%0 Journal Article %T 计及空间关联冗余的节点负荷预测方法<br>Nodal load forecasting method considering spatial correlation and redundancy %A 韩学山 %A 王俊雄 %A 孙东磊 %A 李文博 %A 张心怡 %A 韦志清< %A br> %A HAN Xueshan %A WANG Junxiong %A SUN Donglei %A LI Wenbo %A ZHANG Xinyi %A WEI Zhiqing %J 山东大学学报（工学版） %D 2017 %R 10.6040/j.issn.1672-3961.0.2017.530 %X 摘要： 针对现有节点负荷预测方法对节点空间关联信息没有有效利用的问题,提出计及空间关联冗余的带有估计校正特性的节点负荷预测方法。分析量测信息在时间维度和空间维度上的关联关系,以及二者有机结合的空间关联、冗余的特性,给出二者相互校正的预测原理。对状态量和量测值之间的两种空间关联关系进行深入分析,充分寻求在空间关联拓扑上能间接表征状态量特征的多组量测值方程。依据分析结果建立计及空间关联冗余的预测模型,给出以支持向量机为前期预测模型的预测方法,并对预测方法的优点进行分析。试验结果表明:考虑空间关联冗余的节点负荷预测方法相对于支持向量机模型预测误差明显降低,有利于改善预测结果。<br>Abstract: Aiming at the problem that existing nodal load forecasting methods had no effective use for the nodes spatial correlation information, a new nodal load forecasting method with estimated correction characteristic was proposed, which had considered spatial correlation and redundancy. The correlation between the time dimension and the spatial dimension of the measurement information, and the spatial correlation and redundancy characteristics which combined these two dimensions were analyzed, and the mutual correction prediction principle was given. Two spatial correlations between state and measured values were analyzed deeply to establish measuring equations, which could characterise state features indirectly on the spatial correlation topology. Based on the analysis results, the forecasting model was established, and the forecasting method in which pre-prediction model was support vector machine was given, and advantages of the forecasting method were elaborated. Case studies demonstrated that compared with SVM model, the proposed 山 东 大 学 学 报 (工 学 版)第47卷 - 第6期韩学山,等:计及空间关联冗余的节点负荷预测方法 \=-method could effectively decrease forecasting errors and improve forecasting results %K 节点负荷预测 %K 冗余信息 %K 状态估计 %K 空间关联 %K 支持向量机 %K < %K br> %K support vector machine %K redundant information %K spatial correlation %K state estimation %K nodal load forecasting %U http://gxbwk.njournal.sdu.edu.cn/CN/10.6040/j.issn.1672-3961.0.2017.530