摘要
风能作为一种廉价且安全的清洁能源,在世界范围内愈来愈受到重视.由于风力发电过程中存在的波动性和间歇性问题对电力系统的稳定运行构成严重威胁,因此风电功率的准确预测成为当代电力系统研究的重要课题.针对当前风电场风电功率预测精度较低的现状,提出一种遗传算法优化的小波-BP神经网络风电场发电功率短期预测方法.首先,用Weibull分布函数拟合风速的概率分布,选取了概率分布相似的数据进行训练和预测.其次,将风速序列进行三层尺度的小波分解,将风速的细节信号和近似信号作为BP神经网络的输入,风电功率作为输出,用遗传算法优化小波-BP神经网络的初始权值和阀值进行预测.最后,将预测结果跟采用BP神经网络算法得到的结果进行对比,前者具有更高的精度,绝对预测误差APE最高降低44%,解决了风电场风电功率预测精度较低的问题.
As a cheap,safe and clean energy,wind power has gained an increasing attention worldwide.Since volatility and intermittency existing in wind power generation forms a serious threat to the stable operation of the entire power system,it has been an important subject in the research of power system to make an accurate prediction for electric power of wind farm.To solve the problem that current methods have unsatisfactory accuracy in wind power prediction,a prediction model based on wavelet transform and BP neural network optimized by Genetic algorithm is proposed.The following steps are adopted:Fitting the probability distribution of wind speed utilizing Weibull distribution function,selecting the wind data which has similar probability distribution to train and predict.Then decomposing the selected wind speed series into three levels using wavelet transform,adapting the details and approximation as the inputs of the BP neural network,the wind power series as the output and optimizing the weights and bias of the BP neural network based on genetic algorithm to predict.Finally,compared with the results of the BP neural network,the wavelet-BP neural network optimized by Genetic algorithm has a higher accuracy and reduces the absoloute predition errors by 44%mostly,which overcomes the problem of low accuracy existing in wind power prediction.
作者
杨世秦
孙驷洲
YANG Shiqin;SUN Sizhou(College of Electrical Engineering,Anhui Polytechnic University,Wuhu 241000,China)
出处
《安徽工程大学学报》
CAS
2018年第1期27-33,46,共8页
Journal of Anhui Polytechnic University
基金
安徽省级大学生训练基金资助项目(20161036300164)
