In many photovoltaic(PV)power plants

problems of communication errors

equipment failures and PV power curtailment result in high proportion of outliers in measured PV power data

which is difficult for performance analysis of PV power plants and application of power data. Hence

a new identification methodology is proposed. A probabilistic PV power curve model is proposed based on Copula theory to describe the relationship between the measured global radiation and PV power. Based on engineering experiences and the characteristics of high dispersion

strong randomness and a large proportion of outliers

machine identification models are proposed to identify three typical types of outliers. These methods are verified using measured data of PV power plants and artificial data. The effectiveness of applying the outlier identification methods is investigated through a day-ahead PV power forecasting application. Moreover

the proposed machine identification method is more adaptable and have higher accuracy than the conventional 3-sigma method. This work is supported by National Key Technologies R&D Program (No. 2013BAA01B03).