Short-Term Wind Power Prediction Based on VMD-CNN-LSTM Model
Accurate prediction of wind power is one of the important means to improve the grid-connected stability of wind power.Aiming at the problem that wind power is difficult to be accurately predicted due to the complexity and randomness of meteorological characteristics,this paper proposes a short-term wind power prediction model based on VMD-CNN-LSTM.The overall structure of the model includes multi-meteorological feature series variational mode decomposition(VMD)and reconstruction,multi-meteorological feature information mining by convolutional neural network(CNN),long and short-term memory network(LSTM)prediction results output,and generalization capability analysis.Compared with the traditional method of establishing prediction models only considering decomposition of historical wind power series,the meteorological feature VMD method proposed in this paper has clear physical significance and can track meteorological features to predict future wind power trends.The model is verified on the actual data of a wind farm,and the result of an example shows that the prediction result of this model has high precision and can reduce the influence of multiple meteorological characteristics on the prediction result,which has certain practicability.
Short-term wind power forecastingEmpirical mode decompositionConvolutional neural networkLong and short term memory networks
万方数据
维普
