Deriving hydropower reservoir operation policy using data-driven artificial intelligence model based on pattern recognition and metaheuristic optimizer

Abstract

Robust reservoir operation policies are crucial in ensuring the effective utilization of water resources. However, owing to multiple complicated and changeable factors in practice, it is difficult for standalone approaches to derive reasonable reservoir operation policy. To address the practical requirement, this research proposes a novel artificial intelligence method for deriving reservoir operation policy. The proposed method uses the fuzzy clustering iteration method to identify multiple typical operation patterns from the influencing factors; secondary, for all the samples within each pattern, the novel twin support vector regression (TSVR) is utilized to model the nonlinear mapping relationship between the influence inputs and the target outputs, while the emerging equilibrium optimizer is chosen to determine suitable computation parameters for the TSVR model. The feasibility of the proposed method is fully evaluated on two real-world huge hydropower reservoirs in China. The simulations demonstrate that the developed method can yield better comprehensive benefits than several control methods in deriving reservoir operation policy under uncertain environments. Hence, the experiments confirm that metaheuristic algorithms and pattern recognition techniques can enhance the performance of a standalone artificial intelligence methods in deriving reservoir operation policy.