Abstract:According to distributed photovoltaic system sick running and defraud state subsidies behavior, an anomaly detection algorithm for distributed photovoltaic generation is proposed, based on the outlier analysis based on Grubbs criterion and the characteristics of the distributed photovoltaic system. The algorithm can filter out the users, whose power generation of the unit capacity per day is too high or too low, so that distributed photovoltaic system long-term sick run can be avoided and defraud state subsidies behavior can be curbed. Finally based on the data of the distributed photovoltaic systems in a county, an example is analyzed. And the results show that the algorithm is effective.