Complementary operation optimization of cascade hydropower reservoirs and photovoltaic energy using cooperation search algorithm and conditional generative adversarial networks

The rapid advancement of photovoltaic (PV) and other clean energy technologies has significantly increased their market share within power systems. However, these renewable energy sources are characterized by inherent volatility, intermittency, and unpredictability, which complicate the peak regulation of load demands. This paper presents a novel cascade hydro-solar complementary operation optimization method that leverages uncertain scenario generation to address these challenges. Initially, the conditional boundary equilibrium generative adversarial network model is used to dynamically capture the nonlinear relationships among solar output, irradiance, and solar angle. A clustering algorithm is then used to reduce these scenarios into a subset of representative output scenarios, which are integrated into the hydro-solar complementary operation model. To optimize the operation strategies, the novel cooperation search algorithm is selected as the optimizer. Engineering applications demonstrate that increased solar penetration exacerbates the impact of solar output uncertainty on the power grid. For example, in the spring, when the number of reservoirs is 4, the peak-valley difference of the spring load decreases from 2000.0 MW to 77.0 MW. The proposed method effectively mitigates these uncertainties across various scenarios by reducing peak load demands and enhancing residual load stability. Thus, a viable solution is provided for the complementary operation of cascade hydropower reservoirs and photovoltaic energy systems.