基于数据物理混合驱动的超短期风电功率预测模型

To improve the accuracy of ultra-short-term wind power prediction (WPP), a data-physical hybrid-driven ultra-short-term WPP method is proposed. First, the ultra-short-term WPP model with bidirectional recurrent residual network is constructed, and the prediction results in the test set are used as the prediction template. Then, a polynomial-linear regression model is utilized to fit the wind speed-power curve (WPC) of the wind farm, and the WPC is used to predict at the high fluctuation points. Finally, a dynamic switching mechanism between different models is established according to the wind speed fluctuation threshold, and the template prediction value is modified according to the switching time point, and the prediction value is set to zero for the samples when the corrected wind speed is less than the cut-in wind speed. Experimental validation is carried out with data provided by a wind farm with an installed capacity of 400.5 MW in Jilin, the average normalized RMSE predicted in step 16 of the test set is 0.1589, and the favorable switchover accounted for 90.86% of all the switches, which verified the validity and applicability of the proposed ultra-short-term WPP model.