Data-Driven Robust Scheduling of a Wind–PV–CSP Hybrid System Using Adaptive Uncertainty Sets
Constructing wind and solar energy bases is an effective way to promote the green transformation, and the optimal dispatch of renewable energy bases is essential to their high-quality development. However, wind and solar generation are characterized by intermittency, fluctuations, and unpredictability, which pose new challenges to the economic operation of power systems. This paper proposes an enhanced data-driven robust optimization method to establish an economic dispatch model for a new energy base that accounts for uncertainties in wind turbine and photovoltaic power output. It also develops uncertainty intervals for wind turbine and photovoltaic output using a random forest regression method. Compared with traditional robust optimization methods, the proposed method fully utilizes historical data to establish a more precise and flexible uncertain variable interval model, avoiding the overly conservative issues inherent in traditional robust optimization approaches. Lastly, the proposed method is validated in a case study, demonstrating that the data-driven uncertainty set is more in line with the actual situation.