High resolution wind turbine power output forecasting

Wind turbine generator power output and consumer electricity demand vary independently from one another. This presents a difficult situation for electricity grid managers as they attempt to exactly match demand using wind turbines and conventional generators (e.g. hydro, fossil fuels). Accurate forecasting of wind turbine generator power enhances management of the electricity grid, allowing for more wind turbine generating capacity while maintaining grid stability. This research will calibrate, perform sensitivity analysis, and validate the newest high resolution wind power forecasting model for Atlantic Canada. Resolution of space and time have increased from 10 km and 60 minutes, representing an entire wind farm with one forecast point, to 0.1 km and 5 minutes, representing a single wind turbine. Calibration will be completed by forecasting a variety of wind turbine types and farms and comparing with actual performance data. Sensitivity analysis will compare the incremental gains in forecast accuracy due to increasing resolution. Validation will insure accurate forecasting for a variety of conditions, topographies, and wind farm layout. The final research results will be used enhance and justify the new forecast model as it is productized and sold to wind farm operators, utilities, and grid operators.

Faculty Supervisor:

Dr. Lukas G. Swan

Student:

Nathaniel Pearre

Partner:

Scotia Weather Services Inc.

Discipline:

Engineering - mechanical

Sector:

Alternative energy

University:

Dalhousie University