摘要：【中文】本发明涉及一种基于曲线形态分析的光伏窃电发现方法和系统，其中方法包括：设置正常用户集合的聚类中心、故障用户集合的聚类中心和窃电用户集合的聚类中心；计算各个聚类中心的方差；计算出力上升段实际曲线与拟合曲线的方差与各个聚类中心的方差之间的距离，对用户进行第一次归类；计算出力下降段实际曲线与拟合曲线的方差与各个聚类中心的方差之间的距离，对用户进行第二次归类；判断第二次归类后留在待考察用户集合中的用户的光伏设备的参数是否超过阈值，进行第三次归类。本发明设计了光伏窃电行为的发现算法，且充分挖掘了光伏曲线的数据信息，对设备和用户的其它信息依赖较少，具有较高的普适性。 【EN】The invention relates to a photovoltaic electricity stealing discovery method and a photovoltaic electricity stealing discovery system based on curve form analysis, wherein the method comprises the following steps: setting a clustering center of a normal user set, a clustering center of a fault user set and a clustering center of a power stealing user set; calculating the variance of each clustering center; calculating the distance between the variance of the actual curve and the fitting curve of the output ascending section and the variance of each clustering center, and classifying the users for the first time; calculating the distance between the variance of the actual curve and the fitting curve of the output descending section and the variance of each clustering center, and classifying the users for the second time; and judging whether the parameters of the photovoltaic equipment of the users left in the user set to be examined after the second classification exceed the threshold value or not, and performing the third classification. The invention designs a discovery algorithm of photovoltaic electricity stealing behavior, fully excavates data information of a photovoltaic curve, has less dependence on other information of equipment and users, and has higher universality.