基于类3σ准则的光伏功率异常数据识别

Identification of abnormal data of photovoltaic power based on class 3σ

在光伏电站实际运行过程中,通信测量设备故障、人为因素均会造成测量数据异常。为了有效地识别和处理异常数据,文章利用混合t Location-Scale分布模型中模型参数易于调整的特征,建立了类似于正态分布3σ准则模型的类3σ准则模型,并利用该模型对光伏功率异常数据进行识别。文章对两个光伏电站的实测数据进行分析,研究结果表明,与正态分布3σ准则模型相比,类3σ准则模型具有正确识别率高、适用性好等优点。

In the actual operation of photovoltaic power plant, communication measurement equipment failure or human factors may lead to abnormal measurement data. In order to identify and process the abnormal data effectively, this paper use the model parameters’ s flexible features of the mixed t Location-Scale distribution model to establish a class 3σ criterion which is similar to the 3σcriterion of normal distribution, which can be used to identify the abnormal data of photovoltaic power.Taking the measured data of two photovoltaic power plants as an example, the class 3σ criterion has the advantages of high recognition rate and good applicability compared with the 3σ criterion of normal distribution.