Multi-scale Data Assimilation Method for Surface Temperature Field based on Measured Data of Photovoltaic Power Station

Data assimilation, a key technology for numerical weather forecast, has a very important influence on the forecast accuracy of photovoltaic (PV) generated power. In this paper, a multi-scale data assimilation method for surface temperature field based on measured data of PV power station is proposed. In this method, a multi-scale nested grid of temperature field is established based on DBSCAN clustering. Then, optimal statistical interpolation method is used to make data assimilation with different resolution of the multi-scale nested grid, in order to obtain the multi-scale temperature analysis field. Through the case analysis, this method makes full use of the operating data of PV power plants. It not only further improves the forecast accuracy of PV generated power, but also reduces the amount of calculation effectively, which verifies the feasibility of this method.