Prediction Of Crude Oil Production Base On VMD And Neural Network

The prediction system of crude oil production is influenced by environmental factors,scientific and technological level,national policy and so on,which is related to national economic,military and the formulation of import policies.Pay close attention to China’s crude oil production,explore the reasons for its decline,and make a reasonable prediction to avoid the potential adverse effects.Crude oil production data is non-linear and unstable,and it is difficult to predict accurately.In order to improve the accuracy of prediction,a novel hybrid model named VMD-BPNN is developed to forecast monthly crude oil production.In the empirical analysis,the monthly output of crude oil in China is decomposed into some components and residual series by Variational Mode Decomposition(VMD),then the components is forecasted by BP neural networks respectively.Finally,the forecasting results of the components are summed as the final result of the VMD-BPNN hybrid model.According to the three evaluation criteria,the result shows the proposed method has the best performance among single neural network and EEMD-BPNN hybrid model,and their prediction has high precision and accuracy.But the VMD needs to set value in advance,which is determined the modality number of Ensamble Empirical Mode Decomposition(EEMD).