Wind power prediction method and device based on deep learning fusion model

Abstract

The invention provides a wind power prediction method and device based on a deep learning fusion model, and the method achieves the prediction of the wind power through the real-time monitoring data of the wind power of a Scada system and the combination of the historical wind power data. Real-time wind power monitoring data of the Scada system and historical wind power data are input into a deep learning fusion model constructed by a convolutional neural network, a BiLSTM network and an Attention attention mechanism to extract text features, and finally the obtained features are merged to obtain fusion features, so that the optimal text features are obtained to efficiently and accurately predict the wind power. Through the method, the scheduling operation plan making accuracy of the power supply system is improved, and the error phenomenon of new energy power generation power prediction can be reduced.

Description

technical field [0001] The present invention relates to the technical fields of artificial intelligence, deep learning, natural language processing, new energy, carbon neutralization, and carbon peaking, and in particular to a wind power prediction method, device, computer equipment, and storage medium based on a deep learning fusion model. Background technique [0002] With the rapid development of deep learning fusion models and wind power system scheduling and operation, greater challenges are posed to the accuracy of wind power forecasting. For example, how to optimize the new method of accurate wind power forecasting from the two dimensions of SCADA system real-time data and historical wind power data. Provide reliable decision-making basis for wind power supply planning and safe operation. Current wind power prediction methods include physical methods, statistical methods, and single network model deep learning methods, but these methods generally have defects that lea...

AI Technical Summary

Problems solved by technology

Current wind power prediction methods include physical methods, statistical methods, and single network model deep learning methods, but these methods generally have defects that lead to insufficient and inaccurate wind power prediction. Overfitting phenomenon, either the effect of dimensionality reduction is not significant, or a neural network model with a single spatiotemporal data feature is constructed, and contextual information, data dimensionality reduction or compression, and text feature extraction combined with multiple dimensions are not considered, etc.

Therefore, it can be seen that if the accurate prediction of wind power power is not solved, it will bring many disadvantages to the power supply system, including not limited to increasing various operating costs, indirectly increasing manpower and other expenditures, and even constituting the safe operation of the power dispatching system.

Images