摘要
多风电场相关性研究对于准确把握风力发电出力变化规律,进行风电场出力预测及时间序列建模具有重要意义,因此提出一种基于波动划分和时移技术的多风电场出力相关性分析方法。首先,通过风电场时间序列散点图分析和回归分析得到多风电场整体相关特性;然后,通过波动划分和波动配对算法,提取按波动过程划分的风电场出力序列局部特征并对多风电场波动进行配对;最后,采用基于Pearson相关系数的时移技术和格兰杰因果检验得到最优时移量和时移方向。通过具体算例分析可以证明,所提方法一方面可以精确提取风电场出力波动局部特征,进而更准确地描述风场出力相关特性,另一方面可得到大、中波动对应波动对的最优时移量,其可作为多风电场时间序列建模的约束条件,提高多风电场出力建模精度。
Correlation analysis on multi-wind farms power is significant for grasping the characteristic of wind power, and predicting of wind farm output and its modeling in time series. Thus the correlation analysis method based on fluctuation classification and time shifting technology is proposed. The overall trend of multi-wind farms power is ob- tained by visualizing the time series in scatter diagrams and regressing analysis. Then, the partial characteristics of wind power output classified by fluctuation process are extracted and the power fluctuations of multiple wind farms are paired up by fluctuation classification and fluctuation pair algorithm. The optimal shifting offset and direction are respectively obtained through time shifting technique based on Pearson correlation coefficients and Granger causality test. Case study results verify that the proposed method can extract the partial characteristics of wind power fluctua- tions and describe the correlation of wind power more precisely, it can also obtain the optimal shifting offset of large- and middle-scale fluctuation pairs can also be obtained, which can be taken as the restraints for modeling of multi- wind farms power series to improve the modeling accuracy.
作者
黄越辉
张鹏
李驰
礼晓飞
卫文婷
HUANG Yuehui1 ,ZHANG Peng2 ,LI Chi1 ,LI Xiaofei1 ,WEI Wenting2(1. State Key Laboratory of Operation and Control of Renewable Energy ＆ Storage Systems, China Electric Power Research Institute, Beijing 100192, China; 2. School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, Chin)
出处
《电力自动化设备》
EI
CSCD
北大核心
2018年第4期162-168,共7页
Electric Power Automation Equipment
基金
国家电网公司科技项目(XT71-15-001)~~
关键词
风电
多风电场
相关性
波动划分
时移技术
wind power
multiple wind farms
correlation
fluctuation classification
time shifting
