The technology to harvest electricity from wind energy is now sufficiently advanced to make entire cities powered by wind a reality. High-quality, short-term forecasts of wind speed are vital for making this a more reliable energy source. The investigator proposes to study two topics in wind power forecasting, develop relevant methodologies and analyze their corresponding properties and performances both theoretically and through Monte Carlo simulations studies. The new methods will be applied to a first-of-a-kind testbed wind dataset from northeastern US. The first topic concerns shrinkage and selection of space-time variables in various models for wind power forecasting. The investigator proposes to combine techniques such as partial least squares and the adaptive lasso with space-time correlation information that naturally arises among the predictor variables in the wind forecasting models. The second topic concerns realistic forecast evaluations in the context of wind power forecasting. The investigator proposes to develop new loss functions that are economically relevant for wind power and study their properties and their use in building and evaluating forecasting models in the wind energy arena.
The primary impact of this project is that the new methods for wind power forecasting will provide valuable tools to applied practitioners in wind farming. There is an obvious need for statisticians to be involved in such important problems related to renewable and clean energies for the well-being of our society. In particular, the research topics address crucial and major challenges for advancing the use of energy from wind.
To support large-scale integration of wind power into electric energy systems, state-of-the-art wind speed forecasting methods should be able to provide accurate and adequate information to enable efficient, reliable, and cost-effective scheduling of wind power. Here, we incorporated space-time wind forecasts based on statistics procedures into electric power system scheduling. First, we proposed a modified regime-switching, space-time wind speed forecasting model that allows the forecast regimes to vary with the dominant wind direction and with the seasons, hence avoiding a subjective choice of regimes. Then, results from the wind forecasts were incorporated into a power system economic dispatch model, the cost of which was used as a loss measure of the quality of the forecast models. This, in turn, led to cost-effective scheduling of system-wide wind generation. Potential economic benefits arose from the system-wide generation of cost savings and from the ancillary service cost savings. We illustrated the economic benefits using a test system in the northwest region of the United States as well as using realistic wind farm data from West Texas. Compared with persistence and simple autoregressive models, our model suggested that cost savings from integration of wind power could be on the scale of tens of millions of dollars annually in regions with high wind penetration, such as Texas and the Pacific northwest.
