New energy power system wind-solar power short-term joint prediction method considering correlation

The invention relates to a new energy power system wind-solar power short-term joint prediction method considering correlation, and belongs to the field of electric power and load prediction. The method comprises the steps that power data and influence factors of a new energy power system are acquired, and prediction characteristic quantities are extracted through correlation analysis; decomposing the data by using CEEMDAN, and reconstructing the decomposed data again by using approximate entropy; performing hyper-parameter optimization on the BILSTM model by using a CWOA algorithm, and constructing a prediction model; and inputting the preliminary prediction features to obtain prediction errors, and performing cross prediction again to obtain a final prediction result. According to the invention, combined prediction is carried out on the wind/light power generation power, and the method has good prediction performance and good robustness and is easy to realize. 本发明涉及一种考虑相关性的新能源电力系统风光功率短期联合预测方法，属于电力功率、负荷预测领域。