低调节性能水库库容小，水库水位对出库过程变化敏感，调洪演算中水库水位频繁触及控制水位的上下边界，造成出库流量过程产生剧烈波动，影响最优解的可操作性。基于水库实时防洪调度的操作流程，以最大化实时调洪库容为准则，提出了腾库预泄段、过渡维持段、削峰蓄洪段、安全消落段、期末消落段五阶段划分原则；以最大削峰准则为目标，构建了各分段优化出库过程的确定方法，形成了低调节性能水库防洪优化调度改进分段试算法。以富春江水库为对象，开展了改进分段试算法模拟调度，并与原分段试算法进行对比。实例结果表明，改进分段试算法与原分段试算法相比优势明显：①克服了原方法调洪计算中频繁突破库容、泄流能力上下限的问题，算法稳定性能良好；②减少了闸门操作频次，增强了优化调度方案的可操作性；③削峰效果更加显著，同等条件下削峰率提升13.3%。

The storage of a reservoir with low regulation performance is often small, so the water level of the reservoir is sensitive to the changes of outflows. Therefore, the reservoir water levels frequently approach the controlled upper and lower boundaries in the flood control operation, which causes severe fluctuations in the reservoir discharge process and affects the operability of the optimal operation solutions. This papers divides a flood into five periods including pre-discharge section, maintenance section, flood storage section, safety drawdown section and final drawdown section, according to the process of reservoir flood control operation and the principle of maximizing the real-time reservoir flood control capacity. Then the improved stage-wise trial-and-error method is proposed to determine the optimal discharges of each section, based on the principle of maximum reduction of the flood peak. The Fuchunjiang reservoir is selected as the study case to demonstrate the effectiveness of the proposed method by comparing the results with the traditional stage-wise trial-and-error method. The results indicate that the improved stage-wise trial-and-error method is better than the traditional one. It overcomes the problem that the upper and lower limits of reservoir storage capacity and discharge capacity are frequently broken in the traditional flood control operation method, which brings a good stability performance to the algorithm. It improves the operability of the optimal operation solutions by reducing the operation frequency of the gate. It also increases the flood peak reduction rate by 13.3% under the same conditions of the traditional method.

国家重点研发计划项目(2017YFC0405606)；国家自然科学(51609062)；中央高校基本科研业务费专项(2018B10514)；中国博士后特别资助项目(2018T110525)