Hydrological and Hydraulic Uncertainty Analysis in Probabilistic Design of Flood Diversion Systems Using NSGAII and Bivariate Frequency Analysis

Risk-based optimization is a framework that allows the designer to involve uncertainties in the decision-making process and determine the reliability of a hydraulic structure. This study was conducted to incorporate hydrological and hydraulic uncertainties in the probabilistic design of Karun-4 diversion system in southwestern Iran. The risk-based multi-objective optimization was developed for determining the effect of uncertainty sources on the characteristics of the diversion system. For this purpose, the time series of annual maximum peak flow and maximum flood volume for a period of 58 years were prepared. Archimedean copula function and non-dominated sorting genetic algorithm were adopted. The optimal values of upstream cofferdam height, downstream cofferdam height and the diameter of the first and second tunnels were estimated as 38, 11, 9.2 and 8.8 m, respectively, all corresponding to 25-year return period. The results suggest that the proposed framework could be valuable for decision makers when economic, hydraulic and hydrological uncertainties are expected.