SHORT-TERM WIND POWER PREDICTION BASED ON VMD-BOA-LSSVM-AdaBoost
Aiming at the intermittent,nonlinear,fluctuating,non-stationary and uncertain characteristics of wind power signals,the short-term forecasting method for wind power is established,which is based on Variational mode decomposition(VMD)and butterfly optimization algorithm(BOA)to optimize least squares support vector machine(LSSVM)and introducing adaptive correction to improve accuracy.Firstly,the raw power signal data is splitted into multiple subsequences by using VMD.Secondly,BOA is used to optimize combined prediction model of LSSVM to predict each subsequence.Finally,the prediction value of multiple components is reconstructed through AdaBoost to obtain the final prediction value.Combined with the wind power data provided by a wind farm in Northwest China as an example,the effectiveness of the model is verified.The results show that the combined forecasting model established above can predict the short-term wind power well and has a good forecasting accuracy.
wind power predictionLSSVMVMDAdaBoostprediction accuracy
NETL
NSTL
万方数据
维普
