基于滑动四分位和可行搜索圆算法的风速-功率异常数据清洗方法

Wind speed-power abnormal data cleaning method based on the algorithm of sliding quartile and feasible search circle

受气候条件、运行环境等因素影响,风电场SCADA系统所采集的原始风速和风电功率数据常存在大量奇异点,较难反映风电机组真实性能。本研究提出一种基于滑动四分位和可行搜索圆算法的风速-功率异常数据清洗方法。首先,分析了原始数据的时序特征和异常数据分布特点,将数据分为分散型异常数据和堆积型异常数据两类;然后,运用滑动四分位算法实现了分散型异常数据的识别,提出可行搜索圆(FSC)算法,剔除堆积型异常数据,获得符合风电机组出力特性的数据主带;最后,以我国北方某风电机组实际运行数据为例验证,表明本研究方法能较好地识别异常数据,与传统方法相比,本方法清洗效率高、效果好,具有一定的通用性。

Affected by climatic conditions,operating environments and other factors,the original wind speed and power data collected by the SCADA system for wind power farms can hardly reflect the actual performance of the wind turbine for the existence of many singular points.A wind speed-power abnormal data cleaning method based on the algorithm of sliding quartile and feasible search circle was proposed in this paper.Firstly,the timing characteristics of the raw data and the distribution characteristics of the abnormal data were analyzed.The data was divided into two types:scattered anomaly data and stacked anomaly data.Then,the moving quartile algorithm was used to achieve the identification of the decentralized abnormal data.A feasible search circle(FSC)algorithm was further proposed to eliminate the accumulated abnormal data.And the main data band meeting the wind turbine output characteristics was obtained.Finally,the actual operation data of a wind turbine in northern China was taken as an example to verify the method proposed in this paper.The results show that the proposed method can better identify abnormal data.Compared with traditional methods,the proposed algorithm has certain generality for its high efficiency and excellent results in cleaning data.