Research on Short-term Wind Power Combination Prediction Based on Multiple Factors
In order to improve the prediction accuracy of wind power model,this paper adopts a combined prediction model consisting of Convolutional Neural Networks(CNN),Bidirectional Long Short Term Memory Network(BI-LSTM),and Attention Mechanism(AM).Firstly,consider meteorological factors(wind speed and direction at dif-ferent heights,temperature,humidity,and pressure)was considered to comprehensively reflect the impact of weather conditions on the accuracy of wind power prediction at that time.Then,based on meteorological factors,explore histori-cal wind power time series data and Variational Mode Decomposition(VMD)signals were taken as feature variables for predictive modeling.Finally,only time series data and VMD signals were considered as feature variables for deep learning prediction modeling.It was found that the prediction accuracy of the CNN-BI-LSTM-AM model based on meteorological factors and VMD combined signals as feature inputs reached 97.66%,while the prediction accuracy of the CNN-BI-LSTM-AM model only considering VMD combined signals as input reached 97.71%.Satisfactory results were achieved in the accuracy and stability of wind farm power prediction.
wind power predictionbidirectional long short term memoryvariational mode decompositionconvolu-tional neural networkattention mechanism
万方数据
维普
