Robust multiobjective reservoir operation and risk decision-making model for real-time flood control coping with forecast uncertainty
Xin Huang,Bin Xu,Ping-an Zhong,Hongyi Yao,Hao Yue,Feilin Zhu,Qingwen Lu,Yu Sun,Ran Mo,Zhen Li,Weifeng Liu
DOI: https://doi.org/10.1016/j.jhydrol.2021.127334
IF: 6.4
2022-02-01
Journal of Hydrology
Abstract:Forecast uncertainty in real-time flood forecasting is the primary source of risk regarding reservoir flood control operation, affecting both reliability and safety of the entire system. Informing reservoir operation for real-time flood control coping with forecast uncertainty encourages overall robust control of risk probability, vulnerability, and resilience of upstream and downstream protection areas. Current stochastic programming models or chance constrained programming models (CCP) focuses on reducing either risk loss or vulnerability, thereby failing to ensure systematic control. This study established a robust multiobjective operation and risk decision-making model to inform reservoir operations. The proposed model includes three modules. First, a copula function was applied to capture multidimensional temporal dependences of forecast uncertainties, following which stochastic inflow scenarios were generated. Second, a robust multiobjective optimization (RMOO) model was established with the objectives of minimizing the probability of risk events, Conditional Value-at-Risk of risk loss (vulnerability) and the maximum sequential duration for both upstream and downstream areas. Then, an adaptive reference multiobjective evolutionary algorithm was used to solve the model and obtain noninferior solutions. Finally, the TOPSIS multicriteria decision-making method was introduced to select the compromise solution. Methodologies were verified by application to Xianghongdian reservoir, China. The main findings were as follows. (1) The copula function effectively captured the multidimensional positive temporal dependences of forecast uncertainties. (2) RMOO provided more widely distributed noninferior solutions than CCP, and performed better in lowering vulnerability of the system. (3) RMOO reduced the maximum release by 135.6 m3/s (7.33%) by increasing prerelease before the occurrence of flood peak. The proposed methodologies provide insights for supporting robust control of real-time operations of a reservoir system through simultaneous control of flood risk probability, vulnerability, and resilience.
geosciences, multidisciplinary,water resources,engineering, civil