The Wind Resource: Prospecting for Good Sites

Bruce Bailey, President AWS Truewind, LLC 255 Fuller Road Albany, NY 12203 [email protected]

Talk Topics

 Causes of Wind  Resource Impacts on Project Viability

 Siting Linkages to Resource Assessment  Wind Resource Definition  Energy Production Prediction

 Uncertainty Analysis  Due Diligence Objectives For Financing

What Causes Wind?

 Uneven heating of the earth’s surface  Daily heating and cooling cycles  Earth’s rotation  Weather systems – track and intensity  Position of jet stream  Local influences – sea breezes, slope winds, channeling through valleys, etc. Establishing Project Viability

Wind Resources Determine:  Project Location & Size  Tower Height  Turbine Selection & Layout  Energy Production » annual, seasonal » on- & off-peak » capacity credit  Cost of Energy/Cash Flow  Warranty Terms  Size of Emissions Credits

The wind energy industry is more demanding of wind speed accuracy than any other industry. Power in the Wind (W/m2)

= 1/2 x air density x swept rotor area x (wind speed)3 ρ A V3

Density = P/(RxT) P - pressure (Pa) Area = π r2 Instantaneous Speed R - specific gas constant (287 J/kgK) (not mean speed) T - air temperature (K) kg/m3 m2 m/s Summary of Wind Resource Planning Steps

 Identify Attractive Candidate Sites  Collect >1 yr Wind Data Using Tall Towers  Adjust Data for Height and for Long-Term Climatic Conditions  Use Model to Extrapolate Measurements to All Proposed Locations  Predict Energy Output From Turbines  Quantify Uncertainties Siting Siting

Main Objective: Find and design viable wind project sites

Main Attributes:  Adequate winds  Access to transmission  Permit approval reasonably attainable  Sufficient land area for target project size » 30 – 50 acres per MW for arrays » 8 – 12 MW per mile for single row on ridgeline Siting Attributes

 Winds » Minimum Class 4 desired (>7 m/s @ hub height) for wind farms  Transmission » distance, voltage, excess capacity  Permit approval » land use compatibility » public acceptance » visual, noise, and bird/bat impacts are leading issues  Land area » economies of scale with larger project size » number of landowners Siting Tools

 Wind Maps & Other Regional Resource Data  Topographic Maps Old vs. New Wind Maps of the Dakotas  Transmission Line Maps & Databases  Property Maps  Geographic Information Systems (GIS) and associated data layers

Local Wind Map Showing Transmission and Road Overlays Modern Wind Maps

• utilize mesoscale numerical weather models • high spatial resolution (100-200 m grid = 3-10 acre squares) • simulate land/sea breezes, low level jets, channeling • give wind speed estimates at multiple heights • extensively validated • std error typically 4-7% • GIS compatible Old and new wind maps of the Dakotas • reduce development risks Source: NREL www.windexplorer.com/NewYork/NewYork.htm

Wind data projections available at a 200 m grid resolution statewide Measurement Sources of Wind Resource Info

 Existing Data (surface & upper air) » usually not where needed » use limited to general impressions » potentially misleading  Modeling/Mapping » integrates wind data with terrain, surface roughness & other features  New Measurements » site specific using towers & other measurement systems How and What To Measure

, Vanes, Data Loggers, Masts  Measured Parameters » wind speed, direction, temperature » 1-3 second sampling; 10-min or hourly recording  Derived Parameters » wind shear, turbulence intensity, air density  Multiple measurement heights » best to measure at hub height » can use shorter masts by using wind shear derived from two other heights to extrapolate speeds to hub height  Multiple tower locations, especially in complex terrain  Specialty measurements of growing importance » Sodar, vertical velocity & turbulence in complex terrain Typical Monitoring Tower

• Heights up to 60 m • Tubular pole supported by guy wires • Installed in 1-2 days without concrete using 3 people • Solar powered; cellular data communications Raising the Tower Final Touches / Sensor Orientation Wind Resource Assessment Handbook Fundamentals for Conducting a Successful Monitoring Program

 Published by NREL WIND RESOURCE ASSESSMENT HANDBOOK » www.nrel.gov/docs/legosti/ Fundamentals For Conducting fy97/22223.pdf A Successful Monitoring Program  Peer reviewed  Technical & comprehensive  Topics include: » Siting tools » Measurement instrumentation Prepared By: » Installation AWS Scientific, Inc. » Operation & maintenance 255 Fuller Road Albany, NY 12203

» Data collection & handling NREL Subcontract No. TAT-5-15283-01April 1997

» Data validation & reporting Prepared for: National » Costs & labor requirements Laboratory 1617 Cole Boulevard Golden, CO 80401 Data Analysis

Wind Hours/ Speed Year (m/s) GE 1.5xle - 1.5 MW 0 0.0 1300 1600 1 434.4 2 823.4 1200 3 1,098.6 1100 1400 4 1,228.7 5 1,216.5 1000 1200 6 1,092.3 900 7 900.8 1000 8 687.5 800 9 487.9 700 10 323.0 800 11 200.0 600 12 115.9 500 600 13 63.0 14 32.1 400 Output(kW) Turbine 400 15 15.4(hours/year) Probability 300 16 6.9 17 2.9 200 200 18 1.2 100 19 0.4 0 20 0.2 0 0.0 5.0 10.0 15.0 20.0 25.0 21 0.1 0 2 4 6 8 10 12 14 16 18 20 22 24 26 22 0.0 Wind Speed (m /s) 23 0.0 Wind Speed (m/s) 24 0.0 25 0.0 26 0.0 Speed Frequency Distribution Wind Turbine Power Curve: Output As a Function of Speed

Wind Direction Rose Wind Shear The change in horizontal wind speed with height

Wind Shear is important when extrapolating wind speed data from a met. mast that is shorter than the intended hub height of the turbine

V = 7.7 m/s  A function of wind speed, 2 surface roughness (may vary

with wind direction), and Z2= 80 m atmospheric stability (changes

from day to night) V1 = 7.0 m/s  Wind shear exponents are higher at low wind speeds, Z1= 50 m above rough surfaces, and Wind Shear during stable conditions Profile  Typical exponent (α) values: α » .10 - .15: water/beach » .15 - .25: gently rolling farmland α » .25 - .40+: forests/mountains α = Log10 [V2/V1] V2 = V1(Z2/Z1) Log10 [Z2/Z1]

Predicting Long-Term Wind Conditions From Short-Term Measurements Measure - Correlate - Predict Technique

 Measure one year of data 25 on-site using a tall tower Airport C Regression Airport B Regression y = 1.7278x + 0.7035 y = 1.4962x + 0.4504 2 2  R = 0.8801 R = 0.875 Correlate with one or more 20 regional climate reference stations 15 2 Airport A Regression » Need high r y = 1.0501x + 0.4507 10 2 » Reference station must have R = 0.8763 long-term stability 5 Airport A Airport B

» Upper-air rawinsonde data may (m/s) Speed 60ProjectSite mWind be better than other sources for Airport C 0 correlation purposes 0 5 10 15 20 Reference Station Mean Wind Speed (m/s)  Predict long-term (7+ yrs) wind characteristics at This plot compares a site’s hourly data project site with three regional airport stations. A multiple regression resulted in an r2 of 0.92. Energy Prediction For Sample Energy Production Calculation Appendix 2: Preliminary P50 Energy Production Estimate for Ayia Anna-Kochi Project Area

GE 1.5xle (Altitude Adjusted) WS (m/s) Output (kWh) 1200 Wind Speed Frequency Distribution 0 0.0 1 0.0 1100 2 0.0 1000 cut in 3 1,323.3 900 4 60,484.4 800 Turbine Power Curve 5 168,173.6 GE 1.5xle - 1.5 MW 6 297,296.5 700 1600 7 431,266.7 600 1400 8 539,540.4 500 1200 9 582,520.0 400 1000 10 534,351.4 11 415,751.9 300 800

12 281,729.8 Probability (hours/year) 200 600 13 173,933.0 Turbine Output (kW) Output Turbine 100 400 14 101,064.2

200 15 55,702.5 0 16 29,138.2 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 0.0 5.0 10.0 15.0 20.0 25.0 17 14,472.9 Wind Speed (m /s) 18 6,828.3 Wind Speed (m/s) 19 3,061.0 cut out 20 1,304.1 21 0.0 22 0.0 INPUTS 39 On-Site Turbines 23 0.0 6.738 80 Meter Weibull C Gross Output per Turbine 3,698 MWh/yr 24 0.0 2.034 Weibull K at 50m Net Energy per Turbine 3,258 MWh/yr 25 0.0 1.50 Turbine Capacity (MW) Number of Turbines 39 Gross Plant Production 144,220 MWh/yr Net Plant Production 127,048 MWh/yr Net 24.8% Energy Production Projection

 Multiply wind speed frequency distribution data (annual hours per 0.5 or 1.0 m/s speed bin) by turbine power curve output values (for same speed bins) » Power curve must be adjusted for site air density  Sum the product of all speed bins for the total gross energy production (MWh)  Determine production loss factors and their magnitude » Wakes, availability, electrical, blade soiling/icing, high wind hysteresis, cold temperatures » Cumulative losses are typically 10-15%  Deduct losses to calculate net energy production  Determine net production for different probability levels (P75, P90, P95, etc.) based on uncertainty analysis Micrositing – Predicting Wind Conditions at Every Turbine, and Optimizing Turbine Locations

Software tools (WindFarmer, WindFarm, WindPro) are available to optimize the location and performance of wind turbines, once the wind resource within a project area is defined. Optimization Tools Turbine Noise Emissions

Wind Resource Mapping & Turbine Energy Production

Photosimulations Conclusions

• The wind resource drives project viability. • Wind conditions are site-specific and time/height variable. • Accuracy is crucial. Wind resource assessment programs must be designed to maximize accuracy. • Combination of measurement and modeling techniques gives the most reliable result. • Know the uncertainties and incorporate into decision making. • Good financing terms depend on it.