Using Numerical Weather Prediction for

19721212

Abstract:
Increases in the amount of electricity generated from wind energy will require sophisticated management tools to accommodate its inherent variability and ensure grid reliability. Accurate wind energy forecasts are a key component of these management tools. The effect of wind energy is lost in the noise associated with load forecast errors at low penetration levels of wind energy – that is, when wind energy is only a minor part of the overall electricity generation mix. However, as penetration levels increase, the area control error (ACE) is increasingly affected by the variations (especially the unpredicted variations) in wind energy. Furthermore, as energy markets continue to develop and find ways of handling unscheduled generation, wind energy forecasts become a valuable asset to traders and financial decision makers. In fact, at high penetration levels wind can significantly move the price of the electricity market. Consequently, an accurate wind energy forecast becomes a valuable tool for all market participants - whether operating renewable energy projects, conventional generation or some other aspect of the power system.
This paper presents the use of numerical weather prediction in creating state-of-the-art wind energy forecasts. It discusses the advantages and disadvantages of a physics-based modeling approach and contrasts it with some of the other techniques that are used to generate wind energy forecasts. Finally, the paper highlights some important areas for future development in wind energy forecasting that will enable better management of wind, an unscheduled resource.
Keywords: Wind Energy Forecasting, Numerical Weather Prediction (NWP)
1. Introduction
/ @! B- K9 d  M1 f% }* |Weather-driven, non-scheduled, renewable energy sources require new operational procedures. Conventional fossil fuel power plants can be operated in accordance with the needs of the power system. Renewable energy sources such as wind or solar are variable and thus the operating schedules of such plants are largely dictated by the changing “fuel” supply. This is especially pertinent in the case of wind, photovoltaic solar and run-of-the-river hydro, none of which have inherent storage in their power plant design. To optimize performance of plants that exploit these energy sources, the “fuel” supply must be accurately predicted.
- l1 q3 w1 G7 f1 \9 RWeather forecasting is a complex process that demands sophisticated and customized solutions to achieve the best results. This paper describes a state-of-the-art solution to wind energy prediction ranging from hour-ahead forecasts for market operations through to week-ahead forecasts for short-term system planning.
The consensus in the wind forecasting community is that the best way to model the weather is through the use of non-hydrostatic mesoscale numerical weather prediction (NWP) models. These models are advanced mathematical models that predict the weather through manipulating huge datasets and performing many, many
0 o7 j1 X' C% K1