摘要
为应对多风电场超短期预测模型中输入和输出变量众多、变量间的时空关系复杂等问题,提出一种基于独立稀疏堆叠自编码器的多风电场超短期功率预测方法。该方法基于降维编码、特征预测和重构解码相结合的预测框架,首先设计了一种独立稀疏双层堆叠自编码器提取多维风电功率的空间独立特征,并将其作为预测对象分别预测,最后将特征预测的结果重构解码,获得多风电场功率的预测结果。对实际算例的验证结果表明,独立稀疏堆叠自编码器能增强提取特征的可靠性、独立性和合理性,从而有效提高多风电场超短期功率预测的精度和效率。
To deal with the problems such as numerous input and output variables,and complex spatio-temporal rela⁃tionships among variables in the ultra-short-term forecasting model of multiple wind farms,an ultra-short-term power prediction method for multiple wind farms with independent sparse stacked autoencoder(ISSAE)is proposed,which is based on a prediction framework that combines dimensionality reduction coding,feature prediction,and reconstruction decoding.First,an independent sparse two-layer stacked autoencoder is designed to extract the spatial independent fea⁃tures of multi-dimensional wind power,which are used as prediction objects to predict separately.Finally,the results of feature prediction are reconstructed and decoded to obtain the power prediction results of multiple wind farms.The veri⁃fication results of a practical example show that ISSAE can enhance the reliability,independence and rationality of ex⁃tracted features,thereby effectively improving the accuracy and efficiency of ultra-short-term power prediction of multi⁃ple wind farms.
作者
李丹
王奇
杨保华
张远航
LI Dan;WANG Qi;YANG Baohua;ZHANG Yuanhang(College of Electric Engineering and New Energy,China Three Gorges University,Yichang 443002,China;Hubei Provincial Collaborative Innovation Center for New Energy Microgrid,Yichang 443002,China)
出处
《电力系统及其自动化学报》
CSCD
北大核心
2022年第2期23-30,共8页
Proceedings of the CSU-EPSA
基金
国家自然科学基金资助项目(51807109)。
关键词
多风电场
功率预测
堆叠自编码器
稀疏性约束
独立性约束
multiple wind farms
power prediction
stacked autoencoder(SAE)
sparsity constraint
independence constraint