考虑气温因素的负荷特性统计指标关联特征数据挖掘

发布时间：2018-03-17 13:34

本文选题：大数据　切入点：负荷特性统计指标　出处：《中国电机工程学报》2015年01期 　论文类型：期刊论文

【摘要】：在电力系统负荷特性统计指标和气温日益积累大数据背景下,有效提取数据之间关联特征对电力系统规划和运行具有重大意义。为此,提出一种气温对负荷特性指标影响及其内在关联特征数据挖掘的方法。考虑气温季节特征进行分季度建模,首先通过物理关系和皮尔森相关系数获得气温和负荷特性指标任意两因素之间的相关性特征;然后在多变量时间序列平稳性检验基础上,对水平不平稳的同阶单整时间序列进行协整检验和向量误差修正(vector error correction,VEC)建模以获取其长期同步运动趋势及短期波动特性;进一步通过对变量差分化后的平稳时间序列的向量自回归(vector auto-regression,VAR)建模提取多因素变化量间的动态关系,结合格兰杰因果检验挖掘因素变化量之间的因果引导关系。针对华中某省级电网2006年至2010年负荷特性实际统计数据及相应气温数据的实例分析验证了文中方法的正确性和有效性,方法已在实际电网负荷特性统计分析中得到应用。
[Abstract]:Under the background of power system load characteristic statistic index and temperature accumulating day by day big data, it is of great significance for power system planning and operation to extract the correlation feature between data effectively. In this paper, a method of data mining for the influence of temperature on load characteristic index and its inherent correlation feature is put forward, and the seasonal characteristics of temperature are taken into account in quarterly modeling. Firstly, the correlation characteristics between any two factors of temperature and load characteristics are obtained by physical relation and Pearson correlation coefficient, and then on the basis of multivariable time series stationary test, The cointegration test and vector error correction vector error correction (VECs) model are used to obtain the long-term synchronous motion trend and the short-term fluctuation characteristics of the same order unsteady monolithic time series. Furthermore, the dynamic relationship between the variables is extracted by modeling the vector autoregressive vector auto-regressive VARs of the stationary time series after variable difference. Combined with Granger causality test to excavate the causality leading relationship between factors, the method is verified by analyzing the actual statistical data of load characteristics and corresponding temperature data of a provincial power network in central China from 2006 to 2010. Correctness and validity, The method has been applied to the statistical analysis of the load characteristics of the actual power network.
【作者单位】：智能电网运行与控制湖南省重点实验室(长沙理工大学);国家电网公司;国家电网公司信息通信分公司;国网经济技术研究院;
【基金】：国家自然科学基金项目(51277015) 国家电网公司科技项目([2012]515)~~
【分类号】：TM714;TP311.13

【相似文献】