A Wind Power Prediction Method Considering Wind Power Uncertainty Risk

Abstract

The invention provides a wind power prediction method considering the uncertainty risk of wind power. On the basis of the mixed distribution model, an adaptive mixed uncertainty probability modeling method with the risk value VaR as the training target is proposed. The method first adopts the historical The wind power data generates a training set, and then uses the value at risk VaR as the training target to update the training samples. Finally, through multiple rounds of learning, a high-precision mixed probability model of wind power forecast error is generated, which can effectively improve the economics of power generation.

Description

technical field [0001] The invention relates to the field of power generation, in particular to a wind power prediction method considering the risk of wind power uncertainty. Background technique [0002] After long-term engineering practice and national "blowout" development, wind power generation has been proven to be a reliable clean energy that can be developed and utilized on a large scale. However, with the further increase of wind power generation power penetration rate, the volatility and randomness of wind energy itself bring challenges to the accurate formulation of power generation plans. Under the background that the current forecast error of wind power is generally as high as 10%-20%, on the one hand, the high penetration of wind power sources forces the dispatching organization to purchase a large amount of power generation backup in the day-ahead power generation planning stage to maintain the power balance of the grid operation during the day, and on the othe...

AI Technical Summary

Problems solved by technology

However, in the uncertainty modeling process, the fact that the wind power forecast uncertainty tail distribution (high error interval) has a much greater impact on the economics of day-ahead power generation planning than the head distribution (error low value interval) is not considered. , there is no targeted modeling of the tail distribution of forecast uncertainty, which affects the further improvement of system economy

Images