부산대학교 도서관

With the penetration rate of renewable energy represented by wind power and photovoltaic increasing, the large-scale timing scenarios caused by the uncertainty of their output bring high computational complexity to the optimization analysis of power systems. In this paper, we utilize a Wasserstein distance-based scenario generation method commonly applicable to wind and photovoltaic (PV) power is proposed, which use discretized optimal quantiles to approximate the representation of a continuous probability distribution. The study analysis is based on domestic and foreign data as the research object: the analysis of wind power is based on the time series data of 2015-2019 provided by a Dongshan wind farm in Hubei, China, and the PV is based on a 365-day time series data of 31 photovoltaic power plants provided by NREL integration data sets, which is collected in a region of Washington State, USA. The MMD statistical test was used to compare the proposed method with Monte Carlo Simulation (MCS) for validity testing. The results show that the typical scenario obtained by the proposed method can accurately approximate the original scenarios set, effectively reflect the output characteristics of wind power and photovoltaic power in the region at a certain time, and provide a certain amount of data support for subsequent power system planning and operation optimization.