基于聚类分析的风电场风速-出力典型波动过程关联分析

Correlation analysis of typical fluctuation process between wind speed and wind power

在风力发电系统中,研究风速与功率波动的关联特性能够为风电功率预测提供很好的理论基础,对增强电力系统的安全性与稳定性具有重要的参考价值。针对功率并不随着风速波动而实时线性波动的问题,采用K均值聚类算法和模糊C均值聚类算法分别对风速和功率波动过程进行划分,得到5种不同波动过程,然后采用皮尔逊相关性分析法和灰色关联度分析法分析不同波动过程下风速和功率波动的关联特性。不同波动过程下分析结果显示:K均值聚类时最大相关系数为0.4925,最小相关系数为0.3318;模糊C均值聚类时最大相关系数为0.4868,最小相关系数为0.3293;所有波动过程下,灰色关联度均在0.5附近,表明功率波动与风速波动趋势相似。

In wind power generation system,studying the correlation characteristics between wind speed and power fluctuation can provide a good theoretical basis for wind power prediction,and has important reference value for enhancing the security and stability of power system.Aiming at the problem that power does not fluctuate linearly in real time with wind speed fluctuation,this paper has divided the wind speed and power fluctuation process by K-means clustering algorithm and fuzzy C-means clustering algorithm respectively,and obtained five different fluctuation processes.Then,the correlation characteristics of wind speed and power fluctuation under different fluctuation processes are analyzed by Pearson correlation analysis method and grey correlation analysis method.The results show that the maximum correlation coefficient is 0.4925 and the minimum correlation coefficient is 0.3318 in K-means clustering.In fuzzy C-means clustering,the maximum correlation coefficient is 0.4868 and the minimum correlation coefficient is 0.3293.Under all fluctuation processes,the grey correlation degree is around 0.5,indicating that the trend of power fluctuation is similar to that of wind speed fluctuation.