摘要
提出一种基于时空关联特征与贝叶斯-长短期记忆神经网络(bayesian long short-term memory,B-LSTM)模型的分布式光伏功率区间预测方法。以长短期记忆神经网络(long short-term memory,LSTM)为基础构建近似贝叶斯神经网络,建立考虑时空关联特征的B-LSTM模型,利用其强大的记忆能力和特征提取不同特征尺度的模态分量,并进行分布式光伏功率区间预测。以某地区实际分布式光伏数据集进行算例分析,验证了所提方法的优越性。
A distributed photovoltaic(PV)power interval prediction method based on spatio-temporal correlation features and bayesian long short-term memory(B-LSTM)model is proposed.The approximate Bayesian neural network is constructed by adding a Dropout layer based on the LSTM neural network to establish a B-LSTM model considering spatio-temporal correlation features,and its powerful memory and feature extraction capabilities are used to extract deep features for distributed PV power interval prediction for intrinsic mode function components with different feature scales.An arithmetic example is analysed with an actual distributed PV dataset in a region to verify the superiority of the proposed method.
作者
王海军
居蓉蓉
董颖华
WANG Haijun;JU Rongrong;DONG Yinghua(Nanjing Vocational Institute of Railway Technology,Nanjing 210031,China;China Electric Power Research Institute,Nanjing 210003,China)
出处
《中国电力》
CSCD
北大核心
2024年第7期74-80,共7页
Electric Power
基金
江苏省自然科学基金资助项目(BK20210046)
江苏省333项目(500RC33322003、5002023006-1)
南京铁道职业技术学院“青蓝工程”(QLXJ202111)
南京铁道职业技术学院轨道交通基础设施智能检测研究中心科研平台资助项目(KYPT2023003)。
关键词
分布式光伏
时空关联性
区间预测
distributed photovoltaics
spatio-temporal correlation
interval prediction
