光伏出力序列场景生成新方法OA北大核心CSTPCD

Aimed at a more accurate medium-and long-term power balance verification,a novel method for generating photovoltaic(PV)output scenarios based on transverse and longitudinal fusion and manually specified continuous ex-treme days is proposed. First,the historical data is divided into several extreme days and several typical days according to the characteristic index of daily irradiance and clustering. Then,based on the transverse irradiance day type se-quence and characteristic index sequence in the next month which are generated using the Markov chain,a method to generate the irradiance day type sequence based on the occurrence probability of day type and the allowable day type conversion matrix is put forward to generate the longitudinal irradiance day type sequence and the characteristic index sequence in the next month,and the transverse and longitudinal data are further fused. Finally,the monthly PV output scenarios are generated based on the optimization method and the irradiance-power conversion model. On this basis,the method for generating PV output sequence scenarios based on manually specified continuous extreme days is also formulated. The results of an example show that the proposed algorithm can provide a better reference for operation scheduling.