PROBABILISTIC FORECASTING METHODS FOR WIND PLANTS CURRENT POWER
The share of wind energy in total electricity consumption is increasing, which causes a significant uncontrolled impact on the energy system. This makes the problems for the traditional power plants planning and distribution grids stability, so as electricity market procedures. One of the ways to solve the problem is to improve the wind power short-term forecasting, usually within two days time horizon. To obtain reliable forecasts, many methods have been developed that can be divided into two categories – physical methods that use many physical concepts to achieve the best prediction accuracy, and statistical methods trying to find links between a large number of variables, including the results of physical predictions and historical data arrays. Different models of artificial intelligence and spatial correlations are gaining popularity. A typical step of the generalized forecasting model includes scaling the wind speed to the height of the wind turbine rotor axis. This data is then transferred to the generated power.
Traditionally, the forecast is formulated in the form of a certain deterministic power value. The advantage of such prediction is less complexity. However, the increase in the renewable energy share leads to the new requirements for predictive quality, and the elementary assessment becomes inadequate. The paradigm shift in favor of probabilistic forecasts is expected. Modern forecasting methods can provide more information, and often in the form of uncertainty forecasts. Statistical methods are used to determine such indicators as “reliability”, “sharpness” and “resolution”. The main sources of uncertainty in meteorology are atmospheric unpredictability, an uncertainty of data interpretation, errors in the forecast composing and the forecast interpretation. Forecast uncertainty for application in the power industry contains next procedures: statistical methods of probabilistic forecasts, statistically-based ensemble scenarios, physically based ensemble forecasts.
The choice of prediction methods and the list of calculated results depends on the needs of end-users. It is important to determine which methods are acceptable or may contain restrictions. This is provided by studying the forecasting methodology and the growing requirements for the reliability of electricity supply.
2. Theoretical and applied bases of economic, ecological and technological functioning of energy objects / [V.О. Artemchuk et al.; for ed. A.O.Zaporozhets, T.R. Bilan - Kyiv, 2017. - 312 p. (Ukr.)
3. MöhrlenC. atal. Uncertainty Forecasting Practices for the Next Generation Power System. Wind Integration Workshop, Berlin (DE), 26-29 June 2017.
4. Kuznietsov M. Methods of wind farms electricity forecasting // Kyiv: Vidnovluvana energetіka. – 2010, №.3. – P.42-48. (Ukr.)
5. Sanchez I. Short-term prediction of wind energy produc-tion. International Journal of Forecasting, 22. – 2006, p.43– 56.
6.Ma Lei at al. A review on the forecasting of wind speed and generated power // Department of Electrical Engineering, Shanghai Jiaotong University, 2008.
7. GregorGiebel. The State-Of-The-Art in Short-Term Pre-diction of Wind Power A Literature Overview // Risø National Laboratory, 2016.
8. Ricardo J. Bessa at al. Towards Improved Understanding of the Applicability of Uncertainty Forecasts in the Electric Power Industry // Energies 2017, 10(9), 1402. Electronic resource: www.mdpi.com/journal/energies.
9. Möhrlen C. Applications and Value of Uncertainty Fore-casts. UVIG Forecasting Workshop, Atlanta (US), 21-22 June 2017.
10. Gill J.at al. Guidelines on communicating forecast uncer-tainty. World Meteorological Organization, Tech. Rep. WMO/TD No. 4122, 2008.
Abstract views: 63 PDF Downloads: 158
This work is licensed under a Creative Commons Attribution 4.0 International License.