Comparison of river hydrodynamic models in a quasi two-dimensional approach
Villazón Gómez, M.F. (2005). Comparison of river hydrodynamic models in a quasi two-dimensional approach. MSc Thesis. Katholieke Universiteit Leuven (KUL)/Vrije Universiteit Brussel: Leuven. x, 65, XXVII pp. |
Available in | Author | | Document type: Dissertation
|
Abstract | A mathematical river flood model is a useful tool. It can be applied to manage water resources and to estimate and evaluate damages caused by flood events. The decision in choosing a suitable model that interprets faithfully the natural development of the water resources is becoming more difficult in the recent years. In the present research a comparison between two river hydrodynamic models in a quasi two-dimensional approach is conduced.
The area of study is part of the Flemish basin of the river Dender in Belgium: the Overboelare region, from the confluence between the river Dender and the main tributary Mark till the control structure at Geraardsbergen, over a distance of approximately 5 km. The model is implemented by using the Mike11 river modeling software developed by the Danish Hydraulic Institute (DHI) and InfoWorks-RS made by Halcrow and HR Wallingford.
In order to estimate the development of the water inside the floodplains and the flooded area and therefore the flooded volume, a quasi two-dimensional approach is implemented, in which the floodplains are modelled as a network of fictitious river branches and spills with the rivers. Another approach considered in the comparison is the implementation of the floodplains in Info Works as storage areas through flood cells.
The hydrodynamic quasi two-dimensional models are built up in a very detailed way with an accurate database that is available from a complete river flood modeling study for the Dender basin and a model already implemented in Mike 11. After reduction of this Mike 11 model to the Overboelare region, conversion to an InfoWorks-RS model, improvement of the consistency between both models, calibration and validation, different simulations were conducted, for three historical flood events and nine synthetical events with return periods of 1,2,5, 10,25,50, 100,500 and 1000 years.
Evaluation and comparison of the flooded volumes and areas under the different approaches in both packages is presented. This required the flood results to be mapped in GIS. It is concluded from this comparison that: the model simulation or run times for Mike11 were much smaller and the stability of Mike11 is less sensitive to specified initial conditions, and thanks to the linear relation between storage and flow through the spills for peak values, InfoWorks allows the use of a simply relation (hypsographic curve) to predict the water level inside the floodplain. |
|