基于小波包负荷特征提取和径向基网络的短期负荷预测新方法

A novel method of short-time load forecasting based on wavelet packet feature extracting and radial basis function network

准确的负荷预测是电力系统做出合理调度的重要依据.提出基于小波包能量和神经网络理论的短期负荷预测新方法,将负荷序列进行小波包分解,提取小波包能量作为径向基神经网络负荷序列的输入特征量.大量的预测实例分析表明,所提出的预测方法具有稳定性和准确性.

Accurate load forecasting is the basis of power system dispatching. A novel method short - time load forecasting based on wavelet packet feature extracting and Radial Basis Function （RBF） neural network is proposed in this paper. Load series is decomposed with wavelet packet and the wavelet packet energy is extracted as the input feature vectors of RBF neural network. Results of large numbers of load forecasting eases show that this method is stable and fairly accurate.