A Short-Term Wind Power Prediction Forecasting using Variational Modes Decomposition Based on Long-Short Term Memory

Abstract

The effective prediction of wind power has a great effect on improving the security of the power grid. Therefore this paper presents a new wind power prediction forecasting model which is the combination of variational mode decomposition(VMD) and long-short term memory(LSTM). By using combined model, the accuracy of prediction can be greatly improved. VMD can effectively overcome the instability of wind power data. LSTM can effectively retain more information, thereby greatly reducing the probability of the prediction result falling into a local option state. To a certain extent, the situation of gradient explosion and disappearance is alleviated. Ultimately greatly enhance the prediction accuracy of the results.