Research on Hydro-Photovoltaic Complementary Optimized Scheduling Based on Photovoltaic Forecasting

With the large-scale integration of new energy sources represented by solar power into the grid, how to incorporate the dynamic characteristics of PV power fluctuations into hydropower optimization models is a crucial issue in the optimal scheduling of hydro-PV complementary systems. To fully account for the uncertainty of PV output and effectively enhance the utilization efficiency of solar resources, this paper explores a joint hydro-PV optimal scheduling method that incorporates PV power forecasting. Firstly, the FCM algorithm and CEEMDAN are employed to perform similarity-day clustering and signal decomposition-reconstruction on the original PV power data, respectively, thereby deeply exploring the patterns of PV power fluctuations. Building on this foundation, the CNN-BiGRU-Attention model is utilized to forecast PV power data under three weather scenarios: sunny, rainy, and partly cloudy. Subsequently, a short-term hydro-PV complementary optimal scheduling model is established with the objective of maximizing the power generation of the hydro-PV system. Taking the cascade hydro-PV complementary system in the Shuoqu River Basin as a case study, a comparative analysis is conducted on the forecasted PV power data and the output power of the hydro-PV complementary system under the three aforementioned weather scenarios, exploring the impact of PV integration on the operational modes of cascade hydropower stations in the Shuoqu River Basin. The results indicate that the proposed PV power forecasting model can accurately capture the changing trends of PV data and effectively forecast PV power. Following PV integration, the water levels of cascade hydropower stations rise in the afternoon, and their power output processes and characteristics undergo alterations. The hydro-PV complementary model based on PV power forecasting achieves 100% consumption of solar power with minimal losses in cascade power generation.