A Novel Strategy for Wind Speed Prediction in Wind Farm

The empirical mode decomposition (EMD) is well known for predicting wind speed.However, but the joint application of relevance vector machine (RVM) and empirical mode decomposition in wind speed forecasting is seldom found in the field. This paper proposes a relevance vector machine model based on empirical mode decomposition to predict the wind speed. Before the wind speed forecasting with RVM,EMD algorithm is used to decompose wind speed signal in order to weaken the disadvantageous influences of nonlinearity and uncertainty in wind spped. By the decomposition process, a series of intrinsic mode functions (IMFs) are generated. To each IMF, RVM algorithm is used to construct the model and carry on the forecast espectively. The final predicted result is obtained by the superposition of all prediction results. By the simulation experiment, the comparison of several algorithms is shown. The results showed that EMD-RVM model is effective, and has better forecasting precision