• Developed a novel simulation model to characterize the nozzle-switching process. • Proposed a multi-objective optimization framework for the nozzle-switching strategy. • Revealed the influences of key variables on power regulation evaluation metrics. • Provided recommendations for variable selection to enhance nozzle-switching performance. To mitigate the intermittency of wind and solar power generation, the new power system imposes more stringent requirements on hydropower units’ flexible startup capability and robust power regulation performance. Nozzle switching is a key mechanism for power regulation in multi-nozzle Pelton units, enabling rapid response to load fluctuations while maintaining high efficiency. However, existing studies lack dedicated simulation models that explicitly incorporate nozzle switching, failing to trace asynchronous nozzle behavior and quantify the influence of variables on evaluation metrics. In this study, a hydraulic–mechanical coupled simulation model integrating nozzle-switching strategies is developed to characterize the nozzle-switching process. Based on this model, a problem-driven multi-objective optimization framework using the NSGA-Ⅱ algorithm is established, and the dynamic responses of six-nozzle adjusting regulation and nozzle-switching regulation are systematically compared. Results show that the six-nozzle adjusting regulation achieves shorter regulation time but leads to higher water hammer pressure and lower efficiency under partial load. Nozzle-switching regulation improves efficiency and reduces the water hammer pressure, but it causes longer time and power fluctuations due to the switching mechanism. By applying the proposed multi-objective optimization framework, the regulation time is reduced by 20.7%, and the peak-to-peak power fluctuation is decreased by 54.7% compared with the baseline scheme. These findings provide theoretical and practical insights for optimizing 6–4 nozzle-switching strategies in six-nozzle Pelton turbine units operating under grid-connected conditions.
Lin et al. (Sun,) studied this question.