picture of the institute with institute logo
homeicon uni sucheicon search siteicon sitemap kontakticon contact
unilogo University of Stuttgart
Institute of Hydraulic Engineering

student papers subject

german-icon print view
 
Automated calibration and parameter sensitivity analysis of a numerical sediment transport model in Hydro-FT
tutor
description1. Background:

Sediment transport models are by nature highly parameterized, as these defined by both empirical and physical relationships and parameters for initial and boundary bed topography as well as flow conditions. Sediment transport model calibration can thus be a tedious and highly time-consuming process. Moreover, calibration processes can be subjective, for ex-ample, when objectives are based on visual “best-fit” bed topography outputs. Recently, model-independent calibration tools such as the software PEST are gaining popularity as they can decrease manual effort involved in a calibration process as well as decrease subjectivity. This thesis would aim to automate the calibration process as well as conduct a sensitivity analysis on parameters of a sediment transport model created with Hydro-FT.

The two main goals of this proposed project are therefore: (1) to set up an automated calibra-tion for a numerical model in Hydro-FT based on sediment transport experiments from a 180° channel bend (Yen and Lee, 1995) and (2) to run the automated calibration as well as carry out a sensitivity analysis on sediment transport parameters.

2. Aims of the Project:
  • Review of literature regarding the behaviour and modelling of sediment in curved channels (in particular experiements of Yen and Lee, 1995).
  • Familiarisation with the software Hydro-FT and calibration software PEST.
  • Checking and, if necessary, modifying an existing numerical erosion model of a 180° channel bend in Hydro-FT
  • Defining calibration parameters and calibration objective function
  • Programming of PEST to carry out calibration and sensitivity analysis on Hydro-FT model
  • Interpretation of the results


This Master Thesis will be supervised by Margot Doucet, M.Sc. and Stefan Haun, PhD. Dis-cussions with additional contacts can, however, provide additional information and ideas for the work and are therefore also recommended.

The project can be carried out in German or in English.
contactMargot Patricia Doucet, M.Sc.