A Modeling Method of Wind Power Forecasting Error Based on Meteorological Pattern Recognition

Abstract

The present invention relates to power grid wind power forecasting technology, in particular to a wind power forecasting error modeling method based on meteorological pattern recognition, including obtaining corresponding meteorological models by performing k-means clustering analysis on historical meteorological data, and according to each meteorological model Meteorological data characteristics, training support vector machine classifier, and using it to divide the historical wind power forecast error data into sub-datasets under each meteorological model, and statistically analyzing these sub-datasets to obtain corresponding probability density curves, and then based on the general The distribution model is obtained by least squares fitting to obtain the probability density model of wind power forecast error under each meteorological model, and the modeling of wind power forecast error is completed. This method takes into account the impact of meteorological factors on the accuracy of wind power forecasting, making the modeling results of wind power forecasting errors more accurate; the fitting effect of the general distribution model is better, and the analytical performance of the expression is better; it provides accurate wind power forecasting errors Probability Density Model.

Description

technical field [0001] The invention belongs to the technical field of grid wind power forecasting, and in particular relates to a wind power forecasting error modeling method based on meteorological pattern recognition. Background technique [0002] At present, the problem of wind power accommodation is to judge the future of the system under the premise that the wind power and load forecast results in the next day or several hours are given, and the system uses conventional unit output (including start-up and stop) as adjustment means and meets certain operating constraints. How much wind power can be accommodated. If the wind power prediction error is not considered, the actual value of wind power may be greater or smaller than the planned value of wind power at a certain point in the future, which will lead to the phenomenon of wind curtailment and load shedding, which is not conducive to the consumption of clean energy and will also It will have adverse effects on the ...

AI Technical Summary

Problems solved by technology

This approach essentially ignores the differences in the prediction accuracy of the same forecasting method under different weather conditions. For example, the forecast results may be more accurate in the smooth and sunny weather conditions, but may be more accurate in the windy and windy weather conditions. poor

Therefore, mixing all forecast errors together in statistical modeling will increase the difficulty of coordinating "accuracy" and "analytical calculation" on the one hand, and on the other hand may make the results of stochastic economic dispatch conservative

Images