The TKI PRISMA3 project explored innovative approaches to reduce emissions and lower the costs associated with maintenance dredging in ports. TKI Prisma 3 consisted of a strong consortium comprising the Port of Rotterdam, Rijkswaterstaat, Boskalis, Van der Kamp, Van Oord, TU Delft, and Deltares.
The research was organised into four work packages (WPs):
WP1: Investigating the effectivity of tidal- and river dynamics utilization for the transport of dredged sediment towards sea
WP2: Optimize intake and release of sediment using a sediment transport model
WP3: Beneficial reuse of dredged sediment on a large scale
WP4: Data science for more efficient dredging trips
The executive Summary of TKI PRISMA3 can be downloaded by clicking the image below:
All the data gathered within TKI PRISMA3 can be delivered upon written request. Please contact us through secretariaatzks-dsc@deltares.nl.
WP1: Use tidal- and river dynamics for final transport of dredged sediment towards sea
(the final report and corresponding slide deck of WP1 can be downloaded by clicking the images below)
In this work package, we investigated if it is effective to spread dredged sediment in the Nieuwe Waterweg. If this strategy proves to be effective, it could be an alternative for the current practice of spreading (clean) sediment at sea. In turn, this could lead to reduced sailing distances, thereby lowering emissions.
To study the effectivity of this strategy, we conducted a field pilot in a highly energetic part of the Nieuwe Waterweg near Maassluis (Het Scheur), in September 2024. From 9 until 11 September 2024 we spread 9 hopper loads, totalling 14 kTon of sediment (dry matter). To study the natural sediment fluxes and the fate of the sediment spread at this location, an extensive monitoring campaign was carried out, in conjunction with modelling efforts. For the monitoring, we combined long-term monitoring frames, additional short-term monitoring frames and sailing measurements during the days of the release.
The monitoring campaign provided valuable information on the system behaviour of hydrodynamics and sediment concentrations and fluxes. For instance, it showed that the sediment transport of natural background sediment at the release location is ebb dominant, i.e. seaward. The fate of the released sediment was determined from the monitoring, supplemented by simulations with a sediment transport model of the area. Despite the seaward transport direction of the natural background sediment, the majority of the released sediment (>80%) is transported landward, based on monitoring in the first days after the pilot and modelling of 1.5 month after the pilot. The small part that moves seaward mostly ends up in harbour basins, with only a small fraction reaching the open sea. The released sediment stays closer to the river bed compared to the background sediment. Close to the bed the net transport direction is landward because of the presence of a salt wedge. This explains the difference between the transport direction of released sediment and background sediment.
We found that the sediment transport model overpredicts sediment concentrations compared to the measurements, especially near the bed. As a result, the background fraction at the pilot site is also flood-dominant in the model, in contrast to the observations. We therefore recommend utilizing the valuable and unique pilot dataset to further improve the sediment transport model, thereby enabling an improved understanding of sediment fluxes in the Nieuwe Waterweg.
WP2: Optimize intake and release of sediment using sediment transport model
(the final reports and corresponding slide deck of WP2 can be downloaded by clicking the images below)
In this work package, we evaluated two alternative strategies for the spreading of dredged sediment. The first strategy is to transport sediment with a pipeline instead of sailing it out to sea. The second strategy is to spread sediment at other locations than the current location, i.e. the Verdiepte Loswal, and to investigate if the spreading strategy can be optimized by considering release height, tidal phasing and hydrodynamic conditions.
For the pipeline transport, we compared the power required for transporting dredged slurry (mud) through a pipeline compared to sailing with a hopper dredger. For transport lengths of up to several kilometres (approx. 2-3 km, based on the considered conditions), the power required for pumping compared to the sailing and bottom door dumping methods do not deviate substantially. For longer transport distances, transporting the sediment using a vessel is the most energy-efficient method.
For spreading the sediment at different release locations, we utilized a sediment transport model. This was then used to compare the return percentage during the 3-month modelling period, which is the ratio between the sediment mass depositing in the harbour basins and access channels and the total released sediment mass. The effects of release height, tidal phasing, ambient hydrodynamic conditions, and sediment settling velocity were investigated. Seaward locations, especially offshore ones, show a lower return percentage (typically around 10-35%) than landward locations (typically around 80-90%). The return percentage for the inland locations can be reduced by only releasing fine sediments near the surface (i.e. for near-neutrally buoyant active plumes) and during ebb. However, such optimal conditions are difficult to realise in practice. While the return percentage decreases with increasing distance offshore, this also results in longer sailing distances. We recommend using these results to optimize the release location, timing and method of release to find the optimal balance between sailing distance and return percentage, as an appropriate balance must be established between these competing considerations.
WP3: Beneficial reuse of dredged sediment on a large scale
(the final reports and corresponding slide deck of WP3 can be downloaded by clicking the images below)
To reduce its dredging load, the Port of Rotterdam has set a target to re-use at least 1 million m³ of dredged sediment annually. In the third work package, we explored how beneficial use on this large scale may become a reality.
Therefore, several beneficial reuse cases were compared against the current dredging practice. This comparison was based on a decision-making criteria matrix, combining both quantitative and qualitative criteria. Costs, emissions and time (the quantitative criteria) were computed using a logistical model. Safety, sustainability, nature value, environmental quality and bureaucratic difficulties were judged qualitatively. These were weighted and scored based on the input by a selection of relevant stakeholders. Results show that beneficial reuse can compete with existing practices when all relevant aspects are included in the evaluation.
In a parallel investigation, a potential bottleneck for reuse of dredged sediment into a usable solution was addressed: the slow dewatering of mud. Lab experiments demonstrated that marine worms can significantly accelerate the dewatering and ripening of dredged sediment. The most effective species, Tubificidi, reduced dewatering time from over 7 months using passive dewatering to 1 month, and improved final sediment solids content from 0.58 (without worms) to 0.68 (with worms). The initial solids content in the experiment was 0.42. Marine worms drastically speed up dewatering and ripening of fluid dredged mud and the final product has a significantly higher sediment content and thus is more usable. We recommend to explore if this is a viable option for beneficial reuse on a large scale.
WP4: Data science for more efficient dredging trips
(the final reports and corresponding slide deck of WP4 can be downloaded by clicking the images below)
The aim of the fourth work package was to explore if weekly dredging volumes can be forecasted based on relevant hydro-meteo variables through a machine learning approach. This can enable more efficient dredging, thereby resulting in less dredging trips and thus reduced emissions. We compiled dredging volumes from the dredging database for the port area, while hydro-meteo data were taken from publicly available sources such as Waterinfo.
In general, the correlations between the hydro-meteo data and corresponding dredging volumes were found to be rather low, which meant we could not train a functionally working model. This could be due to multiple factors: the first one is limitations imposed by the status of the data. The second reason could be that anthropogenic influences are not taken into account. Influence of ship traffic, the influence of water injection dredging and the man-made decisions during day-to-day dredging operations were not incorporated. Even though the machine learning model did not reach the desired accuracy, the overall research provided valuable insights into the workability of dredging data in combination with metocean and spatial parameters. The research revealed interesting possibilities for follow-up research, such as incorporating variables which quantify influential anthropogenic parameters and exploring other methods to deal with the input metocean data.
