Evaluation of uncertainty in unidimensional hydraulic models for mountain rivers: impact of channel geometry and roughness coefficient

This project aims to analyze how the combination of two main sources of uncertainty-the geometry of the channel and the roughness-affect coefficient of the hydraulic variables, such as the speed of the flow and the draft, when using unidimensional hydrodynamic models in mountain rivers. The geometry of the channel, obtained from cross sections taken at specific points, duly spaced and/or in critical locations such as curves and hydraulic structures, can be derived from various sources, such as topographic surveys, satellite images, lidar or drone data. On the other hand, the roughness coefficient is a critical parameter that influences flow resistance, directly affecting the speed and load losses. The glue methodology (generalized Likelihood Uncertainty Estimation) is essential for the calibration of these 1D models, since it allows to capture and quantify the uncertainty inherent in these parameters. This approach not only seeks an optimal adjustment, but also provides a more robust analysis of how variations in geometry and roughness coefficient impact model predictions. Quantifying this uncertainty is essential to make informed decisions in the management of water resources and infrastructure design, ensuring that all possible variations in the real conditions of the system are considered.

21st UNIVERSITY RESEARCH PROJECT COMPETITION