水利水电技术（中英文）

【目的】水风光资源长期预报精度有限，如何应对风光出力和径流的不确定性是充分发挥多能源协同优势的关键，为此，提出考虑预报不确定性的水风光互补系统两阶段决策方法。【方法】将互补系统长期调度多阶段决策划分为包含面临时段和余留期的两阶段决策，通过余留期能量曲面表征余留期不同库容条件与未来风光出力及径流所能产生的发电效益；讨论不同预报水平下水风光资源预报信息的利用方式，提出余留期能量曲面确定方法，并分析预见期利用长度对互补系统效益的影响，确定最佳预见期利用长度；以基于调度图模拟的互补系统调度结果作为对照，评估不同预报水平下两阶段模型的适用性。【结果】选择雅砻江下游水风光互补系统进行实例研究，结果显示：(1)两阶段模型中基于径流序列映射模型的预报信息利用方式能较好地表征余留期效益；(2)随着两阶段决策模型预见期利用长度的增加，预见期的边际效应显著降低，互补系统最佳预见期利用长度为3个月；(3)与调度图相比，两阶段模型推迟水库蓄水和消落时机并加深消落程度，使得发电效益和出力可靠性在预报水平较高时整体高于调度图。【结论】结果表明：两阶段决策模型能够灵活结合多个预见期的预报信息，实现有限预报信息下的互补系统长期优化调度。

[Objective]Long-term prediction accuracy of wind and photovoltaic(PV) output and runoff is limited, and how to deal with the uncertainty of wind output and runoff is the key to give full play to the advantages of multi-energy synergy. Therefore, a two-stage decision-making method of hydro-wind-PV system considering the prediction uncertainty is proposed in this study. [Methods]The multi-stage decision of long-term operation of hybrid energy systems is divided into two stages, namely, current stage and remaining stage. The energy surface of remaining stage is used to represent the power generation benefits generated by different storage capacity conditions in the and the future wind and PV output and runoff. Then, the utilization mode of the forecast information of the wind output, PV output and runoff prediction at different prediction levels is discussed, the method to determine the energy surface of the remaining stage is proposed, and the impact of the utilization length of the prediction period on the benefit of the complementary system is analyzed, so as to determine the optimal utilization length of the prediction period. The adaptability of the two-stage model decision-making under different forecast levels is evaluated by comparing the result of the hybrid system operation based on the operation chart. [Results] A case study was carried out using the hydro-wind-PV hybrid energy system in the Yalong River basin of China, the resutts show that(1) for the two-stage decision-making model, the runoff series can better represent the benefit of the carryover stage.(2) The marginal effect of the forecast period decreases significantly with the increase of the utilization length of the forecast period, and the optimal utilization length of the prediction period is 3 months.(3) Compared with the operation chart, the two-stage decision-making model delays the time of impounding and drawdown and deepens the drawdown, and the generation efficiency and output reliability of the model are higher than the operation chart as a whole when the prediction level is higher. [Conclusion]The study shows that the two-stage decision-making model can flexibly combine the prediction information of multiple prediction period, and realize long-term optimal operation of hybrid energy systems with limited prediction information.