A Hybrid Model for Wind-Energy Forecasting

Wind power is one of the leading sources of clean energy in terms of growth (US Department of Energy estimates that by 2030 it can provide up to 20% of electric power in the US (Lund & Mathiesen 2009)). However, one of the biggest challenges for integrating wind into the grid is the inherent time-variability of wind, and the difficulty in forecasting this variability. Various studies have shown that improvements in the wind forecasts can increase the possible penetration rate of wind and lead to very large financial savings in the energy sector (Schreck et al. 2008; Marquis et al. 2011).

Currently, there are short-term (few hours) and mid-term (few days) prediction horizons for wind power output that use different approaches: the former relies on data-driven method while the latter usually relies on numerical weather prediction (NWP) models. Existing approaches for combining short- and mid-term forecasts blend various products in an ensemble sense. However, they miss the potential benefits that can be obtained from their joint analysis. In this talk, we propose novel dynamical methods that interrogate the observed data collected over the past few hours to inform the blending. We first compare observational data to NWP mid-term forecasting outputs to determine the amplitude and temporal (phase) errors. Then, we test blending approaches that are informed by these errors to “correct” the short and mid-term products. The aim is to reduce the amplitude and phase errors of the NWP models, while informing the short term products on the mid-term trends forecasted by the NWP output, which contains information on the large scale atmospheric dynamics and synoptic scales that short-term forecasts lack.

References:

Lund H, Mathiesen BV. 2009. Energy system analysis of 100% renewable energy systems-The case of Denmark in years 2030 and 2050. Energy. 34(5):524–31

Marquis M, Wilczak J, Ahlstrom M, Sharp J, Stern R, et al. 2011. Forecasting the wind to reach significant penetration levels of wind energy. Bull. Am. Meteorol. Soc.92(9):1159–71

Schreck S, Lundquist J, Shaw W. 2008. U.S. Department of Energy Workshop Report: Research Needs for Wind Resource Characterization. NREL Rep. TP-500-43521. (June 2008)