摘要
为了提高风电功率预测精度,降低电网调度的难度,通过对影响风电功率预测的诸多因素如风速、风向、风电功率、温度等进行分析,进而对风电功率的预测方法进行研究和探讨,提出了基于多变量的小波-神经网络模型的短期风电功率预测方法。通过单变量和多变量的风电功率预测的比较研究,证明BP神经网络预测与小波-神经网络预测这两种方法的预测精度不同。而且,对于同一种方法,输入变量的多少也对预测精度产生影响。通过最终的比较研究得出,采用基于多变量输入的小波-神经网络开展风电功率预测可提高预测精度。
In order to improve the accuracy of wind power prediction and reduce the difficulty of power grid dispatching,this article analyzes the factors that influence wind power prediction such as wind speed,wind direction,wind power and temperature,studies the method of wind power prediction, and proposes short-term wind power prediction method based on multiple variable inputs wavelet-neural network.By comparing the single variable input with multiple variable inputs,it demonstrates that prediction accuracy between BP neural network and wavelet-neural network is different.And,in the same method,the number of input variables also impact on the prediction accuracy.Through the final comparative study,the conclusion is that to predict wind electricity power based on multiple variable inputs wavelet-neural network can improve prediction accuracy.
出处
《电力学报》
2011年第6期458-461,465,共5页
Journal of Electric Power
