摘要
提出一种计及误差反馈的短期风电功率预测分析方法。首先构建风电监测数据特征优选策略,将原始输入空间映射到低维表达空间,避免维数灾难;引入随机梯度下降法和误差权值调整方法,建立变权随机森林的预测模型,对预测结果进行误差反馈,实时更新误差权值,提高预测准确率。结合某风场的实际运行数据进行仿真分析,验证所提方法的有效性。
A novel short-term wind power forecasting method considering error feedback is proposed.First,the optimal extraction strategy based on the characteristics of wind power monitoring data is constructed to map the original input space into a low dimensional expression space to avoid the dimension disasters.The stochastic gradient descent method and error weight adjustment method are introduced to establish a prediction model of variable-weight random forests.Error feedback is performed to update the error weights in real time and improve the accuracy of prediction.The effectiveness of the proposed method is verified by the simulation and analysis of the actual operating data of wind turbines.
作者
孙泽贤
孙鹤旭
Sun Zexain;Sun Hexu(School of Control Science and Engineering,Hebei University of Technology,Tianjin 300130,China)
出处
《太阳能学报》
EI
CAS
CSCD
北大核心
2020年第8期281-287,共7页
Acta Energiae Solaris Sinica
基金
河北省科技计划(17214304D)
天津市科技支撑项目(14ZCDZGX00818)。
关键词
风电
特征选取
权重评估
随机森林
功率预测
wind power
feature extraction
weight measurement
random forest
power forecast
