短时间尺度用电行为相关性分析网络模型

发布时间：2018-06-19 01:33

本文选题：相关性 + 随机矩阵理论　；参考：《电力系统自动化》2017年03期

【摘要】：用电行为短时间尺度相关关系是用户精细化管理与智能配用电的重要决策基础。建立了针对短时间尺度用电行为相关性的网络化分析方法。首先,由描述短时间尺度用电功率序列间相关性的皮尔逊相关系数构造用电行为相关矩阵,为减小随机用电行为与测量误差等无序噪声信息对相关矩阵的影响,引入随机矩阵理论对相关矩阵进行去噪,并提出利用特征值谱熵的正则化修正方法。然后,基于相关矩阵构建了相关矩阵分阶与同级网络序列,提出了基于社团结构的相关性聚类方法与基于最小树的相关性等级结构挖掘方法;定义了量度用电相关性的拓扑指标,从而构造了用电行为相关性聚类分析与等级分析的网络模型,并就模型的典型应用场景进行了讨论。最后,通过实例数据验证了上述去噪方法、相关性分析网络模型及指标的有效性。
[Abstract]:The short time scale correlation of power consumption behavior is the important decision basis of user fine management and intelligent distribution power consumption. A networked analysis method for the correlation of short-scale electricity consumption behavior is established. Firstly, the Pearson correlation coefficient, which describes the correlation between the short scale power series, is used to construct the electrical behavior correlation matrix. In order to reduce the influence of random noise information, such as random power consumption behavior and measurement error, on the correlation matrix, the correlation matrix is constructed by using Pearson correlation coefficient. The random matrix theory is introduced to Denoise the correlation matrix, and a regularization correction method using eigenvalue spectral entropy is proposed. Then, based on the correlation matrix, the hierarchical and peer network sequences of the correlation matrix are constructed, and the correlation clustering method based on the community structure and the correlation hierarchy mining method based on the minimum tree are proposed. The topological index to measure the correlation of electricity consumption is defined, and the network model of cluster analysis and hierarchical analysis of electricity behavior correlation is constructed, and the typical application scenarios of the model are discussed. Finally, the effectiveness of the above denoising method, the correlation analysis network model and the index are verified by the example data.
【作者单位】：新能源电力系统国家重点实验室(华北电力大学);
【基金】：国家自然科学基金资助项目(51207051) 中央高校基本科研业务费专项资金资助项目(2016XS02)~~
【分类号】：TM73

【相似文献】