Probabilistic Forecast of Wind Power in High Wind Scenarios of Cold Wave Weather

As the proportion of installed renewable energy in the power system increases, the weather attributes of the new power system become more and more. And the current meteorological changes are becoming more and more complex, with frequent occurrence of cold wave weather, triggering large fluctuations in wind power output. The traditional wind power forecast method is difficult to adapt to the cold wave weather scenario. The large deviation of the forecast brings a great challenge to the wind power consumption. In this paper, we analyze the characteristics of wind power forecast error fluctuation and the mechanism of error generation based on wind power forecast and measured data during cold wave weather. On this basis, we extract the wind power data and numerical weather prediction data under the cold wave weather, and select the strong correlation meteorological factors based on the Pearson correlation coefficient to determine the inputs of the forecast model. Use the generative adversarial network algorithm to expand the samples, so the wind power dataset under the cold wave weather is constructed. Finally, the quantile regression wind power probabilistic forecast model based on the support vector machine is constructed, and an example is set up to verify the effectiveness of the proposed method.