小波-RBF短期风电功率预测

Wavelet decomposition and RBF neural network used in short-term wind power prediction

鉴于风功率预测是风电并网的关键环节之一,风力发电具有波动性、间歇性、随机性特点,首先利用小波变换对历史风功率数据进行分频段分析,然后根据风功率数据高低频的特点分别利用径向基神经网络建立预测模型,最后通过小波重构获得预测信号.通过算例分析,验证了该预测方法具有较高的准确性和实用性.

Wind power prediction is one of the important factory in wind power grid. Wind power has the features of being unstable, intermittent and random. This paper firstly analyzes the wind power history data with wavelet transform from high and low-frequency , then builds the prediction mode with RBF neural network according to the characteristics of high and low-frequency data, and finally obtain prediction signal by wavelet reconstruction. The analysis proves that the prediction method has higher accuracy and usefulness.