基于变分模态分解和径向基神经网络的风电场风功率预测

Wind Power Forecasting Method Based on Variational Mode Decomposition and LSSVM Neural Network

风能具有波动性和不确定性,为了降低风电场短期风功率预测误差,提高风电场发电质量,需要对风功率进行行之有效的建模预测。为提高风功率预测精度,本文提出了一种基于变分模态分解和径向基神经网络的风电场风功率预测方法。以历史风速及风功率数据为输入变量,以风电场短期风功率为输出建立预测模型(VMD-RBF),并与传统的BP神经网络及单一RBF神经网络进行对比分析。试验结果表明,所提出VMD-RBF模型具有最优的预测精度,是一种可行有效的风电场短期风功率预测方法。

Wind energy has volatility and uncertainty.In order to reduce the short-term wind power prediction error of wind farms and improve the power generation quality of wind farms,effective modeling and prediction of wind power is needed.In order to improve the wind power prediction accuracy,this paper proposes a wind power wind power prediction method based on variational mode decomposition and radial basis neural network.Based on historical wind speed and wind power data as input variables and wind farm short-term wind power as output,a predictive model(VMD-RBF)was established and compared with traditional BP neural network and single RBF neural network.The experimental results show that the proposed VMD-RBF model has the best prediction accuracy and is a feasible and effective short-term wind power prediction method for wind farms.