert@noblepower.com Tom Hiester is Vice President of Development for Noble Environmental Power. He has over 30 years of experience in electric power project development with experience in wind, biomass, geothermal, landfill gas and natural gas in the US, Asia, Europe, and Central America. At Noble, Mr. Hiester leads the Project Development Support Group for wind resource analysis, environmental permitting, interconnection, and GIS mapping. In addition, he is the lead developer for the Great Lakes region, negotiates turbine supply agreements, and negotiates from time to time certain project acquisition agreements. Prior to Noble, he led the acquisition, development and construction of 380 MW of wind projects built as rate-based assets for Puget Sound Energy. Mr. Hiester holds a BA in Physics from Colorado College and a Master of Science in Atmospheric Science from the University of Washington. Saskia Honhoff Power Performance Engineer GE Energy Holsterfeld 16 48499 Salzbergen Germany Phone: 0049 5971 980 1199 Fax: 0049 5971 980 2199 Email: saskia.honhoff@ge.com Saskia Honhoff has worked with GE Energy in Germany since 2003, beginning in site-specific loads analysis, and later as a power performance validation engineer. In 2005, Ms. Honhoff graduated from the University of Twente, receiving the prize for the best mechanical engineering masters thesis written that year in the Netherlands. Rob Istchenko Phoenix Engineering Inc. Calgary, Alberta Canada Phone: (403) 250-2436 Email: istchenkora@phoenixengg.com Rob Istchenko has 5 years of experience in the wind industry including site assessment, statistical analysis, and project design. Prior to joining Phoenix Engineering he held positions with Environment Canada and Shell Canada's wind group. He holds an Engineering Physics degree from McMaster University. Clint Johnson Manager, Wind Farm Projects Group Garrad Hassan America, Inc. 333 SW 5th Avenue, Suite 400 Portland, OR 97204 Phone: (503) 222-5590 Fax: (503) 224-3563 Email: clint.johnson@garradhassan.com Clint Johnson is a Wind Energy Engineer with Garrad Hassan America, Inc. As Manager of the Wind Farm Projects Group, he oversees the provision of resource assessment, energy production analysis and wind power project design services to clients in the US market. Mr. Johnson received a Bachelor's degree in Physics from The Colorado College and a Master's degree in Mechanical Engineering from the University of Massachusetts at Amherst. Stephen Jones Director of Utility & Investor Services Global Energy Concepts, LLC 1809 7th Avenue, Suite 900 Seattle, WA 98101 Phone: (206) 387-4200 x232 Fax: (206) 387-4201 Email: siones@globalenergyconcepts.com Stephen Jones is Director of Utility and Investor Services for Global Energy Concepts, a leading wind energy consulting firm which provides analysis, design, testing, and management services for a wide range of clients. Mr. Jones has been working in the engineering consulting and wind power fields since 1986. His work has focused on technical and economic assessments related to facility and system improvements, facility siting and permitting, and evaluations of alternative solutions. He has participated at a variety of levels in independent engineering evaluations for potential investors in both proposed and operating wind power facilities. Mr. Jones holds a Bachelor of Science in Mechanical Engineering (with distinction) from Stanford University. ; Mike Kelly Director, Operations Services Horizon Wind Energy 808 Travis Street, Suite 700 Houston, TX 77002 Phone: (713) 571-6640 Fax: (713) 571-6659 Email: Mike.Kelly@horizonwind.com Jack Kline Consulting Meteorologist 55 Cloverleaf Circle Brentwood, CA 94513-1460 Phone: (925) 240-7855 Fax: (925) 240-7881 Email: windmet@pacbell.net Jack Kline has been working in the wind energy industry since 1982, as meteorologist at US Windpower and then at Howden Wind Parks, and has been a consultant since 1989. He holds a Master of Science in Atmospheric Sciences from Georgia Tech and a Bachelor of Science in Meteorology from Florida State University. Marc LeBlanc Director Garrad Hassan Canada Inc. 130 Albert Street, Suite 912 Ottawa, Ontario K1P 5G4 Canada Phone: (613) 230-3787 Fax: (613) 230-1742 Email: marc.leblanc@garradhassan.com Since 2004, Marc LeBlanc, as a senior wind energy analyst, has been responsible for founding and establishing Garrad Hassan’s Canadian branch office, and subsequently managing and conducting due diligence reviews, energy production assessments, and development services of wind farm projects across Canada. Prior to this, Mr. LeBlanc was a senior engineer in Garrad Hassan’s head office in the United Kingdom, managing, directing and conducting wind energy assessments for wind farm developments in Canada, USA, UK, Ireland, France, Norway and Japan. Mr. LeBlanc has a graduate degree in Mechanical Engineering in Environmental Fluid Dynamics from the University of Waterloo, Canada and a bachelor’s degree in Mechanical Engineering from the University of New Brunswick, Canada. William Marder Vice President Fortis Capital Corporation 520 Madison Avenue New York, NY 10022 Phone: 212-340-5393 Email: william.marder@us.fortis.com David F. Matson Director V-Bar, LLC 1650 St. Lawrence Way Pleasant Hill, CA 94523 Phone: (925) 933-7505 Fax: (925) 933-7505 Email: dfmatson@v-bar.net David Matson is a Director of V-Bar, LLC (formerly Windots) and has been a part of the V-Bar team since 1983. His specialties are data management, software development, instrumentation, and power performance testing of wind turbines. Mr. Matson has processed meteorological data from thousands of stations around the world. He has written a complete software library for comprehensive wind energy analyses. He also specializes in site visits and resource reports. He has installed hundreds of wind monitoring towers and has trained crews for clients to meet state-of-the-art specifications. Ron Nierenberg Consulting Meteorologist 850 NW View Ridge Court Camas, WA 98607-9051 Phone: (360) 210-4066 Fax: (360) 210-4039 Email: ron1230@comcast.net Ron Nierenberg has been working as a wind energy meteorologist for 30 years. In 1978, while working for PGandE he co-authored the original resource assessment report for the Altamont Pass. This was published by the California Energy Commission in 1980 and lead to the installation of 7000 wind turbines in the 1980s. In the late 1980s he conducted the largest publicly funded array loss measurement study, which was published by NREL. Mr. Nierenberg has been consulting for the largest wind farm developers in the United States for nearly 3 decades and is responsible for siting and assessing several thousand MW of wind turbines, in the U.S. and overseas. Andy Oliver Renewable Energy Systems 9050 N. Capital of Texas Highway Suite 390 Austin, TX 78759-7288 Phone: (512) 708-1538 Fax: (512) 708-1757 Email: andrew.oliver@res-americas.com Andrew Oliver has been with Renewable Energy Systems for 9 years and as the Vice President of Technologies & Resource Analysis, Dr. Oliver manages all aspects of wind resource assessment, turbine layout design, energy yield assessment and G.1.S. His team is ultimately responsible for ensuring that all RES projects can be financed from the wind resource perspective. Dr. Oliver holds a Bachelor's degree in Aeronautical Engineering and a Ph.D. in Wind Turbine Aerodynamics from the City University in London. Dana Peck Project Manager Horizon Wind Energy 808 Travis Street, Suite 700 Houston, TX 77002 Phone: (713) 265-0350 Fax: (713) 265-0365 Email: dana.peck@horizonwind.com Dana Peck joined Horizon Wind Energy in 2005 as Project Manager for the Kittitas Valley Wind Power Project; he's based in the Portland and Ellensburg offices. Immediately prior to Horizon, he spent eight years as Economic Development Director for Klickitat County, Washington, where he oversaw creation of the county's renewable energy overlay zone. Mr. Peck’s previous energy industry and government experience include project manager at Kenetech, strategic planning manager for Pacific Power, executive director of the 40- university Southeastern Universities Research Association, and director of government relations for the Western Solar Utilization Network. He has also held Congressional staff positions in both the US Senate and House. Robert Z. Poore President Global Energy Concepts, LLC 5729 Lakeview Drive, NE, Suite 100 Kirkland, WA 98033 Phone: (206) 387-4200 Fax: (206) 387-4201 Email: rpoore@globalenergyconcepts.com Robert Poore is President of Global Energy Concepts, LLC. Mr. Poore has more than twenty-five years of experience in the design, testing, and analysis of wind energy systems. He has been the investor’s independent engineer on multiple projects and led many due diligence investigations, engineering design and test programs, feasibility and economic analyses, wind resource assessment projects, research studies, and training programs. In addition to his technical skills, he is experienced with the financial and institutional considerations of wind projects and has served as an expert witness in arbitrations and litigations. His clients include financial institutions, turbine manufacturers, investors, utilities, insurance companies, law firms, and government agencies. In 1999, Mr. Poore received the AWEA award for technical excellence; in 2004, he served as Chair for Global WINDPOWER 2004. He has also served as a member of AWEA's Board of Directors, and is the chairman of the AWEA board’s Research and Development (R&D) committee. Mr. Poore holds a B.S. degree in Mechanical Engineering and an MBA in Finance and International Business. Patrick Quinlan Director, Wind Systems Business Development Second Wind Inc. 366 Summer Street Somerville, MA 02144-3132 Phone: (617) 776-8520 x21 Fax: (617) 776-0391 Email: patrick@secondwind.com Patrick Quinlan is Director of Wind Systems Business Development at Second Wind, Inc. in Massachusetts. He manages strategic deployment and customer relations for their Advanced Distributed Monitoring System (ADMS) and Enterprise Management Systems (EMS) products. Mr. Quinlan is a licensed Professional Engineer, with a BSME from the U-Massachusetts wind energy program and an MSME from the U-Wisconsin solar energy laboratory, where he developed time-series array loss and wind-hybrid system models. Earlier in Washington DC, Mr. Quinlan was an American Society of Mechanical Engineers Congressional Fellow supporting the House Science Committee member George Brown and later a White House Science and Technology Fellow, supporting the President's Science Advisor Neal Lane. Mr. Quinlan recently worked at NREL and for DOE contractors providing strategic planning for the Federal wind program and DOE Deputy Assistant Secretary. Patrick's prior wind experience includes 15 years of engineering and marketing consulting at Southern California Edison, NEOS Corporation, Wind Power Engineering, Sentech Inc., and AeroVironment, Inc. Gordon Randall Senior Technical Analyst Global Energy Concepts, LLC 1809 7th Avenue, Suite 900 Seattle, WA 98101 Phone: (206) 387-4200 Fax: (206) 387-4201 Email: grandall@globalenergyconcepts.com Gordon Randall is the manager of the Data Analysis Group for Global Energy Concepts, overseeing data collection, quality control, analysis, and reporting for approximately 800 wind resource assessment and testing towers worldwide. Mr. Randall has been involved with the completion of independent wind resource assessment and energy analyses for over 100 operational and proposed wind projects on behalf of developers, investors, and utilities. Mr. Randall will be leading GEC’s east coast office in Massachsetts, opening in early October. Justin Sharp Manager, Wind Asset Management Meteorology PPM Energy, Inc. 1125 NW Couch, Suite 700 Portland, OR 97209 Phone: (503) 796-7063 Fax: (503) 796-6907 Email: Justin. Sharp@PPMEnergy.com Justin Sharp joined PPM’s wind energy group in 2005, after gaining experience in the electric utility sector as an operational meteorologist in Bonneville Power Administration's Weather and Streamflow group. Justin is now manager of the Wind Asset Management Meteorology group at PPM. This group provides meteorological analysis and forecast services (generated in-house and by external vendors) that support the entire wind farm life cycle from Greenfield prospecting through operations. The group provides a bridge from the complex science of wind assessment and forecasting to the needs of commercial end-users. Justin is a specialist in the use of high resolution mesoscale numerical weather prediction models and an expert in Pacific Northwest forecasting. His Ph.D. dissertation explored the dynamics of wind flow through the Columbia Gorge. This has proved useful since the Gorge is the site of numerous wind farms. Richard L. Simon Director V-Bar, LLC 201 E. South Temple, Suite 826 Salt Lake City, UT 84111 Phone: (801) 712-6107 Fax: (801) 906-0119 Email: rlsimon@v-bar.net Richard L. Simon is a Director of V-Bar, LLC. He has worked in the wind energy industry since 1977, at which time he co-authored the first study of wind potential in California. He has directed research programs for utilities, state and federal agencies, and private developers. Mr. Simon played an integral role in the development of wind energy in California during the 1980's. Since that time, he has expanded his work across the United States and abroad. He started a wind energy consultancy in 1983, which has now become V-Bar, LLC. Mr. Simon is well-known and respected in the wind industry and focuses on all meteorological aspects of wind resource and turbine siting, having personally sited more than 5000 MW of operating turbines across the world. Mr. Simon holds a MS in Meteorology. 10 Joel Spenadel Executive Director JP Morgan 10 South Dearborn Street, 12th Floor Chicago, IL 60603-2021 Phone: (312) 732-3416 Fax: (312) 732-2231 Email: joel.spenadel@jpmorgan.com Phil Stiles Senior Product Engineering Manager Acciona Energy North America 101 North Wacker Drive, Suite 610 Chicago, IL 60606 Phone: (312) 673-3027 Fax: (312) 673-3001 Email: pstiles@acciona-na.com Phil Stiles is the Senior Product Engineering Manager for Acciona Energy North America. His responsibilities include fleet support engineering for Acciona’s renewable energy technologies in North America, principally the Acciona Windpower AW77 and AW82 wind turbines. Mr. Stiles joined the wind industry in 1999 as a site manager for NEG Micon, where he held a variety of positions including Applications Engineer, Regional Service Manager, and Asset Manager. Prior to joining the wind industry, Mr. Stiles worked as a mechanical analysis research engineer in the automotive and aerospace industries. He holds university degrees in Physics, Mechanical Engineering, Renewable Energy Engineering, and Business Administration. Pascal Storck President 3TIER 2001 6th Avenue, Suite 2100 Seattle, WA 98121-2534 Phone: (206) 325-1573 Fax: (206) 325-1618 Email: pstorck@3tiergroup.com As the President of 3TIER North America, Dr. Pascal Storck leads all aspects of the company’s operations. His expertise in systems analysis, optimization and hydrologic sciences played an integral role in the development and implementation of 3TIER Group’s unique range of wind, hydro, and solar products and services. During his doctorate research in physical hydrology at the University of Washington, Dr. Storck designed, implemented and monitored a detailed data collection and computer modeling system to investigate flooding events in mountainous watersheds. This work was honored for its scientific contribution by both the American Geophysical Union and the Western Snow Conference. Dr. Storck holds a bachelors degree with distinction in environmental engineering from Cornell University, a masters in civil and environmental engineering from the University of Illinois at Urbana-Champaign, and a PhD in civil and environmental engineering from the University of Washington, where he was also named a NASA Earth Systems Fellow. William Sutherland Vice President — Project Finance Manulife Financial Corporation 200 Bloor St E NT-4 Toronto, Ontario M4W 1E5 Canada Phone: (416) 852-3974 Fax: (416) 926-5737 Email: william _sutherland@manulife.com 11 William Sutherland is Vice President — Project Finance and heads the Project Finance Group at Manulife Financial Corporation. Mr. Sutherland is very well known within the Canadian independent power and U.S. renewable power industries. He and his group have been leading arrangers and providers of financing to the independent power sector over the past eight years. He is a seasoned corporate banker with twenty-six years business development, relationship management and corporate and project finance experience. He has been actively involved in numerous project and structured finance transactions within the power, mining and metals, forestry and pipeline industries and has considerable experience as an arranger, underwriter, lead lender and agent of highly structured transactions. Mr. Sutherland is a Professional Engineer (AEPO) and holds a BSc. (Mechanical Engineering) and MBA from Queens University. Andrew Tindal Director Garrad Hassan and Partners Limited St Vincent's Works Silverthorne Lane Bristol BS2 0QD United Kingdom Phone: +44 (0)117 972 9900 Fax: +44 (0)117 972 9901 Email: andrew.tindal@garradhassan.com Andrew Tindal is a director of Garrad Hassan and Partners Limited and has worked in wind energy for 18 years. Early in his career Andrew researched the influence of wind turbine wakes on the loading of wind turbines before moving to the fields of wind farm energy prediction, technical due diligence work, and short term forecasting for wind farms. He now has responsibility for the technical aspects of the short term and long term energy production services within Garrad Hassan. John Vanden Bosche Principal Engineer Chinook Wind 6571 Lunde Road Everson, WA 98247 Phone: (360) 398-2862 Fax: (360) 933-8070 Email: john@chinookwind.net John Vanden Bosche is a Principal Engineer with Chinook Wind. Mr. Vanden Bosche has 15 years of experience in the wind industry with involvement in a broad range of projects. He has worked on basic engineering research, wind turbine design and analysis, component manufacturing, structural and performance field testing, turbine verification, due diligence investigations, project planning and construction, commissioning testing, operations and maintenance, and design of retrofits. He also has experience with procuring land leases, collecting and analyzing wind resource data, and calculating project performance projections. He has lived and worked on wind energy projects in Tehachapi, California, Palm Springs, California, El Paso, Texas, and South Wales, United Kingdom. Past clients include turbine manufacturers, component vendors, project developers, electrical utilities, government laboratories, and law firms. Mr. Vanden Bosche is a registered U.S. Patent Agent with 10 years of experience obtaining patent protection on inventions representing a variety of technologies. He holds an M.S.M.E. from University of Texas-El Paso and a B.S.M.E. from West Virginia University. John Wade John Wade Wind Consultant LLC 2575 NE 32nd Avenue Portland, OR 97212 Phone: (503) 287-4329 Fax: 503-287-4279 Email: wade.j@comcast.net 12 John Wade is a meteorologist whose principal area of expertise is wind energy site selection and wind farm evaluation. During his 30 years in the wind industry as a wind consultant and scientist, Mr. Wade has been involved in the development of over 50 wind farms, village applications of wind energy, and a wide variety of applied meteorological investigations including: estimation of extreme wind speeds; use of vegetation as an indicator of wind energy potential; use of remote sensing for wind prospecting; cold climate wind prospecting; climate trends in the western United States; utility integration of wind energy; icing effects on transmission lines and wind turbine generators; corrosion impacts on wind turbines; and the use of Sodar for characterizing vertical wind variation. He is the principal author of two US Department of Energy reports: Biological Wind Prospecting and Remote Sensing for Wind Power Potential: A Prospectors Hand book. Mark Young Project Engineer Global Energy Concepts, LLC 1809 7th Avenue, Suite 900 Seattle, WA 98101 Phone: (206) 387-4200 x230 Fax: (206) 387-4201 Email: myoun lobalenergyconcepts.com Mark Young is a Senior Engineer and Testing & Measurements Manager with Global Energy Concepts. Mr. Young has eight years of engineering and project management experience in wind energy applications and related engineering fields. He currently leads the GEC Testing & Measurements group. His experience includes: power performance and loads measurement, acoustic noise measurement, aero-elastic simulation for loads and turbine performance prediction, and component design. He has conducted component design studies, economic analyses, wind resource analyses, site assessment, feasibility studies and other wind energy-specific studies and analyses. Mr. Young has extensive field service, installation, troubleshooting, and testing experience on wind equipment. He has conducted training courses on the installation of wind monitoring equipment and served as the field test engineer for numerous power curve, loads and acoustic measurements on wind turbines. Mr. Young holds a B.S. degree in Mechanical Engineering from the Pennsylvania State University. Mike Zulauf PPM Energy, Inc. / Iberdrola Email : Michael.Zulaufs@PPMEnergy.com 13 INTRODUCTION uononpo.ju To: AWEA Conference Attendees From: Randy Swisher, Executive Director, American Wind Energy Association RE: ANTITRUST COMPLIANCE PROCEDURES It is the policy and practice of the American Wind Energy Association (AWEA) and its members to strictly comply with all laws, including federal and state antitrust laws that apply to AWEA operations and activities. Compliance with the letter and spirit of the antitrust laws is essential if AWEA is to continue to operate at the highest standards of legal and ethical conduct. This policy and practice applies to this event, and in order to ensure compliance by attendees we have included with the program materials our Antirust Guidelines (the “Guidelines”). These Guidelines specifically address discussions among the attendees at AWEA events. All attendees are required to strictly adhere to the Guidelines and violators will be asked to leave the event. We ask that you review the Guidelines and if you have any question about the provisions stated therein please contact an AWEA staff member. American Wind Energy Association 1101 14" Street NW, 12" Floor, Washington, DC 20005, Phone: (202) 383-2500, Fax: (202) 383-2505, Web: www.awea.org “ae Antitrust Guidelines for Discussions at _ American Wind Zhergy Association The American Wind Energy Association Meetings It is extremely important that association members, meeting attendees, and speakers understand that the provisions of the antitrust laws regulate their conduct at association meetings. A thoughtless violation of the antitrust laws by a few members could result in expensive protracted litigation that could destroy the association and/or result in the prosecution of individual members. The most powerful Federal statute, the Sherman Act, provides substantial penalties for violation of the antitrust laws. Individuals can be fined up to $350,000 and imprisoned up to three years for violations. Corporations can be fined up to $10 million. In addition, defendants found guilty of violating the Sherman Act are subject to treble civil damages. oS = w = What You Can’t Do Do not enter into any agreements with competitors regarding or affecting prices. Do not discuss your company’s current price with competitors. Do not agree with competitors on pricing or profit levels. Do not agree with competitors to give or deny cash discounts or promotional allowances. Do not agree with competitors to give or deny credit to a specific customer, or to establish uniform credit terms. Do not agree with competitors to deal or not to deal with any customer or agree on the prices to be charged to a specific customer. Do not discuss allocation of markets. Do not enter into agreements with competitors’ price quotations or bids. What You Can Do Discuss better ways to educate and provide meaningful information to Association members about the industry. Discuss economic trends, business forecasts, and materials availability, emphasizing that each company is free to use this information in the way it sees fit and should make its own business decisions. Provide a properly structured environment for the exchanging of credit. Discuss Federal and State governmental actions and develop industry-wide lobbying efforts. Discuss technological advances and better ways to utilize them. Discuss ways to improve the public image of the industry. These guidelines have been prepared for the American Wind Energy Association by the Association’s antitrust counsel as part of the AWEA Antitrust Compliance Program. ACE ACORE AEP AGA AGC ALM ANPR APPA ASI ASOS ATC AWEA BA BACT Bef BLM BOP BPA Btu CAIR CAISO CAMR CanWEA CAPX CCGT CCl, CDEAC CDF CEERT CEQA Cf CF CFB CFC CFC CFD CH3Br CH;CCl, CH, co, CoD CPC CREB CT Ct DEP DHS DOD Wind Energy Acronyms Area Control Error American Council on Renewable Energy Annual Energy Production American Gas Association Automatic Generation Control Active Load Management Advance Notice of Proposal Rulemaking American Public Power Association Above Sea Level Automated Surface Observing System Available Transfer Capability American Wind Energy Association Balancing Area Best Available Control Technology Billion cubic feet (of gas) Bureau of Land Management Balance of Plant Bonneville Power Administration British Thermal Unit (1000 Btu = 1 cubic foot of gas) Clean Air Interstate Rule California Independent System Operator Clean Air Mercury Rule Canadian Wind Energy Association Capital Expenditures Combined Cycle Gas Turbine Carbon TetraChloride (an Ozone Depleting Substance) Clean and Diversified Energy Advisory Committee Computational Fluid Dynamics The Center for Energy Efficiency and Renewable Technologies California Environmental Quality Act Cubic foot (of gas) Capacity Factor Circulating Fluidized Bed National Rural Utilities Cooperative Finance Corporation ChloroFluoroCarbons (an Ozone Depleting Substance) Contract for Differences Methyl Bromide (an Ozone Depleting Substance) Methy! Chloroform (an Ozone Depleting Substance) Methane (1 of 6 Kyoto greenhouse gases) Carbon Dioxide (1 of 6 greenhouse gases) Commercial Operation Date Certificate of Public Convenience Clean Removable Energy Bond Combustion Turbine Thrust Coefficient Diurnal Energy Production Department of Homeland Security Department of Defense DOE DOT DSCR E10 E85 ECAR EE EHS EIR EIS ELCC ENSO EPA EPC EO! ERCOT EWTS FAA FERC FFT FRCC FTR GHG GL GOES GPS GWh GwP Halons HCFC HVDC HFC 1&C IAD IBL ICAP. ICT IE IEA IEC IEEE IGCC IPP INGA IRP 1OU IPO IPP Department of Energy Department of Transportation Debt Service Coverage Ratio 10% Ethanol 85% Ethanol East Central Area Reliability Coordinating Agreement Edison Electric Institute Employee Health & Safety Environmental Impact Report Environmental Impact Study Effective Load-Carrying Capacity El Nino / Southern Oscillation Environmental Protection Agency Engineering, Procurement, and Construction Expression of Interest Electric Reliability Council of Texas European Wind Turbine Standard Federal Aviation Administration Federal Energy Regulatory Commission Fast Fourier Transform Florida Reliability Coordinating Council Financial Transmission Right Greenhouse Gas Germanischer Lloyd (certification body) Geostationary Operations Environmental Satellites Global Positioning System Gigawatt Hour Global Warming Potential (CO, = 1 and SF, = 22,200) BromoFluoroCarbons (an Ozone Depleting Substance) HydroChloroFluoroCarbons (a greenhouse gas) High Voltage Direct Current Hydro Fluoro Carbons (1 of 6 Kyoto greenhouse gases) Instrumentation and Controls Inter-Annual Variation Internal Boundary Layer Installed Capacity Independent Coordinator of Transmission Independent Engineer International Energy Agency International Electrotechnical Commission Institute of Electrical and Electronics Engineers Integrated Gasification Combined Cycle Injury and Illness Prevention Program Interstate Natural Gas Association Integrated Resource Planning Investor Owned Utility Initial Public Offering Independent Power Producer IRR Iso JPO kPa kw kWh LC LD LIDAR LIOWI LLC LLP LNG LOLP LOT LOTO LPG LTA LVRT MAAC MAIN MAPP. Mcf MEP METAR MISO MMBtu MMS MOU MRO MWS MTEP MTSA MW MWh N2O NAO NARR NARUC NCEP NEPA NESC NESDIS NERC NEXRAD NIMBY NGSA NOx NOPR NPDES NRECA NREL NYISO NYSERDA Os Internal Rate of Return Independent System Operator Joint Program Office Kilopascals Kilowatt Kilowatt hour Letter of Credit Liquidated Damages Light Detection and ranging Long Island Offshore Wind Initiative Limited Liability Company Limited Liability Partnership Liquid Natural Gas Loss of Load Probability Letter of Intent Lockout-Tag out Liquid Petroleum Gas Lost Time Accidents Low Voltage Ride-Through Mid-Atlantic Area Council Mid-American Interconnected Network Mid-Continent Area Power Pool Thousand cubic feet (of gas) Monthly Energy Production Meteorological Aerodrome Report Midwest Independent System Operator Million Btu (equals 1 Mcf gas) Minerals Management Service Memorandum of Understanding Midwest Reliability Organization Meters Per Second MISO Transmission Expansion Plan Master Turbine Supply Agreement Megawatt Megawatt hour Nitrous Oxide (1 of 6 Kyoto greenhouse gases) North Atlantic Oscillation North American Regional Re-analysis National Association of Regulatory Utility Commissioners National Centers for Environmental Prediction National Environmental Policy Act National Electrical Safety Code National Environmental Satellite, Data, and Information Service North American Electric Reliability Council Next Generation Weather Radar Not-IN-My-Back-Yard Natural Gas Supply Association Nitrogen oxides Notice of Proposed Rulemaking (by the FERC) National Pollutant Discharge Elimination System National Rural Electric Cooperative Association National Renewable Energy Laboratory New York Independent System Operator New York State Energy Research and Development Authority Ozone (a greenhouse gas) O&M OATT ODS OEM OPX P&L PCIP PDO PEIS PFC PIP PJM PM POT PPA PPE PSD PSI PSIA PSIG PSS PTC PUC PUHCA PURPA PVC Qc Quad RADAR RASS RAWS RD REC REC RGGI RIX RMATS RMR ROl ROW RPS RRO RTO QF S/S SAR SCADA sD SERC SF. SN so2 SODAR SONAR SPP TCEQ Tef TI TSA Operations and Maintenance Open Access Transmission Tariff Ozone Depleting Substance Original Equipment Manufacturer Operations Expenditures Profit and Loss Principal Controlled Insurance Program Pacific Decadal Oscillation Programmatic EIS PerFluoroCarbons (1 of 6 Kyoto greenhouse gases) Public Involvement Program Pennsylvania-New Jersey-Maryland Interconnection Periodic Maximum Peak over Threshold Power Purchase Agreement Personal Protective Equipment Prevention of Significant Deterioration Pounds per square inch PSI absolute PSI gauge Preliminary Scoping Statement Production Tax Credit Public Utilities Commission Public Utilities Holding Company Act Public Utility Regulatory Policies Act PolyVinylChloride Quality Control Quadrillion Btu Radio Detection and Ranging Radio Acoustic Sounding System Remote Automated Weather System Rotor Diameter Renewable Energy Credit Rural Electric Cooperative Regional Greenhouse Gas Initiative Roughness Index Rocky Mountain Area Transmission Study Reliability - Must-Run Return on Investment Right of Way Renewables Portfolio Standard Regional Reliability Organization Regional Transmission Organization Qualifying Facility Substation Synthetic Aperture Radar Supervisory Control And Data Acquisition Standard Diviation Southeastern Electric Reliability Council Sulfur HexaFluoride (1 of 6 Kyoto greenhouse gases) Serial Number Sulfur Dioxide Sonic Detection and Ranging Sound Navigation and Ranging Southwest Power Pool Standard Error Texas Council of Environmental Quality Trillion cubic feet (of gas) Turbulence Intensity Turbine Supply Agreements UCAP Unforced Capacity (a modification of WFMS Wind Farm Management System Installed Capacity) WGA Western Governors’ Association € Uncertainty WRA Wind Resources Area URL Uniform Resource Locator WRE Wind Resource Grid USACE United States Army Corps Engineers WTG Wind Turbine/Generator USFWS US Fish & Wildlife Service UTM Universal Transverse Mercator UWIG Utility Wind Integration Group WAPA Western Area Power Administration and WECC Western Electricity Coordinating Council OVERVIEWS SMOIAJBAQ Data: What we need, why we need it and how we measure it. John Vanden Bosche Chinook Wind September 17, 2007 AWEA Wind Resource Assessment Workshop ae. DQ onvok Wind Wind Energy Consulting - Founded in 2001 -Main office in Everson, Washington «Seven people including Engineers -Meteorologist John Vanden Bosche - Principal Engineer - Data Analysts Mechanical Engineer +17 Years experience in wind energy -Broad range of clients and projects -Work tends to be for project developers, owners, operators, and banks -Has performed due diligence review or wind resource assessment for 8000 MW of proposed and operating wind projects Presentation Overview - Why do we need data? - Who needs it? + What data to collect? - How to collect it: - Sensor types - Wind shear measurements - Turbulence measurements - Towers - Data loggers Why Do We Need Data? Wind speed measurements: - provide an estimate of annual energy yield. Wind direction measurements: - Used for layout of turbines and wake loss estimates. Wind shear information = enables decisions about tower height. Wind turbulence, vertical wind flow, and other information: ~ Used for wind turbine loads and suitability calculations. Who Needs It? + Utilities - Developers - Banks and Investors + Turbine Manufacturers + Independent Engineers Ideally: - High sample rate measurements - 3 dimensions at hub height - At every wind turbine location - 30 years of data Bad idea: - Collect no data and rely solely on modeling Typical: 10 minute average sampling rate Cup anemometers at 2 or 3 heights, wind vanes at 2 heights ‘One 50 or 60 meter met tower every 2-4 km or for every 10 to 40 MW of Capacity, depending on terrain 1 to 3 years of data Sensor Types - Wind Speed » Cup anemometers- industry standard + Prop anemometers - not recommended, except for use as vertical anemometer + Sonic anemometers - expensive, best way to measure turbulence, not usually used for wind resource assessment campaigns Cup anemometers Max 40 - low cost, acceptable accuracy RISO - expensive, high accuracy ‘Thies First Class - expensive, high accuracy Vector - expensive, high accuracy Met One - possible compromise between cost and accuracy Ice Free: limited accuracy but it can reduce data loss during winter months Consistency between anemometers use for resource assessment and power curve measurement IEC Classification for Cup Anemometers - Based on a horizontal definition of wind - Factors include angular response, dynamic effects, and bearing friction + At least 2 anemometers must be tested for classification - Classification can vary depending on site conditions (turbulence, flow inclination, etc.) lee. free OsSant e— vl _conocte | Angular Response of Cup Anemometers RISO P2546 Vector A100 Thies First Class Bearing Friction Freon tome “108 We aa Classification Results Classification 1EC61400-12-1 All simulations Horizontal wsp definition Cup anemometer Class A Class 8 NRG max 40 06 to24 751083 Riso P2546 131019 5.0to8.0 Thies FC 151018 291038 Vaisala WAA151 1.61024 11.0 to 11.9 Vector L100 131018 4.01045 Class A applies to flat terrain Class B applies to complex terrain Classification result can be interpreted as uncertainty in measurements caused by anemometer operational characteristics Sensor Types - Other - Wind vane for measuring direction + Propeller or sonic anemometer for vertical wind - Temperature + Pressure + Relative humidity + Precipitation < - Leaf wetness we - Tipping bucket - Solar radiation Wind Direction Used for: - Creating wind rose and defining primary wind direction + Designing a wind turbine layout + Determining impact of obstacles - Correlating to other data sources Installation: + Assure proper alignment + Check magnetic declination + Heated sensors are available Vertical Wind Used for: - Determining flow inclination - Atmospheric Stability Installation Considerations: - Avoid tower shadow -Must be perfectly plumb Temperature + Used to determine site air density + Used for estimating downtime for high and low temperature shutdown + Helps to identify icing and freezing factors {nshall © OoWVn Pressure - Used to determine site air density - Indication of weather patterns Relative Humidity - Can be used to help detect icing - Used to determine site air density - Indication of environmental conditions Precipitation - Site weather conditions - Validation of SODAR data Solar Radiation + Site weather conditions - For comparison to Solar Energy Technologies KS Calibrations + Important for understanding & accuracy of measurements + May be required for financing + Traceable to NIST standards + Measnet standard for calibration of cup anemometers + Many analysts use consensus slope and offset for cup anemometers even if calibrated LO Wind Shear Measurements Shear is calculated by measuring wind speed at two different heights on a met tower, separated by at least 15 meters. Shear calculations can be used to extrapolate wind speeds to heights greater than met tower height. Anemometer booms at different heights should have similar orientation and tower shading effects. The lower measurement height should not be low enough to be affected by trees, obstacles, or ground effects Wind shear can be more fully characterized with use of SODAR or LIDAR. Remote Sensing I sopar - uses pulses of sound to measure three components of wind speed at multiple heights up to 200 meters. Current practice is to use SODAR together with a met tower. LIDAR - Uses a laser pulse to measure volumetric wind flow in a similar manner to SODAR, but some data fuggests that LIDAR might be accurate enough to use without a met tower. Both SODAR and LIDAR measurements require an ‘experienced operator and analyst to achieve useful results Turbulence Measurements Turbulence measurement is important to certify that a wind turbine model is appropriate to a specific site. Wake losses are also related to turbulence. Accurate measurement requires a low distance constant anemometer and a high sample rate (i.e. 1 Hz). Vertical turbulence can be as important as horizontal turbulence. SODAR and LIDAR are useful for characterizing 3D turbulence and variation of turbulence with height. Towers Hub height is preferred 40 to 60 meters is typical Tubular Lattice - Guyed - Freestanding Other Tower Considerations Booms and masts should meet requirements of IEC 61400-12-1 Mounting effects should be less than 1% FAA lighting may be required and should be designed to minimize interference with wind sensors Ice survival Installation- Best Practices Locations of met towers DOCUMENTATION!!! ~ North declination ~ Precise sensor heights - Boom orientations ~ Sensor serial numbers = Calibration sheets = Documentation of which sensor Is wited to which Se logger channel = ai = Nearby terrain and obstacles —— a = Site photos = — Documentation is also important when -—=— —— decommissioning a tower Correct wiring - especially for non-standard sensors Data Loggers + Specialized wind data loggers - NRG, Second Wind, AAT, others + General purpose loggers - Campbell Scientific + Older legacy data loggers - Wind runners, totalizers Data Loggers + Must be compatible with the type of sensors used + Must collect data in 10 minute averages with max, min, and standard deviation + Must have excellent atmospheric protection + Must have non-volatile memory + Must be battery powered Data Communication - Manual data collection + Chip collection - Land-line phone communication - Cell phone communication - Satellite communication 10 __American_Wind Znergy Association NN ALLEL! LE! LOTITO SLL L LL LLL LOT LE Data - Modeling Tuesday, September 18 8:45 am — 9:45 am Speaker: Michael Brower AWS Truewind, LLC The Role of Wind Flow Modeling and Mapping AWEA Wind Resource and Project Energy Assessment Workshop Portland, Oregon September 18, 2007 What is an atmospheric model? An idealized representation of how atmospheric parameters, such as the wind, evolve through time and space What are models used for? ¢ To extrapolate observations from a few masts to an entire project area ¢ To enable optimal design of a wind project ¢ To provide a basis for accurate energy production estimates <a AWS Truewind Typical Modeling Contribution to Uncertainty in Energy Production @ Low @ High To Model or Not to Model? That is not the question. ¢ It is usually impractical to measure at every proposed turbine location ¢ Thus nearly all wind assessments must use a model — even if it is a “mental” model ¢ The questions are: Which model? How should it be applied? And what is the corresponding uncertainty in speed and energy? 1 Types of Models ¢ Conceptual (mental) models ¢ Physical models (wind tunnels — not widely used) ¢ Numerical (computer) models * Combinations of the above... Conceptual Models Before PCs, conceptual models were the main tool Examples — “The observed wind at the mast is representative of the ridge line” — “In a mountain gap, the resource should be strongest near the exit.” — “The terrain is flat, so the resource should not vary significantly across the project area” Such models are often successful but must be customized to each site, and may not lead easily to quantitative results Even if not the main tool, conceptual models are essential for validating numerical models: Do the results of the simulation make sense? A good conceptual model is better than a bad numerical model — or a good numerical model wrongly applied Numerical Models ¢ Simpler equilibrium models —e.g., WAsP, MS-Micro, WindMap ¢ Complex equilibrium models —e.g., WindSim, Meteodyn WT — also known as CFD models ¢ Dynamic (non-equilibrium) full-physics models —e.g., MASS, MM5, WRF — also called mesoscale weather models je Combinations (e.g., SiteWind, KAMM-WaAspP) eae! Numerical Models AWS Thuewind solve at least some of the physical equations governing the atmosphere... Equilibrium Models Horizontal + Vertical Transport Transport Force Force Horizontal 4. Vertical Since only Transport Transport Sink: 7 4 t € = equrlt0< iw Ow dels + Pressure 4) Coriolis + Gravity + oe temperature at a point j j Pressure = Density X Temperature X (Gas Constant for Moist Air) SS AWS Trvewind Equilibrium v. Dynamic Models In equilibrium models, the flow is externally imposed and constant, and the model determines only those adjustments needed to satisfy mass and momentum conservation. In dynamic models, wind and other parameters are allowed to evolve under the influence by time-varying forcings, such as solar heating and evaporative cooling. Equilibrium models are fast and offer high resolution but tend to be limited to small regions where non-equilibrium internal forcings are minimal; dynamic models are slow but capture such effects, some of which may be important for wind projects Combinations of such models can be used re <= SS AWS Truewind Key Factors Influencing Near-Surface Winds _ © Topography © Surface roughness (friction) -* Static atmospheric stability (buoyancy) Thermal gradients a Wind Flow over Terrain AWS Truewind Simplified Equilibrium Model ae 8 2 = 2 & Flow Recirculation Behind a Ridge CFD Model WindSim (Vector AS) Surface Roughness “” [ ——Far Upstream |] oe —= ——Far Downstream Transition i —— SS Atmospheric Stability “° ~ Determines if the wind goes over or around an obstacle or is blocked Negative buoyancy : ° SS Mountain Blocking AWS Truewind Mesoscale Weather Model Mesoscale Weather Model Inputs ¢ All models require — Topography — Surface roughness — Site wind data ¢ Mesoscale weather models also require — Regional weather data — Detailed surface characteristics (soil moisture, albedo, etc.) WAsP aa Process oureur: Wino cuumdroLosy OF SPECIFIC LOCATION a a —————— AWS Truewind Meso-Microscale Process topography (e.g., SiteWind) roughness vegetation greenness eerie a = Typical Outputs AWS Troswind ¢ Wind speed map ¢ Wind resource grid (WRG format) _ © Speed/direction distributions _ © Used in plant analysis programs such as WindPro, WindFarmer, WindFarm AWS Truewind Estimating Modeling Uncertainty Direct validation requires at least ~5 masts in representative locations — rarely achieved! Absent enough masts, rely on — Case studies from similar terrain — Predicted range of variation among turbines Account for number and location of masts Beware biases introduced by poor mast placement Map Validation Example’ ~~ y = 0.9727x R? = 0.7848} ry a 9.0 8.5 8.0 75 7.0 Observed Mean Speed (m/s) 65 60 6.0 65 7.0 75 8.0 85 9.0 95 Predicted Mean Speed (m/s) ; | Standard error (o) ~ 3.5% - j Typical range for various situations and models 2% - 10% ieee Interpreting Map Errors ANS Tenn ¢ Not directly interested in the map error, but in the uncertainty (¢) in the array-average speed relative to the masts. ¢ For a given map standard error and an even distribution of M masts among the turbines: oO éx——— VM ¢ Ifall masts are outside the turbine array, then EXO AWS 05 ; Theoretical Dependence of Energy Production Uncertainty on Map Error and Number of Masts 11.0% 10.0% 9.0% 8.0% 7.0% Uncertainty in Energy Production 4 6 8 Number of Masts Mast Placement Guidelines ¢ Modeling errors (co) tend to grow with distance, so more masts are needed for larger projects ¢ To capture the benefit of multiple masts, all masts should be within the turbine array * To limit model biases, masts should be placed in locations typical of turbines — don’t choose only the best spots! scpoerniens —————| Conclusions AWS ewe * Modeling — conceptual or numerical — is a critical step in virtually all wind assessments * Numerical models are convenient and offer quantitative results — but they don’t remove the need for a good conceptual understanding ¢ Proper error analysis and interpretation are key ¢ More masts help reduce uncertainty — but only i if they are properly placed ™ __American_Wind Zhergy Association Data Analysis - Multiple Approaches to Evaluating Identical Data Tuesday, September 18 10:15 am — 11:15 am Speaker: Andy Oliver RES Americas, Inc AWEA RESOURCE ASSESSMENT WORKSHOP Tuesday September 17' 2007 “Data Analysis — Multiple Approaches to Evaluating Identical Data” Dr. Andy Oliver . Data Processing & Quality Control . Data Summaries & Reports . Predicting the Long Term Wind Climate Inter-annual variation and its implication * Selecting a reference station * Regression analysis & “MCP” . Extrapolating Wind Speeds to Hub Height The Three S’s! The Long Term Average Wind Speed is the single most important factor determining the cost of Wind Energy The wind speed Distribution Shape and the extrapolation of the wind speed from measurement height to wind turbine height, the Wind Shear, are also critical Additional Relevant Parameters The Wind Rose (direction distribution) is important for calculating wake (array) losses and topographic speed up factors Air Density (affects power curve) Turbulence Intensity (affects power curve and wake recovery) GOAL: Determine the above conditions at the wind turbine hub height at each meteorological tower DATA SUMMARIES & REPORTS The Raw Data Site Information a Channel Identification Information Data with or without calibration bea factors applied craw 207 164 18 146 Quality Assurance — Icing & & ° o é a Wind speed 6m {mvs} / Wind speed 30 m (mis Temperature |" a Initial checks carried out on data to remove periods of icing Quality Assurance — Instrument Failure Wind Speed (mvs) Initial checks carried out on data to remove instrument failures Quality Assurance — Tower Shadow MaralJSAMMMSh. Shear 1420287102008) 1437 070005 Rapn Arora pees Normalized Wind Speed Variation per Direction for MMM (MastID: xxx) Upper Anemometer at $0.0 metres Lower Anemometer at 35,0 metres nwa SAnmm 035.0 Data from 1600 11/04/2003 to 13:50 03/02/2005 in 19 wind blocks, total samples = 91503, T T NB. Eror bars indicate standard eran | degree tin averaged value ' 1 1 ' ir T 1 ' ' 1 Lower Anem Mean Wind Speed / Upper Anem Mean Wind Speed 1 1 T ' 1 1 1 1 1 1 | + Ideally tower i is designed so that upper instruments are not affected * Tower shadow must be corrected for + Redundant sensors are useful as these data can be substituted DATA SUMMARIES & REPORTS Data Summaries & wauiteler atime ID Monthly Mean Wind Speeds For XXXXX (Mast ID: 136 , Location: XXX.XX Deg West, < Deg North (NAD83) <ommmme LOCation Anemometer Height 56.0: Boom Length: 1.50: Boom Angle: 158 - Wind Vane Height $4.0: Boom Length: 2.00: Boom Angle: 158 Period Total Number of Trem 76949 10 min Periods Averaging Period Mean Wind Speed: — 7Al mis Annual Averige Wind Speed (90%): — 7.26 mis MONTHLY MEAN WIND SPEED (in M/S) FOR 2002 Feb} Mar] Apr] May} Jun Jul} Aug} Sep} Oct} Dec! Mean Speed] 7.49] Actual Items} 3 4464) Expected Items] 4464 % Success] labili 100.0} MONT! D (in M/S) FOR 2003 Jan Mar] Aug] Mean Speed| 09 7.66) 6.79] 6.12 Actual Items} ? i >> | 4320) 4464 ges ties Ice? Instrument Failure?? |/°) aga % Success} 100.0} 100.0] 99. 100.0} 100.0) 100.0) MONTHLY MEAN WIND SPEED (in MIS) FOR 2004 Mar] Apr]_May] Jun] Jul] Aug) Mean Speed 8.03] 7.89] 6.82 Actual Items 4464] 43201 370] Expected Items| 4464] 4320) 4464 % Success 100.0] 100.0] 8.3 Data Summaries & Reports - Monthly Analysis Program Info Average Seasonal Trend For XXXXX (Mast ID: 136) Data Fike: Version: 13.04 Anemometer Height $6.0: Boom Length: 1.50: Boom Angle: 158 ml36USA,000 300 Date: VEINS ‘Wind Vane Height $4.0: Boom Length: Boom Angle: 15 Date: 17/08/2008 Data from 18:30 15/11/2002 - 13:80 03/05/2004 (in 1S data Hocks) 2 == Means 2003 —pé=Means 2008 10000 9000 8000 000 4000 Number of tems of 10 min Data Weighted average wind speed determined for each calendar month (all years) 3000 Take average of these calendar month averages (red line) Apply wind speed to similarly determined wind speed distribution Average Mo From MI36USAxxxInfo025 1s, dated 1705/2004 Example: Mean wind speeds month showing seasonal variation Data Summaries & Reports - Time Of Day Analysis Program Info ‘Average Monthly Diurnal Wind Speed Trends Fer Data File: mi66USAju033 md [rome 1336 ‘Anemometer Height 50.0: Boom Length: 184: Boom Angle: 305 Dare | Date: 1112/2008 2 r ; st Wind Vane Height 48.0: Boom Length: 1 84: Boom Angle: 305 ‘Data fren 16:00 11/04/2003 - 12:50 19/10/2004 (ix 18 data blacks Average Hourly Mean Wind Speed /(/ 9 wo Hour OfDay (GMT-8 howrs (Pacific Time)) Example: Mean wind speeds by hour & month Data Summaries & Reports - Directional Analysis Program Info ‘Wind Rose For XXXXX (Mast ID: 166) Data File: mt 66USAxm0041 xt Version: 140 Anemometer Height 50.0; Boom Length: 1.84: Boom Angle: 185 Date: 06/04/2006 Date: 077022005 ‘Wind Vane Height 48.0: Boom Length: 1.84: Boom Angle: 185 Data from 16:00 11/04/2003 - 16:20 02/04/2006 (in 23 data bhocks) 385.5 MOIS, sis neues Sasa nes “(Ube ph ~ 2838 m5 7 N77 x Ls 3s 38.315 NA 455s ’ ; 7 100-30 j \ sos sss Bsns 1 y ‘002s ‘\ 100-20 28.05 i Vy ss ‘0020 ‘ oo.is INB: Fach tine represents the rss 1 asa Bo.10 pecan te ' 4 moos ninoontarege aad 268-275 ss95 nemometer Boom Jappropriate mvs interval in 1 ma: : the key above vu ‘Anemometer Boom Invers 255.268 8 5.105 \ | INEWind Vane Boom as.2ss +X ross \ i asus 7 sir Goid System UTM, Zone LI North, K , NAD27, Metre (UTM Zone = 11) nas 2 Yasias Mag To Gest Corr: 16.0 dog > L_ oN\7 Easting: 2s2s"~s YL i oc) SS Northing: nosis - 2 sss 198200 = ~ An Ais From M166USAjunlnfoO 22 xis, dated 0604/2006 188-195 les178 1 [wind Vane Boom Inverse Example: Wind speeds and percentage time by direction. The “Wind Rose” Program Info Version: 1334 Date: 11/12/2003 Frequency / 10 min Period: Data Summaries & Reports — Wind Speed Distribution Wind Speed Frequency Distribution For Data Fue Anemometer Height 56.0: Boom Length: 1.46: Boom Angle: 279 Date: 29/09/2004 : 10 25/09/2004 (im 31 data blocks) CALCULATED STATISTICS. | Mean (w/e) = 8.11 Total tems = 140679 “| Anmual Average (m/s) = 803 a3 at = sanan ‘Wind Speed Bin / (m/s) The frequency of occurrence of certain wind speeds PREDICTING THE LONG TERM WIND CLIMATE Why can’t | just measure for a year? The Average Wind Speed (hence energy output) in any given year can be considerably higher or lower than the long term average. 33.7% Capacity Factor 8 AVERAGE WIND SPEED [mvs] .. we Call this the “Inter-Annual Variation” inter Annual Variation in FG United States ~~ Gree Based on 40 years NCEP 200km surface data To be used as a guide only Conclusion On its own, one year of wind data is not enough to determine the long term energy production of the proposed wind farm * Depending on the value of the I.A.V., with 3 to 6 years of on-site wind data we are more likely to be within an acceptable range of energy uncertainty « We can’t wait this long to determine the long term resource, so we: Measure For at least one year Correlate With a long term reference station such as an airport Predict The long term wind resource Basic Concept of MCP Long Term Estimate Historic Estimate Site Measurements aN Foal \ Wind Farm Site Concurrent Period Relationship Reference Stn. —> Time Historic Reference Measurements Se Establish relationship with long term reference - E.g. Site is twice as windy as the reference station Apply relationship to historic reference data to determine historic estimate Weight the Historic Estimate and the Site Measurements appropriately to determine the Long Term Estimate Note that this is a hind cast not a forecast. It is what would have happened in the past. The implicit assumption is that the future will mirror the past 10 Reference Data Some Types of correlation data: Hourly wind speed and direction data 10 minute average at end of hour Daily or Monthly wind speed data 6 hourly Upper Air Measurements (Radiosonde) Wind speeds from a mesoscale model Power output from an operational wind farm! Some Sources of correlation data: + NOAA or Regional Climate Centers - ASOS (Automated Surface Observing System) - RAWS (Remote Automated Weather Station) - Radiosonde Department Of Transportation Power Plants (Dispersion monitoring) State Anemometer Loan Programs Many others — seek & ye shall find. Guidelines to Selecting a Reference Site Stating the obvious: Data must overlap with the site measurements & you must have a longer record than your site! Similar climatology - Generally means nearby - Mountain sites probably won't correlate with a valley reference Consistency - Must have been in same location for entire time - Same instrument heights - Same measurement system (ASOS introduced in mid-1990’s) Find out station history Good exposure ASOS generally are well exposed (next to middle of runway) - RAWS tend to be on western facing slopes and only 2 to 3 meters off the ground - Try to avoid reference stations with very low mean wind speeds 11 Reference Data Suitability Checks Monmy aan Wind Speeds MI Kautay Rterene Staton What to compare inet eee - Before & after system changes Til = - Historical & concurrent periods - Other reference stations Visual inspection - Trends in the data - Step-changes - Difference in wind roses Statistical checks (indicators) - F-test, t-Test, Mann-Whitney test Before you really get going... - Quick correlation: Any relationship? Some Correlation Methods Ratio of Means | Least Squares Orthogonal York Method Matrix Method Y=m.x Y=m.x + b Y=m.x + b Y=m.x +b Probabilistic Simple relative Minimizes error in Minimizes error Minimizes errors Transform windiness perpendicularly in both X & Y between site and reference stn. a | 2 | 3 3 a} ie #3 B897993999999329 BRIRIaaRaIeTRSTARTTZaaaTy 2 susan SSSSS22899223322322929999929999392 3 2 saoacgaca20e0020020900209 8989 Ra2 9 | savssoacasceccace2c9008 2889089820391 | sa9a33333399993999990989a9R9aE2 799 | PePEPEPErerererereerty (117) tT Cert Te sassasecaeaecaccussdsaszauzsze%20399° SO29R9R9999 222229 9RRRZ AERTS TEE zzzz5I eeevaeeez eR eRSeeezezaeezz7R999999991 eoaeeeeeeeeeesaaaaeaaRRzR99088 99999999999999999999999999 PPeEEESTEEPeeeesrrerererrrryy Se9999RRRR9RRR997999999222988R ee39993399399009399889 s82999999e 00999 9RRRRERIVIIFZEIII7I9) een PEIZIZIT 12 The relationship between the site and the reference station usually varies by wind direction We therefore tend to correlate the site with the reference station for multiple directions Normally we would perform a number of correlations around the compass (RES typically uses 12) Averaging Period Wind speed (m/s) Chart shows average wind speeds by time of day on the SAME mast Many regions exhibit this phenomenon (atmosphere more stable at night) Implies that wind farm site will not correlate very well with airport on an hourly basis A correlation of daily (or even monthly) averages may produce a better regression However, a longer averaging period implies less correlation data! Unwise to use airport data for Time Of Day energy distributions 13 Multi-Mast Predictions Wind Speed Distribution The shape of the wind speed distribution is important in terms of energy content and it varies from site to site It is important to remove seasonality as this can bias the shape of the distribution + Use predicted wind speed / direction frequency distribution if the correlation is very good + Else consider using seasonally adjusted wind speed / direction frequency distribution scaled to the predicted mean wind speed The shape of the wind speed distribution can change the relative economics of one turbine type versus another 14 é ® £ o c at} S 5 2 g 2 3 iS e 5 = a .. can have different energy content 39.5% |: Cf= Note: All distributions have been normalized to the same average wind speed 15 Wind Shear * Wind speed generally increases with height above ground (though not always, e.g. Altamont). This effect is referred to as Wind Shear * Wind Shear is critical to the turbine hub height that is selected * The ground coverage (roughness) changes the way the standard extrapolation is made. It is important to correct for this effect. ‘Wind shear’ describes how the wind increases (or decreases) with height above ground Veo =7.5"(80/60)"°° Vo = 8.18m/s a y, (4 .818 Extrapotating to 80m: CU) V, 1 A, Where V is wind speed, H is height and alpha is the wind shear exponent 8 Example: @ =Ln(7.5/7.1) / Ln(60/50) @=0.30 Height (m) 8 8 8 4 6 Wind Speed (m/s) To extrapolate correctly, ‘wind shear’ measurements need to be modified to take account of ground cover (e.g. trees) i yy H-d 4 215 Extrapolating to 80m: ee a7" 9) 00-28 Vy \H-d . Vso =7.5*(80-9160-9) Veo = 8.15m/s Where d is displacement height (equal to 3/4 canopy height) and a, is the revised wind ‘shear exponent * 7.50 iy =Ln(7.5/7.1) | Ln(60-9/50-9) Displacement height = 0.75 * Canopy height = Wind Speed (mm/s) Wind Shear Continued.. * Most in the industry use the exponential shear profile (above) rather than the logarithmic profile * Recommend measuring as close to intended hub height as possible - particularly in low wind speed regimes * Different flavors of shear: Exponent of means ‘Mean of exponents AWind shear changes with speed (& direction) Shear applied on a time series basis 17 Closing Thought * Part of the answer can be found by selecting and applying the right technique to the particular site and wind regime (all sites are different). It follows therefore that: * Part of the answer can be found by ensuring you have the right experience in your team Ultimately though, there is no question that: * Part of the answer can be found through applying common sense 18 __American_Wind nergy Association I eeerereneneeeemerenmenneetieneniammanesaineneineanenncmemamemmammemmamaaell Energy Yield Calculations Tuesday, September 18 10:15 am — 11:15 am Speaker: Thomas Hiester Noble Environmental Power, LLC BZ Noble ENVIRONMENTAL POWER Energy Yield Calculation fora Windfarm Tom Hiester AWEA Wind Resource Assessment Workshop September 18, 2007 Portland, OR Wind Power...the natural choice Noble ENVIRONMENTAL POWER Calculation of Energy For a single turbine we need to know just 3 things: 1. Wind Speed Distribution at Hub Height 2. Power Curve at correct average air density 3. Losses (availability, electrical, etc) For a wind farm we also need to know: 4, Wind speed variation across the site 5. Interaction from other turbines ( ‘Wake’ or ‘Array’ losses) For financing we need to know uncertainty in the energy estimates 6. Energy level that we expect to be exceeded x% of the time Noble rwroumewra ows Different Machines for Different Wind Regimes Power Curve Comparisons Class | (dashed) and Class Il (solid) Wind Turbines 3339993 9 8 lass | machines for high wind speed or turbulent site Percent of Rated Capacity g 3 4 5 6 7 B 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 2% 25 Wind Speed (M/S) Wind Speed Frequency Distribution Frequency (Hours) & 8 8 8 id 0 123 4 5 6 7 8 9 1011 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 26 29 30 Wind Speed (m/s) eZ Noble Overlay the Power Curve lume Wind Speed Distribution a8 Frequency (Hours) 8 Wind Speed (m/s) WZ Noble Se ME Wind Speed Distribution Power Curve TT 8 8 8 Power (kW) Frequency (Hours) 8 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 3 Wind Speed (m/s) Noble ind Speed Distribution | ower Curve 1482 KW Power * Hours = Energy 1.482MW * 376.9 hrs = 559MWh 8 8 8 _ hours Frequency (Hours) Power (kW) 8 4 5 6 7 8B 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Wind Speed (m/s) fmm Wind Speed Distribution Energy —Power Curve 8 Frequency (Hours) 8 8 234 5 6 7 8B & 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 BZ Noble Gross Energy Wind Speed (mis) rte 900 MWind Speed Dsirbuton | © Energy 800 7|__ Power Curve ] | Gross Energy = 6,134 MWh 700 ++ | | Other losses 10% > {|_| @ 2% 3 | ° 7 3 | | __ Net Yield = 6,521 MWh E sot 4 Se > | | | 3 | | . © 400 }—- t t 2 > 8 ° IL 200 ++ > 200 100 ° 2 ° 0123456789 Wind Speed (m/s) 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 1000 800 400 200 z 3 é Power (kW) Noble ENVIBONMENTAL POWER Different Wind Regimes, Different Distribution Shapes Weibull distribution PCW) = (KIC) * (VIC) * etwo Mid-latitude P(v) = Probability density as function of wind speed v storm driven v_=wind speed C= Scale parameter (units of wind speed) Trade Wind ‘oe ‘ e California (thermal and storm) ‘Mean wind speed calculated as: These distributions have the <V> = J v* p(v) div) = C * 1 (1+4/k) same mean wind speed but the tem wid eeeed energy output varies by 20%! T = Gamma function u WZ Noble EHWIRONMENTAL POWER Frequency Distribution: A Statistical Entity that Takes Time to Develop One Day One Week 2% 20% 18% 10%: 16% 10% 14% 10% 12% 1% 10% 10% o% ™. om ox ~ *. eo =. ™ om. 12.3.4 $67 & 9 10111219 1415 1617 18 19.2021 22.23.24 25 2827 12.34.56 7 8 9 10111213 1416 16 17 18 19.20 21 22 23.24 25 26.27 Wind Speed (m/s) Wand Speed (ms) One Month One Year 12.348 67 8 9 1041 1213141516 17 18 1920.21 22 2324 25 28.27 OFFS ASS Ta SONIC MAB MD Wind Speed (mis) Wind Speed ig WZ Noble (DeviRONWENTAL POWER Measurement Strategies Place masts: * Within clusters of proposed turbine sites — 0.5km to 5 km from turbine depending on terrain complexity Measurements can be assigned to turbine sites by various schemes: * Neighboring turbines with no data adjustment or modified by expenence based judgment + Multiple masts impacting turbines by distance weighting scheme. Avoid unrepresentative sites: — Bluff edges, small peaks, valleys, obstructions (trees) Support numerical modeling strategies: — Zones of influence for linearized models — High and low features for mass consistent models nédbural choice WZ Noble TieinoNhaestat POWER Relating Data Among Towers On-site ‘Ten Minute Average Data YY =0.76°X+071 60 80 0 120 40 160 Tower 2 (mi) Ways to Best Explain Relationships Betwee se Observed Winds at Multiple Locations “Linear Regression *Method of Moments *Support Vector Regression *Physics based numerical models »ble MENTAL POWER Estimating a Second Wind Speed Distribution By Scaling Shape of First Tower 26 Distribution Scaled from Tower 2 | | | | | Example: 726 =0.8*T2 en Frequency 012 34 5 6 7 8 9 1011 12 13 14 15 16 17 18 19 20 21 22 23 24 25 M/S ww Noble FHVUIRONMENTAL POWER Energy Calculation for Similar Distribution Shapes as Function of Mean Wind Speed Energy vs Mean Wind Speed With a model or || measurement Note: Relationship is - | determined almost linear! ‘| mean wind speed at each turbine site, it is straight-forward to calculate a windfarm gross energy estimate. eo, NU cr vob le Windfarm Estimate Considers Various Adjustments and Energy Losses ‘Topographic Effects — Dependent on terrain complexity as well as measurement and modeling strategies Wakes — 0% to 10% typical Turbine performance — Blade degradation 0.5% - 1% typical — Icing 0% to 2% depending on weather and wind regime Availability — 3% typical estimate. * First 6 months 90% - 95% typical * First few years 97% - 98% achievable * Some investigators add extra penalty due to tendency for failures to occur during windy periods. Electrical losses = 2.5% - 3% typical not counting transmission losses beyond interconnect of losses associated with availability or curtailment of transmission infrastructure. High Wind Hysteresis — High speed cutout followed by restart. 0% - 2% depending on turbine and wind regime. Sector Management — Shut-down of turbines to prevent excess loading. — Proper siting minimizes this affect to “off-axis” winds. 0% to 3% Ww ww Noble ENVIRONMENTA: POWER Flow Visualization of Internal Structure of Wake of a Large Wind Turbine WIND DIRECTION Figure 6.1 Sequential Pictures of a Single Smoke Trail at 1/2 D and ¢ = 275° IRT701 PNMENTAL POWER Estimate of Most Likely Windfarm Output Project Capacity (MW) 90 283,824 283,824 283,824 ‘This is best estimate of windfarm output. There is a 50/50 chance of the output being higher, or lower. ‘This estimate is therefore called, the P,, energy level. gble Uncertainties Affecting the Energy Estimate * Affecting Mean Wind Speed Determinations Anemometer error (2% - 3% typical) Quality of correlation of site record to reference site (0.2% - 5% typical) Representativeness of reference site record to long term (0.5% - 3% typical) Shear (extrapolation to hub height) (0.5% - 3% typical) Error estimates combined: + Note: These errors add statistically (square root of sum of squares) to yield estimate of the error (standard deviation of a normal distribution) of mean wind speed. * Mean wind speed error * A Energy / A Speed (usually linear) * Affecting Energy Determinations — Wake and topographic effects (3% - 8% typical) ~ Frequency distribution & Wind Rose (0.5% - 2% typical) — Metering and other losses (0.2% - 0.5% typical) * Uncertainty Resulting from Year-to-Year Volatility of Wind — 3% - 6% typical and likely geographically dependent. 10 iN? in? Noble ENVIRONMENTAL POWER Pos: 95% probability annual energy exceeds this level Predictive Intervals Uncertainty analysis used to predict (at ame of estimate) probability that Standard deviation 4 given energy (example) 43.6 level will be GWH/yr) (~15% of exceeded. mean) Noble ENVIRONMENTAL POWER In this example: P50/P95 ~ 1.34 After Some Operation, Uncertainties May Be Reduced and P,,,/P,; Estimates Change After operation, interannual variability remains one of largest uncertainties, but other uncertainties are substantially reduced through operational experience. P,, estimates will be better established, suggesting post- commercial operation financing may enable greater debt capacity. Geographically diverse portfolio may further reduce interannual variability estimate giving higher overall P,, for the portfolio. 11 12 AcTUAL PERFORMANCE S@OUBUIOLJeg jen}joy Validation of energy and uncertainty predictions by comparison to actual production Andrew Tindal AWEA conference Portland, September 2007 Presentation input from: Clint Johnson, Adam Schwarz, Keir Harman, Andrew Garrad ia GARRAD HASSAN Who is Garrad Hassan? Renewable energy engineering consultancy “a SI Lh 2 yr py Y 3 hin é 555 Phe J Lagi Py a ah, oe 7 g f ‘ “ m a “ a ¥ a om i ‘9 ep ok Ie = a 7: Chole wees Portland, OR [oe re Ae per roush. NH y | es ~ San Diego, CA y ; Kistin, =f > Mbnterrey, Mx Offices in 15 countries with 240 staff North American offices shown a GARRAD HASSAN Energy production validation database Contains 156 wind farms in Europe and the US Operational periods from 1 to 14 years Contains 510 wind farm years of energy production Raw metered substation production For a subset of wind farms availability data are available and have been analysed A» HASSAN Distribution of wind farm years 22% 20% g 18% lon 2 2 2 % of population of wind farm producti OS Europe EN Europe GN America | Southern Europe 59% ~~ Northern Europe 31% ~ “US 10% — GH have assessed > 6000 MW in US Year Approach adopted High level — to complement a range of more detailed validations * Actual data are compared to GH Pre-construction . P50 and P90 predictions Some wind farms excluded due to gross issues such as grid curtailment or very poor availability ali. HASSAN Treatment of differing lengths of data * To compare like-with-like, each wind farm year considered separately * Comparison therefore with the 1 year P90 « Asan example if a 10 year P90 value were 88 % of P50 the 1 year P90 would be 85 % of P50 Distribution of Annual Energy Production 510 Wind farm years relative to GH Projected P50 100 ———— a= on __ : aetual production ae ~ —GH Predicted distribution LL 80 + Wind farm years = 510 Average = 93.3% No of wind farm years ¥ # 83 Actual annual production / GH Predicted P50 Distribution of Annual Energy Production 322 Availability corrected wind farm years relative to GH Projected P50 70 g $ Wind farm years = 322 Average = 93.6% & No of wind farm years 8 » 8 3 ° ££ 8 Fk *# 838838 Actual annual production / GH Predicted P50 Summary of results — Whole data base — 7 Whole data base (510 years) Availability adjusted subset (322 years) Average ratio actual/predicted 93.3 % 93.6 % Wind farm years below P90 21.2% 18.0 % Wind farm years below P95 Commentary on results after US results Break out of just US data Existing data base augmented with some public domain data from the Energy Information Administration (EIA) data base EIA data very high level (No availability) and needs to be used with care Data trends indicate probable availability or grid curtailment issues with several US wind farms (often covered by warranty provisions) Data screened to exclude gross issues US energy production data Only 55 US wind farm years in data base of which a third have detailed availability data —- Why so little? Operational assessment not included 12000 + = 10000 | 8000 US MW installed (M' 1987 Distribution of Annual Energy Production - US 55 Wind farm years relative to GH Projected P50 16 ol Aetual production = “4. ——— —GH Predicted distribution Wind farm years = 55 s | | | 1 ] | | Average = 92.1% © No of wind farm years o Actual annual production / GH Predicted P50 Summary of results — US Whole data base (55 years) Average ratio actual/predicted 92.1% Wind farm years below P90 23.6 % Commentary on US results — Availability Turbine and grid availability is a significant issue for some US wind farms — “average” historic number appears to be 93% Levels achieved are substantially lower than those which are generally achieved in Europe aaa 1 © Average monthly availability | —3 month moving average of availability —12 month moving average of availability — No of wind farms in sample i i ! i 2 a 3 4 sigh Years of operation since commissioning | Future availability estimates Projects can and do achieve 97 % availability If availability levels are much lower (93%) in the US do we a) Assume 93% going forward? b) Sort out the problems! Need to understand issues from detailed analysis of the data AND manage assets better y_N GARRAD HASSAN Commentary on US results — Performance Real world turbine issues reduce energy e.g. Control, Bugs More study of complex terrain power curves merited Power [kW] Monthly power curves — using standard SCADA ~~ data- incorrect —— change to control — algorithm revealed 10 wind speed [rvs] 10 Commentary on US results - Measurements Some early data influenced by “stub mount” Left can give 5 to 10 % higher energy than right! This has influenced results in the data base The industry has improved in this area! Commentary on US results — Wind flow/shear Many players putting more focus on measurements: * More masts * Some taller masts * Better sited masts This improves estimates of wind flow over site and shear The industry has improved in this area! 11 Commentary on US results - Windiness Evidence that in several key areas wind speeds in 2005 particularly poor — example below (Oklahoma) This has had an impact on overall data base Annual mean wind speed (%) 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Year Commentary on US results — Key issues (1) Validation data base results indicate estimates have typically been over-predictions But Availability (and grid curtailment) has played a substantial role in this. Industry should not continue to allow this loss of energy to happen. 12 Commentary on US results — Key issues (2) Energy production estimates have played role too, but * Wind measurement quality is improving ...but are developers doing enough measurements? * Methods are improving * Recent windiness has exacerbated under-performance Remember The US data set is still small! tie. Commentary on Global results Same issues as US but different emphasis on different issues in different countries 13 UK case study Can attempt a reconciliation for the UK as: Long data sets available (back to 1993) Can apply windiness adjustment More detailed data available Screened data to remove: * Wind farms with terrible availability! « Wind farms with terrible measurements! iii 14 Distribution of Annual Energy Production - UK 113 Wind farm years relative to GH Projected P50 (screened data set) (actual metered production GH predicted distribution 23 -$—_____—_——_ - a 0.——— Wind farm years = 113, ‘Average = 96.0% Number of wind farm years ‘Actual production / GH P50 Summary of results — UK Whole data base Availability and (113 years) windiness adjusted subset (34 years) Average ratio 96.0 % 101.7 % actual/predicted Wind farm years below 13.3 % 2.0% P90 UK conclusion: If data are screened to consider “modern” data norms and corrected for availability and windiness then projections in line with expectations A» 15 Conclusions — Overall result It is only by looking at such large volumes of data that we can start to answer the question “How well have they done?” “Raw’ results tend to show over-prediction, but... Conclusions — Asset Management Availability a key issue in the US Industry needs to achieve better availability And need to use data to better understand (and review) loss factors 16 Conclusions — Energy Prediction Industry should not be complacent about energy prediction or need for measurements Past analyses did not benefit from the measurements (and techniques) available for current analyses With “modern” measurement campaigns and assessment techniques and considering recent windiness can expect future projections to be in line with estimates ili, HASSAN What next? (1) GH will continue to update the data base with more data. Please provide data for the US! Validation is key 17 What next? (2) Key issues for the industry (and GH) to continue to work on: Understanding loss factors through rigorous analysis of data Validating advanced measurement techniques Understanding power performance in complex terrain More sophisticated wind flow / wake models Ter Wn 2X, © GW eneay orb si YO 18 Sa. Aric: Nery nach __American_Wind Znergy Association ™ How Well Have They Done? Tuesday, September 18 11:15 am — 12:15 pm Speaker: Steve Jones Global Energy Concepts, LLC How Have Projects Performed? Comparing Performance with Pre-Construction Energy Estimates AWEA Wind Resource Assessment Workshop Portland, Oregon Steve Jones Global Energy Concepts 1809 7" Avenue, Suite 900 Seattle, WA 98101 (206) 387-4200 sjones@globalenergyconcepts.com September 18, 2007 oer The Question * How well do operating wind power facilities perform relative to their pre-construction P50 energy estimates? + P50 definition: Over the long term, 50/50 odds that facility will produce higher/lower than P50 estimate P95 estimate: 95% of area under cure lies to the right Energy, GWh per year Example of Unmet Expectations Owners Reporting oy 2° 8S S88BSR3FE FBRSEBB Miles per Gallon Example of Unmet Expectations Drivers rarely see the actual EPA-rated mileage in the real world, according to John DiPietro, road-test editor of automotive website Edmunds.com. DiPietro says most drivers will get between 75% to 87% of the rated mileage, with individual variations based on driving habits and traffic route. Similar History With Wind? “What has been your experience with projects once you do the deal? How well have they performed?” “The ability to predict output is not as good as we would like. We are invested in 26 wind farms, and 21 of them have been in service for some time... Our portfolio has performed at 91% of the P50 level through the end of 2006. That would put the portfolio somewhere between the P75 and P80 forecast. What varies is significant. We have four or five projects that are overperforming and another four or five that are barely operating at a P95 level of output. The science of wind forecasts is imprecise....” Keith Martin, Partner, Chadbourne & Parke LLP, posing question to John Eber, Managing Director of nergy Investments at JPM in Capital Corporation, at Infocast Wind Power Finance & Investment Summit 2007 in La Jolla, California, in February 2007.) E Cc Outline ¢ Approach/Limitations * Operating Project Performance Data ¢ Discussion * Warning Signs ¢ Conclusions Approach Operating North American wind power facilities; wide diversity of locations and turbine technologies Pre-construction estimates only; none “trued-up” using actual results Variety of firms/individuals produced the estimates Actual performance based on revenue meter No data excluded unless confidentiality agreement prevented use Data evaluated for 33 facilities and 125 facility-years of operation 1 to 10 years of operation at individual facilities; no artial years Ceo _>GEC Limitations Not considering potential for reimbursements from insurance or turbine manufacturer — Insurance for lightning damage — Availability warranty liquidated damages Did add-back curtailments ordered by buyer when information available (few cases) Bias towards many “young” facilities with short operating history Year-to-year variability in wind confounds comparison — Somewhat covered by many years of data — Facilities cover wide range of wind regimes and tend to not have concurrent “good” and “bad” wind years _2»GEC Survey Results: Actual Energy Wind Farm Years = = nN nN ao oO a Oo a o a ° "ae Rk 8 8 Actual Energy / P50 Estimate i 9, Average is about 11% below P50 GEC Potential Sources of Difference * Wind estimating process errors or bias * Wind speed to energy estimate process bias * Losses higher than estimated Wind Estimating Bias * Met equipment frequently in best exposed areas, thus more potential for downside than up-side. Models can underestimate impact of terrain. When anemometers are relatively well exposed, more likely to bias estimate high. In inclined flow, typical “NRG #40” anemometer measures more than the horizontal component of wind speed that is useful for turbine rotor. May be experienced on steep ridges. —”GEC Typical Pre-Construction Technical Loss Estimates Array (site dependent) Turbine Availability 2% to 6% BOP Availability 0% to 1% Electrical Line Loss/ Parasitic 2% to 4% Consumption Controls/Turbulence 0% to 2% Blade Soiling/Degradation 0% to 2% Weather 0% to 5% Utility Outage/Curtailment 0% to 5% Power Curve 0% to 2% Total (without array loss) 5% to 17% Survey Results: Availability 20 5 15 + Wind Farm Years ° x 8 Annual Average Availability + Average is about 93% Availability Losses Example + 700 3% —— Faults —@— Power Cure | + 600 | + 500 2% + = = + 400 ra eo + 3005 ° Probability of Fault + 200™ + 100 0 6 8 10 12 14 (16 18 20 Wind Speed (ms) * Lost time # lost energy Source: Big Spring Wind Power Project, U.S. Department —GEC of Energy-EPRI Wind Turbine Verification Program Array Losses ¢ Multiple available models, with varying degrees of validation Model results change materially depending on model used and atmospheric conditions assumed Use of a single model m & a increases potential for 1 2 3 ZA a bias (high or low) to Wake Loss Model Used be introduced Facility Wake Loss (%) Other Losses ¢ Many uncertainties go one way: — Availability: Can easily be 92% but cannot be 102% — Controls: Many ways to be “off” the power curve, but not much chance to be significantly above it — Weather losses can only hurt, not help: * Icing conditions * Hail/thunderstorm shutdowns * Access to site or turbines * Crane use limitations Power Curve Example Percent of Rated Power Turbine Power Performance (10-minute Nacelle Wind Speed, m/s Warning Signs Long-term adjustments to short-term on- site measurement campaigns that result in large upward adjustments to site data Topographic effect adjustments at or above 100% —Implying wind turbines would have better winds than the met tower or towers Low or ignored technical losses — Availability, weather, etc. Conclusions Many sources of losses or other underperformance frequently ignored or underestimated There are projects that run at 97% to 99% availability...but they make up about 1/3 of projects studied There are projects that meet or exceed their P50 projections...but not nearly half of them in our experience Carefully examine assumptions of energy analysis when evaluating likely performance —&GEC 10 2004 Toyota Prius Automatic (fully variable gear ratios) 4 Cylinders 1.6 Liters Regular Gasoline Look up another car > Hybrid Vehicle er New MPG tests are 46 5 5 mote realistic AB comines 45 GQ Combines 54 City Hivy City Hwy L 1918 Muaehe atu koud Soak “You MPG" QD Average based on 86 vehicles. Disclaimer > . 61 Owners Reporting enon t+ H OH Miles per Gallon 11 Contact Information Steve Jones Sr. Director of Utility and Investor Services Global Energy Concepts, LLC 1809 7th Avenue, Suite 900 Seattle, WA 98101 (206) 387-4200 sjones@globalenergyconcepts.com www.globalenergyconcepts.com GEC 12 Pr =a pees q Winb SPEEDS oN ™ Es se on __American_Wind nergy Association teehee deleeentneedeeianetnnn ieee Recent Studies of Wind Speeds in North America Tuesday, September 18 1:30 pm — 2:45 pm Speakers: Dennis Elliott National Renewable Energy Laboratory A national! eboretory af the US Deperiment af Energy (Orfice of Energy Ethcsency & Renewable Energy oFe =1 . “age” MREL national Renewable Energy Laboratory (rae aad Recent Wind Resource Variability Dennis Elliott, Marc Schwartz, George Scott, NREL/NWTC Golden, Colorado AWEA Wind Resource and Project Energy Assessment Workshop September 18-19, 2007 Portland, Oregon * NREL is operated by Midwest Research institute. Battelle 20 Introduction to NREL National Renewable Energy Laboratory, Golden, Colorado Funded by the U.S. Department of Energy, managed by Midwest Research Institute NREL’s National Wind Technology Center (NWTC), south of Boulder, Colorado, employs about 90 people. Background Long-term wind measurement data are rarely available from wind energy development sites Knowledge of yearly wind resource deviation from long- term average is needed to adjust short-term site data Various long-term data sets and methods are used in adjusting short-term site data, such as: — Surface meteorological stations (e.g., airports) — Upper-air stations — Model data such as Global Reanalysis. Global Reanalysis is generally considered to be a consistent long-term data set useful in: — Wind resource assessment studies — Input to numerical mesoscale wind modeling and mapping. Objectives of Study Analyze recent wind resource variability for selected areas of the United States using Global Reanalysis data — North Central (particularly |A/MN) — Southern Plains (particularly TX/OK) — Northern Appalachia (particularly PA/WV/NY) Examine years 2005 and 2006 relative to 20-yr average (1987-2006) Determine largest yearly deviations from 20-yr average Analyze variations and trends over 20 years especially interesting or suspicious features Discuss results and implications — How to estimate confidence and uncertainty of results — Need for additional studies using other long-term data sets SERIES novenst ememante ner eacntony Reanalysis Data Processing 50-yr global climatic data set (1958-2007) produced by the National Centers for Environmental Prediction and National Center for Atmospheric Research Data simulated four times per day on a 208-km resolution grid of wind, temperature, and other variables Incorporates all available rawinsonde and pibal data, and observations from surface stations, ships, aircrafts, and satellites Sigma level (terrain following) surfaces used in NREL’s analysis — five sigma levels in the lowest 1 km above ground + 40-50m, 150m, 300m, 500m, 700m (Levels 1 thru 5) — 150m sigma level (Level 2) wind speeds generally comparable to 50- 80m measurements in well-exposed flat areas For this study, NREL analyzed data from a 20-yr period 1987-2006 for regions of central and eastern United States. Central U.S. - Percent Deviation from 20-yr Average - Level 2 2005 ; ~ 50N - 123 Pees 32,087 | 7 04-13 22 *2¥as” EDI 064 00 -08 14.13 01 o4 0.3 2-58 x 4 a 105W 100W 95wW 0.2 LS 68 18 3.4 97 -104 3.0 45 eyed SIHTDMS 5d. 03 | sow 3B 43 6.0 19 27 2 8 40 40 17 58 28 10 ogp-db ea Se. i a L 105W 100W SSW sow 2005 was mostly below average over southern & eastern areas, above average over Dakotas 2006 was mostly above average Central U.S. - Percent Deviation from 20-yr Average - Level 2 Minimum Year 66 63 - 4.7 4.0) -4.0 46 -40 35 36 5B 62 46 -39 34 -54 68 -79 5,7 6.3 -4.7 65 82 “B7 64 42 42 65 -77 90 OF -104 D7 yay AS 33 48 61 ott bore 867” ct Ni er | 105W 100W 95W g0W Minimum year varies from -3.4% to -15.5%. Least deviation over central areas (KS). Greatest deviation over northern & southeastern areas. oN Maximum Year 5) 57 ShmG8 8 36 44 4ORZ9 7 $265.7 5.1 57 53 46.61 5.0 62 58 76 36 64 74 53.43 69 78 65 44 46 55 7.1 86 58 47 30N “RZ L Ni 1 1O0SW 100W 95W Maximum year varies from 2.9% to 14.6%. Least deviation over Dakotas. Greatest deviation over southern areas & CO. North Central Region Annual Wind Speed Trends - Level 2 X 1987 to 2006 Y -12.0 to 12.0 pct deviation from avg T Percent Deviation from 20-yr Average - Level 2 2006 ow Minimum Year Maximum Year 76 Southern Plains Region Annual Wind Speed Trends - Level 2 X 1987 to 2006 Y -12.0 to 12.0 pct deviation from avg 7 Percent Deviation from 20-yr Average - 2005 04 44 46 (1.6) (1.4) (3.6) (0.3) (19) (5.1) toow Minimum Year Level 2 (Level 4, Mesas) 2006 19 290 44 (27) (3.2) (4.1) (33) (34) (5) 100W Maximum Year Northern Appalachian Region Annual Wind Speed Trends - Level 5 X 1987 to 2006 Y -12.0 to 12.0 pct deviation from avg Percent Deviation from 20-yr Average — Level 5, Ridges 2006 28 Comparison of Regions . non Central 2005 near normal, generally within +/- 1% of 20-yr avg 2006 near normal (+/- 1%) in sw MN & e SD and above (2% to 3%) in IA & e NE Minimum year about 4% to 9% below 20-yr avg Maximum year about 3% to 6% above 20-yr avg Annual trends over 20 years varied across region . Southern even 2005 (0% to -2%) in w TX & OK and much below (-4 to -6%) in c-e TX with record or oreo minimums in some areas 2006 above (2% to 5%) in c-w TX & OK and near normal (+/- 1%) in nw TX Minimum year about 3% to 7% below 20-yr avg Maximum year about 3% to 10% above 20-yr avg Annual trends over 20 years varied across region . Northern Appalachia 2005 much below (-5% to -14%) with record or near-record minimums in many areas 2006 near | (+/- 1%) in central & southern areas, below (-2% to -4%) in eastern areas and above 2 %) in western areas Minimum year about 6% to 14% below 20-yr avg Maximum year about 7% to 11% above 20-yr avg Annual trends over 20 years varied across region, some areas (NY, e PA, NJ, MD, DE) indicate suspicious downward trends —— SS Why Differences between Reanalysis and Measurements/Production Data? + Reanalysis data incorrect + Measurement data quality issues — Exposure problems — Data losses or sensor problems — Tower flow effects — Wind shear extrapolation errors + Reanalysis doesn't reflect local flow conditions and terrain influences + Uncertainties in relating production data to wind resource such as — Turbines not operating at power curve — Wake and array losses or other losses that may not be accurately quantified 15 ys SEP NEY noma tangy tonsirey Reliability and Accuracy Issues with Measurement Data Upper-air station data issues — Standard pressure level data (may be hundreds of meters above surface) Data measured only twice per day Upper-air data may not relate well to measurements at 50m — 100m above ground Stations widely separated (often 300m — 500 km apart) and irregularly spaced Station's wind climate may not represent site's wind climate Surface meteorological station data issues Surface data (at or near 10m) may not relate well to measurements at 50m — 100m above ground Station's wind climate may not represent site’s wind climate Changes in station data (caused by changes in anemometer location, surroundings, or equipment) Pre-ASOS (Automated Surface Observing Systems) to ASOS (1992 — 2003) and post-ASOS (2006 and ongoing) discontinuities Lack of metadata (such as exact locations and long-term history of measurement equipment) Tower data issues - Tower flow effects on measurement accuracy — Data recovery and adjustments for missing data Questionable data due to icing, weather conditions, sensor problems, etc 16 * SEPP ween teste tne tatty Why Compare Different Long-Term Data Sets? A single Ur of long-term data may not provide an accurate or reliable estimate of the annual wind resource variability Comparisons of different types of long-term data sets are needed to: — Evaluate the similarities or differences among the data sets — Identify potential problems or issues with the data sets — Assess the uncertainty or confidence in estimates for specific areas — Choose the most reliable and accurate data set(s) for specific areas if possible. Long-term data sets used for comparison include: — Model-derived data * Global Reanalysis — 208 km, 1958-present + North American Regional Reanalysis — 32 km, 1979-present — Upper-air station measurements (rawinsondes) — Surface station measurements (primarily airports) — Tower measurements (very limited) Conclusions and Recommendations Long-term model-derived data sets such as Global Analysis can be processed to estimate annual wind resource variability over large regions A single type of long-term data may not provide an accurate or reliable estimate of the annual wind resource variability Different types of long-term data sets (including measurement data) should be analyzed and compared to: — Examine uncertainty or confidence of estimates — Identify the most reliable and accurate data set(s) for specific areas. Disclaimer and Government License ‘This work has been authored by Midwest Research institute (MRI) under Contract No. DE-AC36-99G010337 with the U.S. Department of Energy (the “DOE"). The United States Government (the “Government’) retains and the publisher, by accepting the work for publication. acknowledges that the Goverment retains a non-exclusive, paid-up, revocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for Government purposes. Neither MRI, the DOE, the Government, nor any other agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe any privately owned nghts. Reference herein to any ‘specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not constitute or imply its endorsement, recommendation, or favoring by the Government or any agency thereof. The views and opinions of the authors and/or presenters expressed herein do not necessarily state or reflect those MRI, the DOE, the Government, or any agency thereof. 10 AN __American_Wind Znergy Association Recent Studies of Wind Speeds in North America Tuesday, September 18 1:30 pm — 2:45 pm Speaker: Jeff Freedman AWS Truewind, LLC —————————eE7~ __ <== —————————S AWS Truewind Potential Effects of Climate Change on Long-term Wind Speed Trends AWEA Wind Resource and Project Energy Assessment Workshop Jeffrey M. Freedman, Research Scientist AWS Truewind, LLC 463 New Karner Road Albany, NY 12205 ifreedman@awstruewind.com © 2007 AWS Truewind, LUC We've certainly looked at SS COz and temperature... CO2 and Temperature Trends AWS Truewind — CO2 Mauna Loa vise —— CO2 South Pole + --- US Temperature Anomaly (5 yr running mean) ~ Portland, OR Temperature (5 yr running mean) a g4 ° 7 8 Sb 3 Aye saevensereiment fan mn 1S g Keeling et al. 2007 < e 2 SG 4 Eel |t ° PT? we se ee rE = =s f & a ars #L _Lle pay T TT 1990 2000 Decimal Year a - ~ © 2007 AWS Truewind, LLC. et ——= —— AWS Truewind | What about Wind Speed? ¢ Few studies on long-term wind speed trends (but see. e.g. Klink 2002) ¢ Results from downscaling of different Global Climate Models (GCMs) are inconclusive as to future affects on low-level wind | speeds (Pryor et al. 2006) ¢ Some areas observe a net increase in wind speed and other areas show a decrease ¢ Areas within the U.S. that are most susceptible to climate change also contain the greatest wind resource (for example, the Great Plains) (© 2007 AWS Truewind, LLC ane, ——— | AWS Trewind What Observations Can We Use? | ¢ Problem: for surface stations, analyzing trends is complicated by changes in site location, surrounding land use, and | instrumentation ¢ Wecan eliminate first two issues through use of archived rawinsonde (an upper-air sounding that includes determination of wind speeds and directions). | Rawinsonde Stations with Period Of Record >50 Years SS | (through 2006) AWS Truewind 65 stations with long-term continuous record ~ SSa= | Twice-daily rawinsonde data were | acquired from the National Climatic Data Center’s (NCDC) Integrated Global Radiosonde Archive (IGRA; see http://www.ncde.noaa.gov/oa/climate/ igra/index.php) © 2007 AWS Truewind, LLC | AWS ‘ivews nd | | But are these measurements.representative of surface layer (lowest 200 m) winds? | ¢ Problem: lowest standard measurement levels since | 1940s through early 1990s were at 1000 hPa (~ 150 m above ground level, or agl) and 850 hpa (~ 1500 m agl). (925 hPa added during early 1990s as were additional wind levels at 300 m increments to 3000 m) ¢ Problem: for good portion of North America, 1000 hPa (and in some places, 850 hPa) is underground! ¢ Issue: does 850 hPa have any relevance to surface wind speeds? ¢ Answer: depends upon where you are, and when you are. (© 2007 AWS Truewind, LLC: Ci ——S>S= Wind Profiles for Stable and Convective BLs AWS Truewind | a= Sie eee iS eis a a 850 hPa | E at . | _ boundary | layer es 7) surface layer | stable layer, 00 m | Noon 7PM 7 © 2007 AWS Truewind, LLC 850 hPa Wind Speed (m/s) For Selected Upper Air Stations 14 12 Wind Speed (m/s) 10 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 Year RacbNCDCALL. func) Wed May 16 16:37:21 2007 rer, ===: ——S AWS Truewind Increasing wind speeds last 20 years at 850 hPa... Rawindsonde 850 hPa Wind Speed Trends 1967 — 1986 Rawindsonde 850 hPa Wind Speed Trends 1987 ~ 2006 ms ‘decade’ 8 _.and at 700 hPa AWS Truewind Rawindsonde 700 hPa Wind Speed Trends 1967 — 1986 ms" decade"! © 2007 AWS Truewind, LLC | | — AWS Truewind | Rawinsonde 850 hPa Temperature (degrees C/decade) Trends 1987 — 2006 | But it gets warmer nearly everywhere... | | | | | | | } | | | 10 | © 2007 AWS Truwind LLC | ————— AWS Truewind Possible Mechanisms ¢ Cyclic nature of wind speed trends (7?) — could be linked to ENSO or other teleconnections — solar activity (?) — SST anomalies ¢ Global change - working hypothesis is that global warming would reduce the meridional thermal gradient (since higher latitudes would observe greater warming) and hence the pressure gradient which drives the wind. (Zack et al 2007). This assumes that warming would be uniform across all latitude belts--not borne out by GCMs. | — Alternatively: the intensifying thermal gradient between the upper troposphere and lower stratosphere, especially in the higher latitudes, will lead to a strengthening and a poleward shift of the tropospheric zonal jets (Lorenz and DeWeaver 2007) and a concomitant shift in storm tracks. This may be responsible for the observed maximum in wind speed trends over north-eastern Canada 1 © 2007 AWS Truewind, LLC | | | Lorenz, David J and, Eric T. DeWeaver, 2007: Tropopause height and zonal wind AWS Truewind | response to global warming in the IPCC scenario integrations J. Geophys. | Res., 112, No. D10, D10119 Higher wind speeds over most of North America Tronsient < Stobilizotion Figure 3. Ensemble-mean zonal-mean climatological zonal wind for the 20th century (shaded) and t ensembie-mean zonal-mean change in zonal wind (comours). (a) Winter. (b) Spring. (c} Summer. (d) Fe ots Figure 13. The annual average ensemble-mean zonal mean change in the 850 bPa zonal wind relative to 1980 1999 for 2080-2099 (solid) and for 2280-2299 (dotied). The calculation is done for the 7 models in the B1 forcing scenario with data out to 2299. 12 (© 2007 AWS Truewind, LLC —_ — AWS Truewind Conclusions ¢ Wind speed trends over the U.S. and Canada show a general increase during the last twenty years, following a general decline during the previous two decades ¢ Whether this is cyclical—a function of other periodic changes in the general circulation pattern related regional or global teleconnections such as the North Atlantic Oscillation (NAO), Arctic Oscillation (AO) or El Nino-Southern Oscillation (ENSO) —is the subject of ongoing research ¢ If such a link is clearly established, reasonably accurate regional seasonal or inter-annual forecasting of wind speeds, and hence wind power availability, will be possible ¢ This in turn will help the wind energy industry avoid the perception of instability if longer term wind resource forecasting and reliable estimates of the future site-specific power availability becomes a reality. 13 © 2007 AWS Truewind, LLC Dp ne See Wino MEASUREMENT __American_Wind Znergy Association Identifying and Reducing Wind Measurement Bias and Uncertainty: Sensors Tuesday, September 18 3:15 pm — 5:30 pm Speaker: Bruce Bailey AWS Truewind, LLC AWS Truewind Uncertainty and Bias in Anemometry Used in Wind Resource Assessment Bruce H. Bailey AWS Truewind LLC Albany, NY bbailey@awstruewind.com © 2007 AWS Truewind, LLC Agenda Introduction Anemometer Uncertainty Influence of Sensor Geometry Other Considerations The Calibration Process and Uncertainty Real World Conditions Mitigation Methods © 2007 AWS Truewind, LLC Sources of Uncertainty (Typical Range of Impact on Lifetime Energy Production) e Measured Speed e Shear i (4-9%) Measurement Duration, Period of Record e Climate —_ Raterace Station, Quality erceralen | e Resource Model e Plant Losses (2-6%) | Sensor Types, Calibration & Redundancy, ——* Ice-Free, Exposure on Mast, # of Masts (1-3%) [Height of Masts, Multiple Data Heights, | Sodar, Terrain & Land Cover Variability (5-10%) ., Microscale Model Type, Project Size, | Terrain Complexity, # of Masts, Grid Res. —— EEE Eee (1-3%) _ Turbine Spacing (wakes), Blade Icing & | Soiling, Cold Temp Shutdown, High Wind Hysteresis, etc. | 2007 AWS Truewind, LLC eee es AWS Truewind Related Uncertainty Sources Tower Effects - Flow distortion around tower Mounting Effects - Boom length, orientation, and related structure Implications - Flow distortion can have a significant impact on data accuracy and resource assessment uncertainty This topic will be treated in detail separately Tower Influence on Free-stream Speed E . Dimensions in Meters 2007 AWS Truewind, LLC Alternative Sources of Wind Speed Data Beaufort Scale -— Wind Speed Estimated by Visual Effects on Land Features — Accuracy : 15% Without Height Adjustment National Weather Service (NWS) — Measurements of Wind Speeds for Weather Conditions — Accuracy: +/-1 m/s up to 10 m/s 10% above 10 mis Environmental Protection Agency wae78 - Accuracy: 0.25 m/s < 5 m/s 172-207 5% > 2 mis not to exceed 2.5 m/s ne 245-284 World Meteorological Organization Toon — Accuracy: 0.5 mis < 5 mis wara0n 10% > 5 mis Alternative sources and techniques have large uncertainty; the wind industry strives to have stricter requirements (© 2007 AWS Truewind, LLC aati ——— —— AWS Truewind Statistics 101 Uncertainty is The Estimation of Experimental Error in a System + Identifies Errors and Defines the Degree of Error in the System + Bias (B) Systematic Errors =-----------=Q-- + Precision (co) x Random Errors Conuila, Rachael and Uncertainty “ AWEA Wi true Uncertainty estimates errors in a system © 2007 AWS Truewind. LLC AWS True Wind Effects to Include in Measurement Uncertainty + Sensor Type — Cup Anemometer — Propvane Anemometer — Ultrasonic Anemometer — SODAR — LIDAR — Hot Wire + Sensor Geometry within Type Calibration of the Sensor Real World Considerations — Life Considerations — Deployment PROPVANE + Uncertainty Considerations: Fluctuating Wind Directions + Typical Stated Accuracy: +/- 0.3 m/s ULTRASONIC ANEMOMETER + Uncertainty Considerations: Heavy precipitation events; Detection of Reflections . Typical Stated Accuracy : +/- 0.05 m/s to 30 m/s; +/- 3% after CUP ANEMOMETER + Uncertainty Considerations: Overspeeding and Off-axis Flow Impacts + Typical Stated Accuracy: 1-2% Cup anemometers are the industry standard for resource assessment 2007 AWS Truewind, LLC te ————— AWS Truewind Uncertainty as a Function of Sensor Geometry for Cup Anemometers + Variety of Construction Differences - Cup Shapes — Bearing Types — Body Shape — Materials + Variety of Configurations + Choice of Shape and Configuration will Affect the Performance and Uncertainty Tap mncmoricicr Topskape Ot] cop] Rowe Tort] Rotor] Pulses rotation | dum, weight | weight” | por res om Ris Pa4ash Comical ow NRG Maximum 0 [Comical cow Semi-sphoneal | CCW Scmi-sphorieal | COW Comeal cow Even within a given sensor type, geometry may be fundamentally different © 2007 AWS Truewind, LLC = AWS Truewind The Distance Constant Definition: The Length of Fluid Flow Past a Sensor Required to Cause it to Respond to 63.2% (1-1/e) of a Step Change in Speed Effects Indicates How Quickly the Sensor Responds to Wind Speed Changes Measured As How Much Air Must Pass the Sensor for Response Affected by Differences in Sensor Type, cae Geometry, and Material of Composition Hater, Rnpnand i. “tas Aanenny et Oxp Resaniten: In the Past, AWEA Recommended sere Distance Constant < 4.0m Implications: Longer Constants: Slower response Greater Overspeeding Shorter Constants: Faster Response Longer distance constant sensors will have increased measurement uncertainty in fluctuating flow (© 2007 AWS Truewind, LLC <== ———s AWS True wind Speed Range Considerations Starting Threshold — Speed at GE 1.5 MW Power Curve Which Sensor Specification . Performance Begins Measnet Test Range=>! AWEA has Recommended Starting Threshold <1.0 m/s Curve Fit Accuracy Effect on Energy Normalized % FS 8 6 8 2 Ge 16 18 20 22 24 26 28 Mean =7 m/s Wing Speed (m/s) —WT Power —Speed Frequency ~- ENERGY John Obermeier. “Considerations in Anemometer Calibration ” Relevance of Accuracy Specs Depends on Speed Range © 2007 AWS Truewind, LLC AWS Truewind Z - Turbulence Issues Intuwence of Turbulence ntenaty Variation toon Reterence Cave Response to Turbulence Turbulence Can Cause Overspeeding Effects Affects other Parameters and Relationships “Characterisation and Classification of RISO P2S46 Cup Anemometer ~ * Ratio S Th Varies with Sensor Type and Geometry ; 005 01S 2s Turbulence Affects Sensor Performance eet ees arene Excerpt of report Deutache WindGuard Wind Tunnel Services GmbH, AK 02 002 © 2007 AWS Truewind, LLC ———— a AWS Truewind Non-Horizontal Flow Horizontal Mounting of the Anemometer is Important Terrain Can Affect Anemometers by Inducing Vertical Flow Anemometer Types and Geometries Respond Differently to Off Axis Flow Measurement of Vertical Flow May be Needed Adapted From: Pedersen, TF and US. Paulsen. “Classification of Operational (Charactenstc of Commercial Cup-Anemometers ” Resor Nabonal Laboratory Off-Axis Flow Affects Measurement Accuracy (© 2007 AWS Truewind, LLC eee, <== ae AWS Truewind Other Effects + Density and Temperature Effects — Density Change Causes Drag Change — Temperature will Influence Bearing Friction + Heated vs. Non-Heated Sensors — Bearing Friction — Sensor Geometry Uncertainty is Introduced from Force Changes Caused by Temperature and Density Variations Pederson, TF. “Characterisation and Classification of RISO P2S46 Cup Anemometer ~ (© 2007 AWS Truewind, LLC AWS Other System Considerations Data Channel Process — Logger Calibration — Signal Output and Type * Serial, Analog + Voltage, Current Towers: Mounting Effects Long Cable Length — Can cause losses or introduce signal noise — Depends on sensor and signal type Other Sensor Types — Wind Vanes — Pressure Sensors — Temperature Sensors Other System Elements Also Introduce Uncertainty © 2007 AWS Truewind, LLC AWS |: Calibration and Effects on Uncertainty CALIBRATION PROCESS ine = aes + Acquire Sensor Output Over Range of Speeds Ens} + Least Squares Fit #7] Linear Transfer Function a + Statistics of Least Squares Fit 7 Correlation Coefficient (R) Standard Error of Reference Wind Speed, Y + Standard Error Visualization Difference of Measured and Calculated Velocity vs. Sensor Frequency or Reference Wind Speed 0 s+ jesidual, CV [ms] Speed boos Bs28 STANDARDS + ASTM D 5096-02 Standard Test Method For Cup and PropVane Performance IEC 61400-12-1 Standard for Performance Measurements on Energy Producing Wind Turbines Conuila, Rachael and John Obermeet “Anemometer Cabraton ASTM D 6011-96 Uncertainty ” AWEA Windpower Conference. June 2007 Standard Test Method for Sonic Anemometer Performance ISO 17713-1 Standard Methods for Wind Tunnel Anemometer Testing Re e z © 200 400» 6008001000 ‘Anemometer Frequency, / (Hz) AWS Truewind Calibration and Effects on Uncertainty Sources of Calibration Process Uncertainty + Blockage — Flow Blockage due to Tunnel Dimensions — Corrections can be made for Blockage (Corrections Based on Solid/Stationary Test Material) Speed Considerations — Flow Variability Bias - Velocity Bias — Direction Bias — Stress Bias Average Accuracy of Wind Tunnel Calibration Process is 0.5 — 3 %, with Tunnel-to-Tunnel Variability Some uncertainty is inherent eae cae even in calibration tests Uncertainty " AWEA Windpower Conference, June 2007 (© 2007 AWS Truewind, LLC AWS Truewind Calibration Accuracy CALIBRATION vs. NON-CALIBRATION + Calibrated Sensor Accuracy ~1% + Non-Calibrated Sensor Accuracy ~1.5% - 2% + Some Sensors only available as calibrated equipment Compared to overall cost for a resource assessment campaign, calibration helps reduce uncertainty for a modest incremental cost © 2007 AWS Truewind. LLC nae <= —— AWS Truewind Calibration vs Field Environment FIELD — REAL WORLD CONSIDERATIONS + Calibration — Wind tunnel calibrated sensors can behave differently in field application — Sensor calibrated to the same point can behave differently in field situations — Mechanical — Bugs — Wear Degradation/Vandalism | Real world effects can alter sensor performance (© 2007 AWS Truewind, LLC a <a AWS Truewind Real World Considerations Icing Loss of Data Calibration Drift Need for Data Validation and Filtering Diligence is called for throughout the measurement program (© 2007 AWS Truewind, LLC 10 Mitigation Methods Instrument Choice — Robust Sensors — Ice Free Types in Icing Environments Redundant Sensors — Avoid Tower Wake Sector Loss — Control For Data Validation — Continue Receiving Data in the Event of Sensor Loss Periodic Replacement — Minimize Failures due to Degradation and Wear Out Post Calibration — Verify Accuracy Upon Removal Validation and Filtering — Identify (and Remove) Questionable Data — Identify and Document Sensor Issues Measurement program design & execution impacts uncertainty © 2007 AWS Truewind, LLC AWS Truewind Summary Many Factors Contribute to Sensor-Related Uncertainty Sensor Uncertainty Can Be a Modest Contributor to Overall Energy Assessment Uncertainty... if Managed Properly Sensor Uncertainty Can be a Significant Source of Overall Energy Assessment Uncertainty... if Not Managed Properly 2007 AWS Truewind. LLC lie ™ American_Wind nergy Association — ——— Identifying and Reducing Wind Measurement Bias and Uncertainty: Siting Tuesday, September 18 3:15 pm — 5:30 pm Speaker: Ron Nierenberg Consulting Meteorologist Siting and met tower placement to reduce bias and uncertainty Macro and micro views on |] Nicrenberg Consulting |} Meteorologist Macro View = Identify Windfarm Sites « Favorable terrain, ridges perpendicular to the flow, avoid downwind blocks Nicrenberg Consulting }} Meteorologist = Identify Windfarm Sites «Favorable terrain, ridges, blocks « Proximity to transmission lines ¢ Within 10 miles + Appropriate voltage ~115 kV Nicrenberg Consulting || Meteorologist @ Identify Windfarm Sites «Favorable terrain, ridges, blocks ¢ Proximity to transmission lines « Avoid land use conflicts/permitting Nicrenberg Consulting }} Meteorologist = Identify Windfarm Sites ¢ Favorable terrain, ridges, blocks @ Proximity to transmission lines « Avoid land use conflicts/permitting « Low surface roughness on Nicrenberg Consulting }} Meteorologist = Ranking Sites « Estimate of wind speed class on }] Nicrenberg Consulting |{ Meteorologist = Ranking Sites « Estimate of wind speed class « Estimate of size of resource area on }] Nicrenberg Consulting }} Meteorologist = Ranking Sites « Estimate of wind speed class « Estimate size of resource area « Uncertainty of resource, any data available? on Nicrenberg Consulting |} Meteorologist = Ranking Sites « Estimate of wind speed class Estimate size of resource area Uncertainty of resource ¢ Construction and access; «Razor ridges, steep terrain? on |] Nicrenberg Consulting |} Meteorologist = Ranking Sites « Estimate of wind speed class « Estimate size of resource area ¢ Uncertainty of resource ¢ Construction and access « Distance to power line, voltage Ron Nicrenberg Consulting |} Meteorologist = Ranking Sites « Estimate of wind speed class « Estimate size of resource area « Uncertainty of resource « Construction and access ¢ Distance to power line, voltage «Low roughness, terrain complexity Nicrenberg Consulting || Meteorologist Micro View = Atmospheric Stability and Terrain Effects @ Stability = lapse rate, how heavy is air ¢In stable air, there are no vertical movements. Can be thought of as “heavy” air, does not want to go uphill. « Unstable air ~ rainy conditions Nicrenberg Consulting }] Meteorologist = Atmospheric Stability and Terrain Effects ¢ Stability = lapse rate, how heavy is air « Flow goes around terrain in stable « Flow goes over hills in unstable, neutral on Ni¢renberg Consulting |} Meteorologist = Atmospheric Stability and Terrain Effects ¢ Stability = lapse rate, how heavy is air «Flow goes around terrain or over « Downslope acceleration, leeside ridges are favored in stable flow Nicrenberg Consulting }} Meteorologist = Atmospheric Stability and Terrain Effects @ Stability = lapse rate, how heavy is air «Flow goes around terrain or over « Downslope acceleration Frequency of stable, neutral, unstable ¢ Neutral =5%, Stable = 75% in CA «Neutral = 7%, Stable = 53% in WI on Nicrenberg Consulting || Meteorologist Many models restricted to neutral stability. Neutral stability: winds highest on top of ridge Stable flow: highest winds could be in gaps on Nicrenberg Consulting }} Meteorologist Neutral stability: winds highest on top of ridge Stable flow: highest winds could be on downslope on |] Nitrenberg Consulting }j Meteorologist Neutral stability: winds highest on top of ridge Stable: asymmetry, upwind hill less windy on Nierenberg Consulting }} Meteorologist = Real world and the modeled world «Importance of the flow exit: ¢ Model looks at inflow, but the exit is probably more important than inflow Ni¢renberg Consulting |] Meteorologist = Real world and the modeled world @ Model will tend to dampen wind speed gradient due to terrain, so terrain impacts are under estimated. Nirenberg Consulting |} Meteorologist 10 = Real world/modeled world: Solutions «GH recommend met tower within | km of every turbine, or about every 5 to 8 turbines, or one per mile Ni¢renberg ing |) Meteorologist = Eliminating Model Bias «Some models dampen terrain effects, it is critical to site met towers carefully «If met tower is in best terrain, the model will over predict wind speeds throughout the windfarm and if in poor terrain, it will under predict Nicrenberg Consulting |] Meteorologist 11 Eliminating bias and uncertainty: « Measure near hub-height, diurnal effects Diurnal Summary By Parameter Chandier, MN May 2, 1996 -Feb 4, 2003 = Siting met towers to lower bias/uncertainty ¢ Visualize or create turbine layout first, to site met towers on Niérenberg Consulting || Meteorologist 12 = Siting met towers to lower bias/uncertainty ¢ Visualize or create turbine layout first, to site met towers « Site towers where turbines are most heavily concentrated on Ni¢renberg Consulting }j Meteorologist = Siting met towers to lower bias/uncertainty « Visualize or create turbine layout first, to site met towers @ Site towers where turbines are most heavily concentrated « Site towers in most representative areas, avoid best looking locations on Nicrenberg Consulting |} Meteorologist = Siting met towers to lower bias/uncertainty ¢ Visualize or create turbine layout first ¢ where turbines are heavily concentrated ¢ Site towers in most representative areas « Choose one for expected worst site, to define lower bound of the resource on Nicrenberg Consulting || Meteorologist = Siting met towers to lower bias/uncertainty @ Visualize or create turbine layout first @ where turbines are heavily concentrated ¢ Site towers in most representative areas «@ Choose one for expected worst site « Measure vertical wind speed component on steep slopes on Niérenberg Consulting |} Meteorologist 14 m= Siting met towers to lower bias/uncertainty ¢ Visualize or create turbine layout first @ where turbines are heavily concentrated Site towers in most representative areas Choose one for expected worst site @ Measure vertical component @ Measure close to hub height to minimize errors due to shear and diurnal effects on {| Nicrenberg Consulting }} Meteorologist m= Siting met towers to lower bias/uncertainty @ Visualize or create turbine layout first where turbines are heavily concentrated @ Site towers in most representative areas Choose one for expected worst site Measure vertical component ¢@ Measure close to hub height to minimize errors ¢@ Consider Sodar to measure over entire disk Ron |] Nicrenberg Consulting }} Meteorologist n = Windfarm Stats: Capacity = 4 MW 33.3% Nicrenberg Meteorologist 16 ™ __American_Wind Znergy Association Identifying and Reducing Wind Measurement Bias and Uncertainty: LIDAR Tuesday, September 18 3:15 pm — 5:30 pm Speaker: Peter Clive SgurrEnergy, Ltd Mrrr sgurreNERGY Highlighting uncertainty with Lidar Peter Clive, Technical Development Officer, SgurrEnergy Ltd Lidar is a mature remote sensing technology now successfully being used in wind power applications. A compact portable ground-based device can be rapidly deployed to acquire wind data across the entire rotor diameter. Deployments can be made to otherwise inaccessible locations, e.g. forestry, operational turbines. sgurren Wind resource assessment Introducing Lidar Results Wind profiling Limitations of modelling Uncertainty Acceptance criteria Lidar benefits www.sgurrenergy.com Are SgurreNERGY SgurrEnergy Ltd is a leading engineering consultancy with worldwide experience in renewables. Our capabilities encompass the full lifecycle of renewable developments, from feasibility and resource assessment through to post-investment analysis, appraisal and due diligence. inception B_J development P_J implementation J operation SgurrEnergy is an independent Lidar practitioner. Lidar supports the delivery of highly enhanced wind power assessment services by SgurrEnergy. Raraiees Cu ate haere Arar SQGUTENERGY Now: one or a few masts, limited in height + Wind flow modelling fills the gaps between masts + Extrapolation of theoretical wind profiles give wind speeds at heights across the turbine rotor « Restrictions in time of deployment e.g. planning permission The impossible dream: an infinitely tall mast whenever and wherever we want e.g. every proposed turbine location * Wind veer and wind shear measured to top of rotor diameter + Measurements at locations inaccessible to masts + No planning permission Aree sgarres A laser is shone upwards and light reflected by microscopic airborne particulates is detected. Emitted laser beam inclined Beam swept around cone to vertical by a wedge by rotating wedge, scanning the air www.sgurrenergy.com 1: The laser is scanned around a cone 3: For each of these 50 positions the to give a series of values from which the spectrum of the reflected light is wind velocity is derived. measured. win ug 4: The Doppler shift indicates the radial velocity of the particulates and hence the wind carrying them. 2: The reflected light is detected at 50 positions around the cone. ‘ata tom “Measurement ot Turbulence wth» CW Liar EMects of Conca scanning and Probe Volume” Torben Maasiuon and Hane E_ Jorgensen, IA Remote Sensing Exprts Mesting, Riss January 2007 www.sgurrenergy.com Arar SQUITE NERC A polar plot of 50 radial velocities per 1 second scan produces a characteristic “figure of 8” plot aligned in the wind direction. All three wind velocity components, u, v and w, can be obtained from this. Doppler shift => radial velocit DBNOAPRWON >= Crosswind beam +4 Crosswind beam Beam downwind q Time => angular position of beam on coné=> Data trom “Measurement of Turbulence wih 8 CW Liar Emecta of Conical scanning and Probe Volume”, Torben MikReigen and Mans Jorgensen, IEA Remote Sensing Experts Meting, Rise, Jnuary 2007 www.sgurrenergy.com Aras SQUITENE Scans 360° round the cone in one second Takes 50 Doppler spectra => 50 radial velocities in 50 directions Works out wind velocity and turbulence from radial velocities Averages over three scans Next height selected by focussing detector optics Cloud correction at 300m removes spurious returns Produces 10-minute averages and raw data Data download via GSM Data export as CSV SAEs) Aet wind direction ZephIR |° Aran sgarres www.sgurrenergy.com 4 124m * 65m Mawr SQUITrENERGY y=0.994x + 0.256 R?=0.994 210 wind direction, vane on mast [°| www.sgurrenergy.com Aran ee sgurre NE 10 minutes mean horizontal wind speed profile Wind shear profile obtained from a site in Spain. Wind shear profiles can be measured rather than simply extrapolated from mast data, and the cases where they deviate from standard extrapolations identified. height [m] 6 6 7 8 9 0 11 12 13 14 15 mean horizontal wind speed [m/s] www.sgurrenergy.com Arar : sgarr Lidar allows easy wind veer profiling. ‘ae cla wand ersetin oxihoighd Wind directions change with = height. As hub heights increase effects such as the Ekman Spiral begin to become Wind veer profile significant. ‘ obtained from a site in France. Wind veer representing a variation in wind direction of 20° across the rotor diameter has been measured, which seriously impacts predictions of turbine performance. 180 200 wind sirection (deg) Mra SUITE NE RGY Lidar is a uniquely valuable tool for micrositing and model verification. Over-reliance on linear models can lead to incorrect estimates of wind resource. Non-linear CFD models should be adopted and validated using lidar measurements. Wind speed at 60m above ground normalised to wind speed 2km upstream of hilltop ro Linear and non-linear CFD modelling of flow near a Gaussian hill, maximum slope 22°. www.sgurrenergy.com Wind speed resolution The use of light, rather than sound, as a probe allows many more spectra (e.g. 200,000 s*) to be accumulated in a given time period, improving accuracy. In addition the acquisition of 50 rather than the minimum 3 radial wind velocities per height gate imposes a redundancy that improves accuracy. Pulsed lidars’ finite pulse length imposes an intrinsic resolution limit ~ 0.1 m/s. Spectral broadening is mainly a consequence of turbulence in the averaging volume, c.f. measurements at a point. Acquisition of radial velocities at different positions on scan cone also introduces space-domain turbulence related uncertainty. However, turbulence related effects may be a “feature” rather than a “bug”. www.sgurrenergy.com Height resolution Pulsed lidars have constant height resolution ~20 m determined by pulse length CW (continuous wave) lidars rely on focus of detector optics to resolve height. Depth of focus varies with height, altering the effective probe length. Finite height resolution with non-linear wind profiles can lead to bias. Are SgUrre NEAGY Sensitivity of cw lidar to reflected signal for three different height/focus settings Sensitivity (rel. to peak) Hyperfocal distance = 1250m ‘ata rom -Wind ba evaluation a the Canis at ste at Hoveare™ Davi A Smut, Michal Hari Naar Cut} aeety Volume average bias resulting from logarithmic wind profile Assuming neutral stability Neglecting zero-plane displacement Z = roughness length A height g = height gate size (pulse length or depth of focus) Bog = ratio of volume and point values (log profile) --yBiay juasedde. Geometric interpretation www.sguirenergy.com 7 NY Sgurre Net Orientation/position tolerances Lidars can be located next to a mast due to redundancy of data acquired and absence of fixed echoes. They have successfully been used in various orientations, from vertical to horizontal (e.g. staring mode on turbine nacelle). Environmental factors Cloud correction is necessary for cw lidars. Data processing Signal to noise ratios deteriorate above 150m. Deployment duration “Representative” data? www.sguirenergy.com Mar sgurren 3 Vector average/scalar average (constant wind speed and uniformly distributed wind direction) Wind direction standard deviation (for a uniform distribution) in degrees www.sgurrenergy.com Ara SQGUrrENEF Empirical results agree reasonably well - R? = 0.8595 Vector averages accumulated over long periods characterised by large o,,, may cause concern, but wind vectors calculated each second are characterised by low 4), rection (Degrees) Empirical data obtained from a prop and vane instrument courtesy of Kathy Moore www.sgurrenergy.com Arar SQGUITENERGY Correlation to mast measurements A key finding is that if a lidar disagrees with a mast mounted cup anemometer it is because it is measuring the variation in wind flow between the two locations, as required. Poleronce Mast im’s) 60m from mast R? = 0.97 Same lidar now 10m from mast R? = 0.99 ‘ata rom “Practica Experience wth Remate Sensing A Consultancy Perspective” Nol Douglas. IEA Remote Sensing Experts Metng Risa January 207 10 Aree sgurrenerRGy Lidar can be deployed where masts cannot, for example in operational wind farms to investigate turbine power performance + Pup +P Zephi® * Cp cup * Cp Zaphik Windguard power curve measurement using cup and lidar anemometry 6 8 ¥, 65m [m/s] Note that the uncertainties associated with the lidar derived power curve are less than those associated with the cup. This suggests that the lidar introduces less uncertainty into the analysis, i.e. lidar is better than cup anemometry www.sgurrenergy.com Meer SQUCrENERGY The following criteria for acceptance tests have been adopted and easily satisfied in the past by lidar operators under contract 1) 2 week data period 2) R2 value on wind speed correlations >0.96 3) Slope of wind speed correlation: 0.97<x<1.03 4) RMS on wind direction difference <5° 5) Units located adjacent to a tall mast (>40m): * Calibrated instruments + Mounting in accordance with IEC Pt .11 + Sited to minimise differences in wind between locations In addition data availability >95% is typical. The IEA expert committee, on which SgurrEnergy is represented, will publish guidelines by the end of the summer. www.sguirenergy.com 11 Aree SYUNTENE AC + Drives down energy yield prediction uncertainty and reduces the cost of risk mitigation. + No planning permission is needed to deploy Lidar and begin wind resource assessment. + Measurements at multiple locations to supplement mast measurements and validate models. + Deploy to locations inaccessible to masts, e.g. investigate wind shear over forestry. + Wind speeds, wind veer and wind shear profiles at heights up to 150m can be measured. + Investigation of turbine power performance, possibly supporting claims under contract. + Allows investigation of turbulence upstream and in wake of operational turbines. + Rapid short term deployments can be made supporting Noise Impact Assessments. www.sgurrenergy.com sgurr: Questions? peter.clive@sgurrenergy.com Mar 12 Meteorological Tower Configuration and Uncertainty David Baker, President, Phoenix Engineering Inc. AWEA Wind Resource & Project Energy Assessment Workshop Portland, Oregon September 18, 2006 proenitensincerin Company Profile prrenistencineerin Phoenix Engineering ¢ 20 years experience in wind resource analysis and project design ¢ Designed over a third of Canada’s installed wind energy capacity ¢ Project engineering and grid assessments ¢ Active in all Canadian provinces and over 15 US states ¢ Incorporated in both Canada and the US ¢ WindServer™ stores data from over 400 sites and actively receives data from over 200 sites a Presentation Scope ¢ Understand wind flow in vicinity of met tower ¢ Uncertainty and best practices concerning: * Boom length and orientation * Redundant and multiple sensors ¢ Sensor orientation and vertical spacing ¢ Field measurements ¢« Demonstrate poor practices * Quantitative examples ¢ Other Considerations * Boom types, fasteners, lightning protection * Quality Control: monitor and analyze yd S Overview are Addressing Met Tower Uncertainty * Sources and effects of tower uncertainty * Tower Shadow and effects ¢ Shear and effects ¢ Anemometer and effects ¢ Wind vane and effects * Other: boom, guys, cables, lightning * Quality control and data validation ¢« Summary of uncertainty and best practice Concerns and Effect Uncertainty Concerns & Configuration Effect ie Tower Shadow *Boom lengths and offsets «Redundant anemometers eWind vane orientation Shear «Redundant anemometers *Top level sensor configuration Anemometer *Quality control related to tower Behaviour shadow Wind Vane eWind vane orientation Behaviour *Boom orientation *Quality control 5 BET (ony prrenistensineerins Tower Shadow: Basics - Air flow in vicinity of a tower does not represent the free stream Wind speed measurements downstream of tower are unusable Speeds are influenced all around the tower, not just downstream + The behaviour is different for tubular and lattice towers The goal: separate anemometer from tower to achieve an acceptable amount of uncertainty Magnitude: Up to 3 m/s error for shaded directions Tower Shadow Tower Shadow: Types of tower + Two types: ¢ Tubular tower + Lattice tower + Important to distinguish because aerodynamics in proximity of these structures differs considerably e 7 Tower Shadow Bere Tubular Tower Tower Shadow Lattice Tower 2 9 Tower Shadow cere a Tower Shadow: Tubular + lIso-speed plot, with local speed normalized by free stream wind speed of flow round a solid cylinder cross section; analysis by 2 dimensional Navier-Stokes computation Source: Recommended Practices for Wind Turbine Testing and Evaluation, IEA, 1999 A 10 Tower Shadow Tower Shadow: Tubular + Retardation upstream, acceleration on the sides, wake downstream + Least disturbance 45° from wind direction — therefore pointing the boom directly into prominent wind direction not best arrangement + Typical upstream center line deficit: + ~1% deficit for boom length 6 times mast diameter* + ~0.5% deficit for boom length 8.5 times mast diameter* + Recommended Practice: + Boom length: at least 7 times mast diameter - Redundant anemometers with boom offset of at least 135° + Quality control to remove data collected from the wake * Source: Recommended Practices for Wind Turbine Testing and Evaluation, IEA, 1999 Tower Shadow ee Tower Shadow: Lattice + Iso-speed plot, with local speed normalized by free stream wind speed of flow round a triangular lattice mast; analysis by 2 dimensional Navier-Stokes computation and actuator disc theory Source: Recommended Practices for Wind Turbine Testing and Evaluation, IEA, y 4 1999 # 12 Tower Shadow Tower Shadow: Lattice + Analysis depends on solidity of mast, drag of individual members, the orientation of the wind + Least disturbance at 90° - mount sensors 90° to prominent wind + Typical upstream center line deficit for low porosity: + ~1% deficit for boom length 3.7 to 5 times face length* + ~0.5% deficit for boom length 5.7 to 7.1 times face length * + Phoenix has observed larger field of influence + Recommended Practice: - Boom length: at least 7 times mast face length + Phoenix recommends much longer than 7 face lengths - Redundant anemometers with boom offset of 180° + Quality control to remove data collected from the wake * Source: Recommended Practices for Wind Turbine Testing and Evaluation, IEA, 1999 BD 13 Tower Shadow utes engineering Tower Shadow: Lattice + Caution: previous analysis only pertains to triangular lattice, face on to wind, specified lattice porosity (thrust coefficient C;=0.486) + For other dimensions and C,, IEA empirical equation for center line up stream velocity deficit*: A=(0.126C, - 0.006) { 40.08} inh Empirical method for determining C + Where: + Lis the tower face width + Ris the boom length + C, is the tower drag per unit length divided by dynamic pressure and L * Source: Recommended Practices for Wind Turbine Testing and Evaluation, IEA, 1999 Ba 14 Tower Shadow Tower Shadow: Lattice + Danish code prescribes empirical calculation for C; + Definig tas ratio of projected area of all structural members on side of tower to total exposed area, C; can be evaluated”: Type of Tower Expression for C, | Valid Range Square cross section, members _ with sharp edges 4.4(1 ty 0.1<t<0.5 Triangular cross section, round 2 (1 _ ty 0.1<1<03 members . : i Square cross section, round members 2.6(1 7 ty 0.1<1<0.3 * Source: Recommended Practices for Wind Turbine Testing and Evaluation, IEA, 1999 yd 15 Tower Shadow Per onty Tower Shadow: The Bad Tower Shadow (Unfiltered) é k ‘ ‘ 345° or 15° 300° 285° 270° 105° 240° “eee 120° 195" “ygoe «165° Lattice tower with extremely short boom al 16 Tower Shadow Tower Shadow: The Good Tower Shadow Report 345° S 15° 285° 270° 255° 240° 195° Lattice tower with extremely long booms (8.53m or 28’) y 4 Shear prrenistencineerin Shear: Basics + Wind Speed increases with height and assumed to vary exponentially as follows: ; In(U,, /U,,) Z, Zz Z, U, =U, |— Ch ee 7 “| In(z, /z,) Extrapolation of wind speeds to hub height dependent upon shear profile Uncertainty increases with extrapolation distance Shear profile may differ across the site and vertically Shearing up of data requires careful attention to seasonal, diurnal and directional dependencies + Magnitude: Varies from site to site, 1% to 3% uncertainty on wind speed typical 22 18 Shear: Basics + Recommended Practice + Anemometers at multiple heights (typically 3) - Accurately measure height of anemometers + Redundant anemometers at each height (follow tower shadow guidelines) +. Compare shear with all sensor combinations (top and middle; top and bottom; middle and bottom) + Compare shear results to evaluate vertical trends - Larger vertical spacing reduces effect of wind speed uncertainties in shear calculation a 19 Shear prenistensineering Shear: Quantitative Example > [Sm - Assume shear exponent and velocity at 60m: Soper” . a=0.14, U(60m)=7.00m/s ———> U(80m)=7.29m/s un + Calculate actual velocity at 50m and 45m or a ¢ LL U@)-v6) =| oe Zp U(50m) = 6.82m/s U(45m) = 6.72m/s + Assume 2% error on velocity and calculate shear (i.e. top anemometer over-reads and lower under-reads) _ Inf +eW, d-e)U,] - In(z, /z,) Ceo = 0.36 ——+ U(80m)=7.76m/s Oep_45 = 0.28 ——+ U(80m)=7.59m/s + Therefore, error reduced by increasing vertical spacing y 20 2-1 Shear: Quantitative Example Cont'd. - As before: a@=0.14, | U(60m)=7.00m/s > U(80m)=7.29m/s + Now assume -2% error on velocity and calculate shear (i.e. top anemometer under-reads and lower-over reads) _ Inf +e, /-e)U,] a4 In(z, /z,) Oep_5 = -0.079 ——» U(80m)=6.84m/s Qep_45 = 0.00094 ——+ U(80m)=7.00m/s - Again, error reduced by increasing vertical spacing yd 21 Shear ace Shear: Quantitative Example Cont'd. Now assume lower level sensor is shaded with 2m/s under-speed and top level sensor is not shaded + As before: a=0.14, U(60m)=7.00m/s ——-> U(80m)=7.29m/s + Shaded velocity at 45m: U snaiea (45m) = U(45m)—2m/ 5 =4.72m/s + Calculate shear with velocity error: 7 In[U(60m)/U adea (45m) U staat (45m)] —— Factor of 10 error 60-45 = In(60/45) + U(80m)=10.37m/s =1:37 + Therefore, redundant sensor at each height is vital ; 22 Anemometer Anemometer + Consider two common types: cup vs. propeller + Anemometers carry their own uncertainty Magnitude: Ranges from 1% to 2% uncertainty on wind speed where best practices followed Different behaviour in icing environment + More relevant, they have different behaviour in tower shadow Anemometer 23 eee Anemometer and tower shadow + Consider cup anemometer: not symmetrical Direction Description Result -20 WD 1 Cone shaded, cup exposed Over speed WD 2 Both shaded Under speed WD 3 Cup shaded, cone exposed Max under speed Wind speed difference vs. direction Anemometer ee cn Anemometer and tower shadow Consider propeller anemometer: symmetrical wd 3—__,, wo2-—> wot” Wind speed difference vs. direction Direction Description Result 05 WD 1 Propeller partially shaded Under speed oe WD 2 Propeller fully shaded Max under speed WD 3 Propeller partially shaded Under speed Wind Vane prenistensineering Wind Vane: Basics * Operating Principle: potentiometer assumed to vary resistance linearly with angular positions. + Source of uncertainty: + Dead band width (region of null signal) + Non-linearity of potentiometer (worst near dead band) + Nominal resistance of potentiometer Published accuracy of common vanes: + NRG #200P: + 4°, 1% non-linearity, 4° to 8° dead band « RM Young #5103: + 3°, 0.25% non-linearity, 5° dead band + Met One 020C: + 3°, 0.5% non-linearity, 3° dead band a 26 Wind Vane: Basics + Additional uncertainty + Sensor positioning on boom: positioning the north marker relative to an absolute reference + Caution: Lattice towers can alter the wind direction Mitigation: + Phoenix positions dead band parallel to boom and facing tower — accurate portion of sensor towards free stream « Necessary to accurately measure boom heading + Multiple sensors at different levels * QC to adjust wind vane offset 27 Other Concerns ee aE Solution to Other Sources of Uncertainty + Booms + Use stiff booms with rigid attachments to the tower + Flow disturbance: post height 12-15 times boom diameter + Rigidly attach sensors to booms « Accurately measure boom headings + Guy Wires + Avoid positioning sensors near tower guy wires + Sensor Cables - Wrap cables around tower and secure tightly + Vertical mount, top level sensors - Flow accelerates over top of tower: separate sensors + Lightning Protection + Separate lightning rod from top level sensors ; 28 Quality Control Quality Control to Mitigate Tower Influence + Monitor all sensor time series to identify developing problems - Anemometer Quality Control: + Ensure redundant sensors agree + Ensure sensors at different levels provide reasonable shear - Remove tower shadow data (+15° of boom direction) + Wind Vane Quality Control: + Comparison between multiple sensors + Phoenix also uses tower shadow graph: 1. Approximate boom direction from tower shadow plot 2. Compare to actual boom direction measured in field 3. If there is a discrepancy, adjust offset on wind vane data + This addresses uncertainty in dead band location 29 Quality Control pees re Sy Quality Control: Example 1 « Approximated boom directions: 65° and 245° « Actual boom directions: 55° and 235° + Therefore, offset wind vane data by -10° Shadow Graph Sample & & + Wind Speed Difference: WS1 - WS2 (m/s) 6 = & & fai o e 8 8 180 65° ‘ Wind Direction rare) se BD 30 8 8 Quality Control Quality Control: Example 2 + Wind direction differences vs time series Ditetace im WO ad Hm WO rece on WO set 30 WO Daca et WO wd 30m WD Se A750 Db SAY BH Ae 2D wna BD SANS 9 fo VAN 2S ST ta a ee Das Reco a | 60mWD — 45mWD 60mWD — 30mWD 45mWD — 30mWD * Conclusion: 45m WD boom shifted over time then was corrected 2B 31 Reducing Uncertainty cere Uncertainty can be reduced through: + Multiple Sensors + Redundant anemometers with appropriate offsets + Sufficient boom lengths + Vertical spacing between anemometers + Proper wind vane orientation + Stiff booms and rigid connections + Consideration of lightning protection, guy wires, cables + Frequent and diligent quality control 32 Summary Addressing Uncertainty ¢ Uncertainty in energy assessment is inevitable * Itis important to identify sources of uncertainty and take steps as part of the site monitoring and assessment to minimize uncertainty ¢ Uncertainties from various sources can be quantified ¢ These results can be used to better understand project feasibility and financial risks 33 Thank You eee eens, engineering ppreenidtcorsatin USA Phoenix Consulting USA Phoenix Engineering Inc. #103, 2710 - 3 Avenue NE 22302 Moming Lake Drive Calgary, Alberta, Canada T2A 2L5 Katy, Texas, USA 77450 Phone: 403.248.9463 Fax: 403.250.7811 Phone: 1.866.558.9463 34 AWEA Wind Resource and Project Energy Assessment Workshop September 18-19, 2007 Portland, Oregon Reducing Uncertainty in Sodar Measurements Jerry H. Crescenti PPM Energy “I’ve been every where, man...” » sodar — sound detection and ranging > Beeping since 1969 >» Used on every continent @ from the tropics to the tundra @ on ridge tops and in valley floors @ power plants to airports @ offshore platforms @ building tops > Applications @® regulatory air quality operational forecasting emergency response complex flow wind shear and wake vortex NASA launch support wind energy!!! > More than 400 scientific publications > _P Pr Manufacturers Pp» » Atmospheric Research & Technology LLC SC » Atmospheric Research Pty Ltd . 7 ae >» Atmospheric Systems Corporation > Metek > NRG Systems > Remtech > Scintec > Secondwind — Ei | Mention of trade names or commercial products does not constitute endorsement or recommendation for use. “Heavy Duty” Theory of Operation rR cCtTAG AR -2(ac + am) — Cc Am —Pr Ere —> OF Ae ER 2 R 1/3 Ci Cr oO = 0.0039 k P,,— received power P,,— transmitted power E, — received power efficiency E,— transmitted power efficiency c — speed of sound (~ 340 ms") t — pulse length (s) A — antenna aperture (m?) G — effective aperture factor R — range of scattering volume (m) a, — Classical attenuation (m-’) a, - molecular attenuation (m") a, — excess attenuation (m') o — scattering cross section (m*‘) k — acoustic wave number ( 27/2 ) dX — acoustic wave length (m) T — air temperature (K) C;* — temperature structure function “Practical” Theory of Operation >» Acoustic backscatter by: @ small-scale potential temperature gradients @ inversion layers @ wind shear layers @ thermal plumes > Frequency: 1 - 5 kHz (34 - 7 cm) > Minimum three beams to derive u, v, w = Noise power = Average Noise = Doppler frequency N N fo P, = Signal Power W, = Spectral width Af = Frequency step size ~ Nyquist Frequency N Pp Data Validation / Verification Pp > Tower-based measurements @ in-situ ; ca @ scalar average | : a ; @ flow distortion wo | 1 Aine @ turbulent overspeeding pooyucermeene q @ flow inclination Fo bees > Sodar-based measurements = }----/ — @ volume —— Pema @ vector average oe bee @ noise interference \ i @ precipitation \ ; “Fateewrat > Consider spatial variability ¥ @® sensor exposure vv @ atmospheric stability @ complex terrain Pay >, Useful (and /mportant) Sodar Data Pp Horizontal wind speed (ws) Horizontal wind direction (wd) Horizontal wind speed components (u, v) Horizontal wind speed component standard deviations (o,, o,) Vertical wind speed (w) Vertical wind speed standard deviation (o,,) Signal-to-noise ratio (SNR,, SNR), SNR,) Signal intensity (I,, l), ly) Noise intensity (NOI,, NOI,, NOI,) Number of qualified pulses (N,,, N,, N,) Reliability number VV VV VV VV VV WV Pp» Siting Considerations >» Representative location ® acquire data to quantify wind resource and its characteristics @ homogeneous versus complex terrain > Site logistics ® accessible and secure @ level ground ® sufficient drainage @ clear of obstructions @ adequate electric power @® data communications link >» Meteorology tower @ 50 to 60 m tower @ near surface measurements > Time and personnel @ less than one day @ one to two persons > Tools @ digital level @ GPS, compass and/or transit @ voltmeter and oscilloscope >» Antenna orientation @ firm ground or concrete pad @ tie downs & anchor points @ level antenna to +0.5° @ determine azimuth angle to + 2° > Site documentation @ latitude, longitude, elevation @ 360° panoramic photo vista @ site diagram, vista table, log book Installation Deo aocties nauclbook VP Site dd< Maintenance > Site and shelter ®@ overall integrity and security >» Antenna orientation @ periodic check of level and alignment > Computer @ disk space @ clock @ modem > Antenna electronics @ speakers and amplifiers @ ® cables and connectors iN > Antenna and acoustic shield ® acoustic foam lining @ water, snow, ice, dust, sand, other debris ® animal infestation Active Broadband Noise > Wide range of frequencies random or white noise low frequencies > Decreases SNR decreased range backscattered signal biased toward zero > Diurnal, weekly, and/or seasonal patterns > Examples highway and road traffic machinery, industrial facilities, power plants, airplanes generators strong surface wind rain mice Pp» Active Narrowband Noise > Fixed-frequency > Misinterpreted as valid Doppler-shifted frequency e erroneous wind values > May “saturate” signal e no valid wind values > Examples e trucks and forklift beepers e birds e insects (e.g., crickets) Pp» Combating Active Noise > Noise survey @ diurnal and weekly patterns > Qualitative noise survey @ identification of sources > Quantitative noise survey @ noise level meter (< 50 to 60 dB) @ spectral analysis software (e.g., Spectrogram) > Conduct “listen-only” mode @ does the sodar derive a wind profile? > Change transmit frequency > Multi-frequency pulse Spectrogram Example a lei i i | a lL Nh A eM de Ll a * ee La] a] Ah] hi) ee " bY , : Canad — a ~— e - s e PB Be Se ge MT Lea ea ee Sl SS _ 7 Pas >>, Passive Noise Pp» >» Fixed-echo created when main beam or side lobes of acoustic pulse reflect off stationary objects > Return same acoustic frequency @ zero Doppler shift @ wind values of Oms"! > Examples @ buildings and towers @ transmission lines @ trees wy Combating Passive Noise Construct obstacle vista table Tilt oblique beams away from objects Avoid objects taller than 15° above horizon Realign sodar Algorithms to identify and remove fixed echoes May be unavoidable during strong inversions that create ducting Instrument AY Teneo NAN NO Coed Vertical Ang EVE wea Azimuth Reference: Frequency Averaging Interval Ut October 3, 2000 a SO eit Pry caste ets eT wad 00125 cxIRIRD 0 ih) ih) ys ih} 3000 Hz Bit CO) ee ea arto (deg) Site Name: a ce Ys red Petia Paes) VATU wat: Beam | etree TTS Parana Reed Azimuth Orientation Tool ie Zoie UTC Difference: te Seto WeSC pred em cis eee eA) oS ME MIO ELL On) AS eee MIO aay Bad oe en Pieces aaa cee le eves sco eae eee) ene el Ome meee ny dumpster at 100 m Otome ae ee LU tomer ean) aging Waters AAU PGA) .85 7.0 UU) clr ee N 1b ae SY PL Mitutoyo Pro 360 dig. level Garmin eTrex Gi Wa Bonita Other Notes open dirt lot Trem ot okies aiken ols Peat PCN RTS one MS SC ert nd SOM Tar Meet) | ReCoTTeS Orin imo open dirt lot nem os memos > Anechoic shields e reduce radiated side lobe acoustic energy e reduce received side lobe acoustic energy > Materials e acoustically absorbing foam e fiberglass e plywood with sand / lead filling e hay bales Find the sodar in the hay stack! Acoustic Shielding —_ > o o o = nN So c& Ss Height (m) zie a So > So 10 12 14 16 18 20 22 24 Hour (LST) o ND - o o "ew 180) a) 16m/s ~ a yD Pp 0.1 0.01 Attenuation (dB m*") S o °o = 0.0001 Attenuation as a Function of Frequency and Relative Humidity > Classical @ Viscosity, conduction, and diffusion @ «,~f(F) Molecular > Molecular Attenuation @ Excitation of O, @ «,,~f(F, RH) Classical @ Decrease range Attenuation > Exc ess @ Beam broadening and refraction @ a, ~f(F, U,T,, >) @ Reduction of backscatter 2 3 a 5 intensity Frequency (kHz) By Pp Example — Data Availability Profiles > p60 4500 HZ gp 1850 HZ g59 2125 Hz 140 500 500 120 400 400 _ 100 £ = 80 300 300 o ae 60 200 200 40 Relative Humidity —— 0% - 20% 20 100 —— 20% - 40% 100 60% - 80% 80% - 100% 0 0 0 0 20 40 60 80 100 0O 20 40 60 80 100 0 20 40 60 80 100 Availability (%) Availability (%) Availability (%) ~ yD >,» Example — Using SNR to Filter Sodar Data —~ 20 = A SNR >0 SNR > 20 SNR > 25 £ N=1242 N = 1237 N = 1220 SE = 1.43 m/s SE = 1.43 m/s SE = 1.19 m/s 3 15 tr = 0.74 15 [r= 0.74 15 fr = 0.83 2. no Ts £ = - o Ss fo] on £ So Ww SNR > 30 SNR > 35 SNR > 40 N= 1184 N= 1147 N=1112 SE = 0.83 m/s SE = 0.66 m/s SE = 0.54 m/s 15 fr = 0.92 15 fr=0.95 15 1r=0.97 10 10 50 m Sodar Wind Speed (m/s) 9 et | L J 9 & | J 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 50 m Tower Wind Speed (m/s) 50 m Tower Wind Speed (m/s) 50 m Tower Wind Speed ‘/~/s) 0 Height (m) 888 8B o xy s 160 Height (m) iD o & 8 ° o oe —_ Example — Resonance / Ringing . a - a ‘ sind L seth - i___ 3 6 9 12 15 18-15 -10 -5 0 5 10 150 05 1 1:5) 2125" 0 5 10 15 20 25 Wind Speed (m/s) U (m/s) o, (m/s) SNR, pharrell iasilsaaeacialie icicles Sti — aes = L ae meri el 90 180 270 36015 -10 -5 0O 5 10 150.5 1 15 2 2.00 5) 10 15 20 25 Wind Direction (deg) V (m/s) o, (m/s) SNR, 200 160 E 120 ze \ D> | ) ‘o 80 f x 40 0 _ _ 7 n —_i ees i al -2 -1 0 1 20 05 1 LOM CHeiOnO) 10 15 20 25 W (m/s) o,, (m/s) SNR,, Psy > pp = 160 — nN o Height (m) @ So eS o 200 160 = nN o ao oO Height (m) & o Example — Bad Amplifier x a wo | _ 13) | so | st _ 3 6 9 12 15-15 -10 - O 5 10 150 O05 1 15 2 250 2000 4000 6000 8000 Wind Speed (m/s) U (m/s) 6, (m/s) I, | | | | | 7 ; a Cs oe 90 180 270 360-15 -10 -5 0 5 10 150 05 1 15 2 250 2000 4000 6000 8000 Wind Direction (deg) V (m/s) o, (m/s) I 200 | 160 | ( ( | = | I | E 120 = | / D | ‘o 80. t 4 = ( | 40 ( | o 4 4) Ls Lo = — -2 1 0 1 20 O85 1 15 2 250 2000 4000 6000 8000 o,, (m/s) lL, Example — Fixed Echo | 3 | E 120 = | D> ® 80 a | 40 | 0 == — ft = =| | = = | = 0 3 6 9 12 15-15 -10 -5 0 5 10 150 05 1 #15 2 250 200 400 600 Wind Speed (m/s) U (m/s) o, (m/s) \, 200 160 | £120 5 D> ‘> 80. eis 40, 0 | SSS ee pecianieiiahilciiaciieaias 0 90 180 270 36015 -10 -5 O 5 10 150.5 1 15 2 250 200 400 600 Wind Direction (deg) V (mis) o, (m/s) \ 200 | 160 ( E 120 = D> } o> 80 ( x= y 40 0 aondics _ = —_ -2 -1 0 1 20 05 1 Sere 2.5.0 200 400 600 800 10001200 W (m/s) oc... (m/s) l > 200 | | 160 | E 120 | = | D o 80 = 40. | 0 Le i se Le 0 5 10 15 20 25-20 -10 0 Wind Speed (m/s) U (m/s) 200 160 | E 120 | =z a 3 80. = | | 40 \ | — 1 ee 0 90 180 270 360-20 -10 0 Wind Direction (deg) V (m/s) 200 , 160 = 120 =z D ‘oo 80. x= f 40. | 0 L — 5-4-3-210123450 05 W (m/s) 10 200 05 1 15 2 25 30 200 400 600 0, (m/s) I, | | | i ~ Lt _ a - 10 200 05 1 15 2 25 3 0 200 400 600 6, (m/s) Wy | | | L_ ao LO 115 2 25 30 200 400 600 6, (m/s) Iw 800 800 800 Pp» References > Sodar Best Practices (www.iedat.com/sodar.html) >» Standard Guide for Measurement of Atmospheric Wind and Turbulence Profiles by Acoustic Means (ASTM D7145-05, www.astm.org) > Crescenti, G. H., 1997: A look back on two decades of Doppler sodar comparison studies. Bulletin of the American Meteorological Society, 78, 651-673. > Crescenti, G. H., 1998: The degradation of Doppler sodar performance due to noise: A review. Atmospheric Environment, 32, 1499-1509. > Spectrogram (www.visualizationsoftware.com/gram.html) "Pp Summary > Complimentary technology to in-situ measurements > Quantify wind resource across blade swept area > Use “best practices” for siting, installation, and maintenance > Be mindful of active and passive noise sources > Use all available information to screen data >» Consider technology differences and spatial variability when intercomparing sodar with towers >» Understand sodar strengths and limitations Pas Jerry H. Crescent Manager, Meteorology PPM Energy 1125 NW Couch Street, Suite 700 Portlan R 97209 503-796-6997 (voice) 503-796-6907 (fax) 503-956-5434 (cell) Jerry.Crescenti@ppmenergy.com Data ANALYSIS =] 2 S s > 5 s < = a __American_Wind Znergy Association ilaatiedietiaeneaihdinn tedienmenenteniennneneneineietenehtiemmnmntetmiaenteneanenemmtnatn Introduction & Data Processing Wednesday, September 19 8:45 am — 10:15 am Speaker: Matt Hendrickson Horizon Wind Energy Wind Resource and Project Energy Assessment Workshop “Identifying and Reducing Data Analysis Bias and Uncertainty : Data Processing” Matthew Hendrickson “Do you see a man wise in his own eyes? There is more hope for a fool than for him.” -Solomon- Reducing Bias and Uncertainty: Data Processing Data Processing Objectives Data Processing Systems Things to Watch Out For Summary Data Processing Objectives - Turn this 2004-03-06, 14:00:00, 14.5,2.3,15.2,2.5,146,2.4,14.2.23,14.1,2.4,128,23,214,11,213,14,62.40.3 2004-03-06, 15:00:00, 12.9,2.0,13.3,2.1.12.9,2.0.12.6.19,12.4.20,11.5.2.1,202, 11,200, 14.63.9,0.2 2004-03-06, 16:00:00, 12.9,1.8,13.3,1.9,128,1.8,12.6.18,12.2,18,11.3,1.9.194,7,192,10,64.4,0.1 2004-03-06, 17:00:00, 12.6,1.2,12.9,1.2.12.4,1.1.12.2.12,11.8,12.9.8.1.3,188 4, 184,8,63.00.2 2004-03-06, 18:00:00, 122,0.9,12.5,1.0,12.1,0.9,11.8.08,11.3.08.8.7,1.0,180,4,166,4.54.8.0.5 2004-03-06, 19:00:00, 19.8,2.9, 18.3.2 5,18.8,3.4,18.4,3.1,16.4,3.3,15.0,3.1,254,11,78,17,49.205 2004-03-06,20:00:00,25.8,2.3,22.5,1.8,22.0,3.1,22.7.25,16 5.26.17 6,2.6,2.6,350,11,46.0.0.2 2004-03-06,21:00:00,20.7,1.5,19.9,1.4,20.2,1.6,17.1.16,15.4,2.1,11.7,1.5.14,2,13,6,41.5,0.3 2004-03-06,22:00:00, 19.4,1.0,17.8,1.0,15.7,2.1.15.5.10,108,15.9.8,1.0,3,4,347,6,36 7,0.3 2004-03-06,23:00:00,22.0,1.3,21.6.1.1.21.7,1.5,18.0.12,17.7,1.3.12.4,1.1,347,2,338,3,35 7.0.3 2004-03-07,00:00.00,22.6,1.2,22.4,1.1,22.4,1.2.18.4.1.1,18.5,1.3,12.6,1.2.341,2,335,5,35.80.2 2004-03-07,01:00,00,23.4,1.4,23.2,1.3,22.8,1.4,18.3,1.4,180,1.5,12.4,1.5,332_2,330,6,34.1,0.1 2004-03-07,02:0000,23.7.1.6,235,1.6,232,1.7.19.7.15,19.1,1.6,13.9,1.7,392,4,330,7,33.7.0.2 2004-03-07,03:00,00,22.4,1.6,22.3,1.5,22.0,1.7,18.7.15,18.3.1.5,135.1.7.329,3,328,7.33.40.2 2004.03-07,04:00:00,21.6,1.6.21.5,15.21.1,1.6.17.616.17.2,17,12.4,1.6,329,3,328,7.328.0.1 2004-03-07,05:00,00,20.0,1.5,20.2.1.4.19.5,1.5.16.3,1.4,15.8,1.5.11.2.1.5,327,3,328,6,32.00.1 2004-03-07 06:00,00, 18.0.1.4,18.6,1.4,17.3,1.5.15.0.12.14.8,1.2.10.5.1.2.313,4,302.6,30.7.0.2 '2004-03-07,07:00:00, 18.8, 1.6,19.3,1.6.17.9,1.7,15.6.1,7,150,1.7.113,17.317,4.3116,32503 2004-03-07,08:00,00,20 2:2 0,20 6,1.9,19.7,2.0,186.2 118.1,2.1115.9.2.3,323,4,320,7.37.9,0.3 2004-03-07,09:00,00,20 5.2 1,20.8,2.1,20.1,2 2,200.22 19.4,23.17.9.2.5,325,5,322,7.43.40.3 2004-03-07, 10:00,00,24 6,2.6,24.2.2.3,24.1,2.8,2400,2 723.4,30,21.3,3.0,335.6,392,10,47.40.2 2004-03-07, 11:00,00,23 5,2 8,23.0,2.7.225.3.5.23.1:30,21 8,3 7,20.6,3.3,341,8,330,12,49.5.0.3 2004-03-07; 12:00:00, 19.9.2 7,20.5,2.6,19.3,3.0,19.6,28,18.9,3.2.17.6,3.2.330.8,396, 11,508.03 2004-03-07;13:0000, 18.2:2.6,19.4,3.2.17.3.28.17.9,26,16 8.30.16 2,2.9,337,11,394,13,52.1,0.4 2004-03-07, 14:00,00,15.7,2.9.17.3,3.4.15.2,2.9.15.3,30,14.8,30.14.1,3.1,328,10,325,13,52.9,0.3 2004-03-07, 15:00:00, 14.2:2 7,16.2.3.4.13.7,2.9,14.0,29,13.6,29.129,29,323,10.321,13,53.7.0.3 2004-03-07 16:00,00, 12.2.2 1,13.3,2.4.12.32.2.120,22,12.2.2.2,11.1,2.2.301, 10.209, 11,53.8,0.2 2004-03-07,17:00:00,11.0,1.5,11.8,1.7.11.2,1.8,108.15,11.1,16.9.7,1 6,280,7,289,9,53.2.0.2 2004-03-07, 18:00,00,9.7,0.9,10.2,1.0,9.7,0.9,8.9.0.9.95,08,6.9.0.9,273,3,261,3,48.0.0.4 2004-03-07; 19:00:00, 10.0,0.8, 10.1,0.8,10.1,0.8,9.5.0.8.97,0.8,79,0.8,243,2,232,4,42.7,0.4 2004-03-07, 20:00:00, 12.4,0.7.12.9.0.9.125,08,11.8.08,12.0,08,8,7.0.8,238,1,231,2,40.6,0.2 '2004-03-07.21:00:00, 14 3,0.8,15.0.0.9.14.3,0.8,14.0.08,14.1,0.9,10.7,0.9,221,1,221,1,38.7,0.2 2004-03-07'22:00,00,19 5.0.8, 18.9,0.8,19.4.0.8,18.4,0.8,18.4,0.8 13.2.0 7,206,0,203,1/38.6,0.1 '2004-03-07.23:00.00,26.0,0.9,25.0,1.0,26.,0.9,21.6,0.9,21 4,09, 13.7, 1.0,208,0,201,4,38.0.0.1 2004-03-08 00:00:00.27 3,0,9,26.1,0.9.27 5,0.9,21.6,1.0,21.5,1.0,13.4,1.3,212,0,205,5,38.9.0.2 2004-03-08,01:00:00,24 3, .0,22.7,1.0.24 5,1.0,19.2.0.9,19.4,1.0,12.0,1.1,221,1.214,3,38.4.0.1 '2004-03-08,02:00:00,24.4,1.2,23.0,1.2,247,1.2,19.5.1.1.19.7,1.1,13.3,12,220,2.220,4,38.1,0.2 '2004-03-08,03:00:00,24.9,1.4,23.8,1.4,25.4,1.4,18.9.1.2,19.4.13,12.4,1.4,239,2.234,5,37 60.2 £2004-03-08,04:00:00,25.3,1.2.24.2,1.2.25.7.1.2.19.2.1.3,19.7,13,128,1.5,244,2.242,5.37.6.0.1 Data Processing Objectives - Into this Data Processing Objectives ¢ Product — “Cleaned” data set from which to launch further analysis ¢ Mission — To identify and flag “dirty” as much problematic data as possible ¢ Process — Utilize systems whereby every bit of data is examined in the proper context Data Processing Systems The Volume Problem — Typical Project [5 mets] * [10 sensors/met] * [4 fields/sensor] * [52560 records / year] * [4 years] = 42 Million pieces of information that have to be dealt with Data Processing Systems Effective System Design User Control Automation ous \aZ A Data Processing Systems ¢ A good data processing system — does enough automated work to leave the user with time to explore the quality of the data — Is transparent enough for the user to detect failures of the system Things to Watch Out For ¢ Level 1 Diagnostics — Sensor Failures — Icing ¢ Level 2 Diagnostics — Sensor Degradation — Time Stamp Errors — Switching Units — Incorrect Transfer Functions — Incorrect Boom Orientations — Crossed Sensor Cables — Condition Dependant Abnormal Sensor Behavior Summary Quality data processing is essential to the analytic process The most gains (in terms of reducing detectable error) can me made in a quality data processing system There are many subtle data problems that a purely automated system is unlikely to detect. Data systems must free the user for “Level 2” diagnostics The Bottom Line * A poorly designed, unwieldy data system leaves a lot of detectable uncertainty on the table. Time spent pushing data is time that could be spent finessing the data set. Questions? oN ™ __American_Wind Znergy Association MCP Methods and Long Term Adjustments Wednesday, September 19 8:45 am — 10:15 am Speaker: Gordon Randall Global Energy Concepts, LLC Uncertainty in Correlations and Long-Term Adjustments AWEA Wind Resource & Project Energy Assessment Workshop September 19, 2007 Gordon Randall Global Energy Concepts, LLC 1809 7'" Avenue, Suite 900 Seattle, WA 98101 (206) 387 — 4200 grandall@globalenergyconcepts.com oF CONCERTS Peensoe Synthesizing Between Towers on Site ¢ ... will not be the focus of this presentation ¢ Measure enough wind data to correlate by direction — Many sites have seasonal direction differences — In complex terrain, sites will look better or worse depending on wind direction and topographic effects — A few months of data may not be enough for accurate micrositing in complex terrain Long-term Adjustments: What the Uncertainties Are ¢ Period of record of data set * Accuracy of correlations — What's observable in measurements — What hasn't been measured yet * Consistency of reference site — Measurements — Exposure — Changes in wind over the period of record —GEC How Uncertainties are Calculated and Added ¢ Uncertainties generally given as one standard deviation (So approximately 95% of the time result will be within +/- 2 times the uncertainty value) ¢ Independent uncertainties added as the square root of the sum of squares — Large uncertainties quickly drown out others — Small uncertainties disappear, and reducing them further has negligible effect on total _2>»GEC How Uncertainty on Period of Record is Defined ¢ General equation: interannual variability (in percent) divided by the square root of the number of years in the data set Variability calculated as standard deviation in wind speed divided by the average - usually in the 4%-6% range for most of North America Twice the period of record equals 29% reduction in this specific line item uncertainty Accuracy of Correlations No simple, reliable statistic exists to evaluate whether or not a correlation is good A low R? usually means the correlation is bad, but a high R? may not mean it is good Definition of a “high” R2 can vary — what's better: — 0.90 R? on monthly data? — 0.75 R? on daily data? — 0.60 R? on hourly data? Does a Good (or Bad) Correlation Mean Anything? ¢ Time scales are important: — A good-looking correlation on monthly speeds may be close to meaningless — it may just reflect that the seasonal pattern is the same at both locations (e.g., winter peaking) Hardly anything will look correlated using ten-minute data, even between towers on site Hourly data won’t match well between near-ground reference stations and towers at 50+ meters Daily is usually happy medium for evaluating reference stations, but even that is tricky for references with less than hourly measurements Ability to Predict is Important Purpose of comparison to long-term data is to tell whether a period is above or below average, so what’s relevant is whether that can be accurately predicted — Are relationships consistent between years? — Can individual calendar months be predicted between years — winter months are “good wind months’ in much of North America, but is an individual January good compared to an average January? —GEC _ Example — Is this a Good, Consistent Relationship? y =1.10x + 2.37 R?=0.90 On-Site Wind Speed (m/s) 25 3.0 3.5 4.0 Reference Site Wind Speed (m/s) 08x + 2.66 R*° = 0.94 2 So aouaonso y = 1.13x + 2.06 | R? = 0.94 euiie a On-Site Wind Speed (m/s) a @ oO °o = a 2.5 3.0 3.5 4.0 Reference Site Wind Speed (ms) - ° at « First Year » Second Year — Linear (First Year) — Linear (Second Year) ROBAL EMERY CONCIS the Pe eee What's Necessary to Test Predictions? * Seasonal differences will make it difficult to predict winds accurately between calendar months — Seasonal differences in vegetation, direction, and other issues have big effects — Large (10% or greater) errors in predictions from summer to winter could still be OK Having two years of site data is very helpful — predict one year with the data from the other If much less than two years of data available, you start to run into problems —~GEC | Reference Site Consistency: Measurement Changes Little things can cause relatively big effects: — Changes in types of sensors — Changes in recording equipment/methodology In general, default assumption should be that a change makes a reference site inconsistent and inappropriate for use If sufficient site (or other nearby) data are available for periods before and after change, the consistency can be tested —~GEC Example Problem: ASOS Measurement Change Example Relationships: Before and After ASOS Change y = 1.38x + 1.35 R?=0.91 = : aon y=1.30x + 1.19 R’=0.91 Site Wind Speed (mis) a a oa > a 3.5 4 45 Reference Wind Speed (m/s) = * Before = After —Linear (Before) —Linear (After) | Reference Site Consistency: Changes In Exposure * Changes in location of reference sites ¢ Changes in the things around the sites that affect wind speed — Construction of buildings, fences, or other obstacles — Addition/subtraction of other equipment on towers (such as on communication towers) — Changes in vegetation (such as trees growing or being removed) — GEC Example of Changes in Exposure: Trees Growing Near Reference Station Effects of Changing Exposure: False Trends Annual Wind Speed (m/s) BORAT RARE COOKE Few Nene Blue Hill Observatory: 122-Year Dataset BLUE HILL OBSERVATORY ANNUAL WIND SPEED, 8.9 errr rereronrer renner rte 1s WIND SPEED (M/S) a8, 55 50° € asian 1880 1900 1920 v0 Maximum TA ees (16.6 mph), 1923. hisosu Minimo: SF (12.7 mph: 2007 and 2008 Record Mean: 6.7 ms (16.0 mph) Atcha J neon, Aienpha and RvironmenatRewearch Bw Hl Obeevatry Reference Site Consistency: Actual Changes in Wind * Temperatures have changed measurably over the last 50 years — winds may have too If conditions now aren't what they were 25+ years ago, don’t use data from 25+ years ago Greatest advantage to long-term datasets may be that they are long enough for you to conclude that you should not use them —*GEC Synthesize Datasets or Scale Observed Measurements? Different objectives: — Overall annual average wind speeds — Seasonal patterns — Time-of-day patterns — Hourly or ten-minute time series Different starting points: — At least one full year of site tower data: good — Less than a year of site data: fewer options GEC. 10 Problems With Data Synthesis * Near-ground reference stations will not capture hub-height time-of-day patterns Neither near-ground stations nor upper air stations will capture short-term variability It may be possible to correct for these, but if possible, don’t synthesize when not necessary Example: West Texas 2 —_—_—_> UCL at meee of Wind Speed (m/s) 1.2.3.4 5 6 T B 9 101112 13 1415 16 17 18 19 20 21 22 23 24 Hour of Day [Reference site 10m — On-site 80m or CONCEPTS SORA! ERE Deters omnes Problems With Scaling Data If you have significantly less than a year of on- site data, what do you do for the missing calendar months? — Seasonal differences in wind direction can be significant — something must be done to fill in — Shear may also vary by month, with varying vegetation, length of day, etc. Regional tall towers may help, even if no overlapping data — may help quantify seasonal effects, but will require data interpretation (i.e., “guessing”) with high uncertainty 11 Putting it Together - Examples of Uncertainty Introduction * How uncertainty is quantified: generally given in percent, representing standard deviation in wind speed At most sites with modern turbines, roughly a 2-to-1 ratio of uncertainty in energy to uncertainty in wind speed As approximate numbers, 1% higher total uncertainty on wind speed means: — 4.5% to 5% lower energy at the P99 level — 3% to 3.5% lower energy at the P95 level GEC Putting it Together - Examples of Uncertainty Case |: 1 year site data, 25-year reference site Uncertainty Site Data Only With Correlation Period of record + 4.0% 0.8% Correlation accuracy Not applicable 2.0% Reference measurement consistency Not applicable 1.5% Reference exposure consistency ee Long-term wind speed consistency Not applicable Not applicable 1.0% 1.0% Total 4.0% 3.0% 12 Putting it Together — Examples of Uncertainty Case ll: 4 years site data, 25-year reference site Uncertainty Site Data Only With Correlation Period of record 2.0% 0.8% Correlation accuracy Not applicable 1.0% Reference measurement consistency Not applicable 1.5% Reference exposure consistency Not applicable 1.0% Long-term wind speed consistency Not applicable 1.0% Total 2.0% | 2.4% Putting it Together — Examples of Uncertainty Case III: 3 months site data, 25-year reference site Uncertainty Site Data Only With Correlation Period of record 8.0% 0.8% Correlation accuracy Not applicable 4.0% Reference measurement consistency Not applicable Reference exposure consistency Long-term wind speed consistency Not applicable Not applicable 1.0% Total 8.0% | 1.5% 1.0% | 4.6% 13 Summary of Examples Improvement seen by using long-term correlation vs. just using site data: — With one year of site data, 25% reduction in uncertainty observed — good — With four years of site data, 20% increase in uncertainty — why bother? — With three months of site data, 42% reduction in uncertainty — nice, but you’re still in terrible shape Keep in mind these are just a subset of the uncertainties in the process — the smaller the numbers, the more they will be drowned out GEC _ Combined Uncertainties, Effects on P-levels Add these uncertainties to uncertainties associated with: — Anemometer accuracy Tower effects Data quality Wind shear Topographic effects Wind speed over project's life Future changes in wind speed Wind frequency distribution Power curve Losses Air density Ete. 14 Typical Case from Example |: 1 year Site Data, Varying Amount of Data from Reference, Long-Term Energy, 100 GWh/year P50 1 yr 5 yr 10 yr 20 yr 30 yr 40 yr 50 yr 100.0 100.0 100.0 100.0 100.0 100.0 100.0 92.9 93.4 93.6 93.7 93.7 93.7 93.7 86.5 87.4 87.8 87.9 88.0 88.0 88.1 82.7 83.9 84.3 84.5 84.6 84.6 84.7 75.5 Ti. 778 78.1 78.2 78.3 78.3 _—_*»GEC Pi tee Pamence Typical Case from Example |: 1 year Site Data, Varying Amount of Data from Reference, One-Year Energy, 100 GWh/year P50 1yr 5 yr 10 yr 20 yr 30 yr 40 yr 50 yr 100.0 100.0 100.0 100.0 100.0 100.0 100.0 91.5 91.9 92.0 92.1 92.1 92.1 92.1 83.8 84.6 84.9 85.0 85.0 85.1 85.1 Pe 80.2 80.6 80.7 80.8 80.8 80.9 70.6 72.0 72.5 72.8 72.9 12:9 72.9 15 Conclusions * There is no magical source of long-term that will give you a correct answer without significant uncertainties Really long-term reference stations sound nice in concept, but offer no quantifiable reduction in uncertainty or improvement in P95/P99 cases Multiple years of on-site data are extremely important — Help reduce or quantify uncertainties by testing consistency of references — It doesn’t take much before you reach a point of diminishing returns on using any reference —~GEC 16 __American_Wind 4nergy Association MCP Methods and Long Term Adjustments Wednesday, September 19 8:45 am — 10:15 am Speaker: Bob Conzemius WindLogics, Inc. —<$<$—— WindLogics Mark Ahlstrom, CEO mark@windlogics.com The Long-Term Wind Resource: Comparing Data Sources and Techniques for Predicting the Performance of Wind Plants Dennis A. Moon Scott E. Haynes Robert J. Conzemius Overview Sources of long term wind reference data Climatic downscaling techniques Consistency of RNL and ERA40 datasets Correlation of long term datasets with tall tower data Summary of EMCP versus linear MCP at customer sites www.WindLogics.com 1 WindLogics Mark Ahlstrom, CEO mark@windlogics.com Wind Logics Metar Observations =~ Metar stations are numerous and global Typical sampling period is one hour + 10m height is not necessarily representative of hub height Tl WindLogies Integrated Global Radiosonde Archive (IGRA) + Upper air balloon soundings + Decreased station density (although still global) * Typical sampling period is 12 hours www.WindLogics.com 2 WindLogics Mark Ahlstrom, CEO mark@windlogics.com Reanalysis Datasets Combines observations from many sources Dv Par VP + of With physically based model solutions NCAR/NCEP Global Reanalysis (RNL) ECMWF ERA40 Global Reanalysis Yielding gridded data specifically designed for climate study *NCAR/NCEP North American Regional Reanalysis (NARR) www.WindLogics.com 3 WindLogics Mark Ahlstrom, CEO mark@windlogics.com Wind gies Downscaling ¢ Linear MCP y=mx+b X = Reanalysis * EMCP Several different training variables from nearby Reanalysis grid points Long term wind speed TTS TT ENCP non-linear ny regression system Tower Anemometer Logics R? distribution across North America Average R? of 0.65 — most cells exceed 0.7 Weaker correlation in Rocky Mountain west www.WindLogics.com 4 WindLogics Mark Ahlstrom, CEO mark@windlogics.com orrelation of long term datasets with tall tower data * 30 sites with between 3 and 17 years of data - Anemometer height a minimum of 30m AGL Excludes ERA40 reanalysis data Correlation of long term datasets with tall tower data * Method Tower : EMCP Tower Statistical WW vt | relationship a sc pa a based on one year time period observations www.WindLogics.com 5 WindLogics Mark Ahlstrom, CEO mark@windlogics.com _ Correlation of - tower, reanalysis ~_ «> _and observations |. +*Monthly speed errors 7 Los using both EMCP and linear MCP downscaling Correlation of tower, reanalysis and observations NARR and RNL have largest number of R? in upper bins IGRA data has the lowest mean R? 4 8 R’ Distribution against 30 multi-year Met Towers. . for RNL, NARR. Metar, Radiosonde www.WindLogics.com 6 WindLogics Mark Ahlstrom, CEO mark@windlogics.com Summary of EMCP versus linear MCP at customer sites 23 sites throughout North America Energy distribution determined using wind speed and power curve Customer reports include only RNL reanalysis dataset Summary of EMCP versus linear MCP at customer sites + Estimated monthly energy production Montnty MAE, www.WindLogics.com 7 WindLogics Mark Ahlstrom, CEO mark@windlogics.com Summary of EMCP versus linear MCP at customer sites + Example 1 — Wind speed distribution Wind speed histogram for Correlation Period faction of tat occu ences | ssc. rer Wied Spend (mis) Summary of EMCP versus linear MCP at customer sites _ * Example 2 : — Energy distribution Energy Production Histogram for (24-month) Correlation Period Fraction of Total Occurrences Energy Production (MWh) www.WindLogics.com 8 WindLogics Mark Ahlstrom, CEO mark@windlogics.com www.WindLogics.com 9 V-BAR AWEA Wind Resource and Project Energy Assessment Workshop Portland, Oregon September 18-19, 2007 Workshop Paper September 2007 Session: Identifying and Reducing Data Analysis Bias and Uncertainty Title: Shear Extrapolation Authors: David Matson and Allen Becker, Directors, V-Bar, LLC To be presented at the American Wind Energy Association Wind Resource and Project Energy Assessment Workshop 2007 by David Matson Portland, Oregon September 18-19, 2007 Shear Extrapolations David Matson and Allen Becker V-Bar, LLC Presenters at this workshop have or will be talking about many of the uncertainties that affect wind resource and energy assessments. Wind shear is just one of them. I will show some examples of wind shears at various sites being considered for wind development and our methodology for evaluating and using shear to extrapolate hub height wind speed. Wind shear is an important tool as well as a large part of the uncertainty in wind resource assessment. Wind shear is the spatial variation of wind speed, with the wind industry interested in the vertical variation and its use to extrapolate wind speeds to turbine hub height when measurements are taken at lower elevations. The formula used is the power law equation: (u;/Uo) = (21/20)? Where u is horizontal wind speed, z is height, and p is the shear exponent. Solving for p is shown here: p = LN(u)/up) / LN(z;/Z0) When p is determined, extrapolation to hub height uses the following: u=u, *(z/z,)? Note that sites used as examples in this presentation are all in the United States and Mexico, though the exact locations are not provided, as per non-disclosure agreements. The following figure gives an example of changing wind speed with height. The site is in the southern plains of the US. In this case the shear is positive, though different for each layer. Wind Speed Profile from southern Plains, US Height Above Ground, m 5.0 6.0 7.0 8.0 9.0 Wind Speed, mps Figure 1. Variation in wind speed with height. The next figure shows the extrapolation to an 80 meter hub height based on each of the shears calculated from the data in Figure 1. See the table on page 8 for the values. Percent error in the extrapolations ranges from -23% (10-27 m shear) to 1% (40-65 m shear). Example of wind speed change with height —e— Measured —m— 10-27 shear - -A- - 27-40 shear —<— 40-65 shear —— 10-65 shear @- - 27-65 shear 80 70 + | = | os 60 4 | 5 © 50 +—— oO 2 40 8 < 30 ® 20 | r 10 + 0 . - , 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 Measured Wind Speed, mps Figure 2. Figure | wind profile with 80-m extrapolations. The above figure shows how different hub height wind speeds would be predicted based on using the wind shear exponent from different layers. As might be expected using the shears from the higher layers, the results are closer to the measured value than using shears from lower layers, though the smallest error comes from using the 27-65 meter shear (0.2%). Another way to look at shear variations is to consider typical shears values, the extrapolated hub height wind speeds, and the effect this has on gross capacity factor. The following table is an example using a specific wind speed frequency distribution and turbine power curve, for illustrative purposes only. Other sites and turbines would give different results hence this table shows relative differences only. Starting with a 50 m annual long-term estimate of 7.0 meters per second (mps): Shear: 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24 Extrapolated 80 m speed: 7.34 741 7.48 ee 7.62 7.69 7.76 7.84 Gross capacity factor: 39.3 40.1 40.8 41.5 42.3 43.0 43.8 44.6 With each 0.02 increase in wind shear, wind speeds increase 1% and gross capacity factor increases 2%. Wind shear variations around the USA. In the examples below, we illustrate three general categories of vertical shear profiles. If:shear is reasonably well-behaved, or consistent, vertically, then we have a fairly easy task in determining the hub-height wind speed extrapolation. These cases are shown in the “consistent” column below. For cases where the layer shears are “less consistent” (middle column), we need to determine which shear value to use, which depends on site characteristics. In the “inconsistent” vertical shear situation (third column), we must use careful judgment. We have to decide whether to use any of the shear values, or assign a shear that is more appropriate to the true nature of the meteorological conditions at play at the given site. This speaks to the critical need for on-site familiarity. Consistent: Less consistent: Inconsistent: Northeast US, low hill, smooth terrain: Shear Shear Shear Levels(m) Value Levels (m) Value Levels (m) Value 10-30 0.105 10-30 0.153 10-30 0.294 30-50 0.166 30-50 0.209 30-50 0.418 10-50__ 0.124 10-50__0.170 10-50 _ 0.332 Northeast US, high hill, steep terrain: Shear Shear Shear Levels (m) Value Levels(m) Value Levels (m) Value 10-30 0.399 10-30 0.339 10-30 0.168 30-50 0.432 30-50 0.286 30-50 0.240 30-58 0.431 10-50 0.322 10-50 0.191 10-50 0.409 10-58 0.411 Upper Midwest US, smooth flat open terrain: High shear in 10-25 m layer: Shear Shear Shear Levels (m) Value Levels (m) Value Levels (m) _ Value 10-30 0.237 10-30 0.196 10-25 0.623 30-50 0.230 30-50 0.215 25-40 0.167 10-50 0.235 10-50 __ 0.202 40-55 0.249 25-55__0.209 Western Plains US, broad hill; open grassy terrain: Shear Shear Shear Levels (m) Value Levels(m) Value Levels(m) Value 30-40 0.204 10-30 0.149 10-30 0.132 40-50 0.194 30-50 0.100 30-37 0.258 30-50 0.200 10-50 _ 0.133 37-49 0.096 10-50 0.157 More examples... Consistent: Less consistent: Southern Plains, US, gently undulating brushy terrain: Shear Levels (m) Value 10-30 0.383 30-50 0.302 10-50 0.357 Southwest, rough terrain near escarpment: Shear Shear Levels (m) Value Levels(m) Value 10-30 0.102 10-30 0.170 30-50 0.097 30-50 0.116 10-50 __ 0.100 10-50 0.153 California, open smooth hilly terrain: Shear Levels (m) Value 10-30 0.115 30-50 0.095 10-50 _ 0.109 California, rugged mountainous terrain: Shear Levels (m) Value 10-30 0.082 30-50 0.067 10-50 _ 0.078 For sites with consistent shears we would use the “full tower” shear (i.e., shear determined over the greatest possible tower extent). In the case of consistent shears, we are looking at the thickest layer of the atmosphere, and have more confidence in the calculations. At sites which are less consistent, we usually use the full tower value as well, smoothing out midlevel values, which could be caused by lack of sensor level documentation or sensor malfunction. This is one reason why tower installations need precise documentation and site visits to verify sensor operation. At sites with inconsistent shears we would not use full-tower shear if it seemed too high. We evaluate variations in shear, accounting for geographic and vegetative considerations. We base our determination on our knowledge of doing this work for over 20 years. Sometimes we use a value typical of an area, if the calculated shear seems too high or too low. Shears are not always as well behaved as presented in the examples above. During the 1970s and early 1980s, PG&E measured winds in the Altamont Pass and found negative shears above the 30 meter level. More recently another study done in the same area found negative shears above the 80 meter level. These examples show the importance of measuring winds from as high as possible to keep uncertainties to a minimum and point to one reason why we are seeing more 60m and 80m towers. Another conclusion from looking at these examples is the need to measure at several heights, starting at 10 meters above ground level. Special cases: Terrain with substantial tree cover. Northeast US, low hill, smooth terrain, surrounded by trees: Shear Tree- Raw = Adjusted Levels (m) Value Value 30-40 0.558 0.314 40-50 0.759 0.503 30-50 0.646 0.389 Mexico, smooth terrain, surrounded by trees: Long-Term Tree Tree | Extrapolation to Hub Heights (m) Wind speed | Measured | Adjusted | Height Level (m) | (mps) Shear Shear (m) 60 70 80 10 | 6.47 0.226 0.204 5 10.97 11.34 11.67 30} 9.39 50 | 10.54 The above tables show how we use a tree-adjustment formula for calculating shear. The problem of tree influence can be addressed by taking their height into account, and adjust the shear exponent accordingly. It has been established that trees reduce the wind speed, by increasing the “roughness” of the terrain, which has the same effect as reducing the distance of the wind sensor above the ground. Thus, to account for the trees in the shear analysis, the effective height of the anemometers is reduced in accordance with the average tree height in the immediate vicinity, by a factor of three-quarters the tree height. This we term ‘effective tree height’ or ETH. For example, if the average tree height around a tower is 12 meters, the ETH is 9 meters. The formulas below show the method. For the exponent: Pppry = LN (uj/U) / LN ((z; — ETH) / (Zo - ETH)) For the extrapolated wind speed: u = uy * [ (z - ETH) / (z; - ETH) ] etn The tables show the decrease in shear using the tree-adjusted value. In the first example above, the shears seem high after the tree adjustment, with a suspicious value for the 40-50 meter layer. In this case, using the 30-40 meter shear value would be most conservative. The terrain is fairly smooth; hence high shears are a possibility. In the second example, the 30-50 meter shear was used on this project. Towers in tree influenced regions need to have their lowest sensors placed 10 meters above the effective tree height, in order to look at the deepest atmospheric layer possible. Using the sensor cables that come with towers may indicate installation at only those levels. We instruct tower crews to install sensors at heights specific to the surrounding vegetation. Hub height winds available to evaluate lower level shear calculations. There are a few towers with wind sensors at the 80 meter level, as well as lower levels. This gives us a chance to verify extrapolations from those lower levels. Here are two examples. Northeast US, long forested ridgeline: Long- Term Wind | SHEARS... Speed Extrapolated Percent Level (m) (mps) | Levels (m) Value to 80-m Error 40 6.09 40-50 0.319 7.08 3.2% 50 6.54 50-65 0.210 6.83 6.6% 65 6.91 40-65 0.260 729 0.2% 80 7.31 Southern Plains US, smooth flat open terrain: Long- Term Wind | SHEARS... Speed Extrapolated Percent Level (m) (mps) | Levels (m) Value to 80-m Error 10 5.9 10-27 0.149 7.0 -22.7% 21 6.9 10-40 0.149 7.6 -11.9% 40 he 27-40 0.150 8.1 -5.9% 65 8.2 40-65 0.244 8.6 1.0% 80 8.5 10-65 0.174 8.5 -0.4% 27-65 0.204 8.6 0.2% In the first example above, the shears between levels are not as consistent as the second example, though using shears over a thick atmospheric layer give the most accurate results. With only a 50 meter tower you would be off by about 3% extrapolating to 80 meters. In both examples using the shear from the lowest to the 65 meter level gives good accuracy. It should be noted in the second example, that slightly better results were obtained using the 27-65 meter shear (versus 10-65 meter, or “full tower’’). The second example was the source of the first diagram (Fig. 1). Seasonal variation. Here is an example. Upper Midwest US, smooth flat, open terrain Jan Feb Mar __Apr* May Jun Jul__Aug** Sep Oct Nov Dec Annual 0.187 0.151 0.163 0.162 0.177 *windiest month **lowest wind speed month 0.211 0.263 0.277 0.275 0.250 0.234 0.164 ~—- 0.209 Depending on when the most energy is produced, the annual long-term wind speed could be skewed based on the differing shear of the more energetic months. Another example shows more detail including wind speeds at various levels, indicating relative energy by month. Southern Plains US, narrow brushy plateau: Composite Monthly Mean Wind Speeds (mph) Level (m) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov _ Dec __Year 25 6.6 6.8 6.9 73 7.0 6.2 5.8 5.4 5.6 6.1 6.7 6.9 6.43 40 23 7.5 75 8.2 TA 6.7 6.4 6.0 6.3 6.7 74 Tee ee, 60 8.1 8.1 8.1 8.8 8.2 dee 6.9 6.5 6.9 74 8.2 84 7.75 80 8.4 8.5 8.9 9.4 8.9 7.9 7:5 7.1 74 18 8.6 9.0 8.28 Shears: Levels (m) Jan Feb Mar Apr __ May Jun Jul Aug Sep Oct Nov___Dec__‘Year 25-40 0.210 0.210 0.199 0.225 0.206 0.181 0.216 0.221 0.226 0.226 0.226 0.231 0.215 40-60 0.231 0.200 0.195 0.181 0.170 0.207 0.189 0.221 0.236 0.224 0.240 0.230 0.210 25-60 0.220 0.205 0.197 0.205 0.190 0.193 0.203 0.221 (0.231 0.225 0.232, 0.231 —0.212 Wind speeds extrapolated to 80-m, using shears as above... Levels (m) Jan Feb Mar Apr May Jun Jul Aug Sep Oct _ Nov Dec Year 25-40 8.5 8.6 8.6 9.5 8.9 7.6 1 7.0 TS 19 8.7 9.0 8.26 40-60 8.6 8.6 8.6 9.2 8.7 7.8 7.3 7.0 14 19 8.8 9.0 8.23 25-60 8.6 8.6 8.6 Le) 8.7 17 7.3 7.0 14 7.9 8.8 9.0 8.24 Percent error using shear calculation, compared to measured 80 meter value... Levels (m) Jan Feb Mar Apr__ May Jun Jul Aug Sep Oct Nov __Dec__Year 25-40 0.83 1.01 -3.02 1.57 -0.10 -2.86 -0.71 -1.86 -1.09 1.07 0.83 0.15 -0.28 40-60 2.25 0.34 -3.24 -1.53 -2.61 -1.03 -2.65 -188 -0.34 0.90 1.77 0.07 -0.64 25-60 1.93 0.49 -3.19 -0.83 -2.05 -1.44 -2.21 -1.88 -0.51 0.94 1.56 0.09 -0.56 Wind shears are fairly uniform within each month at this site though there is a large variation between months. The values range from 0.170 (May, 40-60 m) to 0.240 (Nov, 40-60 m). The errors in predicted wind speed range from -3.24% (March, 40-60 m) to 2.25% (January, 40-60 m). With this variation, a weighting scheme for calculating hub height wind speed could be used but it may introduce errors with windier months giving the most skewed results. Annual shear errors are quite small, less than 1%, all under-predicting the measured 80-m speed. In this case the lowest layer (25-40 m) has the smallest error, another example showing the need to measure the maximum depth of the atmosphere. Here the annual shear value gives best results. Shear by wind speed categories. This is another technique to evaluate shear calculations. The following table and figure shows calculations from the same site used above in the seasonal evaluation. Shear Results 40-80 meters Wind Shear as a function of wind speed Wind Mean Maximum | Minimum Speed | Hours Shear Shear Shear Mean Maximum = = =Minimum (m/s) 1 0 0 0 0 03 2 554 -0.204 -0.139 -0.287 3 1382 -0.077 -0.025 -0.138 4 1993 -0.047 0.009 -0.140 02 5 2402 -0.007 0.063 -0.110 6 2272 0.019 0.097 -0.097 7 2060 0.042 0.163 -0.104 O14 8 1868 0.059 0.178 -0.100 9 1552 0.069 0.180 -0.086 S 10 1413 0.083 0.187 -0.061 a 0.0 11 1193 0.100 0.196 -0.002 zz 12 1130 0.102 0.177 0.026 = 13 919 0.102 0.154 0.039 O14 14 666 0.100 0.166 0.010 . 15 540 0.103 0.137 0.040 16 390 0.101 0.160 0.051 02 17 191 0.102 0.158 0.044 , : 18 129 0.099 0.120 0.054 7 19 52 0.100 0.173 0.068 03 ’ 20 18 0.103 0.171 0.077 . 21 10 0.084 0.089 0.075 Wind Speed, mps 22 0 Total 20734 0.078 0.161 -0.019 _| Figure 3. Variation in wind shear by speed. This table and figure show: e negative shear at lower wind speeds e above turbine cut-in speed shears increase as wind speed increases e shears are fairly constant above 11 meters/second, until decreasing sample size becomes a factor at 21 meters/second e more samples show larger variation, as shown by maximum and minimum values 10 Shear by wind speed and direction categories. The following table shows calculations from the same site used above for seasonal evaluation. Southern Plains US, 40-m and 80-m wind speeds, 1999 to 2007 Wind Shear by wind speed and direction Wind Direction Sector Speed Overall (mps) 0 30 60 90 120 150 180 210 240 270 300 330___ Average 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 -0.195 -0.196 = -0.160 = -0.176 = -0.234 = --0.157--0.139 0.281 0.215 -0.167,— -0.248 ~—-0.287 -0.204 2 -0.079 = -0.025. -0.064 -0.072._ -0.071 -0.057 -0.066 = -0.138---0.134 = -0.068 = -0.134 = -0.120 -0.077 3 -0.066 0.005 0.005 -0.015 — -0.045 0.009 -0.022 -0.124 -0.115 -0.082 -0.140 -0.111 -0.047 4 -0.024 0.063 0.049 -0.002- -0.023 0.058 =: 0.005. -0.083 -0.061 — -0.067. --0.110 = -0.076 -0.007 5 0.027___-0.097_——0.076 0.002 0.015 0.092 0.032 -0.097__—-0.055 0.021 -0.063_—_—-0.023 0.019 6 0.074 0.163 0.095 0.071 0.031 0.095 0.056 -0.104 -0.014 -0.006 -0.018 0.012 0.042 7 0.102 0.178 = 0.126 = 0.107 -:0.053. 0.102 0.083 -0.100 --0.037. 0.030 — (0.019 0.044 0.059 8 0.100 0.180 0.127, 0.118 0.068 = 0.130 0.093 -0.086 ~—--0.033 0.062 0.044 0.062 0.069 9 0.113 0.187 0.132 0.083 0.084 0.142 0.101 -0.061 0.011 0.074 0.057 0.075 0.083 10 0.099 0.196 0.175.038 0.092 0.146 0.112 -0.002___—02.041 0.113 0.073 0.085 0.100 1 0.110 0.177, 0.135 0.028 0.085. 0.148 0.113 0.026 = 0.062 02.108 0.068 0.089 0.102 12 0.094 0.154 0.085 0.040 0.039 0.149 0.109 0.058 0.074 0.092 0.078 0.086 0.102 13 0.101 0.113 0.166 0.010 0.042 0.127 0.104 0.088 0.097 0.085 0.075 0.091 0.100 14 0.094 0.077 0.114 0.040 0.137 0.107 0.100 0.097 0.092 0.078 0.091 0.103 15 0.084 0.160 0.051 0.157 0.104. 0.110 0.085 0.094 0.064 0.071 0.101 16 0.075 0.158 0.044 0.142 0.106 0.099 0.119 0.095 0.070 0.079 0.102 17 0.087 0.120 0.120 0.101 0.113 0.054 0.107 0.062 0.088 0.099 18 0.088 0.173 0.094 0.110 0.087 0.112 0.074 0.068 0.100 19 0.124 0.171 0.111 0.077 0.085 0.082 0.103 20 0.075 0.089 0.084 21 22 Overall 0.093 0.161 0.114 0.069 0.054 0.126 0.094 -0.019 0.030 0.072 0.043 0.065 0.078 A lot of numbers to digest, but a few things to note: e negative shears at low wind speed ¢ some discontinuities in some direction sectors as the shear increases with increasing wind speed — note sample size table below increasing shears from low to mid-range speeds in most sectors decreasing shears in most sectors at higher wind speeds negative shear overall for 210° column — near mouth of steep canyon note the table below with number of samples for each bin The last row of the above table shows the shear values weighted by the number of samples in the table below. From these we see that winds from the 30° and 150° sectors have higher shear than other directions. At this site, it’s caused by more complex terrain upwind of the tower in those directions. Both 210° and 240° sectors have the most negative shears, probably terrain induced. Most power-producing winds are from the south, as represented by the 180° sector, which shows less extreme changes in shear values than other sectors. 11 Table of sample sizes in the previous table. Southern Plains US, 40-m and 80-m wind speeds, 1999 to 2007 Number of Samples Wind Direction Sector Speed Overall (mps) 0 30 60 90 120 150 180 210 240 270 300 330 Total 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 74 49 73 50 51 48 43 48 31 35 52 54 554 2 133 147 189 157 121 170 170 79 61 76 79 88 1382 3 213 228 207-175 171 245 307 150 65 91 141 144 1993 4 203 243 188 145 255 386 494 189 72 105 122 141 2402 5 172 175 96110 170 445 631 249 84 75 65 102 2272 6 132 127 73 100 103 396 706 257 67 51 48 65 2060 7 130 114 qd 86 92 355 607 236 76 44 51 79 1868 8 148 112 57 53 109 328 374 229 65 39 38 67 1552 9 157 94 39 21 98 303 356 199 64 41 41 48 1413 10 137 55 12 13 71 240 405 130 58 31 41 58 1193 ll 130 45 9 6 52 197 416 147 71 33 24 79 1130 12 100 24 1 2 13 122 433 11s 42 36 31 46 919 13 72 9 1 1 5 56 337 87 37 32 29 46 666 14 43 11 1 0 4 39 283 86 32 25 16 33 540 15 36 3, 0 0 4 23 219 47 16 26 16 15 390. 16 22 > 0 0 2. 12 102 14 7 23 4 13 191 17. 17 1 0 0 0 5 69 16 5 14 2 8 129 18 2 0 0 0 0 3 33 3 2 8 1 3 52 19 1 0 0 0 0 2 9 4 0 2 0 3 18 20 0 0 0 0 0 0 4 0 0 6 0 0 10 21 0 0 0 0 0 0 0 0 0 0 0 0 0 22 0 0 0 0 0 0 0 0 0 0 0 0 0 Total 1922 1442 1023-919 1321 3375 5998 2285 855 793 801 1092 20734 12 Sensor degradation and effect on shear. The following table shows wind speeds from a tower in the Northwest US and the change in wind shear over a three year period, up to when the 30-m anemometer completely failed in December. Level, m Jan Feb___ Mar Apr___ May Jun Jul___ Aug Sep Oct__Nov Dec 10 12.5 12.7 16.9 14.5 13:5 15.2 14.0 13.2 15.2 16.5 13.9 10.7 30 13.3 13.5 17.7 15.1 13.8 15.6 14.1 13.4 15.5 17.1 14.9 11.6 Shear levels 10-30 0.059 0.059 0.041 0.036 0.023 0.020 0.001 0.011 0.018 0.032 0.059 0.074 10 11.3 10.7 16.7 15.9 13.2 13.8 15.1 14.7 14.6 11.4 11.2 10.8 30 11.3 10.9 16.6 15.7 13:2 13.9 15.1 14.7 14.5 11.7 11.6 12.1 Shear levels 10-30 0.002 0.023 -.003 -.012 0.000 0.007 0.000 -.001 -.001 0.023 0.031 0.102 10 15.8 16.7 14.0 15.7 135 13.8 14.0 14.1 PA 14.3 16.6 9.0 30 16.1 17.4 14.2 16.0 13.8 14.3 14.8 14.7 12.4 15.0 17.4 11.6 Shear levels 10-30 0.017 0.035 0.015 0.013 0.018 0.036 0.048 0.037 0.026 0.044 0.043 _ 0.223 This site has fairly low shears for the first year. The 30 meter sensor had some problems in the second year. During the third year shear values look reasonable until December, when the 30 meter sensor failed completely. If wind speeds had been extrapolated to a hub height before the final month, it may have been difficult to tell there was a problem. Using shears from the second year would result in lower than realistic hub height wind speed. For December 2006 the high shear would result in extrapolated hub height wind speeds higher than realistic. Here is an example of the need to monitor sensors, both when analyzing data and during frequent site visits when their physical condition can be checked. If one of three cups is missing it’s sometimes difficult to tell in the data, but if two cups are missing it is obvious. When all the cups are gone, | think you know what the data looks like. 13 Conclusions. Uncertainties are a part of wind resource assessment and will not disappear. Wind shear represents an important part of the evaluation process, with hub height wind speeds extrapolated from what we know of winds at lower levels. Uncertainty in shear calculations is one of the largest contributions of all the uncertainties we deal with regularly. Other uncertainties include, but are not limited to, sensor accuracy, long-term reference site period of record and quality, reference versus on-site speeds, and turbine micrositing. The ideal solution would be to always measure winds at hub height, but this is more expensive and not always practical. In the future, as turbine size increases, turbine hub heights will increase. Measuring winds at hub height will become more important, and continued efforts are required to reduce this large uncertainty. Until hub height measurements are made, sensors should be placed at well-spaced intervals, starting 10 meters above ground level, or 10 meters above effective tree height. For more information on a few other recent studies of wind shear: Evaluation of Wind Shear Patterns at Midwest Wind Energy Facilities, May 2002 GEC: K. Smith, G. Randall, and D. Malcolm, NREL: N. Kelley and B. Smith Wind Shear, Taller Turbines, and the Effects on Wind Farm Development create a need for Taller MET Towers, March 2004 Wasatch Wind: T. Livingston and T. Anderson, Windtower Corp: T. Anderson Towards a Wind Energy Climatology at Advanced Turbine Hub-Heights, May 2005 NREL: M. Schwartz and D. Elliott Wind Shear Characteristics at Central Plains Tall Towers, June 2006 NREL: M. Schwartz and D. Elliott 14 FLow MopeE.inG Bunlepow Mol4 EXAMINING BIAS AND UNCERTAINTY IN THE WASP MODEL JOHN WADE PPM ENERGY WIND CONSULTANT 2575 NE 328° AVENUE PORTLAND OREGON 97212 WADE.J@COMCAST.NET SITE CHARACTERISTICS = About 60 sq miles/155 sq km = 11 meteorological towers mostly 50 meters some 60 meters and one Sodar used. = Closest met tower to another 1 kilometer, most remote 2.9 kilometers. = Longest period of record six years and shortest four months. Only two sites less than a year of data. = Elevation variation of possible turbine locations 367 to 496 meters. = Shear variation 0.14 to 0.22 » Surface Roughness Winter Wheat over the entire area (roughness 0.02 meters based on Sodar verification). Three dimensional view of the area of investigation | Test 1. Using just one Met tower in the WASP model to predict wind speeds at the other ten. Test 1 Results Best }g00d elevation not much terrain up or downwind longitudinally oriented [Characterization oj lof Shear [wel exposed low elevation but great exit for the wind and terrain 2.4[enhancement Longest Record ‘steep gules Up ar ‘wind Topographic accel 65 Lowest Elevation Representative 10,6|9°°% elevation not much terrain up or downwind some enhancement onan 0. 5.7[terrain jinally oniented_good exit low elevation 3.2|good elevation some raphic enhancement good exit ‘0.4]higher terrain behind slopes down to the south ‘0,3[¢xcellent exit some topographic enhancement Test 2. Three Predictors with different RIX Categories. Mean Error Bias 95% Uncertainty 11.39% Test 3. Three met sites, using distance from the predictor as criteria selecting predictec Middle Middle Mean Error 95% Uncertainty locations. Test 4. Using three met towers and terrain similarity to select predicted locations. D__|Center area 2.1% F__|Center Area 1.0% H__|Center Area 2.5% | Center Area -3.5% Low elev good exit | Low elev good exit Mean Error ae Test 5. Using 7 met towers to predict 11 locations. Predicted Predicted Mean Error Test 6. Using Eleven Met towers or “Perfect Prediction.” Note WASP has inherent self prediction error due to the use of a Weibull fit to the actual distribution. This error is generally an over prediction. Self Prediction Error 0.7% 2.0% 0.9% 1.6% 0.8% 1.2% 1.4% 0.9% 0.5% 2.2% Ale|—|z}o]n|m|olo}a|> 1.5% Mean Error 1.07% 1.08% 95% Uncertainty 0.19% Summary of all tests 11 predictors WASP has inherent error due to fitting Weibull to actual distribution 3 predictors grouped by RIX Mean error small but standard deviation of lerrors large 3 predictors grouped by separation from predictor [Sites close together were very different mean error was low, Prediction Using 3 Mets and Site imilarity 3 predictors grouped by site characteristics This appears to be the strongest lapproach Prediction Using Seven Mets 7 predictors grouped by site characteristic Bias increased slightly in this case, but in more complex terrain more towers would reduce both bias and uncertainty Prediction Using 60 meter Speeds erified with Sodar measured [shear 1 predictor 80 m input speed 150 meter series corrected to 80 meters lusing measured shear for each hour [Prediction Using 50 meter Speeds at Sodar site 1 predictor 50 m input speed [Site had Sodar data and the prediction to 180 meters was tuned to get the best self prediction; roughness length 0.02 (wheat [characteristic cover) [Prediction Using 80 meter Speeds 1 predictor 80 m input speed [50 meter series corrected to 60 meters jing measured shear for each hour Prediction Using 50 meter Speeds 0 Sodar verification 1 predictor 50 m input speed Roughness of 0.02 used Impact of Additional Met towers wind | speed prediction bias and uncertainty Characterization Level % Uncertainty Met tower at every Turbine 0.19 ‘Three met towers or one for every 20 sq miles 4.32 Seven Met towers or one for 8.5 miles 2.78 | CONCLUSIONS = The test situation is typical of a location where one would confidently apply the WASP model. a Using one met tower for this 60 sq. mile project area results in bias values that range from -3.5% to 8.3%. = The mean bias was 1.1% and mean absolute bias of 3.8%. The uncertainty was 6.5%. = The best predictors are those that represent the most typical terrain. = Poor predictors were met towers that were enhanced by terrain or had unusual topographic problems (slopes perpendicular to the prevailing wind, or upslope flow). | CONCLUSIONS continued =» RIX was a poor method of determining which met towers characterize other turbine sites. = Distance of a location from the nearest met tower was an even worse means of determining which met tower would predict a given location. a The best means of selecting met towers to represent turbine location was using terrain similarity. =a The more met towers used the lower the uncertainty in the estimate. a Ata site such as this with uniform roughness the WASP predictions of 80 meter speeds from 50 meter speeds are equal to 80 meter predictions based on 50 meter data corrected to 80 meters. ~~ __American_Wind Znergy Association A New and Objective Empirical Model of Wind Flow Over Terrain Wednesday, September 19 10:45 am — 12:15 pm Speaker: Jack Kline RAM Associates A New and Objective Empirical Model of Wind Flow Over Terrain Jack Kline RAM Associates AWEA Wind Resource & Project Assessment Workshop Portland, OR September 19, 2007 Why an Objective Model? * Fluid flow calculations exceedingly complex - CFD challenges ¢ WASP - difficulty in complex terrain ¢ With correct analysis the wind data will reveal terrain / wind flow relationships * Careful data screening / QA required (tower FX, vertical component, failures) Why does the wind speed vary? ¢ Over a wind farm area meteorological forcing not an issue for LT WS variance (with certain exceptions) ¢ Surface roughness effects ¢ Temporal variance (seasonal, T-O-D, stability) ¢ Terrain effects - variance with site exposure Variance of WS with WD Example of WS Ratios by WD 1.200 1.100 1.000 0.900 WS Ratios 0.800 0.700 0.600 4 fl Zi Es s SS & “ges eg SS ss ss Wind Direction Range Basic Concept of Analysis * Obtain digital elevation models ** ¢ Calculate terrain exposures at met sites ¢ Experiment with calculation of exposures (radius of influence, weighting schemes) ¢ Analyze in context of WS ratios. Examination of Terrain Exposure ee SE ST SP First Application - assumption Sector-wise WS Ratios vs U/W Exposure difference at radius = 2000 m Sector-wise WS Ratios “15 -10 5 0 5 10 15 20 Upwind Exposure Difference (ft) First Application - another view Sector-wise WS Ratios vs D/W Exposure Difference at radius = 2000 - 4 1 ee ho WrudS ted at eKeb- lone 0d ge dS Sector-wise WS Ratios 5 0 5 10 Downwind Exposure Difference (ft) ee Application at another site Sector-wise WS Ratios vs D/W Exposure Difference at radius = 3500 m WS Ratios PG 10 20 30 Downwind Exposure Difference (ft) Same site - different view Sector-wise WS Ratios vs U/W Exposure Difference at radius = 3500 m WS Ratios ; 0 10 20 30 Upwind Exposure Difference (ft) Observations on WS & Terrain * Best results usually at radius = 3500 m ¢ Downwind exposure differences dominant * Upwind exposure differences typically have a negative relationship to WS ratios e R42 from 0.88 to 0.95 for D/W exposures ¢ In multiple regression U/W exposure adds <0.01 to R42 Model Overview ¢ Analyze WS data in context of terrain exposures - develop relationship ¢ Calculate exposures at turbine sites ¢ Use observed relationship to calculate WS ratios and WS in each sector (D/W ref) ° Weight sector WS by D/W WD frequency ¢ Normalize Modeled WS vs. Elevation Normalized Modeled WS vs Elevation Normalized WS 0.99 0.98 0.97 0.94 096 098 100 1.02 1.04 1.06 1.08 1.10 Normalized Elevation Modeled WS vs. Mean Exposure 2 Normalized Modeled WS vs Wtd Mean DIW \wvers Exposure WOK cea Uys Normalized WS 5 0 5 10 15 Weighted Mean Downwind Exposure (ft) Mean HH WS vs. Mean Exp. Normalized Mean WS at Mets vs Wtd. Mean D/W Exposure Normalized Mean WS 5 10 15 20 Weighted Mean Downwind Exposure Conclusions ¢ Data shows that terrain influence on WS is dominated by downwind exposure Radius of prime influence is ~3500m_ —) S\ ¢ Model approach may be most effective in complex terrain situations although localized flows can present problems ¢ Can be adapted to a variety of terrain situations UNncertainiydana Bhat | re] WY} er) Wesosezils i L rend) | ; Mpbeage ate ASS3 e ysMamer We “Te — ee 3 Presented by: Justin Sharp 22 oleh eu at “Wind Aeset Management Meteorology "eee ae bh Assessment. Worksop) xe ae Sept: 18-19) 20py, a PPM Energy aren) yt a Bay ee Basics of Wind Resource Mapping SY » Use a mesoscale Numerical Weather Prediction (NWP) model to downscale atmospheric analysis datasets produced by government agencies >» Typical model choices: @ MM5, WRF, RAMS, Proprietary (e.g AWS MASS) » Typical re-analysis dataset options: @ NCEP/NCAR Reanalysis @ 2.5 degree, 1948-present, global @ ECMWF Reanalysis @ 1.125 degree, 1979-1993, global, not publicly available ® North America Regional Reanalysis (NARR) @ 32 km (0.3 degree at lowest latitude), 1979-present, North America ~ Bas An Idiots Guide To Creating A Wind re. PP Resource Map (1) P > Joe Smo’s MMS eroded WRF RAMS NCEP/NCAR ns eo Another Reanalysis Analysis Analysis » Select model and reanalysis set Bay > s Low Resolution Initial and Boundary Conditions per The First Source Of Uncertainty Reanalysis Geopotental Height Field ~ & Bs, An Idiots Guide To Creating A Wind > Pp Resource Map (2) p> First day of month for . . 30 years >» Select days to simulate ®@ Contiguous period: allows better 2005 validation but may be unrepresentative @ Random sample (e.g. simulate every day in a year but select different year for 2001-2005 each day): more representative if enough days are simulated but difficult to validate especially at project sites 20-yr Draw 2005 NCEP/NCAR ister lare INS) “4 og Bs, “tA Time Period Selection = = dant A source of uncertainty - “oat Wind speed anomaly Two different complete years >» Some years are windier than others > Some months are MUCH windier than others >» Shifts in climate regime >» Changes in available data used to create the analysis used to drive Different samples of the same month the mesoscale model a Kistler et al, 1999 Fignrs 4. Iremntnry nf ratty enntraiied and enatyead nhewvatinns Finm hericin finnt, in millions, SATEMP, ADPUPA, ADPSFC, AIRCFT, SFCSHP, SATWHD, SFCBOG. ~~ & Psy — Anidiots Guide To Creating A Wind >, Resource Map (3) Pp 5 km 1km 10 km 200 m > Determine domain size and resolution @ Higher resolution resolves terrain and gradients in meteorological variables better ®@ Theoretically, the higher the resolution the more able the model is to simulate local scale effects 50-levels A 1 km resoin 29-levels YAR 100 x 80 grid 33-levels 2005 NCEP/NCAR Ree ars 60 gp x 75 gp 100 km 50 km x 50 km xX 80 km ~& Pay Example Of Effect Of Model Resolution On Sp A Mountainous Terrain Cross Section Mountain Range appears as a smooth bump at 27 km Details of rolling terrain I . peak anda not represent at 9 km resolution not represented At 9 km » Resolving terrain @ Where is this? » Resolving terrain driven processes @ What is this? >» Land use also affected Troutdale Consequences of resolution ~& Psy Domain and Resolution Configuration > Pp Another source of uncertainty Pp 100 — - —0 » Choice of resolution limited by: on oo. @ Computer power. Scales with gridspacing® f @ A100 km x 100 km domain at 2 km takes EIGHT times 402 as long to simulate as the same domain at 4 km and ar SIXTYFOUR times as long at 1 km than at 4 km. to3 @ Model design (implicit versus explicit processes) 400 @ Fundamental limit on resolution 2 do @ Parameterizations (later) work poorly at certain 3 ok resolutions 5 g 405 >» Vertical levels and interpolation ae @ Vertical levels are typically not fixed in height space . 406 @ Terrain following and changing pressure rook @ Interpolation is always necessary leading to the usual problem of dubious shear assumptions > Domain size and location of boundaries @ Nests need to be big enough to allow air to adjust @ Boundaries need to be sensibly placed 40.7 ~ a Pay An Idiots Guide To Creating A Wind > . Resource Map (4) P Convective sw Radiation - Scheme . . LW Microphysics Radiation Surface Layer — Land Surface PBL Subgrid eddy diffusion > Determine physical parameterizations ® Needed to resolve processes that are smaller than the grid scale @ Numerous choices ® No one choice fits all solution sti 1 km resoln, 29-levels, 100 x 80 grid, Thompson, BMJ, GFDL LW, MM5 SW, Janjic M-O, NOAH LSM, Simple diffusion, YSU PBL PA0|t}s} NCEP/NCAR Reanalysis ~& Bay Parameterizations > p> How much difference do they make? P’ > s Effect can be HUGE >» Choice of parameterization is likely to be single largest source of systematic bias » Effects are both direct and indirect and can be difficult to predict or understand without considerable effort >» Identical simulations apart from microphysics » Temperature difference between Reisner 2 and Dudhia @ R2has super cooled water @ Simple Ice does not > VERY large differences (9°C) after only 12 hours of simulation Pp More on parameterizations » Parameterized processes are one of the biggest causes of error in mesoscale models » The problem is particularly acute for wind energy applications as we need to very accurately determine the low-level wind field. Critical processes that wind speed is a function of are sub- grid scale (in both the horizontal and vertical) and need to be parameterized » The stable boundary layer is one of the most well known challenges to the modeling community >» The lack of accurate soil moisture data is probably also a significant problem in accurate handling of surface physics 8 Psy An Idiots Guide To Creating A Wind tq- Resource Map (5) >» Run the model @ Model numerics and drift @ Blow ups are usually obvious but can skew results if averaged with many good simulations 2005 @ Choice of spin up time ae ad Analysis Wnt 1 km resoln, 29-levels, 100 x 80 grid, Thompson, BMJ, GFDL LW, MM5 SW, Janjic M-O, NOAH LSM, Simple diffusion, YSU PBL 2005 NCEP/NCAR Reanalysis ~ Bay An Idiots Guide To Creating A Wind Lq- Resource Map (6) > Statistical post processing ®@ Done to remove bias but if not done carefully it introduce error and bias especially if . : a Observation quality is poor ne @ Data is over fitted Downscaled ® Training period data is in a different vane A relenae WRE 1 km resoin, 29-levels, 100 x 80 grid, Thompson, BMJ, GFDL LW, MM5 SW, Janjic M-O, NOAH LSM, Statistical Simple diffusion, YSU PBL Black Box 2005 NCEP/NCAR erie I RS) Some Tower 2005 Tower Observation Time series a Psy Other Sources of Uncertainty and > pr Bias - » Land use and roughness data ® Large amount of subjectivity and uncertainty in this data @ Plus it is changing all the time as new crops are planted, trees are harvested and wind farms are built > Microscale effects @ The level of detail desired by the industry is in excess of where one can get with mesoscale models below (see Mike Zulauf's talk) » Validation Ba Psy Ti > Some Notes on Validation -- » There are not enough surface (or near surface) observations » Detailed observations above the surface are almost non-existent » Available observations are not representative of the domain These problems become MUCH more of an issue as resolution increases. The example on the right contains 6600 grid points and just 2 quality surface observations! The observations can’t characterize the large meteorological gradients present. Pp Summary >» By walking through a high level overview of the use of mesoscale models in wind mapping, sources of bias and uncertainty were identified » Major sources of bias and uncertainty that could lead to inaccurate model derived wind data include: The necessity to use low resolution reanalysis data Date sampling methods: sample size, changes in assimilated data type and quantity, climate shifts over time Inaccuracies within the models and their physical parameterizations The impact of domain configuration and resolution Inaccuracies in statistical post processing (data or method problems) Significance to wind energy of microscale detail that cannot be resolved Inaccuracies in land use and other physical characteristics Relatively low vertical resolution requiring interpolation based shear assumptions Huge model data point to observation point ratio Lack of data to verify vertical structure (sodar will help here) ~ a Psy | ddq< Concluding Remarks Pp » Mesoscale models approaches provide an excellent way to quickly get an estimate of the wind resource over large areas at a level of granularity that is impractical using towers >» When used incorrectly, output from mesoscale models can be misleading at a minimum and outright invalid in the worst case >» Even when used correctly, output from mesoscale models is subject to considerable bias and uncertainty. PPM has identified bias in wind maps from numerous sources >» Users of model data MUST have a basic grasp of the causes of uncertainty and bias and demand that providers try to quantify and explain product limitations Contact Information Justin Sharp Manager, Wind Asset Management Meteorology > 1125 NW Couch Street, Suite 700 Portland, OR 97209 > 503-796-7063 (voice) 503-709-9781 (cell) justin.shar, menergy.com i. PPM Energy mad A ScottishPower Company TECHNICAL Loss EsTIMATES sajyewj}sy Sso7 jeojuyde) ™ A exeteve Ai 3 sv seatade aed uh’, American_Wind 4nergy Association Power Curves: The Effect of Environmental Conditions Wednesday, September 19 3:15 pm - 4:45 am Speaker: Saskia Honhoff GE Energy Power curves The effect of environmental conditions Agenda * Power curve measurements * Effect of air density, turbulence, wind shear and blade fouling — Theory: Why does it have an impact? — Data: What does the data show? — Conclusion Power curve measurement: Measurements per IEC 61400-12-1 a Anemometer, one or two side by side Hub height Wind direction, temperature and pressure D: rotor diameter 3/25 GE / AWEA Workshop 19 September 2007 Data comes from a site in USA Flat terrain 4125 GE / AWEA Workshop 19 September 2007 Air density Theory: Why does Air Density impact the power curve? *P,=05xpxV3xS P,: power available in wind [W] p: air density [kg/m] V: wind speed [m/s] S: Rotor area [m2] * At lower air density, the power in the wind is less — Turbine will reach certain power output at a higher wind speed — For pitch regulated machine, rated power remains the same 4 Stall regulated (fixed pitch) Wind Speed 4 Pitch regulated Wind Speed 6/25 GE / AWEA Workshop 19 September 2007 Data: Analysis method * Data is split into 8 groups of equal size « For each wind speed, the 12.5% of the data with lowest air density is in one group, the 2" lowest 12.5% in the second group, etc. * The plot shows each group in a color air density (kg/m?) + A power curve is made for | each group ‘ 10 15 20 wind speed (m/s) 7125 GE / AWEA Workshop ‘ 19 September 2007 Air Density: What do the measurements show? lowest 1/8th air density 2nd 1/8th air density 3rd 1/8th air density 4th 1/8th air density 5th 1/8th air density ~~ 6th 1/8th air density 7th 1/8th air density highest 1/8th air density AEP (% of total period) Wind speed 8/25 B GE / AWEA Workshop 19 September 2007 Theory Air density correction of wind speed ¢ At lower air density, the power in the wind is less — Turbine will reach certain power output at a higher wind speed. * IEC 12-1 recommends correction of wind speeds * Correction assumes constant efficiency B Air density Conclusion Stall regulated (fixed pitch) p Wind Speed Pitch regulated Wind Speed 9/25 GE / AWEA Workshop 19 September 2007 * High air density increases the power curve * At test site, variation is about 5% in energy yield — Variation between sites can be more than for a single site « Accurate Plant Energy Predictions need to account for the relationship between seasonal changes in air density and wind speed + Note: The IEC air density correction has been applied to the measured data shown in the remainder of this presentation 8 10/25 GE / AWEA Workshop 19 September 2007 Turbulence intensity Theory: Why does Turbulence impact the power curve? * Main effect is mathematical — Mathematical averaging effect + Power curves are based on 10-min mean values + Due to curvature of the power curve, the mean power in varying wind speed is not the same as the single power at the mean wind speed — Simplified example: low WS (P~WS3?) + Because controller will not perfectly track wind speed changes there will be some efficiency losses Turbulence Intensity (Tl) What do the measurements show (10 min avg data)? turbulence intenstiy (-) WS, e . lowest 1/8th turbulence | 2nd 1/8th turbulence | 3rd 1/8th turbulence 4th 1/8th turbulence Sth 1/8th turbulence 6th 1/8th turbulence 7th 1/8th turbulence highest 1/8th turbulence 1 0 wind speed (m/s) Turbulence Intensity What do the measurements show (10 min avg data)? 8 | ——— lowest 1/8th turbulence | —— 2nd 1/8th turbulence | —— 3rd 1/8th turbulence 4th 1/8th turbulence | Sth 1/8th turbulence 6th 1/8th turbulence —— 7th 1/8th turbulence highest 1/8th turbulence Wind speed AEP (% of total period) 13/25 GE / AWEA Workshop 19 September 2007 GE / AWEA Workshop 19 September 2007 Turbulence intensity Conclusion « High turbulence improves the power curve for low wind speed and decreases it for high wind speeds. ¢ Low turbulence decrease the power curve for low wind speed and improves it for high wind speeds. * In the presented case high turbulence effectively improves Annual Energy Production — But shape of turbulence vs. wind speed curve is site dependent — The effect will depend on turbine model (through static power curve) GE / AWEA Workshop 19 September 2007 Wind Shear Theory: Why does wind shear impact the power curve? * Wind speed varies over the rotor plane — Varying wind speed could reduce efficiency — Due to P~WS; relationship: higher wind speed in upper halve of rotor outweighs lower wind speeds in lower halve * Standard equations describing wind shear are simplifications — Wind shear can become extreme in complex terrain and at some sites the mean wind shear exponent can be negative * High wind shear often coincides with high turbulence 8 Wind Shear What do the measurements show? Ground 17/25 GE / AWEA Workshop 19 September 2007 wind shear exponent « (-) lowest 1/8th wind shear exponent 2nd 1/8th wind shear exponent 3rd 1/8th wind shear exponent 4th 1/8th wind shear exponent Sth 1/8th wind shear exponent 6th 1/8th wind shear exponent 7th 1/8th wind shear exponent highest 1/8th wind shear exponent i 10 15 wind speed (m/s) 20 25 18/25 GE / AWEA Workshop 19 September 2007 Wind Shear What do the measurements show? ——— lowest 1/8th wind shear | 2nd 1/8th wind shear ~~ 3rd 1/8th wind shear 4th 1/8th wind shear 5th 1/8th wind shear 6th 1/8th wind shear —_——— 7th 1/8th wind shear | highest 1/8th wind shear AEP (% of total period) Wind speed 3 oe ea cna Wind Shear Conclusion * Data shows that high wind shear (positive or negative) reduces Annual Energy production — Due to varying (non-optimal) inflow conditions * Results are not representative for site in complex terrain * Plant Energy Predictions should take into account — Seasonal or diurnal changes in wind shear — Beware of low level Jets 20/25 GE / AWEA Workshop 19 September 2007 10 Blade fouling & degradation Theory: Why does Blade fouling & degradation impact the power curve? Vi * Fouling can be caused by dust or bugs sticking to the blades + Degradation is permanent damage to the blade due to sand or hail + Increased surface roughness reduces lift and increases drag + Extreme impact on stall regulated machines — Stall starts earlier * Rain will wash most fouling away * Blade roughness cannot effectively be measured — Thus, the following analysis takes rain as an indicator for fouling 22/25 GE / AWEA Workshop 19 September 2007 11 Blade fouling What do the measurements show? 1.4 | 1 23/04 30/04 07/05 14/05 21/05 28/05 04/06 ' ‘ Dry period Wet Period 1 ° © —r rain yes/no (average of 10 min) ° a T olL_t 4 09/04 16/04 23/25 GE / AWEA Workshop 19 September 2007 Blade fouling What do the measurements show? —— wet period (clean blades) dry period (dirty blades) AEP (% of total period) Wind speed Blade fouling Conclusion * Blade fouling can significantly reduce power curve * In wet locations, frequent rain will minimize the effect of fouling * For sites with dry periods and insects regular blade cleaning may be cost effective * At sites with sever conditions (e.g. sand storms) the possibility of blade degradation must be taken into account — Blade degradation may be reparable 25/25 2 oe ea cenee Sesame et Thank you Any Questions? 13