A rare daytime meeting, hosted at OMSI, saw 58 show up to hear Bob Baker and Justin Sharp, Ph.D., of PPM Energy talk about Wind Energy Meteorology. Bob’s 30 years of experience in the field, assessing potential terrain for wind energy, and Justin’s physics and theoretical modeling approach complemented their two presentations.
Bob started off and
explained the basics. Wind energy is no
longer the stuff of pipe-dreams from the environmental days of 1970s. The steady-to-dramatic rise in wind energy
capacity is amazing. Wind energy
development began in
The relationship between wind velocity and the power generated is not linear. For every 10% increase in wind velocity, a wind turbine can create 33% more power.
Wind energy assessment involves the detailed geographical mapping of a prospective terrain, identifying wind indicators, and making on-site weather measurements. Three anemometers are mounted atop a turbine to gauge local airflow. Over 140 observation sites are set-up for a prospective area.
The amount of power generated is minimal for wind speeds up to 4 meters per second (9 mph). Above speeds of 4 m/sec, the power quickly builds for every incremental increase of wind speed. Power production peaks near 1500 kW then plateaus at speeds of 12 meters per second (27 mph) or higher.
facilities exist at
Towers are aligned north-south for westerly wind flow in order to maximize generation. Tower anemometers serve a check on the turbine’s performance.
The weight of all the components is amazing. The 1500 MW generator weighs 57 tons. The tower weighs 95 to 150 tons. The rotor blades weigh 38 tons. The most heavy-duty cranes in existence are needed to move these massive components.
Biological-wildlife surveys are conducted before and after the development of a wind energy facility. Contrary to popular belief, bird kills by the turbines are very small.
Justin then picked up on the forecast science. The forecast horizon varies from hours to years and requires different data sets. The MRF (medium range forecast) requires high resolution grid data to predict wind flow over specific terrain. Having good local predictors of wind changes, as well as off-site observations, are highly desired.
How does one deal with error? Depends what time scale you seek. Climatology may work but so does persistence of the forecast, then the use of an advance forecast system (but even then that only works 10-20 hours out in the future). Multiple-member ensemble runs and Baysian Model Averages can quantify forecast confidence levels.
A site needs at least 2-5 years of quality data. Good 10-year wind records are rare. One can now use high resolution, “re-analysis” data sets, fed through Numerical Weather Prediction models (e.g., MM5 model), to create a synthetic 40-year wind record. That synthetic record can then be broken up into seasonal components.
How does wind vary by season? Winter is the most variable. During El Nino years, wind energy potential is at its lowest. During La Nina years, wind energy potential is at its highest. During ENSO-neutral years, wind energy potential has its greatest variability. During warm phase of the PDO (Pacific Decadal Oscillation), wind energy potential is at its lowest. During cold phase of the PDO (Pacific Decadal Oscillation), wind energy potential is at its highest. Based on the rapid increase in computer power, Justin predicts that in 11 years, we will be able to compute a 40-year wind forecast, using ˝ kilometer grid resolution. Given the soaring nature of recent energy prices, their presentations were very well received and many questions were asked.
Note-taker: Kyle Dittmer, OR-AMS President