Optimizing Deployment of Meteorological Towers During Project Development

1
Optimizing Deployment of Meteorological Towers
During Project Development

• European Wind Energy
• Conference Exhibition 2006
• Athens
• Steve Jones
• Global Energy Concepts
• 5729 Lakeview Drive, Suite 100
• Kirkland, WA 98033 USA
• 1 - 425 - 822 - 9008
• sjones_at_globalenergyconcepts.com
• March 2, 2006

2
The Issue

• Central question for investors
• How much energy will the project produce?
• How windy is it?
• Answering the question costs the developer money
• Land lease, tower, instruments, installation,
  maintenance, data acquisition, data evaluation,
  decommissioning
• but also provides developer access to money
• Robust answer provides more funds with better
  terms
• How to balance these competing issues
• Better wind data to reduce uncertainty vs. need
  to conserve development capital
• What guides this balance?

3
Different Uncertainty Distributions

• Example study results
• P50 estimate of 325 GWh/year
• P95 estimate of 280 GWh/year (blue line)
• With less robust measurement campaign, P95
  estimate might be only 265 GWh/year (pink line),
  reflecting less certainty in the answer

P95 estimates 95 of area under curve lies to
the right
P50 estimate
4
Uncertainties (and mitigants)

• Usually Most Significant
• Topographic effects (more towers)
• Wind shear (taller towers)
• Year-to-year variation (long-term correlation)
• Wind turbine reliability (availability)
• Usually Less Significant
• Wind turbine power performance
• Equipment mounting/installation
• Array losses
• Electrical losses
• Many others

5
Topographic Uncertainty
6
How Uncertainty Impacts Value

• Typical analysis
• Best estimate P50 most likely outcome
• 50 chance results will be better, 50 chance
  worse
• 95 confident P95 1 in 20 chance of this
  outcome
• P99 1 in 100 chance
• One-year and long-term estimates
• Investors consider various Pxx levels in
  evaluating down-side
• Uncertainty used in sizing debt
• Both P50 and P90P99 typically considered (with
  different debt service coverage ratios)
• Lower uncertainty can increase debt (and value to
  developer)
• 5 examples presented
• Show range of met data acquisition programs
• Different uncertainties go with each
• Better data acquisition program means less
  uncertainty and thus P90P99 are closer to P50

7
Example Cases

• Single 50-m tower, 1 year data, no long-term
  correlation
• Single 50-m tower, 1 year data, good long-term
  correlation
• Four 50-m towers, 1 year data, good long-term
  correlation
• One 80-m tower, four 50-m towers, 1 year data,
  good long-term correlation
• One 80-m tower, four 50-m towers, 4 years data,
  good long-term correlation

8
Debt Sizing Assumptions

• Using the P50 energy production estimate 1.50
  Debt Service Coverage Ratio
• Using the P99 (one year) energy production
  estimate 1.00 DSCR
• Smaller of the two assumed to size the debt

9
Example Results
10
Balancing Uncertainties

• There will always be some uncertainty
• Overall uncertainty usually the square root of
  the sum of the squares of individual
  uncertainties
• If any one uncertainty remains high, reducing
  uncertainty in the others is not of value in
  reducing overall uncertainty
• Uncertainty in long-term estimates can be reduced
  much more than uncertainty in 1-year values
• 1-year numbers largely result from wind
  variability over short (1-year) time periods

11
Balancing Uncertainty Example
12
Additional Issues

• Terrain Site complexity helps determine
  appropriate met tower campaign
• Turbine Suitability Met data also used to
  confirm turbine suitable to site
• Long-Term References Nearby long-term sources
  (airports, research stations), under proper
  conditions, is often a key element in reducing
  uncertainty
• Competition Others behavior may need to
  influence your own, if for example, competing for
  a power purchase agreement
• Time Met towers need time to gather data.
  Installing them later in development conserves
  capital, but also means less data are available
  for the assessment
• New Results Additional data gathering and
  analysis may increase or decrease the P50 best
  estimate

13
Contact Information

• Steve Jones
• Director of Utility and Investor Services
• Gordon Randall
• Senior Data Analyst
• Global Energy Concepts, LLC
• 5729 Lakeview Drive NE, Suite 100
• Kirkland, WA 98033 USA
• Phone 1 425 - 822 - 9008
• Fax  1 425 - 822 - 9022
• sjones_at_globalenergyconcepts.com
• grandall_at_globalenergyconcepts.com
• www.globalenergyconcepts.com