Applying optimization methods for automatic calibration of 3D morphodynamic numerical models of shallow reservoirs: comparison between lozenge- and hexagon-shaped reservoirs. Issue 1 (15th November 2022)

• Record Type: Journal Article
• Authors: Shoarinezhad, Vahid; Wieprecht, Silke; Kantoush, Sameh Ahmed; Haun, Stefan
• Is Part Of: Journal of hydroinformatics. Volume 25:Issue 1(2023)
• Page Start: 85
• Page End: 100
• Publication Date: 2022-11-15
• Subjects: 3D hydro-morphodynamic models -- automatic model calibration -- parameter estimation -- reservoir sedimentation -- shallow reservoirs; Hydrology -- Data processing -- Periodicals; Geographic information systems -- Periodicals; Geographic information systems; Hydrology -- Data processing; Electronic journals; Periodicals; 551.480285
• DOI: 10.2166/hydro.2022.094
• Languages: English
• ISSNs: 1464-7141

Abstract

Abstract: Understanding the complexity of the siltation process and sediment resuspension in shallow reservoirs is vital for maintaining the reservoir functionality and implementing sustainable sediment management strategies. The geometry of reservoirs plays an indispensable role in the appearance of various flow structures inside the basin and, consequently, the pattern of the morphological evolution. In this study, a three-dimensional numerical model, coupled with optimization algorithms, is used to investigate the morphological bed changes in two symmetric shallow reservoirs having hexagon and lozenge shapes. This work aims to evaluate the applicability, efficiency, and accuracy of the automatic calibration routine, which can be a suitable replacement for the time-consuming and subjective method of manual model calibration. In this regard, two sensitive parameters (i.e., roughness height and sediment active layer thickness) are assessed. The goodness-of-fit between the calculated bed levels and the measured topography from physical models are presented by different statistical metrics. From the results, it can be concluded that the automatically calibrated models are in reasonable agreement with the observations. Employing a suitable optimization algorithm, which finds the best possible combination of investigated parameters, can considerably reduce the model calibration time and user intervention. HIGHLIGHTS: The flow structure and sedimentation pattern in symmetric expansions are numerically studied. Local (GML) and a global (BB–BC) optimization algorithms are used to calibrate the numerical models of two symmetric shallow reservoirs. GML outperforms BB–BC considering the convergence speed (efficiency) without trapping in local minima by having the same predicted parameter values (robustness).