【摘要】 为研究梯级水库连溃的风险,模拟分析洪水演进过程,通过构建超标准洪水、管涌作用下双库连溃的贝叶斯网络模型,并综合Breach溃坝模型、HEC-RAS二维水动力学模型的溃坝和洪水演进模拟方法,以寒葱沟水库和定国山水库为例,推求水库大坝溃坝概率、评估连溃风险和模拟洪水演进过程。计算结果表明:由已知定国山水库大坝溃坝通过贝叶斯逆向推理可知"寒葱沟水库溃坝"的发生概率从3%增长到87%;由已知寒葱沟水库溃坝通过贝叶斯逆向推理可知"寒葱沟水库管涌"的概率从16%增长到87%。超标准洪水工况和管涌工况下,因两水库连溃,洪水抵达CS1—CS6断面时间在0.8～3.0 h; 1 000年一遇洪水工况下,寒葱沟水库敞泄但未溃坝,仅定国山水库漫顶溃坝,洪水抵达CS1—CS6断面时间在21.5～26.5 h。研究可为降低梯级水库大坝风险、制定针对性的应急预案和防灾减灾提供科学的决策依据。更多还原

【Abstract】 To study the risk of sequential dam break of cascade reservoirs, and to simulate and analyze the flood routing process, by constructing a Bayesian network model for the sequential dam break of double reservoirs under the action of extra standard floods and piping, and combining the dam break and flood routing simulation of Breach dam-break numerical model and two-dimensional HEC-RAS hydrodynamic model, this paper calculates the probability of dam break, evaluates the risk of sequential dam break, and simulates the flood routing process, taking Hanconggou Reservoir and Dingguoshan Reservoir as examples.The calculation results show that: Through Bayesian-based backward inference of the known dam break of Dingguoshan Reservoir, the probability of “Hancongou Reservoir dam break” increased from 3% to 87%;while through Bayesian-based backward inference of the known dam break of Hancongou Reservoir, the probability of “Hancongou Reservoir piping” increased from 16% to 87%.Under extra standard flood and piping conditions, due to the sequential dam break of the two reservoirs, the time for the flood to reach the section CS1—CS6 is 0.8～3.0 h; under the flood conditions that occurred once in 1 000 years, the Hancongou Reservoir was opened for flood discharge but the dam was not broken.Only the Dingguoshan Reservoir was over-topped with dam break, and the flood reached the section CS1—CS6 in 21.5～26.5 h.This study can provide scientific reference for reducing dam break risk, and making emergency plan and disaster prevention & mitigation plan.更多还原