Morphological modelling of Longitudinal Training Walls

Exploring schematisation methods in Delft3D-4 for the development of auxiliary bank channels

Master Thesis (2026)
Author(s)

I.G.A. Hille (TU Delft - Civil Engineering & Geosciences)

Contributor(s)

R.J. Labeur – Graduation committee member (TU Delft - Civil Engineering & Geosciences)

J.A. Arriaga Garcia – Graduation committee member (TU Delft - Civil Engineering & Geosciences)

C.J. Sloff – Graduation committee member (TU Delft - Civil Engineering & Geosciences)

Faculty
Civil Engineering & Geosciences
More Info
expand_more
Publication Year
2026
Language
English
Graduation Date
25-09-2026
Awarding Institution
Delft University of Technology
Programme
Civil Engineering, Hydraulic Engineering
Faculty
Civil Engineering & Geosciences
Page Views
32
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

As ongoing erosion in the Waal is causing multiple challenges for the Netherlands, the Room for the River 2.0 programme is looking at solutions to halt this erosion. One of the proposed solutions uses Longitudinal Training Walls (LTWs) to create a multi-channel system that reduces flow velocities while maintaining sufficient water levels in the main navigation channel. An LTW can narrow the main channel during low flow, but widen the total system during high flow. In 2015, three LTWs were constructed in the Waal as part of a pilot for potential application along the rest of the river. This research focuses on the one near Wamel.

The auxiliary bank channel near Wamel showed rapid sedimentation in its deeper sections during the first years. This appears to be linked to an initial response and bank erosion, rather than indicating an unstable system progressively filling up. After the inlet sill was heightened, the channel shifted to erosion over large parts of its reach. The remaining deposition along the stretch where the bank retreated seems to be slowing down as the bank stabilises.

Against the steep slope of the inlet sill and wall, gravity redirects near-bed flow and sediment back downslope towards the deeper main channel, forming a vortex with flow spiralling against the slope, reducing sediment transport into the auxiliary channel. Large-scale two-dimensional models are not able to explicitly resolve this local vertical vortex that plays a key part in the sediment distribution between the main and auxiliary channel and therefore in the stability of the LTW system.

Research into LTWs has grown considerably, but questions remain regarding the morphological stability of the auxiliary bank channel and how to schematise LTWs reliably in the large-scale coarse grids typically used for long-term river morphodynamic modelling. This research investigated different schematisation methods for LTWs in the coarse grid of the DVR model using ten-year simulations of the Waal from 2016 to 2025. A refined model, based on the same formulations and parameters as the DVR model, was developed and used as a benchmark to evaluate the coarse-grid schematisation variants. Representing the wall and inlet sill solely as a 2D weir with a small increase in bed level, as currently done in the DVR model's Baseline schematisation, proved insufficient, since this does not form a physical barrier able to block and redirect sediment once overtopped.
Incorporating the structure directly into the bathymetry instead allows the model's transverse bed slope adjustment to act on the slope of the structure, activating the deflection needed to redirect sediment back into the main channel.

Even with the structure incorporated into the bathymetry, the coarse model struggled to reproduce the lower transport rates into the auxiliary bank channel seen in the refined model, for two reasons. First, the wide grid cells fail to represent the steep slope of the wall and inlet sill, weakening the gravitational deflection. Second, the structure was placed too far into the main channel, misaligning with the upstream bank, allowing longitudinal inflow that degraded the side-weir entrance geometry, and further increased the sediment transport into the auxiliary bank channel.

Two ways of compensating for the missing slope were tested: increasing the strength of the transverse bed slope deflection based on the Koch & Flokstra formulation, and locally reducing the transport magnitude through a calibration factor in the Meyer-Peter–Müller transport formula. However, both resulted in further problems in the main channel. The increased Koch & Flokstra deflection required to match the transport rates observed in the refined model or to correct the missing factor of local steep slopes unrealistically flattened the point bars in the main channel bends. Reducing the transport magnitude instead of the direction resulted in deposition in front of the inlet and an increase in transport rates again over time.

This research recommends incorporating LTW structures directly into the bathymetry, with careful attention to grid alignment relative to the crest and the effect of misrepresenting steep slopes. Local grid refinement, or a sub-grid formulation for the sediment deflection along a steep-sloped longitudinal structure, offers a future solution for representing the local geometry and the sediment distribution between the two channels.As ongoing erosion in the Waal is causing multiple challenges for the Netherlands, the Room for the River 2.0 programme is looking at solutions to halt this erosion. One of the proposed solutions uses Longitudinal Training Walls (LTWs) to create a multi-channel system that reduces flow velocities while maintaining sufficient water levels in the main navigation channel. An LTW can narrow the main channel during low flow, but widen the total system during high flow. In 2015, three LTWs were constructed in the Waal as part of a pilot for potential application along the rest of the river. This research focuses on the one near Wamel.

The auxiliary bank channel near Wamel showed rapid sedimentation in its deeper sections during the first years. This appears to be linked to an initial response and bank erosion, rather than indicating an unstable system progressively filling up. After the inlet sill was heightened, the channel shifted to erosion over large parts of its reach. The remaining deposition along the stretch where the bank retreated seems to be slowing down as the bank stabilises.

Against the steep slope of the inlet sill and wall, gravity redirects near-bed flow and sediment back downslope towards the deeper main channel, forming a vortex with flow spiralling against the slope, reducing sediment transport into the auxiliary channel. Large-scale two-dimensional models are not able to explicitly resolve this local vertical vortex that plays a key part in the sediment distribution between the main and auxiliary channel and therefore in the stability of the LTW system.

Research into LTWs has grown considerably, but questions remain regarding the morphological stability of the auxiliary bank channel and how to schematise LTWs reliably in the large-scale coarse grids typically used for long-term river morphodynamic modelling. This research investigated different schematisation methods for LTWs in the coarse grid of the DVR model using ten-year simulations of the Waal from 2016 to 2025. A refined model, based on the same formulations and parameters as the DVR model, was developed and used as a benchmark to evaluate the coarse-grid schematisation variants. Representing the wall and inlet sill solely as a 2D weir with a small increase in bed level, as currently done in the DVR model's Baseline schematisation, proved insufficient, since this does not form a physical barrier able to block and redirect sediment once overtopped.
Incorporating the structure directly into the bathymetry instead allows the model's transverse bed slope adjustment to act on the slope of the structure, activating the deflection needed to redirect sediment back into the main channel.

Even with the structure incorporated into the bathymetry, the coarse model struggled to reproduce the lower transport rates into the auxiliary bank channel seen in the refined model, for two reasons. First, the wide grid cells fail to represent the steep slope of the wall and inlet sill, weakening the gravitational deflection. Second, the structure was placed too far into the main channel, misaligning with the upstream bank, allowing longitudinal inflow that degraded the side-weir entrance geometry, and further increased the sediment transport into the auxiliary bank channel.

Two ways of compensating for the missing slope were tested: increasing the strength of the transverse bed slope deflection based on the Koch & Flokstra formulation, and locally reducing the transport magnitude through a calibration factor in the Meyer-Peter–Müller transport formula. However, both resulted in further problems in the main channel. The increased Koch & Flokstra deflection required to match the transport rates observed in the refined model or to correct the missing factor of local steep slopes unrealistically flattened the point bars in the main channel bends. Reducing the transport magnitude instead of the direction resulted in deposition in front of the inlet and an increase in transport rates again over time.

This research recommends incorporating LTW structures directly into the bathymetry, with careful attention to grid alignment relative to the crest and the effect of misrepresenting steep slopes. Local grid refinement, or a sub-grid formulation for the sediment deflection along a steep-sloped longitudinal structure, offers a future solution for representing the local geometry and the sediment distribution between the two channels.

Files

Hille26_MSc_Thesis_LTWs.pdf
(pdf | 37.4 Mb)
License info not available