基于大数据驱动的风速不确定性建模

Uncertainty Modeling of Wind Speed Based on Big Data

关于风电场风速和风功率的随机特性,现有研究大多侧重于进行整体分析,而没有考虑信号中高频波动部分和低频稳定部分各自的特性。针对这种情况,本文从气象学中大气运动的湍流流动机理出发,基于大数据分析和数据驱动的建模技术,对风电场风速信号的平均小时风速和湍流部分进行了分析。通过对不同地域的多个风电场历史风速和风功率的海量数据的统计分析,发现了风速波动的内在关系,建立了风速瞬时方差统计参量关于其小时平均值的幂律模型,并验证和初步理论解释了所建立的风电瞬时波动不确定性模型的普适性。

A lot of previous studies have analyzed the random characteristics of wind speed and power in wind farms as a whole; how-ever, they gave little consideration to the respective characters of the high-frequency fluctuant part as well as the low-frequencystable part. Based on this fact, this paper applies the mechanism of atmospheric turbulence in meteorology and analyzes the hourlymean wind speed and the turbulence part in wind speed signal of wind farms with the techniques of big data analysis and data-driv-en modeling. Through the statistical analysis of large quantities of data from historical wind speed in different regions, the inherentrelationship is found in fluctuations of wind speed. And the power-law model is established between instantaneous variance ofwind speed fluctuation component and its hour average component. Universality of the uncertainty model is verified, and prelimi-nary theory explained.