光伏电站高比例异常运行数据组合识别方法

Combined Identification Method for High Proportion of Abnormal Operation Datain Photovoltaic Power Station

高质量光伏功率数据是光伏发电和并网研究的基础,但光伏电站的实际运行功率数据中通常含有高比例的异常数据,且异常和正常数据的界限不明显,传统的新能源异常数据识别方法难以应对此类型的异常数据。分析光伏功率连续型异常数据和离散型异常数据的典型特征,提出了一种高比例异常数据的组合识别方法。首先,基于连续型异常数据在时间上具有连续性的特点,采用相似日同时段均值对比算法剔除连续型异常数据;其次,基于离散型异常数据与正常数据相比较为分散的特点,采用四分位法剔除离散型异常数据。算例分析表明,所提方法能够适应具有异常数据和正常数据界限不明显特征的高比例异常数据条件,有效识别连续型异常数据和离散型异常数据,从而大幅提高了辐照度和光伏功率的线性相关程度。

High-quality photovoltaic power data is the basis of the research on photovoltaic power generation and grid integration.However,the actual power operation data in photovoltaic power stations usually contains a high proportion of abnormal data,and the boundary between abnormal and normal data is not obvious.Traditional identification methods for renewable energy abnormal data are difficult to deal with this type of abnormal data.By analyzing typical characteristics of continuous abnormal data and discrete abnormal data of the photovoltaic power,a combined identification method for the high-proportion abnormal data is proposed.Firstly,based on the continuous characteristics of continuous abnormal data in terms of time,the continuous abnormal data is eliminated by the comparison of mean values of similar days at the same time period.Secondly,based on the characteristics of discrete abnormal data being more dispersed than normal data,the quartile method is used to eliminate discrete abnormal data.The example analysis shows that the proposed method can adapt to the condition of high-proportion abnormal data with no obvious boundary between abnormal data and normal data,effectively identify continuous abnormal data and discrete abnormal data,and greatly improve the linear correlation between radiation and photovoltaic power.