基于相空间重构小波神经网络的短期风速预测

风速具有高度非线性和非平稳性,难以精确预测。对此,利用神经网络逼近非线性函数的能力,结合小波变换多尺度特性,使风速在一定频域尺度上表现出准平稳性,建立了风速的小波神经网络预测模型。为了解决神经网络输入的随意性,以相空间重构理论确定最佳延迟时间和最小嵌入维数,重构风速时间序列,以重构后的时间序列作为模型的输入量对网络进行训练。仿真结果表明,所提基于相空间重构小波神经网络风速预测的准确性能得到了提高。

Short-term wind speed forecasting based on wavelet neural network of phase space reconstruction

Wind speed is a non-liner and non-stationary process.It’s hard to forecast with high precision.Therefore,a wavelet neural network model is set in this paper.The non-liner process of wind speed is forecasted by neural network and the non-stationary process of wind speed is decomposed into quasi-stationary by multi-scale characteristics of wavelet transform.In order to solve the randomness of neural network input,the optimal delay time and minimal embedding dimension are determined by phase space reconstruction,and then the wind speed time series is reconstructed.The model is trained with phase space reconstruction data.The simulation results indicate that the accuracy of the proposed wind speed forecasting based on wavelet neural network of phase space reconstruction is improved.